rust 9*#rustc 1.98.1 (48a229cea 2026-09-01)RL.SAwjL-df1bddb45adbe94bm6GÂn62-64f5f36fb09276941 Xr*-1ad4f876c768bda9<%Z>64V`-2d2ab9fb0fe1def5G g{(3iF-06449a08aaab5e30rustc_std_workspace_core{`"Ei%G-cb0d4d77988d81b5")pE͗E-3ec41d3a5a80a4a4 miniz_oxide8bž LƨJ-e44eaef1807e5fc0adler29lyA6P/D$UB-dca52f01a59c33e7 hashbrownX {e-143c77bb957d9940rustc_std_workspace_alloc _z%-145a318e0772276c std_detectK(0*8D?-f99feda5ce9e492arustc_demangle%jerV-cfd0ab54bb33760ecfg_if}λ՞Co-360da4ba13154744 addr2linecY\G--09755979c57b7e64gimliQ`Dy-9a1fede77fbbf2f0objectF λv.z-7ff85ea729a864ffmemchr]1<=UO-0e3477fb2664a277 jlToKG-549498e8b2e4d924 regex_syntaxg (7"ň-8d0720223fc09a43tests, $ʡ  onepassOwnedDFAw wnew_manyy{ always_match{ never_match{configbuilderto_owned start_map alphabet_lenstride2stride memory_usage to_sparseto_bytes_little_endianto_bytes_big_endianto_bytes_native_endianto_byteswrite_to_little_endianwrite_to_big_endianwrite_to_native_endian write_to_len'a from_bytesfrom_bytes_uncheckedwrite_to  set_prefilter set_start_stateset_transitionadd_empty_state swap_statesremapimpl Fn(StateID) -> StateID remap_statetruncate_statesset_pattern_map shuffleset_universal_starts  special special_mutquitsetflagsstates state_lenpattern_id_slicematch_pattern_len pattern_len pattern_mapquit_idto_index to_state_idstartsmatch_state_indexaccelerator_indexaccelstrans  is_special_state is_dead_state is_quit_stateis_match_stateis_start_stateis_accel_state next_statenext_state_uncheckednext_eoi_state  match_len match_pattern has_emptyis_utf8is_always_start_anchored start_stateuniversal_start_state acceleratorminimalset swaptruncatevalidate   next_state_id prev_state_idtableis_valid table_mutdead     &  & set_start'StartStateIterst empty   new_with_map&to_mapmatch_state_id pattern_id slices pattern_idsfrom_nfa bool_to_int StateIterit  Stateid transitions-sparse_transitions-    is_size_limit_exceeded!nfa%unsupported_dfa_word_boundary_unicodetoo_many_statestoo_many_start_statestoo_many_match_pattern_idsdfa_exceeded_size_limitdeterminize_exceeded_size_limit"source# regex     define_regex_type  new_sparsenew_many_sparseis_match'h find2 find_iter'r2 try_search is_anchoredforwardreverse 32    build_sparsebuild_many_sparsebuild_from_dfas> sparse      !"#$%&'()*+    from_dense   !""""""## ####$$$$ &  try_state5id_len  state_mut sparse_mut  from_dense_dfa       &&()) ,!-  next_eoi-rangenext_at+ $StateMut-2ntrans input_ranges +accel set_next_at insertcontains B  AccelTy ACCEL_TY_SIZE ACCEL_LEN ACCEL_CAPfind_fwdfind_rev)set_lenas_bytesneedlesget&    from_slice   is_emptyEC  as_accel_tys   automaton      Automaton"##$start_state_forward start_state_reverse $!""""" ####$$try_search_fwd  try_search_rev  try_search_overlapping_fwd try_search_overlapping_rev try_which_overlapping_matches"##$HI$!""""" ####$$JJJKK get_matchunsupported_anchored"skip_empty_utf8_splits_overlappingfmt_state_indicator  determinize      runP  cached_stateadd_all_startsadd_start_group add_one_startmaybe_add_state  add_stateget_state_builderput_state_builder       Minimizerin_transitions partitionswaiting P     find_waiting find_incoming_to initial_partitions incoming_transitions)min canonicalizeclearG deep_clone intersectionsubtract  remapper Remappable impl Remappable %WW    search     D find_fwd_impnext_uncheckedD find_rev_impY   find_overlapping_fwdfind_overlapping_fwd_impfind_overlapping_rev  init_fwdinit_reveoi_fwdeoi_revprefilter_restart err  &validate_state_lenset_maxset_no_special_start_states!"""""#matches accel_len!  has_unanchored has_anchored.     goto U start_pattern finish_patterncurrent_pattern_id  add_empty add_unionadd_union_reverse add_range add_sparseadd_lookadd_capture_startadd_capture_endadd_fail add_match patchset_utf8get_utf8 set_reverse get_reverseset_look_matcherget_look_matcherset_size_limitget_size_limitcheck_size_limit      !"   utf8 4 nfa_size_limit shrink captures which_captures  look_matcher c c get_nfa_size_limit  get_shrink  get_captures get_which_captures d get_unanchored_prefix   > is_none is_any      build_from_hir build_many_from_hir    compile         c_concat   c_alt_slice    c_alt_iter  c_cap  c_repetition  c_bounded  c_at_least   c_zero_or_one  c_exactly   c_byte_class c_unicode_class  #c_unicode_class_reverse_with_suffix c_look  c_literal  c_range c_empty c_fail c a a a b b b a a b b b c  is_reverse   U    finish     compile_from i  add_suffix a  pop_freeze pop_root top_last_freeze set_last_transitionerror         size_limit   f word too_many_patterns / exceeded_size_limit invalid_capture_index unsupported_captures  0   literal_trie                   3 4  get_or_add_state   i        c is_leaf chunks  active_chunk active_chunk_start                       PRIME INIT    U  E %    U  t t E %.                                        ! " #          patterns   start_anchored start_unanchored a byte_class_set     group_info  has_capture # # m $ g  look_set_any look_set_prefix_any      into_nfa x   set_starts c c c  set_captures        is_epsilon        ^   matches_unit   matches_byte  ^  {  {       ^  {  {           range_trie                FINAL ROOT    U    C  a  duplicate add_transition add_transition_at set_transition_at  =  2  binary_search    U     push state_id ranges        parts1 parts2 parts3 as_slice         intersects     util alphabet         eoi as_u8 as_eoi as_usize is_byte  is_eoi  is_word_byte    )   singletons    % E  get_by_unit     is_singleton  representatives   elements element_ranges  >                         >   )  set_range add_set    )   remove C contains_range    iter_ranges G                      f                        ^ ) 2 pattern M  get_group get_group_by_name    group_len x interpolate_string interpolate_string_into   interpolate_bytes interpolate_bytes_into   extract 2      extract_bytes 2      U  set_pattern slots  slots_mut   CapturesDebugMap  pid caps    Key              size_hint             )     to_name  pattern_names   all_names   slot     all_group_len slot_len implicit_slot_len explicit_slot_len      add_first_group add_explicit_group   fixup_slot_ranges      small_slot_len    p too_many_groups missing_groups first_must_be_unnamed ~   0       )              escape  DebugHaystack   interpolatestringimpl FnMut(usize, &mut String)!impl FnMut(&str) -> Option bytesimpl FnMut(usize, &mut Vec)  find_cap_reffind_cap_ref_braced  is_valid_cap_letter 2 input's advance_halfadvancetry_advance_half try_advanceinto_half_matches_iterinto_matches_iterinto_captures_iter#handle_overlapping_empty_half_matchhandle_overlapping_empty_matchTryHalfMatchesIter2,finder2 infallible'i2 222 TryMatchesIter2,2К2 2 2 2 TryCapturesIter2, 2К 2   22   Lazy  E Пdatacreateowned E  lookreversedas_repr from_repras_char) singletonGCcontains_anchorcontains_anchor_haystackcontains_anchor_linecontains_anchor_lfcontains_anchor_crlf contains_wordcontains_word_unicodecontains_word_asciiC set_insert set_removeV set_subtractC set_union intersect set_intersect read_repr write_repr available  set_line_terminatorget_line_terminator^matches_inline matches_setmatches_set_inlineadd_to_byteset is_startis_end is_start_lf is_end_lf is_start_crlf is_end_crlf is_word_asciiis_word_ascii_negateis_word_unicodeis_word_unicode_negateis_word_start_asciiis_word_end_asciiis_word_start_unicodeis_word_end_unicodeis_word_start_half_asciiis_word_end_half_asciiis_word_start_half_unicodeis_word_end_half_unicode> check Ϭfwdrev poolPool   with_capacity#with_available_parallelism_capacityAVAILABLE_PARALLELISM  E PoolGuard put      COUNTERTHREAD_ID_UNOWNEDTHREAD_ID_INUSETHREAD_ID_DROPPEDMAX_POOL_STACKSšstacksowner owner_val E get_slow put_value  guard_owned guard_stackguard_stack_transientdiscard value_mutput_imp  aho_corasick  2prefixis_fastbyteset  2  ͷ  2ͷ  2ͷ  2ͷmemmem  2ͷteddy  2ͷ         from_choicefrom_hir_prefixfrom_hirs_prefix 2max_needle_lenͷ PrefilterI 2ͷ 2ͷ  prefixessuffixes  primitives EMAXLIMITZEROSIZE  must߂as_u64as_u32as_i32one_more from_ne_bytes from_ne_bytes_unchecked  to_ne_bytes Error     attempted  index_type_implsIteratorIndexExtwith_pattern_idswith_state_ids  from_input_forward from_input_reverse look_behindanchoredget_look_behind get_anchored  E Ă߂  parse parse_many parse_withparse_many_with case_insensitive multi_linedot_matches_new_linecrlfline_terminator swap_greedignore_whitespaceunicodef nest_limitoctal get_unicodeget_case_insensitiveget_multi_lineget_dot_matches_new_lineget_crlfget_swap_greedget_ignore_whitespacecget_nest_limit get_octalapply apply_ast apply_hir> wire    buffer_too_smallgeneric invalid_usizeversion_mismatchendian_mismatchalignment_mismatchlabel_mismatcharithmetic_overflowpattern_id_errorstate_id_erroru32s_to_state_idsu32s_to_state_ids_mutu32s_to_pattern_idscheck_alignmentskip_initial_paddingalloc_aligned_buffer read_label  write_labelwrite_label_lenread_endianness_checkwrite_endianness_checkwrite_endianness_check_len read_version write_versionwrite_version_lenread_pattern_id  read_pattern_id_uncheckedwrite_pattern_idtry_read_state_id read_state_id  read_state_id_uncheckedwrite_state_idtry_read_u16_as_usize   try_read_u32_as_usize    try_read_u16 try_read_u32 try_read_u128read_u16 read_u32  read_u128 check_slice_len   padding_lenEndian write_u16 write_u32 write_u128LEBE  N          borrow'2 is_from_word is_half_crlf look_have look_need##match_pattern_idsiter_nfa_state_ids   into_matchesUcapacityyset_is_from_wordset_is_half_crlf set_look_haveimpl FnMut(LookSet) -> LookSetadd_match_pattern_id repr_vecto_stateU set_look_needadd_nfa_state_idE Repr 2has_pattern_ids## iter_match_pattern_idspattern_offset_endencoded_pattern_len  ReprVec   set_is_matchset_has_pattern_idsclose_match_pattern_ids  write_vari32 read_vari32 write_varu32 read_varu32     epsilon_closureadd_nfa_states  set_lookbehind_from_start              )skip_splits_fwdskip_splits_rev skip_splits intU8߂߂U16߂low_u8high_u8߂U32߂low_u16high_u16߂U64߂low_u32high_u32߂I32߂to_bits from_bits߂UsizeĂas_u16Ă߂߂   memchr2 memchr3 memrchr memrchr2 memrchr3  X2 spanAearliestset_span(set_end set_anchored set_earliesthaystackget_span get_range get_earliestis_doneis_char_boundary22AGC         AG UCC try_insertGis_full  next_back3continue_past_first_match gave_uphaystack_too_longM  sparse_set resizeU% GCCU   unicode_data f mkwordset  WORD decode decode_last is_boundaryis_leading_or_invalid_byteConfigpre$%&>Builder'(DFA)ms!) TransitionTable&classes* StartTable& universal_start_unanchoreduniversal_start_anchored+ MatchStates++ ,-Flags##$./01StateTransitionIter,2StateSparseTransitionItercur3 BuildError45BuildErrorKindNFA  Unsupported  TooManyStates TooManyStartStates TooManyMatchPatternIDs DFAExceededSizeLimitlimitDeterminizeExceededSizeLimit6734   FindMatches32re, 32  ) Transitions5 & ,-2BB +BSeen%Accels! IterAccels!Accel  OverlappingStatemat-atnext_match_indexrev_eoi  StartErrorQuitbyteUnsupportedAnchoredmode  Runner.builder_statescachememory_usage_statesparsesstackscratch_state_builderStateSetids  Remapper idxmap IndexMapperSpecialmax  min_match max_match min_accel max_accel min_start max_start StartKindBoth  Unanchored Anchored ,Empty  ByteRange!Sparse-Look  CaptureStart+ group_index  CaptureEnd+ӎ  alternates UnionReverseFail Match++af memory_statesf4gp >f4fffg  > WhichCapturesAll Implicit     Compilerparser utf8_state trie_state utf8_suffix ThompsonRef Utf8Compiler Utf8Statecompiled uncompiledUtf8Node!Utf8LastTransition    Syntax Captures Word TooManyPatternsgivenExceededSizeLimitInvalidCaptureIndexUnsupportedCaptures     LiteralTrie  Frames-C5  ,-s    > StateChunksIter-sactive    Transition       Utf8BoundedMap    Utf8BoundedEntrykey   >  Utf8SuffixMap     Utf8SuffixKey      >      Utf8SuffixEntryÜ   >     Innerwxaxxx#f4gyy memory_extra  > ,! Dense   BinaryUnionalt1alt2` +ӎЏ +      SparseTransitions-      DenseTransitions-               !  PatternIter,_marker "  RangeTriefree iter_stack dupe_stack insert_stack  ,-  Anext_id   NextDupeold_idnew_id     NextItertidx      NextInsert    SplitS         SplitRangeOld New Ό         Unit           UnitKind EOI            ByteClasses       !  ByteClassIter " ByteClassRepresentativescur_byteend_byte last_class # ByteClassElementsclass $ ByteClassElementRangesA %  ByteClassSet  & ' ByteSet ( ) * + ,> - . / BitSet   0 1  2 3> 4  5 6  ByteSetIter% 7 ByteSetRangeIter% 8 x  CapturesPatternIter     GroupInfo    > GroupInfoInner slot_ranges name_to_index index_to_name  > GroupInfoError   GroupInfoErrorKindܗ\ TooManyGroupsminimum MissingGroupsFirstMustBeUnnamed Duplicate    GroupInfoPatternNames,   GroupInfoAllNamesxpids current_pid   DebugByte  CaptureRefcapRef Number      Searcher2last_match_end22HalfMatchesIter 22 MatchesIter 22 CapturesIter 22Start  End  StartLF  EndLF   StartCRLF  EndCRLF   WordAscii  WordAsciiNegate   WordUnicode  WordUnicodeNegate  WordStartAscii   WordEndAscii  WordStartUnicode  WordEndUnicode  WordStartHalfAscii  WordEndHalfAscii  WordStartHalfUnicode  WordEndHalfUnicode           LookSet   >   LookSetIter %   LookMatcher lineterm  UnicodeWordBoundaryError     THREAD_ID __rust_std_internal_init_fn   __RUST_STD_INTERNAL_VAL   CacheLine       AhoCorasick _unused       Memchr    Memchr2! !!! !Memchr3! !!! ! !Memmem!!!Teddy!!! Prefilter!!ͷ!!!Choice!! !!! !!! !!! !!! !!! !!! !!! NonMaxUsize! !!!!!!! ! SmallIndex! !!!!>! !!!"!#! $!SmallIndexError!%!&!'!()!SmallIndexIter!rng*!+! PatternID! !,-!./">0"1""23"4" 5"StateID" "67"89">:";""<=">" ?"@""""" " ""߂"""""" "" "" "A"B"""C""D"""E""F"G""H""I""J""K"LM"N" O"" P"" Q""" PatternIDError" "R"S"T"UV" PatternIDIter" "W"X"WithPatternIDIter"","؉Y""Z##[##### # ##߂###### ## ## #\#]###^##_###`##a#b##c##d##e##f#gh#i# j## k## l###  StateIDError# #m#n#o#pq# StateIDIter# #r#s#WithStateIDIter##,#؉t##u###### StartByteMap# # ## NonWordByte# # #WordByte# # #Text# # #LineLF# # #LineCR# # #CustomLineTerminator$ $ $  $ $  $$$$$$$$$$f$$$$AlignAs$$$_align$ $$$$SerializeError$what $DeserializeError$ $ $DeserializeErrorKind$Generic$msg$BufferTooSmall$$ InvalidUsize$$VersionMismatch$$found$EndianMismatch$$$AlignmentMismatch$$$ LabelMismatch$$ArithmeticOverflow$$$\$$$\$ $ ,$ $ $ $ $$$$ $StateBuilderEmpty$ $$$StateBuilderMatches$ $$StateBuilderNFA$ $prev_nfa_state_id$$2$$$$$2$Span$$$%%%% HalfMatch%%  !%"#%$%%%%&'%%%()%*+%,%-%%./% PatternSet%%which0%1%2%34%PatternSetInsertError%%5%6%PatternSetIter%%,7%%8%%%No% %Yes% %Pattern% %9:%;<%=%>?% MatchKind%% @A%BC%>D%E%FG%% LeftmostFirst%  MatchError% %H%I%J%KL%MatchErrorKind%%% %GaveUp% %HaystackTooLong%%&M&N&O&PQ& SparseSets&set1&set2&& SparseSet&&&5& SparseSetIter& &&&&8888[8\8]8^8_8`8a8b8c8e8f8g8h88888l8l8m8m8o8o8q8r8s88x8y8~88888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888 8 8 8 8 8 8 8 8 88888 8 88888 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 88888888888888888888 8 8 8 8 8 8 8 8 8 8 8 8 888888888 8 8 8 8 8 88 8 8 8 8 8 8 8 88888 8 8 8 8 8 8 8 8888 88 8 8 888888888 8 8 8 8 8 8888888888888888 8 8 8 8 8 88 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 88888 8 8 8 8 8 8 88888888 8 8 8 88888888 8 8 8 8 8888888 8 8 8 8 8 8 8888 8888 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8888888888888 8 8 8 8 8 888888888888888 8 8 8 8 8 8 8 8 8 88888888888888888 8 8 88 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8888 8888 8888 8888 88888 8 8 8 888888888888 8 8 8 8 8 8 8 8 8 8 8 8 8 8888 8888 88 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 88888 8 88888 8 8 8 8 8 8 8 8 8 8 8 8 8 8888 8 8 8 8 8 888888888 8 8 8 8 88888 8888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888888 8 8 8 8 8 8 8 8 8 8 8888888888888888 8 8 8 88 8 8 8 8888888888888888888888888888888888888888888888888 8 8 8 8888888888888888 8 8 888888888888888888 8 8 8 888888888 8 8 8 888888888 8 8 8 888888888!8!8!8!888888888!8!8!8!888888888!8!8!8!888888888!8!8!8!888888888!8!8!8!8888888888888888888888888!8!8!8!8888!8!8!8!8!8!8!8!8!8!8888!8!8!8!8!8!8!8!8!8!8!8!8!88888888888!8!8!8!8!8!8!88888!8!8!8!88!8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8"8#8#8#8#8"8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8#8888888#8888888888$8$8$8$8$8$8$888888888$8$8$8$888888888888888888$8$8$8$8$8$8888$8$8$8$8$8$888888888888888888888888888888888888888888$8$8$8$8$8$8$8$8$8$8888888888888888$8$8$8$888$88888888888$88888888888888888888888888888888888888888888888888888888888888888888888888888$8888888888888888888%8%8%8%8%8%8888888888888888%8%8%8%8%8%8%8%8%888%8%8%8%8%8%8%8%8%88888888%8%8%8%8%8%8%8888888888%8%8%8%8888%8%8%8%8888%8%8%8%8%8%8%888%8%8%8%8%8%8%88%8%8%8%8%8%8%88&8&8&8&8&8&8&8888&8&8&8&88888&8888888888888&8&8&88888 8 =0IghtT~g)!bhA;Жg& gO& ʞÑ'[Q))3OEH8-++ ~U$FQCg##$␗ gQ,,=d4;_g,-- h?ٷg,sZH.^]QgLQ 3]Eg@p;rgg34AN`gg,gb+m32QBNg)5\Ytig5 8+Տ& {9Tk Q))׮ǦT=gQ,!-VI砆xg,-2BB +Bi,Cw6ggA-2BB +B]Wg%VS9! >ByggՄ!z'HhQdFʓg ̅-bNFnd'ggqggv>>gg .ӈ7x~kqgSSSSS Uȉ؉^gg Ί-Cgte)!bhA;ЖCCCbuf=C=FFFGlobal ě + ċՋT'  gΌߌY g ֍ !- +ӎ +ӎ Џ +=,3g +afؐf4gpu~? gf4fffgSkg3 gϓ;hgJ ggÔ/iaYڝ~gٕ!g՗ gً^ g Pݵeg  × ܗ     Ƙ ٽѬ>gÕg  b.Zggԙs - C 5 !Y ,- s  QUggΚ- s  DpQ    )iZ>Κ ΚgΚڛ    ӥ(g Ü  [?Sg    _F$wRgĝ mӍg˝˝˝˝˝˝ Ü  ,~Og͞͞  ar g w x a x x  x # f 4 g y y  { u'g ! ӡߡ` +ӎЏ+@vyrQgߢ-+Yyg->,g W  hgg,  /#U"Qť ޥ     4,- )7NA  6%1Y  {*gΩȧ  +mg   (X\ hgШS  ;])gūūūūū© ک Ό F4٣ ggߧg֨g˪  ~_  ګ 4Mr g¬  fll4 gg  :f(Qȱgƭ    `HHQg   &+Qg A Ϻ҉}Q߯  ֓K^g  x2ggν ij2QQ,:;,2gg eݪ9`2Q"П C--kԛg ПšӡC© _g#       ο                           #`E g   m91 g   %QBD|gQ   zqg     ]wkD ggg "0" V A"gg     ɜ2۟ šƲԲlL_jg#Qz%g   ѽ߮g     uoFg     :#KI*g!!!!! cz[g!!!!!! GK<g!!!ځ8g!!!K'`jg!!!!ͷ!|=IgKKKK 7Kphantom7K736!!!!!!!!!!!!!!!!!!!!!!/>g!!!! ݟR{  hg!!!! i#?h gh #CCC=C=#!!!3xČc|gg!!!-?=gQ!!!!  +[T h"""" Vr` hg##"""" X)ES4gg"""" JFgQ""","؉w7{t0gQg###### W-#xgg#### tD:gQ###,#؉2ȳ"gQ####dk؊hg### j1g################$$$s0m g$$ $$$$$$$$$f$$ gg$$$$"g$$$,(gg$$$$ Tg3v~Mg$ $$$$$$$$$$$$$$$$$$$$$\$$ $\$Nm9\'gGEd2c5$$$,$ ǒe0yh g$$$$ oȶ4g$$$$ O-g$$$ $˨ ~gg ^yϊu- $$$$$$*02g 2$$$$4L hg#}J%%%% m gh%%%%sAH. gh%%%%#cZg%%%%e{)6gg%%%, @ήgQP%%%%%%%%mł łgłłłłł%%%%%j+ n g%%%% xrTg%%%% %% %&&&Lru0gg&&&&vg&&&&&576Y Fgg&&&& ?Q҈         !!!!!!!!!!!!!!!!!!!!!!""""""""##########$#$$$$$$$$$$$$$$%%%%%%%%%%%%%%%%&%&&&&                            !!!!!!!!!!!!!!!!!!!"!""""#"#"#########$#$$$$$$$$$$$$$$$$$%%%%%%%%%%%%%%%%&%&&&&&t      !!"!""$%%            !""""####%&   !!!!!!""$#$$$$%%%%%%%%   !!!!"!""$#$$%$%%%%%%%%   !""##$$%%   !""##$$%%   !!!!!!"!""""##$#$$%$%%%%%%%%%%%%&%   !!!!!!"!""""##$#$$%$%%%%%%%%%%%%&%   !!!!!!"!""""##$#$$%$$%%%%%%%%%%%%&%!!!!"!""$$!!!!"!""$$!!!!"!""$$%$%%%%  !"!#"$$   !!!!K C" "C# #C  C" "C# #C !!!!"!"!"!"!#"#"#"#"$ %v!gQ!  g T$T$&88h Q?T"9 #"8#) %TPx& #wZw 'F#&crate::util::primitives::StateID::ZERO "rust-regex-automata-dfa-dense"2řǙșəʙ˙G*R{9 Z_ qyřǙșəʙ˙G*R{9 řǙșəʙ˙G*R{9  ( 1 řǙșəʙ˙G*R{9 \a  s | řǙșəʙ˙G*R{9 \a  s  |%  )  2( 1#) 2řǙșəʙ˙G*R{9 \a ϴs  |% )  2% řǙșəʙ˙G*R{9 řǙșəʙ˙G*R{9    - )  2(    Yht [ g BK * FO!*ϴ ϴluG Sºr} \Ȱ'~u (    g       %#ϴ% % - "      ?L  b kx!   g !"ÜĜŜƜǜȜ_wĤĤŤ :Mpy ggg cgcثq1ȡ  xйdmOX;D&/º ºº $ йrU9ºYboW`mU^ kS \ i Q Zg  N W d  LUbIR _GP ]DM ZAJ W?  "   z_hKT7@#,~       %   #   ϴ   %    %     -    8      Yh tzdϴM1 lO7z[pކ /7 /8( >MY d ( pކ *  *  6  gsx†    [itj9 /, / <(    Ztѐ  R5 ѐ-L   S3T aD< _ g|ő Ľ   %ÜĜŜƜǜȜ_wz m ֑  Ց[ÜĜŜƜǜȜ_w^fb %ÜĜŜƜǜȜ_w  $t o$1   ߍm؎䱎,  D܎m[}OD^:ÜĜŜƜǜȜ_w'r ~)2R_  VwU.X cmA{ /7 / 8 / 8   / 8 ! >ByGL  ^  g   /  8 / 8` / 8(     $ 4 D    a m9 E   kv b,[%ytd k #T mKD yDwQ<4 4D<ӌӌ4ی 4Dی<4 4Ddl4Í $ DÍDލ ލ, . <d T        )!bhA;Жt,& gO& ʞÑ'[++ ~U$FQC! >ByFG <D 4T<Dȑ 4 Tȑ<D 4 Tdl4$ D+  įD  V TDڒ ڒ, . <dT      # 3 ِD,& gO& ʞÑ'[++ ~U$FQC ! >By |p `.elszdBL, 1R lҘ,Ș 4֝Ͷ֝< W8L[$ǝ, Ͷ:ФK d, 4<d 4tJ,԰,ʰ}8τ|>,Ӻ,ɺ 0<4 4DǾ>,,}8<d 4t<d 4t; -< d  4 t`     - \ ԡd  4 t   .,,aFznɈ1ʉˊ= $)qÜĜŜƜǜȜ_w)5\YtiCCC=C=(, /|D $d, 5D $a, 1D$d, 1z d $t&! )<<l4 $  D1  H  \ #>IFT z o Y0D  "94 _`D    t  A   ,"aÜĜŜƜǜȜ_w y " ]ɕV\paddingH ͓T֔Fl4ŏ֔$Dŏ*'DΏ YΏ Ɏݘ옏,͙'dst,l̛֔l4֔$D*'D Yâ̮ ̢͢ݘ,С͙ߡ'*j֔l4¸֔$D¸*'D˸̛Q˸ Ʒݘ앸,͙'*j̛֔,  K  3 !"   I    -  m   - t    .Td 0t  .<d 4 t  .<d 4 t  .<d 4t .d d  0t! !.!\#ԁd #4"t$" $".$d&ܦd &0%t'% '%.'(,,~(Nq7|/OcH|/OMstK ƢlKУ%3=KHd $tLԷ-Է  \  .з2 G @  T #r 1z y oXĖ Ƹ  , K  4 #ɕ[B+׸(׸ t * ׸)  %ݘ N %* 6 ׸) mѸ/      [ÜĜŜƜǜȜ_wv %ݘG+   4Q8 !1! t+ 2 ƴI$0 2m8      ["ȶ.Gh  $& % &4$Q%8#!ҹ1ҹ <L #t": ҹ2 ƴ"X!$'!0' ҹ2)!m̹8* ) * ) (![ "ȶ(.Gh *.0 / 04.Q/8-!d  -t,=t+1 + 1 +. 1T4ػ6ܻ8 ܻ7 84ۻ6Q785!(DŻ 45tӻ3Ż ) ƴ3P2$9209 ! );2m/< ; < ;" :2["ȶ:.Gh <<?AC B C4AQB8@#!%D ?@t>$ & ƴ>T=%$D=0D&' &F=m,G F G F( E=["ȶE.Gh G<JļLȼN ȼM N4ǼLQM8K)!%D JKtI*챼 & ƴITH+$OH0O,- &QHm,R Q R Q. μPH["ȶP.Gh R<߼UWY X Y4WQX8V/!߼%D UVtT0 ߼& ƴTTS1$ZS0Z23 ߼&\Smټ,] \ ] \4 [S["ȶ[.Gh ]d`bd c d4bQc8a5!*D `at_6좽 + ƴ_P^7$e^0e89 +g^m1h g h g: f^[ "ȶf.Gh h\нkmo n o4mQn8l;!н)Dܽ kltj<ܽ н* ƴjTi=$pi0p>? н*rimʽ0s r s r@ qi[$"ȶq.Gh sdvxz y z4xQy8wA!*D  vwtuB엾 + ƴuPtC${t0{DE +}tm1~ } ~ }F |t[("ȶ|.Gh ~¾ 4 ˾G +  _       9 @ G  ' _fmM!##*w%')ݘӶ,͙<4 y͙͙/)4m[}OD^+ݘe͙͙ɕ(ݘ͙͙eɕPg`iݘ͙ ͙Sɕ>U?H͙͙4ɕ6 ݘ c͙͙`ɕKbT ]   u ݘ  A ͙ ͙> ɕ) @ . 7  u O ݘ   ͙ ͙ ɕ    _ O ) ݘo   ͙ ͙ ɕ    9 )  ݘI  ͙ ͙ ɕ    ݘ$͙^͙ɕݘ~͙5͙{ɕf}oxFnwritenw  # !s   ( ) N:; E/F l P Q  [M \  f !g  #qo %r '|)}.K:L/ wVƴȶ6ƴȶ׺ƴȶϽl/fƴ4ȶ;ƴȶƴZȶaD#ƴ ȶ   ƴ ȶ n M ƴ ȶ   ƴ ȶ D5b LLZ<L 5\D wMT   PppIterMutp op end_or_lenopohEC8G S  Q   f   $ L 6 # P6  ?<L 5 \  ADb T  @ $ S @OT`inz$L 6B(1  89@ L  ! EřǙșəʙ˙G*R{9 $ 0F >(Ai x(  #T Wsid  xOV&jqlFD, 1lFD, 1\@4, 1<4 4DP,,=d4;_,s<l 4,E,,}8t $:< $0   [$)$)4 ,  d,F 4-6k.t$4<Ը$2  &)$&'/\ ',,33-5es M,, 0<D 4dX,,-!Pۀ<t$ 44ʀP))3OEH8-,߀o ܃t$܃46T] p)){)&t< $L#D #1T lD ( T     ,  ܈+܈ # X"Έ:Dӈ $ΈO ӈ5 ,, "Qk<n-pUeOل)W5t<ل$LĠ#D/T lD (T   , +V :D$G5ڊ ,,g 8R#]-_Ubل0\4 4DGt,~9<|, 4<= T,<9ݲ!] ݲ!] gl  68β1]I8Ӳ] g e8Ӳ,]   %Y$)50lؐ4ݐ $ؐDݐ  %ÜĜŜƜǜȜ_w0uh ؐ  P$)  ː F  UZ Q l  ty    ,ϐ  ,-- h?ٷ, T , $?ݲ!] ݲ!] /&2(  $6 ) T$0 lL g-\ 3Z,ב Tבݑ($ݑ'בDLT, T($̒D$H Ӓ  ]  @ Ғ     f5!=U Argumentf0m֔49d! "   #   g "# :.] I:] Β;]:)] ǒ  T# $/ $, ' d  gĔ'. &4 ޒ6h   &z%# (%/ ( !=]  ]$ +"0y 6<.] I<] +*#<)]   *O)$# ,)/ , %B f7h>β1]I>Ӳ] /.'>Ӳ,]  .P-(#0-/0)*l4 $3+Dt A8L Q݂3|2,\   P Enumerate<< 0{?i݂21-kp41 Ɠ.64 Qݢ65/75710 ˓85łDϓ95Dٓ:5ł;5 >lL>)=2\\ $eOl[=43D?9Dɔ?(<5TɔD$9D<5t B CB# C864 A 8B @A ŕ& @H7Dߕ D:tٕ D:98 M: )RM ݲ!])Q R8 *Q *! f7TC?Cݲ!]  ݲ!] K?2Aݲ!]  ݲ!]̚ G;2@β1] I@Ӳ] G]F<@Ӳ,] '  FPE=$ HE0 HH>C" "=iBβ1] IBӲ] K?J@BӲ,] ܖ%  JQIA$ LI0 LHBC& =i*S T.+U7 X+VSg+m gUVPDDβ1])IDӲ] ԗPEݲ!]OEDӲ,] ˗-) OVNF$)WN0)WHGC- ?i ˘ \; ֘ ]< ݲ!] [ \]8 ʘ0[ 0!  f7_I?@ ՘0[0 ,`J40^ _`1a 1b^1  gabZKFβ1] IFӲ] ĘZGݲ!]YLFӲ,] %  YVXM$ cX0 c.NC3 ߘAide &d}PIݲ!] ݲ!]  hQ2Hβ1]IHӲ] h]gRHӲ,]  gOfS#if/iTU<ƙmlƙΙ m4lV,Ι7 ]B Ùk  Ùl Ù  PkPjW$)nj :XF:pnU:Z QploYty:qo:q}Z :ro<;t;t t5$r s[ <wl<Lw)v\\)<v^]К<xslК<DӚx(u_TӚ<D$su_d қ=}u ]=| }@ ћ>|> ,`4L>~ \?eu?~? g{aJ.]=IJ] Ǜ{K]zbJ)] = zUyc$=y1=de<=܂=0r fA ޛF"=$ wg4.=L QppIterp opopo2Ыh\ .=  P<< 0{?i݃ifk= DjDD QkDDml DD!D,ڜE ޜ(EsnQݲ!]= ݲ!] y3M]E ]$ o1L.]EIL] ]pL)] E Rq %E3EsrGG aIm E ]E 8 JJ-t6PJ bKk KK gԕuN.]EIN] O]vN)] E Ww %E3EjxGM ʝ KmPβ1]=IPӲ] zPӲ,] = S{ %=3=X|GQ KmÞLȞ $Þ~\Ȟ Þ ݲ!] )T T V!QT cUl|UURβ1]IRӲ] Sݲ!]RӲ,] 4 Z"'5\$lyW ԞM  ݲ!] )Z Z V"RZ d[m [[ g֮Tβ1]ITӲ] Uݲ!]TӲ,] ߞ8 Z"'5P  ݲ!] .` `  fc*] O` a  aa g  Vβ1]IVӲ] Wݲ!]VӲ,] ' Z"'5Yݲ!] ݲ!]ޟ  4c ǟKQXβ1]IXӲ] ۟XӲ,] ҟ S!&44 R}g 1R;j"""ܦp!w!)     ǒ6 =D l  8?E L!S '# % ܖ%'w )*),)˗-. 0 2 %4689@:;<=>:=A@H=B = D% E, F3 EH E J E L EN = P$ =RT`gVn4XFMZip\wߞ8^` b'dDKfRҟh\, m[}OD^2&x& Argumentstemplateccg@#  '&%Y&%řǙșəʙ˙G*R{9 Ĕ% %%Ĕz%t$$)% ĔF%$X$ v$##$##!7""g# ## }## #!!"."C! -!! e  '& ݢݢ݂X ݢřǙșəʙ˙G*R{9 !  ݢ- `?ł$ 7%dj<&J )*ȭ3{` 0 cD/ 0;' Ͷ ":;<A\=>ZM=řǙșəʙ˙G*R{9  ?D$DaEJ@XT4cxZp    Qrc`Au~ ?    )E"E*I %.^ ^- :-  +^% ^- * *  (_6 _ u(*_= =_ 'O_D b_ <'t_K _ '49&8start_id#&9 &:sty%;- < "!_ #` O"%` '%`$)Mx$*Q$+S W!,\` .o`4!0[@!1^ 2` 4` j6` 8`5:n;r-<s- =u>|?~ @+a B>a`DEvE  Fxa Ha JK La Na +Pa RaTU Vb& X.btZ[ \Zb8 ^mb)`6a WbbJ db fbP hbr3^3(3263.252:1100`00///s/'/A/,.--.U--c-,,,1,+ʡ+ ,r+עJ+Q+))f)D)#('''N'U'&&&v(%%%İ$$%$##1#""}""!Ʒ"E!   I c gu$X$+8 5g}9B{c(6U\d! 1a$,ʡѡ-!Odl 0֢[$,-!Odl 0¤ˤ[$,-!Od¥¥tϥ 1ޥϥ]$,-!Odަަt 1]$Ϧ,-!Odt 1§]$,ԧۧ-!Od $tũ K,ɩTũ<D(T\ , +'     - d, $ <|        4 ި٩N<,<,, ݩtMȏvcurrent݆\!d݆$t슮C$L<D(Tή\ή,ڮ #'     - d,$ <l w L   v= <ʫ,<$, wXȏ*qtx|t $݆`'d݆$t̰İ E,İ$Dʰ ETʰ<D(T   , d,$ <|        4 wް? <,< Qk<ȏmptc݆N)sܱB,, 0ԑ $ݲT,,-HĤ &|ҷ "ݸԇ$<,T 5\\˸$hL,\:[\B-D match_indexR state_indexQr"]¹ܹ@$, /׺<$, /˻#^$, /d 4t| 4,۽N$d; ;Dھ  0Tھ\ # I d D  0$ t3   $  $ܡ      5$ L<,, CD; ϽÜĜŜƜǜȜ_w,˼ 4ڼłD,ӽřǙșəʙ˙G*R{9 O$ RՁ#~ޡТn$T'$ ř řǙșəʙ˙G*R{9 b"|,ł$d!$Ut $:\ $  ./< #4OY,,q̫A2?-Ard M\̫yΫDD4 W4DOřǙșəʙ˙G*R{9  ,@Z /7Z /8pZ(Z ZZET`h  ٰT   d $ t ٫ L٫ ٫  4  ! , 9 ٬  @ ݬ  Tܬ  B 8!ά T  $ ,  l  L ( "4  _TryFromIntError q:{ a  D  ݭ 5dݭ Ͻ /  &  t     c8 ! "dү į lį .  į $4 ̶? D   7d ϽR/&  t    c8! "dͰ   Ͱ Dڰ  ? /lڰ  Ͱ  ƴF#!/!lͰ #m % & #, %B&8$! "["ȶ". $d ,ű $ (<ű `   Pȏ(H'$))' F+)UZ Qppʽp opopo2Ы+*,*, 3*-3D /4/ (  - 4F./T  в  1 2 ., 1B280! 0{  Ű  ۱ \5ݘ섫 ,ت ͙< 4 ͙͙RL!ÜĜŜƜǜȜ_wն5i͙͙$͙|ɕU| )͙͙K͙ɕ|ݘJ͙>OX͙͙ɕ ȏAřǙșəʙ˙G*R{9 l xI ߱  ͙͙͙*ɕ$͙ jns ɰ # "#&)R- .A  X    ̶X Ͻq,̶ Ͻnɻƴgȶ>  t  t  = -d ܶ dô  0tô  n   - dݴ , $ݴ < ݴ  ܴ $  l #!   "#B      .  ,ӳ ,ɳ A| #ȏV&y6#B# wT  4  8D d  , .< d      , Dz4ZT  4 .D  4  3ED d  , .< d       %, +DzK=siG   D  $  d : l D (T l  "#&d d , $ <  d τ$ t       ,         F,     ĔL ,   ȏȏBtSr-tiP.τNd , $ < l 4 $ D  "4  wvO  .L  QΪΪChunksѪoҪ chunk_sizeoS #ұ\  k    , << 0{?iWȏnf l D (T d  @# 8  V, ,  3-5dd  @# 8 - a   < , , *-]T  4 .D  ,  C"  ZȏT , DzAd , $ <  d  "@$ 9  I, , ȏ(Ld  @#  J%  ;, , =d ܐ d  0t   K, , y8 l D (T d , $ <   l 4 $ D  2    #  4      X $ ,   @Wȏ;W-Y-\d , $ <   "l #! "#B  R, , ȏ#G##T  4  9D  0  M"  jt L   "J @n ٰ)1xnȕ  /8n  /8hn  nܙ Ͷ Ͷ / 8nқ +   0 9nӝ )*3(n   n n Q`ltDz  ՁͶyN Z%. ٰl6x - d $ t  L   U  # , Z ;,  @  T  B 8!Ę L    D   D4    ƴ K # / D  m   4 B8 !  ["ȶ., t   D  ?/ l   ƴF #/ m   4 B8! [ "ȶ., \ " l ".!  #4 ̶!? D  # 6d Ͻ R#/& % t ' & ' %c&8$! "$ ,Ͷ 'L  W,4\ +  6l +_*! ?4 ̶*@)D  - ud Ͻ)R-/(& / t 1 0 1 /c08.! ". 4 U4  XU4XD Oׅ 3 $ 5 d Ͻ3502' 7 t 9 8 9 7c886! "6 < B4  X <PD  ;  $ = xd Ͻ;=/:!& ? t A" @ A ?c@8>#! ">d ,  $ C$< `  CHB%#DB  &@ FDO TF`E'hm GEy G)( NE HND J OJ ( ̽ H O@I*)T   , L M I, LBM8K+! K^ &{  7    \ Pݘ , ͙< 4 6 M f + ͙ ͙  5 ݘs  ͙    ͙ ͙ ɕ 7 ݘp Ձ ͙ + ͙͙% ɕ "c͙U͙͙+ɕ+BsͶ͙=͙͙ɕ ͙7͙͙ ɕ  ͙8͙͙ɕoȏj  ͙h͙ɕ͙3        g f D HI$w  ƴr  ȶ a ƴ8  ȶF A ̶! Ͻ   ۂ =  ̶ Ͻ v ? ʅ Ͻ|F ϽTŠ I d 4  #*D  H  ׅ _  y   d 4  (D     w L Ļ d  0t t  d   0 t  @     -  t   [     -  t   . t    . t     .d , Š $  <   $l #! "# B  .  , , BvMmՁ>m5`wȏz1##  {$tŠ T  4 .D L t  , .< \  Ͷ           , 8DzX#Ձ ͶX~MT  4 .D  4 @D L t  , .< \  Ͷ             , VDzJ'ՁGͶc},  D  *T D t G ׂ Ͷт Lހ  ހ d   0  t L  Ԇ t   0 +\  *  D  & /    \    *  D  $ /  5  Ͷ t D  "*T  *  D̃ $ /  6 D  $ _\   & (M\ υ tυ DӅ ڰ (TӅ   , F   -  h! !-! d , Š $ #< , $ %#&$%  &%$4"#$  "     ǂ  ' , łD , \ :, d Qq}1$ @9ZGcX rȏ8NRU start_indexd]C˂ Ql ƫ  )   ڰ d @ڰ Š l 4ƈ $ Dƈ  *  ӈ ܈  _݂ , 9TՉ Չ 4 .D  , A   ZȏT , DzAd , Š $ <   "l #! "#Bԋ  R, , ȏ#G##Š T  4 /D  0܎ A ĕ ڔ jt L DHv ٰ)1pvի ٰ)2v (v v v ET`hDzDzٰDٰ (w  $ Hw  %*pwի  &  * w  '  )  5   g | w 2?HQZKTDz]+4Dz=   )DzDz n ) d $ t  L   0 t (  , 5  (  @  T  B 8!Ĕ T  $ , ̳ l D   "4 ̶ ? D   5d Ͻ R/  &  t     c8 ! "l 4  $  D h   PN $) FUZ QltyH ??D @ *   @F/ # l #/"!   , B8!  (4 ̶"!D  $ ;d Ͻ!R$/ & & t ( ' ( &c'8%! "% \  $ -l  ,- 'l ,*+! 04 ̶+@*D  . Cd Ͻ*R./)& 0 t 2 1 2 0c18/! "/ \  $ 4 l    4B3!#53 "@75OT 7`6#hm86y8%$ A69AD ;B; *ܰ  9 B@:&)T  0( = > :, =B>8<'! <\"  ^ ^Cݘ , ͙< 4 ͙͙OIo͙͙ ͙xɕQxřǙșəʙ˙G*R{9 9 E ͙͙%ɕm͙͙>͙ɕoJ͙͙@͙ɕq4  # ͙͙1ɕ+͙}͙ X5b9:X   }    ̶ ϽQ      n!̶ϽH  w̶Ͻb+      t l 4  $ D   ' | "!  ""B  Z   -  t   u     -  t      . \  $  l   ,| "! "" B   .  , , BpA 6MP""y"" n K[  \  4 .D   4 .D       , DzDz3Y\  4 .D  4 @D  ̚ 4 .D  "4 BD         !,  RDzrDz%H~xA\  <  0L +l <   L "` "`԰ ) 1@`` [T# l     %4 ̶ ?D  \  , <   #4  W T  D   D 4 T  4\  E 4    =4  M  D  d  #   $ 4  _LD  t    4   *)(     `< ,  , s< k4 Ͷ Ͷ4YÜĜŜƜǜȜ_wo  -    }R !̶  I &   ;ׁ "܁ $ׁ \ ܁ ׁ /! 5  LÁ , \ \ 59К=w lń 4ʄ  $ń Dʄ |ӄ  & (Mӄ  & (Mń "   !   5ń -D ڰ ^T   p , , \ 8VZ ڰ  lЇ 4Շ  $Ї DՇ |އ  & (MЇ    4Ї )D ڰ ]T 썈 \ $ \ Ē  \  $ l \ψ    ,  Ȉ   Lj    F,  ć  ^d , \ ,ȇ P.7 Y?H~T  ڰ   H\  4 .D  /̗  C"̊  `d , DzA 1 ] l 4  $ D    !5 F F9 ] ] A ] I ] I ] Q ] Ƀ ]    l 4  $ D  !] ] ř]] ] [!] Q] Y]`J   !5 F F  n ] ],΋ ,ċ (+ ~NevX]! ]řǙșəʙ˙G*R{9 ?:left_val0 right_val/  '    [  S  4 .D ߍ 4ߍ  A   jdȍ , DzA l 4  $ D   \  $ l ġ  0 - B| "! ""Bҏ  r , , K Dh""  | r O(  (-:BKU (  +І  , )  - )  ) . *  6    y ~ /<ENWGP Z)2 ;    (   m   d $ t  L   / lͣ /  , 4 ǥ t   \   . ! "    , <  =\  [d ̭ \  ,  ʤ  , < =\[ Q    ܤ *\ܤ  -  ۤ 1  , < = \[    d  $   3T   ʠ Ӣ  ݘ읡 , ͙ 4ע u$ $@ mj/S  =  < t $  Q54  g & gO^ ȏĔUǧ  {  W Ĕ   l ʫ  ,,=d4;_Lҫ řǙșəʙ˙G*R{9    ǧ nw ٨ N, < ǧ Գ \  3,Ǭ  <۬ ǧ ۬ ,  3 L   Z LĔ   T  ǧ ǧ self__tt$׫ ǧ ?-G?۫ 5 _ref__self__tt(  0P  1* -2 << 0{?ik  w  - : ,sZH.^]}_h   w  (  3  4 + 65 řǙșəʙ˙G*R{9 y   ! = J L ox  ³    '  08(  /; F bl (  6   7+7ch gy~jo       (L | 5$  ,    $ , 4 : ( /`  `   `  M       ` I    % 7 ? M       &(0p__self_0 u  w E տ t ( /` S[8199EQ  ` :199EQ,2 @F Y_h;w} g  <    =     ` D M> 4 @  y g  w  ?4@ s & BHOV]k t}      ,>G P* O]  S  ]  t         G P  /7P  )2(8 88BQ] h    (8 @P  A*P  B *  6 ot8      Ucn d 3tFdF F#TF IDFF FFEÜĜŜƜǜȜ_w34AN`EDE QZw! y S%,tTdT T*U GDU܌U UUSÜĜŜƜǜȜ_w34AN`T/DT MVw u Q#*dk$l $$$$$$*0,k4lkDl $l tltll < >(=%L9] <=] 8 9<A] ?8I] @8ɤ]  ?I]@Q]A&8?A(', H0t 4 H(G*D!]B ] ]4D2E)0<]D řE}] ] [B!]?Q]@Y]DfCL Jd 4 J(I+D FGI F-, Ol < O(N.L'  &E0%&K*, R0t 4 R(Q/Dȥ Tdȥ 4̥ T(S0D̥  P QS7 T NPAM1'= L řMA  Lݥ2    _ ¢   ] ]U ! , ,      #   m[}OD^     I R   $ސ |    E  b   K   )              Z     L  A  u ł?,Ƣ     tcV   s bۏ !] 8 RP j -  :j  HL7  e  revsearch ' 0  5 6 ?Ő @`B&   w J Q    5       p .   _    j 9  G Y   K  g    Ed ,t.ԟR$ tا<ݧ $اLݧا) J̨$ , ,łdȧ 25d W  l9  , 1lî9 ܮ ,Ү 1 tѲ<ֲ $ѲLֲѲ\ Il1 ]t< $L\֐ Cl9 ] ] A ] I ] I ] Q ] ɨ ]    t< $L !] ] ř]] ] !] Q] Y]XB䖳\֐ l ] ], , /d  Gۏ ! ]5 [  5 Ő , N E֐   ֐   ֐ k( F  G 34AN`h 2  +  7 puŸ+ 8 ,gb+m ՗ +4   C Lš   ̺՗ Tк   $$$$$$*0 U՗ (<   %b l ԙ L՗ Y  P C*,}v ՗ T 4,    ,՗ Y՗ > _ref__re 08( /; C\f (  H+7ch uzfk    $ <$ #DT.D   %M$)5-: m$!, *, /< [" % 8 o N$ ř9L PU9|m XJ,   T Si5 t %T Ra5 ԡL qR #! 1< >$ lD s[;1 T YD     L !& 2 m ##.    [" .O<$ !<# " &| $!l"O !*   # $6#>KX$`BL #$ V p % ̮ ,Dm[}OD^s< |   hqR)W hOw3N   44  !z ^ 4  Bҧ +   ש D   ` 3 A - ̃T $DЮ   % P$)5_Cԃm  d<  . L [" % 8ĵL   \Ȱ   %ԍ L %* 6 ̵ m !, d< . L  [" .L  \Ȳ   @ $0m!,<    <   & | $lO! [ # /4   *  GOZbo| la } a  / ,D m[}OD^    w<m[}OD^ԍ:   lu< *ԍ]  QZ GN w` X3T  G 4   !  {    k  y  a  o< C J QX_ D ,<<"3$4&bI "rust-regex-automata-dfa-sparse"2 $ /7$ / 8$)2$  )2$  ) 2`$  ) 2(    $ n}   }Xa3< z E#3337T3AD333hD3 _ԍUqLO 33# 442ԍ2 D2&/w&ZN* ,33L4 %\4 44.>ԍ3 U,3 | TH |ILII\IHlH# !#@H3 H C:|KLKK\K !  D$K \<dK4KK DK` K  ?$  K(@  OT `  jo  {  R,K Ĕ4KdKK C. ,K4t tdt݆tcttLdL L!LM I ] ?   ] = >?A!] @= I!] A= I!]C@Q!]DA!]BCDB+)!!]E $] $][&K$ H*1$<]G řH$}]$ ] [E!$]@Q$]AY$]GdFLt\ J \ IJ\ I-,,_ VT_ _V%U0$_p ] N "` M N ]M  f2P.<TO PdQlL ,RO g,QRL/@` ` L"`K+_ t__UfT1Ԫ_0 T42t_ W L_ Y_ ZY)9 Z _G3f X+p_ S WX_ 5t` S  54`\S4`^( `_ <`^@_6]6"`/` \]`[7-tbd b|b d/c8)b94b _cOb9Dbbe\c bKlc# a:Ԃcb4c aR`;Dc4cf|cdc N c h<.c$c QێێRepeatݎelementNJH"h`,cg=\cc*4c PTakeII] fge>c#,dkTddk%j?$ddtddjDi@d/i]A,enTeen%mB$ee!eemElCe4fpdff p%oD,fhthuEh7\h u7l$htFhhU4h tFsGDhitivHh\i sav=rIih4i rFqJDi4ix|idi  j zK)i$j zEq,jyL\ji*4i xyJwMi#|lDl l$l}NTll~ Ml74l _ն}h~=|Ol'4mmo( m @$6   "&'yH5J3""K(= D K  M%I P O !] !] _ be f hrt A H O  y z!z"#${c%ԍ? >%i%%z% LK$m$$$$$g$Ĕ $K $m[}OD^D# # #e"n"""  "I"""!ٙ!e!ٙ!N   -!F! ! ³ !  (!     DO$PDOe P >ٙٙԢ  ٙSٙm  > 6   ۏ !!]V &    4_ <U͙ٙ!ɕ,bͶB|ÜĜŜƜǜȜ_wٙ[  \8Q$e *  g<D 4T<D 4TdŅD,.<d T    dž  Ŀ ,   Q q52[  W  dBL , 1Ͷ쁏<3LɎ[$ ,{7|KdՕ ,˕ 1<d 4t<d 4t/ -c,ޘ ,Ԙ  3W # ܭD $Χ[< , 1 ܲD $ӴX< , 1 D $[< , 1 zd $t !  D  c9D ")2D $c%4UD < ,<ݘ͙2ٙ;@C   s A D $̵ݘ ,͙K D $ݘ ,͙K D $ݘ ,͙K 7  ,     4;$(t( ) ƴF$ 0 d) m/       ["ȶ .6h   4;$1!t % 2 ƴC $0 2m8   ["ȶ.6h  4;$1ā׺<t5 2 ƴS$02m8  [ "ȶ.6h$ # $4#;"$dϽ "t!7t %  % .%T(+ * +4*;)$(D()t' ) ƴ'J&$,&0,).&m//. /. -&["ȶ-.6h/<25 4 544;3$%D 23t1 & ƴ1T0 $6006!"&80m,98 98# 70["ȶ7.6h9<<? > ?4>;=$$%D <=t;% & ƴ;T:&$@:0@'(&B:m,CB CB) A:["ȶA.6hCdFI H I4H;G*$*DFGtE+ + ƴEJD,$JD0J-.+LDm1ML ML/ KD["ȶK.6hMdPS R S4R;Q0$*DPQtO1 + ƴOJN2$TN0T34+VNm1WV WV5 UN["ȶU.6hWX4 X6 "    v }   + 2  9   ;%,y"pw~ Yݘ ,͙4U ݘ ͙ ɕ    N > J ݘ ͙ ɕ    E 5 1 ݘv ͙ ɕ    ,  _ ͙ ɕ  % O ݘ  ͙ ɕ   S C )ݘo ͙ɕ-ݘI ͙ɕݘ#͙`ɕݘ~͙DɕlxO   %   E    -H .  7 8 ,A~B K#L vU V#|ƴTȶ[W3ƴȶ׺ƴȶ Ͻ v   ƴ ȶ  b ƴ, ȶ3 %  ƴ ȶ ƴtȶ{mIƴȶ$,  Ƣ 2 "  I      - ̿У m   -t .T̩d.t  .<͊͊dՊ 4 tՊ  .<d 4 t  .d܏d.t .dܵd‹.t‹ .֋ ,ۈ ,шvt)Mxb5 K k0l / ƢУE c mD5b LLZZݲ!]ݲ!]tN  3β1]IӲ] `Ӳ,] R#/?< 4  4D?Y  P,!-VI砆x`e     Q        , ,-2BB +Bi,Cw6,T , $ݲ!]ݲ!] &3(   $6) T$0,T (${?YTDyT(ݲ!] 9 0  c_       f  (    !O  "  g! "β1]IӲ] ݲ!]Ӳ,]: Z$#0# C?Yiβ1]IӲ] &q%Ӳ,] %S$#'$/'<+t$ +4*4?YL Q))׮ǦT=*) \Ц  P<< 0{?i )(!ch,("., Qܟ .-#/-/%$ 0-D1-D2-ł3-t676# 70&`ݲ!]_ `/,_ ,  [bA 5 0645 &49'D82t8*)(A2HAtDڰ H(G3Tݲ!]ݲ!] ?/3ݲ!]ݲ!]̃ ;+3β1]IӲ] ;`:,Ӳ,]' :S9-$<90<9.C?Yiβ1]IӲ] ?F>0Ӳ,]% >T=1$@=0@92C?YipFGݲ!]E F% E   QJ4 M I J]!Keȭu!LI!KLD5β1]IӲ] Dݲ!]C6Ӳ,] C[B7 %MB1M98D#?YjR3DU1DU(T:TSTݲ!]Q RSq&Q &  İW; a&Q0&  X< &V WX'Y'ZV' YZP=β1]IӲ] Pݲ!]O>Ӳ,]: OZN?$[N0["@C)?Yi5,a bF-cO`-dao-u cd^Bβ1]IӲ] ^ݲ!]]CӲ,]5 ]Z\D$e\0eEF\ $kGlu/?Yjkݲ!]i j%2i 2 QmH M2l m]3neu3ol3nohIβ1]IӲ] hݲ!]gJӲ,]8 gZfK$pf0pLMPuݲ!]t u+8t 8 WwN 5?Yz8v w9x9yv9xysOβ1]IӲ] sݲ!]rPӲ,]' rZqQ$zq0zRSݲ!]ݲ!]  }T3};?Yβ1]IӲ] }|UӲ,] |R{V#~{/~WX4 ?Yu??Y:BBI.5<)  /6 =:^els z - 4 ; '   %   kr"y$ &(:*  ,.50sz2486\c8:'<>@\ , k  5 U řǙșəʙ˙G*R{9  9  E J ^A ^g   ͚ T>ܟ ܟ q  ܟ !* ܟ 6q\łA&EdU   X [G  v N h d~q R ȭ  _     f & U o / &  J S  # o R ?-,?H +    e  2    #  8' 0 B/ 8Hj    . ??  P?! ?   c ?3? ?;?.,e0T1=2&3 ?u? ?}??A E&!I "?$?&Q'V I(?*?,_-a @ .?0?0 2i< 3l  4?6?+ 8t7 9v  :?<?  >?@? DwB P   6  \ h {  E֟  ڰ M+._-İ )/ /^=n 2  @  h o d!/a$ ,-!Odl.[$ ,-!Odl.[$ ,-!Odt/]$ ,-!Odt/]$ ,-!Odt/]$ ,-!O|3,T<, 4<t5"$ _,\X < ,<,,=    MPTX A T $<,Ԯ< ,<,!$(c <, .<t/!D YTg< ,<   /2b  B, , 0<, .L/\ YlY, ,   /-1`   &| "<, .L/)T Y\ L ,\.O   H-JКNL" " @$ , /<$ , /#^$ , /d4tĬ|4*L$d D .T\ # G d ܳD .$ t1  ޡ$  "    ܫТ 3$ L<,, CD;  ,4łD,$jAaՁi '}Xvޡ;Т $T'$ ř "| ,ł$d!$U<, .L/\ Yl[, ,   /-1a  D4.DO ,x9~Hٰ*2ޖ*3  ( KZ f n+Dz4ٰR!-d $tLZ P ,_ ?, @ T B 8!įtl.  &4̶ ? D9dϽ R/ &§t§ § c8 !"l/ !(4̶@ D;dϽR/ &ШԨtԨԨ Өc8!Ũ"d ԚDɻ= -l ƴD#!/!?ܚ#mʩ%Ω& #,ͩ%B&8$!"["ȶ".,$l4 $+D*+"l**)!+4Ī̶)@(DĪΪ,>dϽ(R,/'&.t0/ 0.c/8-!ܪ"-$1 ȫΫ41lΫDѫ4(3Tѫlޫ4$ޫ6Dޫ56Ϋ#2352+81t, $807 L,;7DĬ<K<찬D ;NKD:! ƴ:M9"#=9/=#$?9m A B ?, ABB8@%!ͬ>9[ "ȶ>.,ܬ @׭ G1l׭ DڭG(F&Tڭ, I7 d I,H't׭$ JFH JFH(- EJʭ2 dʭ EWD) ʭ; 4 D?C*Dŭ@ 1CST ,   qx  Lݘ쿤 ,͙<47 N g , ͙ ͙   Z   ͙L ͙͙" ɕ" ?͙r͙͙Hɕ!Hݘ͙r͙5͙lɕf;|͙w͙͙Mɕ&M4Qy ,cݘ >8>G͙~͙ɕO 1K  $͙  # " #-?17  ? 8 > ?a%  d 7  o ̶O Ͻ   l ̶K Ͻ   ɻX ƴ1  ȶ? C   ̶ϽQ hƴcȶڌ \ < tt= -d ܖd .tu   - t      . l4$D    ., ,U YCx U2l4$Dd,.<t   ,2X,l4$D4 3HDd,.<t     ,F<r]!l4$DlD(Tx  ɕC,4 < t <7 !!D  E TL+@;  Om0|4      L+@;  Om0|T  ,  ,ڤTJ!,4 T & ( MԟD x d,4$ %$l#  %$"# A#%&D,#s"  $ ",!"D  4+!=+t =7*!!D *E)T-  ,ڤ!,-M,$L/) & /)(M./D,T.~(1 ̫$ Ϋ QU1d?0^( 0!\&(,'( 4 D5'645  65443'4\3]2 l:2 & :2(M9:8  9 ',8[7/4 ; F  ; &7 cz> L ,LL ɕ" < T Dhpwɕ Uѷ2ICѷ,ɶɕXj̫Lj T- #B 2 B  !+&'npatsE)^2B, 7J##  ڤ G & ڤ. ۭK*ڤR RD Q`̫~Ϋ$Ζ |ΖDӖ7TӖ. c  ,@$,1\ėė4З 8DЗޗ>, ,Dz9l4$DC, ,r8kϜTԜ $ϜdԜlD(Tx 8!ʜ'4 @  @t B !D > TLƝ̝+@;  ǝOm0|4ҝ Ɲ Ɲ  ܝL+@;  Om0|T  ,  ,TJ!,Þ4֞ T֞  ߞ&  ( M Þdɞ x ɞd͙,͙4% &%l#  &%"# A$&*d,$s# $$!#͙,"#͙Dџ  4-"A-,Atۭ,B+!D+>*T/  ,",/M.$L1* à& 1*(M01"d.T0)3  Ϋ$ #Ϋ3A?28) 2"Ο\')͙,ş()͙ 6 D7(87968, 986@5(6\5i4 lա=4 & =4(M<=ҡ;  ҡ<ѡ #(,̡;Z:.4 > \ ?'  >!?: gBȮ d L͙L͙͙Q ɕ! ; T ͙Du}#͙͙+ɕ%͙͙͙͙b͙͙͙?VP͙͙,͙.ɕh͙vL ͙M͙ ͙~C-FIB 2 B !"+' (^*^4B, :p##9   f G T ۭdEr&7c pΫ\4 9DܣJ͙LL9 @ٰ*2x*3˰*3h   Ͷ * 3+  + 4)+4(   ƧŧTco wDzՁͶpyJS#,ٰV%udT ʻLֻۂ \ʻ  dֻ9\ 4 c %O4  KD\ 4  t4  Z D\  G+4 JD\ּ Dl# ּ4 U D !      7нTԽ н dԽ|L  \,J<d %8 <#l %ǿ(˿ǿ6˿$     йͶ\4ͶͶͶ;Ͷa,Ձ łgł &(#ۂ   ] :     Y 5o( 2\l řr$ ř&'| 8/d4ʡ D`  P݂ I%*  G  V[ Q݂ m  v{    d D ł Dd !! 6)L 0Dd&TD1l  5m T"D"1!l#$!#  $#!5 ! o%4  % V v77& 'ȏ,Ͷ 4Ͷ"݂݂݂řǙșəʙ˙G*R{9  ݂#ł}gdY   =:rSZ O start_pattern_lenslK old_start_id new_start_idF-- 2ʡ@  .d $tL\ ܓQ ,a ?- @ T B 8!L D >.  ƴ E # / ? m  4B8 ! ["ȶ.-tD=- l ƴD #/m  4B8![ "ȶ.-\"l".! #4̶!? D#6dϽ R#/&%t'& '%c&8$!"$,Ͷ'Lۂ ,/\+6l+Z*!?4̶*@)D-udϽ)R-/(&/t10 1/c08.!".4U4 X 4QD?ڋ3$5dϽ3502'7t98 97c886!"6<B4 X <PD?; $=xdϽ;=/:!&?tA"@ A?c@8>#!">\$ $C$4X  P CHB%$)DB &F FDU Z Q FlE'ty GE G)( HE KHd 4 K(J*DT - L dϽJ}L/I+& Nx O l# NNODM,! M &   ] :   Pݘ ,͙<4N e ~ C ͙ ͙  R ݘ  ͙   ; ͙ ͙5 ɕ/ U ݘ Ձ ͙ I ͙ ͙C ɕ= @͙s͙͙Iɕ"I$eͶ͙`͙͙6ɕ65͙`͙͙5ɕ55͙`͙͙6ɕ6     ? 9͙͙q͙ɕ  0      D  HK#%   ƴ  ȶ ~ ƴW  ȶe ` ̶@ Ͻ   ۂ \ " ̶ Ͻ  c ͋ Ͻp9Ͻ}H  ϽHd4 (Dw ڋ w  d4 (Dw  Ld .tt d  . t@     - t [     - t .t  .t   .d, $ <  ., ,\)Ձm6 U  dVɥ}d,ɥ$<Lt,.<\Ͷ      ,;ՁͶNtɥCd,ɥ$<4CDLt,.<\Ͷ 졀    π Ȁ   ,VAՁ>ͶYɥs ,Dƫ (TDt?ͶL Ĵd  .  tL܅ ܅t  .)\*  D  $/  ?  \̽ *  D  $/  ? Ո܊ͶtDڰ (T*  D$/ 6D‹ $ŋ?\ & (M\tDڰ (T ,F   - h! !-! ,#$#l#  $#"#A"$d,ɥ$(<8) "Lڤ(])8'!5쐎 '&%7%&; %Ύ    ׊*ܻ ,łD,\,dEeg\$ԉ~5WyF,&Wɕ qux }v "@x##I ƫ    } ڰ ? ڰ ɥڤT3Ï!  Y  , 1dԐ,ِɥ$Ԑ<ِԐԐ"l#0 A @#I,Đ ,!E##ɥ Tؑؑ4.D<,ȑ ,Dzx9~d,ɥ$<ǓB, ,r8ɥk,Dƫ (TDt' ߗ  tߗDڰ  ( T\ ̂     \ؖE  %Ͷ˘' R%j4Ę WM<Ę/L  \ǙǙ1\ .   "` 8--* 2J ZbS  g @`` "`)ǙO4 _D\"Ը\Ě "- К!tĚ#4Ӛ !X DӚǙ\  H0Ǚ4 JDǙ\ D\ĐǙ4 LD,$%$l#  %$"#A#%'#l#  '#"#A&'T + 4+/*DP , #&d *],;)!m   )$(3 ՗.5. LłD,Εܕ\,dȖ pٗhͶz13+!=FͶDwͶbͶͶEc,=7'͙X͙ -   w# &"#.###ƫ ڰ ^ 3 1    t J & 4<( & {9Tk l (,$ ř\Ơ Р+ -Wt) )  #  4  % _".4 WGDt* *    dѢ %t++ & ( M l %t֣,,ѣ ң& ( Ṃ ,t--̣+$ 5 t "$䗤4 _LD/ %Ԥlؤ#  "#Al#  "#A<.|, .4#<P$ d#c$;"!&D "ܤ!T/4˥ _TryFromSliceError' g~W! D˥'  # ,%&D' Ԡ& %'  ř&r} ա£|Ȥ`/" ))׮ǦT=L XT١řǙșəʙ˙G*R{9 'U6łD1_Bƣ.PÜĜŜƜǜȜ_w,̤r,ÜĜŜƜǜȜ_w+_4łLDMV_hqz  start_type{ PIB -%#$###   x ^!ʨ T d g =4 U<  4,ܨ daOܨ=4 dXD \ ,  dd DebugStructddd dd has_fieldsd19qclN <T  5 8+Տ4  D̔$$ì ř T<. 1, pJ&< , dڌ ( t      ( \ԭ ř @Kԫ۬  ,!-VI砆xLܫ (^ ,߬vY sDiqz\2ڌ\ V ڮdܮ ڮ8ܮ/$ WD  ڮ1,[Iڮ:4LD \Į , &/QSU|3Xհ 08 phٱ  p    x(  C O WqzD M3<"ɶ+ɶɶɶҁ  Å~ÅaÅD3dڌ$tL[ ԜX ,` 2l  \ L   +@;   O\A R   I4l 1 (4 _նO D1d `5 ,t   h8 !"XXȈ ܪf8!YA| xY<"ZZȊ ܀f8!)\"p!  "p !J ## @%#OTƷ%`$hm&$y&? '$|< $' +Tl00+Z*+4 *)D,+dϽ)Q,/(&.x/ l#.N/D-!-\ $4l!l4E3!*4̶3@2D5=dϽ2R5/1&7t9 8 97c886!!"6\ $<"l;  < ;J:##=:$@?=OTƷ?`>%hm@>y@+& A>T $A B'lEBd4 "E*D(DFdϽDSF/C)&HI |"HNIDG*!!G$T[L[d L1K,('4 KGJ-D M 4NONPMO, POM.0MJTU  ,TTUJS/!T\W\V WSDV;R0$T]8|R]7Q1T m 2x   ^ݘ ,͙<4w   l 4       ÜĜŜƜǜȜ_wն  ͙> ͙ ͙ ɕ   ͙T ͙   ͙ ͙ ɕ  W   q Ͷz  Z K ͙͙͙UɕOc ͙͙͙lɕElͶS*͙͙͙ɕ Ha͙͙͙ɕSj@IR[dm w{Ba # ')=A\B@"M #'ڌyLڊv;   O   ^ Ʒe ؐ [ Ͻ*   k ̶J Ͻ   { $ Ʒ+ QϽqg ?m9( Y*ZBTreeSet CJfr     ӫ=KVѫK k   =,&_48H/ )1(, ,1= H ak \JJ4J 5DJJ4JEDJJ0 JI{J ,Jv #G\KK4L+DLK' LKbK,Ku >\MM4M+DM|N4O w4NDON$O KdOdOOlO !?NN vt+OLM ,L,L4M p  HH!K  w_|RDR $RTRRRQI,R,Ro5fTWW $W,W|WX"` X f0 8P  ` h[    g  @`` "`)|XDX$X TXlXLX !?XlX !?lY   3 ԌY\YZ TY Y$ ' KTY Y$ ' KY  +Y Y ,Y;"YWXXYX,W,W W ?  ',X4XYUNW    ӷظCĹ] _^0\[[4[+D[ [ [ 3[(|[ c'[14[ _նMD[[Z ,Z,ZÜĜŜƜǜȜ_wն/y fQX9 T]]ӷ$],]|]$^ ř_ l^L^ !?^l^ !?|_D_$_ T_@_ T_ڤ W 6  ^-^  Ԅ_ "_( \&^k4_ _ `>  _0\_ ř  ]^^řǙșəʙ˙G*R{9 l],] ]4^U^,^ÜĜŜƜǜȜ_w?ɕ" B p+ӷڤm4`!  aa`UՄ!z'Hh`,` kddd $dtddle-  > ?1]9] ] A] I] I] Q] ]    fLff !] he ] xe&ظ&)1H X`] : g:<] ř}] ] !] Q] Y]\Bf0f tf^  )f5 jTggӷ$g ,ggg> g&4g̶; Dgg9dgϽ>h &i" liii"F!"|iDi$i$Tii% lii$e%D# i`|i!ԟi#i' i8Tj cd]]'ݘd,c͙c4d x!ۏ !]C m]Ug^yg-͙s eHQV\ & Ő  %\hleLt+ظ ӷL+̶ Ͻ{0Xg e NR |mDm$mTmmmmI,m,mi5f]]Lo_ 0.]I] o\)]o P".?toTo g odoxo?\o$o $o 4oXo  P I #( o< C  RW Q i  qv     o $ooto,o eKA Lo$plp4p e,Dppp A".]]p` 0x p?.]I] p)]p noo29tp n,n n $dd DebugListd dF:_DoP+4 @  <q  "??list | V? ?^*8t{wR`(p a q b ! >By\@q U )  5 nsp ( Մ!z'Hhqzt} %. 7@ {\r4rrr D5\rr  %L#(4Grm\r r) ,r %řǙșəʙ˙G*R{9 ?g1\r řs rrTZ lr Lrm[}OD^XuuB,r ?; ACCEL_CAP}#(y y-9Ay ]g  ACCEL_CAP442m(m! ,m#ARnLn#$p łp&p   $q#$n,n  n",n  + ̔]ٗӗڶPٗٗ$Rl .F8]]m F2.]I] _F)]x SF$ 0 ? t  .  F?HFf  ]]n  F2.]I] _  F)]x  S  F$ 0CC?IFj]]o F2.]I] _F)]x SF$0CC ?JFjl .F8]]$p F2.]I] _F)] SF$0CC?KFjt(F2)t .F8$]]$q  F2.]I]  _!F)] S"F$!0!C#C?LFj]]$r $%F2.]I] $_#&F)] #S"'F$%"0%C(C?MFj't '.&*F8&7+)t)((,F2(2-]]$s ,.F2.]I] ,_+/F)] +S*0F$-*0-C1C?NFj]]$t 03F2.]I] 0_/4F)] /S.5F$1.01C6C?OFj3t3(28F22>9]]$u 6:F2.]I] 6_5;F)] 5S4<F$7407C=C"?PFj]] $v :?F2.]I] :_9@F)] 9S8AF$;80;CBC&?QFj4 ? EE-Gb D# DG DDDD+DOD!"DE&D)A H O |g n u |    | }  AHO !#!(%/'<   J      ( 3 y   o   A   * c   5        [  Q#   I? 9 uG- ?h?  ?q?  ?z ?  ??  ??  ?? m?? ?? !?#? %?'?s3O(]+զA  g    r y     5   v @ G   u |  _f  /7h)1Ͷ)1% Ͷ *2( 6 B M4 @ ͶͶ[,(   *h w *     Ͷ*  %   Ͷ+7 g +1%   dxXaj8A MV)7B8u(&   p'(   (   б7CCC=C=U x9based`fBB RandomStateBk0Bk1Q *3ȱ;=> B(#[     x&$2  .ӈ7x~kq     ` l      I pS8AyN5>HK2;zH { j s|  h q}~  en~{bkx     KTu~III -I-I-ZvH#000.0.Fr  ( -9 D bl8(  +7ch uzfk  # &0*/ FKSH7 DNP   (- řLessGreater%ӡyřǙșəʙ˙G*R{9  &0 >IT*/ AFN 9 FP(    +!  ʅ+  7 pu      ʅaoz o=(=  p? *6bg tyej=    G|G $GG-HGhGLGG D|< ,L,$ řD  d ,t&"\ՓJl'$ řcl | l< W +  L,  d,  t(   #d , :$ @ $d, 4, :d,t($ ř$d,>$ :$ ř%d,>, : %  , g<łE U'c C ՓGNx{d  % '   8 #3 1<  W]L $  ř D   ,  ,   % N #(  4 ?  m l ,  , <   [ !-z l  <  \ , ,x   T A (  D , T  $ WJ$ 4 & D T | |   <  t z4   "  P 91#  N# !/ !# #m) #f  "[  !"-z%\,&,$%&$u)\,*,()*(7 @ ( .P / #h  w.`/7-  $,-0 $ 09, + .'+D2' 221Ծ31!5465 $ 65'K46\,7 ,3473"!l9'D: (| 9<:287)<; $ ;'#K; @ ( ?$P @ #h ?X@/>%  $=>DA'$A3= <& . <C ̴΅C,B'-B(7<D $ D')KD @ ( H*P I #h HXI/G+  $FGJ $J3F E, .'EDL'΅L,K--K/.l N'D O  (| N6O,M7#< P $ P'0KP @ ( T1P !U #h TXU/S2  $RSD!V'$V3R Q3  . Q!X ̷!!΅!X,W4!-W57<"Y "$ Y'6KY"[ ̢""΅"[,Z8"-Zs9"] "t""],\:"-\T;#^l"^&d\&&,&e?,&x&c de܀&T&c7b@&_&D&,&gAT&& $&QgDb$&fB&2hf&&hDCD'jfT,lj,m|,klmksE'i <'  it'zT-nj|-n|-F4.qq.r-y-.Jr3pG+-z pHoH#so/sIJ-zuom- u-.to[ !t-z0w 0t00w,vL0-vsM1y 1\1 1y2xN11D1,1|OT11~41~ 1$ ~'PK}12|1 x<1|}{Q,\22,2R,21bL2ۂ {rEzS21gz?3 2t22.U2/cV3 ̦3l3 34 3W333t3 <31X)3! ř3' D4ZY4 <4  ?z4 4t44.[4/s\5 5\55.]55D5,5^T5545 5& )_O52|5<5`.\66,6a,65bL6ۂ uFb65g?7 6t66.d6/le7 7\77.f77D7,7gT7848 8& )hO70|7,7i.\88,8j,87\L8ۂ uFk87a?8 8l88.m8/on9 <9  t9z;D;,;pT;;;$ q;;;*;u.r&;/  ?z<; ;% (tM'D< D< ;'<; <,<MCv:;( Xw %3xy< <<  <z<[ !#0zI$4 2 1 1U\c 1       k k k k# % '   > 1357mt {;(" %  ,D |  ł( < < m[}OD^ = F4   $  řǙșəʙ˙G*R{9 $ ! z iDZ  Q Z  80d#  ` 1xA;(  a#  Uv-|o (  :e~;TղJ( A$% n$& @ 6+DZA| 5<1sͶrF\WVp<33     | 7 Q <5  ͶL     A   ,7 s Ͷ | N e _  S m  & ^       ? y   ( l   universal_start   !"r{! n f   "T#prev_sid', =FgRN_bj ltuEmxEneeds  "? GQ$$## >##"b"*""!  $ Y΅΅G~3΅ ΅2΅?|<^ۂ 6̟uۂ @=ۂ %    |=<=,=L=($= řD=  >=d>,=t>(>d> >,>$> :>d>>,>,> 5< =% < ,<n @-1`6g$@ ř@D@ @,@,@ H".?@ m lD,D ,D DM ?L( |L>6?,=0#R- XM?R [R lRR[, RZ@RT_4T`_ T$ `_'AK^`TtT̟Z,d 3`ba a ,aٺ,a abłZs7tx{o Փ[GUdg %g'gg2g#-g1:u >|u (4v A wADv Bvw vvwMB7@)*vx  @G?*# C?/ C+,vx E?mv~ Eu w D?[ !D-RxG xtxxG,F.x-FK/yI x\xxI,H0xyDy,yL1TyDzNdzON z$ ON'2KMOy|yH}D,_?TDadba & ba(@M`bς6|ς[,߂_`^A-\,cB,ςfLۂ ^xcI]C]?e le0dE1dGF4 ?lRD,hHTDijhkij  kjiI4ghiDlτqτgql3fJ%˄v f?Rlm + m-LWmpDo tp -<,oPpEnN5͆r LȆq řrqLچt چ< t.sOSL*sQP<u ř  Ͷu׆Q4nR  nTe$f"g!>hiij rr t K  R Y   x|}xf e ,eeٺ,f|gł W Ͷ <|  `     ,}Ͷ<w4 z ͶSR m   A=К#   2 3 : gDEE~H$T[ `n+zF zT[Ŕ̟{w΅=yK*  ۂ  n  ̟      ۂ  Y  [ j 5S$ ř |<,L,4 ??PDɋtË?ͶD, *  G# /   mL ř  lČ,ʌ ,Č <ʌ [ " .P\Ռی,Ռ ,یČ $\,,| Ǎ& (Ԯ   \ͶLŔt  |܏  & (M ׏  ř. Ͷ# l\D ڐ! tڐ̟<6(Ր'  řɐ3  4 ??Pl4 ??PDɔ!lՔ,۔ ,Ք"<۔ɔ !" Ɩ$֖Ɣlʗ# ֗+ #-W#{  l$$!K&ŘD˘,Ř("T˘D֘)Ř*(+)*  +*)#4'()/T&'j%$+5%-΅-0,%1,G&/L. ř/.͙1t100'101(3t302)12G*5\504+ D,9,TD :4 |4<9`:;8--` › W8h?Lƛ7.t›l , ,;/< Lۂ 7;J60͚ 6GÜ=t=0<21<73<> ř ڜ Ͷ>@̚l@0 ?4D Dd ED & ED(5M CE֝% t̟֝?<CB6(ѝ+ A řBŝ7  A4 ??PƞGtG0F81F@9I\I0H: D,M;TD Nϣ4 |ϣH<ߣM`N;L<-ϣ\  WLh?LK=tlϤ ,դ ,ϤO><դϣ LŤۂ KOJJ?Ť J GQlQ0PA1PCB4 ?lPD,TDTDUԃVTWUV  WVUE4STUDXݥqݥSqX3RF%٥v R?PDZl, ,[H<YZ[YIKl\ + \-JW\ D^t_ ܧ-<ܧ,^P_E]L5$aL` řa`Lƨc ƨ<Ш c.bMQLШ*bON<d ř  ͶdèO4]P  ]RƉ!qx ǎَ ݞϧe̊  ,ٺ, <$܋DZ-N W C=LݎͶ<  -KcU Z  H  H &  !   O g  7   <  Ͷ ͶQ  x  W q Ͷ) <   B    <lͶͶ4}[:T@)4ӧpi  Ͷ_cfZ    КV#  E; 4 ?E H ]' `Ŕ̟ k΅W  ۂ }   ̟ E    k  ۂ   ,  7N ,ʛ Λ& (Mϛɛ Ǜ, & (M     ϬӬ Ϭ,ӬϬ Q$0CϬmt[ " % 8 yT] ?*? *<   ֫Ŭ: ܣ  ,DZ, 5wQTɬ q>Ec  ,ѯԽ Q$0Cm̰tа̰߰аׯ[ ". oT] ?*? *<   0 ܍  ,DZ" +mQT vg4;cԲDڲ ,ԲTڲD,T4 w^9\4 " W$ęt 7 " "$dz T 1 ͳ  $t1 1  $ԙl1  4ѴtѴ̟<5 (̴% řŴ,ݳ$lզ1 14ҵܾ ϵ4'  $ն¶tƶ¶1ƶ1$l1 ˷Dѷ,˷!Tѷ˷ !/t̟<· 1(5 ř<4 ?˲u# ѱ ,ޱϲřǙșəʙ˙G*R{9 vD{<X Mj < Lel}sp  /΁\3+̟:զwvS̟ID΁,TDd &D, TD d   Ǻ&   (M        4  $ںT޺ں1$޺Ӻ!$āt11 $ȻԶl1 λ D!t̟<5 (' ř޻.$lզ1 1 Dd & (MѼ$Ѽ$(ͼ) $Ğt1$ֽ!ýtǽý!1 ǽ1 $#l#1 "Ԭt̟"< ľ%9 $ ř%'$4ݾ ?Ƹoh&ʹ  ,$W^-< Jg=< FMQUYݎ]Q" ΁Jb̟pզPU̟l, ,<,L -l4D0'S  , L   GJNQ i4(@@&+8@IS(@  AC   D   D EFFG"p    @*07ENW`i irT]?H*3  1:DR `lu W:             }    /<E} MDMlM 8D' 0MdN M  8D ' 0N tNN 8 D '  0NtNN 8D' 0NtON 8D' 0OtOO 8D' 0OtOO 8D' 0P tP"P! "8 D!' 0M   P P L#9I]@=Q]A>]?@A?4фITфJ ڄ!5 JI!!]B ]]+)E 1<]D řE}] ] B!]=Q]>Y]D\CB H I1]GH9]K  ]F GKA]LFI]MFI]OLQ]PM]NOPN#"4ʅUTDž U %!]Q ]] :8T$1<]S řT}] ] Q!]LQ]MY]S\RB | ]]]]Vݘ},|͙}     T ( ͙ ɕ .    ͙M ɕu   n  ͙ ɕ  l @ ͙ ɕ. E $͙aɕ"͙ɕ}Q͙ɕ?V5͙rɕ3 ۏ !] ! =wZMF<ۏ !]o 3  =Ő >B LŐ MQ   C  { W       O + ` l   .=[SjDt E;,ӐY$,'-)Wʒ$В Tʒ J@,ڒR$,'-)W (-l$4- d:*!u$,ғٓ--/]+̕lѕ$̕ߕѕ4t:*!(t:0!˕G  $,OF-Hw@l$͗4tԗԗ:*!(t:0!G  $,OF-Hw@l$4tә:*!(ٙtߙٙ:0!G  $,OF-Hw@tȝ$ڝ ȝ P@,K$,?(&2:CM(&-5PX #owB     *H>CIT\dkr{  __self_discr __arg1_discr d $tL. d  *3  , \ &           d  &  %T    ݘ,͙4 |_chI c8 ?(  42  $ 6>~ , $  6>$ #/7!h"Y^jr hY^ jrP h <A  S [CCC=C=^c u} $ 6 >h ֍$ҁ0 8h  # Bҁ     ek ~rxҁ] OUbkV\ҁir h  0   ekry     }Rυ>*mYE    "      __self_1   __self_2 D4ކ$   b v>uҁҁKҁ5ҁ (  42< X =`2 >: G @<S ?@   =>? ; ; ,6 (71 48< I:6U9:sy 789 $1 233; H51T45  !2'34 : &!'h"() n*v +&w, -'.0(/0.4 B)H*+,-./ r h nv  w! "#%$%fl z !"#$ PhQYfZbo{[aou 4< IU  -  P Y 4 <I  U       h ֍4<IU6< JP c ~Aτ1:C# S Jυ6"nwclXajυ@         ٜ       -  n     ކ d&T'ކD( % 1  6  <    ٜ  ٜ , υ Q υ v    y      !'-3?τGPZ%H+: JNτ[Nh    $##OτOPτPQτ1 Q? ORτa Ro Sτ S Tτ T Uτ U Vτ!V/?WτQW_  2XτJX .Yτ?KY X LJd MKr w LM#,ZτFZ .[τ?G[ X HFd IGr w   HI#$\τB\ -]τ>C] V DBb ECp uDE#^τ4^!_τ5_h"`τ6`aτ7abτ8bcτ9c;4<7$*;7<=:Fdτ&deτ'ehfτ(fgτ)ghτ*hiτ+i-&.)$*-7.=,F Pjτjhkτk\lτmlXmτi m"# "#!nτn.oτ?oXdrw # pτp.qτ?qXdrw #h rτr֍.sτ?s֍Xdrw #!$% $%# #,~0'1* υυ01//2(3+23# #:">5?8>?= ="!@6A9@A# $#          ! ( / 6 tB τK τT \ d k ττy           HQZcd?vQ vυυ6?H! υ-6 υ   υ             ٜ ٜ    ٜ ٜ ` i r { ττττττ τ)τ2τ;τDτMτVτ_τhτqτzττττττττττττττττ τττ%τ.τ7τ@ J Xgt o__arg1_0 0=     __arg1_1    &b 'ކA ( ) *__arg1_2o + $ 4 5ކ 6K 7* 8 9  B C  F G y J Kh  R Ͳa  / @  l   uřǙșəʙ˙G*R{9 ]b t|xCCC=C=W\ nv؀z ) 2~$ CCC=C=CCC=C=řǙșəʙ˙G*R{9 KKKK 7K7K736$*  =G  P̂  hȀ 09ȀÅͶ*3(s    ss $ s     , 5    ͶpyKV DZ>#a ho(s ux   ؁z    ~$ P̂hȁȁÅͶ#      (sCIP^gpymvX aCL-6 !  HQ[i w b E(           B    $     WfwT]?H*3%u  x  ؂z    ~$       P̂   h  Ȃ    ȂÅ  Ͷ    8s   ss  $ s  qR4Ͷ 9gG  )1 )1% *2  )1% řǙșəʙ˙G*R{9 ^c  u }! řǙșəʙ˙G*R{9 ^c  u }(    <H S8Aº : Fa1(        %      % ! #%    ERdmz  1:  º  ºlNF B          4CT aj )     %    %   !   8  <HR:" (  (4 ?HR( '.6 RZ$qyC   $   ParserBuilderasthiriNbg yp  08Ȟ 000.0.FrTY ks 000.0.FrTY  k t 000.0.FrߧTY  k t# 000.0.FrUZ  l u(       % m|? K @ L x  R9#*(   p    ȟ       # $    9FXao     |^@+            (9 FO   CHirprops썛 -c@   PD (- JY^ Qppʽp opopo2Ы?    $  \ ( ,  1< ,  "CCC=C=gpT 4     D O ,   T   <  _Parse Translaten`    ܰ    %ÜĜŜƜǜȜ_w   {    m $  M F    [ÜĜŜƜǜȜ_wfv %ؒ     $ gԤ    =   y , D D ʋʋřǙșəʙ˙G*R{9  ʋ , #m[}OD^ؒIÜĜŜƜǜȜ_wȏ ParserDFeCؘ.q|ؒ   ( w hirs  Z`ċ܍Վ5ݐCQؘؘ}ؘ <  $ , t     ,ّ  , exprsSW\   " \Ӣ * 쵢  +"/ ?M\  \  0l + \     1  *  ) 4   R D ,  "    '  ?Md ̣ T  14 d -!   "11RefMut2.2.2 .뜿Er Usz,   U<  Ĥ dΤ Τ Tۤ .dۤ 'Τ:  áKT\  D  0TT fD  GT   d ̒ T .d '@  áKS\ ̸ \ĥ ܞ .Slĥ d\  ES&  ҥ dܥ %ܥ& T %/$d (#$ܥd  á#L"R\ '   '0&R f  "&G!R0   $dʦ -ʦ& T -/,d (+,ʦd  á+L*Q\ /   /1. Q gt  *.H)!Q0 ʦe   %ÜĜŜƜǜȜ_w)("Qy~0(0$%#?  ,' 1([ؒ"1.&QN,Md $  w, 4(4 034 5   U   Rppʽp opopo2Ы 352) \ 8# Ƭ  817*Ƭ ,7,+d 624 - 6!-D :6:3.t <  $ </L x   %ÜĜŜƜǜȜ_w<;0w|=;=12#| ?;m 9? : >;[ؒ ">.M D  D  D/C4! BC֭7 Tۭ  $֭ B3,    A5Pۭ2 ֭8  Ao@6P %E@1E78#֭8 G@m9G  C: F@[ؒ "F.9PNCMd® $Ȯ ,® K;4Ȯ Ӯ L  U® Ϯ  Q KLlJ<Ϯ  T  ̊̊MapϊЊD^)cO$ JI=   IH> %MH1M?@# OHm R9tձ SO1 ,  $ RLS=QA,  NH[ؒ"N.M2  QPB$TP0TCD# \3 T \/[Ed  UP[ؒ"U.McZ[o   "ɡZYFO$t ]Oв ^9S,  Y]E^6XGO. p   %XWHO&+_W7_IJ#p aWmVa  [LN `W[ؒ"`.KOO[Mų bV<³  bϳ M  γ VV[VCVq ,V $V" V# V&V     ʦe     [    W$+22   Kcء , ͜,ȡ )OXxZ f ɡɡɡ}ɡ'G!*jɡs4m[}OD^ؒ'   ɡ8Ͷ   ؒ d    \ D    1 m[}OD^ؒ; r  ؒ    7 f AnyCharAnyByte AnyCharExceptAnyCharExceptLFAnyCharExceptCRLF AnyByteExceptAnyByteExceptLFAnyByteExceptCRLF􈨃6 z ؒ     c B ؒ j   ؒ m[}OD^ؒ;D ɡɡB7tؒ}Dnr Q1y? all_anchored26unanchored_prefix9  > ? r F GD O d N O  U?.V `aw"E wvlܞWáɡ- áɡ^>áܞɡqáAOɡáq֩>ɡOɡ]ծg@G/KY g  o     T ̻[  ɡr  ɡ  ɡ ɡ6ɡɡɡɡ T 4    D  T  Od \Ӫ   Ӫ \ߪ ܞ@lߪ ) |  ,  t  0 D  \5ė ի |۫  ,ի ۫   ի2 D  Z 3   ~       Properties 2AHQ, 61ܿοKh]h9)$¡ Eܞ]3$ ! l  !0|    %ÜĜŜƜǜȜ_ww|? m$ "$  řT 4   D  [ؒ4# %ؒ9X ,  "     ԟ    `$ 0   m$ݯ #ݯ L  #0 \   [ؒ" .ݯ   %ÜĜŜƜǜȜ_w w| ݯ m$ $< # , $C3t   [ؒ/ #/I$  $ 0 $ %Lʰ  t  %D4̻  [ؒ".  $ 0  mL  5   9   ԰ [ؒ".H  /  ȯ y   $ x    y&֮   Ԯ m[}OD^ؒؒcgřǙșəʙ˙G*R{9 4ؒ=m[}OD^ؒcbؒk>ؒ &ؒ ~ Yv*e  :one    *t̯    ? z   4/Y=18J, G G L $  Q/=4 .| ?4 8  4 ="l 4  $ >=D Vb<  m  L $  ) =4   [ؒ" .?=d  |   4  8  4  ="L    7T` < m L  $ =\  [ؒ ".@=  =$0 m L  $ =\  [ؒ".A=  =$ 0  m\   , $ , ==̞  [ؒ".B=  = % 1 L $  , $ $5 #=  ![ؒ"!.C=ؾ  #"=$%"0%d ) , $ , )>(=  &"[ؒ"&.D=  ('= %*'1* !T .ܑ , $ .5 -"=  +'[ؒ "+.E=  -,#=$/,0/$% G 2   P2@1&=% 0,[ؒ$"0.F=f31 '''53'54(9'64'6*)4 '748 ( 74 8+9" 35=<(98H(9,-< (;8mt )>;! ), $ , >>=.9  (:8[ؒ+":.G9" ) =</9 %)?<1)?01d )C; ), $ C5 B29į  )@<[ؒ/"@.H9 ) BA39$)DA0)D'4 )EA[ؒ^3"E.I9x! F %  F  8l37=G8; 3= Z8i:yG:< 7 7 7777?7| 7$7u+6/636 6               F   * 18 x   ' !kr#y %?F'} ( ('*.(,n)u.|)" 0)2) 4H , 4 řǙșəʙ˙G*R{9    $ ~ k  ؒ   Ӌ $  u  ؒ    ) ؒ2   z o ؒx  L    ؒ  ) H 4 ؒ=    ؒf ؒ-Z?HӋQ'ؒ(B(%D/ؒ8)ؒ){!,firsty    z second4   " v C,       3   !? [ &?  +!?  #0%?G '3  (7D ); *:7 ,;  .@0? 2E4?(bbp+e   D  R  :   t    } %   k      dk(  (-: ENX(   ȭǔ  ,   ,  8 qv   [ith5 (   ܞ  ,    / x   ,  8    {  ;H Ô/ia { Ybkt#." W۟  08ȯ CCC=C=X] ޟox(   AP\ g) 5ޟ ޟ h2(   ۟  ,Ȱ    ޟ,  8 qv    ѡѡ^lwkդ8Š *2 řǙșəʙ˙G*R{9 _d v(   AP\ g*3 fѦ6(  Š  ,  -  9 rw  ̧ _mxl9H 8͡ (  2> Ipznjnj(  H  +8͡   +  7 pu  ۬  Ygrf4Õ Õ/7( -9 AÕ Õ]g( 6  7Õ*6af sxej  ԱԱ ߳ !( 07 J PƘ{  ` K ` `8 )6  ` ܗ` ܗ8 )    3 ×$ 6 > $ 6 > ȏ$ ȏ6 >   <  ?d  '.5<CÕQԱZb ȏȏ$ s| R?cOQY     o     p  8 Q `z` ( 07*?NS `*e mP%Ƙ&1'9F)%R() &'( ` "!+3";9G$ S#$lr !"# ``= NVcX`9lxFLZ` }`ܗ`ܗOWdYa9my+1?E b × 4 @  s   4 @  <B P V s d ȏ4@ s  ")0+>ԱGPXi  FX dR?mvcl u~k/ _nqz  2; G       ~ | >  %Z e v z  ) _   $ CCC=C=W\ nvH% (! ! 0< G  q{!! (j  k  Κ- s  Dpqm.    ppʽp opopo2ЫΚw p    +  r!  ,8 ;A   k ! ԙs - C 5 !Y } $ 0 "-" ,`փ CCC=C=ΚX] ow CCC=C=[` r{(   AP\ gk w 5փ       8   AP\s] (  ՚     Κ0-  ppʽp opopo2Ыx   řǙșəʙ˙G*R{9 r  ~  u{ {FL     _h  wn|{D (  (-3Κ@ ΚKT^ 1099511628211%#"14695981039346656037`# x# % CCC=C=W\ nv(  3?G    \A& (  `#  *x#   * %   +  7    z  1>GPYIR [)2;  ' l `' ( )1`) ҁ)2(% % % AP\ g C(ҁ (%  `'  *(   + `)   +  7    z % 2?HQZJS \*3< $ i `'   (   `)   8% % % AP\zi`B xC D CCC=C=͞W\ nv(@ @ 3? J  Z?$ (@   `B  *xC   * D   +  7    z @ 1>GPYIR [)2;  ' l F ҁ)1hF XF (E E 3˝? ˝Jҁnjnj (E   F  +hF   * XF   *  6    y ~E 1>GPYIR[(1:#h F  hF   XF  8E E 3˝?r\F E  #,6hF &7XF &7F  */ ; @HMUbF HF /7?e r|`I I ˝ ˝/7`J ҁ)1(G G 3͞? ͞J ˝*ҁ (G  `I  *I   ˝+ `J   +  7    z G 2?HQZJS \*3< $ i `I  I   ˝`J  8G G 3͞?˝s]ڍ 1099511628211%#"14695981039346656037PD KKKK 7K7K736bg y(> > -9A ]g|Q lQ  Q$ TQ CDQ Q  Q Q Q~  Q DQ wN4   "1"'    C\,   M ,܆ ,҆   'K4  +!ݏ  M < , mA4  +!  Q <Ζ ,Ė mA4  +! (  Ѕ Jץ tץ Dۥ ڰ  (Tۥ \ӥ  wKӥ 1 4  XD    | ,  ȏ MVvTWЅ mڰ A{4  + ͷ  O d , pAl 4  $ D $   N 0  M  4 ,ɿ п "-$W۝4 l +t CCC=C=V[ "mu  ? D , 'L̠4  +#T  Pd    K  T , \/W4  +!  D $ , i?4  +!  @ $ , i?4 \ +! Б  9 $ , i?4 t +! Μ  ? $ , i? t  $  ,   $  ,0  T"  o $ ,ޮ ,O4  +  ռ  O d , pA4  +!   I < , mA4  +!   W < , mA    4 l  + !    CV,      (     - j   - 4  +!   L,      1   -   .4  + ! d  O t   .4  +!  .   , , "rV"S S8~g* yMpNp     Ȏ  ؎          Ȏ      xދ    P    h    Ȏڏ    Ǒ        ½  8        % hP2vX?(=ϻpV<"½½I PW(   3ސ  %19x hڅ ` ` `ڋ `rwҁh h >CҁOWУ Boxljk#di  {P h >C O W  % 7 ?  % 7 ?  %1 9)  M $l Chnҁ {     ҁ     #*18FΞOWu~  '0 DMfSυ@υ-wdQT x    r  %  } 8)ކ__self_3  ҁҁ/`ҁ-ҁ    "(.4@ΞHQ[#H ,; KIΞ\Ii   "!!JΞJKΞKLΞ0 L> NMΞ` Mn ~NΞ N OΞ O PΞ P  QΞ Q. ސ RΞER/SΞ@FSY GEe HFs xGH!hڅ TΞ2T` UΞ3U` VΞ4V`ڋ WΞ5W`XΞ6X`YΞ7Y`ZΞ8Z`[Ξ9[`T ;2a <6p v;<: h \Ξ)\h ]Ξ*]^^Ξo+^Y_Ξj,_.)/+./-Х `Ξ%`.aΞ?&aW'%c(&qv '(! P bΞbh cΞc^dΞodYeΞje!"!"  fΞf/gΞ@gYesx ! hΞh/iΞ@iYesx ! jΞj/kΞ@kYesx! #$#$!!-0*1,01! ! :  >3 ?7 >?= =  A4 B8  AB@%- @  C5 D9, 2C?DE!NT ]k !           & - 4 l@ ΞI ΞR Z b i ΞΞw          bk w*3 ?H@R,bktNW.7 CLDV0foxR[+  υ υ O + υ υ q z  ] f o I R [ 7 @ f o x  Ξ Ξ Ξ Ξ Ξ Ξ Ξ ΞΞΞΞ"Ξ+Ξ4Ξ=ΞFΞOΞXΞaΞjΞsΞ|ΞΞΞΞΞΞΞΞΞΞΞΞΞΞ !.     m    b %o & k )I * +c , 7 2 3ކ 4 5k 6K 7+ 8__arg1_3 9 < EI F ] = U   ;  ?    , \   4$ǜ  ,  ܜ  ϙ $ Ξ,  :" ljk# ^c u}(  /;C _i(  "  -9fk x}hm   '   '1" ,1 HMU H 8 EO  D  <    U4 D  X e#    | , D  ڌ148s9  $  d  /  |   K  $    Unique ]8"O 3f*ہ  $  w4    Pppʽp opopo2Ыot  Q      <  t      d    $  6  4  d  ř $ ř& 0  || , $ 3k %QřǙșəʙ˙G*R{9  1Yā2 1_7>Մ +3(  /;C\f(    -9fk x}hm  $   '1 ,1 HMU H 8 EOȍ  D  <    4 D  X [#    | , D  ڌ܏148s/Ώ  $  d  /  |   zK ܏ $ ߑ \ $ ##    ]8"O 3f*ہ      ړ̔ $ l     ? 5!($  ř] T  ř*    | , $ $ 7'!OW̔_̔gpt vr#p# #  l    QG!  ! C    UC T ###&5 S\di#$ $ d ,  $ < d , ɝ$ < x     <   U;l *; 4ɝ$ɝ(  (-3@HQ[   $-7h  h4  FX ס)@6  h4A  סTסޢh '8X '8  +0 < AINVc H 4<D f s}(     2+*  !!! PhantomDatao$z - `  l  $ 1 ,  /#U"wx Zc  èr{l6  #6%  <6 'E# StateID::ZERO̸7  7  7 570  <7 '_ B CCC=C=֨W\ ٭nvC ٭) 2D" 000.0.FrCCC=C=  E$ 000.0.FrCCC=C= Utf8Range^t İ  G CCC=C=ΩW\  n wH CCC=C=W\  n w(@   @ @   $ 4 D ߧ ߧ٭  ٭٭׵ İ İ   4? )H^>İEL٭S٭ZJ CCC=C=W\ ȷnv(H H -֨9 ֨Dȷ ȷ`jL  /7L ҁ)1(K K 0< G s}ҹҁx ҁ*2x ҁ*2(  2Ω> ΩIr|ҁҁ(  x  ,x   ,  8 qv   [ith5 ҁ*2X (  2> Iq{ҁ(    ,X   +  7 pu     Zhsg 44 ҁ*2 ', >F8 (  5A L ҁAnj (     ,   , 8  1 *  6    y ~ 6CLU^MV_+4= $i 43 ', >FP (  2ū> ūI t~(     ,P  1 *  6 ot‚     Ygrf 3Ƃ Ă  $-73 +0 GLT[P '8@HUт HȂ  O \fM(  (-: ENX( 'Hτ ΌL T3 ?  nH ک 3 ?     q"H ©FO5A| t   2V1:r g  {  (     '1 H ,; K\ i |     Hτ Ό/@ΌYesx  H ک/@کYesx H  ©/ @  ©Y  e sx   ")0!<ENV^es{:CLU#,5>GPY c q 8 E   *(  &+8 CLV   &0@ */ FKS H  : GQP    (- ř [gow @ */ AFN   9 FP(  &+1>FOY   #,6 H +: J[ h y" ګ->ګVbpu     -> V  b  pu    JPW^eqz2;DM     &/8AJT bq~    6P  "' ř Uaiqy   &49 JOW\hw           & "  ګ- >ګX] ns {   - >X]  ns { KQX_fm{ 6?  "+4=F P ^mz   kx&256(F F &+8@IS256DH  H H H  GC \G s-256 I I  8J   # J  J  ;8J   P exhaustedݠ PsxJ;  Q J  a ta a+ 4a  SDa a6 La  ? a  ` $` ,` Ͷ!E e  \e $ .IIJe \e ue  e $ 'Ke  e % 'Le  c ,c ,c }Be\j dj $j tj j' j  ƨAj j8 tj  ɧCj j |j ,j  j 4j =Dj j  i is ,i ,i ,i + 0cH'm dm $m tm m m m  mM $m ,m {3o%  o o  n]   :f(n ,n  h,  \  $3l 3 ?     т $Excluded"$tׂ \ׂ   $  %Included"%t \   $ׂ \      t ׂ& 4   YD ,ʃ  ʃ LЃ  $ 1  \Ѓ 1 ă  $  ř  & &t \   $  ' 't \    $옄 \  B  t & 4  YD ,  ř ̄   ř  A ]I ]   ] A!]I!]I!]Q!]!]!!] $]$]* ( 2$<] ř $}]$ ] !$]Q$]Y$]\CD !D " $ # řņ  !"# Ӈ ԇ  * *ց % =  c ڄ+ !]!](ƭ    `HH䐁 , , Dځ  Unboundedn'  ' FͶͶD  aͶcA ۏ !!]3 ]Ͷ&/ AC w Ő CT ħ;  ۞Ws O  ||ħ ħ3         &+ , , (Ɵ     ,@  U *  6 ot˟     _mxl 9(    ,x   + ʦ   Ͷ,     ,8! /5!'   h  |[dm9B  QZ4BMA (  ˳  ,X   , X  l *  6    y ~ 6C OXaPYb.7@  ' l (  ̹ -ض   řǙșəʙ˙G*R{9 v   ! = J  A Ϻ҉}nw   )8 *2(  /; F_i(   8 #PU bgRt   (  (-:BKU(  `  ,8ej w|gl     )`   8  /;gJx   '1` +0 GLT H 8 EO2(  (-:BKU2H   ,8  /;tJ   '12H +0 GLT H  ; HR2 d $ t  L  ,  T -1          t   %  / J      4  ; $- l     t / |     ,   t % / J    4  ; $- l t  ) |     ,1]9]] A]I]I]!Q]"] !"  *4 +  !5 +*!]# ]]+ )&1<]% ř&}] ] #!]Q]Y]%\$B ) *1]()9],   ]' (,A]-'I].'I]0-Q]1.]/01/ 64  6!]2 ]]4 251<]4 ř5}] ] 2!]-Q].Y]4\3B    ]]]]7ݘ , ͙ 4 4 @*͙ɕC(͙eɕ#  ~ۏ !]   <vYLE;ۏ !]n 2  Ő # -Ő .2    o@8| , p    LS*(     Z+@   +  7 pu   %  i'>GP?HQ' 09 t(    Z+@   +  7 pu   %  J:f>GP?HQ' 09 t- /7.)2>CCC=C=řǙșəʙ˙G*R{9  (+ ++AP͸\ ͸gzU aDmm ٙ?0T$d#, F,$  %řǙșəʙ˙G*R{9 jo{$m X?0TϘ Dۘ T#Խ "#1"")Ț w"]T̚!|Ț#4ؚ X! Dؚ$  %řǙșəʙ˙G*R{9  jo${$ !$%m%  %%9"%?0W&c&#$%'mT--+,%?0P,{D+&l!4+*'D" ԇ*A)($.)0.)*"/)m# /3(+"?0Q0(]0,-#1(m\  ',3.,&?0Lț ̛1ț4/,̛+ 2 $341 ř2ڛ0  )Ζ   k ʘ'   y t{<CJ#n3:!A#%!'5d,,  \  ) [  m[}OD^ Ͷd   Ͷ | Ͷ U Qm[}OD^#hym[}OD^?m[}OD^Aڇ|řǙșəʙ˙G*R{9  -  :Z6U֔ڇ#k` TXҖ  ? slot_startr slot_ends   ?   ?L  L ?   ?d ,' x?% ?i' 1 "!?}#/ %?'1%"d  v  h    j E      ԇz   B  P ԇ`u\$ $4< R   ב˰8L<ȩtȩ$Щ Q??Flatten? ?#Tz=d4ЩLީ  ȩ 4ש W =שǩ"  D[.4  p   U ݞ | +    u䍫   ٨,Dݞ\Mhâl|matched whole_match\ group_matches &ޝzמԟe<t$מ.4ê 04ԟY7!D$+$ ^ # R[<fݎ[מiԟ/ _ref__matched_ref__haystack\$$4<ޝCL<t$מ)4L   4ԟY 7"  DW04  p   U | +    u   Ӹչͺ̱,Dݞ\ٹT %ʬlѺs|~\^ ޝמԟjB<t$מ.4̝ 04ԟY7!Dƻ$λ+ϻ$ܻ ^  -6<ݹEݎ[tמiԟ/ $ ř|T "2")D w[T  Pppp opopohEC8GotG  Q   $  $ řd     ͸L  řǙșəʙ˙G*R{9  7#@d  ʲovt ULI\(0 I,o9|ʲ)1 YLp9ʲ7<< 0{?iŷ, Z] (Ȅ ̈́4@ K! - ~ fI(τ  ,7 -  9 rwԄ ֽ _mxl9KKKK 7K7K736ch z( /;C _i(  #PU bgRt    8 /;gJxUtK J)I1<HDOGWbKj   CP+   Cd`fAa   C@&J   tK  L P AN\&L QPVN\&  P<< 0{?iNej     Q  ?    ?l 4I  tl L# +NY< _ k!  L*Z  %ÜĜŜƜǜȜ_wLy~ Zm4I L PAL\[ÜĜŜƜǜȜ_wv %ÜĜŜƜǜȜ_wQLL QPCL\ H \  $ Q<< 0{?inJ4m    řǙșəʙ˙G*R{9 {    J* % >'#  >&( / 6/ 6 6T ! | !, J\ H  #X   P#@"J%* $" G &$V [&g%=pu '% '  \ (%T G  )% | (\,!H% $*J] . /( -  U./ < _ ,-+"Hf   %ÜĜŜƜǜȜ_w+*#Hz 0* 0$%  2*mT 5T G  6)5  5R2,6W4&H* 1*[M#1/RHh6   %ÜĜŜƜǜȜ_w43'Hx} 73 7()  G  83["8.SH H T I] X T; ;+:,N :C9-N %<91<./ ?CCC=C=&l C?x>0N|=9["=.TNĺACCC=C=(l CAx@1N|C(l CC9B2N|TJ FD 8FBE3N'D E# D J <( G5C8 GAO $EJ EJAHJ?8 I9F I   :N;PJ< $JW J5  N APg )5=J7L[H7D 9? )= G=GLHLI?I?KP JPM2BCKBO8494: 4 P W ^           ! (/ 6L!t]J,řǙșəʙ˙G*R{9 z m[}OD^;BK|řǙșəʙ˙G*R{9 p <$Q          y '  m[}OD^D  ÜĜŜƜǜȜ_w g p y   B m[}OD^ U   0 9    %   s = `1  pattern_groupsx}y pattern_indexgroups  e groups_iter  $ӎ ( maybe_name )group, 2  1a 2  8?  =?5% _Q +M  l G  U   k ]1" /yM  R+ pg@ v } \1 _C(  <E1ӎs _ref__pid_ref__group_index0 4  " 1"'(  "V# tDڰ  ( Tb w K)  %řǙșəʙ˙G*R{9  mr ~ ) mɑё  <ɑ$Ց Lё %Ͷ?# ,ɑ4ۑ X Dۑɑ0 W Uk?  waͶl,ʐѐ$m[}OD^#  0 w  :EřǙșəʙ˙G*R{9 řǙșəʙ˙G*R{9 υqu xindices<  ??   ڰ VZ dD *T  υy4  +!(  "P# tDڰ  ( Tb w K)  %řǙșəʙ˙G*R{9  mr ~ ?) mË " l % %?#d wg\   %řǙșəʙ˙G*R{9   ns m*Dϡ X:Tϡٌ͡?ߡ wkqwdѠ,\m[}OD^Ë8  "߄+KËm[}OD^ύ8  E+OSӎXl  j?  ??! d6ڰ Dٌɍٌ 6  ,"[ K tDڰ  ( T\ B WËppʽp opopo2Ы ËÖU?Ğ욮Ӯ   + ےLӮ$֮ w [ 4֮Lɮ WÖ5 Tɮ  ,ǬάÖřǙșəʙ˙G*R{9 Ö߄ÖYے; ڰ Ϙ  ( hFOWt<&$&4 ÖË<ےٽ\޽$ٽl޽ɽ#|ɽ "F"$ ř$ ř  ʼx Ÿ Ѹ   ?jU,ۼ ?řǙșəʙ˙G*R{9 )s$ $\; W Uw?Ԏ ,\Ͷ),ӎ0kˡ!L & (M  3K,yL $t"$ ř & tDڰ  (T4+  &  (M  ř ,    /, , D( T & ( M & ( M*  -0 ř pͶl,\~D( ; ߄ ,&υ) bӎ (ڥڰ 4+#\ Pl G,,߄@گ4+#L Pt V,,߄'*Y|D $T !5 F F H,,=4+#t PI"DoT ܀R,,υ߄,P\$l & (M ʊO,,?|D$TĠ $8t uN( Εq,,>ù  4  + ! ģ CV,쾙  !:     - ^   - 4+! CV,   2  -  .4+ !ԭš CV ,šʚʚ  9 - .4  +!! !.! ",ɘ,$JYǭ6߄?_ӄ\߄eD߄4߄C k6ٻ,"( * -ߤ"  - )   -     +7$ 28$*   k߄ǭ{ӄXaj5> GP.<Gǭ;ӄ z *   ߤ"  !)  8 %gvC#=] \)\1,̸ #6͸;/䷸ tD ) ;T  &  (;N #| x ;!R  υ< _   ;[  l;!&2 EmݸTT ![" %8<;$\ $l4 0 3D U 3, ĺ] ֺ, 6 80$, "+8#ֺ,d ֺ48J,!!" 8+%̠  T8+*  03$$T$%#8$ #L"8,', #'6&8/,)T )/(8'84 &(`,%8 ʼ,,ܼ #,6ݼ+8/$] .T- ř.ʼ($ M+ʼ-H*8(& 18#02"1 2"18.0243 0' 3(08O/3½Y4/ 4/884"  58#765 65 886]&;9,9,8";%$: ř"$8:;7#;T(&;$;=<; <;%;8<&,@D @*?';T1 ]>?$CLC)B(;t9 ]AB ]= >AA ]D=I ]E=I ] GDQ ] HE ] FGH F*),ODO(N+;T !] I ] K ř]]  ] I!]DQ]EY]KXJ;C1 ]MNS, S6R,;0 CRQ-;,9 ]PQ ]L MPA ]TLI ]ULI ] WTQ ] XU ] VWX V/.4 ? 2 !] Y ] [ ř]]  ] Y!]TQ]UY][XZ;C]1;^^52;\26 2% 36 ]37G\4:Y4s__94 1    # z  ] ]  ] ] `ԒL,״řǙșəʙ˙G*R{9 T  Z{ #2υMυ    ^ $4  Q G  ӄ     ' 0     W x $ > /    k    1    @ ! YI0 >υr߄fY&RHۏ ! ]   7υQMË ۏ ! ] d  f ~ P "  DŐ  Ee I  TŐ  U Y'Oy1}RX1 g 1 $ M A   G r   R¹,ǹυǹD͹-T͹6yG&  5υ>owυl _ref__group 08( /;C _i(  7+7di v{fk  %(+ `III4 Khm y   % 1  9 ȶ % 1  9p=BNVܗ % 7?   4 &  S E  r  / 8   R[: (O:  |k      T(+`nv QYfY an"z!" !4=5=JV ȶ 5=JV! /5H p   R Zg  Zbo {   1ܗU^5=JV  #"+3E  , ~9B-6 ?&   Zh @ 6 %    u  V  * ;+Ö Ö19( 1= HÖ Ödn( +2Ö,8ej w|gl  &(͎   ,  ,     ,  ؏? -9# 17#)   jҎ ~\en:C  R[6DOC(!!&+8 CLV 10)\D ($T* \ D  ( $T$4 $$ Q$05 *  ?  \  , H?$*D 5  $ܢ\  , ɕH? $*\D% $T<  $4;4 XU, ?t $q   D  5$l \    ,  PG$! ?\  D!%$T!'ɌT] ?)?$+\!"!D!"%!$T!!$d! !? $!&# !*#"& "D""&5%$l"l!ɕ2yɕb|t# ߫UͶ1 LFaزy߫s3L replacementdappendnsx  cap_refͶ  Ӳ(  <5  8* -2#  ֹE ރ  ڤP ڤr( ڤڤ"ރ  3L3$3$3433jt(;;&+1> IR\(;  `< +P<  *  6 ot;    Xfqf 4;; #,6`<*/ FKSZP<&7?GT;H;M ZdH(==&+1>FOY(=$(=G O1 =  {   l8=1=  n=5;B PYbj { i{  2    g == #,6 H=+: J[ h y(=->Vbpu 8= -> V  b  pu     =JPW^eqz 2 ; D M  #,5>GQ _n{    39"  )1;"Ͷ)1(8" 8"2 > F  Ͷoy (8" 9"  +;"  Ͷ+  7 pu8"     Wep e3 Z"Z" $Z"4Z"ĮZ"5Z#"WZ"[" "`  (["  f4 = Z" \"y    [   g[  @`` "`* \"Y"TZ"Z"u| Y" LY"4Y" /,Z" \cFR Zt sI Ŀy"\y" $y"4y"y"5y"Vz"z" "`  (z") 2 y" {"n  ~  V   gV  @`` "`* {"x"Ty"y"ip lx" Lx"4x" ",y"}\U8D Lf qG,~"( (4"\" " 9 #O1:"  V# / ?" m   ԇ" " [ ! - "  " "T"4" &D"$" řD"   |"" ř"Ͷ'"Q0 ")  "4"( 3"#" $"\R ."4"  Y  %1 4"m :""[ ". """$" řD"  |"T؁"  ""T"4" "(!D"؁ "L" i!G" "%T"4" %($D""# ř$&"Ͷ# "'<"& ř'\"  & ‚"((Â"}">M\l(3 h~"(d"  d") }%" L}"T}"Y    O5 R  $" uͶ_ x & q  $" |P1 KͶ  ^ ej         y ( .5< ,"* *4"\" " <H(1"  % T%* 6 ?" m   Ԅ" " [ "  . ׆"  ݆" "d"D" *T"$" řD"   |"r  "<"" * ,"l" ř"Ͷ$" S1 #*  "4"* Ç."ȇ" $Ç"[Q ȇ)"Ç/" X %1Ç/"m 5""[ " . " " "$" řD"  |"T݈"" "$<""$(#,"݈"L""h#F!" "'<""'(&,"l"% ř&#"Ͷ% ")<"( ř)\"  ( "**‰""?N]m* i"5<Cd҆" d"+ !" LЄ"T"m[}OD^   ȁ . 7 y$"    MͶV۔I 4 = $" &]۔ͶJ۔d      ` U   f ms" s<'" "",bRrqSȕ" $"4"#" ܣ"",> x¢" $"4"*" "",:;," $˷"͸$ٷ"4""%( T" T"  ","  .<"!"\" B "t"*"4" ZD"@"L"H5"4" \"  "  < H (1 """  %" L̼" ߼"T"BͶ ȁ   og   ?  "d"D"( T"*T" T"  ","  .  <"" ?? +!"\"  t "  t"*"4"  Z  D"@"L" I 5"4" \" " < H( 1 " " " !" L" "T"5۔OZͶ  )ȁ" $  mJ  rAH" "W" %pn"$" f<" |"," 2<""K" l" D" <<" "*"" ȁN=""6"L" _Z\"""řǙșəʙ˙G*R{9 ÜĜŜƜǜȜ_w%" L" A xȁk'q@" $"d" ":""1$" 2D"    G","\ J"D"  d"     r"," h U""4"MD""""," "(} 5& /y{!  uI\"  # N(" Ȫ"#PU bgRt"   L" ","3<""M" l"ǵD" }9~ 4"l"$" Q54"p"  %ݬK$)5?t"m"" % ?# " "  "`  (" * 3 "<" ř"     @``  "`  * " " "" " L"m[}OD^ݬ%Fv,"k[HYska} ?U{\   5" "O"ν ij"Ě$" f<" |",".<""K" l" ĚD" {<<" "%"\" ȁN="1"L" _Y\""y"řǙșəʙ˙G*R{9 ÜĜŜƜǜȜ_w!" ĚL" ? vȁi'o?"  "d" ":"0$" 2D"    C","\ J"D"  d"     n"," h U""4"MD""""," "(} 5& /y{   uI\"  # N(܃"  "Ě#PU bgRt"   L"Ě ","3<""ԆM" l"D" }9~ 4"Ěl"$" QĚ54"p"  %K$)5?t"m""? " "  "`  (" * 3 "<" ř"     @``  "`  * Њ" ˆ" "" l؈" Lʈ"m[}OD^<l,"k[HYska} ?K{\   5" "ϓQ" eݪ9`"$" f," TՖ" d"͸t"Ɨd"  ȁ"\"  " lq""L"h\"" C % 1 ?" m#. 4՘" dϘ"  Ę" %řǙșəʙ˙G*R{9 ÜĜŜƜǜȜ_w͸?t " ™" " D" ͸t" ř " $"d"," ͸/ <"" ͸" ř  " "+! ,"\ n  _"" ԗ"$" L" tgȁZ4" :s ,"͸kG,#  ?/ \ED" ȁ$" "  D #  _:B" _#/B$"   "L"  =\""[ " ."  " "  ͗" - ,Ǘ"Y  Cȁ}e4"ȁ$" V?[ ' _ref__finder _ref__caps" !"d" ":"1$"  2D"    D"," \ J"4ś"  T͛"͸ k"," ] J̿"D" d"   ","iV""4"ND"""њ" ," ",-6Y  ]N W pO T\"  [  N("  "#PU bgRt"    4"l"$Ǥ" Q54Ǥ"p"  %K$)5?t"m"ͤ" %řǙșəʙ˙G*R{9 ͸?h " ¥"  "`  (" * 3 "͸<" ř͸"   Z gZ    @``  "`  * " " &-4{ݤ""" L"m[}OD^f8,ݤ"͸[H z?\0  jq3M 5| 4* WT4š8    M6š 4d 5,@Dz $ thisA |L $*5Dz , . z4l g1 4R\ ,   M .. .%t. aC&4.- ?!!!ѧo$z -ljk#. .-T-4-``Va`68Rt R 9šW       .\. !   .   š\1$1 $1410T1$1Dz$2: l2|2 #2?(0t3D3 $3&2l2 I$L5  55 6 6 6 55 a  w _ VG 547 D6 7  88l8 , $8#$8  8 &L86 ,:   )9K Zfz      !F 1 2,2257Dz1 ,1řǙșəʙ˙G*R{9 Dzr>Dz4$2  <2 ÜĜŜƜǜȜ_w  45    GFݱ}L5 * v c DzJ S[ckr {   \ old܊   ٌ  h   qL< < CIT\dkr{ $9T9 @ @ R@ ā@ Ľ? > >  >  =  =  <  < ; |; |: l: l: \9 \9 L9 @8$8$8\98 @ @                         \D DCC$C$D7#) DTF @ @ $P řO O ř PԣO O  ř ON N ř NĕN N ř NM  M ř MM  M ř ML  L ř  LL  ԀL ř  LK  K ř  KJ J ř  KJ ܳJ ř  J|I I ř JlI I ř I|H H ř I\H H ř HlH H ř HLG |G ř G\G G ř GPF řǙșəʙ˙G*R{9 dF$F9|EQ]'  @ @  ř  ř  ř  ř  ř   ř   ř    ř      ř       ř      ř   $ # ř   )(/ . ř+43: 9 ř6?>E D řAJIP O řLUT[ Z řW`_f e řbkjq p řmvut ,+"e.q: }   F    S  A$STS @ @ R,Y %,Y %,X %,X %,W  0,W 0W >V <4V 4V U BU bU RU rT $T ^T zT AYS$S$S\S6 @ @ d%-%.%/%0  01!02">3#<4$5%6&B7'b8(R9 )r:!*$;"+^<#,z=$-A>%$! 2U([[ &+8 CLVh`  8[  [ -9bG s[ [  &0h`&[ H[  ov5 BLa   aaa= ll $llmlt$l ,l2!!}$n\n nDn ;1n-n6 $oLo ĂoDo!9/oonu$n ,n3T!! $qlq qDq!9/q+$q\r qDr!9/r+$r |r rDr! 9 /̬r+q $r lr rDr! 9 /rrq $q ,qRy :4f! !![!$ulu uDu!9/u+u: $u\u ԱuDu!9/uuts$u ,t2S!!$w|w wDw!9/w+w> $xlx xDx!9/xxwy$w ,w2S!!zz $zz*z: Ծzz $zzzym$z$z2!!} | | D|!$|2ܱ|. |  |&D|!$|2|!/ }  }%D}!$}2} / }  }#D}!$} 2}. Ԉ~ })D}!$} 2  }$/  |  ~ ~'D~!$~ 2 ~"~{ $|$|<Vt2c!!B!!!_! | D!$5̭1  ր$Dۀ!$ր6ۀ3   #D!$62 ˁ  !D!$ 62  'D!$ 6  "3     %D!$ 6  Ђ $$=\h!@d!!]! !!l!߃  J\Ƀ$ 1Lt<!$Lԫ ,  ͆g<$$n,/_!$4ۃ!)$dˇS L݇$'*\ۃ!Lˊt؊<݊!$؊L݊|׊ˊ -  h<$$n36f!$4 +$d̞S!L$'*Y!LT\ )  j<܍$Í,ɍn3 5e$؏؏Dݏ +,|ݏЏ\ŏ!L,' )Xͯ!LT (  ΐf<$ې,n, .^$, +,dےVВ!L,' )X!LT (  ǔݓj<$,n, .^$L +,^!Lѕ,ܕ' )X!  ,ֹ,M$DTT4CD2 D#Ϙ  ԗ]<,$X9<tֹQ-! L\ /l  4% /J  D     ,      4 % / J   D,  ћ 4͛% /J ț D›, ›  4% /J  Dݛ ,   ݛ!֚$͙,Ś1-S<~=(07y @ P!1` 08(   /; F _i!(  ` Z+7di v{fk !   ! %DD 0T+$ը ř?lt 5-l M)64HDDѪ     Ѫ@ !|  u !й  %  D $) 5  H mD4!. $ dڪ % ?# T ř  i d L -m[}OD^ B|bit$  ?E h!*!!i!!!S! 08(ó  ȳ /;C\f!(ʳ   ,8ej w|glϳ  ! &'t $$D"}$,$D,/26y!* $D%$,D,%.15y!< S$,Lė֗! _haystack,0htޚ]$Ϛ,Dm(+/gߝD $ߝD>ߝB 4 & (M   4| #  $Н,D1:TW[!̊4 $D>< d 4|̡Ġ $,ՠDܠ3CFJ!ܷD!$DŤϤ<4 & (M      4,  4 & (M    4  ̌  ,   t ^ / dҥ  4ҥ, $,D3O>G+3:!֨4ۨ!$֨Dۨ<d 4, ֨  d      4   ,  <ϩ  թ&  Q  Ω&  4 & ( Mک  4ک,  $Ǩ,DR(07fo!!~ 4 &4 & (. n      4$ _"tҭ ͭ   d  ͭ7   U !~   \T ̎}$׬,Dɬ\vQZTmu| word_before word_after !G'! l $Dñܫ!̱԰x$,D'*.f!  D-   %ÜĜŜƜǜȜ_wv{? m D - %!ÜĜŜƜǜȜ_w?!v t  ! $ 0    mҷ  Ϸ   ķ!?! : ˵ضov}^rx~ !&,Dm[}OD^!%!b@"!> !\ܶ ?!!T  ?!- !2!!O!!!4 &  4ֹD;$  " |%* ( ?V ÜĜŜƜǜȜ_w !D  +L ,  ! ~  $  0    m r!?!t  4ڤD;$ d |$)#J",!D+L ,  !~$0 m fr!?!\T̳    !ڼ&,DƼ\řǙșəʙ˙G*R{9 "34$"+!T")ɕ4$"}!T CGL!R ?!! ?! - ֹ"!!l!sڤ"M!!! 4 &4 & (. n      4$!] t    d  7   U !~   \ T     ~$,D\x T!!!o !U7! 4 &4 & (. n      4$!] t    d  7   U !~   \ T \ԫ    ~$,D\!T0!!v !lN!!D+  !I$0? m!D +!?!   ! v$ 0    m\ o!?!    T 1      LSZ_ !&,D"[!'"!T< jmq!\ @?!!T  ?! U!f!;!!I!!!D+  !I$0? m!D +!?!   ! v$ 0    m\ o!?! \  Ԋ          LSZ_ !&,D"[!'"!T+ jmq!\ @?!!T  ?! U!f!;!!I!!4 &4 & (. n      4$!] \ d \ $,D\t}^ae!j!td   7 U !~ T \ o $,DT+B 'VY]!b!4 &  4ֹD;$  ""|#( ( ?T "7 !D  +L ,  ! ~  $  0    mp r!?!\d  !&,D\"L4$"!0m ! ?!!ֹU"!!8!?t  4ڤD ;$  좀d"|"' # ?S "7 !D  +L ,  ! ~  $ 0   mo r!?!Tׁ\ց|Ӂ  !&,DT"Sɕ4ހ$Ҁ"!8u'! ?!)ڤc"!!G!Nˆ (Ɔ", (͆ "OT afQs҆  " ģ@@ 4@*@0@ N%@6 @?jT?4?ljk#š`""HCԕH DH4H2G[G"PGa HGTGDG4G"'š+X"" N! OO П C--kԛͶͶBLN Ͷ"|DN N4P OO  OJ 04O OLO C )PL P0~SN TL4LS/"WTpY777 6wQ?m!؆"U?"\NNN:N _"$NNMeͶM ͶU/ÜĜŜƜǜȜ_w"ÜĜŜƜǜȜ_w"""""F{ $ߍZ"[ e e 18e   ""U1" ee!e   "%" !eS $ee ""' " "X_#e #ZeY#e g ;#e#e" #Be$e C"!e "`řǙșəʙ˙G*R{9  Ͷ!e ggStoragegfgf~[l(>__rust_std_internal_initu0"d(  #PU bgRtƙ   DzDz" ,"|45oT4š3n"| D 4# C     ɜ2۟666 66poison666M3TrCCC=C= (4D C"2T   J$ i@  OTƷ ` in  z  4   C  & # F $T 6" '  "*$ M"D #$ ř| 2řǙșəʙ˙G*R{9 ~ l44)  KTD4CCC=C="Tr #ͶJ #2"|"";",#,22 UnsafeCell2.\*="#L##]šaƲ& Բl "k")"1Ʒ":"#ۀ##8#yL  "Ę$  Uo?#DtT" "$ G7|T "|    4,  p ` <)\ $4D $4,  "",4#x," @" zmpԲi "##T##ً#K09 "HW7#u"~ "H#_- 2 5 4   T"   |        { WG\, _0<1 l| +? (5  řt܅# 2#4+,5%= 4> =B?K#[m@ >?|A@ A#>-58 9 8:#; 9:)<;6K <V#9,\# C"8 ,Ԟ#  4        P}J [` } Q#?\$#T #$5D#.lƀ3|ƀ 3+? 2("#""D 6""5!T.%!(%%,&!RRRlock6R"6RK6",4  & 6!6!"\ C (,'),5*'\*,+' \ $+,0),4 - \$ #-A,)&l0| 0+? /))̭D /H.)!\#$.,1)̽D 21!5 $1R0 4 ?) PV "#)H()s!!)4? !#7!7#$ 4& 4 (  5 *H I HJ #-KI:ML KJ]MLl M /{B C BD#ECF EDGF  G . f5o\#$4m } +u'' GdN"",4,! R [ ÜĜŜƜǜȜ_w"t g p O X  8  ,   # #Z     #1  ܅#T D # y # #H # řǙșəʙ˙G*R{9 `  #l  #LÜĜŜƜǜȜ_w#888Poisoned888 WouldBlocka ti# " "U #u 4#řǙșəʙ˙G*R{9  C "O #:" # ,}      ~                      (,Բ1resstack_idV@   & +.# # n####b # # #  #V #o # ֟# p G #& Ρ#Т# # # ###4 L #ā$ #[?#Dt\#˃#.,˃Զ# 5 (   ( # T $  A  PU# a jo { ?\#T#+D % ‹""D#/ T)( ,͊#,׊ , $ M ,M\ Ë VТ#v  Ӊʊ",Ђ ,ׂ#iD##W##i#r#{Q#L׉##ƭ##4#z#Ȯ# T#;  : "##&### ####Т##N#U " "1@#^"got "E#\- |\  < ,"nj",4ÜĜŜƜǜȜ_w psL  ݎ: ,"", ,#!RL  Ƒ9 $"", ,#!R4 d 9/Dݓ  d #  E,ד ] Jד<  \" r, ] J܆\  |Ɣ܅#,^K#4ODܒ", (K#"PAت#J#cB"ت#QTÜĜŜƜǜȜ_wljk# #1]9]] A]I] I] Q] ]    ,##   ]8"O 3f*ہ     !   ! !] ] ř]] ]  !]Q] Y]XB šƲԲlL_js22#2.\*="řǙșəʙ˙G*R{9  2#,   )#9K Zfz # 8 W"%2  ̦' ( ')(2* )?J+*Y + <" L# "[$#h{% $&% &  K4 W!Dŝ t]],"",ٝL #s +fN1$ۏ !]l "qřǙșəʙ˙G*R{9 " #h # ##|# #    !(/6B-) Ő   O ### T#+ ۬#1]9]] A] I] I] Q] ]    L##  ###) 8 F O_q!  "! " !] ] ř]] ] !] Q] Y]XBܯ##ܯ#@,گ   )#9K Zfz # گ8 W"گ%2  ( ) ($*)1D+ *Q\,+k , # $ #&%$3F& %S^'&m ' |  _4 W5D ׬]]-"4˧"L"l #gJ =%ۏ !] "řǙșəʙ˙G*R{9 " ####N##`  px  19@GN-  Ő  S C###$ɴ"ɴ "$*$t " ManuallyDrop~$"i_f< LXű"$#D" # ƀ 4ƀ 7o#d#*#1 |#  #; #~D+0AS , 1]9] ] A] I] I]Q]],#d    L#, ԉL G 7  !] ] ř]] ] !]Q]Y]XB|#" /,#V ,< ]]Ҷ"Lȶ#5 ##kL#?,#4 -ۏ !] b" ##Բ  Ő #v## {#t<#$L:"Lb7#\  d 9/4 T #  F,] J<  \ #  s, ] Jܤ4MD", r #~- #*ـ$$pO"_# (Z# #-"X#(a#  X# *6af sxejf# $"" $# ř##řǙșəʙ˙G*R{9 #$#  <#E256X$"(Z$  _$ +7 Bц$y ц$^h ц$(a$ X$ц$"MR _dQqf$ $ $$$ $$ ř$$řǙșəʙ˙G*R{9 |$,$$<$_kindE%(Z%  _% +7 BksZdnj(a% %!LQ ^cPpf% $ $% ř%%řǙșəʙ˙G*R{9 t%,%$<%$E % %( % ! %!0< Gnxnjnj( %  % %!sx gzY %! $($G$% ř% %řǙșəʙ˙G*R{9 | %, %$< %$E%%%(% !%!3? Jnjnjnj(% %%%   !  g^%! —$ M[f[?#Ι$$% ř%%řǙșəʙ˙G*R{9 |%,%$<%$E_&!(Z& #-"X&(a& X& *6af sxejf&!$"" $& ř&&řǙșəʙ˙G*R{9 t&,&$<&$E_'!(Z' #-"X'(a' X' *6af sxejf'!$"" $' ř' 'řǙșəʙ˙G*R{9 l'$ '$<'E)(KKKK 7K7K736hm $h*(*(((( !((!3?G$ $* $e$$o$(( (((  )( $*h*(  ) *(   *  6    y ~((!/<ENWGP$Y'0 9 #$  h ;(\;( $<(<<(-<( U$;(D<( X$$m<(=(;(;($;($<;(řǙșəʙ˙G*R{9 6$ mEpƭ$k$<<($t<($<( w04<(<((<( Qppʽp opopo2Ы U?ݱ$<(<(<.(=( R̵̊̊ϊЊD^)cO$ݱ$,=(<;(L=(ۂ  =(d=(=.(=( 4=(t=()=(%<(<($F4<(t<(Ͷ+$$$}choice?E$Vذ$`$;$ۂ /ܳ$ _ref__needlesT<(4<(  <(D<(<(=('<( ݱ$@$ <(}g$rY(DY( $Y($Y(,1Y,(DY( WTY(Z!(   Literalexactߚ]ؾ)UYe(DZ( X$$$Z+( Z(Z(X(Y($X(͜$Y(řǙșəʙ˙G*R{9 $ SeqliteralsD~d W'$0$$p$8$$M$T$Z(ԓZ(tZ( $>$Z(+Z(Z($$$Z(lits!ۼ$ _ref__kindDd($d($d( $1Dd($d(d(d(b(dc(,b(Db($c($'+/g$Dq($q(4q( $2Dr($r(q(r(p(dp(,p(Dp($p($'+/g$Dt($t(dt( $1tt(t(s(,s(,s($v6$}v(w(u(,u(,u( .d((($(,( -D( ",(%|($( I4((>(d( ,(<$$DÏ($,ŏ(Ï($̏( CDя($ۏ(̏((i(d($,(D($(*-0k$$D($,((4( DD($(((m(d֐($,(D($ǐ(*-0k$$D($,((d( Ct((ёG(,($,(8$$4($,((<( L(L((ג=($($,(7$$((/X( % 7?8ў(  %  7?((  %  7  ?0(  %  7  ?8( % 7 ?8( % 7 ?0( % 7 ? !(6 !$U !Ct !?j !? ! !?(!$+29@GN\ hp$$ —$W`$$ $Yb$ 9   ,  }    ^$3$$$|$?$$ ((/X(T\ 3? n 8ў(3?  r +((OX5A t 0(5A  !t /8( S \5 A  t 8( 5 A     % t 30(W`5A t Ý(! $!)$:  $$a $—$J\ —$h$ $$ $$CU $a$$ $ CQ '  E   c  x$ $% $ $E $ $c !$ 2lѩ(ѩ( (&$Ǫ( ř?t($(ذ$<(4(*-(( R$$Uq?$̨((Z$( ř?Ӱ(\Ӱ( $߰(<(16 Lǰ(  ̰( ( ! ̲( ř ?(d( $(<(  16  L(  ( ز( !Ӳ( ř?ܔ(d( $(<( 16L((( !( ř?Զ(\( $´(<ȴ(16L(( !?̕( ř?յ(Tյ( $(<(16Lɵ(( !?IJ( ř?(d( $(<(16L((׷( !ҷ( ř?(( $(<(27L(Ă( !?( ř ?$( ř( (h(j(j(Ѵh(f(j(s(!$t̨($($<({ $:$C$I>-$$g-"Ġ$$T$y$A E & (?ε(?*(?  w6ذ$$$$m$$$$$?T(4($ (D((D(CT((( $$ (B$(( L(( ,($( @5"(((  (  P͜$((- U(I  X] Q o x}    ?( D(L(d(4( ( D($(T( (<( {o ($(4(,( 2 +( (D(,(!( -#(D(|($( +4(D((,( +<(D(( ((((($d($(͜$( Extractor limit_class limit_repeatlimit_literal_len limit_total 6я l( $ $OPrefixSuffixEhԭ$d(/8řǙșəʙ˙G*R{9 \5 $$]$ $t($$F$a extractor2b  $$$$$$}$A$$$$$1$C())!&+8 CLV)!) Ղ%&0`)" h"3 )D)!&1Ղ%9A "Ubl%`)""*/ ""FKS)!H)Ղ%Ղ%τ%τ%7 DN%P)   (- ř)![gՂ%oՂ%w %`)""*/ "AFN)!)Ղ%Ղ% τ%τ%9 FPň%)d) ,) )|)())  ]$9)) W" !Q\)% ))y)),)řǙșəʙ˙G*R{9 "/a%N%ۊ%4)"d))"*,))d) > )|))F),)$)"<"%(2)2)!&+8υ@IS(2)  3)!MR _dPq2)!υ 3)  82) !2)!-9]G n2)!2) υ&03) h1 2)B2)!&1υ9AR_i%3)&2)!H2)υυov5 BLP2)   (- ř2)![gυoυw ȕ%3)(- >CK2)!2)υυ 9 FP%ܒ5)5)t5)* 5)L 'sM46)  4) ' 5)4)T4)>_Ծ7)|7)υ4 ;D7)FT7) 7)7) 7)$ 'KM78),7)υυN_/8GVυs|8)  ̞8)י% 8)58:)T8)'_9")  9 )!9;),9)&_ܦ;);) ,;));):k):#),:) /%zd@)@) !@)?)T?),?)[6},@) @) ! @)xA)|A),B)B) A4)4B) _]>B)B)Al)TA),A)>ţ%4D)|D)D)C),C)υ,C)X5y} D)}υ{4F)lF)F)E)E)υ,E)X5y} F)}υ{4G)G)G8)G)υ,G) -a_]ZυXa4J)lJ)J)I)I)υ,I)X5y} J)}υ{|L)DL)$L)TL)L) L)$ 'KL)KC),L)υ,L)< ̿O)O)!,O)|O) υO)4O) 2DO)O)lP)LP) P)(P,) !P1) Q)lP)DP)P) TP)P()P)י% >P,) yN)O) O#),N)O)RυPN-R! %%:י%$-<Kυhqy~4̠T)T)!,T)T$)DT)<TT)T?)̆T)י%=T)S)TS),S)?!י%44U)U)\U)!*lU)V)UI)n,Z)Z)DZ)%TZ)Z) $Z)J,Z)YP)DzY) ,Y),Y)υ$'W%,[)[)D[)%T[)[) #$[)I+[)[`)4[) L[),[)υ$'W%pl)(l) !l)!-9 Dm w^h (l) pl)  *6bg tyejl)!% %% % "l)!l) %&0pl)&l)!Hl)%%ov5 BLq) /7(p) !q)!-9 D `j%(q) q) +7ch uzfkq)!% %%% #())!(-:BKUP)  8) !)"/;gJ%x)") '1P) h4 )F)"(3;CυTak%P)+0 GLT)"H)υυ8 EO%P)  !). ř)"]iqy %P)+0ȕ%<AI)")!υυ: GQȕ%())"(-:BKUP) %8) ")"/;aJ%r)") '1P)%* )<)"(3;CυzTak%P)+0%<AI)"H)υυ8 EO%P)  !). ř)"]iqy %P)+0ȕ%<AI)")!υυ: GQȕ%v) !xw)v.)"(v)+B_w) w+)"(w)*ISmallIndex::LIMITx) !x)x0)"(x)+B_y) y))"(y)*ISmallIndex::SIZE|)x|)(|)%|!)|) _ !Z(|)% P|)|/)8|) _ "~ |)% h|)|)"{){)({)ÜĜŜƜǜȜ_w.cs%]% ))י%())ށ') !)"y)(Ɓ)()?י% י%) !)"D0)P)()(̓~))0ƃ) _qIƃ)ڄ)")()()H%0))@)4P)ˆ)"W)()()υ> )"υ"0)x)0) 6@))"ۇQ))()υ>% %)"υ$0)x)0)%4@))"‰Q))(։)υ>% %)"υ"0ߋ)xߋ)0) 6@))"Q)ɋ)()υ>% %)"υ$0))@)4P))"Q)(Ύ)(Ď)υ> )) ()++))%N()% P)9)8ƒ)%\ Β)% hƒ))"̑))()% J%%%4*)#) (),1) !)")()(ԕ)B%40))X) 6h))"ƗY)8)(ٗ)υ?%4())@)ڰ .P))" ))"ɞ[)Dz)ַ%(Ҟ)(ٞ)77mڰ (Ӡ)Ӡ)@٠)ڰ .P٠)Π). ɠ))"k)0)%H)()%??uڰ (Ѣ)Ѣ)@ע)ڰ .Pע)Т) % ̢)O()"[)Dz)%()()**`ڰ %(Ĥ)Ĥ)@ʤ)ڰ .Pʤ)ä) % )O()"k)0)н%H)()**`ڰ %x) 08() ")"/;C%\f%() x)#PU bgRt)" %%& )") '1x)+0 GLT)"H)%%8 EO&p) 08() ")"/;C%\fɄ&()  p)#PU bgRt)" %%& () 08P) 09(Ѿ) "־)վ)E#T` k &&ˈ&(ؾ) () 2+P)  ,  8 qvݾ)# ʉ&  ݉& ]kv݉&j&8)@)) Q4())X)Q$)) ")))"T)`))) ݉&%,5؉6&E 8)p) )50)x)  %řǙșəʙ˙G*R{9 Qqv)x)mޑ&@) x) ) Q ;  0)) %řǙșəʙ˙G*R{9 ޑ&?m)0) W  @) ) P)  ) ř& )) )")LSZQ) &) H)ޑ& )m[}OD^ޑ&&LӋoޑ&)k &ޑ& o?-ms&&&&Aޑ&v) "xw)v.)#(v)+B_w) w+)#(w)*ISmallIndex::LIMITx) "x)x0)#(x)+B_y) y))#(y)*ISmallIndex::SIZE|)x|)(|)%|!)|) _ "Z(|)& P|)|/)8|) _ #~ |)& h|)|)#{) {)({)ÜĜŜƜǜȜ_w.cs&]ѝ& ))י%())ށ') ")#y)(Ɓ)()?י% י%) ")#D0)P) ()(̓~))0ƃ) _qIƃ)ڄ)#)()() H &0))@)4P)ˆ)#W)()()υ> )#υ"0)x)0)%4@))#ۇQ))()υ>% %)#υ"0)x)0)%4@))#‰Q))(։)υ>% %)#υ"0ߋ)xߋ)0)%4@))#Q)ɋ)()υ>% %)#υ"0))@)4P))#Q)(Ύ)(Ď)υ> ))%())+))&L()& P)9)8ƒ)ѝ&\ Β)& hƒ))#̑) )()Ş& J%&ѝ&4*)#)%()*1) ")#)()(ԕ)B%40))X)%4h))#ƗY)8)(ٗ)υ?%4())@).P))" ))#ɞ[)Dz)ַ%(Ҟ)(ٞ)77m(Ӡ)Ӡ)@٠).P٠)Π). ɠ))#k)0)%H)()%??u(Ѣ)Ѣ)@ע).Pע)Т) % ̢)O()#[)Dz)%()()**`%(Ĥ)Ĥ)@ʤ).Pʤ)ä) % )O()#k)0)н%H)()**`%x)%*2() #)#/; F%_i%()  x)#PU bgRt)#&&%& )#) &'1x)+0&<AI)#H)&&%%8 EO&p)Ʉ&*2() #)#/; F%_iɄ&()  p)#PU bgRt)#&&%& ()&*2P) 09(Ѿ) #־)վ)E#T` k݉&&&&(ؾ) () 2+P)  ,  8 qvݾ)# &  ݉& &&]kv݉&j&8)@))&.())X) #M&)) #)))#T)`)))݉&!,5؉6&A& 8)p) )50)x) &E#)Gx)mޑ&@) x) ) Q ;  0)) %řǙșəʙ˙G*R{9 ޑ&?m?)0) W  @) ) P)  ) ř& )) )#)LSZQ) &) H)ޑ& )Ö&o&Ӌޑ&D) ?&1ޑ&K{{ ?-m"J&"&&0&ޑ&e)) "$))))""","؉w7{t0)$) n&)) #$))))###,#؉2ȳ")$) n&! /7"ł)1(! #!#0<Dłmw&(! !  +"  ł+  7 pu!#  ӐUcndӐ2256<&+ &=E(< #<#-9ՁA& &]g&256256012345(ff$&2:CM (f,3;fV^#u}Ba    f$/<ENW_iw f$f *Hff$>CIT\dkr{ *|* *-*|* D*3+**ÜĜŜƜǜȜ_wȏ*D* w&&** d*Ծ*,* 4*/*  |*$* wD*4*x*  PʋJ$) W*D  SXċ d mr  ~  ? * $*<*|*,*% <*$*D* ȏ*^T*4*Վ * D*!*,*D *"*  %I%*6"*m-*$*Z7(**[ÜĜŜƜǜȜ_wȏt %&&" * *=+* **29@&*D*&4*\*$ AD*ʋʋʋZRm[}OD^&ؘ֘Mp&yw7;?. a  7&&&k&CċJՎ&v&ؘؘؘ("*"*$&+17D&LU_("* $  "*#*   Т#*   P#* #*#*#*h$*P$*x$* X$*        (5IO]ciox"*$ &&  &    r {_ hL U9 B&/  $       k P 5  EN \EN\EN\E N\ D M [  D M[  DM[DM[DM [CLZ    AP$a&n&w/8% 30(  9`  +  7 pu$  $ ܇') $$p~׉'~'>( % +7ch uzfk$     #(/ 0"NS `eQr0$   ''  (0  44= 4> u ?}O @Wd A=X B`m D>y CD  ?@ABC 3536W 7` Q 8Yf 95Z :bo <6{ ;< 789:; 30"1+42<I40U34ms 123 3+, 4-<I/+U./ ,"-. 52#p2$s%|O&W d'#Y(an*$z)*%&'() h2P2YbOWdY an"z!" ! h1P2?HOWdYanz 1  4<IUU[ io 1  4 <I  U        0@I4<IU  0$  E&'/8@ &R8A J     * %r @I6? H  # ,         &  \ * 3  ( 1     5     #$ + - 0 " 5 6  = >o &B  /  {  h      $       9 -   >  % O 5uf  !f   u $uJԉvĉv 2 vdv4v C.2DvvDv K2Tv|v "  24 v v$vf  Dv   w?.Ly  y$  '2L  dy   2, !   y i  24y dy4y'% 1DyyDyé'< 1Ty|{"  14 { {${ {D{ C/${1t{{ D{  {?.\|  |$  '1L d|ė| |+\| ),|1||64| E1D||T\|><|1||f4| E1D|d}! }$!}* *8}%"`$ %H}b$ظb)'12J& 'Z (b   )&  g ()#1@`` #"`"1*~6 t~8 818}.(~/ "`- ./H} c- ظc)1128~ d- ظd)212 0 12 3 - 40  g-34,1@`` ,"`+1*78~D~'675 1Ļ~t~< <!1,;<1 ]:;~? L~~ ?%>"1,~9 ]=> ]9 :=A ]@9I ]A9I ] C@Q ] DA ] BCD B$#1]I O HP \ ɕOJPAN%1, !] E ] G ř]]  ] E!]@Q]AY]GXF1C4 wNtM&1D"Dé'MDL'1T*Q# QL(15K L9]JK]H IJA]RHI]SHI]URQ]VS]TUVT*)a t a -!]W (\][ \8e[ظe*^+13J] ^Z_b 4`] g4_`Z,1<]Y řZ}] ] W!]RQ]SY]Y\X1Cxf0f/4f40> 23q/#u u v yyy|jq  ] ] ]]}gq$q,vʟ  yۏ ! ]   Trʟ ɕ&ۏ !] ]':w  ' Ax%extra[ U G  $ & - 0  @Ő  A E RŐ SpW[]O ߨ''é'f 'é''   ظ ' 4 ظ ظk '  's  '  'J é'# ظ' $+     R ƟƟL4 -9 ԠdԠ 46T is˞ݘ앟͙45Le*@D}k άάУL2 -7 ѭdѭ <ޭ+T ^hʫ ݘ읬<͙4Ŭ4Kd)JMQУw | ")4;?"PatternID::SIZEֺdֺ4ں(Dںɺdɺ B)| ",͙/2b" l #)4;?# StateID::SIZEd4 (Dd B)l #,͙/2b#  2416H-KKKK 7K7K736bg 'y(, $,$-9 D' '`j',$, '&0H-' h'3 ,D,$&1'9A'R_i'H-''*/ ''FKS,$H,''''7 DN'P-   (- ř-$[g'o'w 'H-''*/ 'AFN-$-'' ''9 FP'8@ /7(? $?$-9Aʟ Zd'(@ 8@"NS `eQr@$ ʟ ʟ ' 8F')1(E $E$-9 Dʟ ]g'hS')1Sҁ)2(R $RR>$MY dʟ 'ҁJ8+\J+ $J+ ,J+TJ+dJ+$J+MJ+H+ÜĜŜƜǜȜ_w#I+ ,H+TH+dI+$I+* init_value7 match_offsetG2J'ML9+\L+',L+ ,L+TL+dM+$M+CM+J+'K+ ,K+TK+dK+$K+*'.'326w'.`+. .`+d`+ ,`+$t`+'`"+\`+ՓK$l`+(d`+ `$+`+̔,`+ - $`+/ T`+.  `+ ř `r+$a+ ř%`v+ # ?!Tb+.  lc+,c+ ,c+*#, ,v8΁L8$ , 1|D΁$T*, Q<Od ,A΁˷(lłDłD , 1l@$ , /|D΁$T|D΁$T+Y$ ,l*M΁΁|D$T* F4*p$ ,4 !N(T D+J L(D3((%(-:BKU% '1 h2 Dp()@6%cnv~(('8p'8@HU%HJ Wa,\ 4*Fd,{9(TDE$,qd=DT#tùB3ԷN,,>ù|D $T.TD> !|$,41 4a(DT ,D ,; $w $,4C Fu,l,˷((<|$E$J,,,')Y˷(,l,˷((<|$E$Z͙L͙L,')Y˷(,l,˷((<| $N-I$,,')Y˷((\L+ $`$,znA@TD* ʀ^d$Ҁ~rAT\̛DLĂ2 !ӂm$,%,9A;iTT˃ĽDكDك0 !b$%,$9<j(É%(-3@ HQ[(ˉ   ,hʋ  +  7 puЉ%     Zhsg 4ԉ%҉  $-7 h4 ։Fhʋ()@6ډ%cn v~((+0 <AIPhʋ'8@HU%H܉  M Zd @쉍 %ČhL<ό4 Jt<" 4AD8t̟B4(L<̏4܏( +Z̟d?L , 1\̕ݕ9, , 0(ؚݚ%(-3@۔HQ[(  ,P  ,  8 qv% ۔  ^lwk(8% ۔$-7(* <P h4KA%ny۔((+0 <AIPP+0 GLTYan%H۔۔"Q ^h ](\$Ģ $4ĢTܢDܢ % "`"`2@`` U# ,<($$'HKN(  Ī +| "<M$ī *T (H-ϫتw Ϋ7,<($(e03(U((dѮ?L۔, 1\$ $4گ>,۔,x8(\±$DZ($±4DZ|±ױ:,۔,v8(\$($4*,˷(O<R۔,A(˷(L4$۔, 1\$($4*D(OTE$۔,?((\$($4* Q,<,ѷ۔,Ƿ@()P   ljk# \a )s{(  % %2>F) )q{() (   P  1*   )-  9 rw %    ))]kvj)9 %  $-7P '8  )),1 ))HMUZbo ǿ%H ))N [e) *2x (  % %2> Iq{(     ,x   +  7 pu % )   Zhsg 4 6<< 0{?ippʽp opopo2Ы )(  % %1%%%, @ήj )u)C )O)m(   62),8ej w|gl %) ̓)̓)ޕ) &(  %(-ł:BKU( 'H LT3 ?  n   %9? MV_gpyTfo     %  '1H ,; K \ ixH  /@ Y  e  sx   %%.7(1:D Ran   , |]) }] ! }] ٱg |`]$,lT,\$ řL ř ӷ#Ҷ,ȶZ(  %(4<EOl % %8 ( %,4  PX$qv  %  % !+H  %@EKV^fmt}  !ljk#]b ˪)t|(  % %/;C˪) ˪)_i)(    !˪)$QV chSu % )))  %  '1 !˪)˪),1 ˪)˪)HMU %H ))8 EOЯ)( )p  ł % 1 9P    %  h  %M@ h :*W %y %  &  X`łH TD    [ o8nj( )p łNVi2:jFR P 2:1FR(. <BUh  x 4<I U  @ h 09kPXldX `m y     & ʴ)$-5H4PclXa j+=F O  Ӑ 5%    /h ӐF }   =   &  "ʴ)*3=H ,; K!ʴ)\ !i ~"ʴ)" #ʴ)# $ʴ) $  -%ʴ)? %M p &ʴ)&ł/'ʴ)@'łYe sx  P (ʴ)(.)ʴ)?)Wcqvh *ʴ)*.+ʴ)?+Wcqv @ ,ʴ) ,h -ʴ) -\.ʴ)m.W/ʴ)h/     # &0ʴ)ʴ)ʴ)ʴ) "* /8A  %     mvӐӐʴ)ʴ)ʴ)ʴ) ʴ)ʴ)ʴ)$ʴ)-ʴ)6ʴ)?ʴ)Hʴ)Qʴ)Zʴ)cʴ)lv   f  O\   +)?%а - /7а -))2( - & - ->&MYa )))( -  б - +б -  +  7 pu -&   ) ))Zhs)h)6P - - )1 - )2( - & - -A&P\)d    @% ( <-   <-"NS `eQr <-& &&&& ?  ) $) 9<.d.,.l.$._.$ ..$. .D.$. 0\. d.$.$.97 .l. $.     . .  $  . . $. . L.  .   = I. .h p \. >D .\.$.A\. d.$.$.Z7  .l.$. ..$ . .$. .L. .   =I. .h p \. >D .\.$.a\.  d.$.$.z7  .l.$. ..$ ..$. .L.  .   = I.  .h p  \." >"D . . . ....$ц$\.ц$pM4UyyV{J>>)|b)%V256256\ .  L . ! &.ц$\ .'_256 (4<C\ (4<C4  'Õ3ȏ;BD  'Õ3;B$ 'Õ3;Bt> '3;B4 (4<C, (4<C (4<C (4<C$ʄ (4<C| '3;B '3:A" '3:AF '3;B (4<Cl (4<Ct   -        3yJ-v} (4<C! '3:A! ) |?ckr,= )5=D4= )5<C!"  /;CJ" .:ĚBI" .:BI! !0<DKՆ#   )5<C>! !0<"DK[/ #2">"FM   ,8@Gi$   &2ц$:Ai%   &29@̦ % !!*6=DK% !!-9@GNU4( !!(4<C<( !!(4<C<( !!(4<C4( !!(4<C,( !!(4<C<ɞ( !!(4<C\ߞ( !!(4<C) !!'3";B2) !!'3:A) !!(4<C) ""(4<C!) ""+7?F() ""+7?F!) ##+7?F() ##+7?F܃0 $$'3;B- $$'3';B@# $$(4<CE% $$(4<C|u ) ^yϊu-fnu * gٙov< %%(ł4<C %%(4˪)<C<#- &&*)6>E  ,L!,&$$$L%,%Tտ%%t%Dʨ%L%T%tȗ% This crate exposes a variety of regex engines used by the `regex` crate. It provides a vast, sprawling and "expert" level API to each regex engine. The regex engines provided by this crate focus heavily on finite automata implementations and specifically guarantee worst case `O(m * n)` time complexity for all searches. (Where `m ~ len(regex)` and `n ~ len(haystack)`.) The primary goal of this crate is to serve as an implementation detail for the `regex` crate. A secondary goal is to make its internals available for use by others. # Table of contents * [Should I be using this crate?](#should-i-be-using-this-crate) gives some reasons for and against using this crate. * [Examples](#examples) provides a small selection of things you can do with this crate. * [Available regex engines](#available-regex-engines) provides a hyperlinked list of all regex engines in this crate. * [API themes](#api-themes) discusses common elements used throughout this crate. * [Crate features](#crate-features) documents the extensive list of Cargo features available. # Should I be using this crate? If you find yourself here because you just want to use regexes, then you should first check out whether the [`regex` crate](https://docs.rs/regex) meets your needs. It provides a streamlined and difficult-to-misuse API for regex searching. If you're here because there is something specific you want to do that can't be easily done with `regex` crate, then you are perhaps in the right place. It's most likely that the first stop you'll want to make is to explore the [`meta` regex APIs](meta). Namely, the `regex` crate is just a light wrapper over a [`meta::Regex`], so its API will probably be the easiest to transition to. In contrast to the `regex` crate, the `meta::Regex` API supports more search parameters and does multi-pattern searches. However, it isn't quite as ergonomic. Otherwise, the following is an inexhaustive list of reasons to use this crate: * You want to analyze or use a [Thompson `NFA`](nfa::thompson::NFA) directly. * You want more powerful multi-pattern search than what is provided by `RegexSet` in the `regex` crate. All regex engines in this crate support multi-pattern searches. * You want to use one of the `regex` crate's internal engines directly because of some interesting configuration that isn't possible via the `regex` crate. For example, a [lazy DFA's configuration](hybrid::dfa::Config) exposes a dizzying number of options for controlling its execution. * You want to use the lower level search APIs. For example, both the [lazy DFA](hybrid::dfa) and [fully compiled DFAs](dfa) support searching by exploring the automaton one state at a time. This might be useful, for example, for stream searches or searches of strings stored in non-contiguous in memory. * You want to build a fully compiled DFA and then [use zero-copy deserialization](dfa::dense::DFA::from_bytes) to load it into memory and use it for searching. This use case is supported in core-only no-std/no-alloc environments. * You want to run [anchored searches](Input::anchored) without using the `^` anchor in your regex pattern. * You need to work-around contention issues with sharing a regex across multiple threads. The [`meta::Regex::search_with`](meta::Regex::search_with) API permits bypassing any kind of synchronization at all by requiring the caller to provide the mutable scratch spaced needed during a search. * You want to build your own regex engine on top of the `regex` crate's infrastructure. # Examples This section tries to identify a few interesting things you can do with this crate and demonstrates them. ### Multi-pattern searches with capture groups One of the more frustrating limitations of `RegexSet` in the `regex` crate (at the time of writing) is that it doesn't report match positions. With this crate, multi-pattern support was intentionally designed in from the beginning, which means it works in all regex engines and even for capture groups as well. This example shows how to search for matches of multiple regexes, where each regex uses the same capture group names to parse different key-value formats. ``` use regex_automata::{meta::Regex, PatternID}; let re = Regex::new_many(&[ r#"(?m)^(?[[:word:]]+)=(?[[:word:]]+)$"#, r#"(?m)^(?[[:word:]]+)="(?[^"]+)"$"#, r#"(?m)^(?[[:word:]]+)='(?[^']+)'$"#, r#"(?m)^(?[[:word:]]+):\s*(?[[:word:]]+)$"#, ])?; let hay = r#" best_album="Blow Your Face Out" best_quote='"then as it was, then again it will be"' best_year=1973 best_simpsons_episode: HOMR "#; let mut kvs = vec![]; for caps in re.captures_iter(hay) { // N.B. One could use capture indices '1' and '2' here // as well. Capture indices are local to each pattern. // (Just like names are.) let key = &hay[caps.get_group_by_name("key").unwrap()]; let val = &hay[caps.get_group_by_name("val").unwrap()]; kvs.push((key, val)); } assert_eq!(kvs, vec![ ("best_album", "Blow Your Face Out"), ("best_quote", "\"then as it was, then again it will be\""), ("best_year", "1973"), ("best_simpsons_episode", "HOMR"), ]); # Ok::<(), Box>(()) ``` ### Build a full DFA and walk it manually One of the regex engines in this crate is a fully compiled DFA. It takes worst case exponential time to build, but once built, it can be easily explored and used for searches. Here's a simple example that uses its lower level APIs to implement a simple anchored search by hand. ``` use regex_automata::{dfa::{Automaton, dense}, Input}; let dfa = dense::DFA::new(r"(?-u)\b[A-Z]\w+z\b")?; let haystack = "Quartz"; // The start state is determined by inspecting the position and the // initial bytes of the haystack. let mut state = dfa.start_state_forward(&Input::new(haystack))?; // Walk all the bytes in the haystack. for &b in haystack.as_bytes().iter() { state = dfa.next_state(state, b); } // DFAs in this crate require an explicit // end-of-input transition if a search reaches // the end of a haystack. state = dfa.next_eoi_state(state); assert!(dfa.is_match_state(state)); # Ok::<(), Box>(()) ``` Or do the same with a lazy DFA that avoids exponential worst case compile time, but requires mutable scratch space to lazily build the DFA during the search. ``` use regex_automata::{hybrid::dfa::DFA, Input}; let dfa = DFA::new(r"(?-u)\b[A-Z]\w+z\b")?; let mut cache = dfa.create_cache(); let hay = "Quartz"; // The start state is determined by inspecting the position and the // initial bytes of the haystack. let mut state = dfa.start_state_forward(&mut cache, &Input::new(hay))?; // Walk all the bytes in the haystack. for &b in hay.as_bytes().iter() { state = dfa.next_state(&mut cache, state, b)?; } // DFAs in this crate require an explicit // end-of-input transition if a search reaches // the end of a haystack. state = dfa.next_eoi_state(&mut cache, state)?; assert!(state.is_match()); # Ok::<(), Box>(()) ``` ### Find all overlapping matches This example shows how to build a DFA and use it to find all possible matches, including overlapping matches. A similar example will work with a lazy DFA as well. This also works with multiple patterns and will report all matches at the same position where multiple patterns match. ``` use regex_automata::{ dfa::{dense, Automaton, OverlappingState}, Input, MatchKind, }; let dfa = dense::DFA::builder() .configure(dense::DFA::config().match_kind(MatchKind::All)) .build(r"(?-u)\w{3,}")?; let input = Input::new("homer marge bart lisa maggie"); let mut state = OverlappingState::start(); let mut matches = vec![]; while let Some(hm) = { dfa.try_search_overlapping_fwd(&input, &mut state)?; state.get_match() } { matches.push(hm.offset()); } assert_eq!(matches, vec![ 3, 4, 5, // hom, home, homer 9, 10, 11, // mar, marg, marge 15, 16, // bar, bart 20, 21, // lis, lisa 25, 26, 27, 28, // mag, magg, maggi, maggie ]); # Ok::<(), Box>(()) ``` # Available regex engines The following is a complete list of all regex engines provided by this crate, along with a very brief description of it and why you might want to use it. * [`dfa::regex::Regex`] is a regex engine that works on top of either [dense](dfa::dense) or [sparse](dfa::sparse) fully compiled DFAs. You might use a DFA if you need the fastest possible regex engine in this crate and can afford the exorbitant memory usage usually required by DFAs. Low level APIs on fully compiled DFAs are provided by the [`Automaton` trait](dfa::Automaton). Fully compiled dense DFAs can handle all regexes except for searching a regex with a Unicode word boundary on non-ASCII haystacks. A fully compiled DFA based regex can only report the start and end of each match. * [`hybrid::regex::Regex`] is a regex engine that works on top of a lazily built DFA. Its performance profile is very similar to that of fully compiled DFAs, but can be slower in some pathological cases. Fully compiled DFAs are also amenable to more optimizations, such as state acceleration, that aren't available in a lazy DFA. You might use this lazy DFA if you can't abide the worst case exponential compile time of a full DFA, but still want the DFA search performance in the vast majority of cases. A lazy DFA based regex can only report the start and end of each match. * [`dfa::onepass::DFA`] is a regex engine that is implemented as a DFA, but can report the matches of each capture group in addition to the start and end of each match. The catch is that it only works on a somewhat small subset of regexes known as "one-pass." You'll want to use this for cases when you need capture group matches and the regex is one-pass since it is likely to be faster than any alternative. A one-pass DFA can handle all types of regexes, but does have some reasonable limits on the number of capture groups it can handle. * [`nfa::thompson::backtrack::BoundedBacktracker`] is a regex engine that uses backtracking, but keeps track of the work it has done to avoid catastrophic backtracking. Like the one-pass DFA, it provides the matches of each capture group. It retains the `O(m * n)` worst case time bound. This tends to be slower than the one-pass DFA regex engine, but faster than the PikeVM. It can handle all types of regexes, but usually only works well with small haystacks and small regexes due to the memory required to avoid redoing work. * [`nfa::thompson::pikevm::PikeVM`] is a regex engine that can handle all regexes, of all sizes and provides capture group matches. It tends to be a tool of last resort because it is also usually the slowest regex engine. * [`meta::Regex`] is the meta regex engine that combines *all* of the above engines into one. The reason for this is that each of the engines above have their own caveats such as, "only handles a subset of regexes" or "is generally slow." The meta regex engine accounts for all of these caveats and composes the engines in a way that attempts to mitigate each engine's weaknesses while emphasizing its strengths. For example, it will attempt to run a lazy DFA even if it might fail. In which case, it will restart the search with a likely slower but more capable regex engine. The meta regex engine is what you should default to. Use one of the above engines directly only if you have a specific reason to. # API themes While each regex engine has its own APIs and configuration options, there are some general themes followed by all of them. ### The `Input` abstraction Most search routines in this crate accept anything that implements `Into`. Both `&str` and `&[u8]` haystacks satisfy this constraint, which means that things like `engine.search("foo")` will work as you would expect. By virtue of accepting an `Into` though, callers can provide more than just a haystack. Indeed, the [`Input`] type has more details, but briefly, callers can use it to configure various aspects of the search: * The span of the haystack to search via [`Input::span`] or [`Input::range`], which might be a substring of the haystack. * Whether to run an anchored search or not via [`Input::anchored`]. This permits one to require matches to start at the same offset that the search started. * Whether to ask the regex engine to stop as soon as a match is seen via [`Input::earliest`]. This can be used to find the offset of a match as soon as it is known without waiting for the full leftmost-first match to be found. This can also be used to avoid the worst case `O(m * n^2)` time complexity of iteration. Some lower level search routines accept an `&Input` for performance reasons. In which case, `&Input::new("haystack")` can be used for a simple search. ### Error reporting Most, but not all, regex engines in this crate can fail to execute a search. When a search fails, callers cannot determine whether or not a match exists. That is, the result is indeterminate. Search failure, in all cases in this crate, is represented by a [`MatchError`]. Routines that can fail start with the `try_` prefix in their name. For example, [`hybrid::regex::Regex::try_search`] can fail for a number of reasons. Conversely, routines that either can't fail or can panic on failure lack the `try_` prefix. For example, [`hybrid::regex::Regex::find`] will panic in cases where [`hybrid::regex::Regex::try_search`] would return an error, and [`meta::Regex::find`] will never panic. Therefore, callers need to pay close attention to the panicking conditions in the documentation. In most cases, the reasons that a search fails are either predictable or configurable, albeit at some additional cost. An example of predictable failure is [`BoundedBacktracker::try_search`](nfa::thompson::backtrack::BoundedBacktracker::try_search). Namely, it fails whenever the multiplication of the haystack, the regex and some constant exceeds the [configured visited capacity](nfa::thompson::backtrack::Config::visited_capacity). Callers can predict the failure in terms of haystack length via the [`BoundedBacktracker::max_haystack_len`](nfa::thompson::backtrack::BoundedBacktracker::max_haystack_len) method. While this form of failure is technically avoidable by increasing the visited capacity, it isn't practical to do so for all inputs because the memory usage required for larger haystacks becomes impractically large. So in practice, if one is using the bounded backtracker, you really do have to deal with the failure. An example of configurable failure happens when one enables heuristic support for Unicode word boundaries in a DFA. Namely, since the DFAs in this crate (except for the one-pass DFA) do not support Unicode word boundaries on non-ASCII haystacks, building a DFA from an NFA that contains a Unicode word boundary will itself fail. However, one can configure DFAs to still be built in this case by [configuring heuristic support for Unicode word boundaries](hybrid::dfa::Config::unicode_word_boundary). If the NFA the DFA is built from contains a Unicode word boundary, then the DFA will still be built, but special transitions will be added to every state that cause the DFA to fail if any non-ASCII byte is seen. This failure happens at search time and it requires the caller to opt into this. There are other ways for regex engines to fail in this crate, but the above two should represent the general theme of failures one can find. Dealing with these failures is, in part, one the responsibilities of the [meta regex engine](meta). Notice, for example, that the meta regex engine exposes an API that never returns an error nor panics. It carefully manages all of the ways in which the regex engines can fail and either avoids the predictable ones entirely (e.g., the bounded backtracker) or reacts to configured failures by falling back to a different engine (e.g., the lazy DFA quitting because it saw a non-ASCII byte). ### Configuration and Builders Most of the regex engines in this crate come with two types to facilitate building the regex engine: a `Config` and a `Builder`. A `Config` is usually specific to that particular regex engine, but other objects such as parsing and NFA compilation have `Config` types too. A `Builder` is the thing responsible for taking inputs (either pattern strings or already-parsed patterns or even NFAs directly) and turning them into an actual regex engine that can be used for searching. The main reason why building a regex engine is a bit complicated is because of the desire to permit composition with de-coupled components. For example, you might want to [manually construct a Thompson NFA](nfa::thompson::Builder) and then build a regex engine from it without ever using a regex parser at all. On the other hand, you might also want to build a regex engine directly from the concrete syntax. This demonstrates why regex engine construction is so flexible: it needs to support not just convenient construction, but also construction from parts built elsewhere. This is also in turn why there are many different `Config` structs in this crate. Let's look more closely at an example: [`hybrid::regex::Builder`]. It accepts three different `Config` types for configuring construction of a lazy DFA regex: * [`hybrid::regex::Builder::syntax`] accepts a [`util::syntax::Config`] for configuring the options found in the [`regex-syntax`](regex_syntax) crate. For example, whether to match case insensitively. * [`hybrid::regex::Builder::thompson`] accepts a [`nfa::thompson::Config`] for configuring construction of a [Thompson NFA](nfa::thompson::NFA). For example, whether to build an NFA that matches the reverse language described by the regex. * [`hybrid::regex::Builder::dfa`] accept a [`hybrid::dfa::Config`] for configuring construction of the pair of underlying lazy DFAs that make up the lazy DFA regex engine. For example, changing the capacity of the cache used to store the transition table. The lazy DFA regex engine uses all three of those configuration objects for methods like [`hybrid::regex::Builder::build`], which accepts a pattern string containing the concrete syntax of your regex. It uses the syntax configuration to parse it into an AST and translate it into an HIR. Then the NFA configuration when compiling the HIR into an NFA. And then finally the DFA configuration when lazily determinizing the NFA into a DFA. Notice though that the builder also has a [`hybrid::regex::Builder::build_from_dfas`] constructor. This permits callers to build the underlying pair of lazy DFAs themselves (one for the forward searching to find the end of a match and one for the reverse searching to find the start of a match), and then build the regex engine from them. The lazy DFAs, in turn, have their own builder that permits [construction directly from a Thompson NFA](hybrid::dfa::Builder::build_from_nfa). Continuing down the rabbit hole, a Thompson NFA has its own compiler that permits [construction directly from an HIR](nfa::thompson::Compiler::build_from_hir). The lazy DFA regex engine builder lets you follow this rabbit hole all the way down, but also provides convenience routines that do it for you when you don't need precise control over every component. The [meta regex engine](meta) is a good example of something that utilizes the full flexibility of these builders. It often needs not only precise control over each component, but also shares them across multiple regex engines. (Most sharing is done by internal reference accounting. For example, an [`NFA`](nfa::thompson::NFA) is reference counted internally which makes cloning cheap.) ### Size limits Unlike the `regex` crate, the `regex-automata` crate specifically does not enable any size limits by default. That means users of this crate need to be quite careful when using untrusted patterns. Namely, because bounded repetitions can grow exponentially by stacking them, it is possible to build a very large internal regex object from just a small pattern string. For example, the NFA built from the pattern `a{10}{10}{10}{10}{10}{10}{10}` is over 240MB. There are multiple size limit options in this crate. If one or more size limits are relevant for the object you're building, they will be configurable via methods on a corresponding `Config` type. # Crate features This crate has a dizzying number of features. The main idea is to be able to control how much stuff you pull in for your specific use case, since the full crate is quite large and can dramatically increase compile times and binary size. The most barebones but useful configuration is to disable all default features and enable only `dfa-search`. This will bring in just the DFA deserialization and search routines without any dependency on `std` or `alloc`. This does require generating and serializing a DFA, and then storing it somewhere, but it permits regex searches in freestanding or embedded environments. Because there are so many features, they are split into a few groups. The default set of features is: `std`, `syntax`, `perf`, `unicode`, `meta`, `nfa`, `dfa` and `hybrid`. Basically, the default is to enable everything except for development related features like `logging`. ### Ecosystem features * **std** - Enables use of the standard library. In terms of APIs, this usually just means that error types implement the `std::error::Error` trait. Otherwise, `std` sometimes enables the code to be faster, for example, using a `HashMap` instead of a `BTreeMap`. (The `std` feature matters more for dependencies like `aho-corasick` and `memchr`, where `std` is required to enable certain classes of SIMD optimizations.) Enabling `std` automatically enables `alloc`. * **alloc** - Enables use of the `alloc` library. This is required for most APIs in this crate. The main exception is deserializing and searching with fully compiled DFAs. * **logging** - Adds a dependency on the `log` crate and makes this crate emit log messages of varying degrees of utility. The log messages are especially useful in trying to understand what the meta regex engine is doing. ### Performance features **Note**: To get performance benefits offered by the SIMD, `std` must be enabled. None of the `perf-*` features will enable `std` implicitly. * **perf** - Enables all of the below features. * **perf-inline** - When enabled, `inline(always)` is used in (many) strategic locations to help performance at the expense of longer compile times and increased binary size. * **perf-literal** - Enables all literal related optimizations. * **perf-literal-substring** - Enables all single substring literal optimizations. This includes adding a dependency on the `memchr` crate. * **perf-literal-multisubstring** - Enables all multiple substring literal optimizations. This includes adding a dependency on the `aho-corasick` crate. ### Unicode features * **unicode** - Enables all Unicode features. This feature is enabled by default, and will always cover all Unicode features, even if more are added in the future. * **unicode-age** - Provide the data for the [Unicode `Age` property](https://www.unicode.org/reports/tr44/tr44-24.html#Character_Age). This makes it possible to use classes like `\p{Age:6.0}` to refer to all codepoints first introduced in Unicode 6.0 * **unicode-bool** - Provide the data for numerous Unicode boolean properties. The full list is not included here, but contains properties like `Alphabetic`, `Emoji`, `Lowercase`, `Math`, `Uppercase` and `White_Space`. * **unicode-case** - Provide the data for case insensitive matching using [Unicode's "simple loose matches" specification](https://www.unicode.org/reports/tr18/#Simple_Loose_Matches). * **unicode-gencat** - Provide the data for [Unicode general categories](https://www.unicode.org/reports/tr44/tr44-24.html#General_Category_Values). This includes, but is not limited to, `Decimal_Number`, `Letter`, `Math_Symbol`, `Number` and `Punctuation`. * **unicode-perl** - Provide the data for supporting the Unicode-aware Perl character classes, corresponding to `\w`, `\s` and `\d`. This is also necessary for using Unicode-aware word boundary assertions. Note that if this feature is disabled, the `\s` and `\d` character classes are still available if the `unicode-bool` and `unicode-gencat` features are enabled, respectively. * **unicode-script** - Provide the data for [Unicode scripts and script extensions](https://www.unicode.org/reports/tr24/). This includes, but is not limited to, `Arabic`, `Cyrillic`, `Hebrew`, `Latin` and `Thai`. * **unicode-segment** - Provide the data necessary to provide the properties used to implement the [Unicode text segmentation algorithms](https://www.unicode.org/reports/tr29/). This enables using classes like `\p{gcb=Extend}`, `\p{wb=Katakana}` and `\p{sb=ATerm}`. * **unicode-word-boundary** - Enables support for Unicode word boundaries, i.e., `\b`, in regexes. When this and `unicode-perl` are enabled, then data tables from `regex-syntax` are used to implement Unicode word boundaries. However, if `regex-syntax` isn't enabled as a dependency then one can still enable this feature. It will cause `regex-automata` to bundle its own data table that would otherwise be redundant with `regex-syntax`'s table. ### Regex engine features * **syntax** - Enables a dependency on `regex-syntax`. This makes APIs for building regex engines from pattern strings available. Without the `regex-syntax` dependency, the only way to build a regex engine is generally to deserialize a previously built DFA or to hand assemble an NFA using its [builder API](nfa::thompson::Builder). Once you have an NFA, you can build any of the regex engines in this crate. The `syntax` feature also enables `alloc`. * **meta** - Enables the meta regex engine. This also enables the `syntax` and `nfa-pikevm` features, as both are the minimal requirements needed. The meta regex engine benefits from enabling any of the other regex engines and will use them automatically when appropriate. * **nfa** - Enables all NFA related features below. * **nfa-thompson** - Enables the Thompson NFA APIs. This enables `alloc`. * **nfa-pikevm** - Enables the PikeVM regex engine. This enables `nfa-thompson`. * **nfa-backtrack** - Enables the bounded backtracker regex engine. This enables `nfa-thompson`. * **dfa** - Enables all DFA related features below. * **dfa-build** - Enables APIs for determinizing DFAs from NFAs. This enables `nfa-thompson` and `dfa-search`. * **dfa-search** - Enables APIs for searching with DFAs. * **dfa-onepass** - Enables the one-pass DFA API. This enables `nfa-thompson`. * **hybrid** - Enables the hybrid NFA/DFA or "lazy DFA" regex engine. This enables `alloc` and `nfa-thompson`.   ]j^j\jZj!Xj$Wj$Uj%Sj%Qj%Oj%Mj%Kj%Ij%Gj%Ej%Xk[k^kVkKk"t\kT4// / / / /, / / /, /8/& /8/ /*/ / /+ / / /8/ / /8</ /7loggingL/ / /, / / /, /8/ //,/ / / / /, / / /, /8/& /8/ /*/ / /8/ / / /8/'/8,/ / / /, / / /, /8/ //,/ / / / /, / / /, /8/& /8/ /*/ / /8/ / / /8/'/8,/ / / /, / / /, /8/ /\ss A module for building and searching with deterministic finite automata (DFAs). Like other modules in this crate, DFAs support a rich regex syntax with Unicode features. DFAs also have extensive options for configuring the best space vs time trade off for your use case and provides support for cheap deserialization of automata for use in `no_std` environments. If you're looking for lazy DFAs that build themselves incrementally during search, then please see the top-level [`hybrid` module](crate::hybrid). # Overview This section gives a brief overview of the primary types in this module: * A [`regex::Regex`] provides a way to search for matches of a regular expression using DFAs. This includes iterating over matches with both the start and end positions of each match. * A [`dense::DFA`] provides low level access to a DFA that uses a dense representation (uses lots of space, but fast searching). * A [`sparse::DFA`] provides the same API as a `dense::DFA`, but uses a sparse representation (uses less space, but slower searching). * An [`Automaton`] trait that defines an interface that both dense and sparse DFAs implement. (A `regex::Regex` is generic over this trait.) * Both dense DFAs and sparse DFAs support serialization to raw bytes (e.g., [`dense::DFA::to_bytes_little_endian`]) and cheap deserialization (e.g., [`dense::DFA::from_bytes`]). There is also a [`onepass`] module that provides a [one-pass DFA](onepass::DFA). The unique advantage of this DFA is that, for the class of regexes it can be built with, it supports reporting the spans of matching capturing groups. It is the only DFA in this crate capable of such a thing. # Example: basic regex searching This example shows how to compile a regex using the default configuration and then use it to find matches in a byte string: ``` use regex_automata::{Match, dfa::regex::Regex}; let re = Regex::new(r"[0-9]{4}-[0-9]{2}-[0-9]{2}")?; let text = b"2018-12-24 2016-10-08"; let matches: Vec = re.find_iter(text).collect(); assert_eq!(matches, vec![ Match::must(0, 0..10), Match::must(0, 11..21), ]); # Ok::<(), Box>(()) ``` # Example: searching with regex sets The DFAs in this module all fully support searching with multiple regexes simultaneously. You can use this support with standard leftmost-first style searching to find non-overlapping matches: ``` # if cfg!(miri) { return Ok(()); } // miri takes too long use regex_automata::{Match, dfa::regex::Regex}; let re = Regex::new_many(&[r"\w+", r"\S+"])?; let text = b"@foo bar"; let matches: Vec = re.find_iter(text).collect(); assert_eq!(matches, vec![ Match::must(1, 0..4), Match::must(0, 5..8), ]); # Ok::<(), Box>(()) ``` # Example: use sparse DFAs By default, compiling a regex will use dense DFAs internally. This uses more memory, but executes searches more quickly. If you can abide slower searches (somewhere around 3-5x), then sparse DFAs might make more sense since they can use significantly less space. Using sparse DFAs is as easy as using `Regex::new_sparse` instead of `Regex::new`: ``` use regex_automata::{Match, dfa::regex::Regex}; let re = Regex::new_sparse(r"[0-9]{4}-[0-9]{2}-[0-9]{2}").unwrap(); let text = b"2018-12-24 2016-10-08"; let matches: Vec = re.find_iter(text).collect(); assert_eq!(matches, vec![ Match::must(0, 0..10), Match::must(0, 11..21), ]); # Ok::<(), Box>(()) ``` If you already have dense DFAs for some reason, they can be converted to sparse DFAs and used to build a new `Regex`. For example: ``` use regex_automata::{Match, dfa::regex::Regex}; let dense_re = Regex::new(r"[0-9]{4}-[0-9]{2}-[0-9]{2}").unwrap(); let sparse_re = Regex::builder().build_from_dfas( dense_re.forward().to_sparse()?, dense_re.reverse().to_sparse()?, ); let text = b"2018-12-24 2016-10-08"; let matches: Vec = sparse_re.find_iter(text).collect(); assert_eq!(matches, vec![ Match::must(0, 0..10), Match::must(0, 11..21), ]); # Ok::<(), Box>(()) ``` # Example: deserialize a DFA This shows how to first serialize a DFA into raw bytes, and then deserialize those raw bytes back into a DFA. While this particular example is a bit contrived, this same technique can be used in your program to deserialize a DFA at start up time or by memory mapping a file. ``` use regex_automata::{Match, dfa::{dense, regex::Regex}}; let re1 = Regex::new(r"[0-9]{4}-[0-9]{2}-[0-9]{2}").unwrap(); // serialize both the forward and reverse DFAs, see note below let (fwd_bytes, fwd_pad) = re1.forward().to_bytes_native_endian(); let (rev_bytes, rev_pad) = re1.reverse().to_bytes_native_endian(); // now deserialize both---we need to specify the correct type! let fwd: dense::DFA<&[u32]> = dense::DFA::from_bytes(&fwd_bytes[fwd_pad..])?.0; let rev: dense::DFA<&[u32]> = dense::DFA::from_bytes(&rev_bytes[rev_pad..])?.0; // finally, reconstruct our regex let re2 = Regex::builder().build_from_dfas(fwd, rev); // we can use it like normal let text = b"2018-12-24 2016-10-08"; let matches: Vec = re2.find_iter(text).collect(); assert_eq!(matches, vec![ Match::must(0, 0..10), Match::must(0, 11..21), ]); # Ok::<(), Box>(()) ``` There are a few points worth noting here: * We need to extract the raw DFAs used by the regex and serialize those. You can build the DFAs manually yourself using [`dense::Builder`], but using the DFAs from a `Regex` guarantees that the DFAs are built correctly. (In particular, a `Regex` constructs a reverse DFA for finding the starting location of matches.) * To convert the DFA to raw bytes, we use the `to_bytes_native_endian` method. In practice, you'll want to use either [`dense::DFA::to_bytes_little_endian`] or [`dense::DFA::to_bytes_big_endian`], depending on which platform you're deserializing your DFA from. If you intend to deserialize on either platform, then you'll need to serialize both and deserialize the right one depending on your target's endianness. * Safely deserializing a DFA requires verifying the raw bytes, particularly if they are untrusted, since an invalid DFA could cause logical errors, panics or even undefined behavior. This verification step requires visiting all of the transitions in the DFA, which can be costly. If cheaper verification is desired, then [`dense::DFA::from_bytes_unchecked`] is available that only does verification that can be performed in constant time. However, one can only use this routine if the caller can guarantee that the bytes provided encoded a valid DFA. The same process can be achieved with sparse DFAs as well: ``` use regex_automata::{Match, dfa::{sparse, regex::Regex}}; let re1 = Regex::new(r"[0-9]{4}-[0-9]{2}-[0-9]{2}").unwrap(); // serialize both let fwd_bytes = re1.forward().to_sparse()?.to_bytes_native_endian(); let rev_bytes = re1.reverse().to_sparse()?.to_bytes_native_endian(); // now deserialize both---we need to specify the correct type! let fwd: sparse::DFA<&[u8]> = sparse::DFA::from_bytes(&fwd_bytes)?.0; let rev: sparse::DFA<&[u8]> = sparse::DFA::from_bytes(&rev_bytes)?.0; // finally, reconstruct our regex let re2 = Regex::builder().build_from_dfas(fwd, rev); // we can use it like normal let text = b"2018-12-24 2016-10-08"; let matches: Vec = re2.find_iter(text).collect(); assert_eq!(matches, vec![ Match::must(0, 0..10), Match::must(0, 11..21), ]); # Ok::<(), Box>(()) ``` Note that unlike dense DFAs, sparse DFAs have no alignment requirements. Conversely, dense DFAs must be aligned to the same alignment as a [`StateID`](crate::util::primitives::StateID). # Support for `no_std` and `alloc`-only This crate comes with `alloc` and `std` features that are enabled by default. When the `alloc` or `std` features are enabled, the API of this module will include the facilities necessary for compiling, serializing, deserializing and searching with DFAs. When only the `alloc` feature is enabled, then implementations of the `std::error::Error` trait are dropped, but everything else generally remains the same. When both the `alloc` and `std` features are disabled, the API of this module will shrink such that it only includes the facilities necessary for deserializing and searching with DFAs. The intended workflow for `no_std` environments is thus as follows: * Write a program with the `alloc` or `std` features that compiles and serializes a regular expression. You may need to serialize both little and big endian versions of each DFA. (So that's 4 DFAs in total for each regex.) * In your `no_std` environment, follow the examples above for deserializing your previously serialized DFAs into regexes. You can then search with them as you would any regex. Deserialization can happen anywhere. For example, with bytes embedded into a binary or with a file memory mapped at runtime. The `regex-cli` command (found in the same repository as this crate) can be used to serialize DFAs to files and generate Rust code to read them. # Syntax This module supports the same syntax as the `regex` crate, since they share the same parser. You can find an exhaustive list of supported syntax in the [documentation for the `regex` crate](https://docs.rs/regex/1/regex/#syntax). There are two things that are not supported by the DFAs in this module: * Capturing groups. The DFAs (and [`Regex`](regex::Regex)es built on top of them) can only find the offsets of an entire match, but cannot resolve the offsets of each capturing group. This is because DFAs do not have the expressive power necessary. * Unicode word boundaries. These present particularly difficult challenges for DFA construction and would result in an explosion in the number of states. One can enable [`dense::Config::unicode_word_boundary`] though, which provides heuristic support for Unicode word boundaries that only works on ASCII text. Otherwise, one can use `(?-u:\b)` for an ASCII word boundary, which will work on any input. There are no plans to lift either of these limitations. Note that these restrictions are identical to the restrictions on lazy DFAs. # Differences with general purpose regexes The main goal of the [`regex`](https://docs.rs/regex) crate is to serve as a general purpose regular expression engine. It aims to automatically balance low compile times, fast search times and low memory usage, while also providing a convenient API for users. In contrast, this module provides a lower level regular expression interface based exclusively on DFAs that is a bit less convenient while providing more explicit control over memory usage and search times. Here are some specific negative differences: * **Compilation can take an exponential amount of time and space** in the size of the regex pattern. While most patterns do not exhibit worst case exponential time, such patterns do exist. For example, `[01]*1[01]{N}` will build a DFA with approximately `2^(N+2)` states. For this reason, untrusted patterns should not be compiled with this module. (In the future, the API may expose an option to return an error if the DFA gets too big.) * This module does not support sub-match extraction via capturing groups, which can be achieved with the regex crate's "captures" API. * While the regex crate doesn't necessarily sport fast compilation times, the regexes in this module are almost universally slow to compile, especially when they contain large Unicode character classes. For example, on my system, compiling `\w{50}` takes about 1 second and almost 15MB of memory! (Compiling a sparse regex takes about the same time but only uses about 1.2MB of memory.) Conversely, compiling the same regex without Unicode support, e.g., `(?-u)\w{50}`, takes under 1 millisecond and about 15KB of memory. For this reason, you should only use Unicode character classes if you absolutely need them! (They are enabled by default though.) * This module does not support Unicode word boundaries. ASCII word boundaries may be used though by disabling Unicode or selectively doing so in the syntax, e.g., `(?-u:\b)`. There is also an option to [heuristically enable Unicode word boundaries](crate::dfa::dense::Config::unicode_word_boundary), where the corresponding DFA will give up if any non-ASCII byte is seen. * As a lower level API, this module does not do literal optimizations automatically. Although it does provide hooks in its API to make use of the [`Prefilter`](crate::util::prefilter::Prefilter) trait. Missing literal optimizations means that searches may run much slower than what you're accustomed to, although, it does provide more predictable and consistent performance. * There is no `&str` API like in the regex crate. In this module, all APIs operate on `&[u8]`. By default, match indices are guaranteed to fall on UTF-8 boundaries, unless either of [`syntax::Config::utf8`](crate::util::syntax::Config::utf8) or [`thompson::Config::utf8`](crate::nfa::thompson::Config::utf8) are disabled. With some of the downsides out of the way, here are some positive differences: * Both dense and sparse DFAs can be serialized to raw bytes, and then cheaply deserialized. Deserialization can be done in constant time with the unchecked APIs, since searching can be performed directly on the raw serialized bytes of a DFA. * This module was specifically designed so that the searching phase of a DFA has minimal runtime requirements, and can therefore be used in `no_std` environments. While `no_std` environments cannot compile regexes, they can deserialize pre-compiled regexes. * Since this module builds DFAs ahead of time, it will generally out-perform the `regex` crate on equivalent tasks. The performance difference is likely not large. However, because of a complex set of optimizations in the regex crate (like literal optimizations), an accurate performance comparison may be difficult to do. * Sparse DFAs provide a way to build a DFA ahead of time that sacrifices search performance a small amount, but uses much less storage space. Potentially even less than what the regex crate uses. * This module exposes DFAs directly, such as [`dense::DFA`] and [`sparse::DFA`], which enables one to do less work in some cases. For example, if you only need the end of a match and not the start of a match, then you can use a DFA directly without building a `Regex`, which always requires a second DFA to find the start of a match. * This module provides more control over memory usage. Aside from choosing between dense and sparse DFAs, one can also choose a smaller state identifier representation to use less space. Also, one can enable DFA minimization via [`dense::Config::minimize`], but it can increase compilation times dramatically.  dfa-search1<HLs̅sTtLtTs,sss,p,*(r,><t,RPt,hfw,$w lx Types and routines specific to dense DFAs. This module is the home of [`dense::DFA`](DFA). This module also contains a [`dense::Builder`](Builder) and a [`dense::Config`](Config) for building and configuring a dense DFA. ,xx,xKLvc$8< ,iWD D=C,"N\#SL$D%54&D'\ )L%*4-N /HL0T1<2L3$ 4$ 7¬\8<9<;,<L!=L!?<"@D%A$C,#Dd#E$G$H4It$JL dfa-build<$03#&d  ,,,*,+)=,,A??/,WUBD,nl5,,,d,,*dT,B@g,YWk,pnn^,ql,u,,,,l<twfi\O?B=D03"% ,$ ,   < \,NOPQRSTUVWXYZ[\]^_`abcefgh:7 Return a new default dense DFA compiler configuration.BMMҧ !0 Enable state acceleration.GD When enabled, DFA construction will analyze each state to determineJG whether it is eligible for simple acceleration. Acceleration typicallyJG occurs when most of a state's transitions loop back to itself, leavingGD only a select few bytes that will exit the state. When this occurs,JG other routines like `memchr` can be used to look for those bytes which/, may be much faster than traversing the DFA.GD Callers may elect to disable this if consistent performance is moreKH desirable than variable performance. Namely, acceleration can sometimesGD make searching slower than it otherwise would be if the transitions;8 that leave accelerated states are traversed frequently.  2/ See [`Automaton::accelerator`] for an example.   This is enabled by default.T!BMM O$!!8<"A> Set a prefilter to be used whenever a start state is entered.""FC A [`Prefilter`] in this context is meant to accelerate searches by#GD looking for literal prefixes that every match for the corresponding#FC pattern (or patterns) must start with. Once a prefilter produces a$HE match, the underlying search routine continues on to try and confirmt$ the match.%%IF Be warned that setting a prefilter does not guarantee that the search%C@ will be faster. While it's usually a good bet, if the prefilter&HE produces a lot of false positive candidates (i.e., positions matched&GD by the prefilter but not by the regex), then the overall result can'HE be slower than if you had just executed the regex engine without any|( prefilters.((A> Note that unless [`Config::specialize_start_states`] has been(EB explicitly set, then setting this will also enable (when `pre` is)IF `Some`) or disable (when `pre` is `None`) start state specialization.*GD This occurs because without start state specialization, a prefilter*HE is likely to be less effective. And without a prefilter, start state+(% specialization is usually pointless.++B? **WARNING:** Note that prefilters are not preserved as part of,=: serialization. Serializing a DFA will drop its prefilter.,,# By default no prefilter is set.-l- # Example-<- ```- use regex_automata::{-%" dfa::{dense::DFA, Automaton},.# util::prefilter::Prefilter,.$! Input, HalfMatch, MatchKind,4. };..HE let pre = Prefilter::new(MatchKind::LeftmostFirst, &["foo", "bar"]);ܽ/ let re = DFA::builder()/0- .configure(DFA::config().prefilter(pre))0$! .build(r"(foo|bar)[a-z]+")?;0.+ let input = Input::new("foo1 barfox bar");|0 assert_eq!(1%" Some(HalfMatch::must(0, 11)),1# re.try_search_fwd(&input)?,41 );11.+ # Ok::<(), Box>(())<2-22FC Be warned though that an incorrect prefilter can lead to incorrectd2 results!3<3-̞3-3%-3#ʼn-4$-44-44HE let pre = Prefilter::new(MatchKind::LeftmostFirst, &["foo", "car"]);ܗ5-50-5$-6.-|6Ќ-6>; // No match reported even though there clearly is one!l7 None,7#-47č-77.-<8-L8BMM $88M.: Minimize the DFA.::JG When enabled, the DFA built will be minimized such that it is as small; as possible.;;IF Whether one enables minimization or not depends on the types of costsHE size. In practice, minimization can be quite costly in terms of both>GD space and time, so it should only be done if you're willing to wait?HE longer to produce a DFA. In general, you might want a minimal DFA in?  the following circumstances:@@IF 1. You would like to optimize for the size of the automaton. This can@EB manifest in one of two ways. Firstly, if you're converting theAKH DFA into Rust code (or a table embedded in the code), then a minimalAKH DFA will translate into a corresponding reduction in code size, andBFC thus, also the final compiled binary size. Secondly, if you areCIF building many DFAs and putting them on the heap, you'll be able toCHE fit more if they are smaller. Note though that building a minimalDGD DFA itself requires additional space; you only realize the spaceEGD savings once the minimal DFA is constructed (at which point, theE-* space used for minimization is freed).EA> 2. You've observed that a smaller DFA results in faster matchFDA performance. Naively, this isn't guaranteed since there is noGFC inherent difference between matching with a bigger-than-minimalGIF DFA and a minimal DFA. However, a smaller DFA may make use of yourH$! CPU's cache more efficiently.HA> 3. You are trying to establish an equivalence between regularIHE languages. The standard method for this is to build a minimal DFAIJG for each language and then compare them. If the DFAs are equivalentJA> (up to state renaming), then the languages are equivalent.JJHE Typically, minimization only makes sense as an offline process. ThatKDA is, one might minimize a DFA before serializing it to persistentLIF storage. In practical terms, minimization can take around an order ofLKH magnitude more time than compiling the initial DFA via determinization.MM'$ This option is disabled by default.DMBMM Q$MNv6N$! Set the desired match semantics.NOIF The default is [`MatchKind::LeftmostFirst`], which corresponds to theOFC match semantics of Perl-like regex engines. That is, when multiplePHE patterns would match at the same leftmost position, the pattern thatP30 appears first in the concrete syntax is chosen.QQB? Currently, the only other kind of match semantics supported isQFC [`MatchKind::All`]. This corresponds to classical DFA constructionR41 where all possible matches are added to the DFA.RREB Typically, `All` is used when one wants to execute an overlappingSHE search and `LeftmostFirst` otherwise. In particular, it rarely makesTKH sense to use `All` with the various "leftmost" find routines, since theTIF leftmost routines depend on the `LeftmostFirst` automata constructionUGD strategy. Specifically, `LeftmostFirst` adds dead states to the DFAUHE as a way to terminate the search and report a match. `LeftmostFirst`VGD also supports non-greedy matches using this strategy where as `All`lW does not.WW! # Example: overlapping searchWWGD This example shows the typical use of `MatchKind::All`, which is toX report overlapping matches.X; The type of starting state configuration to use for a DFA.wxHE By default, the starting state configuration is [`StartKind::Both`].xlx-x // will panic if you try to use it with an unanchored search.~ let dfa = DFA::builder()A> .configure(DFA::config().start_kind(StartKind::Anchored)) .build(r"[0-9]+")?;$- .range(8..)Ā!-<-.-<-TBMM $$=EB Whether to compile a separate start state for each pattern in thet automaton.ƃ΃GD When enabled, a separate **anchored** start state is added for eachHE pattern in the DFA. When this start state is used, then the DFA willHE only search for matches for the pattern specified, even if there are other patterns in the DFA.ׅ߅HE The main downside of this option is that it can potentially increaseKH the size of the DFA and/or increase the time it takes to build the DFA.HE There are a few reasons one might want to enable this (it's disabledч by default):B? 1. When looking for the start of an overlapping match (using aIF reverse DFA), doing it correctly requires starting the reverse searchGD using the starting state of the pattern that matched in the forwardωKH direction. Indeed, when building a [`Regex`](crate::dfa::regex::Regex),JG it will automatically enable this option when building the reverse DFA| internally.GD 2. When you want to use a DFA with multiple patterns to both search΋GD for matches of any pattern or to search for anchored matches of oneKH particular pattern while using the same DFA. (Otherwise, you would need+( to compile a new DFA for each pattern.)IF 3. Since the start states added for each pattern are anchored, if youGD compile an unanchored DFA with one pattern while also enabling thisKH option, then you can use the same DFA to perform anchored or unanchoredJG searches. The latter you get with the standard search APIs. The formerӏJG you get from the various `_at` search methods that allow you specify a좐 pattern ID to search for.Đ̐  By default this is disabled.l-GD This example shows how to use this option to permit the same DFA toߑC@ run both anchored and unanchored searches for a single pattern.<-̻-ْ  dfa::{dense, Automaton},.+ Anchored, HalfMatch, PatternID, Input,4-ē#-FC .configure(dense::Config::new().starts_for_each_pattern(true)) .build(r"foo[0-9]+")?;ڔ!-KH // Here's a normal unanchored search. Notice that we use 'None' for theؕGD // pattern ID. Since the DFA was built as an unanchored machine, it1. // use its default unanchored starting state.ږ*-%-<-HE // But now if we explicitly specify the pattern to search ('0' beingHE // the only pattern in the DFA), then it will use the starting stateDA // for that specific pattern which is always anchored. Since theיHE // pattern doesn't have a match at the beginning of the haystack, we // find nothing.$-96 .anchored(Anchored::Pattern(PatternID::must(0)));2/ assert_eq!(None, dfa.try_search_fwd(&input)?);ۛJG // And finally, an anchored search is not the same as putting a '^' atIF // beginning of the pattern. An anchored search can only match at the63 // beginning of the *search*, which we can change:$-ܝ85 .anchored(Anchored::Pattern(PatternID::must(0))) .range(5..);<-.-<-BMM T$ݟ2ĠGD Whether to attempt to shrink the size of the DFA's alphabet or not.IF This option is enabled by default and should never be disabled unless%" one is debugging a generated DFA.IF When enabled, the DFA will use a map from all possible bytes to theirHE corresponding equivalence class. Each equivalence class represents aKH set of bytes that does not discriminate between a match and a non-matchA> in the DFA. For example, the pattern `[ab]+` has at least twoɤJG equivalence classes: a set containing `a` and `b` and a set containingB? every byte except for `a` and `b`. `a` and `b` are in the sameߥKH equivalence class because they never discriminate between a match and at non-match.¦ʦFC The advantage of this map is that the size of the transition tableHE can be reduced drastically from `#states * 256 * sizeof(StateID)` toJG `#states * k * sizeof(StateID)` where `k` is the number of equivalenceFC classes (rounded up to the nearest power of 2). As a result, totalEB space usage can decrease substantially. Moreover, since a smallerƩ=: alphabet is used, DFA compilation becomes faster as well.GD **WARNING:** This is only useful for debugging DFAs. Disabling thisܪB? does not yield any speed advantages. Namely, even when this isFC disabled, a byte class map is still used while searching. The onlyFC difference is that every byte will be forced into its own distinctHE equivalence class. This is useful for debugging the actual generatedIF transitions because it lets one see the transitions defined on actualԭ-* bytes instead of the equivalence classes.dBMM U$;1. Heuristically enable Unicode word boundaries.JG When set, this will attempt to implement Unicode word boundaries as ifJG they were ASCII word boundaries. This only works when the search inputְJG is ASCII only. If a non-ASCII byte is observed while searching, then aDA [`MatchError::quit`](crate::MatchError::quit) error is returned.FC A possible alternative to enabling this option is to simply use anJG ASCII word boundary, e.g., via `(?-u:\b)`. The main reason to use thisJG option is if you absolutely need Unicode support. This option lets one߳FC use a fast search implementation (a DFA) for some potentially veryGD common cases, while providing the option to fall back to some otherFC regex engine to handle the general case when an error is returned.ɵIF If the pattern provided has no Unicode word boundary in it, then thisEB option has no effect. (That is, quitting on a non-ASCII byte onlyGD occurs when this option is enabled _and_ a Unicode word boundary is䭷 present in the pattern.)ηַGD This is almost equivalent to setting all non-ASCII bytes to be quitIF bytes. The only difference is that this will cause non-ASCII bytes toGD be quit bytes _only_ when a Unicode word boundary is present in thed pattern.͹չC@ When enabling this option, callers _must_ be prepared to handle<9 a [`MatchError`](crate::MatchError) error during search.޺FC When using a [`Regex`](crate::dfa::regex::Regex), this corresponds;8 to using the `try_` suite of methods. Alternatively, if@= callers can guarantee that their input is ASCII only, then aEB [`MatchError::quit`](crate::MatchError::quit) error will never be returned while searching.  This is disabled by default.ǽlϽ-JG This example shows how to heuristically enable Unicode word boundariesIF in a pattern. It also shows what happens when a search comes across a non-ASCII byte.<-̲-п  dfa::{Automaton, dense},%" HalfMatch, Input, MatchError,4-#-DA .configure(dense::Config::new().unicode_word_boundary(true)) .build(r"\b[0-9]+\b")?;C@ // The match occurs before the search ever observes the snowman%" // character, so no error occurs.-* let haystack = "foo 123 ☃".as_bytes();/, let expected = Some(HalfMatch::must(0, 7));96 let got = dfa.try_search_fwd(&Input::new(haystack))?; assert_eq!(expected, got);GD // Notice that this search fails, even though the snowman characterC@ // occurs after the ending match offset. This is because searchGD // routines read one byte past the end of the search to account forJG // look-around, and indeed, this is required here to determine whether // the trailing \b matches.,) let haystack = "foo 123 ☃".as_bytes();-* let expected = MatchError::quit(0xE2, 8);85 let got = dfa.try_search_fwd(&Input::new(haystack));# assert_eq!(Err(expected), got);KH // Another example is executing a search where the span of the haystackKH // we specify is all ASCII, but there is non-ASCII just before it. This'$ // correctly also reports an error./, let input = Input::new("β123").range(2..);-* let expected = MatchError::quit(0xB2, 1);)& let got = dfa.try_search_fwd(&input);#.41 // And similarly for the trailing word boundary./, let input = Input::new("123β").range(..3);-* let expected = MatchError::quit(0xCE, 3);).#..-<-BMM V$4! Add a "quit" byte to the DFA.HE When a quit byte is seen during search time, then search will returnHE a [`MatchError::quit`](crate::MatchError::quit) error indicating the'$ offset at which the search stopped.FC A quit byte will always overrule any other aspects of a regex. ForJG example, if the `x` byte is added as a quit byte and the regex `\w` isFC used, then observing `x` will cause the search to quit immediately30 despite the fact that `x` is in the `\w` class.IF This mechanism is primarily useful for heuristically enabling certainHE features like Unicode word boundaries in a DFA. Namely, if the inputGD to search is ASCII, then a Unicode word boundary can be implementedGD via an ASCII word boundary with no change in semantics. Thus, a DFAJG can attempt to match a Unicode word boundary but give up as soon as itIF observes a non-ASCII byte. Indeed, if callers set all non-ASCII bytesIF to be quit bytes, then Unicode word boundaries will be permitted when30 building DFAs. Of course, callers should enableIF [`Config::unicode_word_boundary`] if they want this behavior instead.JG (The advantage being that non-ASCII quit bytes will only be added if a-* Unicode word boundary is in the pattern.)EB When enabling this option, callers _must_ be prepared to handle aGD [`MatchError`](crate::MatchError) error during search. When using aFC [`Regex`](crate::dfa::regex::Regex), this corresponds to using the `try_` suite of methods.,) By default, there are no quit bytes set.d # PanicsHE This panics if heuristic Unicode word boundaries are enabled and anyJG non-ASCII byte is removed from the set of quit bytes. Namely, enablingKH Unicode word boundaries requires setting every non-ASCII byte to a quitGD byte. So if the caller attempts to undo any of that, then this willT panic.l-FC This example shows how to cause a search to terminate if it sees aJG `\n` byte. This could be useful if, for example, you wanted to preventA> a user supplied pattern from matching across a line boundary.<-=-EB use regex_automata::{dfa::{Automaton, dense}, Input, MatchError};#-:7 .configure(dense::Config::new().quit(b'\n', true))# .build(r"foo\p{any}+bar")?;)& let haystack = "foo\nbar".as_bytes();HE // Normally this would produce a match, since \p{any} contains '\n'.GD // But since we instructed the automaton to enter a quit state if a=: // '\n' is observed, this produces a match error instead..+ let expected = MatchError::quit(b'\n', 3);EB let got = dfa.try_search_fwd(&Input::new(haystack)).unwrap_err();..-<-$BMM W$$Ƌ=0- Enable specializing start states in the DFA.IF When start states are specialized, an implementor of a search routineKH using a lazy DFA can tell when the search has entered a starting state.GD When start states aren't specialized, then it is impossible to know1. whether the search has entered a start state.C@ Ideally, this option wouldn't need to exist and we could alwaysJG specialize start states. The problem is that start states can be quiteIF active. This in turn means that an efficient search routine is likelyGD to ping-pong between a heavily optimized hot loop that handles mostHE states and to a less optimized specialized handling of start states.DA This causes branches to get heavily mispredicted and overall canHE materially decrease throughput. Therefore, specializing start states-* should only be enabled when it is needed.KH Knowing whether a search is in a start state is typically useful when aIF prefilter is active for the search. A prefilter is typically only runJG when in a start state and a prefilter can greatly accelerate a search.IF Therefore, the possible cost of specializing start states is worth itIF in this case. Otherwise, if you have no prefilter, there is likely no&# reason to specialize start states.B? This is disabled by default, but note that it is automaticallyIF enabled (or disabled) if [`Config::prefilter`] is set. Namely, unlessIF `specialize_start_states` has already been set, [`Config::prefilter`]HE will automatically enable or disable it based on whether a prefilterGD is present or not, respectively. This is done because a prefilter'sFC effectiveness is rooted in being executed whenever the DFA is in aJG start state, and that's only possible to do when they are specialized.A> Note that it is plausibly reasonable to _disable_ this optionFC explicitly while _enabling_ a prefilter. In that case, a prefilterIF will still be run at the beginning of a search, but never again. ThisJG in theory could strike a good balance if you're in a situation where aB? prefilter is likely to produce many false positive candidates.l-HE This example shows how to enable start state specialization and thenۀ?< shows how to check whether a state is a start state or not.<->; use regex_automata::{dfa::{Automaton, dense::DFA}, Input};-?< .configure(DFA::config().specialize_start_states(true)) .build(r"[a-z]+")?;0- let haystack = "123 foobar 4567".as_bytes();>; let sid = dfa.start_state_forward(&Input::new(haystack))?;HE // The ID returned by 'start_state_forward' will always be tagged asЄ@= // a start state when start state specialization is enabled.'$ assert!(dfa.is_special_state(sid));%" assert!(dfa.is_start_state(sid));.-<-KH Compare the above with the default DFA configuration where start statesIF are _not_ specialized. In this case, the start state is not tagged atD؇ all:<->.Ĉ# let dfa = DFA::new(r"[a-z]+")?;%" let haystack = "123 foobar 4567";>.A> // Start states are not special in the default configuration!(% assert!(!dfa.is_special_state(sid));Ԋ&# assert!(!dfa.is_start_state(sid));.-<-͋BMM $?Ќ52 Set a size limit on the total heap used by a DFA.?< This size limit is expressed in bytes and is applied during֍KH determinization of an NFA into a DFA. If the DFA's heap usage, and onlyKH the DFA, exceeds this configured limit, then determinization is stopped and an error is returned.>; This limit does not apply to auxiliary storage used during96 determinization that isn't part of the generated DFA.JG This limit is only applied during determinization. Currently, there isHE no way to post-pone this check to after minimization if minimizationő was enabled.ڑIF The total limit on heap used during determinization is the sum of the(% DFA and determinization size limits.ݒ The default is no limit.l-HE This example shows a DFA that fails to build because of a configuredHE size limit. This particular example also serves as a cautionary tale”HE demonstrating just how big DFAs with large Unicode character classesd can get.<-=-96 use regex_automata::{dfa::{dense, Automaton}, Input};ļ // 6MB isn't enough!ٖ dense::Builder::new()HE .configure(dense::Config::new().dfa_size_limit(Some(6_000_000)))ė .build(r"\w{20}") .unwrap_err(); // ... but 7MB probably is!HE // (Note that DFA sizes aren't necessarily stable between releases.)#-HE .configure(dense::Config::new().dfa_size_limit(Some(7_000_000))) .build(r"\w{20}")?;/, let haystack = "A".repeat(20).into_bytes();C@ assert!(dfa.try_search_fwd(&Input::new(&haystack))?.is_some());.-<›-Λ֛DA While one needs a little more than 6MB to represent `\w{20}`, itDA turns out that you only need a little more than 6KB to represent74 `(?-u:\w{20})`. So only use Unicode if you need it!DA As with [`Config::determinize_size_limit`], the size of a DFA isHE influenced by other factors, such as what start state configurationsžIF to support. For example, if you only need unanchored searches and notJG anchored searches, then configuring the DFA to only support unanchoredߟJG searches can reduce its size. By default, DFAs support both unanchoredԮ and anchored searches.͠<ՠ-=-DA use regex_automata::{dfa::{dense, Automaton, StartKind}, Input}; // 3MB isn't enough!̑.'-ۢ,) .dfa_size_limit(Some(3_000_000)).+ .start_kind(StartKind::Unanchored)L-ͣ.. // ... but 4MB probably is!H.#-'-ӥ,) .dfa_size_limit(Some(4_000_000))..L-Ŧ./.C..-<-tͶBMM Y$¨,ȨG?< Set a size limit on the total heap used by determinization.?.HE determinization of an NFA into a DFA. If the heap used for auxiliaryEB storage during determinization (memory that is not in the DFA butԫGD necessary for building the DFA) exceeds this configured limit, then85 determinization is stopped and an error is returned.ݬ=: This limit does not apply to heap used by the DFA itself.I.(.䲮.Ӯlۮ-=: This example shows a DFA that fails to build because of a=: configured size limit on the amount of heap space used byHE determinization. This particular example complements the example forưJG [`Config::dfa_size_limit`] by demonstrating that not only does UnicodeEB potentially make DFAs themselves big, but it also results in more߱JG auxiliary storage during determinization. (Although, auxiliary storage,) is still not as much as the DFA itself.)߲<-=-KH # if !cfg!(target_pointer_width = "64") { return Ok(()); } // see #10399.ô˴ // 700KB isn't enough!.'-2/ .determinize_size_limit(Some(700_000))L-..! // ... but 800KB probably is!KH // (Note that auxiliary storage sizes aren't necessarily stable between // releases.)Ʒ#-'-2/ .determinize_size_limit(Some(800_000))LѸ-߸./.C..-<-ºʺA> Note that some parts of the configuration on a DFA can have aB? big impact on how big the DFA is, and thus, how much memory is׻HE used. For example, the default setting for [`Config::start_kind`] isEB [`StartKind::Both`]. But if you only need an anchored search, forJG example, then it can be much cheaper to build a DFA that only supportsJG anchored searches. (Running an unanchored search with it would panic.)<-=-K/̲-п+-Ā Anchored, Input,4-԰ // 200KB isn't enough!.'-2/ .determinize_size_limit(Some(200_000)),-L-̏..! // ... but 300KB probably is!K/ٍ/#-'-2/ .determinize_size_limit(Some(300_000)),-L-ܦ./.>; let input = Input::new(&haystack).anchored(Anchored::Yes);30 assert!(dfa.try_search_fwd(&input)?.is_some());.-<-ͶBMM Z$,$?< Returns whether this configuration has enabled simple state acceleration.t& &BM M [$1A> Returns the prefilter attached to this configuration, if any.l& &řǙșəʙ˙G*R{9  &BM M \$"HE Returns whether this configuration has enabled the expensive processċ of minimizing a DFA.d& &BM M ]$):7 Returns the match semantics set in this configuration.t& &BM M ^$%74 Returns the starting state configuration for a DFA.T& &BM M _$1KH Returns whether this configuration has enabled anchored starting states! for every pattern in the DFA.& &BM M `$&GD Returns whether this configuration has enabled byte classes or not.JG This is typically a debugging oriented option, as disabling it confers no speed benefit.& &BM M a$/IF Returns whether this configuration has enabled heuristic Unicode wordIF boundary support. When enabled, it is possible for a search to returnl an error.̆& &BM M b$(GD Returns whether this configuration will instruct the DFA to enter aGD quit state whenever the given byte is seen during a search. When atJG least one byte has this enabled, it is possible for a search to returnl/D& &BM M c$$cddddcdU1?< Returns whether this configuration will instruct the DFA toKH "specialize" start states. When enabled, the DFA will mark start statesFC as "special" so that search routines using the DFA can detect whenFC it's in a start state and do some kind of optimization (like run a| prefilter).ܝ& &BM M e$1DA Returns the DFA size limit of this configuration if one was set.EB The size limit is total number of bytes on the heap that a DFA isHE permitted to use. If the DFA exceeds this limit during construction,:7 then construction is stopped and an error is returned.& &ͶBM M f$9GD Returns the determinization size limit of this configuration if oneEB was set. The size limit is total number of bytes on the heap thatHE determinization is permitted to use. If determinization exceeds thisKH limit during construction, then construction is stopped and an error isl returned.JG This is different from the DFA size limit in that this only applies toKH the auxiliary storage used during determinization. Once determinization&# is complete, this memory is freed.EB The limit on the total heap memory used is the sum of the DFA and  determinization size limits.Ԫ& &ͶBM M g$3L& &BM M h$ hiiiihiUdЇ,첇klmoqrsB? Create a new dense DFA builder with the default configuration.BjjB'$ Build a DFA from the given pattern.JG If there was a problem parsing or compiling the pattern, then an error is returned.Ԃ,&& & &Bj j l$<ߍn(% Build a DFA from the given patterns.ŒʌIF When matches are returned, the pattern ID corresponds to the index of# the pattern in the slice given.ƍT&& & &Bjnn j T$wo+-h# Build a DFA from the given NFA.l-ǒϒGD This example shows how to build a DFA if you already have an NFA inL hand.<-̽-ۓ .܀-̠ HalfMatch, Input,4-ɔє*' let haystack = "foo123bar".as_bytes();EB // This shows how to set non-default options for building an NFA.ҕ let nfa = NFA::compiler().+ .configure(NFA::config().shrink(true))ܧ-ǖ:7 let dfa = dense::Builder::new().build_from_nfa(&nfa)?;/, let expected = Some(HalfMatch::must(0, 6));9...-<֘-t&& & &Bj j$.oppppop֋ ;̰DA Apply the given dense DFA configuration options to this builder.L& &/Bj j $4d74 Set the syntax configuration for this builder usingܲ41 [`syntax::Config`](crate::util::syntax::Config).JG This permits setting things like case insensitivity, Unicode and multit line mode.EB These settings only apply when constructing a DFA directly from adѴ.4& &/Bj j r$4D=: Set the Thompson NFA configuration for this builder using<9 [`nfa::thompson::Config`](crate::nfa::thompson::Config).KH This permits setting things like whether the DFA should match the regexJG in reverse or if additional time should be spent shrinking the size ofdи the NFA.E/d.ʹĹD& &/Bj j $4ć,uu<Btt u߾پDl,xy9HE Parse the given regular expression using a default configuration and! return the corresponding DFA.96 If you want a non-default configuration, then use the>; [`dense::Builder`](Builder) to set your own configuration.l-<-DA use regex_automata::{dfa::{Automaton, dense}, HalfMatch, Input};/, let dfa = dense::DFA::new("foo[0-9]+bar")?;0- let expected = Some(HalfMatch::must(0, 11));JG assert_eq!(expected, dfa.try_search_fwd(&Input::new("foo12345bar"))?);.-<-ԁ& &Bw w x<]IF Parse the given regular expressions using a default configuration and'$ return the corresponding multi-DFA.9/>/l-<-D/;8 let dfa = dense::DFA::new_many(&["[0-9]+", "[a-z]+"])?;/, let expected = Some(HalfMatch::must(1, 3));J/.-<-ԑD& &Bwzz w/ /T ywv l,|}~5.+ Create a new DFA that matches every input.l-<-D/*' let dfa = dense::DFA::always_match()?;/, let expected = Some(HalfMatch::must(0, 0));?< assert_eq!(expected, dfa.try_search_fwd(&Input::new(""))?);B? assert_eq!(expected, dfa.try_search_fwd(&Input::new("foo"))?);.-<-dB{{ |42/ Create a new DFA that never matches any input.l-<-96 use regex_automata::{dfa::{Automaton, dense}, Input};)& let dfa = dense::DFA::never_match()?;;8 assert_eq!(None, dfa.try_search_fwd(&Input::new(""))?);>; assert_eq!(None, dfa.try_search_fwd(&Input::new("foo"))?);.-<-\B{{ }<&  &B{ { ~< \Ȁ 4lookm,Á<,,8&&)!bhA;Ж 8:,KH This is a convenience routine to avoid needing to import the [`Config`]:7 type when customizing the construction of a dense DFA.4B 8B/@= This is a convenience routine to avoid needing to import thẻ%" [`Builder`] type in common cases.<B 8䴊 d<>&׊FC Cheaply return a borrowed version of this dense DFA. Specifically,C@ the DFA returned always uses `&[u32]` for its transition table.4& &)!bhA;Ж &B ֔$ِ"ǎEB Return an owned version of this sparse DFA. Specifically, the DFA=: returned always uses `Vec` for its transition table.ӏۏIF Effectively, this returns a dense DFA whose transition table lives onl the heap.̻D& &B  $%Ó:7 Returns the starting state configuration for this DFA.IF The default is [`StartKind::Both`], which means the DFA supports bothؔIF unanchored and anchored searches. However, this can generally lead toIF bigger DFAs. Therefore, a DFA might be compiled with support for justHE unanchored or anchored searches. In that case, running a search with,) an unsupported configuration will panic.T& &B $/L& & &B Ș$ɘ-GD Returns true only if this DFA has starting states for each pattern.ԙܙKH When a DFA has starting states for each pattern, then a search with theIF DFA can be configured to only look for anchored matches of a specificFC pattern. Specifically, APIs like [`Automaton::try_search_fwd`] canśKH accept a non-None `pattern_id` if and only if this method returns true.30 Otherwise, calling `try_search_fwd` will panic.͜՜DA Note that if the DFA has no patterns, this always returns false.& &B $*KH Returns the equivalence classes that make up the alphabet for this DFA.ϞמEB Unless [`Config::byte_classes`] was disabled, it is possible thatGD multiple distinct bytes are grouped into the same equivalence classFC if it is impossible for them to discriminate between a match and aHE non-match. This has the effect of reducing the overall alphabet sizeHE and in turn potentially substantially reducing the size of the DFA'sҡ transition table.KH The downside of using equivalence classes like this is that every stateĢFC transition will automatically use this map to convert an arbitraryGD byte to its corresponding equivalence class. In practice this has aۣ%" negligible impact on performance.d& & &B ݆$#֤FC Returns the total number of elements in the alphabet for this DFA.IF That is, this returns the total number of transitions that each stateFC in this DFA must have. Typically, a normal byte oriented DFA would¦GD always have an alphabet size of 256, corresponding to the number ofHE unique values in a single byte. However, this implementation has twoۧ2/ peculiarities that impact the alphabet length:FC * Every state has a special "EOI" transition that is only followedEB after the end of some haystack is reached. This EOI transition isEB necessary to account for one byte of look-ahead when implementing things like `\b` and `$`.JG * Bytes are grouped into equivalence classes such that no two bytes inIF the same class can distinguish a match from a non-match. For example,FC in the regex `^[a-z]+$`, the ASCII bytes `a-z` could all be in theB? same equivalence class. This leads to a massive space savings.ʬҬIF Note though that the alphabet length does _not_ necessarily equal theGD total stride space taken up by a single DFA state in the transitionDA table. Namely, for performance reasons, the stride is always theGD smallest power of two that is greater than or equal to the alphabetDA length. For this reason, [`DFA::stride`] or [`DFA::stride2`] areʯGD often more useful. The alphabet length is typically useful only forܖ informational purposes.d& &B ʰ $˰JG Returns the total stride for every state in this DFA, expressed as theձJG exponent of a power of 2. The stride is the amount of space each stateKH takes up in the transition table, expressed as a number of transitions.,) (Unused transitions map to dead states.)GD The stride of a DFA is always equivalent to the smallest power of 2DA that is greater than or equal to the DFA's alphabet length. This´GD definition uses extra space, but permits faster translation betweenKH premultiplied state identifiers and contiguous indices (by using shifts޵,) instead of relying on integer division).JG For example, if the DFA's stride is 16 transitions, then its `stride2` is `4` since `2^4 = 16`.IF The minimum `stride2` value is `1` (corresponding to a stride of `2`)ݷJG while the maximum `stride2` value is `9` (corresponding to a stride ofKH `512`). The maximum is not `8` since the maximum alphabet size is `257`HE when accounting for the special EOI transition. However, an alphabetɹJG length of that size is exceptionally rare since the alphabet is shrunk옺 into equivalence classes.<& &B ɺ $ʺ쉾JG Returns the total stride for every state in this DFA. This correspondsͻGD to the total number of transitions used by each state in this DFA's0IF Please see [`DFA::stride2`] for more information. In particular, thisGD returns the stride as the number of transitions, where as `stride2`ս/, returns it as the exponent of a power of 2.4& &B  $#;41 Returns the memory usage, in bytes, of this DFA.EB The memory usage is computed based on the number of bytes used toؿ represent this DFA.DA This does **not** include the stack size used up by this DFA. To:7 compute that, use `std::mem::size_of::()`.d& &B  $IF Routines for converting a dense DFA to other representations, such as=: sparse DFAs or raw bytes suitable for persistent storage./ /d-/C+( Convert this dense DFA to a sparse DFA.EB If a `StateID` is too small to represent all states in the sparseEB DFA, then this returns an error. In most cases, if a dense DFA isC@ constructable with `StateID` then a sparse DFA will be as well." However, it is not guaranteed.l-<-D/.+ let dense = dense::DFA::new("foo[0-9]+")?;$! let sparse = dense.to_sparse()?;/, let expected = Some(HalfMatch::must(0, 8));JG assert_eq!(expected, sparse.try_search_fwd(&Input::new("foo12345"))?);.-<-,L& &ԍB  $8C@ Serialize this DFA as raw bytes to a `Vec` in little endianJG format. Upon success, the `Vec` and the initial padding length arelϳ/EB The written bytes are guaranteed to be deserialized correctly andDA without errors in a semver compatible release of this crate by aEB `DFA`'s deserialization APIs (assuming all other criteria for the-* deserialization APIs has been satisfied):̱ * [`DFA::from_bytes`]# * [`DFA::from_bytes_unchecked`]HE The padding returned is non-zero if the returned `Vec` starts atJG an address that does not have the same alignment as `u32`. The paddingFC corresponds to the number of leading bytes written to the returnedt `Vec`.l->; This example shows how to serialize and deserialize a DFA:<-IF use regex_automata::{dfa::{Automaton, dense::DFA}, HalfMatch, Input};  // Compile our original DFA..+ let original_dfa = DFA::new("foo[0-9]+")?;GD // N.B. We use native endianness here to make the example work, butIF // using to_bytes_little_endian would work on a little endian target.96 let (buf, _) = original_dfa.to_bytes_native_endian();JG // Even if buf has initial padding, DFA::from_bytes will automatically // ignore it.41 let dfa: DFA<&[u32]> = DFA::from_bytes(&buf)?.0;/0GD assert_eq!(expected, dfa.try_search_fwd(&Input::new("foo12345"))?);.-<-,& &B  $5@= Serialize this DFA as raw bytes to a `Vec` in big endianJԬ0lϳ/Eĭ0D0E0-0̶0#0Hְ0J0F0tβ0l->0<-I0 Դ0.0Gŵ0C@ // using to_bytes_big_endian would work on a big endian target.90J0040/0G0.-<-,& &B  $8C@ Serialize this DFA as raw bytes to a `Vec` in native endianJԬ0lϳ/Eĭ0D0E0-0̵0#0Hְ0J0F0tβ0EB Generally speaking, native endian format should only be used whenEB you know that the target you're compiling the DFA for matches theHE endianness of the target on which you're compiling DFA. For example,IF if serialization and deserialization happen in the same process or onDA the same machine. Otherwise, when serializing a DFA for use in aJG portable environment, you'll almost certainly want to serialize _both_JG a little endian and a big endian version and then load the correct one(% based on the target's configuration.l->0<-I0 Դ0.090J0040/0G0.-<-,& &B  $1,D& &B  4 $+-ɎlHE Serialize this DFA as raw bytes to the given slice, in little endianGD format. Upon success, the total number of bytes written to `dst` islϳ/Eĭ0D0E0-0ɀр0#0JG Note that unlike the various `to_byte_*` routines, this does not writeJG any padding. Callers are responsible for handling alignment correctly.dł # ErrorsւނJG This returns an error if the given destination slice is not big enoughHE to contain the full serialized DFA. If an error occurs, then nothing is written to `dst`.l-EB This example shows how to serialize and deserialize a DFA without dynamic memory allocation.<-I0 Դ0.0GD // Create a 4KB buffer on the stack to store our serialized DFA. WeGD // need to use a special type to force the alignment of our [u8; N]IF // array to be aligned to a 4 byte boundary. Otherwise, deserializingֈ96 // the DFA may fail because of an alignment mismatch.t #[repr(C)] struct Aligned {ˉ _align: [u32; 0], bytes: B,, }DA let mut buf = Aligned { _align: [], bytes: [0u8; 4 * (1<<10)] };ҊGŵ0IF // using write_to_little_endian would work on a little endian target.GD let written = original_dfa.write_to_native_endian(&mut buf.bytes)?;EB let dfa: DFA<&[u32]> = DFA::from_bytes(&buf.bytes[..written])?.0;/0G0.-<-Ў&& & &ݘB  $ڍiEB Serialize this DFA as raw bytes to the given slice, in big endianG0lϳ/Eĭ0D0E0-0̺0ؓ#0J0הJ0d0ǕJ0H00l-E00<-I0 Դ0.0љٙG0G0I090t00̴0Ҝ0,0D0Gŵ0C@ // using write_to_big_endian would work on a big endian target.ϞG0E0/0G0.-<-&& & &ݘB С $ѡƷlӢHE Serialize this DFA as raw bytes to the given slice, in native endianG0lϳ/Eĭ0ФD0E0-0̝0#0E0E0H0̨I0D0J0J0(0J0J0Ԭdܬ0J0ĭH0đ0l-ȮЮE00<ů-ѯI0 Դ0̰.0G0ӱG0I090t0000,0D0G0E0/0G0.-<-ͷ&& & &ݘB  $@#DA Return the total number of bytes required to serialize this DFA.JG This is useful for determining the size of the buffer required to pass)& to one of the serialization routines:ƺ%" * [`DFA::write_to_little_endian`]" * [`DFA::write_to_big_endian`]%" * [`DFA::write_to_native_endian`]ɻGD Passing a buffer smaller than the size returned by this method willJG result in a serialization error. Serialization routines are guaranteed-* to succeed when the buffer is big enough.l-KH This example shows how to dynamically allocate enough room to serializeT a DFA.<-I0.074 let mut buf = vec![0; original_dfa.write_to_len()];JG // This is guaranteed to succeed, because the only serialization errorC@ // that can occur is when the provided buffer is too small. But.+ // write_to_len guarantees a correct size.IF let written = original_dfa.write_to_native_endian(&mut buf).unwrap();<9 // But this is not guaranteed to succeed! In particular,KH // deserialization requires proper alignment for &[u32], but our bufferHE // was allocated as a &[u8] whose required alignment is smaller thanKH // &[u32]. However, it's likely to work in practice because of how mostKH // allocators work. So if you write code like this, make sure to eitherJG // handle the error correctly and/or run it under Miri since Miri willKH // likely provoke the error by returning Vec buffers with alignmentĭ // less than &[u32].C@ let dfa: DFA<&[u32]> = match DFA::from_bytes(&buf[..written]) {FC // As mentioned above, it is legal for an error to be returnedIF // here. It is quite difficult to get a Vec with a guaranteed,) // alignment equivalent to Vec.  Err(_) => return Ok(()), Ok((dfa, _)) => dfa,4-/0G0.-<-@= Note that this example isn't actually guaranteed to work! InFC particular, if `buf` is not aligned to a 4-byte boundary, then theHE `DFA::from_bytes` call will fail. If you need this to work, then youIF either need to deal with adding some initial padding yourself, or use<9 one of the `to_bytes` methods, which will do it for you.d& &B ͓$)!bhA;Ж ʑf=: Safely deserialize a DFA with a specific state identifierHE representation. Upon success, this returns both the deserialized DFAKH and the number of bytes read from the given slice. Namely, the contents-* of the slice beyond the DFA are not read.KH Deserializing a DFA using this routine will never allocate heap memory.IF For safety purposes, the DFA's transition table will be verified suchJG that every transition points to a valid state. If this verification isKH too costly, then a [`DFA::from_bytes_unchecked`] API is provided, which)& will always execute in constant time.FC The bytes given must be generated by one of the serialization APIsEB of a `DFA` using a semver compatible release of this crate. Thosed include:%" * [`DFA::to_bytes_little_endian`]" * [`DFA::to_bytes_big_endian`]%" * [`DFA::to_bytes_native_endian`]%0"0%0IF The `to_bytes` methods allocate and return a `Vec` for you, alongDA with handling alignment correctly. The `write_to` methods do notHE allocate and write to an existing slice (which may be on the stack).IF Since deserialization always uses the native endianness of the targetJG platform, the serialization API you use should match the endianness ofGD the target platform. (It's often a good idea to generate serializedIF DFAs for both forms of endianness and then load the correct one based on endianness.)d0GD Generally speaking, it's easier to state the conditions in which an?< error is _not_ returned. All of the following must be true:GD * The bytes given must be produced by one of the serialization APIs&# on this DFA, as mentioned above.JG * The endianness of the target platform matches the endianness used to" serialized the provided DFA.<9 * The slice given must have the same alignment as `u32`.EB If any of the above are not true, then an error will be returned.dں.0- This routine will never panic for any input.l-JG This example shows how to serialize a DFA to raw bytes, deserialize it" and then use it for searching.<-I0)& let initial = DFA::new("foo[0-9]+")?;63 let (bytes, _) = initial.to_bytes_native_endian();63 let dfa: DFA<&[u32]> = DFA::from_bytes(&bytes)?.0;/0G0.-<-1. # Example: dealing with alignment and paddingHE In the above example, we used the `to_bytes_native_endian` method toJG serialize a DFA, but we ignored part of its return value correspondingGD to padding added to the beginning of the serialized DFA. This is OKHE because deserialization will skip this initial padding. What mattersJG is that the address immediately following the padding has an alignmentC@ that matches `u32`. That is, the following is an equivalent but/, alternative way to write the above example:<-I0)1IF // Serialization returns the number of leading padding bytes added to // the returned Vec.85 let (bytes, pad) = initial.to_bytes_native_endian();=: let dfa: DFA<&[u32]> = DFA::from_bytes(&bytes[pad..])?.0;/0G0.-<-B? This padding is necessary because Rust's standard library doesEB not expose any safe and robust way of creating a `Vec` with aGD guaranteed alignment other than 1. Now, in practice, the underlyingGD allocator is likely to provide a `Vec` that meets our alignmentIF requirements, which means `pad` is zero in practice most of the time.HE The purpose of exposing the padding like this is flexibility for theFC caller. For example, if one wants to embed a serialized DFA into aJG compiled program, then it's important to guarantee that it starts at aHE `u32`-aligned address. The simplest way to do this is to discard theKH padding bytes and set it up so that the serialized DFA itself begins atHE a properly aligned address. We can show this in two parts. The first*' part is serializing the DFA to a file:l ```no_run(% use regex_automata::dfa::dense::DFA;%" let dfa = DFA::new("foo[0-9]+")?;1. let (bytes, pad) = dfa.to_bytes_big_endian();C@ // Write the contents of the DFA *without* the initial padding.85 std::fs::write("foo.bigendian.dfa", &bytes[pad..])?;41 // Do it again, but this time for little endian.41 let (bytes, pad) = dfa.to_bytes_little_endian();;8 std::fs::write("foo.littleendian.dfa", &bytes[pad..])?;.-<-JG And now the second part is embedding the DFA into the compiled programDA and deserializing it at runtime on first use. We use conditional1. compilation to choose the correct endianness.l1̏-%" dfa::{Automaton, dense::DFA},*' util::{lazy::Lazy, wire::AlignAs},̆/4-<9 // This crate provides its own "lazy" type, kind of likeFC // lazy_static! or once_cell::sync::Lazy. But it works in no-allocÁC@ // no-std environments and let's us write this using completely // safe code.96 static RE: Lazy> = Lazy::new(|| {߂&# # const _: &str = stringify! {@= // This assignment is made possible (implicitly) via theσ@= // CoerceUnsized trait. This is what guarantees that our?< // bytes are stored in memory on a 4 byte boundary. You؄=: // *must* do this or something equivalent for correctܚ // deserialization.85 static ALIGNED: &AlignAs<[u8], u32> = &AlignAs { _align: [],)& #[cfg(target_endian = "big")]85 bytes: *include_bytes!("foo.bigendian.dfa"),,) #[cfg(target_endian = "little")];8 bytes: *include_bytes!("foo.littleendian.dfa"),T };d # };:7 # static ALIGNED: &AlignAs<[u8], u32> = &AlignAs {Έ # _align: [], # bytes: [],d163 let (dfa, _) = DFA::from_bytes(&ALIGNED.bytes)݉63 .expect("serialized DFA should be valid");\ dfa< });30 let expected = Ok(Some(HalfMatch::must(0, 8)));EB assert_eq!(expected, RE.try_search_fwd(&Input::new("foo12345")));<-ʋҋC@ An alternative to [`util::lazy::Lazy`](crate::util::lazy::Lazy)?< is [`lazy_static`](https://crates.io/crates/lazy_static) orތDA [`once_cell`](https://crates.io/crates/once_cell), which provideDA stronger guarantees (like the initialization function only beingA> executed once). And `once_cell` in particular provides a moreJG expressive API. But a `Lazy` value from this crate is likely just fineԅ in most circumstances.DA Note that regardless of which initialization method you use, youFC will still need to use the [`AlignAs`](crate::util::wire::AlignAs)FC trick above to force correct alignment, but this is safe to do and:7 `from_bytes` will return an error if you get it wrong.Tё ÜĜŜƜǜȜ_w܂1B ,w͟HE Deserialize a DFA with a specific state identifier representation inEB constant time by omitting the verification of the validity of the30 transition table and other data inside the DFA.KH This is just like [`DFA::from_bytes`], except it can potentially returnKH a DFA that exhibits undefined behavior if its transition table containsĢ invalid state identifiers.C@ This routine is useful if you need to deserialize a DFA cheaplyB? and cannot afford the transition table validation performed by `from_bytes`.l-<-¤I0)1ƥ61GD // SAFETY: This is guaranteed to be safe since the bytes given comeͦ85 // directly from a compatible serialization routine.KH let dfa: DFA<&[u32]> = unsafe { DFA::from_bytes_unchecked(&bytes)?.0 };ڧ/0G0.-<-11 ,ͩiD&& &܂1 &ݘB 0 04 $/1tϾ)& Other routines that work for all `T`./  =41 Set or unset the prefilter attached to this DFA.ǿϿ@= This is useful when one has deserialized a DFA from `&[u8]`.DA Deserialization does not currently include prefilters, so if youEB want prefilter acceleration, you'll need to rebuild it and attachd it here.l& &B  $l,~|& &łB  $D,-t& &B  $$$toG|& & B  $@\& &B  $id1id2@,& &B ##<$ RFHr\& &B 111< $- :<4|& &B  $!D& &B  $|&& & &  W root  length      ׯ0WCCC=C=ÜĜŜƜǜȜ_wB  $ #T& &B  $*& &U<& &11B  $^<"& &B  $f& &łU/ /dHJ(<& & &B ل$4,\& & &B  $(<& & &B  $$,& & &B  $/4& &,,=d4;_&B $'L& &B $B,& & &B  $-K<& &B $-)\& &B $E,\& &1B  $',<& &B  $3D& &B $-99,\& &B  $,ʀ14؀& &))3OEH8-&B ߀ʡ$1& &B $-Uڊ1݊& &B $-"4& &! >By &B $,& & &B $+/ /dg8&&& & &ggg cgcثq1ȡ&B  $Š1/ Ѡ/dԠ/' 'B ʡ $ˡ-֢,l٢' 'B  $-,l' 'B  $-H-t' 'B  $- -t' 'B  $-§-tŧ' 'B ԧ $է-ި<T' 'B  $k' '  $HR5t' 'B  $ ܱ\߱' 'B  $ݲ)L' 'B  $- El' 'B  $- К ܜL' 'B  $̛<' 'B  $*ę' 'B  $c\'' ' 'B ˼ $̼B' 'łB  $\+\' ' 'B  $--l' 'řǙșəʙ˙G*R{9  'B  $#& gOك11ÜĜŜƜǜȜ_w2 ,, =< B < C ' 'B  $ $ $ 1 މ <| ' ' B  $ ə   .$ ' 'B  $ 1 1 B, ' 'B 1̣ 1ѣ 1< $ -  ǣ :<֧ "D٧ ' 'B  $  (/ /d EG iD '' ' 'ݘB ت 0 Ī 04Ǫ $٪  -/ d ' 'B ɳ $ʳ @D '' ' 'ÜĜŜƜǜȜ_wB Ƕ  $ȶ ζ .4 ' '& gO 'B $ ; D ' 'B $ ), ' ',-- h?ٷ'B  $ -X, $4 ' ',,=d4;_'B $ (D ' 'B ۀ$ -/ .\ ' 'B  $ 0 /, l ' 'B  $ - /, l ' 'B  $ - , ' ' 'B $  ' 'B $ ̝ 4 ' 'B $ d ' 'B τ$ 'D ' 'B $ -A6 d ' 'B $ (, / d !# )L ' ' 'B  $ ̟ , 쁞 Ѣ $Ԣ ' 'ͶÜĜŜƜǜȜ_wB  $ /,  \ ¬ & ʞÑ'[ك1Ǭ  ~ 1ÜĜŜƜǜȜ_w2 , ,߾ U  U 2 #/ /d -/ iD '' ' 'ݘB 0 04 $  -/, ׅ ,  d ' 'B $ @D '' ' 'ٖ2B  $   U 22 )4 ' '& ʞÑ'[ 'B $ 6 D ' 'B $ o, ' 'łB $ ?[J[ $$ ' '))3OEH8-'B $ 앉 , ' ' 'B Š $ d ' 'B Ɉ$ #,ދ ؋ / 2d  FL ' 'łB  $ Dȍ ,܍ - ڔ )Lݔ ' ' 'B  $ И $t 3OEH8-   2D  (QQ L֙ $ۙ  8$ ' 'B  $ Ϭ ++ ~U$FQCك1Ԭ   1ÜĜŜƜǜȜ_w2 , Ի ,  5, B  \  ' '1ÜĜŜƜǜȜ_wB  ^<  \  U d '' ' '12B  $ ^< $/ /d <> iD '' ' 'ݘB 0 04 $  -/ d ' 'B $ @D '' ' 'ٖ2B  $  C, 4 '' ' '1B  $  *4 ' '++ ~U$FQC 'B $ 7 D ' 'B $ ?t '' ' 'B $ a{k{ IT ' 'B $ }К} 2\ ' 'B $  > ' ' 'B $ { 4 ' ' 'B  $  ' 'B ċ $ŋ %\ ' ' 'B  $ d ' 'B ʉ$ T  ), D ' 'B  . aT ' 'ÜĜŜƜǜȜ_wB  , ʠ pDؠ '' ' 'ݘB   4  $ 02š \ȡ B  ԡ *d ' 'B $¦ "L  Ԝ =d4;_ 8:5  ɧ / << 0{?iΪΪѪoҪoS #ұ 72 d QQ 79L $ ʫ '$ͫ ' 'řǙșəʙ˙G*R{9 B ҫ  Md  ܀ ,  h?ٷ \  t < Ǯ \Ǯ     ;\ ' ',sZH.^]'B  $ H ' 'L'B  $ Ǻ "պ ' 'B غ  $ٺ <, 왼 T '' ' 'B ɼ  $ʼ <м !g  8 ''' ' 'ggg cgcثq1ȡ'B  $  -QQ L $  7$ ' 'řǙșəʙ˙G*R{9 2B  $ T 2U 3QQ L $  G$ ' '³ B  $ | ,  , KH Returns true if and only if this error corresponds to an error with DFA A> construction that occurred because of exceeding a size limit.  IF While this can occur when size limits like [`Config::dfa_size_limit`] KH or [`Config::determinize_size_limit`] are exceeded, this can also occur GD when the number of states or patterns exceeds a hard-coded maximum. JG (Where these maximums are derived based on the values representable by # [`StateID`] and [`PatternID`].)  FC This predicate is useful in contexts where you want to distinguish DA between errors related to something provided by an end user (for GD example, an invalid regex pattern) and errors related to configured IF heuristics. For example, building a DFA might be an optimization that FC you want to skip if construction fails because of an exceeded size HE limit, but where you want to bubble up an error if it fails for some  other reason. l - < - =- K/ 9.  # let err = dense::Builder::new() '- 2/ .determinize_size_limit(Some(100_000))L - . . ;8 // This error occurs because a size limit was exceeded. &# // But things are otherwise valid. *' assert!(err.is_size_limit_exceeded());  #2  .build(r"\bxyz\b") . 85 // This error occurs because a Unicode word boundary :7 // was used without enabling heuristic support for it. (% // So... not related to size limits. +( assert!(!err.is_size_limit_exceeded());  #2  .build(r"(xyz") . 85 // This error occurs because the pattern is invalid. (2 +2  .-< - ' 'B  $ |Ą , 즄  !$ ' ' 'B  $ Å :х B\Յ Cφ %B щ -|߉ B NJ 3Պ B ȋ 8֋ B Ҍ A B , I B , ߏ %| ,Ə ( 3 >4 ' 'řǙșəʙ˙G*R{9  'B  $ &,ϑ ɑ g D ''' ' 'ggg cgcثq1ȡ'B  $  ly A DFA-backed `Regex`. This module provides [`Regex`], which is defined generically over the [`Automaton`] trait. A `Regex` implements convenience routines you might have come to expect, such as finding the start/end of a match and iterating over all non-overlapping matches. This `Regex` type is limited in its capabilities to what a DFA can provide. Therefore, APIs involving capturing groups, for example, are not provided. Internally, a `Regex` is composed of two DFAs. One is a "forward" DFA that finds the end offset of a match, where as the other is a "reverse" DFA that find the start offset of a match. See the [parent module](crate::dfa) for examples. ,yy,yCTHL,$,$D%,%T%54LL%|̠!,&, 9l,̋(&;9PN      ,   +   , 8 & 8 $  *   +   8   8<  7< ,   +   , 8  8 84 8,  8  8, ' 8D    83< & 8 $ 84< & 8 $ +   8   8    8<  7< ,   +   , 8  8 84 8,  8    83<& 8 $ 84<& 8 $ % T98,98+86 =69JG Parse the given regular expression using the default configuration and:# return the corresponding regex.::HE If you want a non-default configuration, then use the [`Builder`] to; set your own configuration.;l;-;<;-;30 use regex_automata::{Match, dfa::regex::Regex};<<)& let re = Regex::new("foo[0-9]+bar")?;|<Ќ-<$! Some(Match::must(0, 3..14)),=&# re.find(b"zzzfoo12345barzzz"),4=č-=.-<=-=( ( B > <>EZ>GD Like `new`, but parses multiple patterns into a single "regex set."?85 This similarly uses the default regex configuration.?l?-@<@-@33@@52 let re = Regex::new_many(&["[a-z]+", "[0-9]+"])?;AA85 let mut it = re.find_iter(b"abc 1 foo 4567 0 quux");A63 assert_eq!(Some(Match::must(0, 0..3)), it.next());B63 assert_eq!(Some(Match::must(1, 4..5)), it.next());B63 assert_eq!(Some(Match::must(0, 6..9)), it.next());C85 assert_eq!(Some(Match::must(1, 10..14)), it.next());C85 assert_eq!(Some(Match::must(1, 15..16)), it.next());D85 assert_eq!(Some(Match::must(0, 17..21)), it.next());D  assert_eq!(None, it.next());D.- except using sparse DFAs, and return the corresponding regex.HHH3Hē3IlI-I; dfa::regex::Regex, nfa::thompson, util::syntax, Match,4Z-ZZ let re = Regex::builder()Z2/ .syntax(syntax::Config::new().utf8(false))Z63 .thompson(thompson::Config::new().utf8(false))[&# .build(r"foo(?-u:[^b])ar.*")?;[41 let haystack = b"\xFEfoo\xFFarzz\xE2\x98\xFF\n";\.+ let expected = Some(Match::must(0, 1..9));\  let got = re.find(haystack);\.]].-<]-<]B ^^DA Standard search routines for finding and iterating over matches. ^L^!#j@^FC Returns true if and only if this regex matches the given haystack.__IF This routine may short circuit if it knows that scanning future input`KH will never lead to a different result. In particular, if the underlying`KH DFA enters a match state or a dead state, then this routine will returnaIF `true` or `false`, respectively, without inspecting any future input.bdbں.bbHE This routine panics if the search could not complete. This can occurb! in a number of circumstances:ccFC * The configuration of the DFA may permit it to "quit" the search.cEB For example, setting quit bytes or enabling heuristic support fordJG Unicode word boundaries. The default configuration does not enable anye1. option that could result in the DFA quitting.eC@ * When the provided `Input` configuration is not supported. Forf52 example, by providing an unsupported anchor mode.ffGD When a search panics, callers cannot know whether a match exists orDg not.ggKH Use [`Regex::try_search`] if you want to handle these error conditions.glg-h = re.find_iter(text).collect();ـ assert_eq!(matches, vec![ Match::must(0, 0..4), Match::must(0, 5..10), Match::must(0, 11..17),< ]);.-<-LĂ  ,gb+m2B2΂ȿ3 ւٿ3|ق 2 $@B҂=?ܓJG Lower level fallible search routines that permit controlling where theބ41 search starts and ends in a particular sequence.3 Ų3LgG3$3:7 This is like [`Regex::find`] but with two differences:A> 1. It is not generic over `Into` and instead accepts aIF `&Input`. This permits reusing the same `Input` for multiple searchesFC without needing to create a new one. This _may_ help with latency.؈DA 2. It returns an error if the search could not complete where as [`Regex::find`] will panic.ʼnd͉0މHE This routine errors if the search could not complete. This can occur# in the following circumstances:ۊFض3E3J3nj1˸3C3ō5ӹ3GD When a search returns an error, callers cannot know whether a matchӎ exists or not.T((( ( ($$$$$$*0( B  $;0\((( ( ($$$$$$*0(B  $eܪШJG Non-search APIs for querying information about the regex and setting at prefilter.3 Ų3L̩?< Return the underlying DFA responsible for forward matching.HE This is useful for accessing the underlying DFA and converting it toHE some other format or size. See the [`Builder::build_from_dfas`] docs1. for an example of where this might be useful.<( ( (B  $î?< Return the underlying DFA responsible for reverse matching.H3H313<ʮ( ( (B Ү $Ӯ"?< Returns the total number of patterns matched by this regex.Ưlί-<-=-*3?< let re = Regex::new_many(&[r"[a-z]+", r"[0-9]+", r"\w+"])?;$! assert_eq!(3, re.pattern_len());ڱ.-<-\( (B  $>32 LQQ:<L$#$( ( B  32$< d>; Create a new regex builder with the default configuration.B ?)& Build a regex from the given pattern.J//,+6,(( ( ( B  $<s;8 Build a regex from the given pattern using sparse DFAs.J//,+6d(( ( ( B  $<k*' Build a regex from the given patterns.,+6T(( ( ( B / /T $wW!#1. Build a sparse regex from the given patterns.,+6(( ( ( B / /T $w!#n>; Build a regex from its component forward and reverse DFAs.A> This is useful when deserializing a regex from some arbitraryFC memory region. This is also useful for building regexes from other types of DFAs.IF If you're building the DFAs from scratch instead of building new DFAsJG from other DFAs, then you'll need to make sure that the reverse DFA isA> configured correctly to match the intended semantics. Namely: * It should be anchored.1. * It should use [`MatchKind::All`] semantics.! * It should match in reverse.@= * Otherwise, its configuration should match the forward DFA.JG If these conditions aren't satisfied, then the behavior of searches is unspecified.GD Note that when using this constructor, no configuration is applied.DA Since this routine provides the DFAs to the builder, there is no52 opportunity to apply other configuration options.l-EB This example is a bit a contrived. The usual use of these methodsKH would involve serializing `initial_re` somewhere and then deserializingDA it later to build a regex. But in this case, we do everything in\ memory.<-*3.+ let initial_re = Regex::new("foo[0-9]+")?;52 assert_eq!(true, initial_re.is_match(b"foo123"));B? let (fwd, rev) = (initial_re.forward(), initial_re.reverse());85 let re = Regex::builder().build_from_dfas(fwd, rev);-* assert_eq!(true, re.is_match(b"foo123"));.-<-KH This example shows how to build a `Regex` that uses sparse DFAs insteadEB of dense DFAs without using one of the convenience `build_sparse`l routines:<-*3.ԁ4540- let fwd = initial_re.forward().to_sparse()?;0- let rev = initial_re.reverse().to_sparse()?;84-4.-<-|( (B 3 Ų3L $34/1d7/4/J/t/,+64( (Nj4B  $4g=/</FC This permits setting things like whether additional time should be(% spent shrinking the size of the NFA.,+6D( (4B  $4À>FC Set the dense DFA compilation configuration for this builder using [`dense::Config`].GD This permits setting things like whether the underlying DFAs should be minimized.,쌃,( (4B  $4ƃĩȄ<˄B ty  Types and routines specific to sparse DFAs. This module is the home of [`sparse::DFA`](DFA). Unlike the [`dense`] module, this module does not contain a builder or configuration specific for sparse DFAs. Instead, the intended way to build a sparse DFA is either by using a default configuration with its constructor [`sparse::DFA::new`](DFA::new), or by first configuring the construction of a dense DFA with [`dense::Builder`] and then calling [`dense::DFA::to_sparse`]. For example, this configures a sparse DFA to do an overlapping search: ``` use regex_automata::{ dfa::{Automaton, OverlappingState, dense}, HalfMatch, Input, MatchKind, }; let dense_re = dense::Builder::new() .configure(dense::Config::new().match_kind(MatchKind::All)) .build(r"Samwise|Sam")?; let sparse_re = dense_re.to_sparse()?; // Setup our haystack and initial start state. let input = Input::new("Samwise"); let mut state = OverlappingState::start(); // First, 'Sam' will match. sparse_re.try_search_overlapping_fwd(&input, &mut state)?; assert_eq!(Some(HalfMatch::must(0, 3)), state.get_match()); // And now 'Samwise' will match. sparse_re.try_search_overlapping_fwd(&input, &mut state)?; assert_eq!(Some(HalfMatch::must(0, 7)), state.get_match()); # Ok::<(), Box>(()) ``` ,yy4y$ 8 c<   = C$ T N  HL T , < L $  ¬\ < L < , L!L!<"D%$,#d#$$4t$T ,  $ d , ,  ,D ,., ,  ,, ,.,,   %d     { Y\IL9<),  , $,,<%,%%+=%H/&(% return the corresponding sparse DFA.&&9/'C@ [`dense::Builder`] to set your own configuration, and then call'52 [`dense::DFA::to_sparse`] to create a sparse DFA.(l(-(<(-(EB use regex_automata::{dfa::{Automaton, sparse}, HalfMatch, Input};()0- let dfa = sparse::DFA::new("foo[0-9]+bar")?;))0/)J/*.-<*-+ԃ++( (ԍB + <+<, 2a,I/-'/--9/.C4.5ɦ4/l/-//<.-<<-\<ԍB >qT>( (ԍB >/ >/d>L $vNJ wɋ ~/ ~\~.0,~KH Cheaply return a borrowed version of this sparse DFA. Specifically, the96 DFA returned always uses `&[u8]` for its transitions.4 )5\Yti B $2E/74 returned always uses `Vec` for its transitions.HE Effectively, this returns a sparse DFA whose transitions live on theL heap.̇D( (B  $%͆:/I/I/I/H/ˉ,/T( (B  $Ɏ-ĊG/K/I/F/KH accept a [`Anchored::Pattern`] if and only if this method returns true.э)& Otherwise, an error will be returned.=: Note that if the DFA is empty, this always returns false.Ў( (B  $*K/GD Unless [`dense::Config::byte_classes`] was disabled, it is possibleΐFC that multiple distinct bytes are grouped into the same equivalenceJG class if it is impossible for them to discriminate between a match andJG a non-match. This has the effect of reducing the overall alphabet sizeH/0K0F0G0%0d( ( (B ˕ $̕#40ɖEڠ00Dء0;8 compute that, use `std::mem::size_of::()`.dǘ( (B Ԙ $՘ܡNK Routines for converting a sparse DFA to other representations, such as raw*' bytes suitable for persistent storage./ ̸4\-//ÚC0\ format.Eĭ0D0E0-0̺0؝#0HE Note that unlike a [`dense::DFA`]'s serialization methods, this does՞DA not add any initial padding to the returned bytes. Padding isn'tGD required for sparse DFAs since they have no alignment requirements.l->0Ϡ<נ-JG use regex_automata::{dfa::{Automaton, sparse::DFA}, HalfMatch, Input}; Դ0ߡ.0Gŵ0I041 let buf = original_dfa.to_bytes_native_endian();J00Ҥ30 let dfa: DFA<&[u8]> = DFA::from_bytes(&buf)?.0;/0ƥG0.-<Ŧ-,צѦ( (B  $,ԧ@ݺ0\4Eĭ0D0ĩE0-0Ȫ0#0H4D4G4l->0ݭ<-J4Ȯ Դ0.0Gŵ0C044J0ı0ڱ34/0βG0.-<ͳ-,߳ٳ( (B  $/ٴC0\4Eĭ0D0̶E0-0ȷз0#0H4D4G4E0ҺE0H0I0D0J0ϽJ0(0˾lӾ->0<-ĿJ4 Դ0.044J0034/0G0.-<-,( (B  $(,D( (B 0 04 $#%lH0G0lϳ/Eĭ0D0E0-00#0d0J0H00l-E00<-J4 Դ0.0DA // Create a 4KB buffer on the stack to store our serialized DFA.%" let mut buf = [0u8; 4 * (1<<10)];Gŵ0I0A> let written = original_dfa.write_to_native_endian(&mut buf)?;>; let dfa: DFA<&[u8]> = DFA::from_bytes(&buf[..written])?.0;/0G0.-<-(( ( (ݘB  $ -iE0G0lϳ/Eĭ0D0E0-0̡0#0d0J0H0Ĥ0l-E00<-J4 Դ0.0D4%4Gŵ0C0A4>4/0G0.-<-(( ( (ݘB  $-lH0G0lϳ/Eĭ0D0E0-00#0E0E0H0I0D0J0J0(0d0J0H0Ĩ0l-E00<-J4 Դ0.0D4%4A4>4/0G0.-<-(( ( (ݘB  $h.eD(( ( (ݘB 0 04 $<"$#D0J0)0ހ%0"0%0G0$! result in a serialization error.ނl-K0Ѓ a sparse DFA.<-J4Ʉф Դ0.070A4>4/0G0.-<-dĈ( (B ш $҈ڋ)5\Yti1ߋҸeDA Safely deserialize a sparse DFA with a specific state identifierH΃1K1ݍ-1K1IF For safety purposes, the DFA's transitions will be verified such thatIF every transition points to a valid state. If this verification is tooGD costly, then a [`DFA::from_bytes_unchecked`] API is provided, whichϐ)1F҈1БE1d1%1ݒ"Ɗ1%1%0ؓ"0%0GD The `to_bytes` methods allocate and return a `Vec` for you. TheEB `write_to` methods do not allocate and write to an existing sliceǕFC (which may be on the stack). Since deserialization always uses theKH native endianness of the target platform, the serialization API you useJG should match the endianness of the target platform. (It's often a goodJG idea to generate serialized DFAs for both forms of endianness and then.+ load the correct one based on endianness.)d0̘ԘG1?1G1&1J1"1ٛE1KH Note that unlike deserializing a [`dense::DFA`], deserializing a sparseFC DFA has no alignment requirements. That is, an alignment of `1` isTΝ valid.ݝdں.01l-͞՞JĖ1"1˟<ӟ-ߟJ4)11. let bytes = initial.to_bytes_native_endian();52 let dfa: DFA<&[u8]> = DFA::from_bytes(&bytes)?.0;ԡܡ/0G0ܢ.-<-/, # Example: loading a DFA from static memoryףߣDA One use case this library supports is the ability to serialize aGD DFA to disk and then use `include_bytes!` to store it in a compiledEB Rust program. Those bytes can then be cheaply deserialized into aGD `DFA` structure at runtime and used for searching without having to30 re-compile the DFA (which can be quite costly).¦ʦKH We can show this in two parts. The first part is serializing the DFA to\ a file:l1ħ)& use regex_automata::dfa::sparse::DFA;%˪1C@ // Write a big endian serialized version of this DFA to a file.*' let bytes = dfa.to_bytes_big_endian();1. std::fs::write("foo.bigendian.dfa", &bytes)?;٩41-* let bytes = dfa.to_bytes_little_endian();̪41 std::fs::write("foo.littleendian.dfa", &bytes)?;.-<-ī̫JЮ1D1JG compilation to choose the correct endianness. We do not need to employKH any special tricks to ensure a proper alignment, since a sparse DFA has no alignment requirements.l1-ޮ&# dfa::{Automaton, sparse::DFA},̉ util::lazy::Lazy,̧/4ů-Яد<1F1C1̳1±85 static RE: Lazy> = Lazy::new(|| {&1%" #[cfg(target_endian = "big")]ԲB? static BYTES: &[u8] = include_bytes!("foo.bigendian.dfa");(% #[cfg(target_endian = "little")]ȳEB static BYTES: &[u8] = include_bytes!("foo.littleendian.dfa");d1$! # static BYTES: &[u8] = b"";̴Դ-* let (dfa, _) = DFA::from_bytes(BYTES)61\Ҽ1<ѵ1ݵ31EĽ1<-! Alternatively, consider using96 [`lazy_static`](https://crates.io/crates/lazy_static)4߷ or63 [`once_cell`](https://crates.io/crates/once_cell),(% which will guarantee safety for you.Tٸ1ÜĜŜƜǜȜ_w4B ,vH1E1 sparse transitions.K1FC a DFA that exhibits undefined behavior if its transitions contains1GD This routine is useful if you need to deserialize a DFA cheaply andFC cannot afford the transition validation performed by `from_bytes`.d # SafetyB? This routine is not safe because it permits callers to provideJG arbitrary transitions with possibly incorrect state identifiers. WhileEB the various serialization routines will never return an incorrectHE DFA, there is no guarantee that the bytes provided here are correct.EB While `from_bytes_unchecked` will still do several forms of basicKH validation, this routine does not check that the transitions themselvesHE are correct. Given an incorrect transition table, it is possible forJG the search routines to access out-of-bounds memory because of explicit bounds check elision.l-<-J4)115G181JG let dfa: DFA<&[u8]> = unsafe { DFA::from_bytes_unchecked(&bytes)?.0 };/0G0.-<-15 ,t)1/  =41@1D1E1d1l( (B  $G*/ ̸4\g8((( ( (ggg cgcثq1ȡ(B  $^0/ ̸4\/( (B  $-h,l( (B  $-i,l( (B  $-Uj-t( (B  $-k-t( (B  $-k-t( (B  $-l<T( (B  $mml( (  $oo5t( (B  $p\( (B  $)L( (B  $-qEl( (B  $-sК!sL( (B  $̼<( (B  $*ď( (B  $c\(( ( (B  $wB( (łB  $ly+\( ( (B  $-cz-l( (řǙșəʙ˙G*R{9  (B  $5 8+Տ1~Ě1ÜĜŜƜǜȜ_wʽ5 ,#/ ̸4\35iD(( ( (ݘB 0 04 $#%d( (B  $²BDŲ(( ( (ÜĜŜƜǜȜ_wB β $ϲݎղ,2)4( (5 8+Տ (B  $6̸D( ( B  $),( (,-2BB +Bi,Cw6(B  $-ytL(( ( (ÜĜŜƜǜȜ_w,-2BB +Bi,Cw6(B  $ݎ-$4( (,!-VI砆x(B  $̠4( ( (B $̪4( (B  $d( (B  $ݚ#,Ś쿚/ \$&4L( (A-2BB +B]W(B  $-ܣ%Tߣ( ( (B $θ,참 v( (Ͷ B / /d  !#t(( ( ( B / /d %0!#& {9Tk 1}1ÜĜŜƜǜȜ_w5 ,,222"/ ̸4\02iD(( ( (ݘB 0 04 $#%,ڋ,d( (B  $lD((( ( ( (ٖ2B  $ݎseen$(4( (& {9Tk  (B  $5̅D( ( B  $o,( (łB  $S^'$( ())׮ǦT=(B  $( (B  $ĵ,( ( (B ɥ$d( (B  $",ѓ˓/ 5\FL( (łB  $-(t2 2׮ǦT=135 D ;2 ž\ŞQQ79L$8$( (B  $ŧ02 Χgʧ4((( ( (ggg cgcثq1ȡ(B  $ L2 Ū2ΪVI砆x135,Ϊ L62 5\QQ+-L$ԫ'$׫( (řǙșəʙ˙G*R{9 B ܫ $+2 g4((( ( (ggg cgcثq1ȡ(B  $ʹϹǼ$$ʼ) )B ϼ $м,ּD) )B  $) )B  $%,) )B  $ &<) )B  $ 5T) )B  $К\\) )B  $!\) )1B  $iD)) ) )ݘB 0 04 $/1d) )B  $!g8))) ) )ggg cgcثq1ȡ)B  $ ,D]W\tDl4d $6\1,6,2\) )B  $  $$,g8))) ) )ggg cgcثq1ȡ)B  $ Lς̵߂B !4) )B  $Ń-˃(D)) ) )B  $-z,yy,z=C<4$$4t$,,,D,,*\d5$uwd<ԣlLL!bD!)) ) )ͶB "EByك18={=1ÜĜŜƜǜȜ_wΎ6B ,=H#3 II=?I:IIDI) )B J $JK*4K) )! >By )B K $KLDL) ) )B L$LQ#dQ) )B R $RW( * A DFA can search for multiple patterns simultaneously. This DA means extra information is returned when a match occurs. Namely, C@ a match is not just an offset, but an offset plus a pattern ID. KH [`Automaton::pattern_len`] returns the number of patterns compiled into JG the DFA, [`Automaton::match_len`] returns the total number of patterns MJ that match in a particular state and [`Automaton::match_pattern`] permitsA> iterating over the patterns that match in a particular state.IF * A DFA can have multiple start states, and the choice of which startHE state to use depends on the content of the string being searched andHE position of the search, as well as whether the search is an anchoredKH search for a specific pattern in the DFA. Moreover, computing the startMJ state also depends on whether you're doing a forward or a reverse search.MJ [`Automaton::start_state_forward`] and [`Automaton::start_state_reverse`]IF are used to compute the start state for forward and reverse searches, respectively.MJ * All matches are delayed by one byte to support things like `$` and `\b`MJ at the end of a pattern. Therefore, every use of a DFA is required to use! [`Automaton::next_eoi_state`]=: at the end of the search to compute the final transition.HE * For optimization reasons, some states are treated specially. EveryC@ state is either special or not, which can be determined via theKH [`Automaton::is_special_state`] method. If it's special, then the stateKH must be at least one of a few possible types of states. (Note that someMJ types can overlap, for example, a match state can also be an accel state.NK But some types can't. If a state is a dead state, then it can never be any*' other type of state.) Those types are:KH * A dead state. A dead state means the DFA will never enter a matchOL state. This can be queried via the [`Automaton::is_dead_state`] method.MJ * A quit state. A quit state occurs if the DFA had to stop the search@= prematurely for some reason. This can be queried via the,) [`Automaton::is_quit_state`] method.OL * A match state. A match state occurs when a match is found. When a DFAKH enters a match state, the search may stop immediately (when lookingNK for the earliest match), or it may continue to find the leftmost-firstHE match. This can be queried via the [`Automaton::is_match_state`]| method.JG * A start state. A start state is where a search begins. For everyIF search, there is exactly one start state that is used, however, aHE DFA may contain many start states. When the search is in a startJG state, it may use a prefilter to quickly skip to candidate matches LI without executing the DFA on every byte. This can be queried via the -* [`Automaton::is_start_state`] method.!HE * An accel state. An accel state is a state that is accelerated.!KH That is, it is a state where _most_ of its transitions loop back to"KH itself and only a small number of transitions lead to other states."MJ This kind of state is said to be accelerated because a search routine#JG can quickly look for the bytes leading out of the state instead of$OL continuing to execute the DFA on each byte. This can be queried via the$LI [`Automaton::is_accel_state`] method. And the bytes that lead out of%KH the state can be queried via the [`Automaton::accelerator`] method.%%GD There are a number of provided methods on this trait that implement&EB efficient searching (for forwards and backwards) with a DFA using'EB all of the above features of this trait. In particular, given the'FC complexity of all these features, implementing a search routine in(HE this trait can be a little subtle. With that said, it is possible to(KH somewhat simplify the search routine. For example, handling accelerated)JG states is strictly optional, since it is always correct to assume that)KH `Automaton::is_accel_state` returns false. However, one complex part of*OL writing a search routine using this trait is handling the 1-byte delay of a+  match. That is not optional.+d+5++DA This trait is not safe to implement so that code may rely on the,MJ correctness of implementations of this trait to avoid undefined behavior.,+( The primary correctness guarantees are:--MJ * `Automaton::start_state` always returns a valid state ID or an error or\- panics.-JG * `Automaton::next_state`, when given a valid state ID, always returns.JG a valid state ID for all values of `anchored` and `byte`, or otherwise\.6//KH In general, the rest of the methods on `Automaton` need to uphold their/LI contracts as well. For example, `Automaton::is_dead` should only returns085 true if the given state ID is actually a dead state.L0066#==0HE Transitions from the current state to the next state, given the next1 byte of input.11IF Implementations must guarantee that the returned ID is always a valid2DA ID when `current` refers to a valid ID. Moreover, the transition2@= function must be defined for all possible values of `input`.3d3ں.33FC If the given ID does not refer to a valid state, then this routine4IF may panic but it also may not panic and instead return an invalid ID.4FC However, if the caller provides an invalid ID then this must never5 sacrifice memory safety.5l5-66JG This shows a simplistic example for walking a DFA for a given haystack6%" by using the `next_state` method.7<7-79/77+( let dfa = dense::DFA::new(r"[a-z]+r")?;8$! let haystack = "bar".as_bytes();88GD // The start state is determined by inspecting the position and the9%" // initial bytes of the haystack.9DA let mut state = dfa.start_state_forward(&Input::new(haystack))?;9*' // Walk all the bytes in the haystack.Ī: for &b in haystack {:)& state = dfa.next_state(state, b);,:0:KH // Matches are always delayed by 1 byte, so we must explicitly walk the;96 // special "EOI" transition at the end of the search.<&# state = dfa.next_eoi_state(state);<'$ assert!(dfa.is_match_state(state));<<.-<=-T=) )B = $=<=,=Cm=H6>6>>HE Unlike [`Automaton::next_state`], implementations may implement this?FC more efficiently by assuming that the `current` state ID is valid.?74 Typically, this manifests by eliding bounds checks.@d@5@@JG Callers of this method must guarantee that `current` refers to a validAKH state ID. If `current` is not a valid state ID for this automaton, thenA:7 calling this routine may result in undefined behavior.BBJG If `current` is valid, then implementations must guarantee that the IDC96 returned is valid for all possible values of `input`.C) ) C $C be more succinct to use [`Automaton::start_state_forward`] oraC@ [`Automaton::start_state_reverse`]. Note, for example, that thebIF convenience routines return a [`MatchError`] on failure where as thisb%" routine returns a [`StartError`].cdc0ccHE This may return a [`StartError`] if the search needs to give up whendHE determining the start state (for example, if it sees a "quit" byte).dHE This can also return an error if the given configuration contains ane+( unsupported [`Anchored`] configuration.\e)) ) )B e$e4flffJG Return the ID of the start state for this DFA when executing a forward\g search.ggHE This is a convenience routine for calling [`Automaton::start_state`]g1. that converts the given [`Input`] to a [starthJG configuration](start::Config). Additionally, if an error occurs, it ishHE converted from a [`StartError`] to a [`MatchError`] using the offseti'$ information in the given [`Input`].idi0jjC@ This may return a [`MatchError`] if the search needs to give upjFC when determining the start state (for example, if it sees a "quit"kIF byte). This can also return an error if the given `Input` contains ank+ޕ7l))) ) )$$$$$$*0)B l $l>,mwfqJG Return the ID of the start state for this DFA when executing a reverse\rԗ7rrH7r1ǘ7sJ7sH֙7t'7tdt0tuC7uF˛7vI7v+ޕ7w))) ) )$$$$$$*0)B w $wC,xC{HE If this DFA has a universal starting state for the given anchor mode{JG and the DFA supports universal starting states, then this returns that| state's identifier.||FC A DFA is said to have a universal starting state when the starting}JG state is invariant with respect to the haystack. Usually, the starting}FC state is chosen depending on the bytes immediately surrounding the~KH starting position of a search. However, the starting state only differsKH when one or more of the patterns in the DFA have look-around assertions in its prefix.IF Stated differently, if none of the patterns in a DFA have look-aroundӀKH assertions in their prefix, then the DFA has a universal starting state)& and _may_ be returned by this method.сفJG It is always correct for implementations to return `None`, and indeed,KH this is what the default implementation does. When this returns `None`,JG callers must use either `start_state_forward` or `start_state_reverse`ǃ to get the starting state.t # Use case;8 There are a few reasons why one might want to use this:̈́ՄGD * If you know your regex patterns have no look-around assertions inIF their prefix, then calling this routine is likely cheaper and perhaps! more semantically meaningful.HE * When implementing prefilter support in a DFA regex implementation,C@ it is necessary to re-compute the start state after a candidateEB is returned from the prefilter. However, this is only needed whenGD there isn't a universal start state. When one exists, one can avoid! re-computing the start state.l-<-̔-%ٰ1܉ Anchored,4-HE // There are no look-around assertions in the prefixes of any of theҊ30 // patterns, so we get a universal start state.?< let dfa = DFA::new_many(&["[0-9]+", "[a-z]+$", "[A-Z]+"])?;΋?< assert!(dfa.universal_start_state(Anchored::No).is_some());@= assert!(dfa.universal_start_state(Anchored::Yes).is_some());׌ߌEB // One of the patterns has a look-around assertion in its prefix,@= // so this means there is no longer a universal start state.@= let dfa = DFA::new_many(&["[0-9]+", "^[a-z]+$", "[A-Z]+"])?;@= assert!(!dfa.universal_start_state(Anchored::No).is_some());A> assert!(!dfa.universal_start_state(Anchored::Yes).is_some());.-<-) )łB  $ŋEԿ0EB Returns true if and only if the given identifier corresponds to aEB "special" state. A special state is one or more of the following:B? a dead state, a quit state, a match state, a start state or anĒ accelerated state.ߒFC A correct implementation _may_ always return false for states thatIF are either start states or accelerated states, since that informationB? is only intended to be used for optimization purposes. CorrectǔJG implementations must return true if the state is a dead, quit or matchHE state. This is because search routines using this trait must be ableGD to rely on `is_special_state` as an indicator that a state may needFC special treatment. (For example, when a search routine sees a dead state, it must terminate.)IF This routine permits search implementations to use a single branch toKH check whether a state needs special attention before executing the nextØ74 transition. The example below shows how to do this.l-HE This example shows how `is_special_state` can be used to implement aFC correct search routine with minimal branching. In particular, thisFC search routine implements "leftmost" matching, which means that itIF doesn't immediately stop once a match is found. Instead, it continuesқ" until it reaches a dead state.<-̍- .М%" HalfMatch, MatchError, Input,4-ԍ fn find( dfa: &A, haystack: &[u8],ޝ0- ) -> Result, MatchError> {KH // The start state is determined by inspecting the position and theJG // initial bytes of the haystack. Note that start states can neverGD // be match states (since DFAs in this crate delay matches by 1JG // byte), so we don't need to check if the start state is a match.͠HE let mut state = dfa.start_state_forward(&Input::new(haystack))?;" let mut last_match = None;JG // Walk all the bytes in the haystack. We can quit early if we seeFC // a dead or a quit state. The former means the automaton willۢIF // never transition to any other state. The latter means that theDA // automaton entered a condition in which its search failed.41 for (i, &b) in haystack.iter().enumerate() {-* state = dfa.next_state(state, b);ݤ,) if dfa.is_special_state(state) {.+ if dfa.is_match_state(state) {52 last_match = Some(HalfMatch::new(41 dfa.match_pattern(state, 0),Դ i,Ӧ ));41 } else if dfa.is_dead_state(state) {*' return Ok(last_match);ק41 } else if dfa.is_quit_state(state) {FC // It is possible to enter into a quit state afterۨGD // observing a match has occurred. In that case, weC@ // should return the match instead of an error.-* if last_match.is_some() {.+ return Ok(last_match);Ԫ }74 return Err(MatchError::quit(b, i)); }IF // Implementors may also want to check for start or accelEB // states and handle them differently for performanceجDA // reasons. But it is not necessary for correctness.l }L }KH // Matches are always delayed by 1 byte, so we must explicitly walkA> // the special "EOI" transition at the end of the search.׮*' state = dfa.next_eoi_state(state);&# if dfa.is_match_state(state) {-* last_match = Some(HalfMatch::new(,) dfa.match_pattern(state, 0), haystack.len(),| ));L̰7ڰ Ok(last_match),0GD // We use a greedy '+' operator to show how the search doesn't justӱGD // stop once a match is detected. It continues extending the match.FC // Using '[a-z]+?' would also work as expected and stop the search52 // early. Greediness is built into the automaton.*' let dfa = dense::DFA::new(r"[a-z]+")?;ӳ0.-* let mat = find(&dfa, haystack)?.unwrap();,) assert_eq!(mat.pattern().as_usize(), 0);! assert_eq!(mat.offset(), 10);HE // Here's another example that tests our handling of the special EOIB? // transition. This will fail to find a match if we don't callFC // 'next_eoi_state' at the end of the search since the match isn't2/ // found until the final byte in the haystack.,) let dfa = dense::DFA::new(r"[0-9]{4}")?;0.-7Ǹ,7! assert_eq!(mat.offset(), 15);HE // And note that our search implementation above automatically worksKH // with multi-DFAs. Namely, `dfa.match_pattern(match_state, 0)` selectsú)& // the appropriate pattern ID for us.=: let dfa = dense::DFA::new_many(&[r"[a-z]+", r"[0-9]+"])?;0.-7,) assert_eq!(mat.pattern().as_usize(), 1);˼  assert_eq!(mat.offset(), 3);30 let mat = find(&dfa, &haystack[3..])?.unwrap();,7ٽ  assert_eq!(mat.offset(), 7);41 let mat = find(&dfa, &haystack[10..])?.unwrap();,7  assert_eq!(mat.offset(), 5);.-<ȿ-׿) )B  $--JG Returns true if and only if the given identifier corresponds to a deadJG state. When a DFA enters a dead state, it is impossible to leave. ThatHE is, every transition on a dead state by definition leads back to the same dead state.IF In practice, the dead state always corresponds to the identifier `0`.85 Moreover, in practice, there is only one dead state.KH The existence of a dead state is not strictly required in the classicalHE model of finite state machines, where one generally only cares aboutIF the question of whether an input sequence matches or not. Dead statesJG are not needed to answer that question, since one can immediately quitJG as soon as one enters a final or "match" state. However, we don't justGD care about matches but also care about the location of matches, andC@ more specifically, care about semantics like "greedy" matching.FC For example, given the pattern `a+` and the input `aaaz`, the deadEB state won't be entered until the state machine reaches `z` in theGD input, at which point, the search routine can quit. But without theC@ dead state, the search routine wouldn't know when to quit. In aHE classical representation, the search routine would stop after seeingKH the first `a` (which is when the search would enter a match state). ButHE this wouldn't implement "greedy" matching where `a+` matches as many `a`'s as possible.l-KH See the example for [`Automaton::is_special_state`] for how to use this method correctly.l) )B  $--JG Returns true if and only if the given identifier corresponds to a quitIF state. A quit state is like a dead state (it has no transitions otherHE than to itself), except it indicates that the DFA failed to completeKH the search. When this occurs, callers can neither accept or reject that a match occurred.KH In practice, the quit state always corresponds to the state immediatelyGD following the dead state. (Which is not usually represented by `1`,EB since state identifiers are pre-multiplied by the state machine'sB? alphabet stride, and the alphabet stride varies between DFAs.)EB The typical way in which a quit state can occur is when heuristic:7 support for Unicode word boundaries is enabled via the^[ [`dense::Config::unicode_word_boundary`](crate::dfa::dense::Config::unicode_word_boundary)30 option. But other options, like the lower level<9 [`dense::Config::quit`](crate::dfa::dense::Config::quit)EB configuration, can also result in a quit state being entered. TheGD purpose of the quit state is to provide a way to execute a fast DFAKH in common cases while delegating to slower routines when the DFA quits.KH The default search implementations provided by this crate will return a<9 [`MatchError::quit`] error when a quit state is entered.l-K77l) )B  $-.Eκ7IF match state. A match state is also referred to as a "final" state and*' indicates that a match has been found.HE If all you care about is whether a particular pattern matches in theGD input sequence, then a search routine can quit early as soon as theDA machine enters a match state. However, if you're looking for theIF standard "leftmost-first" match location, then search _must_ continueHE until either the end of the input or until the machine enters a deadHE state. (Since either condition implies that no other useful work canDA be done.) Namely, when looking for the location of a match, thenDA search implementations should record the most recent location inIF which a match state was entered, but otherwise continue executing theGD search as normal. (The search may even leave the match state.) OnceJG the termination condition is reached, the most recently recorded match  location should be returned.EB Finally, one additional power given to match states in this crateGD is that they are always associated with a specific pattern in orderDA to support multi-DFAs. See [`Automaton::match_pattern`] for moreIF details and an example for how to query the pattern associated with a particular match state.l-K77t) )B  $-.DA Returns true only if the given identifier corresponds to a startL state<9 A start state is a state in which a DFA begins a search.HE All searches begin in a start state. Moreover, since all matches areB? delayed by one byte, a start state can never be a match state.EB The main role of a start state is, as mentioned, to be a startingA> point for a DFA. This starting point is determined via one of)& [`Automaton::start_state_forward`] orIF [`Automaton::start_state_reverse`], depending on whether one is doing0- a forward or a reverse search, respectively.GD A secondary use of start states is for prefix acceleration. Namely,JG while executing a search, if one detects that you're in a start state,GD then it may be faster to look for the next match of a prefix of theIF pattern, if one exists. If a prefix exists and since all matches mustFC begin with that prefix, then skipping ahead to occurrences of that52 prefix may be much faster than executing the DFA.)& As mentioned in the documentation forEB [`is_special_state`](Automaton::is_special_state) implementationsFC _may_ always return false, even if the given identifier is a startJG state. This is because knowing whether a state is a start state or notGD is not necessary for correctness and is only treated as a potentialGD performance optimization. (For example, the implementations of thisGD trait in this crate will only return true when the given identifierB? corresponds to a start state and when [specialization of startKH states](crate::dfa::dense::Config::specialize_start_states) was enabledFC during DFA construction. If start state specialization is disabledGD (which is the default), then this method will always return false.)l-IF This example shows how to implement your own search routine that does=: a prefix search whenever the search enters a start state.B? Note that you do not need to implement your own search routine<9 to make use of prefilters like this. The search routinesB? provided by this crate already implement prefilter support via?< the [`Prefilter`](crate::util::prefilter::Prefilter) trait.>; A prefilter can be added to your search configuration withJG [`dense::Config::prefilter`](crate::dfa::dense::Config::prefilter) for(% dense and sparse DFAs in this crate.IF This example is meant to show how you might deal with prefilters in aDA simplified case if you are implementing your own search routine.<-- .%74-FC fn find_byte(slice: &[u8], at: usize, byte: u8) -> Option {EB // Would be faster to use the memchr crate, but this is still/, // faster than running through the DFA.C@ slice[at..].iter().position(|&b| b == byte).map(|i| at + i),0Ԋ77ľ7  prefix_byte: Option,07GD // See the Automaton::is_special_state example for similar code // with more comments.H7"7Ġ let mut pos = 0;$! while pos < haystack.len() {" let b = haystack[pos];-7 pos += 1;ك,7.75747  pos - 1,Յ747*7ن47F7݇G7C7-7.7։7=: return Err(MatchError::quit(b, pos - 1));52 } else if dfa.is_start_state(state) {KH // If we're in a start state and know all matches beginJG // with a particular byte, then we can quickly skip toHE // candidate matches without running the DFA through،,) // every byte inbetween.<9 if let Some(prefix_byte) = prefix_byte {ʍKH pos = match find_byte(haystack, pos, prefix_byte) {-* Some(pos) => pos,̎*' None => break, };77lʏ7L܏7K7A7*7&7ڑ-7,77|7L77,0HE // In this example, it's obvious that all occurrences of our patternC@ // begin with 'Z', so we pass in 'Z'. Note also that we need toŔIF // enable start state specialization, or else it won't be possible toJG // detect start states during a search. ('is_start_state' would always // return false.)# let dfa = dense::DFA::builder()FC .configure(dense::DFA::config().specialize_start_states(true)) .build(r"Z[a-z]+")?;52 let haystack = "123 foobar Zbaz quux".as_bytes();ʗ96 let mat = find(&dfa, haystack, Some(b'Z'))?.unwrap();,7!7ߘIF // But note that we don't need to pass in a prefix byte. If we don't,41 // then the search routine does no acceleration.30 let mat = find(&dfa, haystack, None)?.unwrap();,7ך!7IF // However, if we pass an incorrect byte, then the prefix search willӛ# // result in incorrect results.85 assert_eq!(find(&dfa, haystack, Some(b'X'))?, None);.-<-t) )B  $-˳.FC Returns true if and only if the given identifier corresponds to an72/ An accelerated state is a special optimization؞/, trick implemented by this crate. Namely, ifKH [`dense::Config::accelerate`](crate::dfa::dense::Config::accelerate) isܟJG enabled (and it is by default), then DFAs generated by this crate willJG tag states meeting certain characteristics as accelerated. States meetJG this criteria whenever most of their transitions are self-transitions.ɡGD That is, transitions that loop back to the same state. When a smallGD number of transitions aren't self-transitions, then it follows thatJG there are only a small number of bytes that can cause the DFA to leaveEB that state. Thus, there is an opportunity to look for those bytesGD using more optimized routines rather than continuing to run throughƤEB the DFA. This trick is similar to the prefilter idea described inDA the documentation of [`Automaton::is_start_state`] with two main٥ differences:JG 1. It is more limited since acceleration only applies to single bytes.ŦIF This means states are rarely accelerated when Unicode mode is enabled" (which is enabled by default).EB 2. It can occur anywhere in the DFA, which increases optimization opportunities.HE Like the prefilter idea, the main downside (and a possible reason toGD disable it) is that it can lead to worse performance in some cases.EB Namely, if a state is accelerated for very common bytes, then theFC overhead of checking for acceleration and using the more optimizedѪHE routines to look for those bytes can cause overall performance to be52 worse than if acceleration wasn't enabled at all.ث@= A simple example of a regex that has an accelerated state isHE `(?-u)[^a]+a`. Namely, the `[^a]+` sub-expression gets compiled downIF into a single state where all transitions except for `a` loop back toHE itself, and where `a` is the only transition (other than the specialGD EOI transition) that goes to some other state. Thus, this state canٮA> be accelerated and implemented more efficiently by calling anGD optimized routine like `memchr` with `a` as the needle. Notice thatDA the `(?-u)` to disable Unicode is necessary here, as without it,JG `[^a]` will match any UTF-8 encoding of any Unicode scalar value otherHE than `a`. This more complicated expression compiles down to many DFAбKH states and the simple acceleration optimization is no longer available.52 Typically, this routine is used to guard calls toGD [`Automaton::accelerator`], which returns the accelerated bytes forĮ the specified state.tγ) )B ݳ $޳-@= Returns the total number of patterns compiled into this DFA.Ĵ̴IF In the case of a DFA that contains no patterns, this must return `0`.l-GD This example shows the pattern length for a DFA that never matches:<-52 use regex_automata::dfa::{Automaton, dense::DFA};ֶ޶1. let dfa: DFA> = DFA::never_match()?;%" assert_eq!(dfa.pattern_len(), 0);.-<-A> And another example for a DFA that matches at every position:˸<Ӹ-߸582/ let dfa: DFA> = DFA::always_match()?;ع%" assert_eq!(dfa.pattern_len(), 1);.-<-ɺC@ And finally, a DFA that was constructed from multiple patterns:<-58߻>; let dfa = DFA::new_many(&["[0-9]+", "[a-z]+", "[A-Z]+"])?;%" assert_eq!(dfa.pattern_len(), 3);Լ.-<-\) )B  $*B? Returns the total number of patterns that match in this state.EB If the given state is not a match state, then implementations mayTѾͽ.HE If the DFA was compiled with one pattern, then this must necessarily+( always return `1` for all match states.KH Implementations must guarantee that [`Automaton::match_pattern`] can beHE called with indices up to (but not including) the length returned by# this routine without panicking.dں.HE Implementations are permitted to panic if the provided state ID does$! not correspond to a match state.l-DA This example shows a simple instance of implementing overlappingGD matches. In particular, it shows not only how to determine how manyGD patterns have matched in a particular state, but also how to access)& which specific patterns have matched.܂ Notice that we must use-* [`MatchKind::All`](crate::MatchKind::All)%" when building the DFA. If we usedA> [`MatchKind::LeftmostFirst`](crate::MatchKind::LeftmostFirst)@= instead, then the DFA would not be constructed in a way thatHE supports overlapping matches. (It would only report a single pattern2/ that matches at any particular point in time.)GD Another thing to take note of is the patterns used and the order inIF which the pattern IDs are reported. In the example below, pattern `3`C@ is yielded first. Why? Because it corresponds to the match thatGD appears first. Namely, the `@` symbol is part of `\S+` but not partJG of any of the other patterns. Since the `\S+` pattern has a match thatJG starts to the left of any other pattern, its ID is returned before anyT other.<-=-DA use regex_automata::{dfa::{Automaton, dense}, Input, MatchKind};#-C- .build_many(&[C@ r"[[:word:]]+", r"[a-z]+", r"[A-Z]+", r"[[:^space:]]+",d ])?;%" let haystack = "@bar".as_bytes();G6%6D6*6ě6)6,0&6'6(% assert_eq!(dfa.match_len(state), 3);HE // The following calls are guaranteed to not panic since `match_len` // returned `3` above.:7 assert_eq!(dfa.match_pattern(state, 0).as_usize(), 3);:7 assert_eq!(dfa.match_pattern(state, 1).as_usize(), 0);:7 assert_eq!(dfa.match_pattern(state, 2).as_usize(), 1);.-<-L) )B  $-@HE Returns the pattern ID corresponding to the given match index in the given state.B? See [`Automaton::match_len`] for an example of how to use thisGD method correctly. Note that if you know your DFA is compiled with aFC single pattern, then this routine is never necessary since it willC@ always return a pattern ID of `0` for an index of `0` when `id`! corresponds to a match state.DA Typically, this routine is used when implementing an overlapping;8 search, as the example for `Automaton::match_len` does.dں.EB If the state ID is not a match state or if the match index is outEB of bounds for the given state, then this routine may either panicFC or produce an incorrect result. If the state ID is correct and theIF match index is correct, then this routine must always produce a valid `PatternID`.l) )B  $-,JG Returns true if and only if this automaton can match the empty string.HE When it returns false, all possible matches are guaranteed to have a non-zero length.HE This is useful as cheap way to know whether code needs to handle theJG case of a zero length match. This is particularly important when UTF-8HE modes are enabled, as when UTF-8 mode is enabled, empty matches thatEB split a codepoint must never be reported. This extra handling canGD sometimes be costly, and since regexes matching an empty string areHE somewhat rare, it can be beneficial to treat such regexes specially.l-FC This example shows a few different DFAs and whether they match theFC empty string or not. Notice the empty string isn't merely a matterHE of a string of length literally `0`, but rather, whether a match can*' occur between specific pairs of bytes.<-EB use regex_automata::{dfa::{dense::DFA, Automaton}, util::syntax};0- // The empty regex matches the empty string. let dfa = DFA::new("")?;41 assert!(dfa.has_empty(), "empty matches empty");FC // The '+' repetition operator requires at least one match, and so'$ // does not match the empty string. let dfa = DFA::new("a+")?;85 assert!(!dfa.has_empty(), "+ does not match empty");,) // But the '*' repetition operator does. let dfa = DFA::new("a*")?;30 assert!(dfa.has_empty(), "* does match empty");JG // And wrapping '+' in an operator that can match an empty string also/, // causes it to match the empty string too.! let dfa = DFA::new("(a+)*")?;<9 assert!(dfa.has_empty(), "+ inside of * matches empty");FC // If a regex is just made of a look-around assertion, even if theJG // assertion requires some kind of non-empty string around it (such asGD // \b), then it is still treated as if it matches the empty string.JG // Namely, if a match occurs of just a look-around assertion, then the // match returned is empty.-=: .configure(DFA::config().unicode_word_boundary(true))23)& .build(r"^$\A\z\b\B(?-u:\b\B)")?;74 assert!(dfa.has_empty(), "assertions match empty");GD // Even when an assertion is wrapped in a '+', it still matches the // empty string. let dfa = DFA::new(r"^+")?;@= assert!(dfa.has_empty(), "+ of an assertion matches empty");JG // An alternation with even one branch that can match the empty string63 // is also said to match the empty string overall.+( let dfa = DFA::new("foo|(bar)?|quux")?;=: assert!(dfa.has_empty(), "alternations can match empty");C@ // An NFA that matches nothing does not match the empty string." let dfa = DFA::new("[a&&b]")?;IF assert!(!dfa.has_empty(), "never matching means not matching empty");GD // But if it's wrapped in something that doesn't require a match at/, // all, then it can match the empty string!# let dfa = DFA::new("[a&&b]*")?;EB assert!(dfa.has_empty(), "* on never-match still matches empty");IF // Since a '+' requires a match, using it on something that can neverGD // match will itself produce a regex that can never match anything,0- // and thus does not match the empty string.# let dfa = DFA::new("[a&&b]+")?;HE assert!(!dfa.has_empty(), "+ on never-match still matches nothing");.-<-L) )B ޏ$Զ63 Whether UTF-8 mode is enabled for this DFA or not.A> When UTF-8 mode is enabled, all matches reported by a DFA areC@ guaranteed to correspond to spans of valid UTF-8. This includesJG zero-width matches. For example, the DFA must guarantee that the emptyIF regex will not match at the positions between code units in the UTF-8# encoding of a single codepoint.JG See [`thompson::Config::utf8`](crate::nfa::thompson::Config::utf8) for more information.l-IF This example shows how UTF-8 mode can impact the match spans that may! be reported in certain cases.À<ˀ-׀-%- nfa::thompson,̺/4؁-(% // UTF-8 mode is enabled by default.ܘ let re = DFA::new("")?;Ը assert!(re.is_utf8());ׂ&# let mut input = Input::new("☃");)& let got = re.try_search_fwd(&input)?;1. assert_eq!(Some(HalfMatch::must(0, 0)), got);DA // Even though an empty regex matches at 1..1, our next match isGD // 3..3 because 1..1 and 2..2 split the snowman codepoint (which is̃ // three bytes long). input.set_start(1);)91. assert_eq!(Some(HalfMatch::must(0, 3)), got);HE // But if we disable UTF-8, then we'll get matches at 1..1 and 2..2:-6ܭ3ч .build("")?; assert!(!re.is_utf8());)91. assert_eq!(Some(HalfMatch::must(0, 1)), got); input.set_start(2);)91. assert_eq!(Some(HalfMatch::must(0, 2)), got); input.set_start(3);)9Ȋ19 input.set_start(4);)9Ћ assert_eq!(None, got);.-<-<) )B $Œ+֌HE Returns true if and only if this DFA is limited to returning matches  whose start position is `0`.ȍЍ.+ Note that if you're using DFAs provided by,) this crate, then this is _orthogonal_ toB? [`Config::start_kind`](crate::dfa::dense::Config::start_kind).IF This is useful in some cases because if a DFA is limited to producingяDA matches that start at offset `0`, then a reverse search is never.+ required for finding the start of a match.͐lՐ-<-52 use regex_automata::dfa::{dense::DFA, Automaton};'$ // The empty regex matches anywhere9GD assert!(!dfa.is_always_start_anchored(), "empty matches anywhere");֒ // 'a' matches anywhere. let dfa = DFA::new("a")?;EB assert!(!dfa.is_always_start_anchored(), "'a' matches anywhere");$! // '^' only matches at offset 0! let dfa = DFA::new("^a")?;FC assert!(dfa.is_always_start_anchored(), "'^a' matches only at 0");:7 // But '(?m:^)' matches at 0 but at other offsets too.# let dfa = DFA::new("(?m:^)a")?;KH assert!(!dfa.is_always_start_anchored(), "'(?m:^)a' matches anywhere");.-<-) )B  $,KH Return a slice of bytes to accelerate for the given state, if possible.FC If the given state has no accelerator, then an empty slice must be̘KH returned. If `Automaton::is_accel_state` returns true for the given ID,JG then this routine _must_ return a non-empty slice. But note that it isJG not required for an implementation of this trait to ever return `true`HE for `is_accel_state`, even if the state _could_ be accelerated. ThatJG is, acceleration is an optional optimization. But the return values of֛74 `is_accel_state` and `accelerator` must be in sync.DA If the given ID is not a valid state ID for this automaton, then;8 implementations may panic or produce incorrect results.?< See [`Automaton::is_accel_state`] for more details on state/>; By default, this method will always return an empty slice.Оl؞-IF This example shows a contrived case in which we build a regex that weA> know is accelerated and extract the accelerator from a state.<-̚- .ݠ,) util::{primitives::StateID, syntax},4-#-ɡFC // We disable Unicode everywhere and permit the regex to matchGD // invalid UTF-8. e.g., [^abc] matches \xFF, which is not validJG // UTF-8. If we left Unicode enabled, [^abc] would match any UTF-8KH // encoding of any Unicode scalar value except for 'a', 'b' or 'c'.C@ // That translates to a much more complicated DFA, and alsoǤHE // inhibits the 'accelerator' optimization that we are trying to'$ // demonstrate in this example.A> .syntax(syntax::Config::new().unicode(false).utf8(false))䆦 .build("[^abc]+a")?;DA // Here we just pluck out the state that we know is accelerated.EB // While the stride calculations are something that can be relied§GD // on by callers, the specific position of the accelerated state is // implementation defined.4ȃ7IF // N.B. We get '3' by inspecting the state machine using 'regex-cli'.A> // e.g., try `regex-cli debug dense dfa -p '[^abc]+a' -BbUC`.Щ52 let id = StateID::new(3 * dfa.stride()).unwrap();*' let accelerator = dfa.accelerator(id);HE // The `[^abc]+` sub-expression permits [a, b, c] to be accelerated.1. assert_eq!(accelerator, &[b'a', b'b', b'c']);.-<-\) ) )B  $NFñ-Ϭ?< Returns the prefilter associated with a DFA, if one exists.GD The default implementation of this trait always returns `None`. And2/ indeed, it is always correct to return `None`.DA For DFAs in this crate, a prefilter can be attached to a DFA viaGD [`dense::Config::prefilter`](crate::dfa::dense::Config::prefilter).ïEB Do note that prefilters are not serialized by DFAs in this crate.C@ So if you deserialize a DFA that had a prefilter attached to itհB? at serialization time, then it will not have a prefilter after deserialization.lƱ) )řǙșəʙ˙G*R{9  )B Ա $ձkJG Executes a forward search and returns the end position of the leftmostڲEB match that is found. If no match exists, then `None` is returned.GD In particular, this method continues searching even after it entersEB a match state. The search only terminates once it has reached the´FC end of the input or when it has entered a dead or quit state. UponJG termination, the position of the last byte seen while still in a matchܵ state is returned.d0H3!3Fض3޷E3J31˸3C35ӹ3G33䢻 # Notes for implementorsû˻KH Implementors of this trait are not required to implement any particularKH match semantics (such as leftmost-first), which are instead manifest inEB the DFA's transitions. But this search routine should behave as a general "leftmost" search.ؽKH In particular, this method must continue searching even after it entersGD a match state. The search should only terminate once it has reachedJG the end of the input or when it has entered a dead or quit state. Upon˿J99KH Since this trait provides an implementation for this method by default,74 it's unlikely that one will need to implement this.l-41 This example shows how to use this method with a+( [`dense::DFA`](crate::dfa::dense::DFA).<-D/,) let dfa = dense::DFA::new("foo[0-9]+")?;/0HE assert_eq!(expected, dfa.try_search_fwd(&Input::new(b"foo12345"))?);G3JG // the leftmost first match semantics demand that we find the earliestIF // match that prefers earlier parts of the pattern over latter parts.(% let dfa = dense::DFA::new("abc|a")?;/, let expected = Some(HalfMatch::must(0, 3));C@ assert_eq!(expected, dfa.try_search_fwd(&Input::new(b"abc"))?);.-<-&# # Example: specific pattern searchJG This example shows how to build a multi-DFA that permits searching for specific patterns.<-=-- ..-4-#-F-;8 .build_many(&["[a-z0-9]{6}", "[a-z][a-z0-9]{5}"])?;'$ let haystack = "foo123".as_bytes();C@ // Since we are using the default leftmost-first match and bothKH // patterns match at the same starting position, only the first patternGD // will be returned in this case when doing a search for any of the // patterns.//9..JG // But if we want to check whether some other pattern matches, then we" // can provide its pattern ID.$-96 .anchored(Anchored::Pattern(PatternID::must(1)));/, let expected = Some(HalfMatch::must(1, 6));*' let got = dfa.try_search_fwd(&input)?;..-<-0- # Example: specifying the bounds of a searchGD This example shows how providing the bounds of a search can produce;8 different results than simply sub-slicing the haystack.<-D/FC // N.B. We disable Unicode here so that we use a simple ASCII wordEB // boundary. Alternatively, we could enable heuristic support for // Unicode word boundaries.52 let dfa = dense::DFA::new(r"(?-u)\b[0-9]{3}\b")?;*/IF // Since we sub-slice the haystack, the search doesn't know about theB? // larger context and assumes that `123` is surrounded by wordIF // boundaries. And of course, the match position is reported relativeGD // to the sub-slice as well, which means we get `3` instead of `6`.,) let input = Input::new(&haystack[3..6]);/9*9.KH // But if we provide the bounds of the search within the context of theHE // entire haystack, then the search can take the surrounding contextFC // into account. (And if we did find a match, it would be reportedC@ // as a valid offset into `haystack` instead of its sub-slice.)1. let input = Input::new(haystack).range(3..6);ī let expected = None;*9..-<-t))) ) )$$$$$$*0) B  OK<$kJG Executes a reverse search and returns the start of the position of theDA leftmost match that is found. If no match exists, then `None` islϳ/d0H3!3Fض3E3J31˸3C35ӹ3G33l-49?< [`dense::DFA`](crate::dfa::dense::DFA). In particular, thisC@ routine is principally useful when used in conjunction with theMJ [`nfa::thompson::Config::reverse`](crate::nfa::thompson::Config::reverse)A> configuration. In general, it's unlikely to be correct to useFC both `try_search_fwd` and `try_search_rev` with the same DFA sinceHE any particular DFA will only support searching in one direction with܅ respect to the pattern.<-̹-9 .̗/4-#-85 .thompson(thompson::Config::new().reverse(true)) .build("foo[0-9]+")?;//HE assert_eq!(expected, dfa.try_search_rev(&Input::new(b"foo12345"))?);FC // Even though a match is found after reading the last byte (`c`),J9I9#-8:̥ .build("abc|c")?;//C@ assert_eq!(expected, dfa.try_search_rev(&Input::new(b"abc"))?);.-<-̎ # Example: UTF-8 modeA> This examples demonstrates that UTF-8 mode applies to reverseDA DFAs. When UTF-8 mode is enabled in the underlying NFA, then allÀHE matches reported must correspond to valid UTF-8 spans. This includes:7 prohibiting zero-width matches that split a codepoint.ρׁJG UTF-8 mode is enabled by default. Notice below how the only zero-width30 matches reported are those at UTF-8 boundaries:ނ<--%-9Ճ$Ǵ-4-䑄-8: .build(r"")?;2/ // Run the reverse DFA to collect all matches.ȅ&9 let mut matches = vec![];T loop {+( match dfa.try_search_rev(&input)? {Ԇ None => break, Some(hm) => {! matches.push(hm);96 if hm.offset() == 0 || input.end() == 0 { break;52 } else if hm.offset() < input.end() {Έ/, input.set_end(hm.offset());Ă } else {B? // This is only necessary to handle zero-widthFC // matches, which of course occur in this example.C@ // Without this, the search would never advance:7 // backwards beyond the initial match.30 input.set_end(input.end() - 1);7l7L7,0$! // No matches split a codepoint. let expected = vec![ HalfMatch::must(0, 3), HalfMatch::must(0, 0),4č ];ύ" assert_eq!(expected, matches);.-<-ŎB? Now let's look at the same example, but with UTF-8 mode on theGD original NFA disabled (which results in disabling UTF-8 mode on theL؏ DFA):<--%-9ݐ$Ǵ-4-䙑-DA .thompson(thompson::Config::new().reverse(true).utf8(false)):2:ܒ&9쇓:Tס:+::̇Ţ:!:˔9:ԉأ:5:/:Ė:B:F:ŗC:::̘3ʧ:7l7L7,0ę̙$ͨ::: HalfMatch::must(0, 2),ؚ HalfMatch::must(0, 1),ũ:4:":Л؛.-<-t))) ) )$$$$$$*0) B  "XT<ş$ѮˠJG Executes an overlapping forward search. Matches, if one exists, can be<9 obtained via the [`OverlappingState::get_match`] method.ۡGD This routine is principally only useful when searching for multipleIF patterns on inputs where multiple patterns may match the same regionsHE of text. In particular, callers must preserve the automaton's searchʣJG state from prior calls so that the implementation knows where the last match occurred.C@ When using this routine to implement an iterator of overlappingFC matches, the `start` of the search should always be set to the end̥GD of the last match. If more patterns match at the previous location,IF then they will be immediately returned. (This is tracked by the givenGD overlapping state.) Otherwise, the search continues at the starting position given.ʧҧGD If for some reason you want the search to forget about its previousKH state and restart the search at a particular position, then setting the>; state to [`OverlappingState::start`] will accomplish that.d0ʩҩH3!3ŪͪFض3E3J31˸3C35ӹ3G33Ԯlܮ-C@ This example shows how to run a basic overlapping search with aDA [`dense::DFA`](crate::dfa::dense::DFA). Notice that we build theIF automaton with a `MatchKind::All` configuration. Overlapping searchesհC@ are unlikely to work as one would expect when using the defaultDA `MatchKind::LeftmostFirst` match semantics, since leftmost-firstEB matching is fundamentally incompatible with overlapping searches.IF Namely, overlapping searches need to report matches as they are seen,IF where as leftmost-first searches will continue searching even after a̳JG match has been observed in order to find the conventional end positionJG of the match. More concretely, leftmost-first searches use dead statesKH to terminate a search after a specific match can no longer be extended.IF Overlapping searches instead do the opposite by continuing the search@= to find totally new matches (potentially of other patterns).Ͷ<ն-=-̣-2-$Ǵ-4-#-ܸC-<9 .build_many(&[r"[[:word:]]+$", r"[[:^space:]]+$"])?; let haystack = "@foo";.-/-GD dfa.try_search_overlapping_fwd(&Input::new(haystack), &mut state)?;,-I-ƼF-G-ݽI޺-(-ؾ/-G:ؿ,-.-<-)))) ) )$$$$$$*0) )B ^\\tJG Executes a reverse overlapping forward search. Matches, if one exists,C@ can be obtained via the [`OverlappingState::get_match`] method.C:IF matches, the `start` of the search should remain invariant throughoutIF iteration. The `OverlappingState` given to the search will keep trackDA of the current position of the search. (This is because multipleDA matches may be reported at the same position, so only the search>; implementation itself knows when to advance the position.)G:KĽ:>:d0H3!3Fض3E3J31˸3C35ӹ3G33:Aٛ:D:H::ǝ:J:3:<-̥-74 dfa::{dense::DFA, Automaton, OverlappingState},9$Ǵ-4--<9 .configure(DFA::config().match_kind(MatchKind::All))8:%" .build_many(&[r"", r"☃"])?;2:" let input = Input::new("☃");.-:Tס:<9 dfa.try_search_overlapping_rev(&input, &mut state)?;! match state.get_match() {:)& Some(hm) => matches.push(hm),L7,0$ͨ::: HalfMatch::must(1, 0),ũ:4:":.-<-B:G:L:<-̢-7:9$Ǵ-4--<:D:%:2:":.-:Tס:<:!::):L7,0:7 // Now *all* positions match, even within a codepoint,52 // because we lifted the requirement that matches'$ // correspond to valid UTF-8 spans.:::::ũ:4:":.-<-)))) ) )$$$$$$*0) )B  c%b0btFC Writes the set of patterns that match anywhere in the given searchEB configuration to `patset`. If multiple patterns match at the sameJG position and the underlying DFA supports overlapping matches, then all30 matching patterns are written to the given set.GD Unless all of the patterns in this DFA are anchored, then generally96 speaking, this will visit every byte in the haystack.IF This search routine *does not* clear the pattern set. This gives someGD flexibility to the caller (e.g., running multiple searches with theHE same pattern set), but does make the API bug-prone if you're reusingFC the same pattern set for multiple searches but intended them to be independent.DA If a pattern ID matched but the given `PatternSet` does not haveIF sufficient capacity to store it, then it is not inserted and silentlyd dropped.d0H3!3Fض3E3J31˸3C35ӹ3G33l-GD This example shows how to find all matching patterns in a haystack,IF even when some patterns match at the same position as other patterns.<-=-̙-%ٰ1%" Input, MatchKind, PatternSet,4- let patterns = &[ r"[[:word:]]+", r"[0-9]+", r"[[:alpha:]]+",| r"foo",| r"bar", r"barfoo", r"foobar",4:-<:ˀ .build_many(patterns)?;%" let input = Input::new("foobar");85 let mut patset = PatternSet::new(dfa.pattern_len());ށ<9 dfa.try_which_overlapping_matches(&input, &mut patset)?;'$ let expected = vec![0, 2, 3, 4, 6];˂HE let got: Vec = patset.iter().map(|p| p.as_usize()).collect();.Ã.-<-̂채)))) ) )$$$$$$*0) )B ؄ $ل*g8g:4L <T) )B  $iil) )  $ij5t) )B  $jc\)) ) )B  $kfÎ))) ) )$$$$$$*0)B  $Wlf))) ) )$$$$$$*0)B  $mB) )łB  $m/) )B  $-n,l) )B  $-]o,l) )B  $-p-t) )B  $-pە-tޕ) )B  $-qȖ-t˖) )B ږ $ۖ-Yr\) )B ė $ŗ)L) )B  $-s?l) )B  $-ttL) )B  $˚<Κ) )B ֚ $ך*Ğ* *B  $+\* * *B  $-9w-l* *řǙșəʙ˙G*R{9  *B  $ߝkt*** * *$$$$$$*0* B  $xkt*** * *$$$$$$*0* B  $ayԾ**** * *$$$$$$*0* *B  $.z9zԢ**** * *$$$$$$*0* *B Ƣ $Ǣ{{줤**** * *$$$$$$*0* *B ˤ $̤{|ӽϾ"HE Create a new overlapping state that begins at the start state of anyt-,־B,JG Return the match result of the most recent search to execute with thisT state.DA A searches will clear this result automatically, such that if no63 match is found, this will correctly report `None`.L* * B ڶ$|*$Bޡ$@łB $%|&gD*** * *ggg cgcثq1ȡ*B  $ "*** *$$$$$$*0* *ٗB ٗ ***#ٗ *$$$$$$*0* *B***#ٗ; ٗbmX4** *ggg cgcثq1ȡ*B 3 Ų3L -8-C/1|z,zz\z WD (C$WT]$r D$ $ <,,$$|$L!<"D%L%T&,#, D#,333!z,Olo^alT3gE$3;D,/3"%33d\B  z *** * * *1B  $ .  <T* *;B  $$3$* *;B  $%Et* *Ͷб;B  $,d* *Ͷ;B  $,lE|EEDEFFH*H1B $HQUhdh* *ÜĜŜƜǜȜ_wB i $idfa_id4i$inotn** * * 1B n $n dfa_state_idsln|** *ł * 1B  $DͶ;ll* *Ƶ;B ̞ $ў nfa_startLߞ,u|* *Ƶ;B  $<,;UgL* * B  $<4* *B ˿ $п9* *B  $<d* *B  $<d{,{zD{ <0c WD? MIQ=[CHL,$ D L!<"$l !,*24IDV,ffyf?fgrt`bLS U &tCCC=C=CCC=C= ̨TCCC=C=<;%g8*** * *ggg cgcثq1ȡ*B  $ 9;B B $UU$8U ;,@  ; b7db** * *ͶB b $b%bcU c}c*** * * *B c $c d%dincomingDd$e; SSSSS U f=f* *;B g gTo1UpHp* *;B p plss,sB tt* *B t $t-tċuu* *B u $uudu* *B u $uv,v* *B v $vww* *B w $wwDw* *B w $wx Tx* *B x $xy+$y* *ٗB y; y#ٗty#ٗ ٗ$y yEGz=dz*** * * *B z $z ,zdest$z͂9DЂ*** * * *B ق $ڂ ,;$;;d{,{{{0{;D{ C<3"|)  -T  6 g 6,; 68;, L  * *B  $  <  * *B  $ 6\ * *B  $116, * *B ;#;#;< ;$ LNl!#2# * *WBW #;|#;T# ; #%'$;;&y$& ** * *;BW ';|';T' ;$' '1'1'68*6,* * *;BW *;|*;T* ;$* *$&46;  ?ʅ?(D?* *ʅB ? $?-?@.\@* *ʅB @ $@,ALC,CC,CDDv`Ast`->`Hir`->`NFA` process. * Parsing a string into an `Ast` gives it a structured representation. Crucially, the size and amount of work done in this step is proportional to the size of the original string. No optimization or Unicode handling is done at this point. This means that parsing into an `Ast` has very predictable costs. Moreover, an `Ast` can be round-tripped back to its original pattern string as written. * Translating an `Ast` into an `Hir` is a process by which the structured representation is simplified down to its most fundamental components. Translation deals with flags such as case insensitivity by converting things like `(?i:a)` to `[Aa]`. Translation is also where Unicode tables are consulted to resolve things like `\p{Emoji}` and `\p{Greek}`. It also flattens each character class, regardless of how deeply nested it is, into a single sequence of non-overlapping ranges. All the various literal forms are thrown out in favor of one common representation. Overall, the `Hir` is small enough to fit into your head and makes analysis and other tasks much simpler. * Compiling an `Hir` into an `NFA` formulates the regex into a finite state machine whose transitions are defined over bytes. For example, an `Hir` might have a Unicode character class corresponding to a sequence of ranges defined in terms of `char`. Compilation is then responsible for turning those ranges into a UTF-8 automaton. That is, an automaton that matches the UTF-8 encoding of just the codepoints specified by those ranges. Otherwise, the main job of an `NFA` is to serve as a byte-code of sorts for a virtual machine. It can be seen as a sequence of instructions for how to match a regex. <0 0D0 < T  \ ߢ ,, T  D 4 l \<  K"K'=1CTi.$ ߢT$\ L!T!<"LR,LH7:D,,<<b; can be quite useful in the construction process of an NFA.d0KH This returns an error if the state identifier space is exhausted, or if52 the configured heap size limit has been exceeded.L+ +B  $k Add a "union" NFA state.HE A "union" NFA state that contains zero or more unconditional epsilonC@ transitions to other NFA states. The order of these transitionsJG reflects a priority order where earlier transitions are preferred over later transitions.KH Callers may provide an empty set of alternates to this method call, andJG then later add transitions via `patch`. At final build time, a "union"JG state with no alternates is converted to a "fail" state, and a "union"HE state with exactly one alternate is treated as if it were an "empty"T;d0K=5ƒ=L+ + B  $Ts$! Add a "reverse union" NFA state.ҁځKH A "reverse union" NFA state contains zero or more unconditional epsilonC=C@ reflects a priority order where later transitions are preferredEB over earlier transitions. This is an inverted priority order whenJG compared to `add_union`. This is useful, for example, for implementingӄ$! non-greedy repetition operators.K=ԅKH then later add transitions via `patch`. At final build time, a "reverseIF union" state with no alternates is converted to a "fail" state, and aHE "reverse union" state with exactly one alternate is treated as if itԿ were an "empty" state.އd0K=ψ5ƒ=+ + B  $Td俊 Add a "range" NFA state.JG A "range" NFA state is a state with one outgoing transition to anotherJG state, where that transition may only be followed if the current input.+ byte falls between a range of bytes given.d0ҌڌK=5ƒ=L+ +B  $!,p샏 Add a "sparse" NFA state.JG A "sparse" NFA state contains zero or more outgoing transitions, whereIF the transition to be followed (if any) is chosen based on whether theʐEB current input byte falls in the range of one such transition. TheJG transitions given *must* be non-overlapping and in ascending order. (AHE "sparse" state with no transitions is equivalent to a "fail" state.)HE A "sparse" state is like adding a "union" state and pointing it at aFC bunch of "range" states, except that the different alternates haveГ equal priority.HE Note that a "sparse" state is the only state that cannot be patched.HE This is because a "sparse" state has many transitions, each of whichB? may point to a different NFA state. Moreover, adding more suchѕGD transitions requires more than just an NFA state ID to point to. ItHE also requires a byte range. The `patch` routine does not support theHE additional information required. Therefore, callers must ensure thatJG all outgoing transitions for this state are included when `add_sparse`1. is called. There is no way to add more later.dĘ0՘ݘK=5ƒ=dں.HE This routine _may_ panic if the transitions given overlap or are not՚ in ascending order.T+ +B  $-\sܟ Add a "look" NFA state.ǜ>; A "look" NFA state corresponds to a state with exactly oneEB *conditional* epsilon transition to another NFA state. Namely, itԝFC represents one of a small set of simplistic look-around operators.IF Callers may provide a "dummy" state ID (typically [`StateID::ZERO`]),<9 and then change it later with [`patch`](Builder::patch).d0ϟןK=5ƒ=D+ +B  $ $$$! Add a "start capture" NFA state.GD A "start capture" NFA state corresponds to a state with exactly oneFC outgoing unconditional epsilon transition to another state. UnlikeףC@ "empty" states, a "start capture" state also carries with it anHE instruction for saving the current position of input to a particularGD location in memory. NFA simulations, like the Pike VM, may use thisFC information to report the match locations of capturing groups in a regex pattern.HE If the corresponding capturing group has a name, then callers should include it here.I˴=ާ<=JG Note that unlike `start_pattern`/`finish_pattern`, capturing start andHE end states may be interleaved. Indeed, it is typical for many "startéGD capture" NFA states to appear before the first "end capture" state.d0K=DA the configured heap size limit has been exceeded or if the givenɫ$! capture index overflows `usize`.EB While the above are the only conditions in which this routine canĬIF currently return an error, it is possible to call this method with anIF inputs that results in the final `build()` step failing to produce anHE NFA. For example, if one adds two distinct capturing groups with theHE same name, then that will result in `build()` failing with an error.DA See the [`GroupInfo`](crate::util::captures::GroupInfo) type for˯A> more information on what qualifies as valid capturing groups.l-JG This example shows that an error occurs when one tries to add multiple<9 capturing groups with the same name to the same pattern.ñ<˱-ױ- nfa::thompson::Builder," util::primitives::StateID,4-˲Ӳ1. let name = Some(std::sync::Arc::from("foo"));%" let mut builder = Builder::new();쳳 builder.start_pattern()?;ճ2/ // 0th capture group should always be unnamed.C@ let start = builder.add_capture_start(StateID::ZERO, 0, None)?;LԴ // OK?< builder.add_capture_start(StateID::ZERO, 1, name.clone())?;GD // This is not OK, but 'add_capture_start' still succeeds. We don'tGD // get an error until we call 'build' below. Without this call, the+( // call to 'build' below would succeed.?< builder.add_capture_start(StateID::ZERO, 2, name.clone())?;96 // Finish our pattern so we can try to build the NFA.# builder.finish_pattern(start)?;-* let result = builder.build(start, start);ʸ assert!(result.is_err());.-<-C@ However, adding multiple capturing groups with the same name to distinct patterns is okay:<- use std::sync::Arc;ֺ޺-=: nfa::thompson::{pikevm::PikeVM, Builder, Transition},| util::{һ captures::Captures,-* primitives::{PatternID, StateID},T },l Span,4ɼ-ԼܼGD // Hand-compile the patterns '(?P[a-z])' and '(?P[A-Z])'.%=ҽFC // We compile them to support an unanchored search, which requiresKH // adding an implicit '(?s-u:.)*?' prefix before adding either pattern.?< let unanchored_prefix = builder.add_union_reverse(vec![])?;,) let any = builder.add_range(Transition {:7 start: b'\x00', end: b'\xFF', next: StateID::ZERO,D })?;+( builder.patch(unanchored_prefix, any)?;+( builder.patch(any, unanchored_prefix)?;FC // Compile an alternation that permits matching multiple patterns.)& let alt = builder.add_union(vec![])?;+( builder.patch(unanchored_prefix, alt)?;! // Compile '(?P[a-z]+)'.=DA let start0 = builder.add_capture_start(StateID::ZERO, 0, None)?;52 // N.B. 0th capture group must always be unnamed./, let foo_start0 = builder.add_capture_start(1. StateID::ZERO, 1, Some(Arc::from("foo")),< )?;2/ let lowercase = builder.add_range(Transition {41 start: b'a', end: b'z', next: StateID::ZERO,D=>; let foo_end0 = builder.add_capture_end(StateID::ZERO, 1)?;:7 let end0 = builder.add_capture_end(StateID::ZERO, 0)?;&# let match0 = builder.add_match()?;'$ builder.patch(start0, foo_start0)?;*' builder.patch(foo_start0, lowercase)?;(% builder.patch(lowercase, foo_end0)?;# builder.patch(foo_end0, end0)?;! builder.patch(end0, match0)?;$! builder.finish_pattern(start0)?;! // Compile '(?P[A-Z]+)'.=DA let start1 = builder.add_capture_start(StateID::ZERO, 0, None)?;5=/, let foo_start1 = builder.add_capture_start(1=<=2/ let uppercase = builder.add_range(Transition {41 start: b'A', end: b'Z', next: StateID::ZERO,D=>; let foo_end1 = builder.add_capture_end(StateID::ZERO, 1)?;:7 let end1 = builder.add_capture_end(StateID::ZERO, 0)?;&# let match1 = builder.add_match()?;'$ builder.patch(start1, foo_start1)?;*' builder.patch(foo_start1, uppercase)?;(% builder.patch(uppercase, foo_end1)?;# builder.patch(foo_end1, end1)?;! builder.patch(end1, match1)?;$! builder.finish_pattern(start1)?;EB // Now add the patterns to our alternation that we started above.  builder.patch(alt, start0)?;  builder.patch(alt, start1)?;IF // Finally build the NFA. The first argument is the anchored startingA> // state (the pattern alternation) where as the second is the96 // unanchored starting state (the unanchored prefix).52 let nfa = builder.build(alt, unanchored_prefix)?;DA // Now build a Pike VM from our NFA and access the 'foo' captureEB // group regardless of which pattern matched, since it is defined // for both patterns.(% let vm = PikeVM::new_from_nfa(nfa)?;&# let mut cache = vm.create_cache(); let caps: Vec =<9 vm.captures_iter(&mut cache, "0123aAaAA").collect(); assert_eq!(5, caps.len());<9 assert_eq!(Some(PatternID::must(0)), caps[0].pattern());IF assert_eq!(Some(Span::from(4..5)), caps[0].get_group_by_name("foo"));<9 assert_eq!(Some(PatternID::must(1)), caps[1].pattern());IF assert_eq!(Some(Span::from(5..6)), caps[1].get_group_by_name("foo"));<9 assert_eq!(Some(PatternID::must(0)), caps[2].pattern());IF assert_eq!(Some(Span::from(6..7)), caps[2].get_group_by_name("foo"));<9 assert_eq!(Some(PatternID::must(1)), caps[3].pattern());IF assert_eq!(Some(Span::from(7..8)), caps[3].get_group_by_name("foo"));<9 assert_eq!(Some(PatternID::must(1)), caps[4].pattern());IF assert_eq!(Some(Span::from(8..9)), caps[4].get_group_by_name("foo"));.-<-+ +B  $ $ӎ\ $" Add a "end capture" NFA state.EB A "end capture" NFA state corresponds to a state with exactly oneF=A> "empty" states, a "end capture" state also carries with it anH̹=G=F=I˴=<=J=H=G=d0K=Dſ=$=E=I=I=H=H=D=A=|+ +B  $ $ӎ\9 Adds a "fail" NFA state.IF A "fail" state is simply a state that has no outgoing transitions. ItFC acts as a way to cause a search to stop without reporting a match.IF For example, one way to represent an NFA with zero patterns is with aĿ single "fail" state.d0K=5ƒ=D+ +B  $ۅ: Adds a "match" NFA state.C@ A "match" state has no outgoing transitions (just like a "fail"ހFC state), but it has special significance in that if a search entersJG this state, then a match has been found. The match state that is addedHE automatically has the current pattern ID associated with it. This isł:7 used to report the matching pattern ID at search time.d0K=5ƒ=dں.ȄЄC@ This must be called after a `start_pattern` call but before the>; corresponding `finish_pattern` call. Otherwise, it panics.L+ +B  $>+ +B  $, Ul/, Add a transition from one state to another.EB This routine is called "patch" since it is very common to add theJG states you want, typically with "dummy" state ID transitions, and thenIF "patch" in the real state IDs later. This is because you don't alwaysݍEB know all of the necessary state IDs to add because they might nott exist yet.dŽ0ӎێHE This may error if patching leads to an increase in heap usage beyondIF the configured size limit. Heap usage only grows when patching adds a74 new transition (as in the case of a "union" state).dں.ːӐHE This panics if `from` corresponds to a "sparse" state. When "sparse"KH states are added, there is no way to patch them after-the-fact. (If youIF have a use case where this would be helpful, please file an issue. It# will likely require a new API.),+ +B  $$1%IF Set whether the NFA produced by this builder should only match UTF-8.;8 This should be set when both of the following are true:IF 1. The caller guarantees that the NFA created by this build will only=: report non-empty matches with spans that are valid UTF-8.ПIF 2. The caller desires regex engines using this NFA to avoid reportingJG empty matches with a span that splits a valid UTF-8 encoded codepoint.EB Property (1) is not checked. Instead, this requires the caller toHE promise that it is true. Property (2) corresponds to the behavior ofFC regex engines using the NFA created by this builder. Namely, thereעGD is no way in the NFA's graph itself to say that empty matches foundHE by, for example, the regex `a*` will fall on valid UTF-8 boundaries.KH Instead, this option is used to communicate the UTF-8 semantic to regexIF engines that will typically implement it as a post-processing step byDA filtering out empty matches that don't fall on UTF-8 boundaries.ץߥ:7 If you're building an NFA from an HIR (and not using aIF [`thompson::Compiler`](crate::nfa::thompson::Compiler)), then you canGD use the [`syntax::Config::utf8`](crate::util::syntax::Config::utf8)JG option to guarantee that if the HIR detects a non-empty match, then it$! is guaranteed to be valid UTF-8.GD Note that property (2) does *not* specify the behavior of executingHE a search on a haystack that is not valid UTF-8. Therefore, if you'reѩEB *not* running this NFA on strings that are guaranteed to be validB? UTF-8, you almost certainly do not want to enable this option.JG Similarly, if you are running the NFA on strings that *are* guaranteedKH to be valid UTF-8, then you almost certainly want to enable this optionJG unless you can guarantee that your NFA will never produce a zero-widthTЬ match.߬ It is disabled by default.D+ +B $֭;8 Returns whether UTF-8 mode is enabled for this builder.JG See [`Builder::set_utf8`] for more details about what "UTF-8 mode" is.D+ +B  $(FC Sets whether the NFA produced by this builder should be matched inFC reverse or not. Generally speaking, when enabled, the NFA producedKH should be matched by moving backwards through a haystack, from a higher-* memory address to a lower memory address.ñ˱2/ See also [`NFA::is_reverse`] for more details.HE This is disabled by default, which means NFAs are by default matchedײ in the forward direction.\+ +B $!˳=: Returns whether reverse mode is enabled for this builder.KH See [`Builder::set_reverse`] for more details about what "reverse mode"< is.\+ +B  $2KH Sets the look-around matcher that should be used for the resulting NFA.B? A look-around matcher can be used to configure how look-aroundԶ>; assertions are matched. For example, a matcher might carryHE configuration that changes the line terminator used for `(?m:^)` and `(?m:$)` assertions.+ +B $ .:7 Returns the look-around matcher used for this builder.IF If a matcher was not explicitly set, then `LookMatcher::default()` islϳ/+ + +B  $g޺'$ Set the size limit on this builder.GD Setting the size limit will also check whether the NFA built so far޻EB fits within the given size limit. If it doesn't, then an error islϳ/¼2/ By default, there is no configured size limit.t+ +ͶB $,˿-/, Return the currently configured size limit.GD By default, this returns `None`, which corresponds to no configured| size limit.tҿ+ +ͶB  $#IF Returns the heap memory usage, in bytes, used by the NFA states added\ so far.HE Note that this is an approximation of how big the final NFA will be.FC In practice, the final NFA will likely be a bit smaller because ofEB its simpler state representation. (For example, using things like0- `Box<[StateID]>` instead of `Vec`.)d+ +B  $4+ +B  $dԟDY4  KC =M C$r x İL  Utf8Sequencesl l  <  T  \  ڛt  ĝl  l  T    ťL  $  \   L ! < " $ t l ,0 L8 DH dX X l X. vy, Կ  r c̚ T GJ7:'*       \   = : Return a new default Thompson NFA compiler configuration. B  .* 6 3 Whether to enable UTF-8 mode during search or not. F C A regex engine is said to be in UTF-8 mode when it guarantees that I F all matches returned by it have spans consisting of only valid UTF-8. B ? That is, it is impossible for a match span to be returned that   contains any invalid UTF-8.  0 - UTF-8 mode generally consists of two things:  H E 1. Whether the NFA's states are constructed such that all paths to a I F match state that consume at least one byte always correspond to validT  UTF-8. I F 2. Whether all paths to a match state that do _not_ consume any bytes 7 4 should always correspond to valid UTF-8 boundaries.  : 7 (1) is a guarantee made by whoever constructs the NFA.< 9 If you're parsing a regex from its concrete syntax, thenH E [`syntax::Config::utf8`](crate::util::syntax::Config::utf8) can makeI F this guarantee for you. It does it by returning an error if the regexK H pattern could every report a non-empty match span that contains invalidK H UTF-8. So long as `syntax::Config::utf8` mode is enabled and your regexJ G successfully parses, then you're guaranteed that the corresponding NFAG D will only ever report non-empty match spans containing valid UTF-8. H E (2) is a trickier guarantee because it cannot be enforced by the NFAI F state graph itself. Consider, for example, the regex `a*`. It matchesH E the empty strings in `☃` at positions `0`, `1`, `2` and `3`, whereI F positions `1` and `2` occur within the UTF-8 encoding of a codepoint,D A and thus correspond to invalid UTF-8 boundaries. Therefore, thisE B guarantee must be made at a higher level than the NFA state graphI F itself. This crate deals with this case in each regex engine. Namely,F C when a zero-width match that splits a codepoint is found and UTF-8H E mode enabled, then it is ignored and the engine moves on looking for  the next match. F C Thus, UTF-8 mode is both a promise that the NFA built only reportsI F non-empty matches that are valid UTF-8, and an *instruction* to regexF C engines that empty matches that split codepoints should be banned. K H Because UTF-8 mode is fundamentally about avoiding invalid UTF-8 spans,K H it only makes sense to enable this option when you *know* your haystackD A is valid UTF-8. (For example, a `&str`.) Enabling UTF-8 mode andK H searching a haystack that contains invalid UTF-8 leads to **unspecified| behavior**. G D Therefore, it may make sense to enable `syntax::Config::utf8` whileK H simultaneously *disabling* this option. That would ensure all non-emptyK H match spans are valid UTF-8, but that empty match spans may still splitA > a codepoint or match at other places that aren't valid UTF-8. F C In general, this mode is only relevant if your regex can match the% " empty string. Most regexes don't.   , l -  I 9!! ү9! '= : assert_eq!(Some(Match::must(0, 3..3)), caps.get_match());' 'H 9(  let re = PikeVM::builder()(6 ܭ3) ն9)- >)= : assert_eq!(Some(Match::must(0, 1..1)), caps.get_match());* * 9*- >*= : assert_eq!(Some(Match::must(0, 2..2)), caps.get_match());+ + 9+- >,= >, , 9,- >-' $ assert_eq!(None, caps.get_match());- -. -<. -$. B  $. . =- /  Reverse the NFA./ /D A A NFA reversal is performed by reversing all of the concatenated/F C sub-expressions in the original pattern, recursively. (Look around0H E operators are also inverted.) The resulting NFA can be used to match1J G the pattern starting from the end of a string instead of the beginning1 of a string.1 1I F Reversing the NFA is useful for building a reverse DFA, which is most2I F useful for finding the start of a match after its ending position has3F C been found. NFA execution engines typically do not work on reverse3K H NFAs. For example, currently, the Pike VM reports the starting location4% " of matches without a reverse NFA.4 4; 8 Currently, enabling this setting requires disabling the5I F [`captures`](Config::captures) setting. If both are enabled, then the5K H compiler will return an error. It is expected that this limitation will6  be lifted in the future.6 6 .7 l7 -7 7J G This example shows how to build a DFA from a reverse NFA, and then use7  the DFA to search backwards.8 <8 -̲8 -8  dfa::{self, Automaton},8, ) nfa::thompson::{NFA, WhichCaptures},̥9 /49 -9 9( % let dfa = dfa::dense::Builder::new():  .thompson(NFA::config():0 - .which_captures(WhichCaptures::None):  .reverse(true)L: -;  .build("baz[0-9]+")?;;/ 9|; Ќ-;  expected,<; 8 dfa.try_search_rev(&Input::new("foobaz12345bar"))?,4< č-< <. -<= -<= B  $= = K? >J G Sets an approximate size limit on the total heap used by the NFA beingl> compiled.> >I F This permits imposing constraints on the size of a compiled NFA. This?J G may be useful in contexts where the regex pattern is untrusted and one@) & wants to avoid using too much memory.@ @@ = This size limit does not apply to auxiliary heap used duringA2 / compilation that is not part of the built NFA.A AH E Note that this size limit is applied during compilation in order forBD A the limit to prevent too much heap from being used. However, theBE B implementation may use an intermediate NFA representation that isCH E otherwise slightly bigger than the final public form. Since the sizeCH E limit may be applied to an intermediate representation, there is notDJ G necessarily a precise correspondence between the configured size limitE( % and the heap usage of the final NFA.E E& # There is no size limit by default.E lE -F FG D This example demonstrates how Unicode mode can greatly increase theF  size of the NFA.F `G D [`reverse`](Config::reverse) setting. If both are enabled, then the`K >a >a a ,a la -b bH E This example demonstrates that some regex engines, like the Pike VM,bE B require capturing states to be present in the NFA to report matchdc offsets.c cF C (Note that since this method is deprecated, the example below usesd: 7 [`Config::which_captures`] to disable capture states.)d fD A .thompson(NFA::config().which_captures(WhichCaptures::None))f .f& # let mut cache = re.create_cache();g g, ) assert!(re.is_match(&mut cache, "abc"));g1 . assert_eq!(None, re.find(&mut cache, "abc"));h h. -tD ?܃u .u& ?u u, ϝ?v1 ?v v. -Ɂ  util::look::LookMatcher, >4 - ' $ let mut lookm = LookMatcher::new();ǂ' $ lookm.set_line_terminator(b'\x00');  >> ; .thompson(thompson::Config::new().look_matcher(lookm))!  .build(r"(?m)^[a-z]+$")?;& ? 9 6 // Multi-line assertions now use NUL as a terminator.| Ќ-# Some(Match::must(0, 1..4)),, ) re.find(&mut cache, b"\x00abc\x00"),4 č-: 7 // ... and \n is no longer recognized as a terminator.| Ќ-lÆ ܑ-Ն( % re.find(&mut cache, b"\nabc\n"),4 č- . -<ȇ -dۇ B  $    > ; Returns whether this configuration has enabled UTF-8 mode.D + +B    $ ! K H Returns whether this configuration has enabled reverse NFA compilation.\ + +B    ܞ$ ǎ1 ލH E Return the configured NFA size limit, if it exists, in the number of  bytes of heap used.Ύ + +ͶB    ֩$  , ) Return whether NFA shrinking is enabled.T + +B    $ ґ" K H Return whether NFA compilation is configured to produce capture states.?use get_which_captures instead G dّ ??1+ +B    $ 1 E B Return what kinds of capture states will be compiled into an NFA. , ,B    $ - 0 - Return the look-around matcher for this NFA. , ,B    $ '  , ,B  ז  ծ$ؖ 3 L , ,B    $           U   ò <Ʋ B    쒴 G D Returns true if this configuration indicates that no capture states!  should be produced in an NFA.< , ,B    $ 䱶 I F Returns true if this configuration indicates that some capture statesH E should be added to an NFA. Note that this might only include capture' $ states for implicit capture groups.4 , ,B    $ l  Đ < 9 Create a new NFA builder with its default configuration. B  = = : Compile the given regular expression pattern into an NFA. J G If there was a problem parsing the regex, then that error is returned. H E Otherwise, if there was a problem building the NFA, then an error isJ G returned. The only error that can occur is if the compiled regex wouldH E exceed the size limits configured on this builder, or if any part ofH E the NFA would exceed the integer representations used. (For example,> ; too many states might plausibly occur on a 16-bit target.) l - < -F C use regex_automata::{nfa::thompson::{NFA, pikevm::PikeVM}, Match}; ; 8 let config = NFA::config().nfa_size_limit(Some(1_000));C @ let nfa = NFA::compiler().configure(config).build(r"(?-u)\w")?; ( % let re = PikeVM::new_from_nfa(nfa)?;& ?( % let mut caps = re.create_captures();. + let expected = Some(Match::must(0, 3..4));2 / re.captures(&mut cache, "!@#A#@!", &mut caps);+ ( assert_eq!(expected, caps.get_match()); . -< -, ,, , ,B    $ < i D A Compile the given regular expression patterns into a single NFA. I /# / l - < -F ? ; ?= : let nfa = NFA::compiler().configure(config).build_many(&[  r"(?-u)\s",  r"(?-u)\w",D  ])?; ( ?& ?( ?. + let expected = Some(Match::must(1, 1..2));2 / re.captures(&mut cache, "!A! !A!", &mut caps);+ ? . -< -T ,, , ,B      /  /T $ wt!#C I F Compile the given high level intermediate representation of a regularܘ  expression into an NFA. K H If there was a problem building the NFA, then an error is returned. TheG D only error that can occur is if the compiled regex would exceed theK H size limits configured on this builder, or if any part of the NFA wouldJ G exceed the integer representations used. (For example, too many states. + might plausibly occur on a 16-bit target.) l - < -F ?E B use regex_syntax::hir::{Hir, Class, ClassBytes, ClassBytesRange}; ; 8 let hir = Hir::class(Class::Bytes(ClassBytes::new(vec![) & ClassBytesRange::new(b'0', b'9'),) & ClassBytesRange::new(b'A', b'Z'),) & ClassBytesRange::new(b'_', b'_'),À) & ClassBytesRange::new(b'a', b'z'),L  ]))); ; ?ǁF C let nfa = NFA::compiler().configure(config).build_from_hir(&hir)?; ( ?ǂ& ?( ?. ?҃2 ?+ ? . -< -t ,, , ,B    $ $ p H E Compile the given high level intermediate representations of regularȆ"  expressions into a single NFA. I /Ň# / l - < -F ?E ?   let hirs = &[Ή5 2 Hir::class(Class::Bytes(ClassBytes::new(vec![/ , ClassBytesRange::new(b'\t', b'\r'),- * ClassBytesRange::new(b' ', b' '),l ]))),5 ?- * ClassBytesRange::new(b'0', b'9'),- * ClassBytesRange::new(b'A', b'Z'),- * ClassBytesRange::new(b'_', b'_'),Ќ- * ClassBytesRange::new(b'a', b'z'),l ?4 : ; ?K H let nfa = NFA::compiler().configure(config).build_many_from_hir(hirs)?; ( ?& ?( ?ď. ?2 ?+ ?ސ . -< - ,, , ,ϜB    ّ    \Ñ $ڑ 46< > ; Apply the given NFA configuration options to this builder. l - < -+ > ; ?C ?% ? . -<Օ -L , ,?B    $ 4 ͝e 7 /4 / J /t /̘ ԘH E This syntax configuration only applies when an NFA is built directlyJ G from a pattern string. If an NFA is built from an HIR, then all syntax  settings are ignored. l - < -; 8 use regex_automata::{nfa::thompson::NFA, util::syntax}; = : let syntax_config = syntax::Config::new().unicode(false);ƛB ? let nfa = NFA::compiler().syntax(syntax_config).build(r"\w")?;J G // If Unicode were enabled, the number of states would be much bigger.ܜ% " assert!(nfa.states().len() < 15); . -< -4ԝ , ,?B    $ 4 l  I < ,, , ,ϜB      ?  ?\ *!#       Ӯ       :  ,, , ,B    $ $ ˴ ,ߴ <         U D , ,B        P; Q& $ , GIG \ ,, , ,יB    $ ,          U        ? T , ,B      ?  Q0 @& $ , ;= , ,,, ,řǙșəʙ˙G*R{9  , ,B    $ ,  $ $ g d ,, , , RepetitionUgreedyeD:WfB    $ rep  L ,, , ,B    $ $ @4 U   T ,, , ,B    $ $ @4   ,        2s l ,, , ,B    $ $ @4 i L ,, , ,B    $ $  ř         Ug d ,, , , ClassBytes%<\=FB    $ clsȟ l | ,, , , ClassUnicode%TMvyB  ¨  $è @Ѩ ,          U  # ,, , ,@B    $ @ G 4 ,, , ,B    $ anchor4 D L ,, , ,B    $ ,         ;UH < , ,B    $ ,  4 < , ,B    $ 3 4 , ,B    $ E , , ,B    $ $ 1 8 l , ,B    $ c t , ,B    $ Dƃ Ȅ2 L˄ , ,B  Մ  $ք F L , ,B  Å  $ą ,ʅ Յ c T , ,B    $ 4 A D , ,B  ň  $ƈ $Ј 2 L , ,B    $ :  , ,B    $   ,, ,řǙșəʙ˙G*R{9  ,B    $  capture_indexl  $֌ ˎc |Ύ , ,B    $ @l 1 D , ,B    $ 2 L , ,B    $  T , ,B    $ t١    B   , , ,B    $    Œ ţ y    ÜĜŜƜǜȜ_wŒB  < , 7 4 , ,ŒB    $ A  ,, ,Œ ,@B    $ 4          ̧U        ۬UA d , ,ŒB    $ $ Ӯb <֮ , ,ŒB    $ node$ ֱ. Tٱ ,, ,Œ ,@B    $ 4  L , ,ŒB    $ : T , ,ŒB    $  $ ) Dù , ,ŒB  ̹  $ѹ , | , ,ŒB    $  $ǻ l  ν0 ѽ , ,B    $  $ L,fD   $!  L8!  4', ,řǙșəʙ˙G*R{9  ,ۀ3B  ' $')& g *D*,,, , ,ggg cgcثq1ȡ,B  * $* *d     =    & C  $N   \^   Tk   7m <8 ,, , ,B  8  $8 <8 b%  g b@ b ,,, , ,ggg cgcثq1ȡ,B  c  $c  c r    r t% t   B  ,t T   L  - -B    $ ̙ <  - -B    $ ' 4  - -Κ- s  Dp-B    $ ' d  - - -B    $ ֒% ْ  - -B    $ \         U  g @  --- - -ggg cgcثq1ȡ-B    $   Л)    QQ ՛ L $   . $ - -řǙșəʙ˙G*R{9 @B    $ ş$ Κ g @  --- -Κ -ggg cgcثq1ȡ-B    $   <Ԁ  =     C  T   ,      < "  , D z,  n0 ? `bNP @ ,  | $  )  ,- , B  D, / ,/ - -B  /  $/ 4/ $4 -- - -B  4  $4 Ü4 8I 8 -- - -B  8  $8 Ü8 $9 =r = - -B  >  $> Ü> $> D> J  N, N B  DN P ,P - -B  P  $P S0 $S -- - -˝B  S  $S ÜS U ,U  |U $U  Yi Y -- - -˝B  Y  $Y ÜY $Z ^I ^ - -˝B  ^  $^ Ü^ $^ D^ < c      *   format4/   4? I  KM K  R =    \ C  D   4   T   <   $   ߯d   ¬\   L   t   $   \    <       L !  l "  T !  < "  L &  $ EH7:, T# .t7  LG DW  i c $ ('=< SR,  6 J D    wzgjWZGJ7: +DD  J4 DH /̑E  build an NFA from it.E E= : If you want a non-default configuration, then use the NFAE# [`Compiler`] with a [`Config`].F lF -F H H. + let expected = Some(Match::must(0, 0..8));H4 1 re.captures(&mut cache, b"foo12345", &mut caps);I+ ?I I. -N N. + let expected = Some(Match::must(1, 0..3));O7 4 re.captures(&mut cache, b"foo12345bar", &mut caps);O+ ?P P. -T T. + let expected = Some(Match::must(0, 0..0));U, ) re.captures(&mut cache, b"", &mut caps);U+ ?V/ , re.captures(&mut cache, b"foo", &mut caps);V+ ?V V. -c d, @d  assert!(!caps.is_match());d/ @e @e e. -t2 3t"  .build(r"[a-z]+(?-u:.)")?;tJ >u u. + let expected = Some(Match::must(0, 1..5));u7 4 re.captures(&mut cache, b"\xFFabc\xFF", &mut caps);v+ ?v v. - Returns an iterator over all pattern identifiers in this NFA.x xE B Pattern IDs are allocated in sequential order starting from zero,yH E where the order corresponds to the order of patterns provided to they"  [`NFA::new_many`] constructor.z lz -z  ; let nfa = NFA::new_many(&["[0-9]+", "[a-z]+", "[A-Z]+"])?;{8 5 let pids: Vec = nfa.patterns().collect();ԑ|  assert_eq!(pids, vec![ܰ|  PatternID::must(0),|  PatternID::must(1),|  PatternID::must(2),<} 3} }. -<} -D} - -,  /#U"-B  }  $} " ; 8 Returns the total number of regex patterns in this NFA. H E This may return zero if the NFA was constructed with no patterns. In? < this case, the NFA can never produce a match for any input. H E This is guaranteed to be no bigger than [`PatternID::LIMIT`] becausé> ; NFA construction will fail if too many patterns are added. I F It is always true that `nfa.patterns().count() == nfa.pattern_len()`. l - < -+ >ă ̃> ͇A% " assert_eq!(3, nfa.pattern_len()); !  let nfa = NFA::never_match();% " assert_eq!(0, nfa.pattern_len()); "  let nfa = NFA::always_match();% " assert_eq!(1, nfa.pattern_len()); . -< -\Ɔ - -B  ҆  $ӆ ' J G Return the state identifier of the initial anchored state of this NFA. F C The returned identifier is guaranteed to be a valid index into theF C slice returned by [`NFA::states`], and is also a valid argument to  [`NFA::state`]. l - G D This example shows a somewhat contrived example where we can easily( % predict the anchored starting state. < -ƊC @ use regex_automata::nfa::thompson::{NFA, State, WhichCaptures}; 얋 /E B .configure(NFA::config().which_captures(WhichCaptures::None))  .build("a")?;0 - let state = nfa.state(nfa.start_anchored());ь  match *state {' $ State::ByteRange { trans } => {* ' assert_eq!(b'a', trans.start);Í( % assert_eq!(b'a', trans.end);L 7. + _ => unreachable!("unexpected state"),, 0 Î. -< -t - -B    $ ) G D Return the state identifier of the initial unanchored state of thisD  NFA. Đ4 1 This is equivalent to the identifier returned byJ G [`NFA::start_anchored`] when the NFA has no unanchored starting state.̑ ԑF AF ”A A l - G D This example shows that the anchored and unanchored starting states1 . are equivalent when an anchored NFA is built. < -+ >   let nfa = NFA::new("^a")?;= : assert_eq!(nfa.start_anchored(), nfa.start_unanchored());ו ߕ. -< - - -B  Ė  $Ŗ > K H Return the state identifier of the initial anchored state for the givenӗH E pattern, or `None` if there is no pattern corresponding to the given| identifier. J G If one uses the starting state for a particular pattern, then the only@ = match that can be returned is for the corresponding pattern.Й ؙF AF ”A A d 0 I F If the pattern doesn't exist in this NFA, then this returns an error.; 8 This occurs when `pid.as_usize() >= nfa.pattern_len()`. l -Ϝ לG A1 ϟAٝ < -8 A , ) let nfa = NFA::new_many(&["^a", "^b"])?;J G // The anchored and unanchored states for the entire NFA are the same,. + // since all of the patterns are anchored.= AF C // But the anchored starting states for each pattern are distinct,E B // because these starting states can only lead to matches for the켡  // corresponding pattern.ޡ. + let anchored = Some(nfa.start_anchored());@ = assert_ne!(anchored, nfa.start_pattern(PatternID::must(0)));֢@ = assert_ne!(anchored, nfa.start_pattern(PatternID::must(1)));? < // Requesting a pattern not in the NFA will result in None:ߣ< 9 assert_eq!(None, nfa.start_pattern(PatternID::must(2))); . -<ۤ -l - -B    $ M4 t - - -B    $ * ݬ& # Get the byte classes for this NFA. E B Byte classes represent a partitioning of this NFA's alphabet intoڭH E equivalence classes. Any two bytes in the same equivalence class areI F guaranteed to never discriminate between a match or a non-match. (The% " partitioning may not be minimal.) F C Byte classes are used internally by this crate when building DFAs.G D Namely, among other optimizations, they enable a space optimizationE B where the DFA's internal alphabet is defined over the equivalenceG D classes of bytes instead of all possible byte values. The former isԱK H often quite a bit smaller than the latter, which permits the DFA to use( % less space for its transition table.Ѳ lٲ - ? < This example shows how to query the class of various bytes. < -˳+ > "  let nfa = NFA::new("[a-z]+")?;% " let classes = nfa.byte_classes();Դ8 5 // 'a' and 'z' are in the same class for this regex.5 2 assert_eq!(classes.get(b'a'), classes.get(b'z'));˵  // But 'a' and 'A' are not.5 2 assert_ne!(classes.get(b'a'), classes.get(b'A')); . -< -d - - -B    $ * ӷF C Return a reference to the NFA state corresponding to the given ID. 9 6 This is a convenience routine for `nfa.states()[id]`. d ں. K H This panics when the given identifier does not reference a valid state.չ8 5 That is, when `id.as_usize() >= nfa.states().len()`. l - C @ The anchored state for a pattern will typically correspond to aC @ capturing state for that pattern. (Although, this is not an API|Ļ guarantee!)ػ < -A > use regex_automata::{nfa::thompson::{NFA, State}, PatternID}; 캼  let nfa = NFA::new("a")?;ܼG D let state = nfa.state(nfa.start_pattern(PatternID::ZERO).unwrap()); A( % State::Capture { slot, .. } => {+ ( assert_eq!(0, slot.as_usize());L 7. ϚA,ݾ 0 . -< -,ÿ - - -B  ɿ  $ʿ -N . + Returns a slice of all states in this NFA. J G The slice returned is indexed by `StateID`. This provides a convenientH E way to access states while following transitions among those states. l - J G This demonstrates that disabling UTF-8 mode can shrink the size of theK H NFA considerably in some cases, especially when using Unicode characterd classes. < -= -+ > ' $ let nfa_unicode = NFA::new(r"\w")?;* ' let nfa_ascii = NFA::new(r"(?-u)\w")?;. + // Yes, a factor of 45 difference. No lie.H E assert!(40 * nfa_ascii.states().len() < nfa_unicode.states().len()); . -< -4 - - -B    ۝$ & 2 / Returns the capturing group info for this NFA. E B The [`GroupInfo`] provides a way to map to and from capture indexF C and capture name for each pattern. It also provides a mapping fromH E each of the capturing groups in every pattern to their corresponding6 3 slot offsets encoded in [`State::Capture`] states. G D Note that `GroupInfo` uses reference counting internally, such that( % cloning a `GroupInfo` is very cheap. l - G D This example shows how to get a list of all capture group names for̆  a particular pattern. < -8 A : 7 let nfa = NFA::new(r"(a)(?Pb)(c)(d)(?Pe)")?;G D // The first is the implicit group that is always unnamed. The nextK H // 5 groups are the explicit groups found in the concrete syntax above.J G let expected = vec![None, None, Some("foo"), None, None, Some("bar")];  let got: Vec> =B ? nfa.group_info().pattern_names(PatternID::ZERO).collect(); . B ? // Using an invalid pattern ID will result in nothing yielded.K H let got = nfa.group_info().pattern_names(PatternID::must(999)).count();  assert_eq!(0, got); . -< -T - - -B    $ ! 9 6 Returns true if and only if this NFA has at least one: 7 [`Capture`](State::Capture) in its sequence of states. E B This is useful as a way to perform a quick test before attemptingH E something that does or does not require capture states. For example,K H some regex engines (like the PikeVM) require capture states in order to work at all. l - J G This example shows a few different NFAs and whether they have captures\  or not. < -< ? $ ! // Obviously has capture states.  let nfa = NFA::new("(a)")?;  assert!(nfa.has_capture()); F C // Less obviously has capture states, because every pattern has atK H // least one anonymous capture group corresponding to the match for the  // entire pattern. A A K H // Other than hand building your own NFA, this is the only way to buildK H // an NFA without capturing groups. In general, you should only do thisH E // if you don't intend to use any of the NFA-oriented regex engines.I F // Overall, capturing groups don't have many downsides. Although theyJ G // can add a bit of noise to simple NFAs, so it can be nice to disable# // them for debugging purposes.4 ȃ7C @ // Notice that 'has_capture' is false here even when we have an- * // explicit capture group in the pattern. /E A  .build("(a)")?;  assert!(!nfa.has_capture()); . -< -\ - -B    $  D A Returns true if and only if this NFA can match the empty string.H 9 ׋9 H 9J ֌9H 9E 9G Ύ9H 9 l - F C This example shows a few different NFAs and whether they match theF 9H 9* 9 < -; ? 0 9  let nfa = NFA::new("")?;4 1 assert!(nfa.has_empty(), "empty matches empty");F ݔ9' 9  let nfa = NFA::new("a+")?;8 5 assert!(!nfa.has_empty(), "+ does not match empty");, ˖9  let nfa = NFA::new("a*")?;3 0 assert!(nfa.has_empty(), "* does match empty");J 9/ 9!  let nfa = NFA::new("(a+)*")?;< 9 assert!(nfa.has_empty(), "+ inside of * matches empty"); F 9J Ț9G 9J 9 Ĝ9 /2 3) ӝ97 4 assert!(nfa.has_empty(), "assertions match empty");G ɞ9 9  let nfa = NFA::new(r"\b+")?;@ = assert!(nfa.has_empty(), "+ of an assertion matches empty"); J 96 9+ ( let nfa = NFA::new("foo|(bar)?|quux")?;= : assert!(nfa.has_empty(), "alternations can match empty"); C ܢ9"  let nfa = NFA::new("[a&&b]")?;I F assert!(!nfa.has_empty(), "never matching means not matching empty");G 9/ 9# let nfa = NFA::new("[a&&b]*")?;E B assert!(nfa.has_empty(), "* on never-match still matches empty");I 9G 90 ۧ9# let nfa = NFA::new("[a&&b]+")?;H E assert!(!nfa.has_empty(), "+ on never-match still matches nothing"); . -< -L - -B    $ 엑 6 3 Whether UTF-8 mode is enabled for this NFA or not. F C When UTF-8 mode is enabled, all matches reported by a regex engineH E derived from this NFA are guaranteed to correspond to spans of valid̀J G UTF-8. This includes zero-width matches. For example, the regex engineG D must guarantee that the empty regex will not match at the positionsC @ between code units in the UTF-8 encoding of a single codepoint. . + See [`Config::utf8`] for more information.  , l - I 9! ү9 < - -߄. > >4 -  >J > ( 9& 9- >Ç= > D 9ֈG 9̢ ɴ9 9܉- >= >Њ ؊H 9 >ȋ6 ܭ3 ն9- >Ό= >  9- >= >  9̎- >= > ȏ 9- >' > ʐ. -< -< - -B    $  ֑A > Returns true when this NFA is meant to be matched in reverse. J G Generally speaking, when this is true, it means the NFA is supposed toK H be used in conjunction with moving backwards through the haystack. ThatÓ? < is, from a higher memory address to a lower memory address. F C It is often the case that lower level routines dealing with an NFAڔK H don't need to care about whether it is "meant" to be matched in reverseH E or not. However, there are some specific cases where it matters. ForF C example, the implementation of CRLF-aware `^` and `$` line anchors–A > needs to know whether the search is in the forward or reverseI F direction. In the forward direction, neither `^` nor `$` should match֗H E when a `\r` has been seen previously and a `\n` is next. However, inG D the reverse direction, neither `^` nor `$` should match when a `\n`K H has been seen previously and a `\r` is next. This fundamentally changesF C how the state machine is constructed, and thus needs to be altered) & based on the direction of the search. L I This is automatically set when using a [`Compiler`] with a configurationI F where [`Config::reverse`] is enabled. If you're building your own NFAߛ  by hand via a [`Builder`]T - -B    $ . ԜK H Returns true if and only if all starting states for this NFA correspond+ ( to the beginning of an anchored search.ԝ ܝK H Typically, an NFA will have both an anchored and an unanchored startingJ G state. Namely, because it tends to be useful to have both and the costG D of having an unanchored starting state is almost zero (for an NFA).ǟJ G However, if all patterns in the NFA are themselves anchored, then evenG D the unanchored starting state will correspond to an anchored search3 0 since the pattern doesn't permit anything else. l - D A This example shows a few different scenarios where this method'są  return value varies. < -+ > K H // The unanchored starting state permits matching this pattern anywhere7 4 // in a haystack, instead of just at the beginning. A- * assert!(!nfa.is_always_start_anchored());Τ ֤G D // In this case, the pattern is itself anchored, so there is no way# // to run an unanchored search.ʥ A, ) assert!(nfa.is_always_start_anchored()); K H // When multiline mode is enabled, '^' can match at the start of a lineI F // in addition to the start of a haystack, so an unanchored search isħ  // actually possible."  let nfa = NFA::new("(?m)^a")?;- B èJ G // Weird cases also work. A pattern is only considered anchored if all9 6 // matches may only occur at the start of a haystack.Щ"  let nfa = NFA::new("(^a)|a")?;- B I F // When multiple patterns are present, if they are all anchored, then& # // the NFA is always anchored too.2 / let nfa = NFA::new_many(&["^a", "^b", "^c"])?;, ŖB D A // But if one pattern is unanchored, then the NFA must permit an  // unanchored search.1 . let nfa = NFA::new_many(&["^a", "b", "^c"])?;- B . -< -Ů - -B  ޮ  $߮ * = : Returns the look-around matcher associated with this NFA. I Ѯ?˰E ?K ?/ Ȱ? @ = If the NFA was built using a [`Compiler`], then this matcher= : can be set via the [`Config::look_matcher`] configurationB ? knob. Otherwise, if you've built an NFA by hand, it is set via"  [`Builder::set_look_matcher`]. l - K ? < -̜ -. > ? >4 - ' ?' ?  >·> ?! Ҵ?& ?ָ ޸9 ?| Ќ-# ?ع, ?4 č-: ߶?|Ӻ Ќ-l ܑ-( ?4 č- . -< -d - - -B    $ % ۼH E Returns the union of all look-around assertions used throughout thisG D NFA. When the returned set is empty, it implies that the NFA has noI F look-around assertions and thus zero conditional epsilon transitions.¾ ʾB ? This is useful in some cases enabling optimizations. It is notG D unusual, for example, for optimizations to be of the form, "for anyݿH E regex with zero conditional epsilon transitions, do ..." where "..."!  is some kind of optimization. K H This isn't only helpful for optimizations either. Sometimes look-aroundI F assertions are difficult to support. For example, many of the DFAs inI F this crate don't support Unicode word boundaries or handle them usingG D heuristics. Handling that correctly typically requires some kind ofK H cheap check of whether the NFA has a Unicode word boundary in the firstT  place. l - J G This example shows how this routine varies based on the regex pattern: < -? < use regex_automata::{nfa::thompson::NFA, util::look::Look};   // No look-around at all. A+ ( assert!(nfa.look_set_any().is_empty()); H E // When multiple patterns are present, since this returns the union,E B // it will include look-around assertions that only appear in one| // pattern.6 3 let nfa = NFA::new_many(&["a", "b", "a^b", "c"])?;6 3 assert!(nfa.look_set_any().contains(Look::Start)); E B // Some groups of assertions have various shortcuts. For example:% " let nfa = NFA::new(r"(?-u:\b)")?;0 - assert!(nfa.look_set_any().contains_word());9 6 assert!(!nfa.look_set_any().contains_word_unicode());6 3 assert!(nfa.look_set_any().contains_word_ascii()); . -< -d - -B    $ , D A Returns the union of all prefix look-around assertions for everyK H pattern in this NFA. When the returned set is empty, it implies none ofH E the patterns require moving through a conditional epsilon transition5 2 before inspecting the first byte in the haystack. J G This can be useful for determining what kinds of assertions need to beG D satisfied at the beginning of a search. For example, typically DFAsH E in this crate will build a distinct starting state for each possibleF C starting configuration that might result in look-around assertionsE B being satisfied differently. However, if the set returned here isH E empty, then you know that the start state is invariant because there7 4 are no conditional epsilon transitions to consider. l - J B < -? B  ٰB A2 / assert!(nfa.look_set_prefix_any().is_empty()); H ұBE BJ G // pattern. But it will only include assertions that are in the prefixK H // of a pattern. For example, this includes '^' but not '$' even though  // '$' does appear.7 4 let nfa = NFA::new_many(&["a", "b", "^ab$", "c"])?;= : assert!(nfa.look_set_prefix_any().contains(Look::Start));< 9 assert!(!nfa.look_set_prefix_any().contains(Look::End)); . -< - - -B    $ # 4 1 Returns the memory usage, in bytes, of this NFA. D A This does **not** include the stack size used up by this NFA. To3 0 compute that, use `std::mem::size_of::()`. l - I F This example shows that large Unicode character classes can use quite  a bit of memory. < -= -+ > ' A+ ( let nfa_ascii = NFA::new(r"(?-u:\w)")?; H E assert!(10 * nfa_ascii.memory_usage() < nfa_unicode.memory_usage()); . -< -d - -B    $  <   g 8  --- - -ggg cgcثq1ȡ-B    $   T̖  ' D B  $ - T - - -B    $ 5  - -B    $ ,  T -- - -B    $ wt x al , D - -B    $  / \ - -B    $  9  - -B    $   ~ d -- - -B    $ fD         ̵̊̊ϊЊD^)cOÖ řǙșəʙ˙G*R{9 UU       B6 , -- - -B    $  old_to_newT   g 8  --- - -ggg cgcثq1ȡ-B    $   T  ϙ G D Returns true if and only if this state contains one or more epsilonܑ transitions. H E In practice, a state has no outgoing transitions (like `Match`), hasƒG D only non-epsilon transitions (like `ByteRange`) or has only epsilon  transitions (like `Union`). l -Г <ؓ - -+ ( nfa::thompson::{State, Transition},; 8 util::primitives::{PatternID, StateID, SmallIndex},4 - . + // Capture states are epsilon transitions.  let state = State::Capture {ݕ  next: StateID::ZERO,$ ! pattern_id: PatternID::ZERO,& # group_index: SmallIndex::ZERO,Җ  slot: SmallIndex::ZERO,4 -  assert!(state.is_epsilon());   // ByteRange states are not.ӗ"  let state = State::ByteRange {J G trans: Transition { start: b'a', end: b'z', next: StateID::ZERO },4ɘ -Ԙ!  assert!(!state.is_epsilon()); . -< -T֙ - -B    $  d  - -B    $ Τ& ,Ѥ  -- - -B  פ  $ܤ , ̽  g ݭ8  --- - -ggg cgcثq1ȡ-B    $           U$        U  D ? < This follows the matching transition for a particular byte. C @ The matching transition is found by looking for a matching byteF C range (there is at most one) corresponding to the position `at` in| `haystack`. 8 5 If `at >= haystack.len()`, then this returns `None`.< -- - -B    u_~i~$       a d - -B    $ $ 4        ;U7 ? B I F The matching transition is found by looking for a matching byte range; 8 (there is at most one) corresponding to the byte given.d - -B    $   D ? B E B The matching transition is found by looking for a transition thatE B doesn't correspond to `StateID::ZERO` for the byte `at` the given  position in `haystack`. 8 B< -- - -B    ׈ˇ$       ܏a d - -B    $ $ 4        B7 ? B E B? < doesn't correspond to `StateID::ZERO` for the given `byte`.d - -B    ߑ$ 1B $ - - -B    $ %  866 ̵̊̊ϊЊD^)cOFilter predicate==J/V<< 0{?ippʽp opopo2Ы8   8U BU 8QBܩ QB B8 BB4B%B  , d       |       |  9 A > Returns true if the position `at` in `haystack` falls in this  transition's range of bytes. 9 6 If `at >= haystack.len()`, then this returns `false`.< .. . .B    $ D  $        B8 F C Returns true if the given alphabet unit falls in this transition'sK H range of bytes. If the given unit is [`EOI`](alphabet::Unit::eoi), then  this returns `false`.d . .B    $ $ 4        ;B, F C Returns true if the given byte falls in this transition's range ofT  bytes.d . .B    Մ$ $   g 8  ... . .ggg cgcثq1ȡ.B    $    L ܂%    QQ  L $   ' $ . .B    $ tT <4 0  4 c   4   t4   @44   45 I  5 =    5 C  İL5   <5 "  $4 l4 ̜4 ,4 t4 D5 l5 \5 M6 ,6  7 $7  tL ߧ ̯M M ߧB  Q ,Q . .ߧB  Q  $Q Sp $S . .ߧÜĜŜƜǜȜ_w B      T  2 S  S .# .@S$ .#CClT  $T  T tvprb. 4c .. .ߧ .@B  c  $c 4c },        . .ߧU & LȘ . .ߧB  Ҙ  $ט 3 L  . .ߧB    $ 4 y t  . .ߧB    $ from_id<ϱ A, <    . .ߧB    $   C< A,ڴ < Զ ׶  . .ߧB    $   C< A, <˷ & ,  . .ߧ .֨B    $ - 2 L  . .ߧ .֨B    $ - T ֨ ) $  . .֨B    $ A, f l  . .B     /  C  .# . .#C$ xs pred$ uwy{         U ,  . .֨B    $ |  =   . .@B    D 4    z $  ... .ߧ . .@B    trie$ , 4  D  . .B    $  4  . . .@B    $ Tބ ū 3   řǙșəʙ˙G*R{9 ūB      Æ        U        CU C        ޣCC        ޣCC" 4  ūB  r1 И2 4Ә  ūB  Cژ r2 B 4  ūB  C ¥C r3 # D  . .ū . B  Ü  $Ĝ 식 ߧ g 8  ... .ߧ .ggg cgcثq1ȡ.B    $   ܟ ֨ g 8  ... .֨ .ggg cgcثq1ȡ.B    $           U  g 8  ... . .ggg cgcثq1ȡ.B    $   3 T  B C ¥C $<ڥޥϥӥťɥDԴΥ.,ҥåEC֥d 1 A collection of modules that provide APIs that are useful across many regex engines. While one should explore the sub-modules directly to get a sense of what's there, here are some highlights that tie the sub-modules to higher level use cases: * `alphabet` contains APIs that are useful if you're doing low level things with the DFAs in this crate. For example, implementing determinization or walking its state graph directly. * `captures` contains APIs for dealing with capture group matches and their mapping to "slots" used inside an NFA graph. This is also where you can find iterators over capture group names. * `escape` contains types for pretty-printing raw byte slices as strings. * `iter` contains API helpers for writing regex iterators. * `lazy` contains a no-std and no-alloc variant of `lazy_static!` and `once_cell`. * `look` contains APIs for matching and configuring look-around assertions. * `pool` provides a way to reuse mutable memory allocated in a thread safe manner. * `prefilter` provides APIs for building prefilters and using them in searches. * `primitives` are what you might use if you're doing lower level work on automata, such as walking an NFA state graph. * `syntax` provides some higher level convenience functions for interacting with the `regex-syntax` crate. * `wire` is useful if you're working with DFA serialization. $  1  This module provides APIs for dealing with the alphabets of finite state machines. There are two principal types in this module, [`ByteClasses`] and [`Unit`]. The former defines the alphabet of a finite state machine while the latter represents an element of that alphabet. To a first approximation, the alphabet of all automata in this crate is just a `u8`. Namely, every distinct byte value. All 256 of them. In practice, this can be quite wasteful when building a transition table for a DFA, since it requires storing a state identifier for each element in the alphabet. Instead, we collapse the alphabet of an automaton down into equivalence classes, where every byte in the same equivalence class never discriminates between a match or a non-match from any other byte in the same class. For example, in the regex `[a-z]+`, then you could consider it having an alphabet consisting of two equivalence classes: `a-z` and everything else. In terms of the transitions on an automaton, it doesn't actually require representing every distinct byte. Just the equivalence classes. The downside of equivalence classes is that, of course, searching a haystack deals with individual byte values. Those byte values need to be mapped to their corresponding equivalence class. This is what `ByteClasses` does. In practice, doing this for every state transition has negligible impact on modern CPUs. Moreover, it helps make more efficient use of the CPU cache by (possibly considerably) shrinking the size of the transition table. One last hiccup concerns `Unit`. Namely, because of look-around and how the DFAs in this crate work, we need to add a sentinel value to our alphabet of equivalence classes that represents the "end" of a search. We call that sentinel [`Unit::eoi`] or "end of input." Thus, a `Unit` is either an equivalence class corresponding to a set of bytes, or it is a special "end of input" sentinel. In general, you should not expect to need either of these types unless you're doing lower level shenanigans with DFAs, or even building your own DFAs. (Although, you don't have to use these types to build your own DFAs of course.) For example, if you're walking a DFA's state graph, it's probably useful to make use of [`ByteClasses`] to visit each element in the DFA's alphabet instead of just visiting every distinct `u8` value. The latter isn't necessarily wrong, but it could be potentially very wasteful. D 1L     $    $  t $  \  r. M$ =?+-L"  & "1 . Create a new haystack unit from a byte value." "I F All possible byte values are legal. However, when creating a haystack#J G unit for a specific DFA, one should be careful to only construct units$G D that are in that DFA's alphabet. Namely, one way to compact a DFA's$G D in-memory representation is to collapse its transitions to a set of%H E equivalence classes into a set of all possible byte values. If a DFA%H E uses equivalence classes instead of byte values, then the byte given&) & here should be the equivalence class.' B  $' +1 '. + Create a new "end of input" haystack unit.' 'H E The value given is the sentinel value used by this unit to represent(K H the "end of input." The value should be the total number of equivalence)F C classes in the corresponding alphabet. Its maximum value is `256`,)> ; which occurs when every byte is its own equivalence class.* d* ں.* *D A This panics when `num_byte_equiv_classes` is greater than `256`.+ B  num_byte_equiv_classes+ . -D A If this unit is not an "end of input" sentinel, then returns its.3 0 underlying byte value. Otherwise return `None`.,. B  $. 1" /J G If this unit is an "end of input" sentinel, then return the underlying0D A sentinel value that was given to [`Unit::eoi`]. Otherwise return\1  `None`.41 řǙșəʙ˙G*R{9 B  $1 4 2K H Return this unit as a `usize`, regardless of whether it is a byte value3E B or an "end of input" sentinel. In the latter case, the underlying36 3 sentinel value given to [`Unit::eoi`] is returned.D4 B  $4 6& 5K H Returns true if and only of this unit is a byte value equivalent to the6H E byte given. This always returns false when this is an "end of input"l6 sentinel.<7 B  $7 $7 7        ;ܰ8 7F C Returns true when this unit represents an "end of input" sentinel.48 B  $8 :! 8B ? Returns true when this unit corresponds to an ASCII word byte.9 9I F This always returns false when this unit represents an "end of input"l: Cd: B  $: ;  g ;@ ; ... . .ggg cgcثq1ȡ.B  ;  $;  ; F  G FJ G Creates a new set of equivalence classes where all bytes are mapped toG  the same class.,G B  H   I" HG D Creates a new set of equivalence classes where each byte belongs toH  its own equivalence class.TI B  Mg TM . .ÜĜŜƜǜȜ_wB  M  ,M Vi DV .. . .ݘB  V  ɻ$V V Y* dY . .B  Z  $Z Z* Z1 . Set the equivalence class for the given byte.Z . .B  Z  $Z \! [1 . Get the equivalence class for the given byte.\ . .B  \  $\ (]. \H E Get the equivalence class for the given haystack unit and return the]  class as a `usize`.\] . .B  ]  $] ̨` _J G Create a unit that represents the "end of input" sentinel based on the_"  number of equivalence classes.` . .B  `  $` c# aF C Return the total number of elements in the alphabet represented bybJ G these equivalence classes. Equivalently, this returns the total number܎c  of equivalence classes.dc . .B  c  $c i e@ = Returns the stride, as a base-2 exponent, required for theseģf  equivalence classes.f fH E The stride is always the smallest power of 2 that is greater than orgH E equal to the alphabet length, and the `stride2` returned here is thegK H exponent applied to `2` to get the smallest power. This is done so thathK H converting between premultiplied state IDs and indices can be done withi= : shifts alone, which is much faster than integer division. Returns an iterator over all equivalence classes in this set.$n . .  :f(.B  n  $n } oI F Returns an iterator over a sequence of representative bytes from eachp6 3 equivalence class within the range of bytes given.p pH E When the given range is unbounded on both sides, the iterator yieldsqG D exactly N items, where N is equivalent to the number of equivalenceqG D classes. Each item is an arbitrary byte drawn from each equivalenceTr  class.r rJ G This is useful when one is determinizing an NFA and the NFA's alphabetsK H hasn't been converted to equivalence classes. Picking an arbitrary bytesJ G from each equivalence class then permits a full exploration of the NFAtI F instead of using every possible byte value and thus potentially savest"  quite a lot of redundant work.u lu -u uH E This shows an example of what a complete sequence of representativesv( % might look like from a real example.v  = classes.representatives(..).collect();xE B // Note that the specific byte values yielded are not guaranteed!Ăy :ԟy  Unit::u8(b'\x00'),y  Unit::u8(b'a'),y  Unit::u8(b'{'),y  Unit::eoi(3),4z :z  assert_eq!(expected, reps);z z. - = classes.representatives(b'A'..=b'z').collect();~E D~ :  Unit::u8(b'A'), D4 : ՉD . -< -| . .ƭ    `HH.B       ƀ $ɀ $ A24 ,Ɓ < ڇD A Returns an iterator of the bytes in the given equivalence class. I F This is useful when one needs to know the actual bytes that belong toJ G an equivalence class. For example, conceptually speaking, acceleratingȉH E a DFA state occurs when a state only has a few outgoing transitions.C @ But in reality, what is required is that there are only a small݊I F number of distinct bytes that can lead to an outgoing transition. TheJ G difference is that any one transition can correspond to an equivalenceJ G class which may contains many bytes. Therefore, DFA state accelerationɌC @ considers the actual elements in each equivalence class of eachđ  outgoing transition. l -ȍ ЍA > This shows an example of how to get all of the elements in an  equivalence class. < -ŎC D " A% AF C let elements: Vec = classes.elements(Unit::u8(1)).collect();D A let expected: Vec = (b'a'..=b'z').map(Unit::u8).collect();# assert_eq!(expected, elements); . -<ݑ -D . .   &+.B    $ єC tԔ  . . A Ϻ҉}.B    $ ,    ܒ < B  ۖ%  g D  ... . .ggg cgcثq1ȡ.B    $   '   ȱ QQ  Lʠ $Ϡ   " $ . .ȱřǙșəʙ˙G*R{9 B    $ 2    QQ  L $   ҧ" $է . .DB  ڧ  $ߧ +    QQ  L $   Ӵ" $ִ / /DB  ۴  $ 0    QQ  LȺ $ͺ   * $ / /B    $       < B       , B  6 L / /B    $ ,  / < // / /B    $ % 0 d / /B    $ d   , B     &  / /B    $ $ ) 4 / /B    $ $ / D / /B    $ $ ? t / /B    $ ,          ;;, $ / /%  i'/B    $ 8 \ / /%  J:f/B    $ % D / /B    $ c T / /ÜĜŜƜǜȜ_wB    ,  <  D // / /ݘB      2  2  $ s(* < * d / /B    $   g @  /// / /ggg cgcثq1ȡ/B    $   %    QQ  L $    $ / /B    $ *    QQ  L $   5& $ / /řǙșəʙ˙G*R{9 5B    $ T        U 1  Provides types for dealing with capturing groups. Capturing groups refer to sub-patterns of regexes that some regex engines can report matching offsets for. For example, matching `[a-z]([0-9]+)` against `a789` would give `a789` as the overall match (for the implicit capturing group at index `0`) and `789` as the match for the capturing group `([0-9]+)` (an explicit capturing group at index `1`). Not all regex engines can report match offsets for capturing groups. Indeed, to a first approximation, regex engines that can report capturing group offsets tend to be quite a bit slower than regex engines that can't. This is because tracking capturing groups at search time usually requires more "power" that in turn adds overhead. Other regex implementations might call capturing groups "submatches." # Overview The main types in this module are: * [`Captures`] records the capturing group offsets found during a search. It provides convenience routines for looking up capturing group offsets by either index or name. * [`GroupInfo`] records the mapping between capturing groups and "slots," where the latter are how capturing groups are recorded during a regex search. This also keeps a mapping from capturing group name to index, and capture group index to name. A `GroupInfo` is used by `Captures` internally to provide a convenient API. It is unlikely that you'll use a `GroupInfo` directly, but for example, if you've compiled an Thompson NFA, then you can use [`thompson::NFA::group_info`](crate::nfa::thompson::NFA::group_info) to get its underlying `GroupInfo`.  1 1D 1 4 I  K K   =     C  \   \ !  L !  t "  l "  T !  , %  $$  , t L D \  .wy kmY[l>͸ K->HE Create new storage for the offsets of all matching capturing groups.??HE This routine provides the most information for matches---namely, the?KH spans of matching capturing groups---but also requires the regex search@! routines to do the most work.@@HE It is unspecified behavior to use the returned `Captures` value in aAIF search with a `GroupInfo` other than the one that is provided to thisA constructor.BlB-BBEB This example shows that all capturing groups---but only ones thatCFC participated in a match---are available to query after a match has|C been found:C[a-z]+)|(?P[A-Z]+))(?P[0-9]+)$",cc63 re.captures(&mut cache, &"aABCz"[1..], &mut caps);cDc96 assert_eq!(Some(PatternID::must(1)), caps.pattern());d?Dd'>ee.- This is a convenience routine for `caps.pattern().is_some()`.glg-hhC@ When using the PikeVM (for example), the lightest weight way ofhGD detecting whether a match exists is to create capturing groups thati96 only track the ID of the pattern that match (if any):i same pattern ID found in the `Match` returned by `get_match`.qlq-qqCDrKH detecting which pattern matched is to create capturing groups that onlyr41 track the ID of the pattern that match (if any):sxx.-.D=: assert_eq!(Some(Match::must(1, 0..3)), caps.get_match());.-<-L/ /͸ B   $5JG Returns the span of a capturing group match corresponding to the groupʂKH index given, only if both the overall pattern matched and the capturing%" group participated in that match.ẵKH This returns `None` if `index` is invalid. `index` is valid if and onlyGD if it's less than [`Captures::group_len`] for the matching pattern.?DDA [`Captures::empty`], even if a match was found. This also alwaysGD returns `None` for any `index > 0` when `Captures` was created withɆ [`Captures::matches`].BDDA guaranteed to return `true`, `pattern` is guaranteed to return aIF non-`None` value and `get_match` is guaranteed to return a non-`None`TΈ value.݈DA By convention, the 0th capture group will always return the sameEB span as the span returned by `get_match`. This is because the 0thEB capture group always corresponds to the entirety of the pattern'sŠHE match. (It is similarly always unnamed because it is implicit.) ThisEB isn't necessarily true of all regex engines. For example, one canًEB hand-compile a [`thompson::NFA`](crate::nfa::thompson::NFA) via aEB [`thompson::Builder`](crate::nfa::thompson::Builder), which isn'tKH technically forced to make the 0th capturing group always correspond to the entire match.׍lߍ-HE This example shows how to get the capturing groups, by index, from aTƎ match:Վ<ݎ-=-EB use regex_automata::{nfa::thompson::pikevm::PikeVM, Span, Match};A> let re = PikeVM::new(r"^(?P\pL+)\s+(?P\pL+)$")?;ÐJ><9 re.captures(&mut cache, "Bruce Springsteen", &mut caps);ۑ>; assert_eq!(Some(Match::must(0, 0..17)), caps.get_match());:7 assert_eq!(Some(Span::from(0..5)), caps.get_group(1));ݒ;8 assert_eq!(Some(Span::from(6..17)), caps.get_group(2));C@ // Looking for a non-existent capturing group will return None:(% assert_eq!(None, caps.get_group(3));52 # // literals are too big for 32-bit usize: #1039̔)& # #[cfg(target_pointer_width = "64")]:7 assert_eq!(None, caps.get_group(9944060567225171988));.-<-L/ /͸B   $H˩;JDJG name given, only if both the overall pattern matched and the capturing%ƀEJG This returns `None` if `name` does not correspond to a valid capturing'$ group for the pattern that matched.?DDǂEGEԌEBDDEáIETʅEl-¢GD This example shows how to get the capturing groups, by name, from aTE<-=-ĚEŤAEJ>ڥ<E>؎EHE assert_eq!(Some(Span::from(0..5)), caps.get_group_by_name("first"));HE assert_eq!(Some(Span::from(6..17)), caps.get_group_by_name("last"));CEȨ74 assert_eq!(None, caps.get_group_by_name("middle"));.-<-ҩ// /͸ /B   $ $-JG Returns an iterator of possible spans for every capturing group in theɫ matching pattern.EB If this `Captures` value does not correspond to a match, then the)& iterator returned yields no elements.KH Note that the iterator returned yields elements of type `Option`.HE A span is present if and only if it corresponds to a capturing group! that participated in a match.l-ȮЮ;8 This example shows how to collect all capturing groups:<-=->; use regex_automata::{nfa::thompson::pikevm::PikeVM, Span};̱DϰB? // Matches first/last names, with an optional middle name.FC r"^(?P\pL+)\s+(?:(?P\pL+)\s+)?(?P\pL+)$",<=J>IJ=: re.captures(&mut cache, "Harry James Potter", &mut caps);솳D:7 let groups: Vec> = caps.iter().collect(); assert_eq!(groups, vec![  Some(Span::from(0..18)), Some(Span::from(0..5)),Ѵ  Some(Span::from(6..11)),! Some(Span::from(12..18)),<3.-<-HE This example uses the same regex as the previous example, but with aĶJG haystack that omits the middle name. This results in a capturing groupIF that is present in the elements yielded by the iterator but without aTE<-=-Ƹ>ȆDBEFE<=ͺJ>74 re.captures(&mut cache, "Harry Potter", &mut caps);D:EE  Some(Span::from(0..12)),ɥElܑ-  Some(Span::from(6..12)),<3.-<-$/ /͸   ב˰8/B   $¾,տ       U IF Return the total number of capturing groups for the matching pattern.FC If this `Captures` value does not correspond to a match, then this always returns `0`.>; This always returns the same number of elements yielded byJG [`Captures::iter`]. That is, the number includes capturing groups even+( if they don't participate in the match.l-HE This example shows how to count the total number of capturing groupsJG associated with a pattern. Notice that it includes groups that did not=: participate in a match (just like `Captures::iter` does).<-=-63 use regex_automata::nfa::thompson::pikevm::PikeVM;̋DBEFE<=J>7E$! assert_eq!(4, caps.group_len());.-<-L/ /͸B   $&C@ Returns a reference to the underlying group info on which these captures are based.HE The difference between `GroupInfo` and `Captures` is that the formerIF defines the structure of capturing groups where as the latter is whatJG stores the actual match information. So where as `Captures` only givesC@ you access to the current match, `GroupInfo` lets you query anyGD information about all capturing groups, even ones for patterns that  weren't involved in a match.IF Note that a `GroupInfo` uses reference counting internally, so it may be cloned cheaply.l-DA This example shows how to get all capturing group names from theC@ underlying `GroupInfo`. Notice that we don't even need to run a\ԗ7<-C@ use regex_automata::{nfa::thompson::pikevm::PikeVM, PatternID};  let re = PikeVM::new_many(&[ r"(?Pa)", r"(a)(b)",t r"ab",! r"(?Pa)(?Pa)", r"(?Pz)",D?$! let caps = re.create_captures();ġ:&# (PatternID::must(0), 0, None),-* (PatternID::must(0), 1, Some("foo")),&# (PatternID::must(1), 0, None),&# (PatternID::must(1), 1, None),&# (PatternID::must(1), 2, None),&# (PatternID::must(2), 0, None),&# (PatternID::must(3), 0, None),-* (PatternID::must(3), 1, Some("bar")),.+ (PatternID::must(3), 2, Some("quux")),&# (PatternID::must(4), 0, None),-* (PatternID::must(4), 1, Some("foo")),4::7 // We could also just use 're.get_nfa().group_info()'.41 let got: Vec<(PatternID, usize, Option<&str>)> =0- caps.group_info().all_names().collect();..-<-T/ /͸ /B   $lA> Interpolates the capture references in `replacement` with theIF corresponding substrings in `haystack` matched by each reference. The$! interpolated string is returned.HE See the [`interpolate` module](interpolate) for documentation on the%" format of the replacement string.l-FC This example shows how to use interpolation, and also shows how it(% can work with multi-pattern regexes.<-CپE EA> r"(?[0-9]{2})-(?[0-9]{2})-(?[0-9]{4})",A> r"(?[0-9]{4})-(?[0-9]{2})-(?[0-9]{2})",D?&?(?;8 let replacement = "year=$year, month=$month, day=$day";&# // This matches the first pattern.:7 let hay = "On 14-03-2010, I became a Tennessee lamb.";,) re.captures(&mut cache, hay, &mut caps);;8 let result = caps.interpolate_string(hay, replacement);63 assert_eq!("year=2010, month=03, day=14", result);+( // And this matches the second pattern.:7 let hay = "On 2010-03-14, I became a Tennessee lamb.";,E;E6E.-<-/// /͸ / /B   $DӲ\AEIE,) interpolated string is written to `dst`.HE%El-FE(E<-CپE EAEAED?&?(?;E&E:E,E  let mut dst = String::new();=: caps.interpolate_string_into(hay, replacement, &mut dst);30 assert_eq!("year=2010, month=03, day=14", dst);+E:E,E E=E3E.-<-//// /͸ / / /B   $DӲ\d        U4        ͶUEnAEIE)& interpolated byte string is returned.HE%El-FE(E<-CپE EAEAED?&?(?<9 let replacement = b"year=$year, month=$month, day=$day";ۂ&E;8 let hay = b"On 14-03-2010, I became a Tennessee lamb.";΃,E:7 let result = caps.interpolate_bytes(hay, replacement);<9 assert_eq!(&b"year=2010, month=03, day=14"[..], result);+E;8 let hay = b"On 2010-03-14, I became a Tennessee lamb.";,E:E<E.-<-/// /͸ / /B   $DӲ\ϖۉAEIE1. interpolated byte string is written to `dst`.HE%El-ƌFE(E<ƍ-ҍCپE EǎAEAEDӏ?&?(?<E&E;E,E̥ let mut dst = vec![];Ò<9 caps.interpolate_bytes_into(hay, replacement, &mut dst);96 assert_eq!(&b"year=2010, month=03, day=14"[..], dst);“ʓ+E;E,EE<Eʕ9E.-<Ö-֖//// /͸ / / /B   $DӲ\d        U4       EEn?< This is a convenience routine for extracting the substringsIF corresponding to matching capture groups in the given `haystack`. TheKH `haystack` should be the same substring used to find the match spans inӜ this `Captures` value.JG This is identical to [`Captures::extract_bytes`], except it works withɝ `&str` instead of `&[u8]`.dں.A> This panics if the number of explicit matching groups in thisӞJG `Captures` value is less than `N`. This also panics if this `Captures`)& value does not correspond to a match.П؟FC Note that this does *not* panic if the number of explicit matchingHE groups is bigger than `N`. In that case, only the first `N` matching groups are extracted.l-<-6ôE?< let re = PikeVM::new(r"([0-9]{4})-([0-9]{2})-([0-9]{2})")?;â&?(?:E,E쓤D74 let (full, [year, month, day]) = caps.extract(hay);# assert_eq!("2010-03-14", full);왥 assert_eq!("2010", year);以 assert_eq!("03", month);ܥ assert_eq!("14", day);96 // We can also ask for fewer than all capture groups.+( let (full, [year]) = caps.extract(hay);#E왧Eç.-<-< / /͸  EEB    t $57%       Ӹw?EIEǬKEԗEDA This is identical to [`Captures::extract`], except it works with `&[u8]` instead of `&str`.dں.îˮAEJE)EFEHE̮E̱lԱ-<-6ôE?E&?(?ٳ;E,EҴD=: let (full, [year, month, day]) = caps.extract_bytes(hay);$! assert_eq!(b"2010-03-14", full);ߵ assert_eq!(b"2010", year);삶 assert_eq!(b"03", month);ܤ assert_eq!(b"14", day);Ķ̶9E1. let (full, [year]) = caps.extract_bytes(hay);$FF.-<Ǹ-lڸ / /͸  FFB    Et $/1%       lNK Lower level "slot" oriented APIs. One does not typically need to use theseܼLI when executing a search. They are instead mostly intended for folks thatJG are writing their own regex engine while reusing this `Captures` type.͸   Clear this `Captures` value.IF After clearing, all slots inside this `Captures` value will be set toGD `None`. Similarly, any pattern ID that it was previously associatedϿ! with (for a match) is erased.JG It is not usually necessary to call this routine. Namely, a `Captures`HE value only provides high level access to the capturing groups of theHE pattern that matched, and only low level access to individual slots.FC Thus, even if slots corresponding to groups that aren't associatedFC with the matching pattern are set, then it won't impact the higherKH level APIs. Namely, higher level APIs like [`Captures::get_group`] willJG return `None` if no pattern ID is present, even if there are spans set in the underlying slots.FC Thus, to "clear" a `Captures` value of a match, it is usually only<9 necessary to call [`Captures::set_pattern`] with `None`.l-GD This example shows what happens when a `Captures` value is cleared.<-=-6ôEAEJ><ED# let slots: Vec> =B? caps.slots().iter().map(|s| s.map(|x| x.get())).collect();GD // Note that the following ordering is considered an API guarantee.ܒ assert_eq!(slots, vec![ Some(0), Some(17),F Some(5), Some(6),ؔF<3KH // Now clear the slots. Everything is gone and it is no longer a match. caps.clear();@#͒FBFܾFlܑ-lܑ-lܑ-lܑ-lܑ-lܑ-<3.-<-,/ /͸B   $5-* Set the pattern on this `Captures` value.FC When the pattern ID is `None`, then this `Captures` value does notIF correspond to a match (`is_match` will return `false`). Otherwise, it܌ corresponds to a match.DA This is useful in search implementations where you might want toDA initially call `set_pattern(None)` in order to avoid the cost of:7 calling `clear()` if it turns out to not be necessary.l-KH This example shows that `set_pattern` merely overwrites the pattern ID.;8 It does not actually change the underlying slot values.<-=-6ôEAEJ><ED&# assert!(caps.pattern().is_some());#͒FBFGȓFܶFFؔFFFFؔF<3EB // Now set the pattern to None. Note that the slot values remain.ܴ caps.set_pattern(None);@'$ assert!(!caps.pattern().is_some());#͒FBFGȓFFFؔFFFFؔF<3.-<-\/ /͸B   $-IF Returns the underlying slots, where each slot stores a single offset.JG Every matching capturing group generally corresponds to two slots: oneGD slot for the starting position and another for the ending position.GD Typically, either both are present or neither are. (The weasel wordJG "typically" is used here because it really depends on the regex engineFC implementation. Every sensible regex engine likely adheres to thisA> invariant, and every regex engine in this crate is sensible.)JG Generally speaking, callers should prefer to use higher level routines<9 like [`Captures::get_match`] or [`Captures::get_group`].JG An important note here is that a regex engine may not reset all of theHE slots to `None` values when no match occurs, or even when a match ofHE a different pattern occurs. But this depends on how the regex engine$! implementation deals with slots.l-JG This example shows how to get the underlying slots from a regex match.<-̎-&ӀA30 util::primitives::{PatternID, NonMaxUsize},4- E r"[a-z]+",:D?J>.+ re.captures(&mut cache, "123", &mut caps);9DFC // Note that the only guarantee we have here is that slots 2 and 3IF // are set to correct values. The contents of the first two slots are74 // unspecified since the 0th pattern did not match. let expected = &[lܑ-lܑ- NonMaxUsize::new(0), NonMaxUsize::new(3),4:'$ assert_eq!(expected, caps.slots());.-<-,/ /͸ /B   $9KH Returns the underlying slots as a mutable slice, where each slot stores a single offset.EB This tends to be most useful for regex engine implementations for;8 writing offsets for matching capturing groups to slots.=: See [`Captures::slots`] for more information about slots.L/ /͸ /B   $"͸ g @/// /͸ /ggg cgcثq1ȡ/B   $    tQŲA t   $   ͸2   g @/// / /ggg cgcثq1ȡ/B   $ b$   3yJ- řǙșəʙ˙G*R{9 B$ FB  ,   |   F%   g @/// / /ggg cgcثq1ȡ/B   $ ԅ-   QQQQ مL$  *$/ /řǙșəʙ˙G*R{9 B   $-Lć/ /ͶB  · $χ,B  $و6  ވ>  t DA Creates a new group info from a sequence of patterns, where eachGD sequence of patterns yields a sequence of possible group names. TheIF index of each pattern in the sequence corresponds to its `PatternID`,IF and the index of each group in each pattern's sequence corresponds to" its corresponding group index.HE While this constructor is very generic and therefore perhaps hard toFC chew on, an example of a valid concrete type that can be passed toC@ this constructor is `Vec>>`. The outer `Vec`DA corresponds to the patterns, i.e., one `Vec>` perDA pattern. The inner `Vec` corresponds to the capturing groups forEB each pattern. The `Option` corresponds to the name of the  capturing group, if present.FC It is legal to pass an empty iterator to this constructor. It willFC return an empty group info with zero slots. An empty group info isHE useful for cases where you have no patterns or for cases where slotsA> aren't being used at all (e.g., for most DFAs in this crate).d0GD This constructor returns an error if the given capturing groups areGD invalid in some way. Those reasons include, but are not necessarily| limited to:;8 * Too many patterns (i.e., `PatternID` would overflow).=: * Too many capturing groups (e.g., `u32` would overflow).IF * A pattern is given that has no capturing groups. (All patterns must<9 have at least an implicit capturing group at index `0`.)FC * The capturing group at index `0` has a name. It must be unnamed.HE * There are duplicate capturing group names within the same pattern.EB (Multiple capturing groups with the same name may exist, but they# must be in different patterns.)A> An example below shows how to trigger some of the above error| conditions.l-FC This example shows how to build a new `GroupInfo` and query it for7<-2/ use regex_automata::util::captures::GroupInfo;# let info = GroupInfo::new(vec![  vec![None, Some("foo")], vec![None],2/ vec![None, None, None, Some("bar"), None],&# vec![None, None, Some("foo")],D?,) // The number of patterns being tracked.&# assert_eq!(4, info.pattern_len());ć // 2 slots per group$! assert_eq!(22, info.slot_len());.-<-  # Example: empty `GroupInfo`FF7<-2F" let info = GroupInfo::empty();ԣ // Everything is zero.&# assert_eq!(0, info.pattern_len());# assert_eq!(0, info.slot_len());.-<- # Example: error conditionsHE This example shows how to provoke some of the ways in which building a `GroupInfo` can fail.<-2FIF // Either the group info is empty, or all patterns must have at leastܞ // one capturing group.  assert!(GroupInfo::new(vec![$! vec![None, Some("a")], // oǩ vec![None], // ok̪ vec![], // not ok ]).is_err());41 // Note that building an empty group info is OK.GD assert!(GroupInfo::new(Vec::>>::new()).is_ok());EB // The first group in each pattern must correspond to an implicitHE // anonymous group. i.e., One that is not named. By convention, thisKH // group corresponds to the overall match of a regex. Every other group-* // in a pattern is explicit and optional.>; assert!(GroupInfo::new(vec![vec![Some("foo")]]).is_err());GD // There must not be duplicate group names within the same pattern. F-* vec![None, Some("foo"), Some("foo")],F:7 // But duplicate names across distinct patterns is OK. F F F ]).is_ok());.-<-C@ There are other ways for building a `GroupInfo` to fail but areIF difficult to show. For example, if the number of patterns given would̰ overflow `PatternID`.B        /   PPDPPT 5      ݀&# This creates an empty `GroupInfo`.FC This is a convenience routine for calling `GroupInfo::new` with an%" iterator that yields no elements.l-HE This example shows how to build a new empty `GroupInfo` and query it for information.<-2F"FF&F(% assert_eq!(0, info.all_group_len());#F.-<р-,B  CIF Return the capture group index corresponding to the given name in the͂DA given pattern. If no such capture group name exists in the given&# pattern, then this returns `None`.ɃA> If the given pattern ID is invalid, then this returns `None`.GD This also returns `None` for all inputs if these captures are emptyHE (e.g., built from an empty [`GroupInfo`]). To check whether capturesEB are present for a specific pattern, use [`GroupInfo::group_len`].l-JG This example shows how to find the capture index for the given pattern and group name.HE Remember that capture indices are relative to the pattern, such that؇HE the same capture index value may refer to different capturing groupsԥ for distinct patterns.Ĉ<̈-؈=-8A׉߉@= let (pid0, pid1) = (PatternID::must(0), PatternID::must(1)); let nfa = NFA::new_many(&[ϊ'$ r"a(?P\w+)z(?P\s+)", r"a(?P\d+)z",D?" let groups = nfa.group_info();΋63 assert_eq!(Some(2), groups.to_index(pid0, "foo"));FC // Recall that capture index 0 is always unnamed and refers to theԌJG // entire pattern. So the first capturing group present in the pattern'$ // itself always starts at index 1.ύ63 assert_eq!(Some(1), groups.to_index(pid1, "foo"));EB // And if a name does not exist for a particular pattern, None is܎ // returned.52 assert!(groups.to_index(pid0, "quux").is_some());52 assert!(groups.to_index(pid1, "quux").is_none());.-<-D// / /ͶB  ʐ $ːQA \A      IIF Return the capture name for the given index and given pattern. If theӒGD corresponding group does not have a name, then this returns `None`.;8 If the pattern ID is invalid, then this returns `None`.JG If the group index is invalid for the given pattern, then this returnsKH `None`. A group `index` is valid for a pattern `pid` in an `nfa` if and30 only if `index < nfa.pattern_capture_len(pid)`.ƕΕGFHFEFl-˗ӗGD This example shows how to find the capture group name for the given䟘 pattern and group index.<Ș-Ԙ=-8Aәۙ@FF˚%" r"a(?P\w+)z(\s+)x(\d+)", r"a(\d+)z(?P\s+)",D?"F͛.+ assert_eq!(None, groups.to_name(pid0, 0));52 assert_eq!(Some("foo"), groups.to_name(pid0, 1));.+ assert_eq!(None, groups.to_name(pid0, 2));.+ assert_eq!(None, groups.to_name(pid0, 3));.+ assert_eq!(None, groups.to_name(pid1, 0));۝.+ assert_eq!(None, groups.to_name(pid1, 1));52 assert_eq!(Some("foo"), groups.to_name(pid1, 2));Ȟ?< // '3' is not a valid capture index for the second pattern..+ assert_eq!(None, groups.to_name(pid1, 3));ǟ.-<-</ /řǙșəʙ˙G*R{9  /B   $GӎGHIF Return an iterator of all capture groups and their names (if present)쳢 for a particular pattern.բݢGD If the given pattern ID is invalid or if this `GroupInfo` is empty,)& then the iterator yields no elements.ףߣFC The number of elements yielded by this iterator is always equal to@= the result of calling [`GroupInfo::group_len`] with the same9l-GAA<-8A:AGAKAĨJA ABA.BAKAAݫ.-<-l/ /ŷ, Z]/B  Ǭ $ȬHLL      ʼ0JG Return an iterator of all capture groups for all patterns supported byïJG this `GroupInfo`. Each item yielded is a triple of the group's pattern>; ID, index in the pattern and the group's name, if present.հlݰ-GD This example shows how to get a list of all capture groups found inñ41 one NFA, potentially spanning multiple patterns.<-8AͲղF޿El r"a",| r"(a)",D?Ƴ:&E-E&E&E&# (PatternID::must(2), 1, None),4:̵4E/, nfa.group_info().all_names().collect();.ܶ.-<-GD Unlike other capturing group related routines, this routine doesn't63 panic even if captures aren't enabled on this NFA:<-Ƹ<?쏹/EA8Ԗ r"(?Pa)", r"a",˺ r"(a)",d8?< // When captures aren't enabled, there's nothing to return.85 assert_eq!(0, nfa.group_info().all_names().count());.-<-LѼ/ /x Ÿ Ѹ   ?jU/B  ۼ $ܼpC@ Returns the starting and ending slot corresponding to the givenHE capturing group for the given pattern. The ending slot is always oneп)& more than the starting slot returned.KH Note that this is like [`GroupInfo::slot`], except that it also returns*' the ending slot value for convenience.GD If either the pattern ID or the capture index is invalid, then this returns None.l-FC This example shows that the starting slots for the first capturing'$ group of each pattern are distinct.<-8A*' let nfa = NFA::new_many(&["a", "b"])?;| assert_ne!(63 nfa.group_info().slots(PatternID::must(0), 0),63 nfa.group_info().slots(PatternID::must(1), 0),4č-@= // Also, the start and end slot values are never equivalent.KH let (start, end) = nfa.group_info().slots(PatternID::ZERO, 0).unwrap(); assert_ne!(start, end);.-<-,/ /B   $QӎQ<      GHE Returns the starting slot corresponding to the given capturing groupFC for the given pattern. The ending slot is always one more than the value returned.GҩGGl-FG'G<-8A*G|լG52 nfa.group_info().slot(PatternID::must(0), 0),52 nfa.group_info().slot(PatternID::must(1), 0),4č-.-<-$/ /ͶB   ˡ$UӎU"=: Returns the total number of patterns in this `GroupInfo`.C@ This may return zero if the `GroupInfo` was constructed with nol patterns.HAFC `GroupInfo` construction will fail if too many patterns are added.l-<-+>>͇A2/ assert_eq!(3, nfa.group_info().pattern_len());!ǐA2/ assert_eq!(0, nfa.group_info().pattern_len());"A2/ assert_eq!(1, nfa.group_info().pattern_len());.-<-\/ /B   $052 Return the number of capture groups in a pattern.85 If the pattern ID is invalid, then this returns `0`.l-GD This example shows how the values returned by this routine may vary2/ for different patterns and NFA configurations.<-IF use regex_automata::{nfa::thompson::{NFA, WhichCaptures}, PatternID};&# let nfa = NFA::new(r"(a)(b)(c)")?;GD // There are 3 explicit groups in the pattern's concrete syntax and@= // 1 unnamed and implicit group spanning the entire pattern.?< assert_eq!(4, nfa.group_info().group_len(PatternID::ZERO));  let nfa = NFA::new(r"abc")?;0- // There is just the unnamed implicit group.?< assert_eq!(1, nfa.group_info().group_len(PatternID::ZERO));/EAĂ .build(r"abc")?;74 // We disabled capturing groups, so there are none.?< assert_eq!(0, nfa.group_info().group_len(PatternID::ZERO));/EA .build(r"(a)(b)(c)")?;IF // We disabled capturing groups, so there are none, even if there are.+ // explicit groups in the concrete syntax.?G.-<-L/ /B   ڥ$X$B? Return the total number of capture groups across all patterns.FC This includes implicit groups that represent the entire match of ad.l-GG2G<-IG&GGG@G41 assert_eq!(4, nfa.group_info().all_group_len()); G0G41 assert_eq!(1, nfa.group_info().all_group_len());30 let nfa = NFA::new_many(&["(a)", "b", "(c)"])?;41 // Each pattern has one implicit groups, and two-* // patterns have one explicit group each.41 assert_eq!(5, nfa.group_info().all_group_len());/EAG7G41 assert_eq!(0, nfa.group_info().all_group_len());/EAGIG.G?G.-<-l/ /B   $܀DA Returns the total number of slots in this `GroupInfo` across alllGKH The total number of slots is always twice the total number of capturing85 groups, including both implicit and explicit groups.l-GD This example shows the relationship between the number of capturing groups and slots.<-2F&# // There are 11 total groups here.#F FF2F&FD?0- // 2 slots per group gives us 11*2=22 slots.$F.-<€-D/ /B   $ʊ(DA Returns the total number of slots for implicit capturing groups.GD This is like [`GroupInfo::slot_len`], except it doesn't include theтEB explicit slots for each pattern. Since there are always exactly 2KH implicit slots for each pattern, the number of implicit slots is always*' equal to twice the number of patterns.l-GG.+ groups, implicit slots and explicit slots.<Å-υ2F&GKH let info = GroupInfo::new(vec![vec![None, Some("foo"), Some("bar")]])?;0G# assert_eq!(6, info.slot_len());IF // 2 implicit slots per pattern gives us 2 implicit slots since there // is 1 pattern.͈,) assert_eq!(2, info.implicit_slot_len());A> // 2 explicit capturing groups gives us 2*2=4 explicit slots.ĉ,) assert_eq!(4, info.explicit_slot_len());.-<-ъ/ /B   $Ε(DA Returns the total number of slots for explicit capturing groups.GGKH implicit slots for each pattern. (There are always 2 implicit slots for each pattern.)IF For a non-empty `GroupInfo`, it is always the case that `slot_len` isHE strictly greater than `explicit_slot_len`. For an empty `GroupInfo`,ƎFC both the total number of slots and the number of explicit slots isD `0`.l-GG.G<ǐ-Ӑ2F&GKG0G’#GIGGѓ,GAGȔ,G.-<-Օ/ /B   $#Ŗ<9 Returns the memory usage, in bytes, of this `GroupInfo`.IF This does **not** include the stack size used up by this `GroupInfo`.ܗ<9 To compute that, use `std::mem::size_of::()`.d/ /B   $ݘ |tš Υ ɩ-|̩ / /B  ܩ $ / /B      T $uu&vDF      = / /B   $)       \ / /B   گ$,L / /B   $&t / /B   $\       U ; B  \H| B  <<7t B  <> B  <>L / /B   < $)|  >4/ /řǙșəʙ˙G*R{9  /ۀ3B   $* g D000 0 0ggg cgcثq1ȡ0B   $ #ŷ, Z] ,, GB  /   QQQQ L$  Fʉ-$͉0 0řǙșəʙ˙G*R{9 FB  ҉ $׉  BU-L0 0FB   $,B  $ċ8ɋ@+QQLΐ$ӐFA$0 0řǙșəʙ˙G*R{9 ɀHB  $t 1! Provides convenience routines for escaping raw bytes. Since this crate tends to deal with `&[u8]` everywhere and the default `Debug` implementation just shows decimal integers, it makes debugging those representations quite difficult. This module provides types that show `&[u8]` as if it were a string, with invalid UTF-8 escaped into its byte-by-byte hex representation. 4 1f$!!#!g@!!000 0 0ggg cgcثq1ȡ0B ! $! ! ! =!: Provides a convenient `Debug` implementation for `&[u8]`. ! M!J This generally works best when the bytes are presumed to be mostly UTF-8, I!F but will work for anything. For any bytes that aren't UTF-8, they are $!! emitted as hex escape sequences.l!|?1BF1B !d!1/!g!@!!000 0 0ggg cgcثq1ȡ0B ! $! ! 1 Provides routines for interpolating capture group references. That is, if a replacement string contains references like `$foo` or `${foo1}`, then they are replaced with the corresponding capture values for the groups named `foo` and `foo1`, respectively. Similarly, syntax like `$1` and `${1}` is supported as well, with `1` corresponding to a capture group index and not a name. This module provides the free functions [`string`] and [`bytes`], which interpolate Rust Unicode strings and byte strings, respectively. # Format These routines support two different kinds of capture references: unbraced and braced. For the unbraced format, the format supported is `$ref` where `name` can be any character in the class `[0-9A-Za-z_]`. `ref` is always the longest possible parse. So for example, `$1a` corresponds to the capture group named `1a` and not the capture group at index `1`. If `ref` matches `^[0-9]+$`, then it is treated as a capture group index itself and not a name. For the braced format, the format supported is `${ref}` where `ref` can be any sequence of bytes except for `}`. If no closing brace occurs, then it is not considered a capture reference. As with the unbraced format, if `ref` matches `^[0-9]+$`, then it is treated as a capture group index and not a name. The braced format is useful for exerting precise control over the name of the capture reference. For example, `${1}a` corresponds to the capture group reference `1` followed by the letter `a`, where as `$1a` (as mentioned above) corresponds to the capture group reference `1a`. The braced format is also useful for expressing capture group names that use characters not supported by the unbraced format. For example, `${foo[bar].baz}` refers to the capture group named `foo[bar].baz`. If a capture group reference is found and it does not refer to a valid capture group, then it will be replaced with the empty string. To write a literal `$`, use `$$`. To be clear, and as exhibited via the type signatures in the routines in this module, it is impossible for a replacement string to be invalid. A replacement string may not have the intended semantics, but the interpolation procedure itself can never fail.  1̨ 1\ 14IC4 ,tHD9*OL Accepts a replacement string and interpolates capture references with their̽ corresponding values.LI `append` should be a function that appends the string value of a captureIF group at a particular index to the string given. If the capture group63 index is invalid, then nothing should be appended.LI `name_to_index` should be a function that maps a capture group name to aLI capture group index. If the given name doesn't exist, then `None` should be returned.NK Finally, `dst` is where the final interpolated contents should be written.MJ If `replacement` contains no capture group references, then `dst` will be  equivalent to `replacement`.EB See the [module documentation](self) for details about the formatt supported.l-<-*' use regex_automata::util::interpolate; E interpolate::string( "foo $bar baz", |index, dst| {ܩ if index == 0 {$! dst.push_str("BAR");l7T= |name| { if name == "bar" { Some(0) } else { Nonel7T= &mut dst,4č-# assert_eq!("foo BAR baz", dst);<-400 0 0B !0# 0̒0#H0# 00#HͶl Ӳ\4l$/%NK Accepts a replacement byte string and interpolates capture references with& their corresponding values.&&IF `append` should be a function that appends the byte string value of a'HE capture group at a particular index to the byte string given. If the'DA capture group index is invalid, then nothing should be appended.((LH(LH)H))NݢH*MH* H++E¤Ht+H+l+-,<,-,*ڥH,,E, interpolate::bytes(, b"foo $bar baz",-Hܦ-H-.+ dst.extend_from_slice(b"BAR");l-7T-=.H.H.H.ӨH.Hl.7T/=/H4/č-/)& assert_eq!(&b"foo BAR baz"[..], dst);$>B  >>!>!$>B  ?B=dB0 0řǙșəʙ˙G*R{9 йx5{:0B CӲ\CLJL0 0řǙșəʙ˙G*R{9 йx5{:0B L ˆ@L L$M$NS%SB  Td 1" Generic helpers for iteration of matches from a regex engine in a haystack. The principle type in this module is a [`Searcher`]. A `Searcher` provides its own lower level iterator-like API in addition to methods for constructing types that implement `Iterator`. The documentation for `Searcher` explains a bit more about why these different APIs exist. Currently, this module supports iteration over any regex engine that works with the [`HalfMatch`], [`Match`] or [`Captures`] types. $ 1D"L"%,"$,"%T"%""""qs`bNP<>;"2;">,";;"8 Create a new fallible non-overlapping matches iterator.<" " such as a prefilter scanner.>"B 2,>"A+"?6"3 Returns the current `Input` used by this searcher.?"?D"A The `Input` returned is generally equivalent to the one given to@I"F [`Searcher::new`], but its start position may be different to reflectA0"- the start of the next search to be executed.,A"  BA" 2$A"Y"BJ"G Return the next half match for an infallible search if one exists, andB!" advance to the next position.C"CE"B This is like `try_advance_half`, except errors are converted into\C"6C"dC"ں.C"DK"H If the given closure returns an error, then this panics. This is usefulDI"F when you know your underlying regex engine has been configured to notE" return an error.E"lE"-E"EJ"G This example shows how to use a `Searcher` to iterate over all matchesF9"6 when using a DFA, which only provides "half" matches.F"4c"-c"c8"5 let re = Regex::new(r"[0-9]{4}-[0-9]{2}-[0-9]{2}")?;c&"?d"d?"Hd&"He"e/", let expected = Some(Match::must(0, 0..10));eC"@ let got = it.advance(|input| re.try_search(&mut cache, input));f".f"f0"- let expected = Some(Match::must(0, 11..21));fC"Hg".g"g0"- let expected = Some(Match::must(0, 22..32));hC"Hh".h"h":iC"Hi".j"j."-4m"-m"m9"6 let re = PikeVM::new(r"[0-9]{4}-[0-9]{2}-[0-9]{2}")?;nJ">n"n?"Ho&"Ho"o/"Hp"" let got = it.advance(|input| {p0"- re.search(&mut cache, input, &mut caps);p" Ok(caps.get_match()) This example shows how to use a `Searcher` to create a proper" iterator over half matches."<ƍ"-ҍ"-"H쎎"H̰"/4Ύ"-َ"6"H&"?Ǐ"Ϗ?"HF"C let mut it = Searcher::new(input).into_half_matches_iter(|input| {ސ,") re.try_search_fwd(&mut cache, input)<"1"4"1 let expected = Some(Ok(HalfMatch::must(0, 10)));ܑ$"! assert_eq!(expected, it.next());"4"1 let expected = Some(Ok(HalfMatch::must(0, 21)));ƒ$"˃I"4"1 let expected = Some(Ok(HalfMatch::must(0, 32)));$"˃Iٓ"":$"˃I"."-<"-"BC "00# 0$$$$$$*00:"00#݇I %" 2DO"J"H×+"( all successive non-overlapping matches."G"HǘC"HA"> [`TryMatchesIter::infallible`] to convert errors into panics.ՙ"lݙ"-"A"HԽ" iterator over matches.ܚ"<"-"-쎛"H찛"Hқ">4"-"8"H&"?"?"HA"> let mut it = Searcher::new(input).into_matches_iter(|input| {("% re.try_search(&mut cache, input)<"1"3"0 let expected = Some(Ok(Match::must(0, 0..10)));$"˃I"4"1 let expected = Some(Ok(Match::must(0, 11..21)));ܟ$"˃I"4"1 let expected = Some(Ok(Match::must(0, 22..32)));Ơ$"˃I"":$"˃I"š."-<"-"BC "00# 0$$$$$$*006"00#I !" 2"J"H5"2 all successive non-overlapping `Captures` values."G"HC"HB"? [`TryCapturesIter::infallible`] to convert errors into panics.Φ"֦C"@ Unlike the other iterator constructors, this accepts an initialJ"G `Captures` value. This `Captures` value is reused for each search, andI"F the iterator implementation clones it before returning it. The callerF"C must provide this value because the iterator is purposely ignorantF"C of the underlying regex engine and thus doesn't know how to createѩF"C one itself. More to the point, a `Captures` value itself has a fewE"B different constructors, which change which kind of information is:"7 available to query in exchange for search performance."l"-"ǫJ"G This example shows how to use a `Searcher` to create a proper iteratorH"E over `Captures` values, which provides access to all capturing group" spans for each match."<"-̕"-&"ӀAޭ"Ht" Input,4"-"̦"DĮ;"8 r"(?P[0-9]{4})-(?P[0-9]{2})-(?P[0-9]{2})",<"=F"C let (mut cache, caps) = (re.create_cache(), re.create_captures());ۯ"6"3 let haystack = "2010-03-14 2016-10-08 2020-10-22";%"-Ȱ%"" let mut it = Searcher::new(input)1". .into_captures_iter(caps, |input, caps| {/", re.search(&mut cache, input, caps);ܱ" Ok(())\" });"."+ let got = it.next().expect("first date")?;?"< let year = got.get_group_by_name("y").expect("must match");("% assert_eq!("2010", &haystack[year]);"/", let got = it.next().expect("second date")?;@"= let month = got.get_group_by_name("m").expect("must match");'"$ assert_eq!("10", &haystack[month]);ܴ"."+ let got = it.next().expect("third date")?;>"; let day = got.get_group_by_name("d").expect("must match");ڵ%"" assert_eq!("22", &haystack[day]);"!" assert!(it.next().is_none());"."-<"-"""͸BC "000# 0$$$$$$*00 0͸:"000#ӦIո" 2"#"0 0 B ̼"C "00# 0$$$$$$*00Ƚ:"00#I ݽ%" 2$Ѽ"4"""0 0 B "C "00# 0$$$$$$*006"00#I !" 2$"4"$"L"I An iterator over all non-overlapping half matches for a fallible search."M"J The iterator yields a `Result` value until no more" matches could be found."'"$ The type parameters are as follows:"D"A * `F` represents the type of a closure that executes the search."9"6 The lifetime parameters come from the [`Input`] type:"6"3 * `'h` is the lifetime of the underlying haystack."J"G When possible, prefer the iterators defined on the regex engine you'reN"K using. This tries to abstract over the regex engine and is thus a bit more" unwieldy to use."E"B This iterator is created by [`Searcher::into_half_matches_iter`]."2C "bRrqS"(*""L"4"%"2C ""1"2"/ Return an infallible version of this iterator."J"G Any item yielded that corresponds to an error results in a panic. ThisI"F is useful if your underlying regex engine is configured in a way that."+ it is guaranteed to never return an error.T"B 2X+"6"3 Returns the current `Input` used by this iterator."C"@ The `Input` returned is generally equivalent to the one used toG"D construct this iterator, but its start position may be different to8"5 reflect the start of the next search to be executed.,"  B" 2$"!#{"200# 0$$$$$$*00 %"00#I:"C "QQ""$L"$";"$"0 0ݬB " 2$":"2C "g"D""000 0 0ggg cgcثq1ȡ0B " 2$"""2C ""+"6"I"C"IG"ŹI8"I,"  B" 2$"!#x"200# 0$$$$$$*00 %"00#I:"C "QQ""$L"$"'"$"0 0 B " 2$" "G"D An iterator over all non-overlapping matches for a fallible search."I"F The iterator yields a `Result` value until no moreܓ"I"'"ήI"D"I"9"I"6"I"J"IN"ݱI"I"@"= This iterator is created by [`Searcher::into_matches_iter`].t"2C "> x"(*""L"4"!"2C ""-"2"I"J"II"̶I."IT"B 2E+"6"I"C"IG"ŹI8"I,"  B" 2$"!#s"200# 0$$$$$$*00 !"00#I6"C "QQ""$L"$"7"$"0 0B " 2$"6"2C "g"D""000 0 0ggg cgcثq1ȡ0B " 2$"h"2C ""Ԇ+"6"I"C"I˅G"ŹI8"I,ۆ"  B" 2$"!#p"200# 0$$$$$$*00 !"00#Iއ6"C "QQ""$L"$"ˆ#"$ň"0 0 B ʈ" 2$ψ"!"H"E An iterator over all non-overlapping captures for a fallible search."L"I The iterator yields a `Result` value until no more܀"I"'"ήIȌ"̌D"I"9"Iύ"Ӎ6"I"J"IَN"ݱI"I"A"> This iterator is created by [`Searcher::into_captures_iter`]."̃"|"2C ":;,"*,Ő"Ő"tې"$ې"͸L"4"""""2C ""ϓ."2"I"J"IΒI"̶I."IT֓"B 2!x""̤"2000# 0$$$$$$*00 0͸"000#I:"C ǔ"QQÔ""$L"$•":"$"0 0B " 2$"<Ɨ"7"̙"ƙ"2C "g"D""000 0 0ggg cgcثq1ȡ0B " 2$"u"Т"ʢ"2000# 0$$$$$$*00 0͸£"000#I:"C "QQ""$L"$"͸&"$"0 0B " 2$"d 1 A lazily initialized value for safe sharing between threads. The principal type in this module is `Lazy`, which makes it easy to construct values that are shared safely across multiple threads simultaneously. $ 1cL7?< A lazily initialized value that implements `Deref` for `T`.JG A `Lazy` takes an initialization function and permits callers from anyIF thread to access the result of that initialization function in a safeOL manner. In effect, this permits one-time initialization of global resources+( in a (possibly) multi-threaded program.OL This type and its functionality are available even when neither the `alloc`JG nor the `std` features are enabled. In exchange, a `Lazy` does **not**IF guarantee that the given `create` function is called at most once. ItKH might be called multiple times. Moreover, a call to `Lazy::get` (eitherMJ explicitly or implicitly via `Lazy`'s `Deref` impl) may block until a `T` is available.KH This is very similar to `lazy_static` or `once_cell`, except it doesn'tLI guarantee that the initialization function will be run once and it worksIF in no-alloc no-std environments. With that said, if you need stronger KH guarantees or a more flexible API, then it is recommended to use either ! `lazy_static` or `once_cell`.  " # Warning: may use a spin lock  MJ When this crate is compiled _without_ the `alloc` feature, then this type JG may used a spin lock internally. This can have subtle effects that may KH be undesirable. See [Spinlocks Considered Harmful][spinharm] for a more %" thorough treatment of this topic.  VS [spinharm]: https://matklad.github.io/2020/01/02/spinlocks-considered-harmful.html l -  KH This type is useful for creating regexes once, and then using them fromKH multiple threads simultaneously without worrying about synchronization.<-EB use regex_automata::{dfa::regex::Regex, util::lazy::Lazy, Match};OL static RE: Lazy = Lazy::new(|| Regex::new("foo[0-9]+bar").unwrap());/, let expected = Some(Match::must(0, 3..14));85 assert_eq!(expected, RE.find(b"zzzfoo12345barzzz"));<-$/ ClC--kԛBhB y{~ B/ C )IF Create a new `Lazy` value that is initialized via the given function.FC The `T` type is automatically inferred from the return type of the `create` function given.B š # C #L/   35#74 Return a reference to the lazily initialized value.C@ This routine may block if another thread is initializing a `T`.HE Note that given a `x` which has type `Lazy`, this must be called viaFC `Lazy::get(x)` and not `x.get()`. This routine is defined this way63 because `Lazy` impls `Deref` with a target of `T`.dں.GD This panics if the `create` function inside this lazy value panics.IF If the panic occurred in another thread, then this routine _may_ also+( panic (but is not guaranteed to do so).0 0 0B  ƀ 15J C JL/ ""79\4,0 0 0B  $;J C gTJL/ g@B4000 0 0ggg cgcثq1ȡ0B  $oD$cѧ\!ѧ! AtomicPtrLaD$~FQS@B1"$"/ "C "C© _-/24"$" L"4"Ԉ&,& +?/ +C +$+ $+$+ $+RTWY+/ +C +-0-Bš/ J /C /JL// /+-110 0 0B 1 $1<$<0 0řǙșəʙ˙G*R{9  0B <܊ $<?;J ?C ?JT?JL?/ ?g8:@4@000 0 0ggg cgcثq1ȡ0B @ $@A/ AC A# "A$A0 0B A $Ad 1 Types and routines for working with look-around assertions. This module principally defines two types: * [`Look`] enumerates all of the assertions supported by this crate. * [`LookSet`] provides a way to efficiently store a set of [`Look`] values. * [`LookMatcher`] provides routines for checking whether a `Look` or a `LookSet` matches at a particular position in a haystack. $ 1 Lf$\EL68#6JG Flip the look-around assertion to its equivalent for reverse searches.663 For example, `StartLF` gets translated to `EndLF`.77FC Some assertions, such as `WordUnicode`, remain the same since they7GD match the same positions regardless of the direction of the search.D8B (C!@HE Return the underlying representation of this look-around enumerationAEB as an integer. Giving the return value to the [`Look::from_repr`]BIF constructor is guaranteed to return the same look-around variant thatBFC one started with within a semver compatible release of this crate.:7 Checks that all assertions in this set can be matched.œKH Some assertions, such as Unicode word boundaries, require optional (butFC enabled by default) tables that may not be available. If there areݝKH assertions in this set that require tables that are not available, then this will return an error.О؞0- Specifically, this returns an error when theC@ `unicode-word-boundary` feature is _not_ enabled _and_ this set՟/, contains a Unicode word boundary assertion.1. It can be useful to use this on the result ofǠB? [`NFA::look_set_any`](crate::nfa::thompson::NFA::look_set_any)96 when building a matcher engine to ensure methods like̡=: [`LookMatcher::matches_set`] do not panic at search time.LÜĜŜƜǜȜ_wB $ˣ!g@111 1 1ggg cgcثq1ȡ1B  $ 쳧QQLק$ܧ"$1 1 B  $=: Creates a new default matcher for look-around assertions.B C@= Sets the line terminator for use with `(?m:^)` and `(?m:$)`.KH Namely, instead of `^` matching after `\n` and `$` matching immediatelyٶHE before a `\n`, this will cause it to match after and before the byteTJC@ It can occasionally be useful to use this to configure the line:7 terminator to the NUL byte when searching binary data.ĸ̸DA Note that this does not apply to CRLF-aware line anchors such asFC `(?Rm:^)` and `(?Rm:$)`. CRLF-aware line anchors are hard-coded to use `\r` and `\n`.1 1JB  $$'EB Returns the line terminator that was configured for this matcher.û˻A> If no line terminator was configured, then this returns `\n`.KH Note that the line terminator should only be used for matching `(?m:^)`EB and `(?m:$)` assertions. It specifically should _not_ be used for?< matching the CRLF aware assertions `(?Rm:^)` and `(?Rm:$)`.1 1B  $EľIF Returns true when the position `at` in `haystack` satisfies the givenԒ look-around assertion.dں.ʿҿHE This panics when testing any Unicode word boundary assertion in thisKH set and when the Unicode word data is not available. Specifically, thisHE only occurs when the `unicode-word-boundary` feature is not enabled.GD Since it's generally expected that this routine is called inside ofDA a matching engine, callers should check the error condition whenEB building the matching engine. If there is a Unicode word boundaryHE in the matcher and the data isn't available, then the matcher should fail to build.FC Callers can check the error condition with [`LookSet::available`].DA This also may panic when `at > haystack.len()`. Note that `at ==96 haystack.len()` is legal and guaranteed not to panic.<11 1 1B  $kk&kzt11 1 1B !$$DrKH Returns true when _all_ of the assertions in the given set match at the# given position in the haystack.dں.HJKJHJGJDJEJHJJFJDJ9J\11 1 1B  $%kll11 1 1B !$%D}t11 1 1B  $$ʀ%ހDՎU;JG Returns true when [`Look::Start`] is satisfied `at` the given position in `haystack`.dں.?< This may panic when `at > haystack.len()`. Note that `at ==9JD11 1 1B !$!ll8KH Returns true when [`Look::End`] is satisfied `at` the given position in|֘Bdں.?Kϙ9J411 1 1B !$mmwm=C@ Returns true when [`Look::StartLF`] is satisfied `at` the givenƛBdں.?K˜9J\11 1 1B  $DoNoĠ;JG Returns true when [`Look::EndLF`] is satisfied `at` the given position΀Kdں.?K9JLˠ11 1 1B ՠ $֠pp?ҡEB Returns true when [`Look::StartCRLF`] is satisfied `at` the givenܜBdĢں.բݢ?K9Jl11 1 1B  $tt=C@ Returns true when [`Look::EndCRLF`] is satisfied `at` the givenܽBݦdں.?K§9J\11 1 1B  $-x7x?EB Returns true when [`Look::WordAscii`] is satisfied `at` the given˪Bdں.?KЫ9Jl11 1 1B !$K{U{԰FKH Returns true when [`Look::WordAsciiNegate`] is satisfied `at` the given܃Bdں.į?K9J۰11 1 1B  $||˵ұGD Returns true when [`Look::WordUnicode`] is satisfied `at` the givenܞBdƲں.ײ߲?K9Jd0;8 This returns an error when Unicode word boundary tables´>; are not available. Specifically, this only occurs when the30 `unicode-word-boundary` feature is not enabled.|ҵ11 1 1!B  $)GD Returns true when [`Look::WordUnicodeNegate`] is satisfied `at` theĸ! given position in `haystack`.dں.?KϹ9Jd0;K>K3K11 1 1!B  $EJG Returns true when [`Look::WordStartAscii`] is satisfied `at` the givenܨBdں.?K9J11 1 1B  $CHE Returns true when [`Look::WordEndAscii`] is satisfied `at` the givenBdں.?K9J11 1 1B  $ FC Returns true when [`Look::WordStartUnicode`] is satisfied `at` the!ŗKdں.?K9Jd0;K>K3K11 1 1!B  $kuDA Returns true when [`Look::WordEndUnicode`] is satisfied `at` the!ŗKdں.?K9Jd0;K>K3K11 1 1!B  $ iHE Returns true when [`Look::WordStartHalfAscii`] is satisfied `at` the!ŗKdں.?K9J11 1 1B  $HFC Returns true when [`Look::WordEndHalfAscii`] is satisfied `at` the!ŗKdں.?K9J11 1 1B  $V`JG Returns true when [`Look::WordStartHalfUnicode`] is satisfied `at` the!ŗKdں.?K9Jd0;K>K3KԞ11 1 1!B  $HE Returns true when [`Look::WordEndHalfUnicode`] is satisfied `at` the!ŗKdں.?K9Jd0;K>K3K11 1 1!B  $܏<B և/ !#.B 6A> Returns an error if and only if Unicode word boundary data is unavailable.,JB 3|Ê4gD111 1 1ggg cgcثq1ȡ1B  $  !#.dD,JB t1 1!B !_bytes4_att1 1!B !K4Kd 1# A thread safe memory pool. The principal type in this module is a [`Pool`]. It main use case is for holding a thread safe collection of mutable scratch spaces (usually called `Cache` in this crate) that regex engines need to execute a search. This then permits sharing the same read-only regex object across multiple threads while having a quick way of reusing scratch space in a thread safe way. This avoids needing to re-create the scratch space for every search, which could wind up being quite expensive.  1́ 1$ 1^'=: A thread safe pool that works in an `alloc`-only context.((KH Getting a value out comes with a guard. When that guard is dropped, the(JG value is automatically put back in the pool. The guard provides both a)LI `Deref` and a `DerefMut` implementation for easy access to an underlyingD* `T`.**NK A `Pool` impls `Sync` when `T` is `Send` (even if `T` is not `Sync`). This*NK is possible because a pool is guaranteed to provide a value to exactly one+ thread at any time.++NK Currently, a pool never contracts in size. Its size is proportional to the,GD maximum number of simultaneous uses. This may change in the future.,,KH A `Pool` is a particularly useful data structure for this crate because-KH many of the regex engines require a mutable "cache" in order to execute.NK a search. Since regexes themselves tend to be global, the problem is then:.B? how do you get a mutable cache to execute a search? You could://NK 1. Use a `thread_local!`, which requires the standard library and requires//, that the regex pattern be statically known.0 2. Use a `Pool`.0MJ 3. Make the cache an explicit dependency in your code and pass it around.1KH 4. Put the cache state in a `Mutex`, but this means only one search can1 execute at a time.1+( 5. Create a new cache for every search.22OL A `thread_local!` is perhaps the best choice if it works for your use case.2MJ Putting the cache in a mutex or creating a new cache for every search are3LI perhaps the worst choices. Of the remaining two choices, whether you use4MJ this `Pool` or thread through a cache explicitly in your code is a matter430 of taste and depends on your code architecture.55"I55KH When this crate is compiled _without_ the `std` feature, then this type5JI6KI7%I77VI8l8-88HE This example shows how to share a single hybrid regex among multiple8EB threads, while also safely getting exclusive access to a hybrid's9LI [`Cache`](crate::hybrid::regex::Cache) without preventing other searches:41 from running while your thread uses the `Cache`.:<:-:-:&# hybrid::regex::{Cache, Regex},;'$ util::{lazy::Lazy, pool::Pool},t;A4;-;; static RE: Lazy =;:7 Lazy::new(|| Regex::new("foo[0-9]+bar").unwrap());<%" static CACHE: Lazy> =<63 Lazy::new(|| Pool::new(|| RE.create_cache()));==/I=JG assert_eq!(expected, RE.find(&mut CACHE.get(), b"zzzfoo12345barzzz"));<>-$>/ >Cl>"0"B*kh"B y{~ >J>$">/ >C > ?#>DA Create a new pool. The given closure is used to create values in? the pool when necessary.?B"šG>@DKAKAAKH When the `std` feature is enabled, a `Pool` is thread-aware and spreadsBIF its memory out across multiple cache lines. The number of cache linesBHE is determined by the `capacity` parameter passed here. By default, aCJG fixed reasonable number is used. A smaller number means less memory isCHE used, but a higher number means there may be less contention on thisDJG pool in highly threaded environments doing a lot of searches using theE same `Regex` value.EEEB When `std` is not enabled, then the capacity parameter is ignoredEC@ because the underlying pool implementation is not thread-aware.FFDA The capacity must be at least 1. If it's less than 1, then it isG forced to be 1.lGB"šLCHDKIKIIHE This is a convenience routine for calling `Pool::with_capacity` withJJG a number equivalent to the available parallelism for this environment.JJEB If `std` is not enabled or if the query for available parallelismK;8 failed, then this is equivalent to calling `Pool::new`.L#B šQM$MM1M"M"N"P&J QC QےJ$PJLQ/ P  02X(Q?< Get a value from the pool. The caller is guaranteed to haveQFC exclusive access to the given value. Namely, it is guaranteed thatRGD this will never return a value that was returned by another call toR-* `get` but was not put back into the pool.SSGD When the guard goes out of scope and its destructor is called, thenTC@ it will automatically be put back into the pool. Alternatively,THE [`PoolGuard::put`] may be used to explicitly put it back in the poolU&# without relying on its destructor.UUEB Note that there is no guarantee provided about which value in theVGD pool is returned. That is, calling get, dropping the guard (causingVEB the value to go back into the pool) and then calling get again isWIF *not* guaranteed to return the same value received in the first `get`LW call.X1 1 V A1B X $XX</ XC YJXg#%(*Y@Y111 1 1ggg cgcثq1ȡ1B Y $YxZJG A guard that is returned when a caller requests a value from the pool.ZZJG The purpose of the guard is to use RAII to automatically put the value['$ back in the pool once it's dropped.L[2 \ \$\#L\# \2˖V ABNzB \ԟ\\3K \K \K$\KL\2 \\68_%\74 Consumes this guard and puts it back into the pool.]]JG This circumvents the guard's `Drop` implementation. This can be useful^JG in circumstances where the automatic `Drop` results in poorer codegen,^*' such as calling non-inlined functions._1 V A1B_ ƀ _HK `K `K$_KL`2 _""_ "EG\`4``,`1 1 1B ` $`aKK aK aK$aKLa2 a"a>@a!La1 1 1B b $bb_K bK bK$bgbKLb2 bgbNPc@c111 1 1ggg cgcثq1ȡ1B c $c}Ld|dd,d #Td2 RefUnwindSafele UnwindSafeTe AtomicUsize\eDeee=fC,f6df$ddd/db,eTeDftf\fiVXgA-0(((5((ԭ(:L9+)K(<?,/(($(Y(  t(d(t*(;f(*I(F Create a new prefilter from a sequence of needles and a corresponding+( match semantics.+(+D(A This may return `None` for a variety of reasons, for example, if,C(@ a suitable prefilter could not be constructed. That might occur,C(@ if they are unavailable (e.g., the `perf-literal-substring` and-G(D `perf-literal-multisubstring` features aren't enabled), or it might.G(D occur because of heuristics or other artifacts of how the prefilterT.( works..(.J(G Note that if you have an [`Hir`] expression, it may be more convenient/K(H to use [`Prefilter::from_hir_prefix`]. It will automatically handle the0/(, task of extracting prefix literals for you.0(l0(-0(0B(? This example shows how match semantics can impact the matching1H(E algorithm used by the prefilter. For this reason, it is important to1F(C ensure that the match semantics given here are consistent with the2I(F match semantics intended for the regular expression that the literalsĒ3( were extracted from.3(<3(-3(-3-(* util::{prefilter::Prefilter, syntax},ē4( MatchKind, Span,44(-4(4( let hay = "Hello samwise";4(4C(@ // With leftmost-first, we find 'samwise' here because it comes50(- // before 'sam' in the sequence we give it..5K(H let pre = Prefilter::new(MatchKind::LeftmostFirst, &["samwise", "sam"])6( .expect("a prefilter");|6(Ќ-6 ( Some(Span::from(6..13)),7;(8 pre.find(hay.as_bytes(), Span::from(0..hay.len())),47(č-7I(F // Still with leftmost-first but with the literals reverse, now 'sam'Ա8( // will match instead!8K(H let pre = Prefilter::new(MatchKind::LeftmostFirst, &["sam", "samwise"])9(L|9(Ќ-9( Some(Span::from(6..9)),9;(L4:(č-:(:.(-<;(-;(2 2$B ;(L ;(̷L\;(  E %'D<($<(ݱ$>a(\>(Bܳ$$4?(t?(LG(DI(F This attempts to extract prefixes from the given `Hir` expression forEF(C the given match semantics, and if possible, builds a prefilter forLF( them.F(lF(-F(FH(E This example shows how to build a prefilter directly from an [`Hir`]GH(E expression, and use to find an occurrence of a prefix from the regexdG(.G((4Ί(22 2 2B Պ( $֊(D܊($(܋ (dߋ(2 2B ( $(ԏ(<(2 2B ( $(Ǎ2((̭(4܍(Tύ( ̍(($(2 2,B ( $(;($(22 2 2B ( $(# # =(4(22 2 2B ( $($ $ ё(dԑ(2 2B ( $(ג(<ڒ(2 2B ( $(\(j((2 2$$B (L (̷L\(ƭ$@7 EK7 !#($n(((D(2 2Ϝ$B (? (?̶($= = "$n(((D(2 2Ϝ$B (? (?̾( $($(*, 1 ) Lower level primitive types that are useful in a variety of circumstances. # Overview This list represents the principle types in this module and briefly describes when you might want to use them. * [`PatternID`] - A type that represents the identifier of a regex pattern. This is probably the most widely used type in this module (which is why it's also re-exported in the crate root). * [`StateID`] - A type the represents the identifier of a finite automaton state. This is used for both NFAs and DFAs, with the notable exception of the hybrid NFA/DFA. (The hybrid NFA/DFA uses a special purpose "lazy" state identifier.) * [`SmallIndex`] - The internal representation of both a `PatternID` and a `StateID`. Its purpose is to serve as a type that can index memory without being as big as a `usize` on 64-bit targets. The main idea behind this type is that there are many things in regex engines that will, in practice, never overflow a 32-bit integer. (For example, like the number of patterns in a regex or the number of states in an NFA.) Thus, a `SmallIndex` can be used to index memory without peppering `as` casts everywhere. Moreover, it forces callers to handle errors in the case where, somehow, the value would otherwise overflow either a 32-bit integer or a `usize` (e.g., on 16-bit targets). * [`NonMaxUsize`] - Represents a `usize` that cannot be `usize::MAX`. As a result, `Option` has the same size in memory as a `usize`. This useful, for example, when representing the offsets of submatches since it reduces memory usage by a factor of 2. It is a legal optimization since Rust guarantees that slices never have a length that exceeds `isize::MAX`. T 1!!!!!!!!"""""""##### NonZeroUsized ) )C, )))) )| ) )̸ ) )wyegSUAC)/)4)1 Create a new `NonMaxUsize` from the given value.)K)H This returns `None` only when the given value is equal to `usize::MAX`.)B 9E ̩)I)F Return the underlying `usize` value. The returned value is guaranteed) to not equal `usize::MAX`.)BjF %)g@))222 2 2ggg cgcثq1ȡ2B ) $) )|3)̈4)3) The maximum index value.323)643)3:)3E)4)7)6H)E The total number of values that can be represented as a small index.,7)ԁ8)7) The zero index value.$8)9)8A)> The number of bytes that a single small index uses in memory.$9):?)9) Create a new small index.9)9E)B If the given index exceeds [`SmallIndex::MAX`], then this returnsl:)/:)BO ?6);E)B Create a new small index without checking whether the given value< ) exceeds [`SmallIndex::MAX`].<) Using this routine with an invalid index value will result in=I)F unspecified behavior, but *not* undefined behavior. In particular, an=I)F invalid index value is likely to cause panics or possibly even silent>) logical errors.>)>J)G Callers must never rely on a `SmallIndex` to be within a certain range?) for memory safety.l?)Bי%-P A')@I)F Like [`SmallIndex::new`], but panics if the given index is not valid.$A)B Q C%)BE)B Return this small index as a `usize`. This is guaranteed to neverC) overflow `usize`.DC)3 3B C)$C)E!)DC)@ Return this small index as a `u64`. This is guaranteed to neverlE) overflow.4E)3 3B E)%$E)G!)FH)E Return the internal `u32` of this small index. This is guaranteed tőG) never overflow `u32`.4G)3 3B G)%$G)I!)GJ)G Return the internal `u32` of this small index represented as an `i32`.H2)/ This is guaranteed to never overflow an `i32`.4I)3 3B I)%$I)K)J6)3 Returns one more than this small index as a usize.J)JK)H Since a small index has constraints on its maximum value, adding `1` toK7)4 it will always fit in a `usize`, `u32` and a `i32`.DL)3 3B L)$L)NZ)LH)E Decode this small index from the bytes given using the native endianM&)# byte order for the current target.M)MH)E If the decoded integer is not representable as a small index for theN/), current target, then this returns an error.lN)B%W N)S<)QH)MQ&)MR)RJ)G This is analogous to [`SmallIndex::new_unchecked`] in that is does notRH)E check whether the decoded integer is representable as a small index.S)B%Y S)U$)TK)H Return the underlying small index integer as raw bytes in native endian\U)4\U)3 3B U)%$U) U)V,)/ V)\V)4V)V(),V)3 3 3B V) $V)g[ W/)/ W)#W4)LW)3 3 3B X) $X)x\ X/)X)X)/ X)\Y)4Y)Y(),Y)3 3 3B Y) $Y)v] Z2)Z)̧Z)/ Z)#[4)L[)3 3 3B [) $[)a^ [)\ )$\)B ,\)\ )T]),])]>)D])B ,])_ )T_),_)`>)D`)B ,`)b )Tb),b)b>)Db)B ,b)d")Te),e)e@)De)B ,e)l)m)lC)@ Returns the value that could not be converted to a small index.Lm)3 3B m) $m)n*)|n)n)n+)gn@)n)333 3 3ggg cgcثq1ȡ3B o) $o) o)q )QQLq)$q)r()$r)3 3B r) $r)u)u) u) ) u) v), u)8 $u)& u)8,u)$ u), u)8\u)& u)8,u)$ u), u)8$u)& v)8,v)$ v), v)8withiterDv)& v)8,v)*v) v) )8$v), v)8 $v) v) )< The maximum value.v)8v)8,v)8v)& v), v)8 $v) v), v)8 $v) w) w)8Tw)'w)8w)% w)<4 The total number of values that can be represented.w7)8w)8,w)8,w)& w)8,w) w)8Tw)'x)8,x)% x)< The zero value.x)8x)8,x)8$x)& x), x)8 $x) x), x)8 $x) x) x)8Tx)'x)8$x)% x)<8 The number of bytes that a single value uses in memory.x;)8y)8,y)8$y)& y)8,y) y)8Ty)'y)8$y)% y)<; Create a new value that is represented by a "small index."y>)<z)<@ If the given index exceeds the maximum allowed value, then thiszC)< returns an error.{)+ {) {) {)84{)8{)8 {)8 {) {) {)8,{)& {)8,{)({)84{) {), |)8 $|)$ |), |)8\|) |) |) |) 8T|)'|)8 |) |) |)8,|) |)8 |) |) |), |)8 $|) |)8 <|) |) |), |)8\|)<? Create a new value without checking whether the given argument|B)< exceeds the maximum.Ĺ})<})<8 Using this routine with an invalid value will result in};)<7 unspecified behavior, but *not* undefined behavior. In~:)<= particular, an invalid ID value is likely to cause panics or~@)<% possibly even silent logical errors.()<)<< Callers must never rely on this type to be within a certain?)< range for memory safety.ۀ)+ ) ) )84)8)8,)8 )8 l) ) )8,)& )8,)(Á), Ɓ)8 $ǁ) ́) ), ށ)8 $߁) ) )8T)')8 l) ) )8,)<8 Like `new`, but panics if the given value is not valid.;)+ ) ) )84)8)8 )8$) ) )8,)& )8,)(), )8 $) ) ل), )8 $)')8 ) ) ă)8,) Ń)84ƃ) ̃) ˄)84̓) Ӄ) ԃ) ʄ)7invalid T)$ )8L) ) ) ), )8 $)$ )7 valueD)<> Return the internal value as a `usize`. This is guaranteed toA)< never overflow `usize`.ܶ)+ ޅ) ߅) )84)8)8,)8 )8߂D) ) ) )8$)()8,) ) ʆ)8$) )7 ) )8߂D) ) )<< Return the internal value as a `u64`. This is guaranteed toن?)< never overflow.)+ Ň) Ƈ) ͇)84LJ)8ۇ)8,߇)8 )84) ) ) )8$)()8) ) )8$) )7 ) )84) ) )<< Return the internal value as a `u32`. This is guaranteed to?)<M̆)+ ) ) )84)8‰)8,Ɖ)8 ̉)84ω) Չ) ۉ) ։)8$׉)(݉)8) ) )8$) )7 ) )84) ) )<; Return the internal value as a i32`. This is guaranteed to>)< never overflow an `i32`.)+ ) ) )84)8)8,)8 )84) ) ċ) )8$)(Ƌ)8ɋ) ͋) )8$ߋ) )7 ) )84) ) )<- Returns one more than this value as a usize.0)<nj)<= Since values represented by a "small index" have constraints׌@)<> on their maximum value, adding `1` to it will always fit in aA)< `usize`, `u32` and a `i32`.)+ ) ) )84)8)8 )8D) Î) Ɏ) Ď)8$Ŏ)(ˎ)8,Ύ) Ԏ) )8$) )7 ) )8D) ) )<? Decode this value from the bytes given using the native endianB)<M&)<)<= If the decoded integer is not representable as a small index@)<4 for the current target, then this returns an error.7)+ ) ) )84)8̑)8 Б)8lӑ) ) )8,)& ) ) )8)% )7 )()84) ), )8 $)$ ), )8\) ) ) ) 8T)')8l) ) )8,) )8 ) ) Ē), )8 $) Œ)8 <ƒ) ͒) Ғ), Β)8\ϒ)<NB)<M&)<)<? This is analogous to `new_unchecked` in that is does not checkB)<? whether the decoded integer is representable as a small index.ДB)+ ) ) )84)8)8 )8) ӕ) )8,ԕ)& ٕ) ە) )8ܕ)% ޕ)7 )(), )8 $) ) ), )8 $) ) )8T)')8) ) )8,)<< Return the underlying integer as raw bytes in native endian̖?)<4\)+ ) ) )84)8Ɨ)8 ʗ)8\͗) ؗ) ޗ) ٗ)8$ڗ)() ) )8)% )7 ) ) )8$) )7 ) )8\) ) )<9 Returns an iterator over all values from 0 up to and not<)< including the given length.)<)<9 If the given length exceeds this type's limit, then this<)<6\)8) ) )8,)8 )8$) ) )8)& )8,)(), )8$) ) ), ͚)8$Κ)'Қ)8 Ԛ) ך) ۚ)8ؚ)8$)8$)')8)')8q,)8 ), )8 $) ) ԝ) 8 )8) ) ՜)  )8$)$ )8 )& ) )8)8$)'Ŝ)8ǜ)'ʜ)8L̜)(ל)8$ڜ)'ޜ)8)')84) ) ʝ) 8 ) )8 debug_tuple\) ) )8L) ) ) ), )8 $) )8,) ) ) )8$) )84) ) ) )8n4) ) )8$ߝ) )8 ) )8$)')8 )')8Index,) ), )8 $) )8 ) ) )8 ) ) )8!$)84) )8 )% )+ ) ) )84)8 ɞ)8,̞) ў) ) Ҟ)8$Ӟ)$ מ)8,ٞ)& ޞ), )8 $)() )8 ) ) ) )8$) ) )8,) )8߂D) ) )8$) )8 ) )8$)'ğ)8 Ɵ)'ɟ)8IndexMutD˟) ӟ), ԟ)8 $՟) ٟ)8 ۟) ߟ) )8 ) ) ) + ) ) )84)8 )8L) ) ) )8)8$)$ )8,)& ), )8 $)() )8)8 ) ) ) ɠ)8ʠ)8$Π) Ҡ) )8,Ӡ) ؠ)8߂D٠) ) )+ ) ) )8) ) )8<) )7<)8$) )8 ) )8$)')8 )')8N,) ), )8 $¡) ơ)8 ȡ)8̡) ϡ)8 С) ѡ) ӡ) )8!$)84) )8 )% )+ ) ) )84)8 )8,) ) ) )8$)$ )8,)& ), )8 $)() )8 ) ) ) ˢ)8$̢) Т) )8,Ѣ) ֢)8߂Dע) ߢ) )+ ) ) )8) ) )8<) )7<)8$) )8 ) )8$)')8 )')8ND) ), £)8 $ã) ǣ)8 ɣ)8ͣ) У)8 ѣ) ң) ԣ) ) + ) ) )84)8 )8L) ) ) )8)8$)$ )8,)& ), )8 $)() )8)8 ) ) ) )8)8$) ä) Ԥ)8,Ĥ) ɤ)8߂Dʤ) Ҥ) Ӥ)8$)8$) )8) )8 ), )8 $) ) )8 )8$) ) )8,)& )8)(), )8 $) ) ), ˥)8 $̥) Х) )8Tѥ)'ۥ)8$ݥ) ) )8,)8$)8<) )8) )8 ), )8 $) ) )8!$)8,) ), ¦)8\æ)% Ʀ)8 զ)8Dئ) ) )8,)& )8)()84) ), )8 $)$ ), )8\) ) ) ק) 8T)')8D) ) )8,) )8 ) ) ), )8 $) )8 <) ħ) ɧ), ŧ)8\Ƨ)8$)8<) )8) )8 ), )8 $) ) ©)8!$)8,) ), )8\)% )8 )8D) ) ̨)8,¨)& Ǩ)8ɨ)(Ψ)84Ѩ) ר), ب)8 $٨)$ ݨ), ߨ)8\) ) ) ) 8T)')8D) ) )8,) )8 ) ) ), )8 $) )8 <) ) ), )8\)8$ͩ)8<ҩ) ٩)8ک) ݩ)8 ߩ), )8 $) ) )8!$)8,) ), )8\)% )8 )8D) ) )8,)& )8)()84) ), )8 $)$ ), )8\) Ī) ƪ) ) 8Tت)')8D) ) )8,) )8 ) ) ), )8 $) )8 <) ) ), )8\)8$)8<) )8,) )8 «), ƫ)8 $ǫ) ̫) )8!$ګ)8,߫) ), )8\)% )8 )8D) ) )8,)& )8,)()84) ), )8 $)$ ), )8\) ) ) ) 8T)'Ǭ)8Dɬ) Ѭ) ׬)8,Ҭ) ج)8 ٬) ܬ) ), ݬ)8 $ެ) )8 <) ) ), )8\)+ ) ) )8) ) )8$)8$)8 quickcheckT)')8 ArbitraryL)8 í), ǭ)8 $ȭ) ͭ) ֮)8 ۭ)8 arbitraryLޭ) ) )85)& ) )8)8NT)')8Gen)(), )8 $) ) ̮), )8 $) ) )8T)')8NL) ) )85)<9 This error occurs when a value could not be constructed.<)<)<@ This occurs when given an integer exceeding the maximum allowedC)<ʅET)<)<? When the `std` feature is enabled, this implements the `Error`B)< trait.T)+ ) ) )84) ) )8e,)$ )8q,)$ )8~)$ )8L)8)84), )8\) ) ɱ)8|)% ʱ)8$ձ), ڱ)8\۱) ߱) )<8 Returns the value that could not be converted to an ID.;)8)8 )8L) Ų) ˲) Ʋ)8$Dz)(Ͳ)8в) Բ) )8$) )7 ) )8L) ) )+ ) ) )8) ) )8<) )7,)8$)8)'³)8o,ij)'ɳ)8,˳)8 ѳ), ճ)8\ֳ) ڳ) ۳)8$)8$)')8)')8u<)8 ), )8\) ) ) 8 )8) ) )  )8$)$ )8 )& ) )8)8$)')8)')8L)()8$Ĵ)'ȴ)8ʴ)'ʹ)84ϴ) ִ) ۶)8write,) ) ) Ͷ)8 )$ )71failed to create {} from {:?}, which exceeds {:?}3)$ ε)8L) ) ) ), )8 $)$ )8$) )8L) ) )$ ), )8 $)')8)$ )+ ) ) )84) ) )8e,)$ )8q,)8) ) )8,)84), )8$) ) )8t)% )8$·), Ƿ)8$ȷ) ͷ) Ӻ)8 ۷)8 ޷) ) )8)& )8,)(), )8$) ) ɺ)84) ) ) )8)), )8 $)')8,)$ )7Xcannot create iterator for {} when number of \ elements exceed {:?}ϸZ)$ )8L) ȹ) ɹ) Ϲ), ʹ)8 $˹)$ й), )8 $)')8,)$ )% ), )8$) ) )8t) ) )8)& )7 )!)8)8$޺)8D)8 ), )8$) ) )8!$)8$) ), )8 $)% )8 )8 $) ) ) )8)8$)()84) ), )8 $») ƻ) Ȼ) ) 8$ڻ) ޻)7 ߻) )8 $) ) ) )8 ) ) ), )8 $)<D An iterator adapter that is like std::iter::Enumerate, but attachesG)<@ small index values instead. It requires `ExactSizeIterator`. AtC)<? construction, it ensures that the index of each element in theB)<A iterator is representable in the corresponding small index type.D)+ Ⱦ) ɾ) ޾)84ʾ) о) ݾ)8e,Ѿ)$ ־)8q,ؾ)8) ) )8,)84), )8MD) )8 ) ) ) ) 8,)& )8 )$ )8؉)& ), )8$)$ )8$ǿ) ˿)8 ̿)& Ϳ)8DϿ) ؿ)8ExactSizeIteratorڿ) ), )8MD) )8 ) ) ) ) 8 )8 ) ) )8,)& )8 )(), )8MD) )8 ) ) ) ) 8)8؉) ), )8 $)')8$) ) )8,) )8) ) )% ), )8MD) ) )8,)$ )8؉)8$) )8 )& )8D) )8N) )8D)8 ), )8MD) )8 ) ) ) )8!$)8$) ) ) ), )8 $)$ )8 )')8$)% )8 )8 $) ) ) )8)8$)()84) ) ) ), )8 $)$ )8 )')8$) ) ) )8)8$) )8$) )8,) )8 $) ) )- )% )8)8-) )8$) )8؉) )8 $) ) ) )84) ) )% )8$) ) ) ) )8-)$ )8$)% )+))6)QD)6.N!6:N-g))&B,)Q) ̫ c)t)&BN,)N)  ()? )@D)l 1LG Provides helpers for dealing with start state configurations in DFAs. , 1###\k D%,$$$t$\ReR0uwdfRMV$FH46\"#"-* Create a new default start configuration.""KH The default is an unanchored search that starts at the beginning of thel# haystack.#B &6$DA A convenience routine for building a start configuration from an$# [`Input`] for a forward search.%%HE This automatically sets the look-behind byte to the byte immediately%GD preceding the start of the search. If the start of the search is at&0- offset `0`, then no look-behind byte is set.&33 3$$$$$$*03B ','(U+6)DN)# [`Input`] for a reverse search.)*HN*C@ following the end of the search. If the end of the search is at+=: offset `haystack.len()`, then no look-behind byte is set.+33 3$$$$$$*03B +,+18-63 Set the look-behind byte at the start of a search.--JG Unless the search is intended to logically start at the beginning of a.KH haystack, this should _always_ be set to the byte immediately preceding/JG the start of the search. If no look-behind byte is set, then the start/IF configuration will assume it is at the beginning of the haystack. For0'$ example, the anchor `^` will match.0030 The default is that no look-behind byte is set.\1B $1$1232&# Set the anchored mode of a search.22(% The default is an unanchored search.D2łB $2$24+3EB Return the look-behind byte in this configuration, if one exists.|43 3B 4$45&430 Return the anchored mode in this configuration.d54 4łB 5$5<>6>4 4B > /,>?K+K4 4B KТ$K$KNhTN4 4ÜĜŜƜǜȜ_wB N ,NPUeDU44 4 4ݘB U$UUY*dY4 4B Y $YY&gY@Y444 4 4ggg cgcثq1ȡ4B Y $Y ZZLZTpq3TqB qttBv ,v4 4B v $vy&Dy4 4B yƫ $yt 1* Utilities for dealing with the syntax of a regular expression. This module currently only exposes a [`Config`] type that itself represents a wrapper around the configuration for a [`regex-syntax::ParserBuilder`](regex_syntax::ParserBuilder). The purpose of this wrapper is to make configuring syntax options very similar to how other configuration is done throughout this crate. Namely, instead of duplicating syntax options across every builder (of which there are many), we instead create small config objects like this one that can be passed around and composed.  1 14 1$*=*C* $**,*l*,*D*d*{} *o*_aMO;=1*J*G A convenience routine for parsing a pattern into an HIR value with the* default configuration.*l*-*8*5 This shows how to parse a pattern into an HIR value:*<*-%*" use regex_automata::util::syntax;*3*0 let hir = syntax::parse(r"([a-z]+)|([0-9]+)")?;I*F assert_eq!(Some(1), hir.properties().static_explicit_captures_len());*.*-<*-,*4 4B * <*K* K*H A convenience routine for parsing many patterns into HIR value with theԀ *O *l *- * K*H This shows how to parse many patterns into an corresponding HIR values: *< *-L * use { %*" regex_automata::util::syntax, &*# regex_syntax::hir::Properties,4 *- * $*M * r"([a-z]+)|([0-9]+)",̮ * r"foo(A-Z]+)bar",D *? G*D let props = Properties::union(hirs.iter().map(|h| h.properties())); >*; assert_eq!(Some(1), props.static_explicit_captures_len()); * .*-<*-T*4 4&B */ */T* w G*I*F A convenience routine for parsing a pattern into an HIR value using al* `Config`.*l*-*J*G This shows how to parse a pattern into an HIR value with a non-default* configuration:*<*-%*O*!* let hir = syntax::parse_with(* r"^[a-z]+$",;*8 &syntax::Config::new().multi_line(true).crlf(true),<*=@*= assert!(hir.properties().look_set().contains_anchor_crlf());*.*-<*-T*44 4 4B * <*4*l*K*H A convenience routine for parsing many patterns into HIR values using al*ڪO*l*-*J*G This shows how to parse many patterns into an corresponding HIR values%*" with a non-default configuration:*<*-L*٥O%*O&*O4*-**:*Ot* r"\W",*O4*:B*? let config = syntax::Config::new().unicode(false).utf8(false);;*8 let hirs = syntax::parse_many_with(patterns, &config)?;G*ǧO* assert!(!props.is_utf8());*.*-<*-|*44 4 4&B */ */T*&w: D !\$*%*$.*+ Return a new default syntax configuration.%*B ,6*(;*8 Enable or disable the case insensitive flag by default.)*)F*C When Unicode mode is enabled, case insensitivity is Unicode-aware.*B*? Specifically, it will apply the "simple" case folding rules as** specified by Unicode.***D*A By default this is disabled. It may alternatively be selectively+>*; enabled in the regular expression itself via the `i` flag.,*B $,*,*10*,>*; Enable or disable the multi-line matching flag by default.-*-E*B When this is enabled, the `^` and `$` look-around assertions will.H*E match immediately after and immediately before a new line character,.H*E respectively. Note that the `\A` and `\z` look-around assertions are/G*D unaffected by this setting and always correspond to matching at the/#* beginning and end of the input.0*0D*ȹO0>*; enabled in the regular expression itself via the `m` flag.T1*B $1*1*9:*2F*C Enable or disable the "dot matches any character" flag by default.2*2K*H When this is enabled, `.` will match any character. When it's disabled,3F*C then `.` will match any character except for a new line character.4*4I*F Note that `.` is impacted by whether the "unicode" setting is enabled4K*H or not. When Unicode is enabled (the default), `.` will match any UTF-85G*D encoding of any Unicode scalar value (sans a new line, depending on5H*E whether this "dot matches new line" option is enabled). When Unicode6G*D mode is disabled, `.` will match any byte instead. Because of this,7G*D when Unicode mode is disabled, `.` can only be used when the "allow7E*B invalid UTF-8" option is enabled, since `.` could otherwise match8* invalid UTF-8.8*8D*ȹO9>*; enabled in the regular expression itself via the `s` flag.9*B $9*9*?**:6*3 Enable or disable the "CRLF mode" flag by default.;*;D*ȹO;>*; enabled in the regular expression itself via the `R` flag.<*<5*2 When CRLF mode is enabled, the following happens:<*F*C `\r` and `\n` as line terminators. And in particular, neither will>$*! match between a `\r` and a `\n`.$?*B $?*?*H4*@D*A Sets the line terminator for use with `(?u-s:.)` and `(?-us:.)`.@*@L*I Namely, instead of `.` (by default) matching everything except for `\n`,AF*C this will cause `.` to match everything except for the byte given.A*AK*H If `.` is used in a context where Unicode mode is enabled and this byteBE*B isn't ASCII, then an error will be returned. When Unicode mode isCK*H disabled, then any byte is permitted, but will return an error if UTF-8C/*, mode is enabled and it is a non-ASCII byte.D*DI*F In short, any ASCII value for a line terminator is always okay. But aDH*E non-ASCII byte might result in an error depending on whether UnicodeE#* mode or UTF-8 mode are enabled.E*EH*E Note that if `R` mode is enabled then it always takes precedence andFH*E the line terminator will be treated as `\r` and `\n` simultaneously.G*GC*@ Note also that this *doesn't* impact the look-around assertionsGB*? `(?m:^)` and `(?m:$)`. That's usually controlled by additionalH-** configuration in the regex engine itself.|H*B $H*$H*L0*I7*4 Enable or disable the "swap greed" flag by default.I*JK*H When this is enabled, `.*` (for example) will become ungreedy and `.*?`J* will become greedy.J*JD*ȹOK>*; enabled in the regular expression itself via the `U` flag.TL*B $L*L*P7*L2*/ Enable verbose mode in the regular expression.M*MG*D When enabled, verbose mode permits insignificant whitespace in manyMG*D places in the regular expression, as well as comments. Comments areN=*: started using `#` and continue until the end of the line.O*OF*C By default, this is disabled. It may be selectively enabled in theOH*E regular expression by using the `x` flag regardless of this setting.P*B $P*P*W-*Q8*5 Enable or disable the Unicode flag (`u`) by default.Q*QG*D By default this is **enabled**. It may alternatively be selectivelyR?*< disabled in the regular expression itself via the `u` flag.R*RG*D Note that unless "allow invalid UTF-8" is enabled (it's disabled bySH*E default), a regular expression will fail to parse if Unicode mode isTE*B disabled and a sub-expression could possibly match invalid UTF-8.T*TK*H **WARNING**: Unicode mode can greatly increase the size of the compiledUF*C DFA, which can noticeably impact both memory usage and compilationVH*E time. This is especially noticeable if your regex contains characterVH*E classes like `\w` that are impacted by whether Unicode is enabled orWG*D not. If Unicode is not necessary, you are encouraged to disable it. Set the nesting limit used for the regular expression parser._*`K*H The nesting limit controls how deep the abstract syntax tree is allowed`I*F to be. If the AST exceeds the given limit (e.g., with too many nesteda5*2 groups), then an error is returned by the parser.a*aG*D The purpose of this limit is to act as a heuristic to prevent stackbI*F overflow when building a finite automaton from a regular expression'scD*A abstract syntax tree. In particular, construction currently usescI*F recursion. In the future, the implementation may stop using recursiond0*- and this option will no longer be necessary.d*dH*E This limit is not checked until the entire AST is parsed. Therefore,eI*F if callers want to put a limit on the amount of heap space used, theneG*D they should impose a limit on the length, in bytes, of the concretefG*D pattern string. In particular, this is viable since the parser willgH*E limit itself to heap space proportional to the length of the pattern\g* string.g*gI*F Note that a nest limit of `0` will return a nest limit error for mosthJ*G patterns but not all. For example, a nest limit of `0` permits `a` butiI*F not `ab`, since `ab` requires a concatenation AST item, which resultsiJ*G in a nest depth of `1`. In general, a nest limit is not something thatjE*B manifests in an obvious way in the concrete syntax, therefore, itj)*& should not be used in a granular way.Tk*B $k*,k*r+*l+*( Whether to support octal syntax or not.l*lH*E Octal syntax is a little-known way of uttering Unicode codepoints inm@*= a regular expression. For example, `a`, `\x61`, `\u0061` andmI*F `\141` are all equivalent regular expressions, where the last examplen* shows octal syntax.n*nK*H While supporting octal syntax isn't in and of itself a problem, it doesoJ*G make good error messages harder. That is, in PCRE based regex engines,oA*> syntax like `\1` invokes a backreference, which is explicitlypH*E unsupported in Rust's regex engine. However, many users expect it topF*C be supported. Therefore, when octal support is disabled, the errorqI*F message will explicitly mention that backreferences aren't supported.r*r(*% Octal syntax is disabled by default.,r*B $r*r*s!*s.*+ Returns whether "unicode" mode is enabled.\s*4 4B s* $s*t**t7*4 Returns whether "case insensitive" mode is enabled.t*4 4B t* $t*u$*u1*. Returns whether "multi line" mode is enabled.tu*4 4B u* $u*v.*v;*8 Returns whether "dot matches new line" mode is enabled.v*4 4B w* $w*w*w+*( Returns whether "CRLF" mode is enabled.Dx*4 4B x* $x*x'*x=*: Returns the line terminator in this syntax configuration.y*4 4B y* $y*z$*y1*. Returns whether "swap greed" mode is enabled.tz*4 4B z* $z*{+*z8*5 Returns whether "ignore whitespace" mode is enabled.{*4 4B {* ${*|*{**' Returns whether UTF-8 mode is enabled.D|*4 4B |* $|*|#*|%*" Returns the "nest limit" setting.t}*4 4B }* $}*}*},*) Returns whether "octal" mode is enabled.L}*4 4B ~* $~*~7*,~*44 4 4B *&$*<*ӃG*L*44 4 4      empty_min_rangeLalB * $*<*f*L*44 4 4TranslatorBuilderf=#8=;B * $*<*։**<*B d 1 Types and routines that support the wire format of finite automata. Currently, this module just exports a few error types and some small helpers for deserializing [dense DFAs](crate::dfa::dense::DFA) using correct alignment. $ 1$$$$$< =C<L!t"<"d#$d,̒D*(\d/z|&&D&,B$&'*g'@'444 4 4ggg cgcثq1ȡ4B ' $' '()|((44<<4,B 55F6,B $678l7,B $78B8B D8,89A|9B D9,:;b;B L;<;==t=,B D=>>>,B $>?g?,B \?$?@ntA,B \A$AB+|BBC,gC@C444 4 4ggg cgcثq1ȡ4B C $C CԃDR<R4 4 4B R,RWHW4 4 4B W,W]@]4 4 4B ] ,]cU|c4 4ٖ2B d/ c ,di9i4 4B j ,jqFq̪qqB/ q  rT4 4,ÜĜŜƜǜȜ_wB  ,expected_labeltȃ$b\44 4 4ݘB ,3|4 4B Ƣ,ߙ[4 4PB  ,˞gٞ4 4ݘB 2 24  3  "¡3СBmd4 4PB  ,expected_versionɦʫplث4 4ݘB 2 24    (*Ʈ*ԮBУz|4 4,ÜĜŜƜǜȜ_wB ѱ ,ʱ$ܱ ,2U K̑4 4PB  ,]4 4B 2 24    )+z4 4,ÜĜŜƜǜȜ_wB  ,$½vl4 4,ӵPB  ,$' ,2PG4 4B ,Yt4 4B 2 24< G #%|4 4,ÜĜŜƜǜȜ_wB  ,$LPUPUP|4 4,PB  ,$LPUPնUPqd4 4,ÜĜŜƜǜȜ_wPB  ,$qd4 4,ÜĜŜƜǜȜ_wPB  ,$sl4 4,ÜĜŜƜǜȜ_wB  ,$+D4 4B , +D4 4B ۭ, -L4 4B  ,|4 4,ٖ2B /  , at_least_lend$68k,PB   $k,PB   $k,PB   $PP:\B non_padding_len|4666#%L4 4B ׁ  ʁҁ%L4 4B   'T4 4B   GEd݅2c5 $L4 4B   $L4 4B   ˈ&TΈ4 4B   وˉ$LΉ4 4B   ؉$L4 4B ˊ  Ɗ&T4 4B    1, 1 1 +1 61\ 12C,,2$2$2$|2$D2D2 $2<2 <2"L2%L2&T2&,2#f$2|2C\258&) ,2l2$2F2ל2nj22d2vxL2,2$$$$$$$ )K)K)C))<) L*!<*"$*4*L),)L)D)\))l))Esu)ef$*VX0)0404 4 4B 0 $01g1@1444 4 4ggg cgcثq1ȡ4B 1 $1 1T33$3B 4%D44 4B 4 $44)d44 4B 4 $45)d54 4B 5 $56)L64 4B 6 $66)L64 4B 6 $67'L74 4B 7 $78=l84 4B 8 $8,89@94 4řǙșəʙ˙G*R{9 1B 9 $9;@;4 4ٗB ;; ;;t;; ٗ$; ;*,<*d<4 4B < $<<$<4 4 ^yϊu-4B < $<AA(AB B;dBB $BC,C5 5B C $CC&DC5 5B D $DF-gF@F555 5 5ggg cgcثq1ȡ5B F $F FHH3DHB $HJ)J5 5B J $JK)K5 5B K $KK)LK5 5B K $LLblL5 5B LPM#P̉M#PP5 5ͶB n1 D $ G[5 5ͶB  Qn2D$v5 5ͶB  QQn3D$?5 5ͶB  QD$Hw5 5ͶB  QQD$Qw&5 5ͶB  QQQD$Q141$$%%%%%%%%%%, RangeBounds\$LL!f$Lqs\gA.0'2'(A'=: Create a new search configuration for the given haystack.(B4(\( 2| (H:,! Set the span for this search.--JG This routine does not panic if the span given is not a valid range for-HE this search's haystack. If this search is run with an invalid range,.IF then the most likely outcome is that the actual search execution willT.ͽ.//>; This routine is generic over how a span is provided. While/<9 a [`Span`] may be given directly, one may also provide a0DA `std::ops::Range`. To provide anything supported by range0,) syntax, use the [`Input::range`] method.11,) The default span is the entire haystack.11DA Note that [`Input::range`] overrides this method and vice versa.2d2ں.22HE This panics if the given span does not correspond to valid bounds in30- the haystack or the termination of a search.3l3-33FC This example shows how the span of the search can impact whether a4KH match is reported or not. This is particularly relevant for look-around4GD operators, which might take things outside of the span into account5(% when determining whether they match.5<5-6=-6-6&ӀA7>47-77-* // Look for 'at', but as a distinct word.7%" let re = PikeVM::new(r"\bat\b")?;8&?8(?88>; // Our haystack contains 'at', but not as a distinct word.9 let haystack = "batter";9941 // A standard search finds nothing, as expected.:%-:->:'>;;DA // But if we wanted to search starting at position '1', we might;DA // slice the haystack. If we do this, it's impossible for the \b>=: assert_eq!(Some(Match::must(0, 0..2)), caps.get_match());>>FC // But if we specify the span of the search instead of slicing the?DA // haystack, then the regex engine can "see" outside of the span?)& // and resolve the anchors correctly.@0- let input = Input::new(haystack).span(1..3);@->A'>AA.- 0`, no match is found.]HE 2. The regex `a` is compiled with `Anchored::Yes` and searches `aba`]GD starting at position `2`. This reports a match at `[2, 3]` since^HE the match starts where the search started. Since there is no `^`,_IF there is no requirement for the match to start at the beginning of_ the haystack._HE 3. The regex `a` is compiled with `Anchored::Yes` and searches `aba``JG starting at position `1`. Since `b` corresponds to position `1` andaGD since the search is anchored, it finds no match. While the regexaHE matches at other positions, configuring the search to be anchoredbJG requires that it only report a match that begins at the same offsetb&# as the beginning of the search.cGD 4. The regex `a` is compiled with `Anchored::No` and searches `aba`cEB starting at position `1`. Since the search is not anchored anddIF the regex does not start with `^`, the search executes as if theredIF is a `(?s:.)*?` prefix that permits it to match anywhere. Thus, ite# reports a match at `[2, 3]`.eeEB Note that the [`Anchored::Pattern`] mode is like `Anchored::Yes`,f<9 except it only reports matches for a particular pattern.glg-ggDA This demonstrates the differences between an anchored search andgEB a pattern that begins with `^` (as described in the above warninglh message).hjGD let input = Input::new(haystack).span(2..3).anchored(Anchored::No);k->kHE // No match is found because 2 is not the beginning of the haystack,l  // which is what ^ requires.l'>mm  let re = PikeVM::new(r"a")?;mJ>nHE let input = Input::new(haystack).span(2..3).anchored(Anchored::Yes);n->oKH // An anchored search can still match anywhere in the haystack, it justoGD // must begin at the start of the search which is '2' in this case.p=: assert_eq!(Some(Match::must(0, 2..3)), caps.get_match());pp ҫRqJ>qHE let input = Input::new(haystack).span(1..3).anchored(Anchored::Yes);r->rC@ // No match is found since we start searching at offset 1 whichsHE // corresponds to 'b'. Since there is no '(?s:.)*?' prefix, no matchs // is found.t'>tt ҫRtJ>uGD let input = Input::new(haystack).span(1..3).anchored(Anchored::No);v->vHE // Since anchored=no, an implicit '(?s:.)*?' prefix was added to thewJG // pattern. Even though the search starts at 'b', the 'match anything'w-* // prefix allows the search to match 'a'.x.+ let expected = Some(Match::must(0, 2..3));x+?xx.-=: assert_eq!(Some(Match::must(0, 0..8)), caps.get_match());Ĉ̈?< // When 'earliest' is enabled and the regex engine supports>; // it, the search will bail once it knows a match has beenlӉ // found.63 let input = Input::new("foo12345").earliest(true);->Ҋ=: assert_eq!(Some(Match::must(0, 0..4)), caps.get_match());.-<Nj-DB 2$ Ӓ2͌/, Set the span for this search configuration.DA This is like the [`Input::span`] method, except this mutates theҍ span in place.>Q<Q `std::ops::Range`.dں.HQ0Ql-Ԑ<ܐ-֎R)& let mut input = Input::new("foobar");(R input.set_span(2..4);(% assert_eq!(2..4, input.get_range());<-Dڒ5 5(B R RT 2($2 .0<?< Set the span for this search configuration given any range.EB This is like the [`Input::range`] method, except this mutates theԖRKR×HQIQTޘͽ.dں.ERؙBRER!܌RJG This also panics if the given span does not correspond to valid bounds3Rl-<-̜֎R)R(Rҝ input.set_range(2..=4);(R<-L5 5Ւ(B R ʞR͞ 2Ւ($A -/,)ޥGD Set the starting offset for the span for this search configuration.GD This is a convenience routine for only mutating the start of a span*' without having to set the entire span.dں.ƧΧJG This panics if the span resulting from the new start position does notFC correspond to valid bounds in the haystack or the termination of a\ԗ7l-<-֎Rɩѩ)R(R input.set_start(5);Ȫ(% assert_eq!(5..6, input.get_range());<-L5 5B  2$u %EB Set the ending offset for the span for this search configuration.ʬҬEB This is a convenience routine for only mutating the end of a span*R˭dӭں.HE This panics if the span resulting from the new end position does notFR\ԗ7l-<-¯֎R)R(RȰ input.set_end(5);(% assert_eq!(0..5, input.get_range());<-<5 5B  2$ .$! Set the anchor mode of a search.ŲB? This is like [`Input::anchored`], except it mutates the search܌ configuration in place.l-Ƴ<γ-ڳ52 use regex_automata::{Anchored, Input, PatternID};)Rʴ30 assert_eq!(Anchored::No, input.get_anchored());! let pid = PatternID::must(5);/, input.set_anchored(Anchored::Pattern(pid));=: assert_eq!(Anchored::Pattern(pid), input.get_anchored());<-dǶ5 5łB Զ 2$ٶ@ )DA Set whether the search should execute in "earliest" mode or not.B? This is like [`Input::earliest`], except it mutates the searchܲRҸlڸ-<-֎R)Rٹ# assert!(!input.get_earliest());쁺 input.set_earliest(true);" assert!(input.get_earliest());<ʺ-d5 5B  2$ ν"C@ Return a borrow of the underlying haystack as a slice of bytes.l-<-֎RѼټ%:,) assert_eq!(b"foobar", input.haystack());<-Dս5 5 2B ޽ 2$߽-* Return the start position of this search.ƾξB? This is a convenience routine for `search.get_span().start()`.DA When [`Input::is_done`] is `false`, this is guaranteed to returnFC an offset that is less than or equal to [`Input::end`]. Otherwise,2/ the offset is one greater than [`Input::end`].l-<-֎R%:! assert_eq!(0, input.start());0- let input = Input::new("foobar").span(2..4);! assert_eq!(2, input.start());<-,5 5B  2$+( Return the end position of this search.@= This is a convenience routine for `search.get_span().end()`.HE This is guaranteed to return an offset that is a valid exclusive end$! bound for this input's haystack.l-<-֎R%: assert_eq!(6, input.end());0R assert_eq!(4, input.end());<-5 5B  2$2/ Return the span for this search configuration.JG If one was not explicitly set, then the span corresponds to the entire range of the haystack.GD When [`Input::is_done`] is `false`, the span returned is guaranteed<9 to correspond to valid bounds for this input's haystack.l-<-&# use regex_automata::{Input, Span};%:<9 assert_eq!(Span { start: 0, end: 6 }, input.get_span());<-D5 5B  2$'=: Return the span as a range for this search configuration.JRԶRHE When [`Input::is_done`] is `false`, the range returned is guaranteed<Rl-<-֎R%:(R<-L5 5B  2$&;8 Return the anchored mode for this search configuration.FC If no anchored mode was set, then it defaults to [`Anchored::No`].l-<-5R)R3R!R/R=R<-d5 5łB  2$"A> Return whether this search should execute in "earliest" mode.l-<-֎R%:#R<-d5 5B  2$EB Return true if and only if this search can never return any otherd matches.JG This occurs when the start position of this search is greater than the end position of the search.l-<-֎R)R assert!(!input.is_done()); input.set_start(6);S input.set_start(7); assert!(input.is_done());<-<5 5B  2$5JG Returns true if and only if the given offset in this search's haystack63 falls on a valid UTF-8 encoded codepoint boundary.IF If the haystack is not valid UTF-8, then the behavior of this routine is unspecified.l-EB This shows where codepoint boundaries do and don't exist in validT><-֎R":'$ assert!(input.is_char_boundary(0));(% assert!(!input.is_char_boundary(1));(% assert!(!input.is_char_boundary(2));'$ assert!(input.is_char_boundary(3));(% assert!(!input.is_char_boundary(4));<-5 5B  2$ C '2g@555 5 5ggg cgcثq1ȡ5B  2$ "l;2Q4Q\ %$B 2 L#! Returns this span as a range.,5 5B ˷($GD Returns true when this span is empty. That is, when `start >= end`.D5 5B ($ԉ$! Returns the length of this span.KH This returns `0` in precisely the cases that `is_empty` returns `true`.5 5B )$-EB Returns true when the given offset is contained within this span.FC Note that an empty span contains no offsets and will always returnT false.D5 5B  $ $ +KH Returns a new span with `offset` added to this span's `start` and `end`\ values.45 5B  $ C& g@555 5 5ggg cgcثq1ȡ5B  $ $\4%,5 5SB  $?' '#1L6 6SB  $*( #\4$,6 6SB  $)  $$B A)  ʀ#$̀B * %*66 6 6B  $A+ ׂ%!66 6 6B  $, t܋Č:@= Create a new half match from a pattern ID and a byte offset.ˌB̟42  >2 7@SGD This is like [`HalfMatch::new`], but accepts a `usize` instead of aˎHE [`PatternID`]. This panics if the given `usize` is not representable as a `PatternID`.$ǏB #3  -3 "Ɛ/, Returns the ID of the pattern that matched.DA The ID of a pattern is derived from the position in which it wasˑIF originally inserted into the corresponding DFA. The first pattern hasFC identifier `0`, and each subsequent pattern is `1`, `2` and so on.<6 6B  $줕 The position of the match.ړFC If this match was produced by a forward search, then the offset isGD exclusive. If this match was produced by a reverse search, then the offset is inclusive.46 6B  $T?41 Create a new match from a pattern ID and a span.Ŝ͜B? This constructor is generic over how a span is provided. While<Q՝Rdں.! This panics if `end < start`.lƞ-؞EB This shows how to create a match for the first pattern in a regex)& object using convenient range syntax.؟<-+( use regex_automata::{Match, PatternID};/, let m = Match::new(PatternID::ZERO, 5..10);ؠ*' assert_eq!(0, m.pattern().as_usize());쇡 assert_eq!(5, m.start());䩡 assert_eq!(10, m.end());<ʡ-(B( (T(n: y: /1ت<@= Create a new match from a pattern ID and a byte offset span.BSʤ<Q싥RC@ This is like [`Match::new`], but accepts a `usize` instead of aHSʦSdں.B? This panics if `end < start` or if `pattern > PatternID::MAX`.̧lԧ-EB This shows how to create a match for the third pattern in a regex)S<- use regex_automata::Match;" let m = Match::must(3, 5..10);̩*' assert_eq!(3, m.pattern().as_usize());S䝪S<-$ߪ(BS ST (<  < )+"ի/ڧSDSڬFC originally inserted into the corresponding regex engine. The firstKH pattern has identifier `0`, and each subsequent pattern is `1`, `2` andT so on.<6 6B  $گ׮'$ The starting position of the match.<9 This is a convenience routine for `Match::span().start`.,6 6B  $ԝ%" The ending position of the match.Ȱа:7 This is a convenience routine for `Match::span().end`.6 6B $.ݱ&# Returns the match span as a range.>; This is a convenience routine for `Match::span().range()`.,6 6B  $$! Returns the span for this match.$6 6B ($63 Returns true when the span in this match is empty.GD An empty match can only be returned when the regex itself can match the empty string.D6 6B $Լ%" Returns the length of this match.ֶ޶KÕS÷6 6B Ƿ $ȷ|̉)DA Create a new set of pattern identifiers with the given capacity.HE The given capacity typically corresponds to (at least) the number of(% patterns in a compiled regex object.dں.KH This panics if the given capacity exceeds [`PatternID::LIMIT`]. This isIF impossible if you use the `pattern_len()` method as defined on any ofJG the regex engines in this crate. Namely, a regex will fail to build byHE returning an error if the number of patterns given to it exceeds theGD limit. Therefore, the number of patterns in a valid regex is always'$ a correct capacity to provide here.B D85 Clear this set such that it contains no pattern IDs.,6 6B  $.KH Return true if and only if the given pattern identifier is in this set.D6 6B  $0JG Insert the given pattern identifier into this set and return `true` if85 the given pattern ID was not previously in this set.KH If the pattern identifier is already in this set, then this is a no-op.JG Use [`PatternSet::try_insert`] for a fallible version of this routine.dں.@= This panics if this pattern set has insufficient capacity to store the given pattern ID.46 6B $jJS8SKSd0JG This returns an error if this pattern set has insufficient capacity toST6 6ÜĜŜƜǜȜ_wB  $IF Return true if and only if this set has no pattern identifiers in it.D6 6B  $IF Return true if and only if this set has the maximum number of patternIF identifiers in the set. This occurs precisely when `PatternSet::len() == PatternSet::capacity()`.JG This particular property is useful to test because it may allow one toJG stop a search earlier than you might otherwise. Namely, if a search isIF only reporting which patterns match a haystack and if you know all ofHE the patterns match at a given point, then there's no new informationHE that can be learned by continuing the search. (Because a pattern set/, does not keep track of offset information.)<6 6B $@= Returns the total number of pattern identifiers in this set.6 6B  $FC Returns the total number of pattern identifiers that may be stored in this set.HE This is guaranteed to be less than or equal to [`PatternID::LIMIT`].KH Typically, the capacity of a pattern set matches the number of patterns30 in a regex object with which you are searching.D6 6B  $(A> Returns an iterator over all pattern identifiers in this set.HE The iterator yields pattern identifiers in ascending order, startingd at zero.$6 6%%%, @ή6B  $0|1̧g@666 6 6ggg cgcثq1ȡ6B  $ (̞QQQL$'$6 6B  $-L6 6FB  $3Pʃ,L̓6 6B ׃ $܃lłٱ!έKH Returns true if and only if this anchor mode corresponds to any kind of anchored search.l-ѮٮIF This examples shows that both `Anchored::Yes` and `Anchored::Pattern`%" are considered anchored searches.ѯ<ٯ-.+ use regex_automata::{Anchored, PatternID};)& assert!(!Anchored::No.is_anchored());ΰ)& assert!(Anchored::Yes.is_anchored());>; assert!(Anchored::Pattern(PatternID::ZERO).is_anchored());<-\6 6łB Փ$*ƲIF Returns the pattern ID associated with this configuration if it is anIF anchored search for a specific pattern. Otherwise `None` is returned.l-<-.Sô˴-* assert_eq!(None, Anchored::No.pattern());.+ assert_eq!(None, Anchored::Yes.pattern());!R޵<9 assert_eq!(Some(pid), Anchored::Pattern(pid).pattern());<-<6 6łB ȶ $ɶt6̶6 6B  $|Ą.ڄ1. Create a new error value with the given kind.EB This is a more verbose version of the kind-specific constructors, e.g., `MatchError::quit`.B $%52 Returns a reference to the underlying error kind.$Ĉ6 6 6B Ɉ $ʈċ2HE Create a new "quit" error. The given `byte` corresponds to the value͉KH that tripped a search's quit condition, and `offset` corresponds to the63 location in the haystack at which the search quit.؊85 This is the same as calling `MatchError::new` with a" [`MatchErrorKind::Quit`] kind.$ˋB$Ћ 4ڋ+ÌGD Create a new "gave up" error. The given `offset` corresponds to the96 location in the haystack at which the search gave up.͍Ս8S$! [`MatchErrorKind::GaveUp`] kind.<ŽB  4ʎ2JG Create a new "haystack too long" error. The given `len` corresponds to41 the length of the haystack that was problematic.8S-* [`MatchErrorKind::HaystackTooLong`] kind.B Ǒڔ9JG Create a new "unsupported anchored" error. This occurs when the callerFC requests a search with an anchor mode that is not supported by theɓ regex engine.ߓ8S1. [`MatchErrorKind::UnsupportedAnchored`] kind.łB$%|٪Ӫ&g«@ū666 6 6ggg cgcثq1ȡ6B ɫ $ʫ Ы111T1&&&&-=-C<-",-TVD-4 -%| - 0- -B D -4-4-6 6B - $- new_capacityd--,-6 6B - $--$-6 6B - $-*-d-6 6B - $-t- /- -B D -#4-4#-6 6B #- $#-Td#-'&-D(-6 6B (- $(-)!-)-6 6B )- $)-*%-D*-6 6B *- $*-.4-4/-6 6B /- $/--/-42-D4-6 6B 4- $4--4-6-,6-6 6B 6- $6-6.-$6-6 6&&&& ?6B 6- $6-8*-d8-6 6B 8- $8-9#-g9@-9-666 6 6ggg cgcثq1ȡ6B 9- $9- 9-<'-҈QQ=-L=-$=-=%-$>-6 6҈B >- $>-1d11$1).d.B! .#.L.ц$B).. .$ .ц$ .>.4.6 6"B .",.C.\.6 6"B .",.!..ͶB $.9.\.6 6B .(,. .&,.&.B  '. 52 The configuration used for compiling a dense DFA.  JG As a convenience, [`DFA::config`] is an alias for [`Config::new`]. The IF advantage of the former is that it often lets you avoid importing the  `Config` type directly.  LI A dense DFA configuration is a simple data object that is typically usedA> with [`dense::Builder::configure`](self::Builder::configure).HE The default configuration guarantees that a search will never returnC@ a "quit" error, although it is possible for a search to fail ifC@ [`Config::starts_for_each_pattern`] wasn't enabled (which it isFC not by default) and an [`Anchored::Pattern`] mode is requested via [`Input`](crate::Input).,4=0IħTDTT%d#Ŀ<ϴ%%t-s[s[s[& &B s[[]g[][]&&& & &ggg cgcثq1ȡ&B [][]gggB ņLI A builder for constructing a deterministic finite automaton from regular expressions.*' This builder provides two main things:MJ 1. It provides a few different `build` routines for actually constructing@= a DFA from different kinds of inputs. The most convenient isNK [`Builder::build`], which builds a DFA directly from a pattern string. TheMJ most flexible is [`Builder::build_from_nfa`], which builds a DFA straight from an NFA.:7 2. The builder permits configuring a number of things.JG [`Builder::configure`] is used with [`Config`] to configure aspects ofHE the DFA and the construction process itself. [`Builder::syntax`] andNK [`Builder::thompson`] permit configuring the regex parser and Thompson NFAKH construction, respectively. The syntax and thompson configurations only.+ apply when building from a pattern string.HE This builder always constructs a *single* DFA. As such, this builderIF can only be used to construct regexes that either detect the presenceGD of a match or find the end location of a match. A single DFA cannotJG produce both the start and end of a match. For that information, use aJG [`Regex`](crate::dfa::regex::Regex), which can be similarly configuredLI using [`regex::Builder`](crate::dfa::regex::Builder). The main reason toOL use a DFA directly is if the end location of a match is enough for your useKH case. Namely, a `Regex` will construct two DFAs instead of one, since a>; second reverse DFA is needed to find the start of a match.NK Note that if one wants to build a sparse DFA, you must first build a denseKH DFA and convert that to a sparse DFA. There is no way to build a sparse+( DFA without first building a dense DFA.l-LI This example shows how to build a minimized DFA that completely disables Unicode. That is:IF * Things such as `\w`, `.` and `\b` are no longer Unicode-aware. `\w`HE and `\b` are ASCII-only while `.` matches any byte except for `\n`IF (instead of any UTF-8 encoding of a Unicode scalar value except forJG `\n`). Things that are Unicode only, such as `\pL`, are not allowed.JG * The pattern itself is permitted to match invalid UTF-8. For example,JG things like `[^a]` that match any byte except for `a` are permitted.с<Ձ-݁- .A̮/4Ȃ-ςӂ#-85 .configure(dense::Config::new().minimize(false))A9  .build(r"foo[^b]ar.*")?;2/ let haystack = b"\xFEfoo\xFFar\xE2\x98\xFF\n";ʄ0H9..ԅ؅.-<-,쏆<ІhtT~tކ4ކ䑇DIiIiIi& &B Iijgjj&&& & &ggg cgcثq1ȡ&B jj=: A dense table-based deterministic finite automaton (DFA).KH All dense DFAs have one or more start states, zero or more match statesKH and a transition table that maps the current state and the current byteKH of input to the next state. A DFA can use this information to implementMJ fast searching. In particular, the use of a dense DFA generally makes theKH trade off that match speed is the most valuable characteristic, even ifMJ building the DFA may take significant time *and* space. (More concretely,NK building a DFA takes time and space that is exponential in the size of theNK pattern in the worst case.) As such, the processing of every byte of inputNK is done with a small constant number of operations that does not vary withOL the pattern, its size or the size of the alphabet. If your needs don't lineOL up with this trade off, then a dense DFA may not be an adequate solution to your problem.52 In contrast, a [`sparse::DFA`] makes the oppositeGD trade off: it uses less space but will execute a variable number ofLI instructions per byte at match time, which makes it slower for matching.MJ (Note that space usage is still exponential in the size of the pattern in the worst case.)>; A DFA can be built using the default configuration via the0- [`DFA::new`] constructor. Otherwise, one can>; configure various aspects via [`dense::Builder`](Builder).A> A single DFA fundamentally supports the following operations: 1. Detection of a match.&# 2. Location of the end of a match.LI 3. In the case of a DFA with multiple patterns, which pattern matched isė reported as well.LI A notable absence from the above list of capabilities is the location ofIF the *start* of a match. In order to provide both the start and end ofIF a match, *two* DFAs are required. This functionality is provided by a(% [`Regex`](crate::dfa::regex::Regex). # Type parametersNK A `DFA` has one type parameter, `T`, which is used to represent state IDs,NK pattern IDs and accelerators. `T` is typically a `Vec` or a `&[u32]`. # The `Automaton` traitLI This type implements the [`Automaton`] trait, which means it can be used for searching. For example:<-I0%˪1)& let expected = HalfMatch::must(0, 8);MJ assert_eq!(Some(expected), dfa.try_search_fwd(&Input::new("foo12345"))?);.-<-/ bhA;Ж-/<4<d,t/Fjmtt& &B tɴ2 $"|/  gO#%D,<t<R/TZRR' 'BZ R )T / ʞÑ'[+-D , |ȕ $ȕ  L l 4 ܙ \ܙ Ͷқ +қ ӝ )ӝ d:/(Tn DGd:d:' 'Bn  d: -\ / ȥ ~U$FQC#%L 4 tի \ի  \ ca/Tvȥ caca' 'Bv  ca=b/QJwȥ gmp=b=b''' ' 'ggg cgcثq1ȡ'Bw  =b=b 5, ␗Թ LĒ ĩ <  )Ď ' 'B    g''' ' 'ggg cgcثq1ȡ'B   )9 sZH.^] T   - << 0{?ippʽp opopo2Ыg''' ' 'ggg cgcثq1ȡ'B   /: L  ,  6 ³ ŘgoqŘŘ''' ' 'ggg cgcثq1ȡ'B  ŘŘ  <9 An error that occurred during the construction of a DFA.  JG This error does not provide many introspection capabilities. There are 1. generally only two things you can do with it:  GD * Obtain a human readable message via its `std::fmt::Display` impl. NK * Access an underlying [`nfa::thompson::BuildError`](thompson::BuildError) OL type from its `source` method via the `std::error::Error` trait. This error KH only occurs when using convenience routines for building a DFA directly  from a pattern string.  NK When the `std` feature is enabled, this implements the `std::error::Error`T N, ;T  3]E $ 555' 'B  5g''' ' 'ggg cgcثq1ȡ'B   , >` when the `alloc` feature is enabled. For mostIF in-memory work loads, this is the most convenient type that gives theE"B best search performance. When the `alloc` feature is disabled, no̕r default type is used. # When should I use this?IF Generally speaking, if you can afford the overhead of building a fullH0E DFA for your regex, and you don't need things like capturing groups,IF then this is a good choice if you're looking to optimize for matchingJG speed. Note however that its speed may be worse than a general purposeK/H regex engine if you don't provide a [`dense::Config::prefilter`] to the underlying DFA. # Sparse DFAsFC Since a `Regex` is generic over the [`Automaton`] trait, it can beH. E used with any kind of DFA. While this crate constructs dense DFAs byK H default, it is easy enough to build corresponding sparse DFAs, and then   build a regex from them:  <  - * 3 ,  1; . // First, build a regex that uses dense DFAs.!,x ) let dense_re = Regex::new("foo[0-9]+")?;! !I F // Second, build sparse DFAs from the forward and reverse dense DFAs.". + let fwd = dense_re.forward().to_sparse()?;".M + let rev = dense_re.reverse().to_sparse()?;# #A > // Third, build a new regex from the constituent sparse DFAs.#? < let sparse_re = Regex::builder().build_from_dfas(fwd, rev);$, $K; H // A regex that uses sparse DFAs can be used just like with dense DFAs.%4 1 assert_eq!(true, sparse_re.is_match(b"foo123"));% %. -<& -& &F C Alternatively, one can use a [`Builder`] to construct a sparse DFA&G_ D more succinctly. (Note though that dense DFAs are still constructed'J G first internally, and then converted to sparse DFAs, as in the example\' above.)(<(+-(*;3(J(AY> let sparse_re = Regex::builder().build_sparse(r"foo[0-9]+")?;)KסU)+( assert!(sparse_re.is_match(b"foo123"));**.-<* -**' # Fallibility+C+JRG Most of the search routines defined on this type will _panic_ when the+JG underlying search fails. This might be because the DFA gave up because,FC it saw a quit byte, whether configured explicitly or via heuristic,KPH Unicode word boundary support, although neither are enabled by default.-GD Or it might fail because an invalid `Input` configuration is given,.74 for example, with an unsupported [`Anchored`] mode..<.GKD If you need to handle these error cases instead of allowing them to/HE trigger a panic, then the lower level [`Regex::try_search`] provides/%" a fallible API that never panics.0"l00-0?0FN.1J^G `\n` byte, and handle the error returned. This could be useful if, for1HE example, you wanted to prevent a user supplied pattern from matchingܝ2 across a line boundary.2-<2;-2=K-3G[D use regex_automata::{dfa::{self, regex::Regex}, Input, MatchError};3334;8 .dense(dfa::dense::Config::new().quit(b'\n', true))4#.4!4'0$ let input = Input::new("foo\nbar");5Hc.5Gs.6=.7..71. let got = re.try_search(&input).unwrap_err();7.88.-<8 - ȡ ( 3 AN`" P 8 P 8 #3v ##' 'B8 #3%g gIL((( ( (ggg cgcثq1ȡ(B8 ȸ!JG An iterator over all non-overlapping matches for an infallible search.OL The iterator yields a [`Match`] value until no more matches could be found.۴IF If the underlying regex engine returns an error, then a panic occurs.'ήIѵյGD * `A` represents the type of the underlying DFA that implements theĝ [`Automaton`] trait.+( The lifetime parameters are as follows:B? * `'h` represents the lifetime of the haystack being searched.>; * `'r` represents the lifetime of the regex object itself.FC This iterator can be created with the [`Regex::find_iter`] method.H\Ӹ323 3gb+m߸?A< 3ח PM 522323 gg69((( ( (ggg cgcثq1ȡ(B 32C@ A builder for a regex based on deterministic finite automatons.MJ This builder permits configuring options for the syntax of a pattern, theJG NFA construction, the DFA construction and finally the regex searchingMJ itself. This builder is different from a general purpose regex builder inϽMJ that it permits fine grain configuration of the construction process. TheFC trade off for this is complexity, and the possibility of setting aGD configuration that might not make sense. For example, there are twoԬ different UTF-8 modes:ǿ˿JG * [`syntax::Config::utf8`](crate::util::syntax::Config::utf8) controlsMJ whether the pattern itself can contain sub-expressions that match invalidT>MJ * [`thompson::Config::utf8`](crate::nfa::thompson::Config::utf8) controlsKH how the regex iterators themselves advance the starting position of the74 next search when a match with zero length is found.JG Generally speaking, callers will want to either enable all of these or̐ disable all of these.IF Internally, building a regex requires building two DFAs, where one isKH responsible for finding the end of a match and the other is responsibleHE for finding the start of a match. If you only need to detect whetherHE something matched, or only the end of a match, then you should use aGD [`dense::Builder`] to construct a single DFA, which is cheaper than building two DFAs. # Build methodsJG This builder has a few "build" methods. In general, it's the result of'$ combining the following parameters:# * Building one or many regexes.1. * Building a regex with dense or sparse DFAs.JG The simplest "build" method is [`Builder::build`]. It accepts a singleIF pattern and builds a dense DFA using `usize` for the state identifier representation.MJ The most general "build" method is [`Builder::build_many`], which permitsMJ building a regex that searches for multiple patterns simultaneously while52 using a specific state identifier representation.<9 The most flexible "build" method, but hardest to use, isKH [`Builder::build_from_dfas`]. This exposes the fact that a [`Regex`] isIF just a pair of DFAs, and this method allows you to specify those DFAsd exactly.l-LI This example shows how to disable UTF-8 mode in the syntax and the regexLI itself. This is generally what you want for matching on arbitrary bytes.<-=-->34-3236ܭ3&34̮3.3 ¯3.63 // Notice that `(?-u:[^b])` matches invalid UTF-8,85 // but the subsequent `.*` does not! Disabling UTF-8" // on the syntax permits this.@= assert_eq!(b"foo\xFFarzz", &haystack[got.unwrap().range()]);.-<-I<BN,( (B g((( ( (ggg cgcثq1ȡ(B MJ A sparse deterministic finite automaton (DFA) with variable sized states.KH In contrast to a [dense::DFA], a sparse DFA uses a more space efficientNK representation for its transitions. Consequently, sparse DFAs may use muchJG less memory than dense DFAs, but this comes at a price. In particular,EB reading the more space efficient transitions takes more work, andOL consequently, searching using a sparse DFA is typically slower than a denseD DFA.EB A sparse DFA can be built using the default configuration via theOL [`DFA::new`] constructor. Otherwise, one can configure various aspects of aNK dense DFA via [`dense::Builder`], and then convert a dense DFA to a sparse(% DFA using [`dense::DFA::to_sparse`].OL In general, a sparse DFA supports all the same search operations as a denseDUMJ Making the choice between a dense and sparse DFA depends on your specificLI work load. If you can sacrifice a bit of search time performance, then aMJ sparse DFA might be the best choice. In particular, while sparse DFAs areMJ probably always slower than dense DFAs, you may find that they are easily" fast enough for your purposes!TMJ A `DFA` has one type parameter, `T`, which is used to represent the parts?< of a sparse DFA. `T` is typically a `Vec` or a `&[u8]`.܌TLTT<-J4%˪1/0G0.-<-J/ 5\Yti(*$$$$$<$$$$<$d$,$/T ( (B \\/ Ƀ8+Տ%'L4ޖ<ޖL\|/TɃ||( (B |ɧ]TЧ/ ۧ{9Tk +-D,|$˰L˰l4ԋ\Ͷ+ԁ)Ă֯//TۧLO֯֯( (B ֯`,i,Cw6\հհtDlٱ4ٱd1t$1\1|,1AAA( (B A\g$VS9*̘1g11))) ) )ggg cgcثq1ȡ)B 11,l4,3 , >By L/4/g3kU,gg) )B, gp*rTp6 p6pqz'Hhp=?34q Dq q 76gp6g 7 7))) ) )ggg cgcثq1ȡ)Bp  7 7ys,zdFʓ},}L}|<|<|<) )By |<ܿ:7 Represents the current state of an overlapping search.KH This is used for overlapping searches since they need to know somethingOL about the previous search. For example, when multiple patterns match at theҧOL same position, this state tracks the last reported pattern so that the nextLI search knows whether to report another matching pattern or continue withMJ the search at the next position. Additionally, it also tracks which state'$ the last search call terminated in.JG This type provides little introspection capabilities. The only thing aNK caller can do is construct it and pass it around to permit search routinesJG to use it to track state, and also ask whether a match has been found.Ϋҫ?< Callers should always provide a fresh state constructed viaNK [`OverlappingState::start`] when starting a new search. Reusing state from63 a previous search may result in incorrect results.xʭbNFnd'! ȵӵڸ*ͶĶ<D}D}D}* *B D}(~g(~(~*** * *ggg cgcثq1ȡ*B (~(~.....* *B .tttt** * *B t tHE An error that can occur when computing the start state for a search.GD Computing a start state can fail for a few reasons, either based onIF incorrect configuration or even based on whether the look-behind byteKH triggers a quit state. Typically one does not need to handle this errorFC if you're using [`Automaton::start_state_forward`] (or its reverseJG counterpart), as that routine automatically converts `StartError` to aܔ [`MatchError`] for you.IF This error may be returned by the [`Automaton::start_state`] routine.NK This error implements the `std::error::Error` trait when the `std` feature| is enabled.JG This error is marked as non-exhaustive. New variants may be added in a semver compatible release.:yTgq$IF An error that occurs when a starting configuration's look-behind byte is in this DFA's quit set.yD! The quit byte that was found.$GD An error that occurs when the caller requests an anchored mode that isn't supported by the DFA.yt0- The anchored mode given that is unsupported.$łGGG* *B Gۅgۅۅ*** * *ggg cgcثq1ȡ*B ۅۅ4v>>Tl$tͶ%Ͷ///* *B /g*** * *ggg cgcثq1ȡ*B &4&7x~kq&t'4'(( ((;ԡ7t7|9,9;;=<=>,> B(Bg69*** * *ggg cgcثq1ȡ*B& |D^* *B g*** * *ggg cgcثq1ȡ*B ððððð* *B ðoooo** * *B o oԱԱԱ** * *B Ա Ա** * *B  ԴD -C   !4!ʅg*** * *ggg cgcثq1ȡ*B =\=ʅtet?B? uses internally to translate a regex's high-level intermediate># representation into an [`NFA`].>?IF The primary function of this builder is to abstract away the internal?KH representation of an NFA and make it difficult to produce NFAs are that@KH internally invalid or inconsistent. This builder also provides a way to@HE add "empty" states (which can be thought of as unconditional epsilonALI transitions), despite the fact that [`thompson::State`](nfa::State) doesAKH not have any "empty" representation. The advantage of "empty" states isBNK that they make the code for constructing a Thompson NFA logically simpler.CCNK Many of the routines on this builder may panic or return errors. GenerallyCNK speaking, panics occur when an invalid sequence of method calls were made,DOL where as an error occurs if things get too big. (Where "too big" might meanENK exhausting identifier space or using up too much heap memory in accordanceEA> with the configured [`size_limit`](Builder::set_size_limit).)FtF # OverviewFF ## Adding multiple patternsFFA> Each pattern you add to an NFA should correspond to a pair ofGJG [`Builder::start_pattern`] and [`Builder::finish_pattern`] calls, withGKH calls inbetween that add NFA states for that pattern. NFA states may beHMJ added without first calling `start_pattern`, with the exception of addingH capturing states.IēI ## Adding NFA statesIIMJ Here is a very brief overview of each of the methods that add NFA states.I%" Every method adds a single state.JJB? * [`add_empty`](Builder::add_empty): Add a state with a singleJ63 unconditional epsilon transition to another state.KHE * [`add_union`](Builder::add_union): Adds a state with unconditionalKGD epsilon transitions to two or more states, with earlier transitionsL preferred over later ones.LJG * [`add_union_reverse`](Builder::add_union_reverse): Adds a state withMGD unconditional epsilon transitions to two or more states, with laterM,) transitions preferred over earlier ones.NNK * [`add_range`](Builder::add_range): Adds a state with a single transitionNKH to another state that can only be followed if the current input byte isܭO within the range given.OHE * [`add_sparse`](Builder::add_sparse): Adds a state with two or morePJG range transitions to other states, where a transition is only followedPJG if the current input byte is within one of the ranges. All transitionsQKH in this state have equal priority, and the corresponding ranges must beQ non-overlapping.ROL * [`add_look`](Builder::add_look): Adds a state with a single *conditional*RIF epsilon transition to another state, where the condition depends on aS! limited look-around property.SJG * [`add_capture_start`](Builder::add_capture_start): Adds a state withTHE a single unconditional epsilon transition that also instructs an NFATMJ simulation to record the current input position to a specific location inUNK memory. This is intended to represent the starting location of a capturingTU group.VFC * [`add_capture_end`](Builder::add_capture_end): Adds a state withVHVWMVWLI memory. This is intended to represent the ending location of a capturingTXVXMJ * [`add_fail`](Builder::add_fail): Adds a state that never transitions toY another state.YOL * [`add_match`](Builder::add_match): Add a state that indicates a match hasYLI been found for a particular pattern. A match state is a final state withZ no outgoing transitions.ZZ-* ## Setting transitions between NFA states[[LI The [`Builder::patch`] method creates a transition from one state to the[KH next. If the `from` state corresponds to a state that supports multiple\LI outgoing transitions (such as "union"), then this adds the corresponding\NK transition. Otherwise, it sets the single transition. (This routine panics]OL if `from` corresponds to a state added by `add_sparse`, since sparse states^$! need more specialized handling.)^l^-^^IF This annotated example shows how to hand construct the regex `[a-z]+`_# (without an unanchored prefix)._<_-_-_==`"=t`A4`-``%=`LI // Before adding NFA states for our pattern, we need to tell the buildera(% // that we are starting the pattern.a=bMJ // Since we use the Pike VM below for searching, we need to add capturingbOL // states. If you're just going to build a DFA from the NFA, then capturingc&# // states do not need to be added.cC=d.+ let range = builder.add_range(Transition {dNK // We don't know the state ID of the 'next' state yet, so we just fille# // in a dummy 'ZERO' value.e4=De=eOL // This state will point back to 'range', but also enable us to move ahead.fOL // That is, this implements the '+' repetition operator. We add 'range' andg,) // then 'end' below to this alternation.g)=gJG // The final state before the match state, which serves to capture theh! // end location of the match.h96 let end = builder.add_capture_end(StateID::ZERO, 0)?;i'$ // The match state for our pattern.i# let mat = builder.add_match()?;i52 // Now we fill in the transitions between states.j! builder.patch(start, range)?;j builder.patch(range, alt)?;jGD // If we added 'end' before 'range', then we'd implement non-greedyk // matching, i.e., '+?'.k builder.patch(alt, range)?;k builder.patch(alt, end)?;k builder.patch(end, mat)?;lNK // We must explicitly finish pattern and provide the starting state ID forl // this particular pattern.m#=mMJ // Finally, when we build the NFA, we provide the anchored and unanchoredmKH // starting state IDs. Since we didn't bother with an unanchored prefixnOL // here, we only support anchored searching. Thus, both starting states areo // the same.o+( let nfa = builder.build(start, start)?;ooMJ // Now build a Pike VM from our NFA, and use it for searching. This showspHE // how we can use a regex engine without ever worrying about syntax!p(?q&?q(?q.+ let expected = Some(Match::must(0, 0..3));r/, re.captures(&mut cache, "foo0", &mut caps);r+?rr.- ; explicit capture groups, are included in the Thompson NFA.D ϢD A Only capture states corresponding to implicit capture groups areH E included. Implicit capture groups appear in every pattern implicitly5 2 and correspond to the overall match of a pattern. C @ This is useful when one only cares about the overall match of aF C pattern. By excluding capture states from explicit capture groups,I F one might be able to reduce the memory usage of a multi-pattern regexK H substantially if it was otherwise written to have many explicit capture\ئ  groups.Ujm$ 9 6 No capture states are compiled into the Thompson NFA. J G This is useful when capture states are either not needed (for example,J G if one is only trying to build a DFA) or if they aren't supported (forڨ  example, a reverse NFA). l # Warning 7 4 Callers must be exceedingly careful when using this٩8 5 option. In particular, not all regex engines support> ; reporting match spans when using this option (for example,٪7 4 [`PikeVM`](crate::nfa::thompson::pikevm::PikeVM) orQ N [`BoundedBacktracker`](crate::nfa::thompson::backtrack::BoundedBacktracker)). < 9 Perhaps more confusingly, using this option with such an? < engine means that an `is_match` routine could report `true`? < when `find` reports `None`. This is generally not something? < that _should_ happen, but the low level control provided by!  this crate makes it possible. K H Similarly, any regex engines (like [`meta::Regex`](crate::meta::Regex))H E should always return `None` from `find` routines when this option is˯E B used, even if _some_ of its internal engines could find the matchE B boundaries. This is because inputs from user data could influence߰I F engine selection, and thus influence whether a match is found or not.I F Indeed, `meta::Regex::find` will always return `None` when configured  with this option.AVY, ,B    rgrr,,, , ,ggg cgcثq1ȡ,B  rr I F A builder for compiling an NFA from a regex's high-level intermediate̅  representation (HIR). L I This compiler provides a way to translate a parsed regex pattern into anO L NFA state graph. The NFA state graph can either be used directly to executeN K a search (e.g., with a Pike VM), or it can be further used to build a DFA. O L This compiler provides APIs both for compiling regex patterns directly from? < their concrete syntax, or via a [`regex_syntax::hir::Hir`]. @ = This compiler has various options that may be configured via!  [`thompson::Config`](Config). M J Note that a compiler is not the same as a [`thompson::Builder`](Builder).I F A `Builder` provides a lower level API that is uncoupled from a regexL I pattern's concrete syntax or even its HIR. Instead, it permits stitching7 4 together an NFA by hand. See its docs for examples. / , # Example: compilation from concrete syntax O L This shows how to compile an NFA from a pattern string while setting a sizeO L limit on how big the NFA is allowed to be (in terms of bytes of heap used). < -̏ -- * nfa::thompson::{NFA, pikevm::PikeVM},t A4 - ; ?C ? ( ?& ?( ?. ?2 ?+ ? . -< - # # Example: compilation from HIR N K This shows how to hand assemble a regular expression via its HIR, and then$ ! compile an NFA directly from it. < -F ?E ? ; ?) ?) ?) ?) ?L ? ; ?F ? ( ?& ?( ?. ?2 ?+ ? . -< -D ;h 4 t 4 ̌ <  T  T # \ III, ,B  Ig,,, , ,ggg cgcثq1ȡ,B  죔 \ Jǔ ,Ҕ         , ,B      g  ,,, , ,ggg cgcثq1ȡ,B     dŞ Œ/iaҞ ܞ <ܞ ݤ@ , @| 4  Œg  ,,, ,Œ ,ggg cgcثq1ȡ,B    ğ L˟ Yڝ~۟ D۟  T ޟ, ,B  g,,, , ,ggg cgcثq1ȡ,B  | D g՗ Š ,Š  $ , ,B  g,,, , ,ggg cgcثq1ȡ,B  ̞  ً^ L , <͡ ͡ , ,B  .g..,,, , ,ggg cgcثq1ȡ,B  ..QIF An error that can occurred during the construction of a thompson NFA.JT1TGTIF * Access an underlying [`regex_syntax::Error`] type from its `source`OL method via the `std::error::Error` trait. This error only occurs when usingLI convenience routines for building an NFA directly from a pattern string.IF Otherwise, errors typically occur when a limit has been breached. ForJG example, if the total heap usage of the compiled NFA exceeds the limitIF set by [`Config::nfa_size_limit`](crate::nfa::thompson::Config), then building the NFA will fail.TPݵe$ Õ, ,B g,,, , ,ggg cgcثq1ȡ,B t ÕٽѬ>ȏÕBÕBÕBu   u   u u  ȏÕBW  ȏu u   u u  ÕBW  6v 6v   6v 6v  ÕBW | Õd, d, l Õd, d,  Õd,  ÕT, Ծ Õ02 ÕÕ, ,ÕÕB Õg,,, ,Õ ,ggg cgcثq1ȡ,B ! \" b.Z$ 4$  L% %  /'/'/', ,B!  /'k ,k  !Y k k 4k  m. \m  ppʽp opopo2ЫΚp ,p   r! 4r  'g''--- - -ggg cgcثq1ȡ-Bj  ''dà ,ʃ   QUփ \փ   4  8,8,8,- -B  8,L-L-L-B Դ |  Dp˚ ՚ \՚  @- 4  ppʽp opopo2Ы 4  @-g----- - -ggg cgcثq1ȡ-B  -- T  Κ )iZ>D $  l $  000Κ0Κ00- -ΚΚB  0  Κ̊ t ӥ(d# <#  |# D#  Ԁ% %  222- -B  22g22--- - -ggg cgcثq1ȡ-B  22% %  [?Sd' <'  ( (  d) )  c5c5c5- -B%  c566g6666--- - -ggg cgcثq1ȡ-B%  6666888B @ l@ _F$wRdB exp)` or `(?exp)` - A named capturing group.( % * `(?:exp)` - A non-capturing group.9 6 * `(?i:exp)` - A non-capturing group that sets flags. B ? Only the first two forms are said to be _capturing_. CapturingG D means that the last position at which they match is reportable. TheJ G [`Captures`](crate::util::captures::Captures) type provides convenientM J access to the match positions of capturing groups, which includes looking& # up capturing groups by their name.   # Byte oriented J G This NFA is byte oriented, which means that all of its transitions areL I defined on bytes. In other words, the alphabet of an NFA consists of the  256 different byte values. G D While DFAs nearly demand that they be byte oriented for performanceN K reasons, an NFA could conceivably be *Unicode codepoint* oriented. Indeed,M J a previous version of this NFA supported both byte and codepoint orientedO L modes. A codepoint oriented mode can work because an NFA fundamentally usesK H a sparse representation of transitions, which works well with the large' $ sparse space of Unicode codepoints. J G Nevertheless, this NFA is only byte oriented. This choice is primarilyJ G driven by implementation simplicity, and also in part memory usage. InI F practice, performance between the two is roughly comparable. However,M J building a DFA (including a hybrid DFA) really wants a byte oriented NFA.L I So if we do have a codepoint oriented NFA, then we also need to generateH E byte oriented NFA in order to build an hybrid NFA/DFA. Thus, by onlyO L generating byte oriented NFAs, we can produce one less NFA. In other words,N K if we made our NFA codepoint oriented, we'd need to *also* make it supportM J a byte oriented mode, which is more complicated. But a byte oriented modeܬ  can support everything.   # Differences with DFAs M J At the theoretical level, the precise difference between an NFA and a DFAL I is that, in a DFA, for every state, an input symbol unambiguously refers J G to a single transition _and_ that an input symbol is required for each L I transition. At a practical level, this permits DFA implementations to be!N K implemented at their core with a small constant number of CPU instructions!M J for each byte of input searched. In practice, this makes them quite a bit"O L faster than NFAs _in general_. Namely, in order to execute a search for any#K H Thompson NFA, one needs to keep track of a _set_ of states, and execute#J G the possible transitions on all of those states for each input symbol.$N K Overall, this results in much more overhead. To a first approximation, one$E B can expect DFA searches to be about an order of magnitude faster.% %K H So why use an NFA at all? The main advantage of an NFA is that it takes&N K linear time (in the size of the pattern string after repetitions have been&M J expanded) to build and linear memory usage. A DFA, on the other hand, may'I F take exponential time and/or space to build. Even in non-pathological'O L cases, DFAs often take quite a bit more memory than their NFA counterparts,(L I _especially_ if large Unicode character classes are involved. Of course,)K H an NFA also provides additional capabilities. For example, it can match)J G Unicode word boundaries on non-ASCII text and resolve the positions of*  capturing groups.* *O L Note that a [`hybrid::regex::Regex`](crate::hybrid::regex::Regex) strikes a+O L good balance between an NFA and a DFA. It avoids the exponential build time+M J of a DFA while maintaining its fast search time. The downside of a hybrid,O L NFA/DFA is that in some cases it can be slower at search time than the NFA.,M J (It also has less functionality than a pure NFA. It cannot handle Unicode-K H word boundaries on non-ASCII text and cannot resolve capturing groups.). l. -. .O L This shows how to build an NFA with the default configuration and execute a.  search using the Pike VM./ \d{4})-(?P\d{2})-(?P\d{2})")?;5& =6 6; 8 let haystack = "2012-03-14, 2013-01-01 and 2014-07-05";6. + let all: Vec = vm.captures_iter(6' $ &mut cache, haystack.as_bytes()7 ).collect();7, ) // There should be a total of 3 matches.7  assert_eq!(3, all.len());70 - // The year from the second match is '2013'.86 3 let span = all[1].get_group_by_name("y").unwrap();8( % assert_eq!("2013", &haystack[span]);9 9. -<9 -9 9G D This example shows that only the last match of a capturing group is:I F reported, even if it had to match multiple times for an overall matchl: to occur.: <: -:> E; ;) & let re = PikeVM::new(r"([a-z]){4}")?;;& ?<( ?< ܰ<  let haystack = b"quux";<1 . re.captures(&mut cache, haystack, &mut caps);< D=: 7 assert_eq!(Some(Span::from(3..4)), caps.get_group(1));= =. -<> -> ar Bv 7 B TD  GGG- -B>  G , { u' 4   t     ܸ l    T   t  ̩ d   \  |ދ Lދ  T $  l <  ڏ dڏ  Ǒ dǑ      d  ^^^B t   A state in an NFA. M J In theory, it can help to conceptualize an `NFA` as a graph consisting ofM J `State`s. Each `State` contains its complete set of outgoing transitions. G D In practice, it can help to conceptualize an `NFA` as a sequence ofN K instructions for a virtual machine. Each `State` says what to do and where| to go next. L I Strictly speaking, the practical interpretation is the most correct one,I F because of the [`Capture`](State::Capture) state. Namely, a `Capture`N K state always forwards execution to another state unconditionally. Its onlyI F purpose is to cause a side effect: the recording of the current inputH E position at a particular location in memory. In this sense, an `NFA`I F has more power than a theoretical non-deterministic finite automaton. M J For most uses of this crate, it is likely that one may never even need toL I be aware of this type at all. The main use cases for looking at `State`sN K directly are if you need to write your own search implementation or if you0 - need to do some kind of analysis on the NFA., @vyrQBBL J G A state with a single transition that can only be taken if the current3 0 input symbol is in a particular range of bytes. / , The transition from this state to the next., v K H A state with possibly many transitions represented in a sparse fashion.J G Transitions are non-overlapping and ordered lexicographically by inputT  range. J G In practice, this is used for encoding UTF-8 automata. Its presence isG D primarily an optimization that avoids many additional unconditionalF C epsilon transitions (via [`Union`](State::Union) states), and thusG D decreases the overhead of traversing the NFA. This can improve both, ) matching time and DFA construction time.v  v v BX 'w @ = A dense representation of a state with multiple transitions.'w  'w 'w BX $ ? < A conditional epsilon transition satisfied via some sort ofC @ look-around. Look-around is limited to anchor and word boundary| assertions. I F Look-around states are meant to be evaluated while performing epsilonJ G closure (computing the set of states reachable from a particular stateJ G via only epsilon transitions). If the current position in the haystackH E satisfies the look-around assertion, then you're permitted to follow  that epsilon transition.=T B ? The look-around assertion that must be satisfied before movingt to `next`.$ l > ; The state to transition to if the look-around assertion ist satisfied.$ , F C An alternation such that there exists an epsilon transition to allG D states in `alternates`, where matches found via earlier transitions) & are preferred over later transitions. E B An ordered sequence of unconditional epsilon transitions to otherB ? states. Transitions earlier in the sequence are preferred over& # transitions later in the sequence.T \ E B An alternation such that there exists precisely two unconditionalJ G epsilon transitions, where matches found via `alt1` are preferred over  matches found via `alt2`. G D This state exists as a common special case of Union where there areG D only two alternates. In this case, we don't need any allocations toE B represent the state. This saves a bit of memory and also saves an5 2 additional memory access when traversing the NFA.l B ? An unconditional epsilon transition to another NFA state. This  is preferred over `alt2`.$ l E B An unconditional epsilon transition to another NFA state. MatchesF C reported via this transition should only be reported if no matches# were found by following `alt1`.$ < 3 0 An empty state that records a capture location. I F From the perspective of finite automata, this is precisely equivalentE B to an unconditional epsilon transition, but serves the purpose ofۀJ G instructing NFA simulations to record additional state when the finite9 6 state machine passes through this epsilon transition. D A `slot` in this context refers to the specific capture group slotE B offset that is being recorded. Each capturing group has two slotsF C corresponding to the start and end of the matching portion of thatT΃ V݃ J G The pattern ID and capture group index are also included in this stateF C in case they are useful. But mostly, all you'll need is `next` and\  `slot`.clڅ 0 - The state to transition to, unconditionally.$څ  0 - The pattern ID that this capture belongs to.T  ɆG D The capture group index that this capture belongs to. Capture groupB ? indices are local to each pattern. For example, when capturingB ? groups are enabled, every pattern has a capture group at indexD G\ ڋ B ? The slot index for this capture. Every capturing group has two@ = slots: one for the start haystack offset and one for the endC @ haystack offset. Unlike capture group indices, slot indices areF C global across all patterns in this NFA. That is, each slot belongsA > to a single pattern, but there is only one slot at index `i`.$ڋ $ H E A state that cannot be transitioned out of. This is useful for casesČK H where you want to prevent matching from occurring. For example, if yourG D regex parser permits empty character classes, then one could chooseF C a `Fail` state to represent them. (An empty character class can beJ G thought of as an empty set. Since nothing is in an empty set, they can  never match anything.), J G A match state. There is at least one such occurrence of this state for2 / each regex that can match that is in this NFA.ސ 乐  The matching pattern ID.Tސ P`P`P`- -B  P`ggggg- -B  gJhJhJhJh-- - -B  Jh Jh ? < A sequence of transitions used to represent a sparse state. L I This is the primary representation of a [`Sparse`](State::Sparse) state.K H It corresponds to a sorted sequence of transitions with non-overlappingL I byte ranges. If the byte at the current position in the haystack matchesI F one of the byte ranges, then the finite state machine should take the  corresponding transition. +Yy" 7 4 The sorted sequence of non-overlapping transitions.\ zzz- -B  z{g{{--- - -ggg cgcثq1ȡ-B  {{f|f|f|f|f|- -B  f|}}}}-- - -B  } }ܸ > ; A sequence of transitions used to represent a dense state. M J This is the primary representation of a [`Dense`](State::Dense) state. ItL I provides constant time matching. That is, given a byte in a haystack andL I a `DenseTransitions`, one can determine if the state matches in constantL  time. H E This is in contrast to `SparseTransitions`, whose time complexity isO L necessarily bigger than constant time. Also in contrast, `DenseTransitions`- * usually requires (much) more heap memory. >, H E A dense representation of this state's transitions on the heap. Thisԧ  always has length 256.\ - -B  g--- - -ggg cgcثq1ȡ-B  ҅҅҅҅҅- -B  ҅-- - -B    ) & A single transition to another state. L I This transition may only be followed if the current byte in the haystack4 1 falls in the inclusive range of bytes specified.T W l * ' The inclusive start of the byte range., \ ( % The inclusive end of the byte range.  1 . The identifier of the state to transition to.$ oooooo. .B  o  . .B  nhnn.. . .B nin nnnn.. . .B   ԩ / , An iterator over all pattern IDs in an NFA. 2 / This iterator is created by [`NFA::patterns`]. L I The lifetime parameter `'a` refers to the lifetime of the NFA from which& # this pattern iterator was created.\ /#U"    * <  !!!ѧo$z - DgDD... . .ggg cgcثq1ȡ.B  DD@ L@ ߧ4B 4B  ٭C $C  ٭D" TD  E$ \E  İ̕G TG  H dH  ߧ. .ߧߧB@  dH ,H  ֨)7NJ \J  ȷ֨. .֨֨BH  K TK  6%1YL ,L  L [\x00-`]t 1 => [a-z]  2 => [{-\xFF]t 3 => [EOI]< - E B Where EOI is the special sentinel value that is always in its own  singleton equivalence class.$ ~_Bkx 7 B D  ::::::. .B  :  . .B  ````.. . .B  ` `ȴȴȴ.. . .B  ȴ ȴ.. . .B   l D  4MrBBx x   x x  BY   y  y    y  y  BY"  IIIIII. .B  I  ƶƶƶƶƶ. .B  ƶ    .. . .B     .. . .B   .. . .B   Yy =: 7 A representation of byte oriented equivalence classes.= =N K This is used in a DFA to reduce the size of the transition table. This can>O L have a particularly large impact not only on the total size of a dense DFA,>  but also on compile times.? ?L I The essential idea here is that the alphabet of a DFA is shrunk from the?L I usual 256 distinct byte values down to a set of equivalence classes. The@N K guarantee you get is that any byte belonging to the same equivalence classAJ G can be treated as if it were any other byte in the same class, and theA' $ result of a search wouldn't change.A lA -B B6 3 This example shows how to get byte classes from anBN K [`NFA`](crate::nfa::thompson::NFA) and ask for the class of various bytes.C C C" AC% AD8 ϸAD5 AE ҹAE5 AE E. -E!!A> let re = PikeVM::new(r"^([0-9]{4})-([0-9]{2})-([0-9]{2})$")?;"J>""52 re.captures(&mut cache, "2010-03-14", &mut caps);#D#:7 assert_eq!(Some(Span::from(0..4)), caps.get_group(1));#:7 assert_eq!(Some(Span::from(5..7)), caps.get_group(2));$;8 assert_eq!(Some(Span::from(8..10)), caps.get_group(3));$$.-<%-%%%" # Example: named capturing groups%%IF This example is like the one above, but leverages the ability to name&=: capturing groups in order to make the code a bit clearer:&<&-&>E''PM let re = PikeVM::new(r"^(?P[0-9]{4})-(?P[0-9]{2})-(?P[0-9]{2})$")?;'J>((5Z(D)DA assert_eq!(Some(Span::from(0..4)), caps.get_group_by_name("y"));)DA assert_eq!(Some(Span::from(5..7)), caps.get_group_by_name("m"));*EB assert_eq!(Some(Span::from(8..10)), caps.get_group_by_name("d"));**.-<+-D+͸k-T- .. >,> {͸{{/ /͸͸B+ {ׄ"@= An iterator over all capturing groups in a `Captures` value.IF This iterator includes capturing groups that did not participate in aKH match. See the [`Captures::iter`] method documentation for more detailsʃ and examples.܃HE The lifetime parameter `'a` refers to the lifetime of the underlying `Captures` value.ב˰8$ F7, << 0{?i/ /BȄ g?B/// / /ggg cgcثq1ȡ/Bτ z ։FC Represents information about capturing groups in a compiled regex.LI The information encapsulated by this type consists of the following. For each pattern:LI * A map from every capture group name to its corresponding capture groupTы index.܋MJ * A map from every capture group index to its corresponding capture groupL name.LI * A map from capture group index to its corresponding slot index. A slotNK refers to one half of a capturing group. That is, a capture slot is eitherЍJG the start or end of a capturing group. A slot is usually the mechanismMJ by which a regex engine records offsets for each capturing group during a\ԗ7IF A `GroupInfo` uses reference counting internally and is thus cheap toTÏ clone.Ώҏ*' # Mapping from capture groups to slotsKH One of the main responsibilities of a `GroupInfo` is to build a mapping͐MJ from `(PatternID, u32)` (where the `u32` is a capture index) to somethingGD called a "slot." As mentioned above, a slot refers to one half of aGD capturing group. Both combined provide the start and end offsets of30 a capturing group that participated in a match.ߒMJ **The mapping between group indices and slots is an API guarantee.** ThatDA is, the mapping won't change within a semver compatible release.IF Slots exist primarily because this is a convenient mechanism by whichĔGD regex engines report group offsets at search time. For example, theKH [`nfa::thompson::State::Capture`](crate::nfa::thompson::State::Capture)ؕNK NFA state includes the slot index. When a regex engine transitions throughOL this state, it will likely use the slot index to write the current haystackKH offset to some region of memory. When a match is found, those slots are×NK then reported to the caller, typically via a convenient abstraction like a [`Captures`] value.NK Because this crate provides first class support for multi-pattern regexes,HE and because of some performance related reasons, the mapping betweenƙMJ capturing groups and slots is a little complex. However, in the case of aMJ single pattern, the mapping can be described very simply: for all captureNK group indices `i`, its corresponding slots are at `i * 2` and `i * 2 + 1`.JG Notice that the pattern ID isn't involved at all here, because it onlyB? applies to a single-pattern regex, it is therefore always `0`.ÜMJ In the multi-pattern case, the mapping is a bit more complicated. To talkEB about it, we must define what we mean by "implicit" vs "explicit"ם capturing groups:KH * An **implicit** capturing group refers to the capturing group that isKH present for every pattern automatically, and corresponds to the overallJG match of a pattern. Every pattern has precisely one implicit capturingԟNK group. It is always unnamed and it always corresponds to the capture groupt index `0`.HE * An **explicit** capturing group refers to any capturing group thatJG appears in the concrete syntax of the pattern. (Or, if an NFA was handơOL built without any concrete syntax, it refers to any capturing group with an䖢 index greater than `0`.) Some examples:ʢ΢HE * `\w+` has one implicit capturing group and zero explicit capturing\W<9 * `(\w+)` has one implicit group and one explicit group.MJ * `foo(\d+)(?:\pL+)(\d+)` has one implicit group and two explicit groups.EB Turning back to the slot mapping, we can now state it as follows:KH * Given a pattern ID `pid`, the slots for its implicit group are alwaysȥ# at `pid * 2` and `pid * 2 + 1`.EB * Given a pattern ID `0`, the slots for its explicit groups start&# at `group_info.pattern_len() * 2`.٦KH * Given a pattern ID `pid > 0`, the slots for its explicit groups startNK immediately following where the slots for the explicit groups of `pid - 1`D end.MJ In particular, while there is a concrete formula one can use to determineϨJG where the slots for the implicit group of any pattern are, there is noJG general formula for determining where the slots for explicit capturingKH groups are. This is because each pattern can contain a different numbert of groups.ĪJG The intended way of getting the slots for a particular capturing groupDA (whether implicit or explicit) is via the [`GroupInfo::slot`] orԫ  [`GroupInfo::slots`] method.JG See below for a concrete example of how capturing groups get mapped toTĬ slots.ϬlӬ-FF7<-ɭKH use regex_automata::util::{captures::GroupInfo, primitives::PatternID};#F FޮF2F&FD̯?կ,F&F96 // We can query the number of groups for any pattern.63 assert_eq!(2, info.group_len(PatternID::must(0)));63 assert_eq!(1, info.group_len(PatternID::must(1)));ѱ63 assert_eq!(5, info.group_len(PatternID::must(2)));63 assert_eq!(3, info.group_len(PatternID::must(3)));1. // An invalid pattern always has zero groups.85 assert_eq!(0, info.group_len(PatternID::must(999)));ĪFó$FHE // We can map a group index for a particular pattern to its name, if // one exists.ȴA> assert_eq!(Some("foo"), info.to_name(PatternID::must(3), 2));:7 assert_eq!(None, info.to_name(PatternID::must(2), 4));ŵ(% // Or map a name to its group index.B? assert_eq!(Some(1), info.to_index(PatternID::must(0), "foo"));B? assert_eq!(Some(2), info.to_index(PatternID::must(3), "foo"));.-<-30 # Example: mapping from capture groups to slotsJG This example shows the specific mapping from capture group indices forLI each pattern to their corresponding slots. The slot values shown in this,) example are considered an API guarantee.<-K[#F FѺF2F&FD?Ȼ̻A> // We first show the slots for each pattern's implicit group.@= assert_eq!(Some((0, 1)), info.slots(PatternID::must(0), 0));ϼ@= assert_eq!(Some((2, 3)), info.slots(PatternID::must(1), 0));@= assert_eq!(Some((4, 5)), info.slots(PatternID::must(2), 0));ѽ@= assert_eq!(Some((6, 7)), info.slots(PatternID::must(3), 0));C@ // And now we show the slots for each pattern's explicit group.ھ@= assert_eq!(Some((8, 9)), info.slots(PatternID::must(0), 1));B? assert_eq!(Some((10, 11)), info.slots(PatternID::must(2), 1));޿B? assert_eq!(Some((12, 13)), info.slots(PatternID::must(2), 2));B? assert_eq!(Some((14, 15)), info.slots(PatternID::must(2), 3));B? assert_eq!(Some((16, 17)), info.slots(PatternID::must(2), 4));B? assert_eq!(Some((18, 19)), info.slots(PatternID::must(3), 1));B? assert_eq!(Some((20, 21)), info.slots(PatternID::must(3), 2));LI // Asking for the slots for an invalid pattern ID or even for an invalidKH // group index for a specific pattern will return None. So for example,NK // you're guaranteed to not get the slots for a different pattern than the // one requested.85 assert_eq!(None, info.slots(PatternID::must(5), 0));85 assert_eq!(None, info.slots(PatternID::must(1), 1));.-<-LL&߷Bz 8 B     / /B  e!ge!e!/// / /ggg cgcثq1ȡ/B e!e!"""B t 5&E37*\ ߤ"lߤ )l d bgbb/// / /ggg cgcثq1ȡ/B bbeeeB ̄85 An error that may occur when building a `GroupInfo`.DA Building a `GroupInfo` does a variety of checks to make sure theNK capturing groups satisfy a number of invariants. This includes, but is notKH limited to, ensuring that the first capturing group is unnamed and thatA> there are no duplicate capture groups for a specific pattern.tқګ!Ĥ$ |||/ /B |E}gE}E}/// / /ggg cgcثq1ȡ/B E}E} zl*&|  l < t< l <  < L < d$ v~v~v~/ /B v~g/// / /ggg cgcثq1ȡ/B $MJ An iterator over capturing groups and their names for a specific pattern.=: This iterator is created by [`GroupInfo::pattern_names`].IF The lifetime parameter `'a` refers to the lifetime of the `GroupInfo`Æ)& from which this iterator was created.Z]+ ppʽp opopo2Ы6660 0B 6g14000 0 0ggg cgcثq1ȡ0B Վ ŌHE An iterator over capturing groups and their names for a `GroupInfo`.96 This iterator is created by [`GroupInfo::all_names`].̍ЍI[)[?jUT  $ \ ؏?,؏ řǙșəʙ˙G*R{9 Zg000 0 0ggg cgcثq1ȡ0B͎ 6{ `.!L"I This boxed closure is unlikely to be inlined _and_ it infects the public"N"K API in subtle ways. Namely, unless you declare the closure as implementing#J"G `Send` and `Sync`, then the resulting iterator type won't implement it#H"E either. But there are practical issues with requiring the closure to$J"G implement `Send` and `Sync` that result in other API complexities that$9"6 are beyond the scope of this already long exposition.%I"F 4. Some regex engines expose more complex match information than just%J"G "which pattern matched" and "at what offsets." For example, the PikeVM&I"F exposes match spans for each capturing group that participated in the'K"H match. In such cases, it can be quite beneficial to reuse the capturing'M"J group allocation on subsequent searches. A proper iterator doesn't permit(O"L this API due to its interface, so it's useful to have something a bit lower(K"H level that permits callers to amortize allocations while also reusing a)B"? shared implementation of iteration. (See the documentation for*L"I [`Searcher::advance`] for an example of using the "advance" API with thed*" PikeVM.)*"*J"G What this boils down to is that there are "advance" APIs which require+M"J handing a closure to it for every call, and there are also APIs to create+F"C iterators from a closure. The former are useful for _implementing_,L"I iterators or when you need more flexibility, while the latter are useful-9"6 for conveniently writing custom iterators on-the-fly.-"-&"# # Example: iterating with captures-"-J"G Several regex engines in this crate over convenient iterator APIs over.L"I [`Captures`] values. To do so, this requires allocating a new `Captures`/K"H value for each iteration step. This can perhaps be more costly than you/G"D might want. Instead of implementing your own iterator to avoid that0J"G cost (which can be a little subtle if you want to handle empty matches0=": correctly), you can use this `Searcher` to do it for you:1"<1"-̷1"-1&"ӀA1"H2" Input, Span,42"-2"24"1 let re = PikeVM::new("foo(?P[0-9]+)")?;2'"$ let haystack = "foo1 foo12 foo123";3"3("?3&"?3":4;"8 let mut searcher = Searcher::new(Input::new(haystack));42"/ while let Some(_) = searcher.advance(|input| {40"H5"HD5" }) {5K"H // The unwrap is OK since 'numbers' matches if the pattern matches.6A"> matches.push(caps.get_group_by_name("numbers").unwrap());,6"06"36" Span::from(3..4),ԗ7" Span::from(8..10),ܲ7" Span::from(14..17),<7"37"7."-<8"-D8"2PM 528"9",9";"t;"Ͷ2km0 0B8" 22g000 0 0ggg cgcثq1ȡ0B8" 2| O"L An iterator over all non-overlapping half matches for an infallible search."L"I The iterator yields a [`HalfMatch`] value until no more matches could beT" found."'"ήI"D"I"9"I"6"I"J"IN"ݱI"I"H"E This iterator is created by [`Searcher::into_half_matches_iter`] and4"1 then calling [`TryHalfMatchesIter::infallible`].|"2C "%pnB| `B 2"{}"2Cg"g000 0 0ggg cgcثq1ȡ0B" 2} J"U"O"U"'"ήI"D"I"9"I"6"I"J"IN"ݱIł"Iڂ"ނC"@ This iterator is created by [`Searcher::into_matches_iter`] and0"- then calling [`TryMatchesIter::infallible`].\"2C "ijB} `B 2"{}"2C\"g|~000 0 0ggg cgcثq1ȡ0B܃" 2W} K"H An iterator over all non-overlapping captures for an infallible search."K"H The iterator yields a [`Captures`] value until no more matches could beT"ۑ\ʝ"Ν'"ήI"D"I"Þ9"I"6"I"J"IN"ݱI֠"I"I"F This iterator is created by [`Searcher::into_captures_iter`] and then,") calling [`TryCapturesIter::infallible`].""d"2C "eݪ9`BW} `B 2"{}"R2C\"g|~RR000 0 0ggg cgcثq1ȡ0B" 2RRl A look-around assertion.HE An assertion matches at a position between characters in a haystack. OL Namely, it does not actually "consume" any input as most parts of a regular MJ expression do. Assertions are a way of stating that some property must be /, true at a particular point during matching.  2/ For example, `(?m)^[a-z]+$` is a pattern that:  LI * Scans the haystack for a position at which `(?m:^)` is satisfied. That LI occurs at either the beginning of the haystack, or immediately following  a `\n` character. 30 * Looks for one or more occurrences of `[a-z]`. KH * Once `[a-z]+` has matched as much as it can, an overall match is only 41 reported when `[a-z]+` stops just before a `\n`.HE So in this case, `abc` and `\nabc\n` match, but `\nabc1\n` does not.MJ Assertions are also called "look-around," "look-behind" and "look-ahead."NK Specifically, some assertions are look-behind (like `^`), other assertionsNK are look-ahead (like `$`) and yet other assertions are both look-ahead and look-behind (like `\b`).Ԃ # Assertions in an NFAIF An assertion in a [`thompson::NFA`](crate::nfa::thompson::NFA) can beNK thought of as a conditional epsilon transition. That is, a matching engineJG like the [`PikeVM`](crate::nfa::thompson::pikevm::PikeVM) only permitsGD moving through conditional epsilon transitions when their conditionIF is satisfied at whatever position the `PikeVM` is currently at in thelNKH How assertions are handled in a `DFA` is trickier, since a DFA does notMJ have epsilon transitions at all. In this case, they are compiled into theOL automaton itself, at the expense of more states than what would be required̓ without an assertion.$#`E                              ,KH Match the beginning of text. Specifically, this matches at the startingԾ position of the input.C@ Match the end of text. Specifically, this matches at the endingԹ\auw<IF Match the beginning of a line or the beginning of text. Specifically,JG this matches at the starting position of the input, or at the position+( immediately following a `\n` character.,JG Match the end of a line or the end of text. Specifically, this matchesDA at the end position of the input, or at the position immediately preceding a `\n` character.tLI޼\J\JG immediately following either a `\r` or `\n` character, but never after a `\r` when a `\n` follows.O< J\D\KH preceding a `\r` or `\n` character, but never before a `\n` when a `\r`  precedes it.*L" GD Match an ASCII-only word boundary. That is, this matches a position!B? where the left adjacent character and right adjacent character!GD correspond to a word and non-word or a non-word and word character.  |#"41 Match an ASCII-only negation of a word boundary.EY[\%#IF Match a Unicode-aware word boundary. That is, this matches a position#B\$G\v%%63 Match a Unicode-aware negation of a word boundary.G[]t'%KH Match the start of an ASCII-only word boundary. That is, this matches a&JG position at either the beginning of the haystack or where the previous'KH character is not a word character and the following character is a wordt' character.14qd*(GD Match the end of an ASCII-only word boundary. That is, this matches(FC a position at either the end of the haystack or where the previous)KH character is a word character and the following character is not a wordt*\ "L,*GD Match the start of a Unicode word boundary. That is, this matches a*J\+K\t,\'t.,EB Match the end of a Unicode word boundary. That is, this matches a-DA position at either the end of the haystack or where the previous-K\t.\0.FC Match the start half of an ASCII-only word boundary. That is, this/KH matches a position at either the beginning of the haystack or where the//, previous character is not a word character.20DA Match the end half of an ASCII-only word boundary. That is, this1EB matches a position at either the end of the haystack or where the10- following character is not a word character.42JG Match the start half of a Unicode word boundary. That is, this matches3C@ a position at either the beginning of the haystack or where the3/\54HE Match the end half of a Unicode word boundary. That is, this matches4GD a position at either the end of the haystack or where the following5&# character is not a word character.n I!I!I!I!  I!I!0 0B  I! ! g !!000 0 0ggg cgcثq1ȡ0B  !!%  %%%%0 0B  %L%L%  L%L%01 0 1B  L% L%[Y@= LookSet is a memory-efficient set of look-around assertions.ZZGD This is useful for efficiently tracking look-around assertions. ForZOL example, a [`thompson::NFA`](crate::nfa::thompson::NFA) provides propertiesܯ[ that return `LookSet`s.<\m91 l`\IF The underlying representation this set is exposed to make it possible\A> to store it somewhere efficiently. The representation is that]<9 of a bitset, where each assertion occupies bit `i` where] `i = Look::as_repr()`.^^HE Note that users of this internal representation must permit the full^EB range of `u16` values to be represented. For example, even if the_KH current implementation only makes use of the 10 least significant bits,_?< it may use more bits in a future semver compatible release.$`  }>}>}>}>  }>}>1 1B [ }> [?  ??B  w?  w?w?w?w?1 1B [ w?@@  @@11 1 1B [ @ @A> An iterator over all look-around assertions in a [`LookSet`].2/ This iterator is created by [`LookSet::iter`].\QBD| d  zb  zbzb1 1B  zbb g bb111 1 1ggg cgcثq1ȡ1B  bbҳ)& A matcher for look-around assertions.ͫѫNK This matcher permits configuring aspects of how look-around assertions ared matched.l-ìOL A `LookMatcher` can change the line terminator used for matching multi-line*' anchors such as `(?m:^)` and `(?m:$)`.<­-ʭ-.> ?>4ʮ-Ѯծ'?'?>ȯ>?!Ҵ?&?а԰9?|Ќ-#?±,?4č-:߶?|Ќ-lܑ-ϲ(?4č-.-<-\ݳzq D  h  hh1 1B ó h_i g _i_i111 1 1ggg cgcثq1ȡ1B ʳ _i_i~ тJG An error that occurs when the Unicode-aware `\w` class is unavailable.MJ This error can occur when the data tables necessary for the Unicode awareHE Perl character class `\w` are unavailable. The `\w` class is used toLI determine whether a codepoint is considered a word character or not whenNK determining whether a Unicode aware `\b` (or `\B`) matches at a particularlӅ@IF This error can only occur when the `unicode-word-boundary` feature isl disabled.]wkD  B~ 7B      1 1B Ɔ  g 111 1 1ggg cgcثq1ȡ1B ͆ g.aL"&eeB"     e      "Oee gg"gfgf~[l(!e      ""Ne"e "  L  / ɜ2۟ B  UB  np     /J g 111 1 1ggg cgcثq1ȡ1B  i#\{#ѽ߮ \#perf-literal-multisubstring'#)#4#<#  r   r r 1 1B Z# r   g   111 1 1ggg cgcثq1ȡ1B a#   a <{$uoF ц$Ba 7ц$B  \$  ц$$     2 2B Z$  q  g q q 222 2 2ggg cgcثq1ȡ2B a$ q q  4{%:#KI* B 6B  %  e   e e 2 2B Z% e   g   222 2 2ggg cgcثq1ȡ2B a%    < %cz[!!!B ;B!! ! %!! %!! !!  2 2B! %! !  !g!  222 2 2ggg cgcثq1ȡ2B! %! !  W <%GK<!!!!BW ?B!! !%!!%!!%!! !!  2 2B!%! !  !g!  222 2 2ggg cgcثq1ȡ2B!%! !  i&4{&ځ8!\&|&"&5&:&E&<&!!  !!    2 2B!Z&! !  f !g!f f 222 2 2ggg cgcثq1ȡ2B!a&! !f f i',{'K'`j!\'ތ]'')'4'<'!!h !!h h 2 2B!Z'! !h  !g!  222 2 2ggg cgcثq1ȡ2B!a'! !  ((0(- A prefilter for accelerating regex searches.(C(@ If you already have your literals that you want to search with,C(@ then the vanilla [`Prefilter::new`] constructor is for you. ButD(A if you have an [`Hir`] value from the `regex-syntax` crate, thenL(I [`Prefilter::from_hir_prefix`] might be more convenient. Namely, it usesL(I the [`regex-syntax::hir::literal`](regex_syntax::hir::literal) module toO(L extract literal prefixes for you, optimize them and then select and build a( prefilter matcher.(H(E A prefilter must have **zero false negatives**. However, by its veryK(H nature, it may produce false positives. That is, a prefilter will neverK(H skip over a position in the haystack that corresponds to a match of theG(D original regex pattern, but it *may* produce a match for a positionI(F in the haystack that does *not* correspond to a match of the originalJ(G regex pattern. If you use either the [`Prefilter::from_hir_prefix`] orH(E [`Prefilter::from_hirs_prefix`] constructors, then this guarantee isL(I upheld for you automatically. This guarantee is not preserved if you useN(K [`Prefilter::new`] though, since it is up to the caller to provide correct?(< literal strings with respect to the original regex pattern.(l( # Cloning(N(K It is an API guarantee that cloning a prefilter is cheap. That is, cloningG(D it will not duplicate whatever heap memory is used to represent the( underlying matcher.(l(-(9(6 This example shows how to attach a `Prefilter` to theK(H [`PikeVM`](crate::nfa::thompson::pikevm::PikeVM) in order to acceleratel( searches.(<(-̲(-&(ӀA#(ʼn-̗( Match, MatchKind,4(-(C(@ let pre = Prefilter::new(MatchKind::LeftmostFirst, &["Bruce "]) (L (> 9(6 .configure(PikeVM::config().prefilter(Some(pre))) ( .build(r"Bruce \w+")?;!&(?|!(Ќ-!$(! Some(Match::must(0, 6..23)),!8(5 re.find(&mut cache, "Hello Bruce Springsteen!"),4"(č-".(-<"(-"("J(G But note that if you get your prefilter incorrect, it could lead to an#( incorrect result!#(<#(-#(-#&(ӀA$#(ʼn-$(ֵ]4$(-$($( // This prefilter is wrong!%C(@ let pre = Prefilter::new(MatchKind::LeftmostFirst, &["Patti "])%(L%(>&9(]&(]&&(?'9(6 // We find no match even though the regex does match.|'(Ќ-l'(ܑ-'8(]4((č-(.(-<((-L((|=I!!!)()(̿)()(!!KKKK 7K7K736l*(*()(<*(!!*(*(̬*(t*(!!j !!j j 2 2B!((! !j , !g!, , 222 2 2ggg cgcثq1ȡ2B!((! !, , Ɲ(4֝(/>!!!!!!!!!!!!!!!!B!!B!!B!!B!!B!!B!!B  !! !  B!!]4(!!  !! !  B!!]<(!!\ \ !! !\ \ B!!]<(!!  !! !  B!!]4(!!  !! !  B!!],(!!O O !! !O O B!!]<ў(!!  !! !  B!!]\(!!% !!% % 2 2B!(! !% * !g!* * 222 2 2ggg cgcثq1ȡ2B!(! !* *  -)* A `usize` that can never be `usize::MAX`.)O)L This is similar to `core::num::NonZeroUsize`, but instead of not permitting3)0 a zero value, this does not permit a max value.)M)J This is useful in certain contexts where one wants to optimize the memoryI)F usage of things that contain match offsets. Namely, since Rust slicesL)I are guaranteed to never have a length exceeding `isize::MAX`, we can useK)H `usize::MAX` as a sentinel to indicate that no match was found. Indeed,N)K types like `Option` have exactly the same size in memory as ad) `usize`.)6)3 This type is defined to be `repr(transparent)` for=): `core::num::NonZeroUsize`, which is in turn defined to be$)! `repr(transparent)` for `usize`.\)ݟR{ !!"B 7"B!!%d)!!"@ @ @ @ !!@ @ 2 2B!)! !@ )A !!A A A A 2 2B!)! !A A !h!A A 22 2 2B!!!)!YA ʹYA !A A A A xB xB !!xB xB 22 2 2B!)! !xB  xB B !!B B 22 2 2B!)! !B  B C !!C C 22 2 2B!)! !C  C A +)( A type that represents a "small" index.)M)J The main idea of this type is to provide something that can index memory,J)G but uses less memory than `usize` on 64-bit systems. Specifically, itsM)J representation is always a `u32` and has `repr(transparent)` enabled. (So@)= it is safe to transmute between a `u32` and a `SmallIndex`.))N)K A small index is typically useful in cases where there is no practical wayL)I that the index will overflow a 32-bit integer. A good example of this isK)H an NFA state. If you could somehow build an NFA with `2^30` states, itsJ)G memory usage would be exorbitant and its runtime execution would be so M)J slow as to be completely worthless. Therefore, this crate generally deems!L)I it acceptable to return an error if it would otherwise build an NFA that!N)K requires a slice longer than what a 32-bit integer can index. In exchange,"J)G we can use 32-bit indices instead of 64-bit indices in various places.")"O)L This type ensures this by providing a constructor that will return an error#J)G if its argument cannot fit into the type. This makes it much easier to$M)J handle these sorts of boundary cases that are otherwise extremely subtle.$)$L)I On all targets, this type guarantees that its value will fit in a `u32`,%I)F `i32`, `usize` and an `isize`. This means that on 16-bit targets, for%F)C example, this type's maximum value will never overflow an `isize`,&G)D which means it will never overflow a `i16` even though its internal'$)! representation is still a `u32`.')'K)H The purpose for making the type fit into even signed integer types like'M)J `isize` is to guarantee that the difference between any two small indices(K)H is itself also a small index. This is useful in certain contexts, e.g.,)) for delta encoding.)))) # Other types)))M)J The following types wrap `SmallIndex` to provide a more focused use case:*)*D)A * [`PatternID`] is for representing the identifiers of patterns.*I)F * [`StateID`] is for representing the identifiers of states in finite+0)- automata. It is used for both NFAs and DFAs.+)+) # Representation+)+J)G This type is always represented internally by a `u32` and is marked as,N)K `repr(transparent)`. Thus, this type always has the same representation as-M)J a `u32`. It is thus safe to transmute between a `u32` and a `SmallIndex`.-)t-) # Indexing-)-D)A For convenience, callers may use a `SmallIndex` to index slices..)d.)5.).O)L While a `SmallIndex` is meant to guarantee that its value fits into `usize`/I)F without using as much space as a `usize` on all targets, callers must/L)I not rely on this property for safety. Callers may choose to rely on this0O)L property for correctness however. For example, creating a `SmallIndex` with1O)L an invalid value can be done in entirely safe code. This may in turn result1')$ in panics or silent logical errors.T2)i#?h!!BA 6B!! !3)!!F F F F !!F F 2 2B!2)! !F 2) G !g! G  G 222 2 2ggg cgcثq1ȡ2B!2)! ! G  G /H !!/H /H B!! !H !!H H H H 2 2B!2)! !H I !h!I I 22 2 2B!!!2)!YI ʹYI !I I I I I I !!I I 23 2 3B!2)! !I  I :J !!:J :J 33 3 3B!2)! !:J  :J SK !!SK SK 33 3 3B!2)! !SK  SK Ԩl)jB)? This error occurs when a small index could not be constructed.j)jN)K This occurs when given an integer exceeding the maximum small index value.k)kI)F When the `std` feature is enabled, this implements the `Error` trait.|l)3xČc|!tl)Ll)!!^ !!^ ^ 3 3B!l)! !^ Q_ !g!Q_ Q_ 333 3 3ggg cgcثq1ȡ3B!l)! !Q_ Q_ ` !!` ` ` ` 3 3B!l)! !` a a !!a a 33 3 3B!l)! !a  a q )tq)-?=!q)q)!!a !!a a 3 3B!p)! !a 3b !g!3b 3b 333 3 3ggg cgcثq1ȡ3B!q)! !3b 3b  G)D The identifier of a regex pattern, represented by a [`SmallIndex`].)K)H The identifier for a pattern corresponds to its relative position amongJ)G other patterns in a single finite state machine. Namely, when buildingK)H a multi-pattern regex engine, one must supply a sequence of patterns toH)E match. The position (starting at 0) of each pattern in that sequenceN)K represents its identifier. This identifier is in turn used to identify and3)0 report matches of that pattern in various APIs.)L)I See the [`SmallIndex`] type for more information about what it means for')$ a pattern ID to be a "small index.")))& Note that this type is defined in theH)E [`util::primitives`](crate::util::primitives) module, but it is also:)7 re-exported at the crate root due to how common it is.L) +[T!!B 8B!!%T)!!bc bc bc bc !!bc bc 3 3B!)! !bc )c ""c c B"" "ad ""ad ad ad ad 3 3B")" "ad d "h"d d 33 3 3B""")"Yd ʹYd "d d d d e e ""e e 33 3 3B")" "e  e /f ""/f /f 33 3 3B")" "/f  /f Ig ""Ig Ig 33 3 3B")" "Ig  Ig  @)= The identifier of a finite automaton state, represented by a) [`SmallIndex`].)N)K Most regex engines in this crate are built on top of finite automata. EachN)K state in a finite automaton defines transitions from its state to another.J)G Those transitions point to other states via their identifiers, i.e., aL)I `StateID`. Since finite automata tend to contain many transitions, it isD)A much more memory efficient to define state IDs as small indices.)L)ݑ^%)" a state ID to be a "small index."<)Vr`""B 8B""&T)""g g g g ""g g 3 3B")" "g )3h ""3h 3h B"" "h ""h h h h 3 3B")" "h i "h"i i 33 3 3B""")"Yi ʹYi "i i i i i i ""i i 33 3 3B")" "i  i aj ""aj aj 33 3 3B")" "aj  aj {k ""{k {k 33 3 3B")" "{k  {k Pv)""""""""""""""""v)v)Mv)""w)w7)M(w)""x)x)M x)""y)x;)M y)""{/)y>)Mz)zC)M{)M{)B""n 1)|B)ցN})N})};)ςN~:)N~@)؃N()N)?)Nۀ)Nh)B"" "o ");)هN )B""(~q %)A)N)N@)3 3B")"ڰ  )ۇ!)ن?)N)̎N0)3 3B")" " )‰!)?)ÐN)M0ω)3 3B"։)" ׉)!)>)N)N0)3 3B")" " ))0)Nnj)׌@)NA)іN)N@)3 3B"Ď)" " Ŏ)̑;)B)N&)M)@)N7)ʚNhӑ)B"" "ix )7)B)N&)M)B)NДB)N)B"" "Ly )Ɨ$)̖?)NX)4X͗)3 3B"ٗ)" " ڗ))') )B""))"g"@))333 3 3ggg cgcثq1ȡ3B")" " ))ߝ')""/)"""!X)0)""ɞ#)(̞)3 3 3B"Ҟ)" " Ӟ)){ *)""/)"#"/)H)3 3 3B")" " )Y| *)))""/)"""!X)0)""#)()3 3 3B")" " )i} -)))""/)"#"/)H)3 3 3B")" " )i~ )"") )B"" "())"""P)()""զ.)@ئ)B"" "())"""P)()"".)@)B"" "(¨)ͩ)"""P)()"".)@)B"" "())"""Pګ)(߫)""0)@)B"" "()Hձ)");)NH)3 3B"Ʋ)" " Dz))x)) )"g"@))333 3 3ggg cgcثq1ȡ3B")" " ))P·)"۷)޷)B"" ")޺)""QQ"H) )""#) )3 3B")" " )ǿ2)""?̿)@@Ͽ)Qڿ)",.))B"" ",M ?)""?)@@)^)""QQ"35H) )""&.) )3 3&B")" " )2 <)N)C)NP)ʅE)B)NP)Nt)X)ES4""B2 8B""%x)""~ ""~ ~ 3 3B")" "~ p "g"p p 333 3 3ggg cgcثq1ȡ3B")" "p p  ""    3 3B")" " B B ""B B 33 3 3B")" "B  B  l) JF""B 9B"" "p)"" ""  3 3B")" " ^ "g"^ ^ 333 3 3ggg cgcثq1ȡ3B")" "^ ^ U))""?)w7{t0""#%())""P))"" ""?)##  3 3B"Ѿ)" #  ##?Ig)#g#nq  333 3 3ggg cgcثq1ȡ3B#ؾ)# #  Pv)################v)v)Mv)##w)w7)M(w)##x)x)M x)##y)x;)M y)##{/)y>)Mz)zC)M{)M{) B##  1)|B)ցN})N})};)ςN~:)N~@)؃N()N)?)Nۀ)Nh)B##܏ ");)هN )B## # %)A)N)N@)3 3B#)# )ۇ!)ن?)N)̎N0)3 3B#)# # )‰!)?)ÐN)M0ω)3 3B#։)#  ׉)!)>)N)N0)3 3B#)# # ))0)Nnj)׌@)NA)іN)N@)3 3B#Ď)# # Ŏ)̑;)B)N&)M)@)N7)ʚNhӑ) B## #u )7)B)N&)M)B)NДB)N)B##V )Ɨ$)̖?)NX)4X͗)3 3B#ٗ)# # ڗ))') )B##&))#g#@))333 3 3ggg cgcثq1ȡ3B#)# # ))ߝ')##/)###!X)0)##ɞ#)(̞)3 3 3B#Ҟ)# # Ӟ)1 *)##/)###/)H)3 3 3B#)# # )a *)))##/)###!X)0)###)()3 3 3B#)# # )q -)))##/)###/)H)3 3 3B#)# # )q )##) )B## #())###P)()##զ.)@ئ) B## #())###P)()##.)@) B## #(¨)ͩ)###P)()##.)@) B## #())###Pګ)(߫)##0)@) B## #()Hձ)#);)NH)3 3B#Ʋ)# # Dz))x)) )#g#@))333 3 3ggg cgcثq1ȡ3B#)# # ))P·)#۷)޷)B## #)޺)##QQ#H) )###) )3 3B#)# # )ǿ2)##?̿)@@Ͽ)^ڿ)#$&))B## #,/ ?)##?)@@)^)##QQ#35H) )##&.) )3 3&B#)# # )ڄ <)N)C)NP)ʅE)B)NP)Nd)W-#x##Bڄ 8B##&x)## ##  3 3B#)# # u #g#u u 333 3 3ggg cgcثq1ȡ3B#)# #u u  ##    3 3B#)# # 8 8 ##8 8 33 3 3B#)# #8  8 . \) tD:##B. 9B## #p)##ӡ ##ӡ ӡ 3 3B#)# #ӡ Q #g#Q Q 333 3 3ggg cgcثq1ȡ3B#)# #Q Q U)|)##?)2ȳ"###%())##P))##ϣ ##?^)##ϣ ϣ 3 3B#Ѿ)# #ϣ  ##?A^)#g#^a  333 3 3ggg cgcثq1ȡ3B#ؾ)# #  !IF The configuration used to determine a DFA's start state for a search.OL A DFA has a single starting state in the typical textbook description. ThatOL is, it corresponds to the set of all starting states for the NFA that builtLI it, along with their epsilon closures. In this crate, however, DFAs have41 many possible start states due to a few factors:MJ * DFAs support the ability to run either anchored or unanchored searches.MJ Each type of search needs its own start state. For example, an unanchoredGD search requires starting at a state corresponding to a regex with a;8 `(?s-u:.)*?` prefix, which will match through anything.JG * DFAs also optionally support starting an anchored search for any oneEB specific pattern. Each such pattern requires its own start state.LI * If a look-behind assertion like `^` or `\b` is used in the regex, thenNK the DFA will need to inspect a single byte immediately before the start of 1. the search to choose the correct start state.  OL Indeed, this configuration precisely encapsulates all of the above factors. IF The [`Config::anchored`] method sets which kind of anchored search to JG perform while the [`Config::look_behind`] method provides a way to set DA the byte that occurs immediately before the start of the search.  NK Generally speaking, this type is only useful when you want to run searches JG without using an [`Input`]. In particular, an `Input` wants a haystack GD slice, but callers may not have a contiguous sequence of bytes as a KH haystack in all cases. This type provides a lower level of control suchMJ that callers can provide their own anchored configuration and look-behind byte explicitly.l-KH This shows basic usage that permits running a search with a DFA without" using the `Input` abstraction.<-̖- . util::start,74-0- let dfa = dense::DFA::new(r"(?-u)\b\w+\b")?; let haystack = "quartz";>; let config = start::Config::new().anchored(Anchored::Yes);.+ let mut state = dfa.start_state(&config)?;*' for &b in haystack.as_bytes().iter() {)6,0&6'6.-<-JG This example shows how to correctly run a search that doesn't begin atKH the start of a haystack. Notice how we set the look-behind byte, and as0- a result, the `\b` assertion does not match.<-̀- ._74-0__%" let config = start::Config::new()  .anchored(Anchored::Yes)! .look_behind(Some(b'q'));._2/ for &b in haystack.as_bytes().iter().skip(1) {)6,0&6 // No match!(% assert!(!dfa.is_match_state(state));.-<-KH If we had instead not set a look-behind byte, then the DFA would assumeOL that it was starting at the beginning of the haystack, and thus `\b` shouldFC match. This in turn would result in erroneously reporting a match:<-- ._74-0__-* // Whoops, forgot the look-behind byte...>Ӫ_._2_)6, 0 &6 +( // And now we get a match unexpectedly. '6!!.-; A hack to align a smaller type `B` with a bigger type `T`.DA The usual use of this is with `B = [u8]` and `T = u32`. That is,FC it permits aligning a sequence of bytes on a 4-byte boundary. ThisFC is useful in contexts where one wants to embed a serialized [denseMJ DFA](crate::dfa::dense::DFA) into a Rust a program while guaranteeing the# alignment required for the DFA.MJ See [`dense::DFA::from_bytes`](crate::dfa::dense::DFA::from_bytes) for an$! example of how to use this type.<$$$$42 "$$ /185 A zero-sized field indicating the alignment we want.4$$܇' d>; A possibly non-sized field containing a sequence of bytes.,$$  $$$$_D2CgK$g$QT  444 4 4ggg cgcثq1ȡ4B$$ $  ̠DA An error that occurs when serializing an object from this crate.KH Serialization, as used in this crate, universally refers to the processHE of transforming a structure (like a DFA) into a custom binary formatOL represented by `&[u8]`. To this end, serialization is generally infallible.NK However, it can fail when caller provided buffer sizes are too small. When30 that occurs, a serialization error is reported.GD A `SerializeError` provides no introspection capabilities. Its onlyHE supported operation is conversion to a human readable error message.JG This error type implements the `std::error::Error` trait only when theDA `std` feature is enabled. Otherwise, this type is defined in all configurations.t,($%$%$$,  $g$  444 4 4ggg cgcثq1ȡ4B$$ $   )LI An error that occurs when deserializing an object defined in this crate.))K_*H_+KH represented by `&[u8]`. Deserialization, then, refers to the process of+HE cheaply converting this binary format back to the object's in-memory,DA representation as defined in this crate. To the extent possible,,GD deserialization will report this error whenever this process fails.--IF A `DeserializeError` provides no introspection capabilities. Its only.H_..J_/D_/_0Tg3v~M$$B 7B$$ $0$$  $g$  444 4 4ggg cgcثq1ȡ4B$/$ $  00Nm9\'$$$$$$$$$$$$$$$$$$$$$$$  $$<0$$00$$,t1$$1$1$$,d1$$1$1$$,|1$$l1D1$$T2,2$$t2$$l2D2$$T2,2$$2$$2L2$$t2<2$$l3$$3D3$$,3$$3$3$$,L3$$33$$3$3$$,<4$$44$$4$4$$,  $g$  444 4 4ggg cgcثq1ȡ4B$0$ $  $$$$$;҅ ,-ǒe0y$$'B҅ 8'B$$ $L-$$'l  $$l l 4 4B$,$ $l   $$    4 4B$,$ $ w  $h$w w 44 4 4B$$$,$Yw ʹYw $w w w w b  b  $$b b 44 4 4B$,$ $b  b   $$  44 4 4B$-$ $     $$  44 4 4B$-$ $   " @oȶ4$$B" 8B$$ $<@$$  $$  5 5B$?$ $   $g$  555 5 5ggg cgcثq1ȡ5B$@$ $  s FO-$$Bs 8B$$ $; methods and/or superfluous parameters in the common cases.85 An `Input` permits configuring the following things:MJ * Search only a substring of a haystack, while taking the broader context63 into account for resolving look-around assertions.EB * Indicating whether to search for all patterns in a regex, or to.+ only search for one pattern in particular. :7 * Whether to perform an anchored on unanchored search. 52 * Whether to report a match as early as possible.  KH All of these parameters, except for the haystack, have sensible default MJ values. This means that the minimal search configuration is simply a call HE to [`Input::new`] with your haystack. Setting any other parameter isl  optional.  GD Moreover, for any `H` that implements `AsRef<[u8]>`, there exists a JG `From for Input` implementation. This is useful because many of the IF search APIs in this crate accept an `Into`. This means you can JG provide string or byte strings to these routines directly, and they'll85 automatically get converted into an `Input` for you.GD The lifetime parameter `'h` refers to the lifetime of the haystack. # Organization;8 The API of `Input` is split into a few different parts:FC * A builder-like API that transforms a `Input` by value. Examples:,) [`Input::span`] and [`Input::anchored`].GD * A setter API that permits mutating parameters in place. Examples:41 [`Input::set_span`] and [`Input::set_anchored`].HE * A getter API that permits retrieving any of the search parameters.>; Examples: [`Input::get_span`] and [`Input::get_anchored`].NK * A few convenience getter routines that don't conform to the above namingFC pattern due to how common they are. Examples: [`Input::haystack`],(% [`Input::start`] and [`Input::end`].HE * Miscellaneous predicates and other helper routines that are useful<9 in some contexts. Examples: [`Input::is_char_boundary`].OL A `Input` exposes so much because it is meant to be used by both callers ofOL regex engines _and_ implementors of regex engines. A constraining factor isNK that regex engines should accept a `&Input` as its lowest level API, whichKH means that implementors should only use the "getter" APIs of a `Input`.)& # Valid bounds and search termination@= An `Input` permits setting the bounds of a search via eitherIF [`Input::span`] or [`Input::range`]. The bounds set must be valid, or=: else a panic will occur. Bounds are valid if and only if:C@ * The bounds represent a valid range into the input's haystack.IF * **or** the end bound is a valid ending bound for the haystack *and*@= the start bound is exactly one greater than the start bound.MJ In the latter case, [`Input::is_done`] will return true and indicates anyKH search receiving such an input should immediately return with no match.KH Note that while `Input` is used for reverse searches in this crate, theIF `Input::is_done` predicate assumes a forward search. Because unsignedNK offsets are used internally, there is no way to tell from only the offsets,) whether a reverse search is done or not.Ԣ # Regex engine supportMJ Any regex engine accepting an `Input` must support at least the following\  things:  &# * Searching a `&[u8]` for matches. GD * Searching a substring of `&[u8]` for a match, such that any match!85 reported must appear entirely within that substring.!B? * For a forwards search, a match should never be reported when"NK [`Input::is_done`] returns true. (For reverse searches, termination should"# be handled outside of `Input`.)"#JG Supporting other aspects of an `Input` are optional, but regex engines#LI should handle aspects they don't support gracefully. How this is done is$MJ generally up to the regex engine. This crate generally treats unsupported$LI anchored modes as an error to report for example, but for simplicity, in%KH the meta regex engine, trying to search with an invalid pattern ID just&'$ results in no match being reported. 2$$*0$$$$&&D&$$RT&$&$$&D&$$łt'D'$$  2$$$$  5 5B$&$ $2 |:7 A representation of a span reported by a regex engine.EB A span corresponds to the starting and ending _byte offsets_ of aJG contiguous region of bytes. The starting offset is inclusive while theHE ending offset is exclusive. That is, a span is a half-open interval.KH A span is used to report the offsets of a match, but it is also used toJG convey which region of a haystack should be searched via routines like [`Input::span`].KH This is basically equivalent to a `std::ops::Range`, except thisNK type implements `Copy` which makes it more ergonomic to use in the contextNK of this crate. Like a range, this implements `Index` for `[u8]` and `str`,NK and `IndexMut` for `[u8]`. For convenience, this also impls `From`,52 which means things like `Span::from(5..10)` work." 4L$$,) The start offset of the span, inclusive.,$$t*' The end offset of the span, exclusive.$$     %%  5 5B$$ %    %%    5 5B%% %   %h%  55 5 5B%%%%Y ʹY %    !  !  %%! ! 55 5 5B%% %!  ! <9 A representation of "half" of a match reported by a DFA.OL This is called a "half" match because it only includes the end location (orLI start location for a reverse search) of a match. This corresponds to theՅLI information that a single DFA scan can report. Getting the other half of96 the match requires a second scan with a reversed DFA.܆GD A half match also includes the pattern that matched. The pattern isNK identified by an ID, which corresponds to its position (starting from `0`)OL relative to other patterns used to construct the corresponding DFA. If onlyLjOL a single pattern is provided to the DFA, then all matches are guaranteed to엉 have a pattern ID of `0`.5 m%%<%%lʋ4ʋ%%2- 2- 2- 2-  %%2- 2- 6 6B%% %2- ʼn -  %g%- - 666 6 6ggg cgcثq1ȡ6B%ˉ% %- - P/  %%P/ P/ P/ P/ 6 6B%҉% %P/ /  %h%/ / 66 6 6B%%%։%Y/ ʹY/ %/ / / / q1  q1  %%q1 q1 66 6 6B%܉% %q1  q1 ;8 A representation of a match reported by a regex engine.KH A match has two essential pieces of information: the [`PatternID`] that96 matches, and the [`Span`] of the match in a haystack.IF The pattern is identified by an ID, which corresponds to its positionHE (starting from `0`) relative to other patterns used to construct theNK corresponding regex engine. If only a single pattern is provided, then all74 matches are guaranteed to have a pattern ID of `0`.řəKH Every match reported by a regex engine guarantees that its span has its96 start offset as less than or equal to its end offset.H sAH.%%<%%T$%%S4 S4 S4 S4  %%S4 S4 6 6B%ؚ% %S4 ߚ 4  %g%4 4 666 6 6ggg cgcثq1ȡ6B%% %4 4 u6  %%u6 u6 u6 u6 6 6B%% %u6 6  %h%6 6 66 6 6B%%%%Y6 ʹY6 %6 6 6 6 8  8  %%8 8 66 6 6B%% %8  8 ʿ A set of `PatternID`s.LI A set of pattern identifiers is useful for recording which patterns haveHE matched a particular haystack. A pattern set _only_ includes pattern85 identifiers. It does not include offset information.l-$! This shows basic usage of a set.<-0- use regex_automata::{PatternID, PatternSet};" let pid1 = PatternID::must(5);" let pid2 = PatternID::must(8); // Create a new empty set.˻&# let mut set = PatternSet::new(10); // Insert pattern IDs. set.insert(pid1); set.insert(pid2); // Test membership.Ѽ  assert!(set.contains(pid1));  assert!(set.contains(pid2)); // Get all members.|Ќ- vec![5, 8],ϽA> set.iter().map(|p| p.as_usize()).collect::>(),4č- // Clear the set. set.clear();$! // Test that it is indeed empty. assert!(set.is_empty());<-̉[ #cZ%%T%% ,%%)A  %%A A 6 6B% % %A +B  %g%+B +B 666 6 6ggg cgcثq1ȡ6B% % %+B +B C  %%C C C C 6 6B% % %C D  D  %%D D 66 6 6B% % %D  D  C@ An error that occurs when a `PatternID` failed to insert into a `PatternSet`.NK An insert fails when the given `PatternID` exceeds the configured capacity of the `PatternSet`.DA This error is created by the [`PatternSet::try_insert`] routine.n e{)6%%L%%|D%%E  %%E E 6 6B% % %E fF  %g%fF fF 666 6 6ggg cgcثq1ȡ6B% % %fF fF A> An iterator over all pattern identifiers in a [`PatternSet`].OL The lifetime parameter `'a` refers to the lifetime of the pattern set being iterated over.@= This iterator is created by the [`PatternSet::iter`] method.̍ %% @ή%6%%<< 0{?ippʽp opopo2ЫH  %%%%H H 6 6B% % %H }I  %%%g%nq}I }I 666 6 6ggg cgcثq1ȡ6B% % %}I }I +( The type of anchored search to perform.OL This is *almost* a boolean option. That is, you can either do an unanchoredOL search for any pattern in a regex, or you can do an anchored search for any pattern in a regex.؉܉NK A third option exists that, assuming the regex engine supports it, permits85 you to do an anchored search for a specific pattern.HE Note that there is no way to run an unanchored search for a specificMJ pattern. If you need that, you'll need to build separate regexes for eachd.d0KH If a regex engine does not support the anchored mode selected, then theIF regex engine will return an error. While any non-trivial regex engineLI should support at least one of the available anchored modes, there is noLI singular mode that is guaranteed to be universally supported. Some regexюIF engines might only support unanchored searches (DFAs compiled withoutGD anchored starting states) and some regex engines might only support.+ anchored searches (like the one-pass DFA).:7 The specific error returned is a [`MatchError`] with aѐGD [`MatchErrorKind::UnsupportedAnchored`] kind. The kind includes the/, `Anchored` value given that is unsupported.ɑ͑HE Note that regex engines should report "no match" if, for example, anJG `Anchored::Pattern` is provided with an invalid pattern ID _but_ whereOL anchored searches for a specific pattern are supported. This is smooths outLI behavior such that it's possible to guarantee that an error never occursJG based on how the regex engine is configured. All regex engines in thisɔOL crate report "no match" when searching for an invalid pattern ID, but where<9 searching for a valid pattern ID is otherwise supported.֕lڕ-GD This example shows how to use the various `Anchored` modes to run aGD search. We use the [`PikeVM`](crate::nfa::thompson::pikevm::PikeVM)KH because it supports all modes unconditionally. Some regex engines, likeȗLI the [`onepass::DFA`](crate::dfa::onepass::DFA) cannot support unanchoredl]<-=--&ӀA*' Anchored, Input, Match, PatternID,4ٙ- E r"Mrs. \w+", r"Miss \w+", r"Mr. \w+",Ú r"Ms. \w+",Dך?&?'$ let hay = "Hello Mr. Springsteen!";1. // The default is to do an unanchored search.FC assert_eq!(Some(Match::must(2, 6..21)), re.find(&mut cache, hay));>; // Explicitly ask for an unanchored search. Same as above.74 let input = Input::new(hay).anchored(Anchored::No);FۑaGD // Now try an anchored search. Since the match doesn't start at the41 // beginning of the haystack, no match is found!85 let input = Input::new(hay).anchored(Anchored::Yes);1. assert_eq!(None, re.find(&mut cache, input));֟ڟJG // We can try an anchored search again, but move the location of whereGD // we start the search. Note that the offsets reported are still inIF // terms of the overall haystack and not relative to where we started // the search.ʡC@ let input = Input::new(hay).anchored(Anchored::Yes).range(6..);HE assert_eq!(Some(Match::must(2, 6..21)), re.find(&mut cache, input));עۢIF // Now try an anchored search for a specific pattern. We specificallyKH // choose a pattern that we know doesn't match to prove that the search-* // only looks for the pattern we provide. let input = Input::new(hay)85 .anchored(Anchored::Pattern(PatternID::must(1))) .range(6..);1aåLI // But if we switch it to the pattern that we know matches, then we find // the match.a¦85 .anchored(Anchored::Pattern(PatternID::must(2)))aH՘a٧ݧ.-<- łm%%%%%%%%%%łB٨IF Run an unanchored search. This means a match may occur anywhere at or+( after the start position of the search.שߩ@= This search can return a match for any pattern in the regex.%%ł%%łEB Run an anchored search. This means that a match must begin at the! start position of the search.@a%%ł%%łˇ JG Run an anchored search for a specific pattern. This means that a matchIF must be for the given pattern and must begin at the start position of| the search.ˇ %% %ˇ ˇ łB%%aL%%J J J łJ  ł%%J J 6 6łłB% % %J  łK  ł%g%K K 666 6ł 6ggg cgcثq1ȡ6B% % %K K %M  ł%%%M %M %M %M 6 6łB% % %%M eM  łeM  ł%%eM eM 66 6ł 6łB% % %eM  eM ׷;8 The kind of match semantics to use for a regex pattern.JG The default match kind is `LeftmostFirst`, and this corresponds to theDA match semantics used by most backtracking engines, such as Perl.:7 # Leftmost first or "preference order" match semanticsKH Leftmost-first semantics determine which match to report when there areNK multiple paths through a regex that match at the same position. The tie isOL essentially broken by how a backtracker would behave. For example, considerջNK running the regex `foofoofoo|foofoo|foo` on the haystack `foofoo`. In thisOL case, both the `foofoo` and `foo` branches match at position `0`. So should'$ the end of the match be `3` or `6`?KH A backtracker will conceptually work by trying `foofoofoo` and failing.GD Then it will try `foofoo`, find the match and stop there. Thus, theLI leftmost-first match position is `6`. This is called "leftmost-first" orJG "preference order" because the order of the branches as written in the̿:7 regex pattern is what determines how to break the tie.KH (Note that leftmost-longest match semantics, which break ties by alwaysKH taking the longest matching string, are not currently supported by thisJG crate. These match semantics tend to be found in POSIX regex engines.)IF This example shows how leftmost-first semantics work, and how it even%" applies to multi-pattern regexes:<--&ӀAtA4- E r"foofoofoo", r"foofoo",|:D?&?GD let got: Vec = re.find_iter(&mut cache, "foofoo").collect();.+ let expected = vec![Match::must(1, 0..6)];..-<- # All matchesLI The `All` match semantics report any and all matches, and generally willNK attempt to match as much as possible. It doesn't respect any sort of matchJG priority at all, so things like non-greedy matching don't work in thisL mode.MJ The fact that non-greedy matching doesn't work generally makes most formsMJ of unanchored non-overlapping searches have unintuitive behavior. Namely,JG unanchored searches behave as if there is a `(?s-u:.)*?` prefix at theMJ beginning of the pattern, which is specifically non-greedy. Since it willLI be treated as greedy in `All` match semantics, this generally means thatNK it will first attempt to consume all of the haystack and is likely to windĄ up skipping matches.GD Generally speaking, `All` should only be used in two circumstances:OL * When running an anchored search and there is a desire to match as much asLI possible. For example, when building a reverse regex matcher to find theLI start of a match after finding the end. In this case, the reverse searchDA is anchored to the end of the match found by the forward search.IF * When running overlapping searches. Since `All` encodes all possibleNK matches, this is generally what you want for an overlapping search. If youOL try to use leftmost-first in an overlapping search, it is likely to produceMJ counter-intuitive results since leftmost-first specifically excludes some52 matches from its underlying finite state machine.OL This example demonstrates the counter-intuitive behavior of `All` semantics52 when using a standard leftmost unanchored search:<-̟-&ӀAֵ]4->?< .configure(PikeVM::config().match_kind(MatchKind::All)) .build("foo")?;)& let hay = "first foo second foo wat";&?B? let got: Vec = re.find_iter(&mut cache, hay).collect();74 // Notice that it completely skips the first 'foo'!0- let expected = vec![Match::must(0, 17..20)];..-<-OL This second example shows how `All` semantics are useful for an overlappingIF search. Note that we use lower level lazy DFA APIs here since the NFAMJ engines only currently support a very limited form of overlapping search.<---* hybrid::dfa::{DFA, OverlappingState},$Ǵ-4--KH // If we didn't set 'All' semantics here, then the regex would onlyKH // match 'foo' at offset 3 and nothing else. Why? Because the stateNK // machine implements preference order and knows that the 'foofoo' andMJ // 'foofoofoo' branches can never match since 'foo' will always match-* // when they match and take priority.<:)& .build(r"foo|foofoo|foofoofoo")?;&?.-(% let input = Input::new("foofoofoo");̨ let mut got = vec![];Tס:GD re.try_search_overlapping_fwd(&mut cache, &input, &mut state)?;)& let m = match state.get_match() {Կ: Some(m) => m,T1 got.push(m);,0Ě:: HalfMatch::must(0, 6), HalfMatch::must(0, 9),4:..-<-: j+ n%%%%%%  Report all possible matches.3%%GI%%Q Q Q Q  %%Q Q 6 6B% % %Q  yR  %%yR yR B%% %R  %g%R R 666 6 6ggg cgcثq1ȡ6B% % %R R S  %%S S S S 6 6B% % %S S  S  %%S S 66 6 6B% % %S  S OR KH Report only the leftmost matches. When multiple leftmost matches exist,HE report the match corresponding to the part of the regex that appearsĝ first in the syntax.><T OR %%OR OR %% HE An error indicating that a search stopped before reporting whether a match exists or not.LI To be very clear, this error type implies that one cannot assume that noJG matches occur, since the search stopped before completing. That is, ifJG you're looking for information about where a search determined that noOL match can occur, then this error type does *not* give you that. (Indeed, atMJ the time of writing, if you need such a thing, you have to write your own search routine.)HE Normally, when one searches for something, the response is either anKH affirmative "it was found at this location" or a negative "not found atNK all." However, in some cases, a regex engine can be configured to stop itsNK search before concluding whether a match exists or not. When this happens,OL it may be important for the caller to know why the regex engine gave up andOL where in the input it gave up at. This error type exposes the 'why' and thed 'where.'MJ For example, the DFAs provided by this library generally cannot correctlyIF implement Unicode word boundaries. Instead, they provide an option toIF eagerly support them on ASCII text (since Unicode word boundaries areLI equivalent to ASCII word boundaries when searching ASCII text), but willHE "give up" if a non-ASCII byte is seen. In such cases, one is usuallyHE required to either report the failure to the caller (unergonomic) orNK otherwise fall back to some other regex engine (ergonomic, but potentiallyd costly).OL More generally, some regex engines offer the ability for callers to specifyMJ certain bytes that will trigger the regex engine to automatically quit if they are seen.JG Still yet, there may be other reasons for a failed match. For example,LI the hybrid DFA provided by this crate can be configured to give up if itOL believes that it is not efficient. This in turn permits callers to choose a different regex engine.NK (Note that DFAs are configured by default to never quit or give up in thisKH fashion. For example, by default, a DFA will fail to build if the regexNK pattern contains a Unicode word boundary. One needs to opt into the "quit" behavior via options, likegd [`hybrid::dfa::Config::unicode_word_boundary`](crate::hybrid::dfa::Config::unicode_word_boundary).)*' There are a couple other ways a search)& can fail. For example, when using theOL [`BoundedBacktracker`](crate::nfa::thompson::backtrack::BoundedBacktracker)KH with a haystack that is too long, or trying to run an unanchored searchނ/, with a [one-pass DFA](crate::dfa::onepass). xrT%%˪)B  ˪)B%% %!փЃ%%˪)8U  %%8U 8U 6 6B% % %8U U  %g%U U 666 6 6ggg cgcثq1ȡ6B% % %U U V  %%V V V V 6 6B% % %V W  W  %%W W 66 6 6B% % %W  W ,) The underlying kind of a [`MatchError`].NK This is a **non-exhaustive** enum. That means new variants may be added in  a semver-compatible release.:͉ Lru0%%%&%%%&&$ݗC@ The search saw a "quit" byte at which it was instructed to stoptUd%%DǘEB The "quit" byte that was observed that caused the search to stop.$%%l30 The offset at which the quit byte was observed.4%%4GD The search, based on heuristics, determined that it would be betterɚEB to stop, typically to provide the caller an opportunity to use an쓛 alternative regex engine.EB Currently, the only way for this to occur is via the lazy DFA andIF only when it is configured to do so (it will not return this error byl՜ default).%%lC@ The offset at which the search stopped. This corresponds to theĝ96 position immediately following the last byte scanned.4%%|GD This error occurs if the haystack given to the regex engine was tooEB long to be searched. This occurs, for example, with regex enginesIF like the bounded backtracker that have a configurable fixed amount ofHE capacity that is tied to the length of the haystack. Anything beyondѠA> that configured limit will result in an error at search time.%%T74 The length of the haystack that exceeded the limit.%%EB An error indicating that a particular type of anchored search wasҢ=: requested, but that the regex engine does not support it.FC When this error occurs with [`Anchored::Pattern`], then one commonGD cause is that the underlying regex engine was built without supportHE for anchored starting states for each pattern. For DFA engines, this  typically means enabling theif [`hybrid::dfa::Config::starts_for_each_pattern`](crate::hybrid::dfa::Config::starts_for_each_pattern)45gd [`dfa::dense::Config::starts_for_each_pattern`](crate::dfa::dense::Config::starts_for_each_pattern)IF configuration option, depending on which regex engine is being built.اHE Note that this error should not be returned by a regex engine simplyKH because the pattern ID is invalid (i.e., equal to or exceeds the numberKH of patterns in the regex). In that case, the regex engine should reportͩ a non-match.<%%t0V$%%łW  &&W W 6 6B& & &W Z  &g&Z Z 666 6 6ggg cgcثq1ȡ6B& & &Z Z _  &&_ _ _ _ 6 6B& & &_ _  _  &&_ _ 66 6 6B& & &_  _ -T -v&&Ԗ -$ -&&Զ -$ -&&g  &&g g 6 6B& -& &g gh  &g&gh gh 666 6 6ggg cgcثq1ȡ6B& -& &gh gh ܦ-L-76Y F&&&T--&&-,-&& -4-&& j  &&j j 6 6B& -& &j m l<-&&҈?&&T҈Bm GT҈B&& &<-<-&&Tdk  &&҈&g&MOdk dk 666 6҈ 6ggg cgcثq1ȡ6B& <-& &dk dk ?A?I?[]?n ?_ ?( ? ?[ ?1 ?K ?^ ? ? ?^ ? ? ?#r?^q?fq?j?* ? ? ? ?' ? ?+ ? ? ?4d? ? ?| ? ?; ? ? ?Q ? ? ?f ? ? ? ? ?0 ?@ ?v% ?% ?- ? ?1 ? ?5 ? ?@ ?i- ? ?0+ ?}* ?* ?) ??) ?( ? ( ?x' ?' ?% ? ? ?|% ?$ ? ?^$ ?# ?" ?=" ?! ?P! ?! ? ?1 ? ?r ? ? ?; ? ?I ? ?\ ? ? ? ?? ?K ? ? ?( ? ?b ? ? ?' ? ?h ? ? ?W ? ? ? ?% ?v ?3e? ? ?g ? ?[ ?h ? ? ?9 ? Ӷ ? ?- ? ? ? ?C ? ?R ? ? ?K ? ? ?Q ? ?,d? ?o ? ?W ? ? ?Y ? ? ? ? ?Q ? ? ? ?Wc?^b?fb?=b? ? ? ?! ? ? ?D ? ? ? ? ? ?U ? ? ? ?) ? ?r ? ? ?a? a?a??` ? ?A ? ? ? ?N ? ?] ? ? ټ ? ? ? ?_?_??" ?M^?U^?Ř?& ?Z?Y?Y?? ?V?U?U??. ?~ ? ? ? ???>?>??T ? ?u ?2 ? ?$ ?ի ? ?? ? ??ܨ ? ?o ?9?:?? ?+?*?*?? ? ? ? ?ۡ ? ? ? ?N ?S ? ? ?"? ? ?] ?b} ?| ?){ ?x ? x ?=s ?p ?m ?xm ?i ? ?5g ? ?*d ? ?c ? ?Za ?H ?G ? ??@F ?E ?E ?#E ?D ?eD ?D ?C ?(C ?B ?sB ?B ?A ?YA ?A ?@ ? ?E@ ?? ?,? ?"?3= ? ?< ? < ? ٲ?: ?: ?9 ?8 ? ?7 ?7 ?/7 ?6 ?5 ?45 ?z!?4 ?4 ? ?0 ? ?/ ?7/ ? ? ?W. ?- ?- ?, ?, ?A, ?+ ?;+ ?) ?y( ? ?L?& ?% ? ?L? ?% ?$ ?J$ ?# ?# ?!# ?" ?e" ?! ? ?l! ? ? ?? ? ? ?? ? ?1? ? ? ? ? "?% ? '?=?q ? ?E ? ?^ ? ? ?[ ?  ?k ? ? c? ?8 ?K ? n?Ln??? 7? ?*? ?' ? ?j ? ? ? ? ? ?L? ? ? ?d ? f? ? l?l? ? w?w? ?+ ? ? ? ?} ?gq ?l ?` ?P\ ?ZL ?C ?#> ?2 ?/ ? ? ?? ? ֜?ݜ?5 ? ?? ? ? ?? ? ? ?? ? ? ? ? ? ˍ? ? ?? ? ?? ?H ? ?| ? ? ?J ? ? ? ? ?Z ? ? ? ? ? ??L ? ?ş?s ? ?? ? ? ܢ?? ? ? ?? ??#?+?(~?? ?? ????G?ۅ?j ? ?? ?? ? ?? ?+ ?????? ? ? ? ?s ? ? ???? ? ? n? ? ? ?0 ?e ? ? ?D ?J ? ??? ? ??O ? ? ?: ?B ? ? ?ݿ ? c? ? d? d?ֽ ? ?Q ? ? ?a ? ?Ⱥ ?y ? ?\ ? z? z? ? ? ?: ? ? ? ?W ?_ ??[ ?R ? '? ?p ?x ?? ? ? ?4 ? ?o ?$ ? ??ج ? ?? ?n ? =?=? ? ????( ? a?a? a?Ԥ ? e?e?e? f?ݣ ? ?? ?R ? ??j ? ??] ? ?? ? ?P ? ?? ? ?e ? ??k ? ? ?R?ɛ ? ? ? ? ? }? ?1 ? ?c ? ? ?X ? ? ?M ? ? ?4 ? ??#?+?w?? ?? ?@ ? ?ё ? ?9 ?h???1????u ?u ?????)s ?Vl ?i ?;e ?c ?b ?G` ?\ ?RY ?W ?-Q ?N ?/ ?], ?e* ?O' ?& ?'" ? ? ? ? ? ? ?R ?G ?R ? ??7?????8?u?? ?h???<????r????P?X???  ? ?  ??  ???  ??y?? ϡ ?f?  ??ϫ?  ???  ?  ??  ?v?  ??  ??  ??  ?? ͟ ?? ֨ ?&?  ??  ?U?  ??4???(?Ԡ?y?$??\????  ??;???5?W?_? ??? ??@?H??i????Ͻ? ??.??8?̚?  ??.??  ?]???A?ǖ??9?A??E????Д?w?v?? *?*??? + ?G?? 8 ?T? c ?Lc ?B?J?'?g??@??C??È?  ?L ???-??}?+?  ?L ??Ȭ?Ь?2?ܫ???66??D? 4 ?? 9 ???B?J?:?_???=??+?3?7?r?z?D?5?׃? S ?K? Y ???v|?Pz?Tk?g?b?_?Y?'Y?oS?O?{K?zE?;>?&6?0?*?!???)?Q ? ?  ? ? ?d ? ?  ? ?/ ??c?  ??  ??  ? ?e? e?xd?d???Z?  ??  ? ?a?a?"a?{?`?`?`?+?  ?? ?\?[?[??C[?Z?Z??  ? ??z?4X?W?V?W?4V??=?=?#=?<?<?<?"<?8:?9?9? 9?8?8?8?8?? ; ?L; ?2??P? V ??{???F?4?߹?Z???-??e?  ? ?J0?R0???-?-?\??+?+??m?Z*?b*??פ?M)?(?(??£?M?  ??'?&?&?V?&?p%?x%?)$? #?}"?"?{??¡?f?f?ԟ??۞?[?  ?ݝ??  ?L ?!?!???; ?C ?i????|?Ay?n?j? ?b?`]?T?N? ? ?.K? ? ? ?wF? ? ƈ?B? ? ? ? ? ?5-?:'??R?? ?L?-? ?L??? ?L??5 ?= ??_?????e!?>? ?=??*?0?>?=?d???U?????b??????D?*?d?l?E}?;?u?}????? ??0?P?X???ļ????H?l?Ӹ? !?L!?? !?L!?v? ? ? ?? 0? 0? 0??z?????#????Ъ???~??Ş???´?ʴ?? Y"?Y"?:? x"?y"?? ~"?~"? ? "?"?? "?"? "?"? "?"? ø"? ν"?ս"?ki? "?"?6h? "?"?^?y]? "?"?̱?Ա?? "?"?[? "?"?U?T? "?"?ή?֮?? "?"?4R? "?"? "?L?K? "?ɚ"?&?.?R? "?"? "?J?9:?J9?8? ?L?g4?3?2? @?L@?1????!??????? ?$?$?"?!? !??O??l????!? ?L?-?g?o?b?????_i?' ?' ?? ??? ?O? ?? ?G? ?? ϗ?q? ?e? ?P? ?@? ?2? ?"? ¬?? ?? ?|? ?g? ?T? ?? ?? ?n? ?\? ?? ?? ?N????@? ???*?0?? Y?LY????5? c?Lc?|?|???????E?#? ?? ?????:? ?L?#|?_{?g{? ??B?ߦ? #?r? #??Х?z?y?y?q ??[?? $?ˢ? $?t?)?%y?ax?ix? ?g??I? %?ܟ? %??:?w?v?v? ??)?Ɲ? %?Y? %???u?u?u? ?a??C? %?֚? "%??4?xt?s?s?f ?|??^? &?? &??O?,s?hr?pr? ??ݕ?z? '? ? '??k?e?d?d?, ??|??2}? c(?{x? p(?w?4w?{r? q?p? (?'p? (?o?ro?zn? n? (?m? (?4m?l?3a?m`?u`?* ?l?k?j?[?Q[?Z?Z? Z?Z?Y?Y?Y?Y?a? )?L)?J?I?I? G ?I?rH?zH?G?G?8G?@G?F?F?_[?Z?Y?Y?W?T?S?8S?XR?Q?CE?~D?D?Q_ ?C?fC?nC?N?M? o)?Lo)?B?B?B?3b ?kM?-=?1<?;?;?:?:?e:?m:?9?9?C6?G5?4?4?3?4?{3?3?2?3?/?T/?.?`.?-?+?*?)?H̜)?)?!)?"(?L'????p ????@$?>#?)?H)??J?R?^ ?c"???? ? ??v??????)?H̜)??E?F?p ??4?<?u ???"?d ?b ?)?H)?<?v?~?Q ??G?L?S? ?????­ ? ?'??,?k??p?}??h?? U??? Z?LZ????7 ?3???5??r?w?? *?? *????Ȼ ??X??B??0????$??D? *?? *?? *?p?w? ??? ?? '?L'?[?b? ?P?W? ?? C?LC??`? ??U?<?m? ?$?ٷ?|?ض?`?\??1?b??F?? ??í??f???X????ܧ?y?f?????\?E??(?/?m?t???h?o??*? 1?L1?x?,???D??? ??0???? ? ???0?ؙ? F?LG?&?٘??'???i??? T?LT?;??I??6?Ւ?O?ȑ??.?ُ??)?Ў?x????*??+?I? ?L?c? ??Z???8?q??}?? ?p???n??Ɓ?'? !2? !2? !2?? 2? 2?~? ϝ2? 2?\}? 2? 2? I+?I+?w?H+? L+?L+?v?K+? N+?N+?Nt?M+?b?:b?va?`?_?9_??K,?2)?&?#? ????Q???O?0?? ? ? ?L?^??'?.?k?r?} ????)?[? ?L??O??b? ?? ??ŭ?̭?- ?>??????h? ?.?5?4 ????Z?a?{?????u???3?:?+B ?????h?????P??M??X?_?fF ?e? ?L?̘???}I ??k???&?-?K ?? ??O?$?Ag? f?f?R ?e?d?d??V?U?U?U ?HU?T?T?G? I?DH?KH?Z ?G?'G?.G?"? ӫ?L?jF?E?E?gh ???W? ?D?N???_???c???? 9-?L9-?C?C?dk ?J???F?Mjw{ """"####fully compiled DFAs'hybrid::regex::Builder::build_from_dfas dfa::sparsedfa::Automaton%Unicode scripts and script extensions Input::range hybrid::regex::Regex::try_searchExamples>nfa::thompson::backtrack::BoundedBacktracker::max_haystack_len*hybrid::dfa::Config::unicode_word_boundary hybrid::dfahybrid::regex::Builder::dfacͱc%https://www.unicode.org/reports/tr18/ccղcclazy DFA’s configurationٳccdfa::regex::Regexͱc5nfa::thompson::pikevmcc$BoundedBacktracker::max_haystack_lencܴc hybrid::regex::Builder::thompsoncc1https://www.unicode.org/reports/tr44/tr44-24.htmlc5cٵcnfa::thompson::NFAܴccmeta::Regex::search_withccnfa::thompson::backtrack5cٵcc util::syntaxccmeta regex enginecc meta::Regexccanchored searchescc,nfa::thompson::backtrack::BoundedBacktrackercconfigured visited capacityܸchybrid::regex::Builder::syntaxcc'nfa::thompson::Compiler::build_from_hircc.ֹchttps://docs.rs/regexܸccdfa::dense::DFAccdfa::onepass::DFAc.BoundedBacktracker::try_searchֹccnfa::thompson::Buildercchybrid::dfa::Buildercc%https://www.unicode.org/reports/tr29/.ڻcCrate featuresccuse zero-copydeserializationcchybrid::regex::Regexcۼchybrid::regex::Regex::findڻcc%meta regexenginecc(construction directly froma Thompson NFAcc%https://www.unicode.org/reports/tr24/ ۼccAvailable regex enginescǾclazyDFAccAutomaton traitcc Input::earliestcc9configuring heuristic support for Unicode word boundariesǾcchybrid::regex::Builder::buildccUnicode general categoriesc cShould I be using this crate?cchybrid::dfa::Configcc dfa::regexcc Input::spanccccnfa::thompson::Configcc4Unicode’s “simple loose matches” specificationc cc nfa::thompsonccccnfa::thompson::pikevm::PikeVMc nfa::thompson::backtrack::Configc regex-syntaxcUnicode Age propertyccc Thompson NFAccInput::anchoredcBoundedBacktrackerc2nfa::thompson::backtrack::Config::visited_capacityccutil::syntax::Configccnfa::thompson::Compilercccmeta regex APIs cc dfa::densecc dfa::onepassc8nfa::thompson::backtrack::BoundedBacktracker::try_searchcchybrid::regex::Builderc$c$ constructiondirectly from an HIRcc builder APIc regex crateccdfa::dense::DFA::from_bytescc hybrid::regexccmeta::Regex::findcc!manually construct a Thompson NFAcc$hybrid::dfa::Builder::build_from_nfacc$Unicode text segmentation algorithmscc API themescccc cݱccccccccccc ݱccccղcccٳccrate::util::syntaxthompson::Config::utf8thompson::Config dense::DFA sparse::DFA"dense::DFA::to_bytes_little_endianhttps://docs.rs/regex/1/regex/$dense::Config::unicode_word_boundary dense::Configcrate::util::syntax::Configcc regex::Regexcc!documentation for the regex cratecc!crate::util::syntax::Config::utf8cc5 crate::hybridcc crate::utilccsyntax::Configcc$05 hybrid modulecccrate::util::primitivesccsyntax::Config::utf8cc05cc crate::util::primitives::StateIDcc crate::util::prefilterHcc0c dense::DFA::from_bytes_uncheckedcc!crate::util::prefilter::PrefilterccHdense::Config::minimizedense::DFA::to_bytes_big_endiancc" crate::dfacHc crate::nfacdense::Builderccccrate::dfa::densecc!crate::nfa::thompsoncc onepass::DFAccccrate::dfa::dense::Configcc crate::nfa::thompson::Configcc one-passDFAccc0crate::dfa::dense::Config::unicode_word_boundarycc"crate::nfa::thompson::Config::utf8ccdense::DFA::from_bytescc,heuristically enable Unicode word boundariescccccccccccccccccccccDFA::from_bytesAutomaton::accelerator$https://crates.io/crates/lazy_staticdense::Builder::configureConfig::unicode_word_boundarycConfig::dfa_size_limitBuilder::build DFA::stridecccrate::util::lazyc DFA::configccccccc5cccrate::MatchError::quitc%Config::prefiltercAutomaton::try_search_fwdcc util::lazy5cccccccStartKind::Bothc5c%cDFA::to_bytes_native_endianccccrate::dfa::regex::Regexcregex::BuilderDFA::to_bytes_little_endiancc. crate::InputccccMatchKind::LeftmostFirstc.ccBuilder::syntaxcDFA::write_to_big_endianccccrate::util::wire::AlignAsConfig::starts_for_each_pattern.c$ccnfa::thompson::BuildErrorccc%ccBuilder::build_from_nfaccDFA::from_bytes_uncheckedc once_cellc$self::Builder::configurecccccccConfig::start_kindcc DFA::stride2c lazy_static Config::newcccccccccccrate::MatchErrorcConfig::determinize_size_limitccConfig::byte_classesccccrate::util::lazy::Lazyccccccccccc$c0cMatchError::quitc%cccccDFA::newcutil::lazy::Lazyc0cc ccHc!ccrate::dfa::regex0c!crate::dfa::regex::BuilderHcDFA::to_bytes_big_endianccccHccMatchKind::AllcBuilder::thompsonccDFA::write_to_native_endianccccrate::util::wireAnchored::Patterncthompson::BuildErrorc ccConfig::specialize_start_statescc%c"cBuilder::configureccDFA::write_to_little_endianccc"https://crates.io/crates/once_cell self::Buildercc%ccccccccccccccccccccc parent moduledense::Config::prefilterccccccccccccccc5Builder::build_manycc%ccccc$5Regex::find_itercccccccBuilder::build_from_dfasc5cc %cHcccrate::util::iterccHccrate::util::iter::SearchercccHc util::iterccc.cccutil::iter::Searchercccc.cc c Regex::findccccc%.ccRegex::try_searchcccccccccc%cccccccccccWccccsparse::DFA::newdense::DFA::to_sparsedense::Config::byte_classesccHcccccccccccccccccc5ccccccccc5ccccccc%5cccccc%cHccccccccccccccHccccccAutomaton::start_statecccrate::dfa::dense::DFAcAutomaton::pattern_lenAutomaton::match_patternstarting configurationdcrate::MatchKind::LeftmostFirstdcdcŁdddAutomaton::is_accel_statecŁdcrate::MatchKind::Alld%̅Ђdcd%̅Automaton::is_match_stateЂddense::Config::accelerated%c̅dcdcAutomaton::is_dead_stated2crate::dfa::dense::Config::specialize_start_statescd.cc!dccdc.cAutomaton::next_eoi_statedccrate::dfa::dense::Config::quitcd.OverlappingState::get_matchccdccφdcd%cAutomaton::start_state_forwarddcstartconfigurationφdnfa::thompson::Config::reversedccdccdňdccAutomaton::match_lendc start::Configd%crate::dfa::dense::Config::start_kindcňdcdcccӉddcccd!Automaton::next_statecӉd#crate::MatchKinddcccc!dccdcccAutomaton::is_start_state!d%crate::dfa::dense::Config::acceleratecd%c !cHcdcdc!cHAutomaton::is_quit_stated$crate::dfa::dense::Config::prefiltercdcHdcccˎdccAutomaton::is_special_state$dcspecialization of startstatescˎdOverlappingState::startcc ӏdccdcdcAutomaton::start_state_reverseӏdcdense::Config::quitd%crate::nfa::thompson::Config::reversedcdccccƑddddcdcƑdcddcd crate::Anchoredd%d-5 Glushkov NFAAhttps://en.wikipedia.org/wiki/Glushkov%27s_construction_algorithmccrate::dfa::sparse::DFAcrate::dfa::sparseccthompson::pikevm::PikeVMthompson::pikevmhybrid NFA/DFA ThompsonNFA7https://en.wikipedia.org/wiki/Thompson%27s_constructiond5dccϔddcdddוddd5dc ϔdcdcdddוddNregex_syntax::ast ccrate::dfa::onepassfull DFA crate::metaregex-syntax::astregex_syntax::ast::Ast c dregex-syntax::hir::Hirc one-pass DFAddregex-syntax::ast::Ast˜d d regex_syntax::hircddbacktrack::BoundedBacktrackerЙdc  d regex_syntax::hir::HirdcPikeVMpikevm::PikeVMdЙddc dregex-syntax::hirdddddc  cddؘdcdddd dcdd˜dؘdcddcBuilder::set_size_limitBuilder::start_patternBuilder::finish_patternBuilder::patchBuilder::clear crate::util::captures::GroupInfopbcc nfa::StateŝddBuilder::add_matchddcbccthompson::StatedŝdbBuilder::add_failddcccccrate::nfa::thompson::Compiler؟ddbcBuilder::add_capture_endddabccc$ccthompson::Compiler̠d؟dbBuilder::add_capture_startdbdabc cccd̠dbBuilder::add_lookddabcdbBuilder::add_sparsecddBuilder::set_reverseacBuilder::add_rangedd.cNFA::is_reversedabccBuilder::add_union_reverseܤddc.Builder::set_utf8ddaabccBuilder::add_uniondܤd StateID::ZERO. "ddabcBuilder::add_emptyddUcrate::util::capturesb٦ddadddUdɧdb٦dpdddUddɧd bpb{4c cfcrate::nfa::thompson::StateWhichCaptures::NoneWhichCaptures::Allccf$crate::nfa::thompson::State::CaptureddcccfState::Capturecddccc$dConfig::which_capturesdcdcccConfig::reverseddc dthompson::BuilderccdcConfig::capturesdddܸccrate::meta::RegexŬdcdcddcrate::nfa::thompson::backtrackܸcdŬddccd3crate::nfa::thompson::backtrack::BoundedBacktrackerdܸcdcccrate::nfa::thompson::pikevmܮdddcc$crate::nfa::thompson::pikevm::PikeVMd,ܮdddcWhichCaptures::Implicitdd,4d dcédǰdd,4dc dédǰdccregex_syntax::ErrorConfig::nfa_size_limitcccddcccdcdNFA::start_anchored Config::utf8Config::look_matcher  State::Dense NFA::statedӳd NFA::patterns State::Sparsed NFA::statesddcalphabet::UnitddcִdPatternID::LIMITddcddalphabet::Unit::eoiddcdִd NFA::new_many޵ddddcdddc !crate::hybrid::regexd޵ddګddcrate::hybrid::regex::Regexddګdcrate::util::captures::Capturesָdcdګ`dccָdd`dccc State::Uniondd,`cccBuilder::set_look_matcherdɧddd,cc ddɧdd d,ӳdcd dɧd ! Unit::eoicrate::nfa::thompson::NFAByteClasses::iterByteClasses::representativesByteClasses::elementsc˪c¬d˼ddddc˪c¬d˼dddd˪cc¬Captures::set_patternthompson::NFA::group_info thompson::NFAdCaptures::extract_bytesCaptures::extractCaptures::get_groupGroupInfo::all_namesdddinterpolate module˿ddGroupInfo::pattern_namesdddCaptures::iterd˿dcGroupInfo::slot_len޵dddddcrate::nfa::thompson::BuilderddcGroupInfo::group_len޵ddd˼ddCaptures::group_len!ŬdddcGroupInfo::slotsd޵dd˼dd5crate::nfa::thompson::pikevm::PikeVM::create_capturesdŬddGroupInfo::slotdd˼dPikeVM::create_capturesddŬdnfa::thompson::Stateddc.Captures::emptyddnfa::thompson::State::Captureddcc.Captures::matchesdddCaptures::slotsddcc. Captures::allddddCaptures::get_matchddc%crate::nfa::thompson::NFA::group_infoddddddddddddddddddd הmodule documentation ה dהdEdfa::regex::FindMatchescrate::dfa::regex::FindMatchescSearcher::advance Searcher::into_half_matches_itercSearcher::into_captures_iterc Searcher::newddc$TryCapturesIter::infallibleddSearcher::into_matches_iterddcc TryMatchesIter::infallibled%dddcTryHalfMatchesIter::infallibleddddd%dcddddcFhttps://matklad.github.io/2020/01/02/spinlocks-considered-harmful.htmlSpinlocks Considered Harmfulddddo˼dLook::WordStartAsciiLook::WordEndAsciiLook::WordStartUnicode  Look::EndLFLook::StartCRLF Look::EndCRLF cdLook::WordStartHalfUnicode'crate::nfa::thompson::NFA::look_set_anyLookMatcher::matches_set LookSet::iterLook::WordEndHalfUnicodeLook::WordUnicodeNegatedd  Look::StartLFccdNFA::look_set_anyddd Look::WordEndHalfAsciiLook::WordUnicodedd ddd Look::End dd ccLook::from_reprddLook::WordStartHalfAsciidLook::WordAsciiNegatedd d dd Look::Startdd cd ddLook::WordEndUnicodedLook::WordAsciidd d ddLookSet::availabled˼ddddddd dddd dddd˼ddddd ddd dddddcrate::hybrid::regex::CachePoolGuard::putcddCacheddcdddddddcdddd3ddddPrefilter::newcPrefilter::from_hir_prefixccregex-syntax::hir::literalregex_syntax::hir::literalPrefilter::from_hirs_prefixddddddcddcdcdddddd!dddcccddd$SmallIndex::MAXSmallIndex::newSmallIndex::new_uncheckedutil::primitivesccddddc!!c dddcdc"! Config::anchoredConfig::look_behindddd#d$https://docs.rs/regex-syntaxcConfig::unicoderegex-syntax::ParserBuilderregex_syntax::ParserBuilderEhttps://docs.rs/regex-syntax/*/regex_syntax/struct.ParserBuilder.html$cdddddcddddd"crate::dfa::dense::DFA::from_bytesccdcc dense DFAsdenseDFAdcccdcdddcc cdcddInput::set_anchoredInput::get_spanInput::get_anchoredccPatternSet::iter#MatchErrorKind::UnsupportedAnchoredcrate::dfa::onepass::DFA c2crate::dfa::dense::Config::starts_for_each_patterncٳcInput::set_spanddcccPatternSet::try_insertdd&dfa::dense::Configdcٳc Input::newdd Match::newccddc+dfa::dense::Config::starts_for_each_pattern%d%cddd޵dHalfMatch::newdcddc3crate::hybrid::dfa::Config::starts_for_each_patternddd޵d Anchored::Noddcd,hybrid::dfa::Config::starts_for_each_patterndd!%c޵d Anchored::YesddcdMatchErrorKind::HaystackTooLongddcInput::is_donedd%cMatchErrorKind::GaveUpdddcdcccInput::is_char_boundarydd%cMatchErrorKind::Quitdd%%ܮddcd$dcc Input::endddcccrate::hybrid::dfacdd%dܮdc dddcc Input::startddccccrate::hybrid::dfa::Configdddܮddd%ccInput::haystackdd$cc%1crate::hybrid::dfa::Config::unicode_word_boundarydcd%ddcccddd% dddccdddd cdddcٳcP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQgP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQP  ####!!!PQPPQ$P  ####!!!PQPPQrust-regex-automata-dfa-denseConfigpreminimize match_kind start_kind starts_for_each_pattern byte_classes unicode_word_boundaryquitsetspecialize_start_statesdfa_size_limitdeterminize_size_limit accelerate Builderconfigthompson dense DFA )assertion failed: self.is_match_state(id)) dense::DFA(    i:  START-GROUP(anchored) START-GROUP(unanchored) START_GROUP(pattern: )   => i MATCH( i): , state length:  pattern length:   flags:   ) transition table#assertion failed: self.is_valid(id)#starting table ids MatchStates slices pattern_ids pattern_len match states no match states to index)assertion failed: dfa.is_match_state(sid))Flags has_empty is_utf8is_always_start_anchored flag bitset StateTransitionIterlenitStateSparseTransitionIterdensecur BuildError kindDeterminizeExceededSizeLimitlimitDFAExceededSizeLimitTooManyMatchPatternIDsTooManyStartStates TooManyStates Unsupported NFARegexforwardreverse0reverse search must match if forward search does02forward and reverse search must match same pattern20assertion failed: start.offset() <= end.offset()0 FindMatches redfarust-regex-automata-dfa-sparse(internal error: entered unreachable code($transition length should be non-zero$,expected transition length to be <= 257,pattern ID length fits in u32 sparse::DFA(   i:  sparse transition tablesparse starting table ids%start states for each pattern enabled%invalid pattern ID StartStateIteri StateIter idsparse state transitionsSeensetinvalid accelerator index &Accels must contain valid accelerators&:7expected accelerator bytes written to be a multiple of : accelerators Accels() IterAccels accels$no quit in start without look-behind$OverlappingStatematatnext_match_indexrev_eoiUnsupportedAnchoredmodeQuitbyteD> Q A> >A* *A  quitRunnernfabuilder_statescachememory_usage_statesparsesstackscratch_state_builderStateSetidsRemappermapidxmap IndexMapper stride2*assertion failed: !dfa.is_match_state(sid)*Specialquit_id min_match max_match min_accel max_accel min_start max_start maxspecial state ids)expected to write certain number of bytes)8expected to write multiple of 8 bytes for special states8Anchored Unanchored Both start kind Match pattern_id Fail UnionReverse alternates Union CaptureEnd group_index next CaptureStart LooklookSparse transitions ByteRange transEmptystates start_pattern captures memory_states utf8 look_matcher size_limit nfa_size_limitshrinkwhich_capturesNoneImplicitAllCompilerbuilder utf8_state trie_state utf8_suffix parser ThompsonRef startend Utf8Compiler statetarget Utf8State compiled uncompiled Utf8NodelastUtf8LastTransitionUnsupportedCapturesInvalidCaptureIndexindexExceededSizeLimitgivenTooManyPatternsWordCapturesSyntaxFramechunksunionsparseStateChunksIteractiveUtf8BoundedMapversioncapacityUtf8BoundedEntrykeyval Utf8SuffixMap Utf8SuffixKey fromUtf8SuffixEntrySparseTransitionsDenseTransitions PatternIter _markerNextDupeold_idnew_idNextIterstate_idtidx NextInsert rangesSplit partitions NewOld(start range must be less than usize::MAX( ByteClassIter classesByteClassRepresentativescur_byteend_byte last_class ByteClassElementsclassByteClassElementRangeselementsrange ByteClassSet ByteSetbitsbyte set2expected to write multiple of 8 bytes for byte set2 ByteSetIter bByteSetRangeItera matchtoo few matching groupsCapturesPatternItercapsnames GroupInfo GroupInfoInner slot_ranges name_to_index index_to_name memory_extra GroupInfoError Duplicate patternnameFirstMustBeUnnamed MissingGroups TooManyGroups minimumerrGroupInfoPatternNamesGroupInfoAllNames group_info pids current_pid $)assertion failed: !replacement.is_empty()) CaptureRef capNumberNamedSearcherinputlast_match_end["unexpected regex half find error: 5 to handle find errors, use 'try' or 'search' methods[Vunexpected regex find error: 5 to handle find errors, use 'try' or 'search' methodsVassertion failed: m.is_empty()TryHalfMatchesIterfinder  HalfMatchesIterTryMatchesIter MatchesIter TryCapturesIter CapturesIter Z!unexpected regex captures error: 5 to handle find errors, use 'try' or 'search' methodsZLazydataWordEndHalfUnicodeWordStartHalfUnicodeWordEndHalfAsciiWordStartHalfAsciiWordEndUnicodeWordStartUnicode WordEndAscii WordStartAsciiWordUnicodeNegate WordUnicode WordAsciiNegate WordAscii EndCRLF StartCRLF EndLFStartLFEndStart LookSetIter LookMatcher linetermUnicodeWordBoundaryErrorPool PoolGuard +regex: thread ID allocation space exhausted+ CacheLine stacksowner owner_val poolvalue AhoCorasick _unusedMemchrMemchr2Memchr3MemmemTeddy Prefilter is_fastmax_needle_len SmallIndex invalid small indexSmallIndexError attempted SmallIndexIterrnginvalid PatternID valuePatternIDError PatternIDIter WithPatternIDIterinvalid StateID value StateIDError StateIDIter WithStateIDIter look_behind anchoredCustomLineTerminatorLineCRLineLFTextWordByte NonWordByte multi_line dot_matches_new_linecrlfline_terminator swap_greed ignore_whitespaceunicode nest_limit octalcase_insensitiveAlignAs_alignbytesSerializeErrorwhatDeserializeErrorStateID PatternID ArithmeticOverflow LabelMismatch expectedAlignmentMismatch alignment addressEndianMismatchfoundVersionMismatch InvalidUsize BufferTooSmallGenericmsg padding of  is bigger than 7 - padding of  is bigger than extra  bytes-41expected end of initial padding to be aligned to 4endianness checkversion numberStateBuilderEmpty* invalid span  for haystack of length * HalfMatch offsetspaninvalid match span PatternSet whichPatternSetInsertErrorPatternSetIterPatternYesNo LeftmostFirst MatchError HaystackTooLongGaveUp SparseSets set1set2 SparseSetIter #"eˉe؉eeeeeeĊeeeeeeϋe݋eeeeěeԌeeeečeލeeeeeeӎe܎eeeeȏeՏeeeeeېeeeeeeБeڑeeeeeeeؒeeeeeɓeۓeeeeeĔeeeƕeϕeٕeeeؖeee×eؗeee˜eߘeeeeeeeəeeeۚeeeeee͛eeeeeeee̜eלeߜeeeeeeeeeeeƝeНeeeeeeeϞeٞeeeeeeПeޟeeeeeee̠e֠eeeݡeeeeeeeeԢeeeeeeeeȣeգeeeeeeeeӤeޤeeeeeeĥeӥeݥeeeeeeeͦe٦eeeeeeeeΧe٧eeeeeʨe֨eeeeeeeeũe۩eeeeeeeŪe٪eeeeeeϫeݫeeeeeeeeͬe٬eeeeeeϭeeeeeeeޮeeeeeeeeeeeeʰeeeeeeeƱeڱeeeeeeIJeزeeeeeeɳeԳeeeeeeƴeҴeeeeeeeeeeeǷeٷeeeeeeeeĹeݹeeeeeɺeۺeeeeeeeûeϻeeeeeeeeļeeeeeeeeeƽeҽeeeeeeeeľeҾeeeeeeͿe׿eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee0W4=~5Zv5Y90ʹq}c+61_98X4I `2s;؀U}Ms7(s n>Fɱ{s:.@8\;xAd\uᙶE^q={Mg{ۦ؛wzh5Z`mU91D\9yGcuד*A.'~ֆxMx>y:4$#VD*^DN_{b?%4;ͼeC3׉WV2}R~ĿӐۙٿ$'.gE|$= 0')׍hh[0neqP-|Q&OF~+ha `[;D \}P}zN|M khYt3 /죤AɱdAu%zV6]٩c>ٚow{xٵehjT$%<n<ֿmru0)h %w,6>* ֪|zަ{\Ύ$r$Y=sVc =M?LشsYPt 8-"gM|%bEx;hY\n1r ^/_$;tȿM-ľ[ ǫy a:Qn\rF)(}vA쩐t~?"oƑnTQűT!pѮQ4!˘ iz[_pb~ٳt:-t8i wJX'=y!RUGճ @yǷeپdYr`Fb%R)a>]YFX ;Rrl_>գ'z75gf_1= ԐBEG(Wq:sڥ}Zr@&WVte}q&bQƌyL`n ?F]gP>Lk >$n⁅qU=M=E@"^7U,(@/Q6&*3?2͓Un|WM]BSW!Zv}UCcn iP@N^.:d[wޣ߰a_(F2uPǯ7Qx?w:qQ?I (!qHk}yKOy7VPפG,NslWڜ)Am9Nú_mm2v\u<ij`B3V)a9{Zhfʹ9 V/b[qq]N 3UH,TFQJ!(ho:rWkFy<ňjnH5-;P=% (dx)U/tz.?fKY)VI[ஶ CGfmλ#lěBEn yk,# RىALIPL~Mf1΂xZ+Q40{&?$Gt@O-w<wP10t޲#a>($vgJ%ǰ73?r]A5sSؕ4ʢ}Ru eF`٫ee(Ww@A SԟVNJfeFfKw#Ʃ\{:k40{v&;?:J./izY_R%07=o(0DK6H>]/\|-.&ŕ~sU h;}l ǁݛIyuB+<[2pHrMJ)=ZR&Ss8RzHd:c&n)kH֝D)싏V^<"I?8y[<ŀ;QU/?SO賛xWK{is'=rԄoJ+;|R_(Z>>*,nB1p͈sXFw a?ў S1O0Ãv_TPoiwX.jP5 j@0%ٟ&y$n!$.Hc؞dr Y5 ecd£*_o *1gH= 7pʲ0H[]֜uϸVͣ-]ڭz@V#U?c2`fnu0}I+Ƚ>Db)cqO>7 ;VM%RQʨ P4N4nJ4hkZ 4;6|%%옭uoIyLљ+G_OttQ^ɹr 26[_Ywfi 73k x㊰a_N7nv6]DREc\F}S-J>QO"Xؙeq.!A"rE_g[֙<=k?AjFgTETpeM|*ܹw\2kBџlf1{o]Z`AsWx}ks6JEnw/.M7}szk/=ԕfJ a:o"">CiORx/V2A C_WG%B+6jBKèП팋QiDV5ȃuhiy ( oԥNc?d$ l&JKiRvyuqγh.W_8g# nm^֍DKQ;hr[L@WUѾp~3LTY5%ȭ)z5 `qvޅc+eVeր3Y"|,vϮ!R&`R@WhoXd`:uT;ѩ^a+)#B0Run=ƵHCbHS2LYLOo]I$4^5q!~Ɩdd@WcY$6G]RHX`}V}Vs6"@RI1}}oun-kwpe!DZx'cmb}[V鹫7Iuy N<\r9FUS0|^Fhߐ8 08/@,#N$6Ë)CUۍ8TX[&UВd>QgapSO C-DUV4N`{𴸬= ʙEtO5쌒zdQ8}WBk\Хk~zω{?M0Rd5=<]E|O>w ,ˆ=ղ7=P O'Km'c^#3Ծ}c:lNBe&')Q}/MF'FDmFr`{)`Y 5KKR`ZM 3ޘ*K4l&Tm Su1=&2݂"M7;n$e2f|ś˶KQ 8gR(iI\ ݈cR]0=y0Х7/Q7O|(KTJ0XD:sڲQ$LvH'`@65@o\rڐPC&նcͫmk+Zp{ay9.>ݓ!zBW8"d#ƹªIp"wYG6'i8pɱ}ΑZHRՆ׃$9#|4kZ);^42ki5G_.d?xl>*!_csDTÖ-Hoa@~O NN9_F91GP=CYp3Տ\X-mhCŁ hbDwU*;tQxv,3󹂐Sw@r;CA0O&BdwgbOGH5|/;j%HFɑ-gRҲy/u܋Y=aCD!l㺂Ng>oR,ony̵\畁,lB޽RPz!Cr5?/ai$ 僌 7+~yj `j.Pd3 2mrz2Yx ޓF?Eζ ;D%j,Ѱ}fYBYs_퍦^Ө]׮$#rt%M>aQ cz/8.Ts?pp~9@)%rc8q5vϩS $!c v_\Wc CvC6] e;L2.纕e¢{?a"B\%Gvyl J^?w;8yi>8۞0=%;15"'Wv:6Kd@cC  P嫑Nb "(9Y qVc/?;zZJB3Z>j:D0 abN,e ێAe hH9#p*m2:]WQ (A v<q8a1Z\)3Yp4<ܻA8D*y1 Ei Z#%̓~Wwƣ7JI-.޼$@s'GK[T/2_Ln#n-.3lm :~b" eB 'l FM#OOm>SUvFdsp;:,QZ- tŗa_h| m,W{X7Rc8 =M ӁC2fR\ɺN~QP((!k`%DuHttڽba4ǿiGYu |FvjYZ#:3YK'nT[ Hɹ.hEa#AT_gڀZF -1'Q*?FXK}GAk|߸75t޾v`F]RF^哘L0rYs3 \)boWkInED\4)ܶ30,YIj΅Yrg+m!I#S͎"_EZ7}qɃ, ipwh BS%Aca^=mĔrAϧS5]rIp\"C5$Q9u%?U׻hC\r P)R@UG뻞dT#6o9Wa$K n$2IQڃC=9EeAöъd |gYk'3BpݗfKWplzWنUpnl:^c_e~XV01Vb?K]|`x-]YM XSDD؎`3O I@;}lxȫ`䧣oiVtž${lb4k: rh 7$rsN=};L28xp>cVRZ&ֺ.𱐌! q.KHMh{F] Õ]ODH {NQ/ܛNOwtq_¯š4}W]I|cgkXXN7<ċ)Dy 8_q$IvpAXLN{+VV.Mɧ"FI8VnJ_3+^zK|5!!ÓASS9?R_TS$5tb#826[Uuqdw & l\GBjDҐ%p+fsw*dҟp%a򾧄Nvj[+*ך9.r?ζ .ZkFgڤX'Ԅ\i4_9^HM8DK롻Od8"3A^̶ē͚ҟHƧBÖLHYR5 G'ZJf- N}V#3V`HȨDU߁:s9R;TCI2- \r/*lNdY-dIx_ PۡȬIژ ;nTˑk{c[*:bۀ;Saq6~rv13Ş ; moC!)()Wb걙wm^y9]iafQ?J>|_(.Z^+s䶹}gCx>kqVZ3X=cc-q0WƱ\ٚOƝzVzW=T60a$Ћ=vGtȻPΠB|KjF6[ Gcw d=:\Be$}&mFcR¸FNN%mݦjR$39B!NhlU5DD'k1t .}:EV8)k@w%+{\ + /'?Bzs$bN?` fC|[N8'.~Gܤи]ɩ9 *%! V>?O+%Ʀ'|u̴0/jv?pG^5NOUDSձ&і֪bi+= nDN$phxQſDv UyvvZ~&̗5'- `=J{?b\S0TR2h@m$J6@fŋB *;Tg?ww([Lj(+|r,B}:yП8Olj)!YI=\nX _ARd}6=vE [_m@6RG+e|ۮPZ~LKyIĆx^lX3W5 ,I@1gXnΉySWbAT9~|_OQ FwE[uhmBM:4HImdE>϶Gopjc@' \!e$"Q?|:m@ͥC,dÌqO?*`֯77'AJ ~@'f\麮QqXǙHKŌTXI)erI qW-CSz(ߋw65Ŝq}H5]kNw`~dr' "SQ>̥HQ/$p&f_<~ܰYUT˫2[S{ṁ]DpTt1p\|q02:݁iOLJTǔ"؜lwlт҇lg(n`L &ϝ`#^(WMoVwB ^-&!C1XE\KD^:'2~rŠ伝_-l\ rnyϮ<ͱC$ t7GPY’RQh7n9tv^V?\ڳd0*^]5e,#rHoA0wO^5Wlү]m*"ronla.@Ԭ,cc,qV E_g$hgG ]thҧߟ% H;I@Cq q8NKFBE9؈uŤ{rOXk{=ώv!Rw բ ڴ/ޚMGj̳c Kj.Vӫu/o[oͲ@f4!,V'+b1CɂӼhS9a+!ƨ5/йΥ )%0AQ8oD{$2\jO V)T(e@rD`V? OiPRG󋕁hAJ R:Z9gS*I"2˗Lm(+%8OYnDdkEc"֕A@`®Ws'0!P[p%uW]F^Q}f~D(ND }9;'nتλXd8/ҴG= hm4_NΙ8dk9=7 wǍBG ƛ!_r0Bh< _ChJVsySJj jbv\hwj,qG~%z}18t52_ +S[R.>ص4eZ&4_?TVHS]ݔkXCws{5n}7a38/!M]uAru_Ά?-6BC]ƿUqdTj W @zذ8~+AeO!ı7 L\#?0x[?ق>e{ufJ5bW->HXkvl>}{5>&9uטz4Nw]\E|'*/ e;X*q~/HBI#araDEk}d|QOA]s)ĩOF,Fʥ{+UP{J;zeEgbEo>X}i6^äpaZ5]Q)D\ |rƪD8&W+gTG7w\#D*NsP;SЌ5_,uXXl؋C6s޻ $f͍6ٔ|@!) |AuthM#䥚D\ިXyiftW hv9\!!ݴDmq`!FS< DX) '1(|50yN.ePzz|L0Yz[Cgc]U5փ7{Q} aɝ19s?Ч!N`R<nlP(e=S_ \Y4i- É6p'!@|͙~wfO)=D?m.qObRghV`}I37 b]L$CS&gPY s|ZqCD.|^ /mp3C0A<Uɂ>;fSp_ ̊)s©+N*R9.)^gD&PEف8tU%`YA46skh$M 1 "ԀT[˿OZ[ckGMm_JjF?J]O[.18=j^vi_Q?HSȸP7}ᮙiffB=RPe$5ܣ[S~y$\)NضŹ5uxq3.b(I\?|DϦ\,bic뷯)=d7i)쎸Q/,p3s//8'=nɆ3;߬Ѡ2s59FPFmEa 5Rp @)}x['{2tdW[aa\Ւ, C;[,n@,뙉toᵩ@,OxSwu]|Kx_?&_O-@} $ gfK2gPmQ(rnM(OWc2JՆD{0u&0<>!> 鴦5ѻB*A;1T1,734mqUhJF{c{ax/\ع䞮 4$¾n^?|ѰUد.9=&*z ͙?&{ wrApLN@3ŋtR_;#ⲼR. 2r_.+;l3"3?9nt-kl XM_e  | K T!fvhۘr.[z{$$"Ɵ }h~=* PIJ2↖~dԩ\b&nW0V9,4M@_PEF8Ejsl{.3`Nyxޑ,GˌCc9Sy+ 5!AP]{WTi$ Ɨ <׆G\gF\ay6/` ԅF H4)fbmE*O s!`6Fkξ>D nFPsE|پ&rW@>&>Q6g7dߛ6TT .Ȑ)sG`&DЮ#Z1V {t\f|N  _nv폥qUr n9hi^ӭHIo}Ot` ʮCH[cXZjuE+jK5?a{XTKL@w:E* 󑦜'οA>ogڲ^J+|<ĊzugWrk(ɔ(6#ygA&U{֌,ǔjvre6f[e;dORʰ ;MWE ;: ηUƾ1Y#1B] DGc'}9.UNOm^by^eҧx‹"H!<=\qK^ YLBYy]PP'#2+V٦Cg!ɽ"ttB;\+joFXHu%4q)8Iz"褼0rҐ8Gvk&jɚb/6|~h'RLoNx\ؽUo˧h8##EV~UFHVܶCf ^[ZN;%?[:٘dw?H;~TgT*:i){&vT3d|^p:`F]ֲ+<|׎b|X<|alU-bAgbJ FW ܀ !Gڎ] =\c0n4I2䁯I+{"\c4)Uc"q :a(kj KZN1BFg#zMժI`S@ZMj2jFQ8SwxǛXB[=^iAӟy̌2 դUPe_03/ig F漡`;e`?_ Ynӻ:tYxte AnQ6wJɴ||pd:46R;K38Gm9-F:K0}ҋ+3%hT€2Oda_@$5Bv<3թ*ȄE}[l$JX-nINb4OvG4-G 96tƓ! xh?non3M(g7 ^:-a V@ ݫ&$Fs%zaqYϘY< %o#H3s bPf$.@ ngp@6,.6rȬJ񷃆C駼Cw{R)nIw3DcUpjn[e`V:A?2FoQjY-\Bvv71BY1eSMTjauk{yGYYt1#ɭwa2m#0 OG ⇙^K0'%I`6uT8P\g>팀CԜy! ~βxj*z]=<2 =.$YCzp!$E&wAIr c[[-;B2m$7TP/)a(:7:&DHNAtCo;ʅxNAa?Cr ZcnF0R/K9ظ:V%roէCTHʚJHh=nbuqM@. GXBe` 7P fΚ|<٦8,oZlAy2Y{u9~B%: /ܫei@;2 AQm[>d`y襤xB-Q~V=5?4OlD=7"GoG1m<1b~t<Fw0ZtgE]^+x0/V"/?c_dνS9]-\܆oglI]W~)ldq8,|?DCK!cA`Z+ˌǿIOS<iT,"T~j'3_N1c1`>G\'0wTtRvv) v;ȏYuF?}b"!'z(̇61\li/ {MTbݡorA {@z_Aa~VYxv\*:vIBzǴ4cJ+-MޡzWZ OPIo }C$t3UPwth@kl({.Y6R?N~7>̢x\d^-:e cw-P:) $iZL//"Ǽ4H)nZ0;X ˌ['dsc =N5$Tp:*iTɚY.ϴ)(M\slz#<3G=*:.M&q`4D~ɉhށ?Ш Sq1p}C+ Ћd0 |'[E! WTY bmFCm/4~ S,n6gu5[ oqnN@#XH7p!KXEMs8/t~[x=Cb/OR:9o\h ܣSB5=d+Qu7^-|?^̪g$R@B\7I?M]KmُCOu >8=X߭K&ֽ@L]@Pm ";A]DPc7Vx/fYvCk“`c e2ұcpuvɡw<hNc(ձ ³-{$EMc K}#Vnجn\ qz#{gt 6!`7 u WQ.77D.EFMDĦ>|1kkKE`ܫgI;F*;P(+a;IXKwh^&f.J^dj~]QyS5h3vdEу5ě6A3zK+ !T73w|/t5ʝт9RKUf&"tKF |jVT[fz!J`ESZS`,.Y]U҂4dsczꝑ$5m]|PꫜzƂZDc~mv : (^Ñ +ڡ`|ZG\]qtyI~ {QX l dᗟgOj|~}KjU gQEc0, 7R4;oO .pE"%JiZV|y]"[Or&VEʡU!wK@P H8YHG'?Hijɻ\J kqv:>(.abF`KCDd00u: AF)H. c۷SV_;`6 DAss I;8 PN2"'-^thjoU& "W:;f ̇m^ w#$5T馗dwdj; bo(Җk5Z r-)rŁUcY >eZK%W$_SVیJ*P$-7ńzɔQm cR,Ɵ/Y $zz4CF$v#.u"Pd?/mgh|RJz3.8ȓ{3JorHym!SyL&e|y^^ ]A:fp4zOܲxw:{Ծm<\ѝ=vvۧRPMѠgW%HCRx{O5GhpxZgc60oOzFYfsƪ~ݔbo\is@N,.T`$L-RƱGã ƛvnQvэMϭ)R{+SVF0k?m+WLMcXNEkFeb'e^ z 8(0b$y [pf!'A(fem`4 ᘉh>ě+ r<ۍ+px?^ֲ~H{.`j Qkh2=GP%\eibN,O'qhvmԨz'tť4 {[^HQKJ^h HmaK:_!Ct(<-itjOSFӥhXhJX3-[9.>{&_ $EY&G'<\*aƨ6ݑ%1EXA2v,)LH[ iBkA~4q|Nrϕ8 poU!̿UgD+YwN5牳1"{C!‚+DbwqѢ;Mډ;.x8Ki:.AR]eh_b3$cNrVލ~y# uv4q{1-,dwceD-L- ZwB )Z lh'UwݼU!Әc^, }buZh^͐ @ `# Y7d5 +Q^޹'7bzg܂yL U98&a:0hFF>⣃q9CE3}$֙#'IoᘚpJi%-_بbНls ~KF|x$6!e7t:$>$$4|H"nhv=Jc EF\ ay.WyCl%<0{˪ ;*E|;yoH :FF~e4'%ay\-4Al+/wl<<`!t-Le=$J :.,08?Qh;/[=>$@7o:<58#U0f_8Zj,LLoBߣ ϘR$aDUR.U'e[Oq }SʂLfx/vx%On"\#^q^y_K=O/=|_IOf{T?6{Qjj`*JeEgi(5WQ&ik$^IuF+e=:dzCP(?:ҘfX,AJdO=8e䦯~KF;ֶk&mQtXu)V,,WY*7Wאьo>Y)KWz7JoYREE[) gwH"0l=b;KmOICPD$lyo?nD}QyqTBº|>wBeȧqduYXTA5促ևr_Vnj3Q4b'BDvmkm%S?Vc zkm F0Q^>#t% ѩ8G Ǯwq Cwǣ&<=λ J2]UҜ ]s.xJd'z\ZOnf?S%vy Y\Xc6ہ~5YDGyCȻNj=IqH/cGLˆW->$r7s 1 D"++0,w4#iwTf,зJf`HQu([f6BR ^_Ǝʴ/_tQD#z 5}❮ AJjWd2 [brd'GtUGї͹>Mҝ%X[[>ոDM#v}Ua%G^xbO'rꃒ-r8V4 nstzŸєDނkLCW"C؎&u9,~%\C=5*!=|6XaҶzj(A0+j[a%"gb|FKU $Kv<4oRu/3ϊdW`v-RS+AY1մ.AVRע"0K׻13)30X+骠~B4i\t< p7/Z^<5x=i'Jӏ/%OA~т蘩n rw`ҸWw~ii\|mI,DP&1RHT c%/2jkC+*s;U篼]E NG0x 43|B<uosƣ)dTK;|BP2}D+hv'AϕV!(.mэrR[,:o~J7IcجЫ*2Ńů|-!!!T A*Qkq=ۑgSnRH8,UKRxuoCBRR! h}\2{'g[d "@)mj!po'e6BnrB}͗H{;F=lr >^J4TmmԹ)YF** Ϙ٨^K6x{1X>Pls(A!@M~0l+:ƚ6\a҄tJ6 AX"o"ܟu9pd;RkP[x}/2B3MHk]7 fE1rWl' ʲƓ.ԟN;~Nv1(^Qe;]ߝu. Րnrb%rDxO3KwxD[:ES`^ps:{{f[eG CxC2"WhS:T PAPd=F.*kQhlp?Tzj`FС OmE+"[m Zxl@Llm"#jm)Nj`8S)>30Q PXv:￫WzvXV D6W([b8?k73A>{wfa] Kcsv{$T  <-];j&0fnkZy% 7В(K ZMIF^P Bv_c /u@ظ.>(2cu`.Vqnʮ%dV\IϫMW Aݛ&o[{8cHFv0i$~S7g"DE,: ,i<]?,zp8u%ܑӠJуe;T(ܶ"?^W.I te'˔1eiL\"q|{UpuP ,J;3[4VJfGShaTP}w]TTa7ǹ$AvNw5[x@o)aV(}m60/ J^|زgq=r覣.~o,[?a7_(- +y(_s{yZs}7|׀":"vۧ$eة|Nbd;)6+']A>~q5#}˸%Rʕr#ܕ ebd쮗]AV0ΌW]<ٻd:#&|$BdXr#+w*?5 Wӻ)ErRYdʹх 7d51YJݎ C0/o0\f{; ̆ >c#8M%S6az}w6g~Jټk=Ka6sQj1hI%UrXhƅv3I148^h6ym&aꃗ-`WK~Q=BF "NH?] )P BMmkPfu.;/n{~ 5" ޻G O2sŻx+СikqJM3>8u$E$JA\;JS6Ksm#L8gݷǥ^*4E/OQN%fV&9Y4{{CV0ؖ}a^w^[B(rmPUcfpŠ _VrpS,wHпR{dϸr8#3-[o>m* 4Jy~fyj[v v+Fd!R,p[_On}sm0pCr! >H^)*ӭcW텡p#P[*<ՄE U+2ߋ2@́~-AmYl'qQ-$2w!? 1զƪ:m>P3N؝~NZbs_-Ҙ;3 MydϗO\;wPB}J%ק$ͪ&[uCInWS71SXB_G}=a6FE<ҡg Dшdq'!:q]Z9Y"S8EL0Fݼ& -&T&E0wG(% ۳2qtfWsoYk/Ѣovk3&Pczw>丼DrS"Ueոwe/ K&p#!i*'G 2t&Y3Km.kǽPPZP[_d־1Y+R$:kG[eq(q)n*ʯoZ$eM#fo*4KBV3g㳋9^[HΙxx5sSYpO:6mF:l;+ginsڥxH>H.KsM?zD_oplM(vG#%ёs^EHvf#ؘA g^jg#9yPZe*{Jw2¬D2-J' k~job0NH\GH+A(BԺY'9tji 'W!*3ls)*)ȓ4ЃÙګ׃>S7#y6k3H2%?CprƆ:V t d*X0@&gRYn߇|H auɪRYքr#%s7|Rz*4"zH͔֟әd}/;*2DAИ)(1'Q-\ynI2@h_E=Yw^^\lHBkF;'~ޓbAղ| EZ$H-ďԳ]*RT֦e:ǐt)&d9c)U}f{2-GHqf>GJ\~`PJ`gKS՚gxLSa oؓ;@S~˦^x!7ůHN(k1Ƭ/l|I~%'Dg†D;Vfl-r[Ywkj$Ev5݁&=@;m)TpEUY'kKBU9Dq[40D,hʉ;tGFdI\JݖVܲ.{]3Ɯ[Arq0v و`Tܯ))]GOyha6s>MɐDž2b7+NNy;O0`_0_(˱cqUhu_xEPٌ{%@>{˨?HD|h>~Dz7Ntv=Ͷ& B}K5g 58>E.f=_W[̮:iMpX„0^CtlRTvU2 _F8E/V\aPcuWq~Aa9,cSsADL T?@y?]^RQhYe\i#3Rwg.*-ELj-uNm༊|mLSL`D/9,]Fj1guBVDN ?YCMg[u2e y岗He8Zw*iBUbHp雀ie"m8bUpgVޱO6~[8Sԋ=_5?R}&PQ҆v~Ǫe(5= .~l_yZydәr-EU!10ռ=LN'J{eKg7SF v8'HR|~sT2H5j,Q`Lڽ^w Iږ+ޝ!v>N*j:~im|q˺g=m}1Ʒ2ꢱu$gp"EHSL˽uigf0x%i Ss˃ /  P%2Z#>wg?[}GD>GᓙvM #ǘ TwTTE\^հuUN/Dl= :mRF iV3gA@>%'1IaN/Y#qfNZy㛥;f#X(5,7of."b8="Fgl~KQms5Qk7= oP#%#v߯qe&@SԏCGDXl6ѻ k 5P9`u0Zs2yJVbxZ~h͈7S&G@34.c \]G-= JB&O$CRԻZޖVõ2PJQk{&uӼ@)Ђ(8H>Då=A#ux2' N@y ;խ?_0r֧(})ΰ |5իjQL(p6ll UFcn&QF\KdwZq=#i[r1rI!eʂ4lZdGB]? x7auJA, ?o =k 懪ǁWR+hOQm~\Qkӂ1蟥uHɢ$g 鞤R(Iuʅ,0wVӓ(ٷA̽%WH}59}\\/\Z/sh>-3SPpK3kNI9\D5u"W%%$|ou"gWT @gqnx>Ř5t)&rĸ. ΔQǤ;o5'C ͟Ζ<6yNc%q,-XPEޖsY C΃RDAW#AV %`If -~x>e;n ]!k֢?90Ջɹ ucD4߻t^b'ZI2n w\k N(|%À:XR;9 {,nSU*[ M~b8u娖7$/פx7 !#x |~` eh nл &-no% J\l󊴞Z:0{ڀp2n!uG[sѿyj&]OkLA`*؎WNIOQeDFsUO&g>sǺQ/aX$8sWNa^eeZ@OB8= n78Q`L8"?FaHh C| MmΎOͯ;gX>C,bK8NONo޵D 1]oK7~Xj5ppQwXa6ָpa$lt7^{JyN_+&wQY_xI~W< 1Ͱ`!-_cl)G 5@Q:zn۲F :~h5 L#//N syDRxf={5v'Ddg% ΁S'hka.""Y[VF׺<;.):;jgF@i;W`1 Ғb =dP2<%Lͮ"> Ϳ1e5Nls >ydV$I\H7CZ ` ./taj:oU@$p\K&B\_:M Gcn!xa b=H_:k zFuй 833_Z1a|, jńj24?YKrR-qͺI:WO^R^̽8@wmYI趟T =x?O)Ta@|tjr ;qٺt.V^bb&VY|eQ(ˎ6zAn..'=2I"B}[_W) 6i<ځ`Hn̯-gP`Kְ.l$Յ` m|@KKk9EX7<]™|)*k`ns^ kF}Wj&7SEc?^=ㇿ^ vJ[o['.mP=Չ x@wd_.°Jj͍, .ZD:n9B%sȱ\4,5REOм6$n [Hۖ ߮HK <{1lbf`.%~`#"PM7;Qg߮$Y]]Wj 9`< N}M3q& sXmMڮٶ;M%a ilDD n>;즹34:[ؠ mQ޺ 2>Wœk*笗;?D|#.M;~zDK7\GU1*hA\|7у[ 7]*K 3Nr> a,eOGxӘgqٯh9MEZ/ cuLP٪(3@T 9%RA 8sY)'EFt'mG?iwy@H*IE /1$OOP͛m)|gNH3@:as8?sR +Ũ%? CWjB"U 5gFӥn`"-~RC,\d_:1K֯T?xu$k\J (V`qxQ W1H % )3 _wH%@u c۫|k*;^9Cdg[ ;3m?dW%<>Yu=N-_2t@OCJiE( PgS&"- >UrCdbSKn2FuR~=)\-kyӘ&!cMtݼm8!n!7NN K_@IXt]NGtL #cq.҃`Ju?񷧞(tk HCRwT k|^y1;~R{HXFr#uDTIDe81sňexneZ𚧐'6g%+ȳH(@8't3@Z6 U)Y,2HKa%  GIcRZpF?y%Z)ъcLA2b%ޮF^hk[q犀uw1tαbP9K Ь>xcUsKk4d:ۗNy 2Kw//pPbۃ,4ɓ˅k5cs;#pY/>W\jn&W˝a{2j@SG&H4`]t7%>߰D`>UGF\/ ipU_uwH۞y Y9{$]t=AJ9M?o^{{NF^$v#g(^QPWB!6OLC:V +MI_֦o`/ZGOԩ-fj7o5 }5`&hNGG]]p TmdliA$ 7H1r5FfvķqaNɧN Xevͭ>YXӶiVj-KrREqw*jұj}e GhkOD\R~X\J>dORH^#HX-'I|'3#!*~29OXU %}|>6I]Ai`n;OT,EfU[ܲ5Lw)my+lH^yAZae_Ϭqalp9bq'5 [Dƙ1Xg'!I:UF~[I UlJvg ]h(UgTB] HW(JxH|B sT ^]J}#Y`Dy>]C*r ~5) ٌ)P'*#a !8F4CP"lR@RocQ2; \3L?:>o ׁ$ͶU9@hz{l-M ,d{T+h҄ ݡ${pSFcp Vh$sbquEz3~Nl U~Q t)÷+asx0&3>ҹ1X;bmO-KNxk,[!y-ME@JT]2NFQ1w YܩVnĜkCK?;P= Gk 79 63ݩidz. kKVTP%yOKPR&Yo`vgs)?}>k='"ƄJ'C[VJУV|)!(e-Q ^O %5eӴKTTOf&nQl̚ ?RƂ"eS|avZc*_7^V.p8exU"wD KScieFf{(oa0UNĪ9,ep B88%u[l>9wgDFAdUbηH>]Ɍ3 )[2?M&ȍYɍU-ph WE ck]yTwz LӅ1{[.]%qcjo+ɑûfƾI]UgA;UA4=Fu2BA΃Z?ӱ(4kkaZ̴/o?W j+1EAUO^Xue]HrccJU^ p4}= xb~YF&| yTC}sʞ&!j}QjF]ϔTvSF֐<3>~:-!W } 9h&F:+k;U{m G}=l6LUW œ縖="P?LƠWn6qOHT %ұt~db]\H*w JU$b4?; fhR@G`[ˢ4`]\] ܏ }C`|KXc[gELߋU _2* ۱%i.몃!`Sh a_YcC~|;&jt#\"4s<>j%8z9C%r)oeԢA&6ϟ_m]E[5\&ך6#iܗe>)uɹCs-iy-Z+xtU,PSnNV<8 Iw!Dᯅ-XY!C؎,AOnŔH2X Yyo@c"5oRZz"U5/5.*T={]i]!e\Ӻ8t쥰);t@Yew,X^(Ze5x$gBgr 6Ks^ё C5:Op7Q'wHD| SƃOsnߪ,$@'S;FDʆA'BM"f;ݍlISډ|7_)J#'f8LO6ھU-Aa~xHz7]]riҤdB>(=,Mؤ/*%Co_ZC{:Ю\T|o ;c}qO^ 4TJzgj26CsMcA4OcTO|F"T4Cm`5"psX2gwz g@ J#K!3M-vc 'b/': Acg"8?v4ˌ }[c)KgElq6bu;zL84!];OSt9Ѣ"Z3 =RwՃ#/J-Z]Tc< [l 8kmuEa|`@M93Vb* [2}'0KxvՕ~0NɌL"I׶[0}'iz޻LӈW RD~bI.)fAYbrtgԻc@m!tY dt,RNMNM7\5[yRmo,HJ`{qaYBM}MnZT Ty_Q,|%=HlXh n%:שFBX#AlM@3΢-{byEގcf}r<0 co-xvקIuÂ΄+%Oaִ K.wCIp~\I|hb|3GҎa5Ɏ11@ˆPڻinOJu}<ЩH F&gHB/s:Ӕ.uwW)-jc(zHQ|+XlǗAoto[u.=~d;%ˈ0Su1>E>}z$ZD+tq_YFasUEC1'^ڑ.Xwg4! e#ȶ xYeZuv^^&7T\8g-ԿWaT"je >Z1Q0Qc#)iY]x|SYz+7u]Saj3Ts_>(*bذ[neZIrC "DD !k M=7T $xVn>n P=Ӧ[ˆ*Ru*H!:ݞak[T eA`^"XdKc|mV/WO̢#ЈDXX;N:p0mhn? 18V+A?ZXY+k3>9Qʂ5Q]LdɣZޥr;s,Fl!JFiM$?ӳ |}qKe}e[_xο}ڜWvoD37,~Ʉ{_JݲQibs1;MH5bᙤ;[2Jk|uMX( ${>+TmH܅Lu!O"@lm*6S;˓7--IM1xݑpI"EH,;;>GƟޟe9o8~lStQL9P@4rc'WE6<Ĩ 0T*Bڗl]LpeAWѯ*｡ 9T"Hl.T":9UU^a`9_QLy9` Ftɼvڼ2䙇(ܦ5j+UɅKpH$i+gCaKPY Es/'~.CGFR'D4ӓ^G!HrW*#_+ɀ_HpґAp쀮NVw-r^7EJs 'N6 9T4H6_ߥ53EtT,ƺs Vf=dsV{xr<t أ+PcLmR9@9'&F̀Ne^.*F"! sHJ-dJ=ٗ4ճ)-OgA-+F?'rظ]+HpJeXa >gCu,:5-(c:t!ς.47[㍅1Ag!+6y,R$|~߭vEcMa#l"3>Pnp(;cz _oL3%4nc'h᝵Sje?~$TQZKzRgP1\`dGM|JUT9d)EhzC`13yPY wLӔAWLFNLRnlBӭ#P<0tt-wE. ;$fFvfVg *+2gI-OM7B:ǰ P Ԏ=Fd̃<(/ nVO!@Zaˍca1Wt4i$B7{q$=,h N..hoFi@RgV&c,WRyz4jih2(/v\2ɣUD8Y?.F'eq:L z>| 36Ĵl؇C?T_#s^sLB WrU LTm0g rq򄅕pB2t`(mA;[*j\8Hgڂuى\T碲n#]v/+iKJ &uk.׎K|(?b×2ґ qXڮúo@pͧsܢ2քヽ>)8\1GK>l]2A#C6ê:v\S=/eu37\@jBv$ grr%u*AFSQ5;$BwADZq p$$4RO&E=VUSg6lVd'e, N,:]H-rHBvl_r_\5Se _hk6\&^3bNεv&$roNobqĨ,zKirP G:d88?Aoa2itl1Mܵ*\o|Q)`JbxشQ_%{;u9:XӄGe0}C(%yH68'X@|zcYK v#Pᓃ,{^u,F׸K#ƺ5JA'w_BNuW,#YyG9M`'Wq%}=8w(i_)"ۿaFkLdR[<_diYolug~DKTX;S,~Bz?y ~ꃃ&tef5ͬ#W(՜L0~f5kz}f14DoC xKյkǏiP$Vqwe|3If>1XHvۜ~~P. |mWކ1hTS#lckxNZc&TaIBt2ͭ ?xthŖ)SWMw&]NKSJZcwZ8+1|&k階. ڞ#t92}4[?~q6;nV?(YBp ꑝ23{>pdX%>W8Yݩߩcb^N!OPO-lکeXn <+89OHgXsћ4+:R=L\ȞaXZp1 NxrUX? iG4ںp4%_ljIZc5ܧ\D5U/cr2V컂ն@Ckuk+;Ti$qFEIrs$H.MlO7H|$Ͱ[RxX€(3Ŋu`\b>?zN #c3^ZfдI2E^Oypi9 _؇r3gX0羨Au<dzv$@:n Iᗿ>60੅_x5\W2xH\k]Qڦ PFeyQ^yq&$ǿ]]ZRYYDËeTcØ]s}#1 -$>xy%E B%0S+sH)VZH2yazI,haz`8~|~=*刣iO$6B؊CI-vHRzn j ; `믗MQ} |4?uӜ=g<9~~q(R(n:ДH/ʲ.M! PKQg :h;> U,}s5%(|^"/3oa%'nQɫ*|0b`'fySНIlmJ;AТ̗̓ )k 1փ6-ª^f62UP2Ӯ7["V݂耢qua\VPa9Ň`Wͷ[n"K"0$Tm[AU!\RV*-.*ͺˡsOwBYe|SWЍη* Zo#XBZUwZMN7d~B꭫>'F^c^KSiߙU 49=u!NQ&.hm$ 13:T08 5tR,1է6/&ǥȊ0/A=7\,KdwPUE鉫>6}jb5\XW+Cnc ĔE3ýۋrep/h]_*4tyFl.jqEEA5OU]4zDb"}TaG&tV7ö;JFoO:dDuJQfɊcטY!fa{m-Uy P88G/$ЧObIk_t Ϗ 3 tX0ZDoUŻ΁x=;J#=E$7gOܻ3i|[˷*ڈƶL)+뎑]}ޑ8_M#NuN}LtL%dn; %Qr $M[J ߮ 9Tf3P%[7ɺhݣq<zgWEѶqٲFu7 GqcuѤ5uC13aML&8bIWĭeL:-1$hBݣ :geTֳ n"s<ߟ*<M#ѳ~"ЧdDX09klk_|z1 pCc9d7ģʄ u\,ݛR".N&T.UbqKqLtkk >?~̢Se-~.d[oō-&lݭY2G4w"7![sWhNJ^/BBڄ S8-7Zuɘj*czu$+Є,)5V&1ga7yGVPXIDs=VG}Zf6ZvOfk{L8V{lgpm j&~XnM "q+Xng jyOjcѬҳGPn wBU<%FY⸒ުbl$0|S]^{7=j`<:ߏ8^)<Ίx[mYli֝Ze7Fj?iдߴ$梾x~G"p (Z"b%uŬz#^1SGD )@J)Ra EubA m@62Y[8ϒmy&<XCLٝGɌku2W7U™+Yܤ?|ρ|Cļgܡ:\i!"o-ӿ?# @kM(cƔ޿v* 4 Qⱸڰrǚ&7epс1!7b~̂Ftnc!5AW)Jgh6>%iQC*0&J_G6֘Rz?k*QI˾g`vE<-Acpv[U,HުoJ  ?9VnZ9&CY!oӘ9Еr6xL!]6gBiv#ˋWu<`eлfyڶb}S"iE;/EJZPB($ݐN0ݳ˼USxzÇg'+Y#{%ܙŀ"m )xAX@%u;ƨBFl_ ~2rwrn4HZ A *)>^*HK zWXrH(;^)m@0(Z| k<ſt72e@78 m2& Sv;Nz^bZz gJXvD,SC: Zfb;i2h`(pmehOM!\n2գnvI&BhyXnz$1mSdhqu_ųx8412-up Qرghm.pEbF7Ul1<ozÁ)RC=\G*W?g3)-VHhdvGhYJ4/B%We*JbN$*k&V㏛Rغ 0k3)\K$FY|@_/JXaMRp Bϫm /,X*LbHGee(o{nIDHҗ@ƩY(#(-ԼGx_S.| ACÂfAb Ȟ;R&D;D] cVlw `-F?Thv=7.fO )LBΊ z C7lB/EA.7Ȉ%v+ R (ys.~s.l3fto/[_@A]'R+~>O-riVSnrSMKIF]%\2 1ѨZV>YtC\ͅlKmu+WnS͹ʒˑM$rFȷ}qĭUῸVMp~TE .ipr~d\ N_sx SO.(*L#pJrBWP > a{S'lKtg> e0`K<㵊 ~g[SLiJ  Sz6"O/sH8O.r}# _w%sMY& لS9%"gEqyU{Be΋Wx (}MJS0go*Ur%\ʇotY!f#` ܄KtJ]5rsqC?{._6e<. @56=hnRh.Iӈ["(\ o$L"א{ĭQS"6: q Ѵ)±h1.  <@= L )(&*0pDpھ$0q11-umbS;X=%iN뺳b0AϾ6l= uplYw,'\ED[C FSi Ky-&;f!Lj :B[*.G}jY6鏠4 -߂RKOH N-hBtk|0&-c sؕS)22T~FNHܑ`蟎 i2ԼF/t9.|Ai"P`oW'#N_p; B݂}p*9cZmTpGz))Sr1P}eCtC}ȡ!K]w]2l0,qI29x}6R(4rCwxL暈m$=cQt*k=nvpU-+. ^bړc%-M)!\ HͥLY-DmT I}t*(kvjj93~+ttƒ<+"10OnVb⳦XVroԶͪu/t'B$#H~ |Kpvk$@Rn,]G^Ra]QSI-el2E} H=mYdj"U +72v.Pd)b3xN5}Пi2S.fwM{ JrW򙅺o{0weRd,.:8jCDtwSbFJqSUh3%KL+ٍT2f s \v > Z_ mcT%E_:`db(ehjsm yT%K_@`jb.ehjymo !%'+ikFGH W_gmqu  -/13579;=?ACEMGIKOQSUWY[]_acegmoqsuwy{}! ^1 e)3 "  r R & J 01> (.!  #!u u!*0.b * ,GYdnso ( > b g t s d V 1 1 1 1 1 2 (2 =2 6 6 6 6 6 6 6 6 6 7 7 67 M7 e7 y7 7 7 7 7 7 8 8 .8 E8 I8 M8 U8 X8 ^8 d8 k8 r8 x8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 9 9 29 9 : H> t(I %kT o"i In } 9S > @L "T [   z C  v L f y =  y  b   O  I  N + ' Y     0 ` R .    c  V  l       ' Q' a. % P3& 3. :1 65 @ zA c g 0h h Sj j k k k ql 0m m An n p ^p p <q q r r r is s /t t !u u v `v v Jw w -x x (y y z z z { | } n} } @~ ~    k  4   d  <   =   e <  ? c  y + (   K ]    x   < 8 Y "  e l '  q + H   5 T 5 ;  E  Z  H     @  ( % p @ ] y p  q  7 0 R ]  ?3  U  J  _  B        ' + / 3 9 P a t    x 6 w g   L " M%    /   R 8 b  9 g     9    0 N         , C Z q u z }                          " % , 2 8 ? U j   I  K {  ?   \   x  t  /# &$ %% %( t,+ S. (. (2 >/5 Z8 8 <; օ; :&M S LS T XT U (V V V ]W W +X X Y ~Y Y DZ Z [ z[ [ 2\ \ ] ] 3^ ^ X_ _ `` ` a a b c c Fe e Ef f f gg Mh h h ti j j ^k k l vl l zm m n rn o o Yp p fq q (r r r \s s s t t t u `u #v v v v Aw ^w x `x #y iy y 9z z { x{ { F| | Y} } } ~  2 N k     +    ~     # ' + 2 4 : > B H ] r    Q    z  B ; 2 c ~  p -  W    V         !^ '  9 V v E  x 6 Bw   % t.##/F4@$GE ;U~]X5c nr=3Ot=DLP=|[7S.n"VpP %,36?BIQTZ_fhlsuyp0^fpC*-39?BGKPUX[_cikqsw~-hxM/X"-,6U#aMg[t5=SL8Q6             1   Ne'J~,A<I$1 3333"3(3+3.35383=3A3E3L3Q3S3W3^3`3d3h3l3r33384456=@JAD,FC GI6M PR Y[H{ ~y^"ih `kh  "%+25;AHLPW^gjpvzf/,1$;'o :[yu"%  ]u:\~[F   V   z  9   k  ,YsxzY5_-D  #T>](rGe#(+-37;ACGMoR  !!!! """<#^##$$%v%%%&&'''(U)W)])c)g)n)s)u)y))))))*g**P+++(,|,,,--.d0g0m0s0v0|00000000001 1111 1&1.11171@1C1I1O1X1[1a1g1m1s1z111u468#;k>BEI N''QWW]$b e*kFs-;{yu3,t)#[:DcJ\MZy1N 3 =PfxOKh 7Yx0;t &(.28>EKaw5w8KG73/w0z5qH 1H_w.29;AEKu  k6q}NKkvIbO17 Hfxs)b54oR1]'[dp"yQ?\c$$$$$%% % %%%%%$%&%*%0%v'%$,:%04!7B<3?&I&@NVAW3[i&d(jnn&n(rvvR+w~d+~;b7n0]|4| '/z sy9+'w*; /1L]XO$9*WKqDNg$G'n [x,T||)0N*f*4d!|:"S"U"Y"]"a"g""k$R%d/9; )>%)D;ISUTVZ/ZNZZ\]^^L__h``ta}bbb-eKejeeVfgh8hhhijjkkOlnnnnYoo&qCqqqdr(s\tyttu*u*N 9SBMi>TZW;?CIhruI R L & 82m NLii[$L g %J  =b  J2TU@j!_!Lpqa#R\{P\[j%          v-/9j#q"(>jC  $!x!!""%"+"-"1"7"M""##y##s$$%0%%&\&&V'''(((?))9***-+0+6+8+<+B+X++.,,-~--%.r.u.{.}....)/r//E001i1m1t1|1111111111112 2 222 2'2.2\12$;r;;<%?EhE&HMbNOSTJUUZVVWWXX$YYYYZZ{[#\cccccccccdmee>f}ffRggg hkkjl!mmnpkqrNsss#tett2uMuuvvvvw wbwwwwx"xdxxxxxyyyzzz%{A J/28:>DFLPV}c{} ?]Ca;{\tv|2 ! >]+Jo ;GM k          # % ) - 1 7 M   b    B&tPP%u xG 9!F7c- Z!"v""?##:$$$F%%%h&&='e''''U(( )U))!**, ,,,,!,(,,,4,6,=,?,C,I,O,W,Y,],c,g,m,r-x-|---------------7.U./>/w//0s001t11o23Z334@444-5K55)6l667l7088999::;F;;6<<1==>>@?Z??@^@@A}AA?BB CxC9DD[EEEF GGGEHHIfII'JJ[KLsLL`MM[NNOmOOOPQhQQRRXSSTTTQUUVbVVWsWW)XXX9YvYxY|YYT[$]^ _____`f```8ajaaa*bbb%ckccc.dndddfeee8f~ff gLggg hLhhh=iiii5jvjj+kqkkk:lzlll:mmmnZnnnoyoo>ppqwqq0rrrgssstutytttttttOu \.2 L4. +}p{0S{cSr+RoX}_~Uu_Oy  u  M o-J R`=CMvLAT}'{:   r  D {     S   " t  L#>Zu;d?7!T!!!!"7".../4/Q/n////Z00161S1n112y22222 3(3D333I4f4444E55O666667'7777n888|99l::::;;?!?A???o@AmAAAABBBBBC$C?C_CCCvD$ZDZbZZ[h[$\__$``hJaoAoaooo)pwwxx8xTxpxxyxyyyyy:zzz{6{T{q{{{{|y|||}"}@}^}|}}}+~\~~~d{8d:#=]4U-t a7Ss1Mi(/Oz#DCc;[?Zz/k (Cw$jo0P%=">[w!=Zv i8 %:Vt'3:Vv K zfF >\zKk"Tq;Pp'r (Hv\|3f6pEYsGa~$?\wIiNjDa6&Da0{%VGd/Du=]E`|.:k3S9Sl' G    _      [ { 0 P  J,d,,,(-F-d------.2.N.j...../8/45 588N9g 9;:<<=>@NA9BEF_FuGI"KKL1LLLM2MSMMO&PFPPPsQQRR>R R?UUUHVhVW>WWWWYXcYYZXZZZ[[;[[[\\\,]^^^&__`6`S`p`````:aiaaaa,bHbdbbbc,cIcicccddde$e{eeRfffffgfgg=hphhhh1iQij'jjjjPkkl9lllm)m]m}mmm nnnWolooovvvwjwwwx(xIxxxHyhyy.zzzz{L{l{{{{m|||}2}}}2~-M-J{:FdD @'Dx'4TnFf9Y =Db~lT;Wt f(姬#zPp<Xx.NOl3Qo"@^{RM?Y~N \q3d Y)C9]._{%V  l . Xro$S !m"#C$%%'`(.):*l+,-../m///W0w001!11 2K5555 6,6c666 7-77a8~8889&9======j >#A<A\AAABBCWCC$D_DE EmE'FTFFFG:GGdHHHI"IIJJcJJJJK'L@LYLyLLLM%N>NWNpNNNOOO0PPPPPxQQQQQRqR%MS^o____@```aaBb[bbb,cEcccd/dddee9eeeffBj\jjjkk=kkkulllm:mmm:9n{|4|T||||}}} ~~~EZ{ z Fa@Uv/Ozqr7Lmn 1Q7q7;5 uy^  !Bge A7!Ij)yaJdF[|f.Po x5z/ osX:{_]9/AbiQ:Tt6KlVp@_Xm3S~Mt v xb ,  :Tn &@t! ZvK'Aa1Mj+Fc *Fb8 #C-M'&F-Gg #C"xX$!Uuy> S t  G$e$$$$/%D%d%&:&&&t''' (-+K+h+++,+,K,-!---[.p.."/3 4'4G444t55 66?66B8`8|88899K;;%<<<t^=OPQRRS"S|SSSgTTTU2UUdmmn(nnnnnoooYpnppp$/r~~ d:Zu t{(7)I9i57Sox p H 1 F3    lR $ HC$014( .  # !uG}  s!*..Tb(, %`''*v,AY^Q_F`pb4ehjmbn o sg   @Ba i e 4uky$:ap : _OGmfS9  J  S   cA h v j =r z { +  J    k         .    w 5  M 0  k     / F =  l  $ ; 5S 9T CU ^ _ Mg h j l r s Zt :u }v w :x = y } ~       9   4  h 3E4hj! :0 TZ!"I#%&')+1Q( g>XfU ~X[UC" k =%jlX1&Fn=yN "YZ^~_`Qbeeghikzn;op>r0t8&y/VV&  vAd:[l5/< d V 5D""'%% ((O++.6/; 2ThWYczeKfrsGt uuvvowwqxxyyz,8cP)Z D    %'-)-D. /3:5 69 :>t ?E}G HS__L`Daxbc>efhij lmmt=`8clb   9$o  ^  +t.!! "./11Y2t3144u5w6777789: ;>1??@@@@AA AOCCZZI[_w`oQoowx^yzz{|}T~~M/Dqc  ?3SK9jWrV^f+@|(y f[KK``8l]& YF2"qiB>?kM;aC/ 7     k @ +t,-.444444444!LL-MBMO6PPQR-R7UUXV.WWWXYZZ[+[[\\|^_ ``caa$btbcYccdee|ffvgghhAijjjkkkXmmmmmngo|oXvvwzww#x8xxXy zzG{\{{{}||!}}(=rg40B0 sDAVI)41`h>H| CjE:Ql*P#zQB K  +5?IS]gq{./}//g0012z5566s677Z889<-==@LAADPENFF*GHHIJsJJKiLLMNNO@PPQQaR^_P`^araaaaaa)eeijjk-kkll)mm{/|D|||}}~Uj@qPe ?^G\~,A'&Ius,/tVkYCmk$'dF[9Cr oTQSZahqzr3=6W3swPeN c >$$$?%T%*&&''$+++&,;,--k..374456.69889<;<<sO{OORSSSwT U!UWmnnnnnoip~p_mn~~tJs|9a%+++                           +                                                  + +  +                      '     '                              '             '     '                                      +                '  '                                     ))***               *******)))*'))* )))'))'''''))'  '  '  ******************'   '  '''')))))))' ' ' '            ''              ''   ******  '' ''   **)**'' n x w h G1 ,2 E3 9 "9 9 = BI S =i n U}  ޕ  P N  9  + > < e ?  r 4  G k ? " v C  y    R% % - }1 |5 @ zb Mg g j j 9k k l l Im m Zn n p p Uq r s s s Ht t Pu u !v v w cw w Fx x Ay y 5z z | } } } [~ ~ -   O ā !  ݂ W à 8 V 6 ~ $ ш X | 9 D ލ A A d  v 5 ˒ . U ; % ! x ڝ 0 a 0  N m N T ^  s c Ȧ 8 Y A , ҫ  E b ~ ) P 5 k ̳ m n ( c x [ ; * c - o ! u H g n w ) ?   g C Y     W  j / " # ( /+ . . )2 4 7 s8 : L R S qT AV V W vW W DX X "Y Y Z ]Z Z 8[ [ [ K\ \ -] ] ^ _ y` ` b b )c d _e e ^f f g fh h i j l 2n n &o p rp p q q Ar r cs s s t t yu *v v v Fw ww yx y y Rz z ${ { { _| | r} ~ ~  7 S p    D  1 L a v . ҅ i ' G 4 Z  S J g : F  p ݑ 3  Ԕ [  4 Q %#/_4S@DTA]cUnsq&ɦqiVtP,lG ;|*ojHv(~\,E7eGp:EDN)yOUWPk "   ]  f'A~@c3F YB$013346V=@DEGILPWRX9[z}ӂ,#*cy+)ڝ\`8|EjSbT\'-ut_ 0#  o  !   R     Er48N- x]i3uS}1j !%"#w##)$$6%&&,''()))+***h++@,,,---- .84T68;L>B'EDIMPMWWs]awekrz;7S*|c ufʨ0L97ϹhIyOeMQd`P2H*IN-mi^V0a0)  gNMkJv@v;Du#+04|7<Q?IMZV\[!d jmmru} ##D; [@H^` t8`@ `fG2|&!8$%b.8:<CGRSUYZ4Z[g]^__abbd2ePeff hQhijjkl2nnnno+q\q}rattu~^t=kZӇQml̊IK –ԗdޘn`b+H֦;̪ЫQKL l#;a^!(K7oF"O7rUuz;ja!>  (      FHRp;5V, hj=$6`$R:|  s!-"""X##S$$_%%&&&D'l''(n((n)):**s,-m.V///7000711123r34X444c5A666.77H88799:;^;;N<<I==7>> ?B??@w@@5AAAXBBCRDEtEFGGG^HHIII@JJtKLLLyMNtNN/OOO8RR+T?YYl[<]^_w__`~`` aaab0bbb=ccdFdddd~ee fPff$gdggg$hRhhiUii jMj|jjCkkkRlllm@mm,n`n o6oopqIr ssTuY^dVHa.@&> W 3j"r<]:v|*l0J42ck$Uf!eYl5?R  5      2 k   <   ) 'B^yS! ?Y??@AIAoAAAAA*BFBqBBBBB C)CwCCY*ZJZZ[___`n)oyoorwwwx xt48%Ca *Yv/w`~;kN2L5[z#If (D` %RnHeN+I8 nLiKyuIdFAok=Wp3 ~  f      h +$,L,,,--K-i------.7.S.o..../J4~55>8n888+9P9;#<<=}>@7A"BEEHF^GHJ KKKILLkMOP^PPQVRUU VVuWX8YYYBZZr[[\M^^^>__`:`W`t````Aamaab0bLbbbc1cccdddHfHH:II%JLJJKKL)LBLLMMM'N@NYNNOOPhPP~QQRR^___`x`aabDbbbbbbc.cLc}ccccd6dgddddeQeeijDjjtkklRmmz{{{l||}D~~" Jg݅;/4uinOÖnɘ04j̜I((^#*6ɣ'\cPDک#L=h/ٮ5Wܯh*.dƷA  V".T[ H}4 <-X%GpkT#Ny%P{8D$>Xr*95j4^)y5Rn/Kg.J [e?^/ [:D<W ; $L$j$$|%q&&'+2+P++c,X--.334_445W68G8e88+9 ;P;=<<DO{PPZQQcRRR:SSTJU,mmm@nn&oppqrK~~~%K^dÇ!qaV@m׏M ;WEzے܉  % ( 0 8 ] b p o ` Q 1 1 1 1 1 2 #2 92 6 6 6 6 6 6 6 6 6 7 7 27 I7 a7 u7 7 7 7 7 7 7 8 *8 A8 E8 I8 P8 U8 [8 a8 g8 n8 u8 {8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 9 .9 9 9 > oI T ii n } 3 :  t = p F ` s 7 s ]  I D H % " T  + Z L ' ^ P  f         K [ % =& J& . 1 5 @ mA tA c {g h +h h h Nj j k k k kl m *m m m ;n n o Xp p 6q q r tr r r cs s )t t u u v Zv v Dw w 'x x "y y z yz z z { | | | h} } :~ ~    f Ā .  ^ 6  Ƅ / 6 ܅  _ ݆ 7   ~ , 9 ]  s % " F W   s    | ٕ ) 6 ז 3 ڗ T  _ њ g ɛ ! r   k Н  & B  . M á " / Т ݢ 5 ? T ѥ B  | է : " ԩ Ȫ  ǫ j ʬ Ѭ ެ 4 ; X t Į ˮ j k * 1 $ + L W 9 O н D  ƾ  Y  <      # ) - 1 5 L _ r s $ 1 r b H H  ) M 2 B I P ] y 4 b     + I         ( @ W m q u z                           " ) / 5 ; Q f   D   G v  :    X    n r  n )# !$ $ % ( n+ M. . . 2 85 T8 8 8 7; ; ; 4M R 9S FS T ET RT U V "V V V WW W %X X Y xY Y >Z Z [ u[ [ ,\ \ ] ] .^ ^ ^ R_ _ _ M` Z` ` a a b c c @e e @f f f Tg ag Gh h h `i mi j j j j Yk k l pl l l hm um m n ln o o Sp p aq q "r {r r r Ws s s s s t mt tt t t u Fu Mu Zu v v v v v v (w /w ) 6 - L R _ y ڎ U j  ' p ɐ ِ R  y P Γ ޓ ˕      б ߻ Q q A t < }  n#/4@EUx]Rcnr7ߙ.}6FξJƿ7vU1 {M)h P\i?L "(/3:?FLQW]afjnsw{k+SY `j= z'*06<?DIMRUX]aeimquy"(dtH*S~)~'~$1P\<IUbWo19@OH2<K1rt              ,    I  {_!Dx(/<7E$1 33333%3(3+31353;3?3C3G3L3Q3U3Y3^3b3f3j3n33334456=@DAD&F=GI0MPRY[A{~ tڃXdbÔ֕Zeb "(.28>DHLSZagmsyu j} 4VtoXcp5UyU@ {  P   t  3   e  &SnstT/Y(> O9X#mB` %(+059=AEIjM  !z!!""""7#Y## $$%q%%%&&& '''(Q)U)Y)_)c)i)n)s)w){)))))*b**K+++$,w,,,---a0d0j0p0s0y000000000011 1111#1)1.141:1@1F1L1R1X1^1d1j1p1v1}11p4668;g>BE}IN!QWW]bek s({sonΛU5?]DUG֪Us,H-Ӳ8KǷarIFb2Tr+5ho "&*04:@G]s0fsq2DA1!.|*  r*t/lC-D[r *.49?CGqf1myJ1Ff rD^J+ Carm#V]y //iL+W!U_isK~3:V_$$$$$$%% %%%%%%$%(%,%q% ,5047<<?I NV;W[cdxjnnnr{vvLw~~~^~ \hWv/pwAŸ!ơ*tĢMZgt÷#3"q5)FXRI3QFk>IaA!hOVr|&GNovx%+IHUb/Z` w5"O"S"W"["_"c"""f$%%r//99;;=>DDHHSSTTAVNVYY Z*ZHZZZZ\]]^^^^F____b````janadbkbxbbbee(eFedeeeePfg gh hh2hhhhhiiijjjkkkkIlnnnnnn?oFoSoo qq!q=qqqJrQr^r"sCtJtWtsttuuD %Iāс(5O#0=Çɇ/<Ie -:}߉6CPUƋӋ܌S7;AEcmpEݖܜ̝LF 2,gHFֻcbUzE`C7\{D ,NQKXe Zp}HLR_lMS`mgmzx~]N=JWu0=J=JWUd       m t      c j w   p( *&3frm $9e?  !t!!"""'"+"/"3"H"""#t##n$$%+%%%&W&&Q'''((((:))4****+-+2+6+:+>+S+++*,,,y--!.o.r.w.{..../%/n//@001e1i1p1w1|1111111111112 2222#2*2W2; ;n;; <?EEcEHM]NOSTDUUTVVWWWXXY~YYYwZZZh[u[\\cccccccccdhee:fyffNggghkkelmmnoWqfqvrr:sIsssstSt`tt!u.uHuuv v}vvvww]wwwwxx_xxxxxyyyzzz!{;9F,/48<@DJNRy^wx:CXg|>\6wXptx~. 9X&Ek6BH f           # ' + / 3 H   ]    =!oKK prA32Ap}2]( T! "`"p"""(#8###4$$$@%%%b&&8'`''''O(()O))**,, ,,,,$,(,/,4,8,=,A,E,K,Q,W,[,_,e,i,n-t-x-~--------------2.P./:/r//0n001o11j223V334;4~44(5F55$6g667g78+88999{:::A;;$<1<<<,==>>>6?Z^ZZZ [U[d[\_____ `````ca o=o]ooopwwwwx4xPxlxxxyfysyyyyy)z6zzzz{1{O{l{{{{ ||t||||}};}Y}w}}}}&~V~~~`v4`߂5LJ9YЈ0Q(o]}Ԑޑ2OoŒ -IeԔ$+Kv @Ϙ?_ϙ7W ;Vv+gӜ $?s ՞fǣAk,L!˦9ȧ:WsȨ9Vrǩ9e4 }!5Qo#/5RrFsa9WuGgOl4Ll"fm$DqXx.a~2lAUoC]z ;XsEeJf:@\.2!?]}+ov!NRB_}(@q9YA\x *6g/O5Oh# C    [      W w , L   ,F,`,,,#-A-_-}-----.-.I.e......3/455R8889I9b9;5<<=>@IA4BEFZFpGIJKK L-LLLM.MOMMO"PBPPPoQQRR:RR9UUUDVdVW:WWWWTX^YYYTZZZZ[7[[[q\\\\']^^^^_"__`2`N`k`````6aeaaaa(bDb`bbb c'cEcecccddde eweeMffffff gbgg8hkhhhhh-iMij#jjjjjKkkkl5lVllll m%mYmymmmnnnSohooodvvvvwdw{wwx$xExxxDydyy*zKzzzzz{H{h{{{{i|~|| }.}}}.~̂)IÃv}ȇχ (Fsw4:A_}NJˊˋҋ<@Nō<"@t#ِ0PőڑjBbƓޓ5Uʔؔ}ƕ598?]zh7Soߪb$vLl7Tt*JJh.Lj;YwELIv|9MTyFJ\Vm+/\` U%?f/5Y{*U[w!NR     W [ h *  ?FSnV]jh@OFpe   R!|!!0"Z"i"r###$0$?$$$%]%%%&&'!(M(\(()*))'*6*/+Y+h+Z,,,]---..../g/~//S0s000112G5{5556(6_6667)77\8z88899<=9===f>@A8AXAAABBBBOCCD[DDDEE\EiE$FPFFFG6GG$H`HzHHHHII JFJ_JJJJKK#LjXjjjkk9kkk`lqlllm6mmm6n{{{0|P||||}}} ~}~~0AVwvB]}ӂȃ<Qr+Kvjjًn"3Hi؍jޑ-Mܒ3ȓ\m3ĕ5zԖ.mqDVȞ :{٠ס`ˢ^9̤0ХΦBbç"qʨר] C`תͫBWx#_|ɮ'Iket/tα(gk>P۸2sѻ޻zϼܼXyý V~1|Ŀ(:ZaM3Pp2GhOl 9[ET/OzI 6ap 8crr^~(6Pj"<p |VrG#=]-If 'B_{ &B^~3?)I#"B)Cc?t{ }{Qqu) : O p  B$`$}$$$(%@%`%&6&&&_'p'''((+F+d+++,',G,,---F.W.l...34#4C444p5566;66=8[8x88899>;E;;<!<<<<o=OPPlQQwRRRRSuSSScTTT U.UU_mmmn$n{nnnn oooUpjpppq rz~~~ `6Vˀpoȃv#Շ2Ό%EƎ5e13Ok+nt{Ó˓ӓۓ #+4<EMV^fnv~ǔϔהߔ (19BJR[cks{ŕ͕֕ޕ&.6>FNV^fnvȖЖؖ!)19AIQYajrz×˗ӗۗ $,5=EMU]emu}ǘϘט "*2:BJRZbjrzę̙ՙޙ&.6>FNV^fnv~ȚКؚ $,4<DLT\dlt}țЛ؛ (08@IQZbks|ƜΜ֜ޜ&/7?GOW`hpyʝҝڝ &.6>FNV^fnvžʞҞڞ %-5>GOW_goxȟП؟ (08@HPX`hpx ʠҠڠ "*2:BJRZbjrzɡѡڡ'/8@IRZclt|Ģ̢Ԣܢ )2;CLU^foxɣѣ٣ '08AIQYaiqyɤѤڤ "*2:BJRZbjrz¥ʥҥڥ $,4<DLT\dlt|Ħ̦Ԧܦ $,4<DLT\dlt|§˧ӧܧ#+4=EMU]emu}Ũͨըݨ %-5=EMU]emu}ũͩթݩ %-5=EMU]emu}Ūͪժݪ %-5=EMU]emv~ʫҫ۫ '/8@IQZbks|ȬЬج (19AIQYaiqyɭѭ٭ "+3;CKS[cks{îˮӮۮ #+3;CKS[dmu}ǯϯׯ߯'/7?GOW_gowǰϰװ߰'/7?GOW_gow±ʱұ۱&.6>GOXaiqy²ʲҲڲ #+3;CKS\dlt}Ƴγֳ޳&.7?GOW_gowɴѴٴ !)19AIQYbks|ǵϵ׵ߵ'/7?GOW_gowȶѶٶ !)19AIQYaiqy·ʷҷڷ "*2;CKT]emu}Ÿ͸ָ޸"+3<EMV_hqzŹιֹ߹ (08@IQYaiqyɺѺٺ !)19AIR[cks{û˻ӻۻ #+3;CKS[cks{ü˼Ӽۼ %.7@IR[dmv~ý̽Խܽ )2:BJS\dmv¾˾Ծܾ $,4<DLT\dlt|Ŀ̿Կܿ&.6>FNV^fow!*2;CLT]env '/7?GOW_gow #+3;CKS[cks{ #+3;CKS[cks{ #+3;CKS[cks{ #+3;CLT]emu}'/7?GOW_gow'/7?GOW_gow'/7?GOWbmu} #+3;CKS[cks~'/7>EMU\cjrz#+3;CKS[cjqx  '.6>FNU\dkrz #+3;CKS[cks{ !)19AIQYaiqy $,4<DLT\dlt|&.6>GOW_gow %-5=EMU]emu|#*18@HPX`hqz !(/6=DLT\dlt| $-4;BJRZbipw &-5<CJR[dkry&.6>" - 5 Z i t s d B1 1 1 2 !2 (2 A3 6 6 6 6 6 7 7 /7 F7 7 7 7 7 7 7 '8 >8 M8 X8 ^8 d8 k8 r8 x8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 9 9 9 = *I S %i n <} ŕ ײ 7 5  % # L & Y  . R & w { ] *  `     9% % ?& - d1 c5 o@ rA ab 4g g h h i j k k k ql m 0m m m An n p p  4 =  e . ? c y + ō ( ( K ]   + < ٖ " g ך l ϛ t  s ՝ $ + H   5 T $ 5 Ң ; ˣ E  Z ٥ J  ݧ @ ͨ ( ک Ϊ ' ͫ Ϭ Ӭ 9 @ ] y ɮ Ѯ p  / 7 ) 0 R T U ֽ  J ̾ _ B   ' + / 3 p # &  K  W  \ / y G N R b  r U ^   & f  N     % S j }                     % , 2 8 ? U  w    ?    p Q  " # % ( + - i. . 2 4 7 Z8 8 : ; L R ;S S GT XT V (V V V ]W W +X X Y ~Y Y DZ Z [ z[ [ 2\ \ ] ] ^ ^ _ _ O` `` ` a b c c Fe e Ef f Vg gg Mh h bi ti j j j l l jm n rn o o Yp p fq q (r r r ^s s s s s rt vt t t Ku Ou `u %v v v v v -w 1w Aw ^w Kx Ox `x ay ky y 9z z { x{ { F| | H} Y} } } ~   2 N k    # + ܁  ~    ' + 4 : > B H ] r  Q   /  z Չ B + ; " 2 P T c ~ " p - W đ  l  Д      C s  q  8  #/F4:@DT(]bWgtK=DS8@Swy          E   N(e'J~B1A=$01 333"3.383=3A3E3S3W3`3d3h3l3334i6==@kDEFzIL|P>RX [z} iJ`h֝ %+5;APW^jpvDҲHmc,Q:yIx;Ce\[F  V   z  9   k  ,Ysz5_D Q];e-#-37;CGoR ! """^##$$%%&&''(W)])g)u)y)))))*g**P++(,|,,,--.g0m0v0|00000000 11 1&11171C1I1O1[1a1g1m1s1z1 4<668:4>AE+IMP4WWZ]a^egkrz":{:cJ\Mܪyܰ3 kO0x0mt (.28>EKa5huw8KG7#/0z5 @Wn2;AEUQF>In  f[b54oR1]']dp"8?\#$$$$% %%%%%&%*%+/4d7;8?IMAVC[dimmTrug}~~  +|6u|E '̡/¢EO\i~G[x3qDgG'nT[xLt)NCJW^d|!M"U"Y"]"a"" $%%J.t/89:;<=BDGHfRSSTUCVYYZZ/ZZZ[N]]^^^^___``alaibmb}bbdee-eKeeeff gh hh8hhhhiiQjjkkkk-nnnnnnDoHooqq&qCqOrSrdrHtLt\tytuu~H Ɓ\*9S%2BLJχ1>Mi"/8ETȋ؋ތ?C13LƘiVH~h{JҢm/m"٨823S޻k "HiE2 VBMZz6r3PTap=QUbqko||v#R?L\{2??Lj%      r v     h l y  v-/(9lt#q">j $!x!!%"-"1"M""##y##s$$0%%&\&&V''(((?))9***0+8+<+X++.,,-~--'.u.}.../r//E001m1t1111111112 222 2'2:;;1?D EyH@MNNwSSTJUUZVVWWXX&YYYyZZj[[\bccccccsd.eef9gtgghjgkllmynoq9rrsssUtet#u2uMuvvvw wwwx"xxxFyyzz%{ŻA;2:>FLP+O|_a;{ v||8>jSeXgc>bjGk       % ) - 1 M    B&tPR% xG 9!rc Z!"x""?###:$$F%%%h&&?'g'''U((U))!**,,,,!,,,6,?,C,I,O,Y,],c,g,o,r-|----------U.>/w//0s001t11o223Z34@444K5)6l667l7 808899:;F;;&<6<<<1==>>>:?@???@^@@A}AA?BBxC(D9DDD[EEEGGGG>HEHHIfII'JJJK[KKLsLL`MM[NNOmOOORRT;YxY|YY6[C[T[]]$]^^^____`f```jaaa,bbb%ckcc.dndddfeee8f~f gLggg hNhhh=iii5jxjj+kqkk:lzlll#kD!]coS1&-JJR <M}aAT}'}:    u {     S   $ v  #>Zu; !7!i!!"_...../4/Q/n////0 1161S1112i2y22222 3(333394I4f444455Q6666'77p888899n::::;;>?A??@A1AmAAAAABDB[BBBBB C$C_CCY Z$ZDZZZX[h[______````n$oaoonwwwwxx8xTxxxyhyxyyyy+z:zzzz{6{T{q{{{ ||{||||}"}@}^}|}}}-~q~~f}fż#]ԈɉU_YF7s1Miٔ!(OŖDӘc֙[?z2kל ($#-CäPDϦ@ϧ"A[zϨ!@Zv˩#;lª8,U:Vt'w:vr  >\zk%Tqp*kr Hy|6f6HYv Gd~ $B\i!Nj>Da06&D3t{PVGdGu]E`|.=kS 9Slf  b     { P +,J,,,(-F-d------.2.N.j..../E4|55:8V8889N9;<<=x>@2ABEECFYGHJJKK1LLSMO PFPPQ>RUUUhV]WW3YYY=ZZZ[[u\\H^^^^ _&__`6`S`p````=aiaaa,bHbbbc,ciccddd$eeUffffgg@hphhhQiFjjSkkkl{͇ԇ-u{?FdɊϊЋ؋>DRɍ@''sqfxĕ˕7==Db ;Wt !(t#p<xN#O3To%@^JRyRY~HN*`t-^d,\)Fj29~Y_{PV/     Y ] l   DHX[_oTj2HQ g >!T!"2"^#u##$$$I%_%&& ($((())+1+F,\,I-_-p..//0!12255,66-758a889<====@@#A\AABBWC$DbDDDE^EmE*FFF:GG(HdHH"IIJJJJJKK'L@LyLLMM%N>NWNNOOPPPP{QQQqR^____``aabBb^b{bbbb,cHceccccd2dOdddde9eeijBjj\kkdll:mmv{{{T||},~~4{ FvOŅ6n&mpQV̓`7ȕrSO. !B۠١ AԤHզIj)̨ިթ J|'fԮ.PįhxlMI&:ӻѼ9̿@Ab:tlV @HXSP!8w#:y , :Tn &!1f0Z'a1Pj+Ic*F CM'F-g C"@$Su#- t $G$e$$d%Y&&c''+-+K++K,@--J..33 4G445?68B8`889 ;C;K;<%<<<@OyPPXQnQ_R{RR"SST2U(mmm(nnoppqqG~~~~ yGYa nҌI>ʎ<i57S/rxÒǓϓדߓ'08AIRZbjrzÔ˔Ӕ۔ $-5>FNW_gowɕҕڕ "*2:BJRZbjr{Ė̖Ԗܖ %-5=EMU]fnvǗϗחߗ (19AIQYaiqyØ˘Әܘ&.6>FNV^fnv~șљڙ "*2:BJRZbjrzĚ̚Ԛܚ (08@HPX`hpyě̛ԛܛ $,4<EMV^goxœʜҜڜ "+3;CKS\dlu}ƝΝ֝ޝ"*2:BJRZbjr{ƞΞ֞ޞ!)1:CKS[ckt|ğ̟ԟܟ $,4<DLT\dlt|ƠΠ֠ޠ&.6>FNV^fnvš֡͡ߡ #+4<ENV_hpxȢТآ %.7?HQZbkt}ţͣգݣ #,4=EMU]emu}Ť֤ͤޤ&.6>FNV^fnv~ƥΥ֥ޥ (08@HPX`hpxȦЦئ (08@HPX`hpxǧϧا '09AIQYaiqyɨѨ٨ !)19AIQYaiqyɩѩ٩ !)19AIQYaiqyɪѪ٪ !)19AIQYairzƫΫ׫߫ #+4<EMV^goxĬ̬Ԭܬ $-5=EMU]emu}ŭͭխݭ'/7?GOW_gowǮϮ׮߮'/7?GOW`iqyï˯ӯۯ #+3;CKS[cks{ð˰Ӱ۰ #+3;CKS[cks{Ʊαױ߱ "*2:CKT]emu}Ʋβֲ޲'/7?GOX`hpy³ʳҳڳ "*3;CKS[cks{Ŵʹմݴ %-5=EMU^goxõ˵ӵ۵ #+3;CKS[cks{ĶͶնݶ %-5=EMU]emu}Ʒηַ޷&.7?GPYaiqyɸҸڸ '/8AIR[dmv~ʹҹ۹ $,4<EMU]emu}źͺպݺ %-5=ENW_gowǻϻ׻߻'/7?GOW_gowǼϼ׼߼!*3<ENW`irzȽнؽ %.6>FOX`ir{Ǿоؾ (08@HPX`hpxȿпؿ "*2:BJRZbks| &.7?HPYajr{ #+3;CKS[cks{ '/7?GOW_gow'/7?GOW_gow'/7?GOW_gow'/7?HPYaiqy #+3;CKS[cks{ #+3;CKS[cks{ #+3;CKS^iqy'/7?GOW_goz #+3:AIQX_fnv~ '/7?GOW_fmt| #*2:BJQX`gnv~'/7?GOW_gow %-5=EMU]emu} (08@HPX`hpx "*2:CKS[cks{ !)19AIQYaiqx &-4<DLT\dmv~$+29@HPX`hpx )07>FNV^els{ ")18?FNW`gnu} "*2:B42 9 *9 N9 9 = XI T Si n i}    f s d - ` 6 P c  c A  / , 0 E ( N D   | P z @ V  k   o   5 E {% & G& - 1 5 @ @ b bg g (h {h h 0j qj j jk k Al l 'm m m n n o Fp p q }q q &r ~r r Ms s t t u u u Jv v .w w x x y my z iz z z e{ { | | J} } ~ ~ ~ [  G  } H  Å K % ~ ߇ b 3 6 > c  m n ߏ 6 ?  i M ^ Õ 3 + z ʗ 5  O W  T | М U ȝ  : ٟ  ̠ / , ڢ % ֤ D l ŧ *  >  R ݫ ۬  P m R S ܰ  б ? ٲ 8 r 0 ͼ = ƽ  : u H + ` . f O  > B @   = "  Z q , A g  j   8 x L b   0   ? g  0    P    X  X # $ $ % ( X+ 7. . . d2 5 58 8 8 '; Q; M R 'S CS S "T OT T ,U V kV V 9W W X nX X PY Y (Z Z Z _[ [ \ \ \ c] ^ H^ 9_ o_ _ !` W` ` ba a b b c e e 0f f f *g ^g )h h h 6i ji i j *j @k k k Zl l l |df IJr/47YLV).emm.+?G(F_9c(|QL*H|)xJLs   "   v  qiG j,]9(z33"4@4 56=@'ADF+GIMPRXg[{}Y3KPĕ@SPmuOm${8ceq^=e9miX|Z^7M0~l_  ,   Q   |  I  o"^aR2{i85+(~Ge1OU3  !c!!!""#D###$%_%%%&&&''o())))Q**)+++,f,,, -h--1]4668;]>BEkIMQxWW]bekr{a]y\C#L#yަ:,lΩΪ0b9u sݵ0Է>]0λ>Sy:aPXnp}6 &+u_rP_Qy(dS@"\}ph1V)>WY5O9 'PWU >q'w_1 #t= CLc0fNB8%+#047-<y?INVW[SdRjmVnnrKvv7w}V~`ނJMvEeR՝ϞtݟР\ѢWdq.i!4609AAS3.G,vVf!j^Ai`8 \qR_'I c|"M$%.|/89:;=>9CDHHRSTT_UKVY Z"Z@ZrZZ[]]^^*_p__>```Naaubbbd%e>e\eee5ff6gh*h{hhhiiijjkk-lWnnnnn-oPoo+pq5qqq0r[rrasTtktt v",t΁%2G-:چ9F]yȈ*7d̉ @MFËЋp:QՐ\^2QӖw͜t=աأr/e թ!а$s],456¾`̿3Xu!<8[W:GI,Ub eRGz'l2\i~  ]j.w/6 Ds GTm:Gv GT;\        ~     $ t   Pr'r#05| x\0T(   c!!?""""c##K$$%"%%%%F&&.'''(f((()))*x**J++++n,,V--..."/1//001T162:;f;;;s?DE[EHMLNNSS.UU*VVV@WWWnXXhYYYMZZZ@[r[[\cdYeue!fFffGggghjkTl mmnoFqaqerr+sDsessst:t]ttt+u@uuu v`vvvvwKwjwwwwxMxlxxxxyyytzzz{pݼ!C^Tfg~+0PUt/K$gI > 4Z| %1e- U  ? z  D }   *pU0z>aZlx<Gzy>o :!!R"k"""#3###$i$$+%o%%A&&'L't'''2(((()))d*-!.?../h// 0]00 1^11Y222E33314m4455556V667V77(8r8 999b:::0;;;.<<<={= >>>$?N?o??B@@@`AA!BBBOCC0DDD=EEEFF+GGGHHHIII JkJJRKKKKLL,MM;NNNVOOTPPHQQRR1SSStTT8UUUIVVVZWWXkXX!YIZ>[K[\]]]^^(____```a@aaabFbb cRcccdYddd eeefefff7gwggg7hhhh$iiiii j`jjkXkkklelll%mVmmm nAn]oopppUqqrrrr8sstyuuԃ׈z6wW ͳ/pξ WI8fMf4[ 3a8`Gg5]JFbbS,W<4J}*{5j~-)>^e y  \ z & Y   A {      a   92Mi&{Go!G!y!!!""{.... /'/D/a/~////@0001)1F1c111162q222222383^3334A4Y4v4444,5_55-6^666667i777O88889G::::;;>?,???Q@@#A^AzAAAA B5BQB|BBBBBC4CJCCCoDYZ6ZVZ|ZZZ+[a[\O____X``:an5oLoooqpwwwwx,xHxdxxxy;ypyyyyyz3z}zzz {){G{e{{{{{|_||||}3}Q}o}}}~;~~~Ga -_=r΂$܇3Hm *?k hUlÐ*G^ԒA]y̔ #:].&N/F Oe>z˜7m؝.ߡ9c&; Mݧ2Okݨ1Nj۩1]z#Ыl`/Kg%-Ja~> bH1Om6V2d#F[ F3XGgDw![{(UKg9Ur4Pl4T-^zTq7Ul!O=%Xu:R(H"Up,H{>Gaz 2    B o     F f  ; y +:,X,o,,,9-W-u----- .'.C._.{.....)/Q455H8889?9Z9;-<<=>@AA,BEERFhGIJKKLLsLLM(M=MMOP1PPPVQ~QQR(RRUUU3VSVW)WWWW;X?YYYLZZZZ[&[[[3\\\\]b^^^ ___*`F`c`````%aJayaaa bbWui+ތ&~ +َ8nڏ&?ԑQג <Qؓ+DӔdܕ$זUsIKg~ګ ¬e3[/Mc9w/`w&Db3Qo)c2e+j55Eg%Km6w7V{Bu 7o =Y    - e   )Pfk@g}`8J>hz]   J!t!!("R"d"j####($:$$$$U%%%&&&(E(W(()%))*1*'+Q+c+R,|,,U---|..../V/x//:0b000 1r12@5u55566U6n6677~7A8r88899<=(===M>@A1AGAAAlBBBBGC{CDHDDDDE:EfEF5FFFG%GGGTHrHHHH III:JXJnJJJKKL5LNLdLLLMMN3NLNeN{NNNOO%P;PPP\QQQQPe]'Vl‚+K`DŽ#:]BY|ϋ]}0BWQy|ԑ'<˒ "/j|!D͖'ܘYڙ]ޚ1QsН2֟[b@pۢ Inܣ1oĤ'ʥŦ*Z[ԨE9Xoƪ1QfάSt$Ac:q>DZ!{ԳSԴWص+Kڷkȸ*κSۻZټ8hӽAfԾ)g"RK5)H_!AVCd{1Sy"Q '>a8\.Yk0[m aEm}0Jd~6j \t;j.5L&%A^z ;Ws:Vmz.8x 1;Ru. cY f^K`d7 I ^  %$X$v$$$%:%O%%%&&&2'm'''' +>+\++++!,6,, ---.T.f.{..3342444W5556)66 8S8q88889";\; <h<<P=KOPPdQQmRRRR SdSSSJTrTTUUU3mmmmnjnnnnnooDpdpyppqrR~~~~OoEڀWRhn͇+{4 Hx ͐,GczБ\\.2 9 $9 H9 9 = PI T Ki n a}   ^ f W S ) C V V   z  " l  ] # ? z " H <  t t H e 2 H  ]   Z   ' 7 m% % A& - 1 5 @ @ b Yg g "h fh h "j kj j \k k 3l l !m jm m n tn o 8p rp q oq q r xr r ?s s t t t yu u ; M R S =S S T IT T U V ]V V +W W W `X X BY Y Z zZ Z Q[ [ \ x\ \ U] ^ 5^ 0_ Z_ _ ` Q` ` Ta a {b b c e e "f zf f g Xg h h h i di i j j 7k rk k ,l xl l m lm m m Nn n o 5p p Cq q r ]r r r s hs s s s t xt t t u Qu u /v v v v v 3w Gw w x Qx x %y y z oz z C{ { | || | J} } } x~ ~  : V s     e w  O  O d y ] څ &  b Y  t ̊ t ۋ - n nj $ V h B Y ۏ S u ΐ ( Y ʒ ӓ { & M  Y I  P m A#/y4p@DTZ].c}nqmGhNžcC#mIa65{nJ`C=e ')|KC( X ` `(i+: @aY ] o5Fy}`BcJ0y2{Y    |  i  U[9\O{3st334:456v=@@DFGIMPtRXY[{}K%=B2EBdlF_m ]z\TA/_g[; N=P)?"p^Q     C   n  ;  aPSzI$mm[*/{%u>\$IO&   5!!!""#>###$$Q%%%&&&''a())))D**+++,Y,,,-[--1P4o668 ;T>BE]IMQjWW]aekr{SOkN5FpЦ,^"\3 gӱQe"dη0z/"Mu¿ K,[=zRhyjwup%GQd"VuCs\VM7Swtg_$P 5J|L(B, yq JIzGu+hiV#f/5> U{"SE2%+047 <k?IMVVu[EdDjm(nnr=vv w}(~D<?h7_?~kʟS|ˢlQ^k(;&(q"+ 3;M b9&mMXa0.V~R2IGLYz65i"D$%.v/89:;<=!CDGHRSSTBUEVYZZ7ZPZZ[]]^^_N__0`j``@avaobbbde5eSelee,ffg h!hmhhhiiijjkkl5nnnnn oJoo pq,qxqqrUrr?sNtbtt~m#Pȁ,>'4͆؇3@Tp$1@É:G9\ʋc-K̐SU)K ʖߗnxk4̡ʣd&Wǩ°eT&'(R%Jg.*MI1A%O\\$a#t;)VcuWdvq~)>mANd4ARAN^2]sw. q      x      n {  Bd!d*vkO*uGW  V!!9""""V##>$$$%g%%%9&&!''''J((())*k**D++++a,,I--.../+//000G102:;8;;;f?D E-EHM?NNSS UUVVV+WWW`XXZYYY0Z{ZZ&[l[[\cdPeoef@ff>gygghjkGllmno=q[q\rzrs>sPssss%tWttt%u7uuuvSvvvv wBwdwwww xDxfxxxxwyyygzzz{̻gԼ =XKYZu#GHk"E^<zw}5'Mo$W O  9 q  7 w   !gL'wm4WxOR ~m 6Z-t KAz /!!G"e""" #-#y## $c$$$i%%3&& 'F'n'''$(((")))V*-.9.. /_///P000Q11L2228333(4d44 5/556M666I77"8e8899U:::';y;;(<l<<=n==>>?E?\??5@@@RAABwBB!CC*DrDD/EEEzFFGGGGzHH;III]JJLKKK=LLMM0NNNKOOJPPQ|QQ|R&SlSSiTT UeUUVvVV,WWW=XXX!Z8[E[[]]]^^____ ```a:aaa b2bbcIcccdPddddeef\fff.gnggg.hThhi`iiijWj~j kOkkkk\lllmBmmmn8nSoo pRpp'qqqhrr.s{stauui΃nшqq:r! b|I;}2X?X&U-X*Z>a'Ww<@JvTEI3~!<o!r,dx u1QX l  O t  F   ; u      W   3)D`{ m9i!>!p!!!""f....//;/X/u////10n00 1 1=1Z1s111!2k2222223/3I3333;4P4m44445J556X66666 7Y777?8x88897:v::::;> ?&?s??A@@AUAqAAAAB,BHBsBBBBBC+CDCCC_DYZ-ZMZgZZZ[[[[ ____)``*an,oFoooaptwwwwx#x?x[xuxxx&yjyyyyyy-zmzzz{ {>{\{y{{{{|O|||| }*}H}f}}}}5~y~~7mR$=.lȂӇ *B]$9e_Lfݐ!>XΒr8TpÔ4N(Hܙ &@F_8qœݜ.g|ɝ (֡0Z 5٥ uGԧ )FbԨ (Ea}ҩ (Tq]Z &B^| |$A[uu 5R8(Fd&P,[x@U3z-zH7a>nLuOB^}0Li+Gc%N'UqKh .Lf<*Ol4L|BLg#Br8>Xq ,    3 i     7 `  5 p +1,O,i,,,0-N-l-----..:.V.r..... /K455?8y88869Q9;$<<=~>@8A#BEEIF_GHJ KKKLcLLL"M7MMOP+PxPPFQxQQ R"RRUUU#VMVV#WWWW+X9YYYCZ|ZZZ [ [[[\y\\\\O^^^^__!`=`Z`w````aDapaaab3bObibbbc4cNcccidddd eVee$f]ffffffAgkghHhxhhhh i6ii jwjjjj"k[kkl$lClllllmMmbmmmmnn2o\oqoo6vvvvvDwowwwx-xxxyMyyy8zszzzz{<{Q{{{{I|r|||}v}}~bՂ2Eȅ~܇5ONlYΌ u%Ɏ/h}ʏ 9ΑAyΒ6Kғ">͔TӕǖLj9B^{kʫzU#U&D]3n)Wq;Yw *HfZ"R a"/6a|8^'d.M l}3b$f*7     _    J]w[7atrW/D5_tT ~  A!k!!"I"^"a####$4$$$$L%v%%&&&(<(Q(( )))*+*+H+]+I,s,,L-v--s.....G/r//+0\0001c1275o55556L6h6667o7;8i88889<="=o===>@ A(AAAAA\BBBBCuCCBDkDDD E%E`EE/FFFFGGGKHiH~HHHIII1JOJhJJJKKL,LEL^LLLxMM N*NCN\NuNNNOOP5PPPNQQQQ.RVRR^ ____`E`` aa*bFbdbbbbc0cNcccccd8dsddddeekee9fi+jGjjjj k"kkklhlzlllm~mmn{{{{$|9||||j}}}}]~~~8J_Nހ!Mf܂EZ4N<IvƋMw֌*<QAsҎvˑ!6Œdv{ >ĖxӘIʙMΚ(KלcǝӞ)KݠRۡ0jբ@hӣ(f!|ĥۥ$QΧR|Ψ5۩0Oi!K`Jkҭܮ8Zt%k8r˳CĴGȵ}"Eѷ[˹ !~CջJӼ(bͽ8`˾ ^t~I Bl% ?Yt;P:[u(Jd K8R)V%Pe{'Rg}R}6gn'A[u-dy8k5a{,F 8Uq2Nj1Mg L(u 2 i+2Lf(uTFWX EZT~1 C X  $O$m$$$%4%I%%&&&'g'y'''+5+S++++,0,,-q---N.`.u..334,4y44G5y556#668J8h88889;S;;X<<@=EO|PP[QQdRRRRSTSSS:TlTTUUtU-mmmm nZnnnnnoo4p^psppqrL~~~~?i ?ԀGL_eć"r.ҍo Bo#>ZtS}M62 9 ,9 P9 9 > \I T Wi n m}    j w h 1 d : T g g E  3 0 4   G  P H # T E [  p     : J % (& - 1 5 @ @ b gg g h 5j sj j ok k Fl m m %n n o Kp p q q q :r r Rs s t t u u u Ov v 3w w x x y ry z nz z j{ { | O} } !~ ~ ~ `  L  M  ȅ  P * G C h  r s ; D  n R c ȕ 0  ϗ :  T \  h ՜ Z ͝ ? ۟ Ѡ 4 * % ۤ I %  q ʧ /  Ʃ f  v  U r W X  ձ  D ޲ = t 2 ϼ B ˽  ? z J - e  h T  C E  B ' , v 1 C ~ o   : } N d   5  D i  5   U   ]  ] # $ $ ( ]+ <. . i2 5 :8 8 ,; S; M R 2S S 6T T @U pV V >W W X sX X UY Y -Z Z Z d[ [ \ \ \ h] ^ J^ >_ _ 8` ` ga a b b c e e 5f f f Jg .h h h Mi i ,j Ek k l _l l Sm m n an n o Hp p Vq q r pr r -s s s t Wt t t 7u u ev v v w Uw w 7x x :y y !z z z V{ { -| | %} } } ~ ~ * H d      x Ѐ ʁ b ւ W l i 8 Ά t  k & ފ ٌ  ΍ 6 v Ď P g  a } ֐ ;  l ݒ @ ۓ 8 _ ( [ " c  T#/4@DTm]Acnq4ƙ+z{Ʋ3#!vV6\.tIC9Ow19<$^Q[. 3jrr 0Lj++>e-n/.}^`  '  {  nLo1b-|33'4B456=@,ADF0GIMPRXl[{}^8PUɕEXU rzTr)=ejBgP nc<R5"qd  1   V     N  t'cfW7n= 7-- Lj6QW8  !h!!!""$#F###$ %d%%%&&'''t()))*V**.+++,k,,,-m--1b468;b>BEpIMQ}WW] bekr{fb~aH(N%~?1qөӪ5d; z x5ַCb5ӻCU~?cRZp$J;+zdwUd V{-iU E'aum6X.C\^:T> ,R\Z@v,|d6(yBHQh5hSG:%+(0472<~?I NVW[XdWjm[nnrPvvFCU]3L.{[k#ocCke:^,Kh~"R$%S/9;=CHySTVY'ZEZ}Z\]^^/_{_C``SaNbbbdCeaee:ffg/hhhiijkk2lbnnnn8oop:qqq;rs-tptt}3L߆b~vӉ,Ku͌?Sܐce9Sؖ|ҜyBڡݣw4jک#հ)xb19:;Ǿeѿ8]z&A$=`\? I[>j Wa,97Xi; Fur,9@a     L    S  U ww_}a2Y-  !h!!A"""h##P$$ %$%%%K&&3'''({((.))*}**L+++s,,[--../3//"001Y182:k;;;x?D`EHMQNNS3UU/VVVeWWsXXmYYYdZZT[[cd^ewe&fHffLggghkk[lmmnoKqcqjrr0sFsosstDtt uEuuuevvvvwPwlwwwwxRxnxxxxyyyyzzz{5`Ykl05UZy4M&lN%C9_*6j2 W  A   I    /u Z5@c\ z>[~C t >>+?S?q??G@@AeAA&BBBTCDDBEEEFF-GGGHHHNIIJpJ(KKPLL1MM=NNNXOOVPPMQQ RR3SSSvTT=UUUNVVW_WWXpXX&Y[\^___``$aBaaabbcWcccd^dddee$fjfff?.???V@@(AeAAAAAB:BVBBBBBCC9CLCCCtDYZ;Z[ZZZF[\j___t`?an:oNooovpwwwx1xMxixxyFyyyyyzzz{.{L{j{{{{{d||||}8}V}t}}}~=~~~Lfɀ2BtЂ)ć8Jr,Am%mZnȐ/L`֒#Fb~єƕ(<b0+P4H"Tg@М!<oݝ0ģ>h (=%O7TpŨ6Soĩ6b%իq b4Pl*2OcC gM6Tr;X 4i(H]H!5]LiF|#`}-WRl@Zw9Uq9V/cYv<Zn&Q?']z<T-J$Zu1J#@Lf 4    G q     K h  = ~ +A,],q,,,>-\-z-----.,.H.d......./S455M888 9F9_9;2<<=>@FA1BEFWFmG IJKK LLxLL M*M?MMOP3PPP[QQQR*RR UUU8VUVW+WWWW@XAYYYQZZZZ[([[[Y\\\]d^^^___/`K`h`````*aLa~aaa bAb]bqbbb$cBcVccc~ddddekee9feffffffVgsg$hPhhhhh!i>iijjjjj7kckkl2lQllllm"mUmjmmmmnnGodoyoo>vvvvvXwwwwx x5xxx0yUyyyFzzzzz{D{Y{{{{]|z|||}}}~jǂ%:Z݅%Cd\zn -(-ގ =pߏŐ+A֑Vޒ >SÓړ0FՔi&ܖZxNPlê ߫Ǭj8]4Re;|1ey+Ig8Vt+h7g-o77Ji'Mr;y"<[ Gw9t?d   H   0kpGe=LCm|b   O!y!!-"W"f"o###$-$<$$$$Z%%%&&&(J(Y(()'))$*3*,+V+e+W,,,Z---..../[/z//?0d000 1w12E5w55566Z6p66777C8w88899<=*===R>@A6AIAAAqBBBBLC}CDJDDDEME F7FFF G'GGG[HwHHHHIIIAJ]JpJJJKKL:LSLfLLLMMN8NQNjN}NNNOO*P=PPPaQQQQAR^R S^____0`M``(aa:bTbrbbbb$c>c\ccccd(dFddddde&eeeNfi;jUjjjjk*kkkHlllm&mmm"n{{{,|A||||~}}}}q~~Rgb)[nǂ0Mb̄(<bD^~ԋb DYčV{~ۑ)>͒$D~#!FҖ,`d6Sx՝7`gȡErݢNp6tɤ)̥Ǧ ,_`J@]q˪6ShӬZyFh V@̱&ٳZ۴^ߵ0M߷p͸/ۺX_=jսFhپ.l!$WP:0Ma&CXJi}6X6,@f=^3^m 5`o fJo5Oi ;lwy=o3:N( *Fc %@\x#?[o0:}3$@Tz0h[ k`Mbi K `  '$]${$$$%<%Q%&'&&&G'''( +C+a+++,#,8,,---..h.}..34 44444\5556+66"8X8v88899$;a;<m<<U=MOPPiQQrRRRRSiSSSOTtTT UUU5mmmnnonnnnnooIpfp{ppqrT~~~~Tq"G܀\Tmƃs!҇06 !J}"Ґ1Lh|Ցca9 = CI S >i n V}   Q O :  , ? = f @  s  5  I l @    $  w D   z    S% % - ~1 }5 @ {b Ng g j j ;k k l l Km m \n n p p Wq r s s s Jt t Ru u #v v w ew w Hx x Cy y 7z z | !} } } \~ ~ .     P  "   X  9 X 8 &  Z  ~ ;  F  C  C   f   x 7 0 W   = '      # z b 2 P o P  V  `  u  e  :   [ C  .     +  Q l  n p * e y \ <  + d .  p " v K     o x * @   h     Y  k 0  " # ( 0+ . . *2 4 7 v8 ; L R S rT BV V W wW W EX X #Y Y Z ^Z Z 9[ [ [ L\ \ .] ] ^ _ |` ` b b ,c d be e af f g ih h i j l 5n n )o p up p q q Dr r fs t zu -v xw zx y y Uz z '{ { | b| | u} ~ ~ G  4 0 l ( H  5   [    T  K  ; I  q 6  \   !     5 R &#/`4T@DTB]cVntq'rjWuQ-mH<},qIy+ _/F8hJs=HGQ!{W Qn  _  h)CBe5 [3346W=@DEGILPXRX:[z}/$+dz,*]a9}FkVcU](0 xwb3&  r  $   U     Hu7;Q0 {`k5wU~m !&"#z##,$$9%&-''),***i++A,,.--94U68;M>B(EEIMPNWWt]axekrz<8T+}fxi3O:8jJNTgcS5K-LQ.nj_W2b1*  OOmLxBx=v+04}7<R?IM[V]["d jmnru} $$E> ^AI_ a ucC!gH3})9$%c.8:<CGRS!UY[h]__abd ffRhikjl3no]q~rt~aul[oJL!eoac,I<RLMm%<b_"*L:r#P8V$A  )   IIS=7X. >!!g"=##$$$J% &v&&p''-()Y))S**r+H,,---.//_0001:;J?DHYM'NNSUeUUuVVW1XXAYYZ [[dGeehjk/llmno3qRrsstgu,vv9ww;xx_yyz\Dh y zU$QWl~q )|6W{ a g   ]Ajl? &8b&T<~  u!/"""Z##U$$a%&&&F'n' (p((p))<*-n.X///900091112 3u3!4Z444d5C666077J88:99:;`;;P<<K==:>>??@z@@8AAA[BBCUDEwE FGGaHH"IIICJJwK!LLL|MNwNN2OOO:RR-TYn[?]y__`` aaabbb?ccdHdddee fRff&gfggg&hhiWiijOjjEkkkTlllmm.n8oopqKrsUuZ_eWIb/A'? 4k#s=^;w}+m1K5dl%Vg"[n7AT  7   4 n  >   ,T!!4"e.001 23355X6@77w89u:7;>Z??@KAA,BsBBxCCYZ[_`nzooswx%ySzz5||}5~~l8lvneAh]|t8=(Zi]GTQEZ@|59 ,2a>O3O8& 'O;oLvGBl4   i   i +',,,/K4o8,9 KJLLlMO_PPQWRU VVvWX9YZs[[\N^?__`Daabbbccd=ee]f(ggHhhji_jj[kkullm%noo6vv+wwbxxyyjzz{1||K}}bfD ] kY@y+VAz<g)+g/w{}7EcR #o5    +/_~ k!I"#$$v%&<( )*H+s,v-..//0:125E66F7;88.9<V==@AuAABjDDE/FFSGGAHH;II(JJKKLLMMMNOOiPPQRR^`y` a bb ccciddReei!jjukklSmm{{{m||}E~~#!h<05vjoPo15kJ)_$+ 7]dG& @0i+/eB!W#/U\~7 0 qlVPR9E:65z \f@_ \;E=X< $$}%r&&'++d,Y--.3`445X688,9;><<EOPQR;SSTKU-mmAnn'opprL~~&LbWBNH9 = CI S >i n V}  ߕ  Q O :  , ? = f A  s 5  I m A % x E  {    T% % - 1 ~5 @ |b Og g j j ;k k l l Lm m \n n p p Wq r s s s Kt t Su u $v v w fw w Ix x Dy y 8z z | "} } } ]~ ~ /   Q Ɓ # ߂ Y Ń : Y 8 ' ӈ [  < G D Î D  g  y 8 ͒ 1 Y > ( $ c 3 P o R W a  v f ˦ ; \ C  , R m γ n p * e y \ < + e . s % w L p y + @   h     Z  l 1 " # ( 1+ . . +2 4 7 x8 ; L R S sT CV V W xW W FX X $Y Y Z _Z Z :[ [ [ M\ \ /] ] ^ _ }` ` b b -c d ce e bf f g jh h i j l 6n n *o p vp p q q Er r t |u zw |x y y Vz z ({ { | c| | v} ~ H  4 0 Ӆ l ) I 6 \  U L ; I  q ޑ 6  ] " 6 S '#/a4U@DTC]cWnuq(˦slY wS/oJ> },qIz, `0G9iKt=HGQ"|W Ro  `  h*CBe5 [3346W=@DEGILPXRX:[z}Ԃ/$+dz,*]a9}FkVcU])0!xwb3&  r  $   U     Hu7;R1|`k5wU~m !&"#z##,$$9%&.''),***i++A,,.--94V68;M>B(EEIMPNWWt]axekrz<8T+}fxiͨ3O:8ѹjJPTgcU5K-LQ.nj_W2b1*  POmLxBx=w+04}7<R?IM[V]["d jmnru}$$E_BJb a ucC!g{4~):$%e.8:<CGRS#U[j]__ab"ffThimjlo_qrt:cwn]qъJL!Ö՗eߘoac,¡Iצ<ͪѫRLMm%<b_"*L:rxRY&C  +  LLV=7Y. >!!h"=##$$$K% &v&&p''.()Y))S**s+H,,---.//_0001:;J?DHYM'NNSUfUUvVVW2XXBYYZ [[dGeehjk/llmno3qRrssthu-vv9ww;xx_yyz]ʼDh y zU$QXl~q )|6W{ a g   ]Ajl? &9c&T<~  u!/"""[##U$$a%&&& (p((p))<*-n.X///900091112 3u3!4Z444d5C666077J88:99:;`;;Q<<K==:>>??@{@@9AAA\BBCWDExE FGGbHH#IIIDJJyK"LLL}MNwNN2OOO:RR-TYp[A]z__`` aaabbb@ccdHdddee fSff&gfggg&hhiXiijOjkFkkkTlllmm/n8oopqKrsWu\afXJc0A'? 4k#s>^;w}+m1K5em%Vg"[n7AT  7   4 n  >   ,T!4"01355@7798;Z??@KA@AxCCZ[`zooxTz6|}~9vnBhޖҗ]|t=ܤi]۪QEЬ@9 aO38& PpvGl4     i ,',,/4o84,9JLLlM_PPQWR VVvWXZs[[\@_`abcc=ee(gghji_jjkkkulklm%nooZvv+wwbxxyzjz!zz{1||K}}f^DkY@Ē+WɗɨB<g+ hSw~8FS p    -.//0:15E66F78.9V==@AuAABDDEFSGHAHH;IJ(JJKKLLMMMNOiPPRR`y``a btaba cacacaidadReei!jjukklSmm{{m||}E~~#hޅ`05joPoʘ15k͜J)`%, 7ʣ]dH& @0ޯi³+/eǷB!X$/¾U\~8  0 rl9E:!z\g@_ \;>< $}%r&&'+d,Y--.`445X68,9?<<OR;SSTKUAnn'oppp~~&bW؏N!Iݒ9 = .I S )i n A} ʕ ܲ < : %  * ( Q  + I ^ 3 W + |  b /  e    >% % - i1 h5 t@ fb 9g g i j %k k k vl 5m m Fn n p ^p p Aq q r s ns s 4t t z z { }{ { K| | ^} } 0   U  3 ~ ى F ? 6 & u Ǐ 2 \ ɑ  q  G w u E x = #/K4?@DT-]cAn_q̅3]پUѿB`<X3s'h[ut4bjuH1#lxQ3\&10:U eMkxAL<W6  I   R- j,OE334n6B=@@pDEFILPCRX%[z}ynOemHֲLq$h1V?~N}@Hu: a~`K   [     >   p  1^ x$Y:dIUa?isV  !!!""c##$$"%v%&'')*k**T++,,,--$4@68:8>AE0IMP9WW_]acelkrz'#?hOaR~8%ز=#PpfT57}9|=PL<34w5:HYUJBMr  k:9tW6b,bu'a+/4h7;=?IMFVVH[ dirmnnYruvvl}~0G,ѡ4Jy'L]`}vNlL,s }Rfh!$$%N.8:<BGkRS U[S]^_ab ff=hhVjkoHqqir~tJ`WFX܋57 PʘmZLlNעr4¦r'ި=78Xp 'MnJ 7#[~;A'V o*    {24>'!uBnC  (!|!Q"'#}##w$$4% &`&&Z''((C))=**\+2,,---.v//I001:$;r;;5?DE}HDMNN{STOUU_VVWXX+YYYZZ[wd2eehjkkllm}noq=rrsitQuvv$ww&xxJyyzF/{Sce?<BnWi\kg!BfnKp Q   F*xTV)}"L>&F7h- _!"}""D##?$$$K%&m&&'Z((Z))&*-Y.B/{//#0w00#1x11s2 3^3 4D444O5-6p667p7488#99:;J;;:<<5==#>>?@c@@!AAADBB C}C>DD`EEFGJHH IkII,JJ`K LxLLeMM`NNOrOOPQhQQ$RRXSSTTTQUUVbVVWsWW)XXXYX[(]c__j```naaabb)cocc2drddjeeel~lllmn"oyoo>ppq|qq5rrrgs@uEJoPBή~۷4ӼMv,*wtzV^(pI&bhtX6 OWAR EX!+>  !    W  ( z  ?!"0Z01345+779";E??5AAB]BBcCCZl[`eoox>z |}~#a؈͉Yw,SɖHטg_~(EǤTHӦ=ƪ<0+z $otL :!m:Za2Wj    T ,,/Z89J5LLWMJPPQBRUlVaWWZ^[[\*_`abmcc(eegghUiJjjk^llmnoojvwwMxxlySzzp{||6}}QӊHT͍D+wj|A,'t|RRbh"0n=Z p  lJi V!4"x# $$a%&'((*3+^,a-.//0%150661789A==@`AABC^CC+DDqEF>G*HH&IJJJK}LLMN OTPPRuR_d` bbbiccSdd=ee jj`kkl>mm{X||}0~~zSɅp qUZГ;̕xZ V 5I"HۤOܦq0)ȯ|rTP- A @ӿGi xp \W:<$0% e8GQ+JkG&(y'x $h%]&&'+O,D--.K445C689)<<PpQR&SST6Um,nnoppq~} ֌MBΎÏ91ǒ{^iJi7B0{_j] {O Z  ^  " gQOp{-8]j#m #GR]" - !!c""# #Y### $\$g$$$ %%Y'''B(M(X((((9))))) *m** +i++,k,,.v/////000,00005d6888T:_:j::*;9;;<<\<f<p<z<<<<=======-><>L>M@X@@@CCC DHS^U tM#hzr}~Ek ep{.MXcT_ X c     3 > D e   S i    - |     ;F"-r\r} <GR^nZI5@(3>IT`#z"Ch BMXcny"u.hs~(<h  j        W!b!m!x!!!"X""d#$#%&&&&&&&''''2'='I'd'''''7((+*b*x***++^+t+++,),`,v,,/-..U.`.k.v.......///)/4/?/K/[/k/:0q0|000001@1K1V1a1l1w11111111 222+2;2K2V2a2q22222222Q3\33 423*353A3g33344!414@4P4_4m4{44455l5544445'5a5w555#6.6o66 777*767F77 88*8r888:::::;;D;~;;;<+=g=r=}====\>>>>>>>>>6?A?L?W?b?m?y???????B@c@@@A$A/A:AEAQAaAAAAAAAAA-B8BCBOB_BBBBC'CiCtCC *  I& *h h )m m r z | 8  5 ~ . ܢ ݬ 0   \     % . 8 ES QT !V _ Y` `g li j l tm r s t Yu v ;w Yx R}  5 , ^ ǰ ;hK0[HaNJ  ;o6r -YfsTa~/9;>DHSTMV ZZ^_`wb'eehhjknRo q]rVt$ÁЁ'4/<;H,9BOŋҋWd|^k_lyIV<IIV    v  )%2~"%(+$/;EQTWZt[\s_t-u vE|#%3*80<<2DDTKK@[M[]]^^3_(`RbeshjamnuփوyH./11s23C445ZZc[_`wxry5zz|\ g RiEhEpl@2l!֨&sݻۼS9 o'V.kuv0{5K\!96lr~8Je FwS<EGz0    y?-$]'(D*-/:0B12+34467;K3LL>QS#UM] ^^_Jdqtvw}9~EM$щ z9ۯ?S-?K^1v?  #&&(''(s)**,C--.AGGXhiijRkllnUofppqs ttttxyz}dH"g#þ_2QukA&! #''p0 19T=P?AyCYEEFJGtMPPVXFY-^_)cdhhiMjkklZmnnoZpqqr7sstwuvvwwxxyzj{GBͶm`-FV!9O:;=:@HPW\lm oJprsucysX&YSHRIOJKKMGN2O`PbQRRrS"TDUUVcWCyx}~' RVnkO3HVV?u" =J>aCDIMoN7PlQFRWWXYZ[\fI{>%ζֹ`MC{"[h AD`@i uQ&t+k,H8Y@@(AABCrDXEFHBJKtLDO4RRSfTmUBV)WWXYsZG[\'djykclmmo]qcuxH|&} q66:֨ Ͳ{nyHe|B`hf v  C   3  _    j H I! ! r" q# q$ -% 0 I9 : \D E RO O P Q R S lT ]U DV Y Z Z [ \ ] . ճ h 7     | G      q  V     ;    " " =# # M% & ' v( j) * * + 1 p2 3 3 39 : h< < s= > > M? @ P P VY]ehjq8`;bØ*W{`-// 7J<P|&~EمN9}б2QQvwFxx|/}6ǰp    /Q,&':+,-0N12446@8P9:<=@@=ABD)FQG\N`gHhz{N||׃m׏3BE4.8=ܟ̠~أ$}٭/8ĴzG Z.TdKUgDB^ c!p"be|C}t~wfΪFPG!!$J%{>>`??;bbh]i3s\  o A   }  d R   * % * @ @ PA B B C F G H H H I 6J J ^ O_ }` ` va 1b `c c Id d Xe +f f g h h h i ]j k m Jo p q s Ht u v w x y [z ~{ | } H ~ n { р \ i \ ͐ ґ  X ޖ Ә c . s | Ҡ O X o ٰ 5 β  X ƻ  = y F    y  0- - 8/ / 0 Q4 4 ]6 6 7 @ )B C &D tE dF H {I J K M cM Q MR R S S T U V W W Z y_ _ 9g eh j bk UaΕIz { } p(+2\;V~ ~ K L yN N VP MR >S T T % † E l yt t Cu u u 4v v v %w vw w x ix x y Wy y y z 4{ | U| | | U} rM(oJ%l ~   _   U Z M ? ߃ k Rl l 9m o nr s t ;v 0 d ، K  Ó  I ؄ , Ѕ q ɇ  k aVV<=,=@=1'1 G*G6>܍X Y Z 5 K Mt ct t .<jz7 x y Gz O Y a u Ű  qv.{3IZ74jp|6HcDuQ:CEx.     w=m$['(B*-/80@12)344678K1LL_CDIWMmN5PjQDRWWXYZ[\bG{t|<#̶Թ^~KAy BYf (?B>go sO(+2[;@&AABCpDVEFH@JKrLBO2RRSdTkU@V'WWXYqZE[\9bhwkalmmo[qauxF|$}o448Ԩ˲ylwFcz@^fdZ  ?   P   M   ~ h F G! ! p" o# o$ +% % G9 : @ E F O UP Q LR =S T T BV X Y Z [ \ ] ^ V g # 5      z E      o  T     9     " ;# # K% & ' t( h) * * + .- n2 3 3 O4 : f< < q= > > K? @ @ P Q W] chjku+;9bHdߦx=|!i9<$~CׅE} {α0˶QUwDxxay-{Ű n   -O*&'8+,-012426>8N9:<=@@;ABD'FGQHL`gFhAy{L||v}Տ1@_2,6;ڟʠh֣{׭-6´xExX,R b5SeBT\ a!n""efA}r~uڍd̪D^!$H%&>^??W@b%d[ij  m t   { A b    ( 1 * 0 @ NA B B C ZD G H H H I 4J J K M_ {` ` ta /b ^c c Gd d Ve )f f g h h h i [j k k Ho o q mr s t :v }w x y Ez |{ | } ~ l y π Z g F Z  Б  “  H M , q z Р M V , W 3 3 ̲  ӳ Ļ Z f ; w D   w  ! - 6/ / 0 1 4 [6 6 7 19 'B C $D rE bF H yI J K M aM P KR R S S T U V W W Z w_ _ 7g ch j `k l `Vqv/{4J[ 85kq}7Id EvR;DFy/    x>n$\'(C*-/90A12*344678K2LL=QS"UL] ^^_`qtvw}8~DL#Љ8گ>R,J]0>  #&&'''(r)**,B--.AGGWhiijQkllnToeppqs tsttxyzV{cGɝ[_¾^1Pt,.@%! #''/1 7-<O?AxCXEEFIGsMPPVXEY,^_(cd~hgiLjkklYmnnoYpqqr6sstvuvvvwxxyzi{O|AMl_,EU!9N: ;=9@HPQ\|lmnIprsuUv|W%X QINJKKMFN1O_PaQRRqS!TCUUVbW\z~փ& l͕DUmjN2Gc"U>tAz =5>`CDIXMnN6PkQERWWXYZ[\bH{|| =$Ͷչ_LBz!CZg )@C?~ h tPs+j,G8z>@'AABCqDWEFHAJKsLCO3RRSeTlUAV(WWXYrZF[\:bhxkblmmo\qbuxG|%} p5~ 59ը  ̲zmxGd{A_ge[  @   Q   ^    i G H! ! q" p# p$ ,% % H9 : @ E F L N N QO O P Q R S kT \U CV X Y Z [ \ ] ^ W h  $ D ȇ 6      { F      p  U     :     " <# # L% & ' u( i) * * + /- o2 3 3 P4 : g< < r= > > L? @ @ P Q Lt t X] chjku,;:bId˜Bu u y>|"j9<%~D؅F~ |ϱ1̶QUwExxby.|ưo   .P+u 3v v &'9+,-012436?8O9:<=@@_??X@b&d\ijU~ 2r^    n u    T | B c    ) 2  Y L * 0 > @ OA B B C [D G H H H I 5J J K N_ |` ` ua 0b _c ރ c Hd d We *f f g / h h h i \j k k Ql l 8m m Io p q s Gt u v ~w x y Fz }{ | } ~ m z Ѐ ׄ [ h G c ׌ J [ ̐ ё  W ݖ N - + r { Ѡ  N W - X 4 4 Ͳ  Գ Ż [ υ g  < x E p   x  ! - 7/ / 0 1 4 \6 6 7 29 (B C %D sE cF H zI J  K M bM P LR R S S T j U V W W Z x_ _ 8g dh j ak l y{1щڐac77kYlmmn)?G} cAAP>?,8D9llmvzzłܒ  .EADYH?JLN8bb"cc ddd9j{ّ^ߙb>XXٴ\ݵ.H"R~a4VhVV#( =%=9=M=141G7G=C>L N N JO Y Z Z B X w Zt pt t t ;<w DfAc>` ;!+k Jl l 1m m x y Tz \ C ̘ \ f n Ұ  8VlVV', =)===Q=181G;GAG> L N N NO Y Z Z F \ { ^t tt t t ?<{ !HjE gBd ?!/o Nl l 5m m x y Xz ` Č G И ` j r ° ְ  9 > _I T Zi n q} "    n  { l 5  h > X k   k    7  3 7     L '  ^ H   ^   y     C S % 5& - 1 5 @ b rg h >j j tk k Kl m m *n n o Pp %q q r [s s !t t u u u Rv v <w w x x y uy z qz s{ | X} } *~ ~ ~ i  U     V  &     U  /      L  k    {  |  >   G  w W    C  W  _    c   %    -  .  L  (   t  2    o    b  c      G  D      j  Y  H      E *        t       :  f  f  !# $ ( f+ E. . r2 '5 C8 8 /; +M R S T yV V GW W X |X X ^Y Y 6Z Z [ m[ [ $\ \ \ q] &^ I_ ;` ` pa b c c %e e 8f f f 7h h Pi i Pk Vm dn n o Kp p Yq q r sr 6s t v w y y ,z z z a{ { 8| | 4} } ~    g   l =   }  n  /         U  f  @   q      A b 1 d +  l  ]#/4@ETp]Jcnq~,0&eE!k=!XFB'T|DGi\o"w w5Uw44"2w#ac ,   sOr4g 03,456=@DF5GIMPRXq[{}c=SZH]Zu }Yu,oGSs#fAU:vi 6   Y  !   S  y, fkbByH2Om;=  !"/##$$%i%&'y(*[**3++,p,,r--g468;g>BEuIMQWW]beks {kgfM-D6v:%}:H8!B)W 0 i |i[2nH*fzr;1_c=WC$ 1aei;~EMVk:,-0475<?INV[[d\jmrUv}yTWO"lD# >>8A"If6O~^pp$]$%b/9;=DHST'V\]>_R`babIfghijAloqst@NLfh>|E |7 o }g6>=@j;`+F)BeaDNw9NC     hbf^!m!!"m##U$$%%P&&8'''~(3))***+x,,`--.//'00 1^1;;}?DHMVNNS<UU8VVWX|XYvYYgZW[[dae+fhkk^lmmnoNqmr3sttunvvUwwWxxy~z{^no3:_7oQ *H>d/;o7   N  4z_:C h_^ A!!W""##$p$$v%O&&%'<(())*n*+..r//0g001h11c22O33;4w44!55`66 7`78|899l:::;;<<==>>?R@@ApAA1BBCDDMEEFHHYIIJ{J7KK[LL<MMBNNN]OO[PR8S{T[\^__``-aaabbc`cc#dcdddee-fsffAggghAhh2iwii*jjj kfkkkolll/mmOnboppqyr?su\}?' gY"nUp>BQ?R*l]6aR 4=73Hho'   f 0  F   f   @Q!"N0 23:5;6s7Y8W:;?[@CyD[\Tao{pyzm|~QoQ.wg0* v/lR@-bQ e2> +!26(#  L P   ,3/}LMMP`QQR=V WWEXZ[b\]_/aabcdpe>f[g)h&iij<kmnLoo]wxx5yy{b||}~%y0[5nYo=B*O}FR `/D00|15_66789=W>AvBBFGGHIJKLMNOPfQFRS5`aeSfjkQlmm'n|}}v~!g5gcg[M&cg;}ipNSy iS; ]a5uahFKqYC+kBkO8$"mvn  !%&&P'(,,-7..4a55689x<`=nSTTTUtnnoNppY'a& l,j HwCMDȰ \RT%4_)`Sbethjbm g @1 ?3 F9 x = F c X @ dh ij r z { ̄  ˑ x ߕ = q џ ɡ % p  ] ( ż @ # ^ n 9 0 ~ _  $ <; S T U 3^ X_ g h j vl r \s t u #v v x #y } ~  ގ Δ Ld~#[I  @;$01r384ԝ[]x-#G+M!"<#%&'()+.1DZ1̷KY;wqK}NHxs)yQ#0%]4ȟzɢK$ ,T|0A4!g"YNZ]L_h`tadjeghij+n op r=su~N͇>ZII@_!1tP\[   j(!7""%%'(*B++%..)/.2S)WYbcme>fNs#ttuvvbwwdxxyyzû VCM  7  u  %='e'' )*m,-7./3-559:>Z?E GG9Y^ __`8a*bcdfLhivjk:mmZnost}{0S+X_Ubv_r D   f!!!]./0p12F334G5O66 777n8~9l::>#??@AAABBBBCACCYdZ[_&`nCoolwrxyyz{y|}+~~d{:dłx?!6]cڐU˒1|%Eә=\/dy *2=̧>w̨= iR tXo|M"=R'*v^3rEsa?Kc0'1D?:5)   _  ] 2 +f,,.B4y5788;=@EJKLLM4MO(PPuQ RRUUJV WWW0YyZZ[[[\\E^^_`:aaafbbKccde}eRffhg=hh3i jjjPkk9llJm_mmmnYono.vvvlwwx*xxJyy.zz9{N{{{o||}}Z/;ŅL"ez ߐ6ˑn3Hϓ;ޔ hq|RZ0 nQ"} O'^q5YC_',   tQ/N ;!"[##$F%& (()+C,F-m...o//Y0012l5556e6672888<==@>AA_D"E'FFGGHIIeJJK[LLMrNNO2PPxQQSR^_B`aa[bbEcc/ddeei}jj kkkwllmms{!|6|||}}~G\|cقBW13sߋt9Nލpm3’s;ОgҢeإ!˧yҩfݪH]Ѯq"5ȹ_ʽ]iV8Mra 5~Mt vzd av.-xCMF d%/(I%=PBW{@ U $$$1%F%&&v''*++,-,--].r.3)44v5 6 68~88;;<=OvPUQ\RRS~SSiTTU%mmm nnnno[pppqD~~f<рD^k+9q & s uv ~ dnYern-k-\ ^''>rG'!.0Q2)4p68:>@Yz_owVyz|M~ـ}ϑΓ2P@_rDXVg?*a7d4Z( +,c4O0UQYt^`\abbduf`hskNvyzj,8,*ӪA;=GIt  2S8<@DGFHIKMOQ^8ai{9T nM8kp7$+3284;]OEmd~dZ  y  s hv ~ WԞaYden ^ O Q'y'>eG!.0D24c68:>@Ym_nwIyz|@~̀wp‘%C3Re7KIZ2 T*W'Mt +,V4O#UDYg^_OabbdhfShfkAvym]+ ƪ4.0:|<g  2F8<@D:FHIKMOQ^+ai{,Gaz@+^c*$+3%8';PO8mW~WM "-9=AEIMQUY]aeimquy}    ! / 3 : K [ j {   4 I ] p  $ = ^ n r v }   $ 7 I P T a o s {    < W u 4DLPXmq *2IQVe 3FUex  5Kaw )<Yltx#+3;CXmz (05:BJRZbjnv  $,4<DLZbj*2:BLPX`hpu !%-5=A_dhp{DHVZbnsx}'/4<AIYckp|!)19AINS[cw  %*/49>CHMRW\afkpuz  2:?DINV^fnv~%*2:?GOW_gow !&.6>OW_muz #+3;CKS[ckpx    ' / 7 < D L T Y a i y ! !!,!4!,C,H,M,R,W,\,a,f,k,p,u,z,,,,,,,,,,,,,,- -%---M-U-q-y-~-----------. .....".'.,.1.6.A.F.N.V.^.f.k.s.............////-/>/F/N/U/Z/_/d/i/n/s/x/////////////////00 00000$0)0.03080=0B0G0L0X0`0e0m0u0}00000000 11-1=1V1m1}1111111112242J2R2j2r2w2|22222222222222222222222222233 3333!3&3.3:3B3X3f3v33333333 44/474<4A4H4M4\4j4o4w444444444444445 5555*525E5J5O5a5i5v5555555556 686F6W6\6k6z66666666666666677777$7,7:7B7G7L7`7h7z77777777777788 88%8-858A8Z8b8}888888888888899 9999 9%9*929:9B9Z9_9d9l9q9y9~9999999999999: :::":*:/:4:9:>:C:H:M:R:W:\:i:u:z:::::::::::::::::::;; ;;;;;#;(;-;2;7;<;A;F;K;P;U;Z;_;d;i;n;s;x;};;;;;;;;;;;;;;;;;<<&<<>>>>">'>,>1>9>>>F>K>S>X>`>e>j>r>z>>>>>>>>>>>>>>>>>>>>????#?+?@L@Z@j@o@w@|@@@@@@@@@@@@@@@@@@@@@ AAAAA"A'A,A4A?ALAZAjAyA~AAAAAAAAAAAAAA BBB%B9BABXB`BeBuBBBBBBBBBBBBBBBBBBCCCCC$C)C1CBCQCYC^CfCkCsCCCCCCCCCCCCCCCCD DDD D(D0D8D=DEDJDODTDYD^DcDhDmDrDwD|DDDDDDDDDDDDDDDEEE!E;EZE_EdE}EEEEEEEEEEEEEEFF FF'F4FEFJFRFjFrF}FFFFFFFFFFFFFFFFF GGGG#G+G0G8G@GHGPGUG]GbGjGoG~GGGGGGGGGGGH+H3H8H@HEH\HvH~HHHHHHHHHHHH I)I1I6I>ICIKIPIUI]IbIjIrIzIIIIIIIIIIIIIIIIIJJJJJ#J+J>JCJHJMJRJ`JJJJJJJJJK KKK,KGKLKQKlKtKyK~KKKKKKKKKKKKKK LL$L,LJLRLkLsLLLLLLLMM M(M6M;MCMKM]MfMpMuM}MMMMMMMMMMMMMMMMMMNN$N,N5N:NBNJNSN\NeNjNrNzNNNNNNNNNNNNNNNNN OOO"O+O4O9OAOIOROWO_OgOoOwO|OOOOOOOOOOOOOOOOOOP PPPP&P.P3P;PCPKPSPXP`PhPpPxP}PPPPPPPPPPPPPPPPPPQ QQQ"Q*Q2Q7Q?QGQOQTQ\QdQlQxQ}QQQQQQQQQQQQQR R R@R\RvRRRRRRRS SS$S8S?SPSaSvSSSSSSSSSSSSTT-T@TZTpTuTTTTTTTTU'UEU`UyUUUUUV1V6V=VBVJVWV\VaViVqVzVVVVVVVVVVVVVV"W'WDWIWNWSW[WcWkWpWxWWWWWWWWWWWWWWWWWX XXX X(X0X8X=XEXMXUX]XeXjXoXtXyX~XXXXXXXXXXXXXXYY(Y0Y8YAYOY\YmYrYzYYYYYYYYYYYYYYYYYYZZZ-Z@Z]ZbZjZrZzZZZZZZZZZZZZZZZ [[[[&[.[3[;[C[K[S[[[o[t[y[~[[[[[[[[[[[[[[[[[[\ \\\!\)\.\3\<\D\M\U\Z\_\d\i\n\s\{\\\\\\\\\\\\\\\\\\] ]]] ])]1]:]H]M]R]W]\]a]f]n]v]{]]]]]]]]]]]]]]]]]]]^^^^^#^(^-^2^7^<^A^F^K^P^U^Z^_^d^i^n^s^x^}^^^^^^^^^^^^^__#_5_=_E_N_V___d_l_t_|_______________`` ````"`*`/`7`<`G`T```l`t`|`````````a aa!a&a.a6a>aCaKaSaXa`ahapaua}aaaaaaaaaaaaaaaaab bbbb#b+bCb[bsbbbbbbbbbbbbbbbbbbc$c4cKc_cdclcqcyccccccccccccccccccccccd dd(d0dBdYdadfdndddddddeee+e8eKege}eeeeeeeff%f6fGfLfSf_fdfifnfsfxf}ffffffffffffffffgg0gJg`g{gggggggggh(h0hLhhhphhhhhhhhi*i?iGi`iwi|iiiiiiiiij/j7jTjYj^jcjjjjjjjjjjjjk k#k+k3k;kCkHk[kikzkkkkkkkkkkkkkkll lllll$l)l.l3l8l=lBlGlLlQlVl[l`lelmlrlwl|llllllllllllllllllllmm&m7m?mGm`mzmmmmmmmmmmmmmmnn3nYnun}nnnnnnnnnnnnooo!o-o2o:oBoGoOoWonowoooooooooooopp!p)p1p6pEpJpRpZp_pgp{pppppppppppqqq+q>qRqeqnqsq{qqqqqqqqqqqqqqrr rrrrr#r(r0r5r:r?rVr^rfr}rrrrrrrrrrrrrr sss$s-s8sAsJsYsisnswsssssssssssssst ttt,t=tBtKtTt]tjtstttttttttttttttttuuuu(u-uwFwKwSwXw`whwmwuwzwwwwwwwwwwwwwwwwwwwx xxx$x,x5x=xExJxRxZxbxjx|xxxxxxxxxxxxxxxxyyyyy>yCyKySy[yjyyyyyyyyyyyyyyyyyyz zzz%z-z5z=zEzMzRzZz_zgzozwzzzzzzzzzzzz{&{3{;{F{N{V{^{f{n{v{~{{{{{{{{{{{{{{{{{{| |||#|+|4|<|D|H|P|X|]|s|{||||||||||||||}1}5}=}E}W}_}g}o}w}{}}}}}}}}}}}}}}}}}~ ~~~!~)~I~Q~Y~d~h~p~x~~~~~~~~~~~~~~ */MRn{ %-5?GLT\dlu}ɀрـހ "*/7?HPX`hpxǁρׁ߁+3;@HPbjrzǂ߂#(05=BJOT\nzǃЃ؃  (08N[uÄӄބ $)1AIW\dwЅ&+08IUZlqɆцކ !*2FNai{‡ˇׇ܇ "*3;CKShpxLjψ׈߈ "-2BGOTY^fkpxʉ݉ $,16>RZbjrwNJ̊Ԋ(08=EJR[chpu}ʋϋ׋ 8A\dŒnjό܌ (05=BI`hpxÍȍЍՍ 6>IQV^ckpwԎ܎ $)1HPYafns{ǏϏ׏ߏ';@GOX`invΐڐ&.3;@HMRZs{̑ԑّ &.38@EJRW_glqyђْ&+3;CHPYapu}Óӓۓ$)16>CKX`emrzȔ͔Ք ',49>FRW_diqv{ȕ͕Օ #(05=PU]bglty~$)19RZcpy~ŗʗҗڗ $,16>CHPU\ains{ØȘИ՘ژߘ '?GOW\dlu}șЙՙݙ %*27>Tg|ɚњ֚ޚ+4CXa{ݛ08@L_glt|ܘ֜ޜ "'/7FNS[`hw|ŤʤҤפߤ )2;@HMUclqyĥɥѥڥߥ */7F`irzIJɲѲֲ޲ 8=EJRW_z̳ѳٳ ".6>CHPUZbgow|´ǴϴԴܴ #(05:BGOT\dinvĵɵεֵ 5=EJRZ_goxŶʶ϶׶ܶ',18=EJRW\dy~ķɷηַ3;DMRZ_glty~¸ʸ۸"*/7?FMT[agmsyŹ˹ѹ׹ݹ!(/6=DKRY`gnu|ºɺк׺޺$+29@GNU\cjqxŻ̻ӻڻ  '.5ELSZahov}ÿʿѿؿ߿ %,3:AHOV]dkry !(/6=DKRY`gnu|$+29@GNU\cjqx†”›¢©°·¾  '.5ELSZahov}ƄƋƒƙƠƧƮƵƼ %,3:AHOV]dkryǀLJǎǕǜǣǪDZǸǿ !(/6=DKRY`gnu|ȃȊȑȘȟȦȭȴȻ$+29@GNU\cjqxɆɍɔɛɢɩɰɷɾ  '.5ELSZahov}͙̈́͋͒ͧͮ͠͵ͼ %,3:AHOV]dkry΀·ΎΕΜΣΪαθο !(/6=DKRY`gnu|σϊϑϘϟϦϭϴϻ$+29@GNU\cjqxІЍДЛТЩазо  '.5ELSZahov}ԄԋԒԙԠԧԮԵԼ %,3:AHOV]dkryՀՇՎՕ՜գժձոտ !(/6=DKRY`gnu|փ֊ִֻ֑֦֭֘֟$+29@GNU\cjqx׆׍הכעשװ׷׾  '.5ELSZahov}ۄۋےۙ۠ۧۮ۵ۼ %,3:AHOV]dkry܀܇܎ܕܜܣܪܱܸܿ !(/6=DKRY`gnu|݃݊ݑݘݟݦݭݴݻ$+29@GNU\cjqxކލޔޛޢީް޷޾  '.5ELSZahov} %,3:AHOV]dkry !(/6=DKRY`gnu|$+29@GNU\cjqx  '.5ELSZahov} %,3:AHOV]dkry !(/6=DKRY`gnu|$+29@GNU\cjqx  '.5ELSZahov} %,3:AHOV]dkry !(/6=DKRY`gnu|$+29@GNU\cjqx ) s xv ~ gqY eun0!n0_ >uG)!.0T2,4r68:>@@@@@AAZ}_owYyz|O~܀ёѓ4RNV^w~t F[XA-d9f6\* +-}444444444O2USYw^`^abcdwfbhkkkPvzz|m/‹:/,֪C>@ILw  !+5?IS]gq{2U8<@DIFHIKMOQ^Vaja~aaaaai{;VONU\clu~nr9$+3487;kOsO{OOmWmf~nw~\ m9 n9 o9 p9 q9 r9 s9 t9 u9 v9 w9 x9 y9 z9 {9 |9 }9 ~9 9 9 9 9 9 9 9 9 K V W X          iA jA kA h j j j j j j j j j j j j ar br cr dr er fr gr hr ir jr kr lr mr nr or pr qr rr z { { { { { { { | | | | | | | #| (| -| 2| 7| -       l m n o p q r s t u v w x y z { |    ! " # $ % & ' o p                 ®   { 7 O 2 o p 3 8 X Y Z [ \ ] ^ _ ` ?   r s t       $ $ $ $ $ $ $ $ $ ; ; 7S ;T iU nU sU xU }U U U U U U U U U U U U U U U ^ _ _ _ _ _ _ _ _ _ _ _ Qg Rg h h j l l l l l l l l l l r ^t ct ?@ABiDEF^_`hY^ hijklmno{|}~WGL*+,-ot$)_%`%a%b%c%d%e%f%g%h%i%j%k%l%m%n%o%lܝݠdefghi;<=>?.38=?P"""""""""""ZZ^^_`UbZbeegghikk=opq@r4t9t  {34Ҍ^_`a(˱̱ͱαϱбѱұӱԱձֱױرٱڱ۱ܱݱޱ߱"GHIf>C]n ^ _ ` a 7F""""##)%%%%%% ((((((Q++,, ,,.@/E/J/O/T/M2N2O2P2Q2R2S2T2U2TTTTTrWwW|WWWYcc|eifjfkflfmfnfofpfqfrfsftfufvfwfvs|sItuuuvvvswxwwwuxzxxyzz01qrstuvwR1234 !"#$%&'()*+,\ F            %%%'''3)8)=)-F.-/.///0/1/2/3/4/5/6/7/8/3334<566666 6!6"69:::::::::?????????????E,H-H.H/H0H1H2H3H4H5H6H7HU__R`S`T`JaOaTabbbbccccHeIeJeKeLefffffhhhiiijjjjlll lmmttttttttttttttttttttGHIJKb<Bgm nd  ;012345 b g -v! "/1v34w5779 ;3??QCCZK[y`Sooxz{}~ՂOƈFseےA5ĘUMlB*h]b:n( [$sDO E 9   m B v,.#LLDM8PPQ/RUZV0WWZ-[[\_`avb[cceegxghCijjommn~oww:xxZy^{{|#}}?i22FXK+bj@~l<R %DM //i0156u6789/==NAAREF,GHIuJJkLLNNBPPQcR_R`+eej/kkl+mmF||}}~lsgA ^CҒ)(ijklmnopqrstuvwxȠΠ͡ww| .̦38vëm [cdefghijklmnopqrƻżoܽot&+0f];EtSt5?8Y5ge $V%,&&'+=,--.94450689<<SSyT#Unnoop~vLČ;'z y j Y +{1   h5r$*%'o('-./ 1z123c;;$"<<$<b==&>>>F??@i@@'AAAD#B#B"CC CEDi  ؐ ݓ E*2eqrHsW{@o""7#   fhj œהN~ !h"#>$%%'[())5*g+,-.BHUOoq9 P Y  lA h j sr z <| .  }   ( q  ! Ŭ î %  8 T 7 q = a D  u  $ ; 8S @T U ^ _ Sg h j l r ht Au #w Ax [y }  ڀ ց   G g N Rn 34o=N[i""X#%&')+1XCnGamc p\Q.y.p% qj@BFnD U"Z^_`_beghik>oqEr>t |5׌b-  'JkHbs/< b  8G" #*%% ((R+,.Y/V2TWYcexfsNtuuvv}wwxxyyz28xW5-a G    %'B)-K.9/4A5#69:?E8HV__U`YabcMefhij%lmmtLgHssj@6o l 2{!"/1{34|577779;8??VCCZP[~`Xoox$z|}~ ڂTˈ/LxjF;ʘZRqr+GƦ/mbKg?s-`)xI?T%J>    r G {,.(LL-MJM=PPQR5RU_V5WWWZZ[2[[\\_`a{b`cceeg}ghHijjjXmtmmnngoov wzww#x@xx_yG{c{{{}||)}}(DƒnŠ77sKّA]ݓP 0goE qAl*W*zIR . /}//n001z556#6z67$78 94==SAAWEF1GHIzJJpLLNNGPPQhR_W`0eejjk4kkll1mm/|K|||}~~UrxÃPmFGd,Hג.Ó.yԠҡ|/ϥͦ=Ũ{ȫVs`s̻ʼty '5kFc@Jy#oXy:D=^:PlN k $$?%[%1&&''++&,B,--k..>4456668 9<<RSSS~T U)Unnnnooipp{Q߀Ɍ@,O_RSU]\Y[^afVZXTbge`NWcQP                                                  "#x     "#}|                           oklrsq  u  """""####   """####                             wwwwwwW&W%WX$$$WËÖW֋ X$$$WW Ĕ TW Ĕ X!W_"""_ȏ_&_%_ MMM֍MM 60ڔ_ _ _ _&__%_f&ff ffffffffffffffffffffΞfff  $ 1 HHH SetValZST 1 1 1UC C C$$$$$j jj$ $$j$ $1Mut Ur1/Q2=11Immut |~1$g00a0."> mg111NodeRefheight@;(c10j4V=@;(c1U11U@;(c@;(cUU@;(c@;(cUU      ggggggggggg g#g#g%g%g'g)g)g*g+g1g3g5g7g7g;g>g>g@gBgDgGgJgLgQgQgTgg^ggpgvgyg}gg g g U ̵̊̊ϊЊD^)cOppʽp opopo2Ы ? ̵̊̊ϊЊD^)cOppʽp opopo2Ы  U  $e ;ppʽp opopo2Ы;e ppʽp opopo2Ы eppʽp opopo2Ыeee ב) ڔڔ   釕U  釕 U 釕 ҈ 釕   U  ߉U'ʼn   ʼn  1111111111111 1 SplitResultleftkvright棑)HK1U   DormantMutRef     nݜw  W 1  1      ׯ0W1 1 1棑)HK1U   珕     nݜw1 棑)HKU   珕     nݜw  W 1  1      ׯ0W 棑)HKU   珕     nݜw  W 1  1      ׯ0W1111CD  Um 1  11111 釕  釕   111111 釕 1111DCloned,:^/RevFœrMND,:^/D)DD݂D D II]ێێ ݎ NJH"111199ˣÜĜŜƜǜȜ_w 3f*ہõõĵ AllocError ]YO^`U9  88ArcInner8strong787871%wxdU1Ђݠ P11檕1檕檕1D DDD͞DDDD;33$ MaybeUninituninitW췜 έ㭕W췜 $  㯕   RawTableInner  bucket_mask K ctrl K  growth_left K items KTGפ$  ljk#Ͷ;$$22#2.\*="22#2.\*=" 22#2.\*="ǯ22#2.\*="ߧ22#2.\*="22#2.\*="22#2.\*="11޺1.1.?3 *r11޺1.1.?3 *r 11޺1.1.?3 *rǯ11޺1.1.?3 *rߧ11޺1.1.?3 *r11޺1.1.?3 *r11޺1.1.?3 *r11Ӡ2.2.2 .뜿Er11Ӡ2.2.2 .뜿Er 11Ӡ2.2.2 .뜿Erǯ11Ӡ2.2.2 .뜿Erߧ11Ӡ2.2.2 .뜿Erɡ11Ӡ2.2.2 .뜿ErİP"00Weak0 -0-n?QQ--RcInner------6>)\;;CCC=C=  ScalarRange    `S$҇2; 1CCC=C=ȷ٭ǯCCC=C=ޟCCC=C=CCC=C=CCC=C=$'88Ⱦ8 787=֍B$88Ⱦ8 787=֍B88Ⱦ8 787=֍B88Ⱦ8 787=֍B88Ⱦ8 787=֍B888787871%wxd888787871%wxdljk#έ㭕W췜ɕljk# InternalNodeedgesy1ljk#˕˕y1ljk#˕˕yljk#˕˕yljk#LeafNodeparent parent_idxkeysvals7BCJ-1ljk#ЕЕЕЕЕ7BCJ-1ljk#ЕЕЕЕЕ7BCJ-ljk#ЕЕЕЕЕ7BCJ-((RawVec+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ;((Օ+ ,XՔ((Օ+ ,XՔ ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ֨((Օ+ ,XՔΩ((Օ+ ,XՔ((Օ+ ,XՔΚ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ͞((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔĕ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ((Օ+ ,XՔ  ScopeGuard dropfn   ڻ>  U   TableLayout  K  ctrl_align Kk%      ڻ>   RawTable & K  K   K:1{ho  U      ڻ>   UřǙșəʙ˙G*R{9     hash_builder  & jS     & jS   & K  K   K:1{ho釕777Os7677Simple7677 SimpleMessage7677Custom76VN7O CustomOwner6 fDIntoIter (AzFoqf???=?=?й=?=? =?=_O ???=?=?й=?=? =?=_O    "<< 0{?i<< 0{?i1ꗕA 1  k Q\?1A 1  k Q\?1A 1  k Q\?A 1  k Q\?HHHN:棑)HK1d`.jG棑)HK1棑)HK1棑)HK1ПšӡC© _Ͷ"ll CopyOnDroplsrclllllnȘΉqllllllllnȘΉq$jj MergeStatejjjjjjAdjjjjjjjjAd$  VacantEntry Ü  handle  dormant_map     DSs1   Vacant   Occupied =? 21Dropper A1 A A ACC FillGapOnDropCread=CN=C=&6j:dr  DropOnPanic   |z.     |z.㯕33 DropGuard323num_init2izQ:C33323Ã2izQ:C33323Ã2izQ:C33323Ã2izQ:C33323Ã2izQ:C33323Ã2izQ:C֨33323Ã2izQ:CΩ33323Ã2izQ:C33323Ã2izQ:C͞33323Ã2izQ:C33323Ã2izQ:C33323Ã2izQ:C@@@@? ՝h0~@@@@? ՝h0~  0HY,H*W1  0HY,H*W1  0HY,H*W  0HY,H*WD; BB SetLenOnDropBBB local_lenBD;D ;UD@U DҁQՎUՎUҁDҁQˏˏҁIͶUIͶU> range_stack 1(cvȏ" ;6666666 Bi֨ߧÕ PanicGuard وj֣ ݒ ; Ȅ  ֨      ;釕1y[[nn m[}OD^UnȈ# m[}OD^Uvvvvv Handle@idx$CU@;(c8A11ve̚@蚖$CU@;(cߜ1_zTl1››U̚@蚖$CU@;(cߜ1̚@蚖$CU@;(cߜ1ž1U̚@蚖$CU@;(cߜ̚@蚖$CU@;(cߜž٣U̚@蚖$CU@;(cߜ̚@蚖$CU@;(cߜžååU̚@蚖$CU@;(c1 11ڨڨU@;(cf Z1ʪ1ʪʪU߫ʪUǫ@;(c11UU@;(cѭѭѭU׮ѭU@;(cUU1 3f*ہЕ 3f*ہҕ 3f*ہԕ 3f*ہՕ1α1ְα1α1α66 77767o67 error_drop67 outer_drop6*I]@DUMڔNNk k$CCC=C=$ll$ll$mm$mm$mmm$mm$ll$jj$jj$jj$jj$jj  U jj$j$ $ $U j$jj$ww$A AAA͞Ag333333֨3Ω33͞333iBBB DefaultHasherB ldaDppʽp opopo2Ы$$ $US鱕U鱕;;B ?U$$U$DU UiiiiiI I楕AAFːI I楕;ģ;<Rm[}OD^UƖ333$DD)DD݂D DDģD++QQppʽp opopo2ЫѡPːQҁSR<ƖƖƖUƖĝ<ƖQQꃕSppʽp opopo2ЫR ͶR楕ͶS̵̊̊ϊЊD^)cO鱕R̖QģҁSģQՎR QˏQBB==J/V R Q ii'iiR݂ŕR) R•$$$$$$–11 – ? Ö  ڔڔ 釕U  釕멕U 釕2   釕2 Ԗ ʼnӖ ʼn ʼnԖ ʼn2   NJ2 NJՖ   ֋   D;D;D D1힖11ܫ1Ԯ  ډ ҈߈  ን ҈ʪǫѭ쮖DҁDҁ㑖DÖÖj绖j$ü伖RƖǖS鱕QՎQˏԑIIÒҒ<ƖƖ˖˖$"mřǙșəʙ˙G*R{9  řǙșəʙ˙G*R{9  UD 1ِ 1  ʖ      1C C C  19ߧ9評"6nᘖn͙ ؔ ٔC;CCCCCCCCCCC֨CΩCCC͞CCCCCCĕC232$3wwxwwxxxwwwwwxwwx$x$w$www֨wΚwwwwwwwwwwĕwwwwxxxxwxwWXͶWͶWXXljk#WÖWWXWWWWWXWXWWWWXWXWWWWWXXX XWWWWW____ Utf8Error valid_up_toy error_leny"C*uH__"_ն_ն__նͶHkk$jj$??hڦڦۦ Sip13Rounds I%1111#ppʽp opopo2Ы#ppʽp opopo2Ы#ppʽp opopo2ЫDD  XXXX"ppp opopohEC8G 1D DDD͞DDDaͶDF;FFFFFFFFFF֨FΩFFΚFFFFFFFFFĕFFppppppppp֨pΚppppppĕpp 1 1  ### 1 1 1 1      .Ͷ 釕  .)M;M;MMMMMMMMMMMMMMMMMMM֨M֨MΩMΩMMMMMMMMMMMMMMMMMMMĕMĕMMMM11111111111111111111111CCCCC CCCCCCCCCCCCCCCCCCCCCCCCCCCCCC֨C֨C֨C֨C֨CΩCΩCCCΚCΚCΚCΚCCC͞C͞C͞CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC͕͕̕̕ΕΕϕϕЕЕҕҕԕԕՕՕpppp11ߜ1ߜ1ߜߜ1888888888111ߜ1ߜ1ߜߜ11111 釕 1119$9999911ߜ1ߜ1ߜߜ19911 1ǯ1ߧ111111 1 1ǯ1ǯ1ߧ1ߧ111111111111    1LB)BB݂BB)BB݂B)BB݂""' 1 1  PڔPP݂PģP PPP<< 0{?i̵̊̊ϊЊD^)cOΖP<< 0{?iڔPPP<< 0{?ippʽp opopo2ЫP<< 0{?ippʽp opopo2ЫP<< 0{?ippʽp opopo2Ы֨P<< 0{?iP<< 0{?i˖PPݢP P<< 0{?iPP<< 0{?iP檕PDŽPPPPPPppʽp opopo2ЫĕPב)PΪΪѪoҪoS #ұPPppp opopohEC8GPppp opopohEC8GP,,=d4;_P,,=d4;_P))׮ǦT=P,!-VI砆xP###,#؉2ȳ"P###,#؉2ȳ"P###,#؉2ȳ"P""","؉w7{t0PʽA 1 CPzE1PݒP݂P P P)PPPPPP̿PPNNNN1C  㯕333333֨3Ω33͞333@@11ˆnn;n;nnnn n nnnnnnnnnnn֨n֨nnn͞n͞nnnnnnnnnnnnnnnnnDΖDDDꃕ<˖=ב)jj$""""""""";ss###R;RR RRRRRRRRRRRR֨RΩRRΚRRR͞RRRRRRRRĕRRRR   Bģ#= 0ц$; 1    .%T;        w hث/ȏł֨ΚÕ";;####֨######֨###; 1֨ΩΚ͞ǩĕΩΩCC)nnoͶ"hhhLKɖppʽp opopo2ЫKɖXXͶXXXXXXXXXXX@@@@L*o***MoooΚoll$DDDDDDDDD֨DΚDDDDDDĕDXXͶXXXXXXX˝˝(noDDDDͶ˕͕̕ΕϕЕҕԕՕ333$3$>111:$:::://1 YYYYYYYYΞYτYYYYYYYYΞYτYY/:$::::::22ߧ22::22 2ǯ2ߧ2221;;LBBB BB")pppppp$MڔFJɖI I楕;ģ;;;˖<—<ڔ<<Ö<ppʽp opopo2Ы<ppʽp opopo2Ы1<<<<ė<ppʽp opopo2Ы<ŗ<ppʽp opopo2Ы<Ɨ<ppʽp opopo2ЫΩ<ppʽp opopo2Ы<<ppʽp opopo2Ы͞<ppʽp opopo2Ы<<<ב)<ב)<˖<<<݂< <<<<<jj$DDDDDDDD֨DΩDD͞DDDD*D;DD DDDDDD֨DΚɕDD͞DDDDDDDDDɕMD;DD DDDDD֨DD͞DDDDDDDɕDD++++Mn++@@$$DDXDDDDD 釕 D++1EEE;EE EEEEEEEEEEEE֨EΩEEΚEEE͞EEEEEEEEĕEEEEEEEEΚEEEEE````` ```` `` ``````````````````` ` ` `2`<`ܛ<`;``P`O`C`D``P`P````````E;E EEE``` ``` `` ``````````ʪ`ߪ`````ȫ`2`<`ܛ<``P`O`C`D`P`P`P```````h;h hhhhE9EE9$99h /h /:$::::11ǯ1ߧ11111111 11QQSQQQėQŗQƗQKQב)Q˖QQڔQ Q楕Q—R̖ȏȏȏȏȏiiQ݂Q݂Q Q)QQQ$$Q˝˝Xo     K      V          T     !   R     D                                              G        E       ^     [     3    Y p        W         U          S            Q                                                         F              P      (     4    ~ 4 Z q     M    ?? "?0ap["~ei_core_intrinsicsderive_eq_internals0׷{/Ț"L\e,}Vderive_clone_copy_internals0,| SR? fmt_internalsfmt_arguments_from_str0m~ii?06t; ]panic_internals02 ?0u?0ï4 \H)O_0S. T?0閩 e, 0^ŕ!^4?05EMBYi$crate::format_argsL]]0-?~,0 wgswriteln˗,ȱ]024_  ?06\ $crate::panic::unreachable_2021LpR`[Z0,]Q?0.a aKm 0/{<ɿr<l0 .5-M$crate::assertɊT]]generic_assert_internals0H2 0s=ha  ?0ryrP:xLx0׃V?0 ِ=?0:.Ee D_?0n/& |?06\rFQ^ g+?0bޠe, 0ס#Q?0><<o$"*u  06EXW1>?0P|Ke,  0/ȩe5,  0SMoV t?0ל p$& e 0z%L܉0ooߏwq, gg0S"L~ J?0דn}< ?07穇6Z$crate::panic::panic_2021"`[Z0ׇ֨?w 0Et §?0ל|cT~9 0\:yh ?0ך/wD4 0CNq,% gg0ג0|tq,  gg0L;ڲkN)uO)0?kFEc $?0רq,l)L gg08u0E$@`?0&A?X~* ڡ?07L|Jr< 0sl;K2@ 0 LKUt ?0׊[Rg  ?0ע^#w] |&?0 \&ux $?0ymZhe, 0D!taz$crate::format_args_nlDݲ!]]0㑉m ?0׀Nw  6?0hb1'4& 0_yxiNJq, gg0׈]qK ?0!"QEqH,^gg0k׬dS6?0^c ~ 0۷5 $ ?0 0!&~ J0Ԍ}XOe-, )0׆1@\o$ʼnJ   0+ 0^|yw8k ?02C5@ "?0 Z#n?0nBJ "`0q]|Q e, -0> zs p?0#byBF4c 0׭` Do045q,"lgg0wwJq,&gg0BbA"?0שUe5!=0BW6=e, %40Bgrx $ܻ?0~S])?0/|Y6B&0bML2)0RөMe,G )0J&0׿bڷ?0כּIQP $?0Do,1|m ?0ג\} M?0 <<]#0טOgv 2?02N[7}k[Z0p•qݲ 0וeD 0;VGr ?03v/q,gg0הgo kY?0cFmxes,y0ϐYm~G?0Tmb y0sse,ó0~R`~@0 J 7?0_L@0Te,! }06j!QL ?09 # 0B5+L)08c ?0>W;  e?0׏+FM "?0a~=v$u=ii0X4kq, gg0cJz?0ne&ЫT?023r|_a6[ݲ!]v0ז|f o?0'quo"4M0˸1[Z"[Z0\/3g]D0הѩEl ?0.jP x?0X~ zϑxq,a$,gg0י e, `!0P<)0Yc7kr $?0c}qcq,2)gg0׵,;Z!0״ wLq,E b gg06,;b?0׭ڠK ?0D*b ?0wX+h?0J"q(ؾ)[!gg0g-%f ?0מg.e,$0٦mV assert_ne{ ]0w5Tb}q,͎#gg0!q ?0|VmP ?0׹R[Bo]g!0Z- ?0׌S->} L?0z" *dj%0㤜'N\]("0ѢH> ?0Ti6Sq,(($gg0, $YĞLE0.y0y%0_X<8 ?0HPZo$)  0ז%%?0=q3p h?0usÔ ?0ת5 1q,!l&gg0"uO DՂ?0y.U "?0Ě؈>7 "`'08ۑ j'0ל㏜`Uc7~ )0d^FkQm?0ת}O@ ?0׿qdjk~_*0L~hj|ԍWݲ!] 0ʲF?~=*0w02N&%0ם7 \Y]&0דlË)?0׃gqw N|:0גT<ǽ[r< %0I308 3p ,o?0Y3??0׿\!~e,470 ?0 XIڛq:, 5gg0|-CՌ ?0@  :Ԣ5+[Z0|R r60u q,o6gg0שe4LB1 R?0 d-Ix:~?0k1~ B970Qq,57gg0- ?0&(70ר =e,90)e^# ?0׮.^ ?0׍j= 80+Fo$)u0  0"ta?0׀+Ɠ?0`$E V90vp2 e()P;0ʣ:PHLȂ !80׹;xw D̀?0K"}Rݲ!]+0躞ȘH ]:0_DX$T-_0@?o$ q2  0Ęt:N*0קooWy?0 8p"`;0 TaJ i?0לZIi{5 ;0״jۄb^LK:0בx Te,;=0ן5A  ?0רoFN ?0BPc/ L 0=0s P 9 ?0~?0ePq,0?gg0ӭ ?0&%wRL >0=Ӂ6~?0ׅBv~]sq,?gg0ע3|֤r<D<0{Վ ?0^ 2qg,@gg0c!q,<-@gg0Tm%81.B0C {"6[Z0,\SAW0N$0}o,iQ ?0SP|q,a&Bgg0D4Te, BD0-/ͯ<B0כ GTQ μ?0׷ֱ@B  ?0GW%?0׷6~ P?0לv9Q&C0׋3q ?0CTEZN~)*F0פuXJE2q, Dgg0Tg%  D?0}d+J]!B08$e,,F0~\E`q, Dgg0׭P~7*=?0N;v~y+eJ,FG0d!Q%ܖ$B0DP{v10o$=  0Z` &F0-5 TF0_(C?0b#5~xH0״3|7)=,F0׾% f ?0s 2*0.$crate::thread::local_impl::thread_local_innerhQaethread_local_internals gf0FCb;Zq,sGgg0ף/I(( H0׌z Ylu_A 0P#)nH0?u ?0ә  D?0!$ BMN?0Az ?0xo<-?I02 g?0 W^e,"*^K0קNrSkQ I0w) |?0"x* 0  ?0Ad"`J0Zq39$։-ii0טѵ!U "?0+=p* ?0׸BJM+ [K0v+e,SM0a| r?0R`q,܃"Kgg0׮q,Kgg0י̊ "@?0#+*IL0)cQN7L0oQ̶"e,%N0!{Jxq, Lgg0׹#N* ?0JUKE)M0Mǵ ?0S(Se,l)O0*N4te,> O03觠9?0y܉Nj N0ǩuȾ)N0/O?Hq, Ngg0O L*&[ ?0~=`.Bݲ!]@0׉=G#I?0]Hq, Ogg0B $Nǒ?0c$E]M0RZ|X ?0Sq,pPgg0t\PX  Z?0ךmX\71* 0q_  ?0{1y; ?0;޺-Asq,͆lQgg0״P'V6q,Qgg0Mfn3H' ?0eew. ?R0Wwq 뭵̓~)]06#h ?0i\ |q7"?0#N6 t?0Aso$fJ  0tYdpđo$ K  0)JpHA; $?0׸+]/?0E;aI S0p+e,U0ҷeՍ[]?0)5% T0Xm/K  ?0@hd_0eJ[Z0a0[ $?0U29W4?0O׎~4N 0n2P"`V0޼yhVR0_>Hq,Ugg0xٞKz  ?0k!40 ?0c63Ļ8l&)V0k% { ?0·Mxt{$2)9ii0_~ ٙ?0ֹfLF U0ׇН ?0.mZ&)W0 "a'* |?08bse,K Y0׻E'< ?0c.?K ?0w-:]LV0k[= "`wY0̕& T0ל{ ڪ?03&;e,[0]x$m ?0װsx 5_  ?0: ɛĠ\qr,p Zgg0xq,YZgg05W~pZ0?"kq,Zgg0R{WKh ?0׭:7'?0-*$)p>ii0Fz2셔 [0׵ ?0{Vƺ ߘ?0RxBF G J\09[F ̫?0AkO@U $?0Q.N+:?0׊'Dl~_0ץb>"  G?0:wY]Z0׿΢In  ?04Zݲ!]hO0!$M[ $?0h+l4?0S-q,8"3^gg0ׯ Ԏnc9[0[ c݋K]W\0KA2_ ?04{>?0Lne,Z$.a0׾(u6@q, o_gg0#$zd  ?0i-8#o$2)W  0:3-3a0;xFW5 ?0!ԣIue,E fb0OgOR`0_  ?0rT7e(Ѿ)c0bk! ra0>?$V~ ͡?0P64/ ?0mFGz$?0מg Ԟ?0ד<+F  l#?0%X;F b0==^?0׆yxx b0&Yo]q,bgg0q, igg0ץ3WB ؁?0\CjT)/0׸ЬAq,% Ejgg0p|J Tn,6j0sD@Hi^ ?0 Kwq()jgg0A G%~ qm0ס<~! v?0}Z>呥?0sB?M@  ?06'QU e,F m0#e<, n0ophv,Ee, |0&ii|Ix08#*F  ?0׽Y8 $?0s<~}0΍b -?0׍qN\q[Z03sЯAܛKu 0Wc^d\)ɎT]Wx0~58x&ơqy,|gg0׺E.,x  ?0עCT[8"a}0GjY~_d!\?01\9\ $ ?0m@@u"`U~0MV1 ?0̘1jη L?0_лQ4+ ?0XI ?0O5\?0'{] ?0ױk ~ F0|  ,`?0זIr<d{0׎2 ?0ׅ!,q,gg0g0+~ D0ל /בq5, (gg0ׇso$ߚAx  0םfٹ| Ę"?0ױ%  S]?0Q+Vq,gg0c;㍗ Ԩ?0׵x?T dq,"Xgg0#t%T7G0p)qm5D^?0L5ie,Z&0QOT;q, .gg0ij $r?0ӕ}e!Ƃ0'`wV8?0T\"Ll)0C,fd ?0h&c?,̩ 0˷'q()gg0i2ee, 0ɫn $?0ogt_Hݲ!]u0 ,W l7?08+ 0 Kzse, 0"3Qɛg&#0ξ>ouBqI,&gg0 ?3yߋ  ?0\*6 $?0GS5dإ 0DH{/1e,އ0ט-E <[?0$-bXZƲF|0וزsqx,gg0פ_] ǹ  ?0! lQ?0ׅ;,M0NCD8h D ?0 nUV2} :m[Z0.J!tC?0/k%5J ~҉0g`v(q,τegg0>GcnX']?0MèG%9?03Ro"-06JyƠCe,-$0{&ΥG D?0lKϡd 0׌F9o\+F"`b0TzPe,0~3K ?0bBRv_ ?0-*YA"o$ {  0.rz  ?0הzMCE),0ם1Pokb[?0׎2):c0oO_w?0ׯ(m1z:XrGec~E 0;R3?0׻_YB?0#971?0חi\we,H 0׊8ݲ!]0ע?j{ ?0T& 2?0'Ba}>[Z0׮"D׷ ?0ךrf?0Obp2^?0[Wo$!z  0װw0)?0M"G ~q]0׽3[e,;0 m4NtpN08@Xo$  0׈;Bb?0[>q,(gg0ך ! Ȉ ?0g&d~)ʞ08a-?0Dk s?0>hH%/?0ׇE/ c0=l{^ ?0o$   0qe⒱ ?0 6'r} $?0epK?0K\:e,2)0$28q,@ &gg0ׯ(?a:=?0טi@E0z1q&Ңl ?0;}zF?0kHAT/?0\;cJnA?0u@Y@f Dݬ ?0ד  Ţ0"e,0@ V?0SnW'[Z0H H?0ZFĎʣ0+t]lUu~[ 0׭B7K+ ?0Bmq,@*gg0|N") ?0׈n/( ?0ם7Y6YH0F )z0 ")L 0<Ty94?0׾↬k `?0!+O:$&0t$ _?0C@ς+%q(ؾ)gg0F ?0U+ D?0׎~ ` p?0 g^e,&ר0׮tB[Q}o$   0<ԴN&0׆l:Lo yT?0nsB Uݲ!]}0BCܣ6u ?0ׯstÏ+T0ױ˲.1e,L08r5xo$=l  049!q90)G3 u]0&"`.q,a#7gg0%!lN|C0~|t  D ?0׽ǵF5e, 0+F $?0kdk?0'.L)0דF`Ee,% 0KXmS(??0I^0Pe()0,\+5 ?0{$$_lj)o$ :  0Ԉcf?09vjG ?0O+zE %053i$ml ?0ٴ쮿ydxy0VrjVq;, ugg0Ip'6 ߬0Hi,e,׮0װT}# ?0;Bx ?0ט<8 ,?03ʌL,0B.ªQ$0`Х  ?0צlŠq,a'.gg0g9YRvYďp0דׇ'/e,08\Rz4*( 0Pv9^"R~)]0$$96u?0op?0׸d ,ү0aJ:P% ?0׭B `T2 ?0.[_o$   08p D?0ʥTTY Vݲ!]U0גe,?Ʋ0yr< 0 ߃V`]mN0׹yFqQݲ!]20 #en ?0:NJNe,Գ0ý/6ki?0sBKb'D p0Ѻ'q,;xgg0{)0t\ Ԉ ?0י*S.ey,0׼4"?0׹dh* ?0%ϾqW ?0`{\0RՔ5&[Z0׳Ł›[ D?0sϺ r ?0Ħ.Qr ?0גＺȾ)k07ryA?0_nq, gg0*Q ?0ץE&  ޛ?0ל:QY?Ļq,/gg0=0Ta  ׯ?06!Nq,"*gg0 Զ  ?0gw@'QKo$   0ש!ҳ ?09jno5$ >  0&YYz?0"-3f1߃, 0צ)OaOe,ؚ0cZ9r<G0p!wʶ?0>f7Z ?0?+ "`U0<:L0$Y8}+  ϑ?0c,p%De, U0#_4F)09≿ ?0',~l)O0.119Ϛ2f0F|-OW?0*^̎e()0_Ub T?0טXYfT6q, gg0N'hCF ;?0׌FnFݲ!]'0 >0Oƙ& 0׏R< Mq, gg0ȓ|60I38  ?0(NeI,Z0iM SG ķ?0F!֚S  0רYL0z߇; Y?0;0]  ?0v?B!'V04mڑwex,0Es7B ?0N_ _M?0׿B cY  ?0)ve`,0׬lG?2?0ׯ+%MJq, -gg0H,\4 0OQVBO 3?0ׁl<*(0 "[Z0bԸc ?0Pt(+q, %@gg0_K>4{ ?0$$s4e, 0^)5p) 07Fia'iT2)0x}xs o0חHJ ə?0ׂ%BQq,G gg0v30 Ș?0K:I ?0{S_; ޒ?0׳K$L)0g'Do ?01ۡձ& ǟ?0BbU>ݲ!]0-CEpu "`0ס bL̮e,8V0דҍR ?0[ȥ9 e,0 18_^?0 \+  ?0׈+i)FCq,ʳbgg0>}L*05.n?0Ӓzc7xN $?0t̝ X0yw T)0al?B4 -0:mo|7?0׻[ ́?0o/!0|N?0u#lL:?0Hur_e), d0*aX1 lg?0r2ĉ Lf0skK?0׀2iwq, =gg0f :T\?0^SGe\,0]tj_ ԰ ?0׆n/?0׶xWFp05KY  1?0]BȂ\Ǹ"`@0HYgx70@ey)]0(XQ"= ?0לDBa?0&:l.0VM.iV&0dze,Z%04[9k;q, =gg0נnbӭ ?0>83xH ?0+}?,0~fr<2)0מz.~r  ?0ײ¬bNo0&8ؤX\e,O0``! ?0ת>NfB_Q&0[yoCRe,(0T%r9YM0 {8 ox?0\a11 $?0ŰT{G?0׶։Sr<)0a%)kq ?0}lgF) ?0+^ ?0BOcQD @ 0\whe,<0#_D\>9?0::++?0LmmT'e,P0%g. (L !0PCHn \?0׭2Zj  ?0`U[!cXݲ!]0H8y"a”?05;L&0=I oL=]0ץLmq$60 r0Pq ?0ן<١7b]0לQ",0ס)wN|0׋TTe, 0ɗZ ?09 ?0̵\(ye,@ 0!  ?0X_J {L 0NU+Ҽ {?0t$ ?0NT/\ "?0~~Lљ$]m0סGL0א.s7 ?0ZЃ~#ݼ 0*6  ?0׼&|>el,,0۝o?0~9O0׹HڴSr<[a0׹(c Tr,0r ͬ?0YA>Qo$@  0߱FS(W*?0pw %j0חS'e] ?0uM\Y!0׸.݁*0:]U3 $?0?VC^T)0liF  ?0;Gji6~ 0׋_lL!j?0FjOW&0D0 ?0pʗe(Ѿ)F01OSa7?0ׁOs})L F0b7\> f?0׀g=J^e,R0׷/:ae, U0׃eKҫ  ?0ו+dߞ70ףQMTݲ!]0;~G k0ND~"ZL0{T;o-+?0ש q{ e,=0Pxgl:0@Z  ?0%![ j]0OZa 8Ӄ"0ײoeP ̆?0SUD9e,Z#0זd"=N| 0N'  ?09[)W^u$)ii0FS(aOq, gg0/#Ƒ ?0h=Ye, 00E̾R ʝ?0װQvK ?0Pm:N0˹`e;, 0 |nD@ ?0 "`0srTF6p]?0=> a#  ?0z7,a?0?*~> ?07[~0ֈ.XT]S0׷zt?09 F6=$0f'Zre,Z'0ePPq,Zgg0L1lE $?0 (H^f,0\mu? 0pr<)+0t8;*z0q³L @?0!-ʜL?0hfo= 0/Цmq()gg0HuĻEre, 90׈,3N6?0`Tݲ!]p0׿a;m,K 0~WC5-0ׂ>$q, Hgg0,:N|T0ϓ@2  ?0QT>i?0a|j. :02Zn ?0Cϯ 5`do$;E  0ׄ|T> [t0|kȘM>?0eo_w a?0~Ygil)?0ס`{J; ?0ل͕ ?0B)ҖNt0~Y ?08'1f  ?0;>BFpe, ;03Zb 0ז;. ?0՞Rs0r,ۑq()gg0KOC Fe, D0V0Oj, ,m?0bF5n?0נ:HT 0,碋QS9Y2EO ?LU  _ANGQKW C;Mn=I  :0׬ːlW, debug_assertӭ'͜S]0Ee`}kE0fs}>0x6U Ք5]0jhblydebug_assert_ne'b]0(9g W't0](۲Ԣ5o0B>9'0l`q9 "02?Y, 001yXš4&G0oG͹wGar0׳t,YKX; "0S1 &0B_hp "e0ٖZ1b0lh5NJ΃&0Tse\0QI z"R0ן_XXtRŚ]0-b" Z"0"`ACZ'0VCPvߌva0MՆ c.'/&1,&&&@!&*&%&/5&*& &/'4&)&#&3&.&a(&#& 3&'-&'&)"&22&F,&&&T!&O1&g+&&& &j0&*&%&C5&/&)&)$&A4&.&(&[#&\3&-&'&"&2&,&3&j(&&&^!&.'W5&/&)&`4&.&(&s#&P'&!&1&+&&&!&#&$3&J-&'&="&K2&&&r!&v1&+&$&& &/'D%&z5&/&y/'x4&.&3&-&/(&"&2&,&d'&!&1&+&&&"!&1&+&%&J &"0&B*&$&4&C/&g)&#& 4&N.&(&C/'Y-&'&G"&5&,&)0'/+&%&S &20&[*&$&5&]/&z)&#&,&'&!&1&+&W&&4&.&)&#&3&-&W(&"&2&-&x'&"&2&',&&&6!&01&M+&%&l &P0&}*&q/&-&'&y"&2&,&2'& &/& *&$&4&/&)&#&3& .&2/'3&/',2&&&J!&E1&!&1&+&k&&!&0&*&%& &$&4&/&&)&#&3&!.&=$&V4&.&(&n#&k3& (&"&2&: &0&.*&$&4&*/&D)&#&3&5.&&& &0&*&!&$&b)&#&Z2&j,& '&+&8&& &0&*&4&.&(&#&3&-&>(&"&2&,&&!&%&1&+&M&&%&a &F0&n*&$& 5&g/&)& $&(4&w.&(&A#&x2&,&('&!&3&-&'&$"&'2&;,&%&{ &`0&*& %&95&v/&)&$&0&*&%& &0&*&$&4& /&!)&#&3&K,&&&Y!&S1&(&"&2&,&F'&!&0&*&%&5 &0&*&$&4& /&5)&#&3&+.&t(&#&3&6-&'& &0&*&0%&f5&/&)&Q$&i4&.& (&"&2&,&Z'&!&$&4&4/&S)&#& /'N%&/'`$&.&(&#&3&-&4(&"&2&,&i'&,&&&0' +&%&.'F*&$&5&k)&.'X.&(&(#&.''&V"&d2&t/'!&1&+&B&& &0&*&h%&5&/&)&z$&3&-&M(&"&s'& "&p5'$0'&1&>+&M/'A0&d*&$&5&R/'$&q.'(&7#&>3&w-&~%& &/&*&)&#&#&2&2&,,&$&24&+&f&& &0&&%&R5&/&)&8$&F4&.&f3&-&'&"&2&,&<'&!&1&+&p&& !&0&+ & 0&*&$&4&/&+)&#&3&&.&o(&#&3&1-&'&3"&<2&U,&&&]1&&& &x0&*&)&G$&-&(&"&,&!&1&+&&&%&? &N)&#&3&?.&(&w!&q1&+&)&& &0&*&I%&~5&/&)&.&(&4&M/&.'4&S.&/'/3&^-&'&L"&_2&/&)&p$&4&.&(&#&3&C(&"&,&.'1&4+&%&W &70&s*&$&)&$&4&h.&(&2#&43&m-&'&j"&i2&,&'/'1&+&R&& &0&*&4&.&3&-&R(&#&2&-&}'&"&2&6,&&&;!&51&X+&%&q &U0&*&%&*5&l/&)&!&1&+&/&)&$&2&)*&4&%/&?)&3&0.&~(&3&;-&'&A2&&&h!&l1&&& &0&:%&p5&/&V$&s4&.&}#&{3&-&"&2&,&!&1&+&+&%&'&!&l.'+&.&& &0&*&Y%&0'/&)&f$&4&.&-0'#&3&-&.'"&2&,&n'&c/'1&,&&& 0'1&*+&%&.''0&P*&$&}/'R/&p)&#&80'].&(&c-&'&["&/'z,&'&!&1&0'G&& &0&.'m%&5&/&/'$&4&.&g.'#&3&-&/'"&2& -&/'"& 2&,&.'+!&2&'&"2&&&E!&:1&%&v &45&s0&!%&M5&/&3$&K4&.&d#&-&"&,&!&1&-&"&,&U'&1&&&1&+&/'_,&&&$&5&H/&4&##&'&Q"&/'t,&'&+&=&& &0&*&c%&5&/&)&u$&3&-&H(&"&v.'-&/'"&2&&&!1&9+&\ &<0&_*&$&5&b/&)&#4&m.&(&93&r-&'&n2&'&!&1&\&& &0&y%&5&/&$&4&.&#&3&.&$&4&.')&#&3&.&\(& #&.'!-&'&."&W/'@,&&&O!&J1&/'%& &n0&.'%&H5&/&,/'-$&<4&.&/'U#&W3&-&.'"&2&,&h/'!&1&+&0'!&0&*&.'% & &*&\5&)&B$&[4&(&-&(&"&3*&4&I)&3&:.&+&N &=*&4&>/&4&-0&$& 5&X/&#&4&c.&-#&h-&e"&,&!&1& &.'s%&5&$&"&",&1!&H+&K0&x*&%5&)&|.&(&H3&'&/'.&e(&,-&/'62&P,&X1&/&)&.'P4&.&(&i#&]/'-&(&"&@-&'&8"&F2&Z,&&&m!&g1&v+&&& &0&5%&u5&*(&"&2&,&/'!&1&+&=0'!&1& +&/'D &0&8*&/'4&9/&])&.'3&D.&(&7/')3&T-&'&/'T2&e,&'&.'{1&+&3&&n/'0&*&T%&0'/&+1&C+&%&i*&$&)&r.&(&<#&C3&|-&'&o"&s2&#'&!&1&+& &%& &/&*&4&.& )&#&.&K#&-&~"&,&!&+& &/'%&|.'/&_#&a3&-&'&"&2&,&A'&!&1&+&u&&!&0&*&%&0 &0&*&$&4&/&0)&#&3&&&c!&b1&q+& && &0&*&+%&a5&/&)&L$&d4&.&(&x#&p3&-&(&"&2&,&3&E-&2&,&.'U*&2&&&/'R+&G/'f &)&$&-4&.&(&F#&M3&-&'&t"&}2&,&-'&!&1&+&a&& &0&*&%& &.&b+&%& &e0&*&%&>5&{/&)&$$&74&.&(&P#&R3&-&'&"&K'&1&z&&0&%&$*&4&:)&3&y(&20'{+&/' &0&*&?%&k5&/&)&[$&n4&.&(&/'u3&-&%(&.'2&,&_'&!&1&+&&&/' 1&+&I.&(&O-&'&B"&O2&o,&|!&1&+& &0&*&^%&5&/&)&k$&}4&.&(&3&-&9(&2&,&/',&&&%+&%&K*&$&u)&`"&.')&$&/'2&-&7'&!&1&.'&&C0'?1&]+&!/'Z0&l+&&&=/'}0&/'/&4&//&X)&5&&&$'&j&cQ&*4'e&*&H&\&H&}& ;&t&o& s&)'&x&Qi& '&&_&C&&&I0'{&L&n&e&'&&0&H&&&y~&<&&&4'*'_4'j&`!'&۠&a&['&–&rV&3 '&&M&'&2'B&^o&'1'\&}'&&R&&ƿ&Au&,'&&k&"'&ѡ&a&#'_&,'&o&J&'&K&Mf&'&C&o]&'0';5'S&\&X&"' O&'p&='' &g&&DT&} '&d&7-'&&&&{q&+('7&h&'&o&K&&I&%G&&ݼ&&&r&=)'&K&'A&&&&&1[&!& &K~&<&N&s&^*'&&3j& 'z& '&&_M&'R&Xy&*7&&&t'&ۛ&3'2'r&&&5'4'g&A&3'!'K3'&g2''1& '&bN&u5'&C&72'1'y&'8&*&&'ۜ&'&/&>&W&&ю&3'P&&JD&&&z&''%&ߦ&}f&'p&]&\&>& 1'0'/v&,'Ԣ& c&&&O&3'D&dE&&1'{&ͱ&''&Ч&g&@'& 'y&e&@&&&&$&m&'&&P&5'f&&{F&&z&|&:&&&Er&('&è&}h& 'K&'&&A&&&@&m&J$'&j&yd&Q&'&Y&G&&}&}&;&&i&~s&&ܩ&i&Q ' &&`&~'&&V& 'F& &L&'x&'&l& \&'X&x&kR&&۾&~&?=&&&j&!'P&&La&'&&V&h '&B&M&U&R&6C&H&&y&&=&%'&e&2'&?&\&'&&=S&'P&&I&'f&&>&&&a&R'&ŗ&W&''&P&D&&&z&V8&&&"p&&'&x&S&J&M'&u&,'q&tl&##'&&b&'}&˜&X&'&&O&'&S&0E&Gg&'{&?&&B&v&-'G&&m&#'&m&c&Y'Z&'c&R&&&?F&L&&e|&&:&F&R&q&('&d&>h& 'A&&GK&O'&I&@&W&A&w&/.'R&/&Ad&',&&TG&#& &;}&3&r&)']&9&i&'n&d&_&'u&&tU& '&V'&&A&&O&x&[6&B&&n&&3&q[&'&v<&ٴ&It&*'&&j&%!'&&`&$'&&V& '&ύ&M&)'&&B& &&y&7&A&߯&o&%'&m&e&'&&&&S&'&‰&3'&r&N1'& &u&&k&V"'&&W&Y'x&N&'&C&J&fz&w&'&&7]&C'g&6&ī&6l&"'&h&tb&B&$&KX&P' &/&NO&`'&&D&L&H&8&~& &p&k''Y&:&f&m'F&̝&]&*'&&rT& '&&'&&v&d-'&T&L&2&Z&Y&c'&'&&ݻ&$|&9&q&Y(''q&&^&'&M&J&'&&w@&& &w&&&1&+&'Q&s'&}&:&@&ϲ&r&j)'"&&h&u'3&&&P_&' &&J&'&*&OY& '&ʏ&P&'s&&E&&&{&9&&&Sq&''%&&g&o'@&8&^&'Y&?&T&b '&NJ&J&&H@&W&ط&_w&.'&\&am&#'&ۣ&c&'&T&Z&'&&P&'5&‡&F&&&Z&m&Q&'&[H&&&&=&&&*+'&&!'N&'6& '&&b&&&'&'*&'&X&'y&*' E&f{&69&&x&&o&g&'&&5^&^'&&eJ&'!&&r?&&&&-'&&Ķ&u&&&&&v&&&l&&ic&&Y& &,&&E&&&:&'&< ''&&#&&w&&m&x&d&I&&=&8&s&&[&CV& '&k&0M&'#&&&&sI&Y&C&B&v=&i&&t&Y+'7&;k&&"'\&#W& 'H&z&3N&g'&&tC&&&6z&7&&u&o&&'V&&f&'&&]&'8&M&{S&( '&3&%& Y&'O&P'E&{&f9&&&$q&''&&'&g&d^&4& 'J&&J&P&}&@&&|&&r&&h&&^&&T& 'G&&&t}&];&&&Os&*'k&&i&'&̟&_&N'"&ƕ&U&6 '&ٌ&L&'/o&%'&=&Me&'&&[&M')&H&R&'&b&H&&&~&=&:&W&F\&|'t&+'SW&&>&&u&|&8&b&'&&oZ&'J&c&N&'&&D&&&6{&8&N&I&p&'&C?&]&&cv&-'&$&l&[#'&&:c&)'7&ř&Y&3'ŭ&#' &'& &Q&&G&&D'&&!_&|'E&& U&d '|&&WL&'M& &XA&&&x&5&&x&)n&$'&&&&G&&&;&|&ӳ&*'&&aj& 'J&t&+`&'Q&5&f '&E&.'&ZB&&7&&I%'&L'&t&+'t&a&j'&}&R>&,&n&e,'6&?&&&.w&qP&'&5&&0&:$'&&Q&O4'&I&H&&m&~&M;&&&?s&)'&&vi&A '&&_&t&Ɍ&&{0'&q&n&=e&'&#&R&H&&˾&~&<&&ɴ&4'+'4'j&!'@&&O&'p&[''*&7g&&bT& ':&&T-'&&چ&ͻ&q&I('T&.h&'&&7K&&h&DG&&&0&#&r&Z)'0&L& 'A&;&"&ä&ƚ&a[&?&)&i~&0<&l&t&}*'&&Qj& '& '&&M&'p&}y&^7&&&' &&3'3'&2&&+5'4'&_&3'"'o3'&2''O&'8&N&5'ׄ&C&W2'1'&z&F8&F&;&'&3'&L&>&X&&&P'o&؅&mD& & &z&-''I&&f&7'&]&&3?&>1'0'Sv&,'&*c&&& P&4'c&E&&'2'{&&''&&g&_'0& '&&8@&&ȷ&&L&m&3'&&P&c'&&F&&&|&:&&&sr&-)'&&h& 'K&'=&ۂ&HA&&&h&n&u$'&&d&Q&'&&G&&&}&;&&&s&& &i& ':&+&`&'&%&3V&& 'q&5&L&'y&<'&&6\&='&&R&I&&&f=&*&G&+k&!'&>&{a&'!&&&W& '&j&#N&&z&dC&o& &y&&e& &'ޥ&f&Z'&m&\&'(&=&kS& '{&#&I&='&&B>&F&(&b&~'&&W&'#'&y&:D&&&z&8&>&ΰ&Mp&&'&&T&.J&y'&v&,'&l&K#' &Ģ&b&'&&X&'&&O&'4&~&TE&sg&'&?&&q&w&-'p&&,m&#'&&c&'9Z&'&}&&D&kF&u&D&|&R:&o&&r&('&&mh&4'l&&uK&v'&m& A&&o& x&S.'v&Z&id&<'U&&}G&L&8&d}&_&s&)'&h&Ai&'&&_&'&&U& '&'&ڃ&B&&x&x&6&k& &n&D&\&[&'&<&&tt&*'&&j&P!' &ˠ&a&K'&&V&# '3&&M&W'E&B&B&8&׹&y&7&l&&o&%'&&e&'K&/&&ؒ&-S&'@&&;3'V&&t1'&8&1u&&k&"'&&W&'&N&'@&D&t&z&&'&3&_]&o'&a&&dl&#'&&b&m&L&sX&s':&S&uO&'&C&D&z&o&&9&&9&q&''&_&g&'k&&%^&N'&ѓ&T& ' &&'&&v&-'&~&z&]&&Y&'&' &&&U|&9&q&(''&&^&='&u& K&?'&&@&&1&w&B&&]&Z&SQ&'&+}&;&d&&r&)'M&)& i&'^&T&_&'5&G&dU&T 'l&K&F'&&A&w&&tx&K6&&Ю&n&$'^&&d&t'![&l'&*R&I'O&&KH&w&e&;~&f<&>&(&9t&N*'[&s&#j& '&d&`&' M&'&Hy&&ϯ&\&&u&U,'&&&l&"'l&db&'&;X&&&N&&&D&ܺ&&{&8&&}p&&'Ϧ&f&]'&]&'&4T&m 'UJ&'&J&pY&#'&&;P&'&&E&&&|&9&6&&kq&('E&&h&'a&W&^&'x&_&T& '&&J&&g@&w&&ww&.' &}&m&#'9&& d&'&t&Z&'&ސ&P&%'V&&F&&&Z&&Q& '&sH&&&&=&&&I+''&1&!'d&'L& ' & &5b&&&'&'@&'&Y&'&@' E&{&V9&&&&&g&'&(&T^&~'$&/&|J&'@&&&?&G&&-'լ&Qm&&ˣ&c&'&&_Z&S&&aP&%&t&/F&j&:&&B&('u&&'&&l'&ٔ& '1&Ӌ&>'&x&U& '&&GL&'&p&JB&&?&x&6&2&<&o& %'&-&e&'u&&[&'&8&[R&x'& &cI&&&&&C&&7&1&-&%'e&"'q&>&^&+'&k&I&&>&ڶ&u&&&&2&v&7&&l&&c&'&Z&$&B&&E&<&:&'&R '=&&9&&w&&m&&1d&4I&&=&Y&s& &z&bV& '&&OM&'B& &&&I&y&b&b&=&&ӵ&t&x+'W&Zk&F"'{&CW& 'h&&RN&'&&C&&:&Vz&8&&&o&:&'v&&=f&' &˜&']&'W&m&S&G '&T&D&?Y&'O&o' F&{&9&&&Cq&''&'&0'&&^&I&, 'i&&J&o&&@&&|&&5r&&h&&_&&T& 'G&&ս&}&};&&ó&ns&&*'&̩&i& '&&`&n'A&&U&V '6&&L&'No&9%'&]&le&'&˛&[&m'H&h&R&'&&I&&3&~&/=&Y&w&^\&'t&+'tW&&>&&u&&X&b&'2&&Z&'i&&O&'&&D&<&8&V{&8&n&h&p&'ـ&b?&}&&v&'-'&D&l&{#'<&"&Yc&I'V&&Y&S'ڭ& $'&'&!&Q&&G&ͼ&e'#&&@_&'e&&,U& '&&wL&'l&@&yA& &&;x&5&&&Jn&$'&̑&&& H&&&<&&&*'&:&j&!'j&&K`&'q&K& '&[&M'&{B&&7&&a%'&d'&t&+'&a&'&&s>&M&&,'U&?&G&Ow&P&'&ODHT  W)Jj KZ USֽ@ }CU -Y`Q/}߮ 9'`@6Z- )4!˘ ' ~yj `jGNvj< _CGO|S3MHq o ބC `>1ʆ `v%8.:x.N&?o^{ժI`S _uwH*W?g? %)/ϝ`BI2:~"@ gORH^#?# Vtə&?$ [x@o)a(#+w*?5{uɟ!m, B(- +y0ꃗ-`Wrĸ.óvR.:V1@5}Пir<3|%}4[?8p==U }>k6@cA`Z+= ě6A3 @`M8D!REc˰n ?F]C_WGsG`ƹªIl JfF &G@3#zM qnN@# OWAREهЩH Fq’RQhPKTX89OH!{m#UL~^iM 1hD[B@,3`1XEGnY`+Z^[[J[WL]@P v]LY-GNb "rqH/cG BPbjjT(e@gԻc@mPgψBi\hwj,{.`j g l@L $m@;5Xn{2j@ST4Cm`5'q``gpm j$s 1 "gEqzRuՋ}]ڭz@{b"k{kfK}U a~&KQ D|  oؓѻ  2t@O2G7`&q.҃`JuKR`Zar@ɰΉyS(.ab F**` &@ 5@QoU@$p ?RƂ{]8q2 $ đ8@KDž d2 [  ThvoΊ z soQb}@4v9]E N* |U. 2tc1`C шd(ƣ Z`As19EkFZ Yr`FV/b[漡` sM?Y{%@>tJ6j g{ۦ؛-Hoa@~U9@OO7H6xX,yW` C<`! haz`8^U<'W6 G}SHp*zK+ d;%|"I,p}'N@Zq 8O η[ ӫ \ x_`LEζ MN5$m P$&uӼ@)Z@g7 l\wvZC FOS<i? ?LƠWnza|`@@yN$la.@]! ee0ɀeif``]:m@ P EK9o}# _؀U :[ؠ ?` ~Nw W[a. ܀ 6ٔ|@!qtyI~ a, ܠC"H!\ oHܑ`),ӖO OY1qb &q`u {@zR k“` .aI\,3,1qG#N_0q^9kvK7aubanƎ nm< C a[  c =O a?bR[!dLSa Ac/%uA0iO] g4! a+ B5F^QOM!\A ȱ Ys3T9)8bmE*Og D' !!-8\!!^"+&DЮ#ZE"% ;M%aEY:3&"dzv$F"m )xAF~+h1~,AMO[.1838/!Mf0KQ&.?k73A> f4!u$$4ސ1ga7yG7|Z9A[q :a(k V92v <=kR_(Z>>*z]AiA1`S!D!e3& REE DAiGY?FaI0&JA _YFaEEA5SGDCaQN!\0 n4HZcSsA TK/ bƸUVr$V}Ua gVaP)Ta6+7u]Saj\4;n ]!Ͱ`!VPa9ňca^- dA8G|eOda_@ ;(fmc'h᝵,gXl&k!J "Hl.kwƑnTW\Ȟ$LBvpB ' (9Y qVs Qⱸ\IB \^ݗfKWA<%Lͮ">RJz7 rS"U:ஶ0;T(ܶ"?k"guʢ}RuI,3󹂐S1Ǯwq "褼u e"m!  _IQ&gHB/srp; B݂1WBkJ.pEbF79c"֕(Z"b@(,S["pQ ^M@3΢-_/Ѣo6hb{nuu91T2Hbw UbHgڂq6" `nєDނ J⣃ 9jbv"# jBS͎"(8B5i73/9,tu[)r,B}:6~WtL #cm.3cԬ,cc,^,_C;WN n Plr [ %<DoNBi#qU 9kfcGʋdC Wm 5'C G|CļRRD c!'_ DA$OOP8%zr~[) gw 8-"gMTP#C,h#2_F %)x?%'gV&cU ( )(Mq z*. GU+C!,-; yk,#5C-DC.#VaB;1cK] [u2e 3C0A3 F^g5# $6Ë<4.c.H+U1Eu4CߨR_;#Ⲃy&K6=hnt7@itH ϏAcheƧBÖL'}UCcnDcUp j.F0]'Gt (vG#\KmOICP ynIftW[7GPYO;gX>C@Ye6CsMc#^ySŤڲ^JK K_^Ct`js_fAbEGFĒ6bc-0 rb% }#Y`D@b}gCx>c|(?b×y,bic>j:vwkC6+o3lm$jkC% ixBvl_rj5ppQw|ZqCHXFr#uDjqFU!&i:g#LEdCtC}6,dwce qwe|F{c{t{L,|<M#i/O¢ Kr=c"U N3;f#DgLh@("wՃ#/:c< [=g~Dy ~ꃃiTo Yë^([Cgc]Jf"`mUCvD\ިYNɌLFV=5?)  cW`ù\1͚ҟ%Q xNZc&T5uC1>8۞0j.9ɧ"9wg&*0p ˢ4L&dݧccֿ|4C cueة77Jc!'c^#Wӥn`c+䁯I+ ❮ as8?zDz#OdHh C|N_!j3c>)uɹCs-pd4s'3֥-) |?|rf/:n_e5促և !=|6 o*4L2Ńů9 W!*$v#g(V5ȃLdɣQʣֆSs˃M7ö;Vͣ7#睈6] e;aBA΃Z?]JуeZz gJ T;(6#yP >գ']DPc7 ã Q k|KԼ+fCbۃ,4R@RocQI.%~]i ڐPC&y#E ߩcbo@c"¯y)-WwMd@co&u 9c?z!CrB&Pf$ m#0 C53Mcÿ /j+c5+/`)kn nB c;:4$#V;s5at nnT[SlclC# ^W=.$Y "lH}#!w@r;2\!e$"bExVm䨢,зJ # Ic{; # $v* {F6 DA >aQ UhI= Jy;*2v{$T Qkh2h ݄)(ND ʮCHA +iD3,2hS:ީ yՋ kdr\Ύ$rMYLOռ2ޤA$;wgE3 >gCu,,R$#ӱ(4^V %`ͭ ['dsk `']E ƣ)d. TQ'dySC: qu_1փ6 ;:Z l;xa T' <9`<:$>$$ ;?D|C@$㱊(=lǗAoy Vf=ds}GD&DHN o]I$n|Id'@^K }MɐDN$P>OD߰D`}6=vi[1ݡ$Ui$D!.B=h>ě+b !Sy}ī{8Izt H)$<"CD;Y\Xc6 lNdY-O>GJ\Dw JAɱd=a d JI-a7ǹ$A&>UrCdbdE>϶,뙉2L%dnPd= Ǒ$ 1ѨZΧw>丼D9Si KD;D]l -$dwHmau ecde̊ĝ $J7ńz- J*BdNy 7Ȅh| $6b&vT3d| дߴ$<.J mg%Do@Lru$+Є5F+h҄ TriҤdBq$ԤO1\`d1פGZ]T<4D~v  Sq1y Á)RC==nΰ>R(B$A`AUcd:KyIĆx YЧ$"Z+ I  lRT\DEdIDhM#䥚X9 D$ y'd vs0焹jF$v3 FK?[ I9\D5yd>Dw8"d# ׋=?D(9h$sXF2$%6LUwT||7hT€2 HȨDU/d\upґApbuZ 6W,zעAwad; Ϙ \B Q ^/E V & لS% c  %7 /  NP=% Re[Z.e $Etgb| ]A:%GTEP%yNBe&'z$j%qxٵehB(ceeI8eS F^哘M=EJ?8tU%b" eB C!["9F" Z#%؈c e2 P$-, %WH}` eh~H^_Eb\)%]="Py@6RG+e+|EDD0?X};˓P G:d8}s5%sRECfCE է6/F]^%iD 1]oXX^%HgA4_%A`^"0O)\ED[R}cůc~ d+UHfE?]'.g)_r0BhrXhƥg嬅kPle ~h5 fVi4JGi^#qW L0rBrEJf#rt%MTEnyq '7aFgULtkk  mѿy > go*Ur%֪|K|%Ui.~E) |&bEL'd|~E/E+ WQ. 0neq./2 ŚiNTU ˪  aDU 4%f* T1Bŀ ޚocfpŠ6FE%$:i%N/%dW%<> ľ_~KSYz >eqaxvՕ~0E؈uŤi-Q~( ##E Җge J{eK5y㛥Z  R_Z1a*+;GAXӂ1蟥>rq򄅕lez=|śŨ%?Y!WeBaWkK ,)5V&yŅ|s I2E^?`٫ee(LqΡE#lOH䭥_Vp dEubFܙŀ,Fʥ{]'R+~>O۳E8 h=$e0v+"RӯX0yǷeZjuE+jC mw-l\ KTmm^ %}**#%sluzDbI^sM#fKSQ5;%\2 AϾ6vpU-A\M2ȅiɅnj ɥunyEف([R %%$D%?VL5Q` ԅFՅHeT c%/2# T*e `E 9B%sȎHA%m!Gڙ Tk6JɌku2NXXN7Ȉ%vy}3y]>5gGw 籜Ew=GP%i hI%U;Fe|6mEa:BM 1 CP̚ G(% 10%nT_谴QvэS Tzj 8EL-U6$M[J }kE1qW)-uyNc%q^?|y$ r%tOnŔ`'Wq%}=]uAx&<?.F'[n& n"21#@4UkuPre6U "% f_ QG  2"'-^ 'l FM#]fOQ Fwig F K&p#=Hp|F"&|>` FtKJ &uv^Iu ? C %fV&|wfpzW. F7 ~AdGHqf~pftֶk& m ^h~"@k3\ʇ&-c%%ay h^&f.J zP'X] >'F\pD)lxj*z]  RFghq(0UI:UF~4 JrWukcX7EKVY_Z5 -${[&,L ErRY?<*]f6sR[<c2`fns$bNq}c+cLx*qfaDz75gf_1Of&&8hY)rWkFyd$ l&Y'k4ںp4)r B0ZsݰJ3s&mLvH{;fBv$ 3L`g;9 :c/4fk!7 Wx O:6mSGj&j'-L 4Jy~fyzZD Dskm F0Q mv/P t=:JC10v`DpxMP[_D+F? 4&UvԢ& *F"=FdF0,  Seyq&O&~%J4DVFV p~9@Y0b$] )&d]FtƓ vk&x AfsnY뱷R=vkF; eDFH2ya\үYC|^dݼU! .4{Ӿ&V+{л &-nCf,EfU[J2fy4%DĦ> 9=?If8m}1Ʒ2FVΆ%+[φH+ I>f OPIZ PQ҆(2%?C2JՆD{j'Άie׆-2+V٦Cl %ObfrTNJ`pT7,Mؤ/`&e>&>Qd:c&kCJ{[|b9"v&z !6OLCWކ1hUٓ@F#&|G\'0D \؍(G[3 ~)Il"*bC.@ J/U`Y90-1.>(2cBZ0'L]a4'2 =B4g@z"U5/5l"3>'W_8g#W8.%8* ;e kOP">z ~@'",64@n& =V9CǕF4CPGȢEGYN1BFg AG'HYbLERgJG53lzRjKM䈭,EQ&R:9Qg: AhSG<F*TGEzdT8+xdUM W @lɀXZ گYG>UGF\1Xg2uT8P\g \GH}r^aq9.)^gۭ8?6_'4GShaT#b8 2v<]GRϯd֤te'˔;%c7v$k\Jg$hgG`uigJ(Ug9CjNJ'pPlRܹgX~T_dVogJ ngp@6 mj!poV յkǏK{is'=u?~qeea#>wgDNvyǵ/zG3?{.o }C$[ \G9Go'- '6QbN& i*X0ζ 4ds D͋gˌǿI> ^֍DKd'z 5gFA'BY2G8jCDy{ Y- B2䙇(G t0g k͜DV<]?8YHG' ,{^~e4' lۋgszKèq l3"+MerI q'MIF^ 9GE 2@h_ꢱuGOg!m LٝGM4ǿI[/l6 ]thaiEz' >*v:6KnRr &G c '-_ɇ'}^b'Zz(̇6L J_G6 GIrAϧS/G/$ЧO碲s^eҧc b|Xڒ c'u 7*d'v#L#/'ql +_@<~V()'to YE U+rߚߧ G ⇙ '<\ <oSQ's3懪;b,Ij義< Kgi> Nsgx9&g)}' S "o ]z>˸qwFu'-kbǁ GÙT6q{,GLbHG`2lb}* +'9w:{'-R &?X l H%1 ^Hng#b2$8Sw ҋܬۗP-gEa] 9yPcCr  HX-'I|G= hm= GJ-;OR^Fh7.KX|H h"#ʀ>hʉl!Muz XREgi(5 o(Җ# nLWbϗO\L!]6B$#H~c³-{$E gW%HD ./&ȗk%k(ɔO 1$hg%+ȳEB/H)H"0l Q W1H /2䑐ޠ2a4nJ4hk"Ǽ4H)nh 2Oy7VP z 8(\ u9K 8H>D 2(Xm8! qV?(J^/BBgO\:$Bdo8H۱(0DY aqqEhp@+EtD(5e GhkEH,mYI趟3;mJ k OKPRˑL~c M(`b k N(H.Ml5:;Q`!=R(q}6R(4<xcUsoWQqX$T.IYHm$g ]h8_9^H y_QZjj`* nS$ ah[Hx<` ]c=4؞drNc(ļ hYe lg(n@lXg1eKdw.8cjDM@ר;jhܚ l 'KmVlmɼooQ kNkd]s~AтduHtz4jihWdhݦ%`58S]GOyha6jc(zHvOA~ (.(S$n-˂H|Y C΃RNKgoͲ@t h\%:ש]@(Q ‰HubkAZ{_tŗay0,la_(F2Gpne В( Fߐac (j B rߕ=>c}׍hh[-;ȏYH #ǘi;Wٓ/nAWGƟޟ@÷(!P2W#AvGhYC|0$SQ>̥H20!P!\ HF15Yu"Wɬhio*G㰈EN%HF7@_/Z Z1ڽ^==N-_ZKwfm.p8) 8*V :RHa!V#@F3ԆDqγh.&È<~E|r!b <ňj v$ňwxH>HW2t`(mn5xzȌAbqĨ御HpH,T-0;fT3HU,HުӈWJQ/$p&f3!R&6@Fnh^MGtȻnISACvѢ׈]TC(jE<%a(: ΑZHR_0_(˱ wGh#Q}"I#']d(W `R<5h3 刺.v;;(9ghT{| P٪(3eQs`~ H}ת;[)`Jթ*ڈw pd|KX%h͆|rhflg& ȡ$rFȷN ᱾`#R Őp+ Iaţ( H*IE7I 8XꉓH6L)K.Iӈ["PR 2. P?] )P 0e -%AG/F Η3 )[ L )( ]fɼ m H4)fTpE, ipw+v3I F :YM XK} aɝ[I U5P )  OUDp PiIy d  i2*^<"Iot7%<,)F +N*ի;*E A2v,)L @)U |)*Y HIo> mo,HJ`Va$K n:\3L?KײMWNIOGdOhgGDCPi[ EPٌB޽RPA9R;T1RT֪WnDU)tDH<UɂczUTXI)&W>Y)V. DX) bj+UɅK:S\I;!`]o ߏ8^5VۦJ !Sg`vE{(oa0bε [c)K28GrHU7J`PGz))Sr14h t ji 'f!Lj ^b lMZrmuhFqnl DpTt18Lt. =j^vi_w<w?5t)&@u cALz _[%L{V0|6+|` m|@I1eiL+L#i'64 INvSFnR Qb )w)2z)Jݎ CPߍI罎BI"w ɢ@\i4J.$lGMm_CIpiwB E&̗5V )'tB蘩 Kcs ܯ))Q Iw1;|8IgZ5]Q)DrcySНIl{qɃ*ghV)Siߙ'N6 ڰ˪Ͷ J b" ݍ*A%SQ l. Ε?m.ݼ& -&T/aT1*TYRwg.*-I /*d~`4 a C=5* 4uTY7*VD8&H8> l.jq(;^0<>l M=7ߚNAJrU"_-CjˌC֠@w:E*F F}9-F:ʻ mBG s,YIj$~2oI*D0'Kj 9w([Lj( =M Ӂ߁M ;kO}[ݳPLV|~}KjU 2PJkTQWDOHT %|hbDwU*/#P<;‡TLʊr X0Y5 3&Y`?_ Y kLA`*؎ޑ8_bJ ̳c KjqAd*6۝e^B ԕfJ vgJCK/h cq:fP'KXc>Wœk*=ֿAl Vqnʮ  p}N/Y#q'rdsp;H3s b 6Ksm#@|tjr73x+yBSxŵ+Qzʄ$czꝑ Rz?ڊ}J^<w~ !~Ɩcte(J<i jńj,[L@W/_tQ ;.*/́o*~zA*E%Ja~"ƄJ' ɏޡg@ J#*\H*Pꫜ ΔQǴVnJ_[ Gq[Hۖ ߮/Y $0 3Zr혊*bؖ*HK z0oOzI 7^{Jyģ$5 W@ʲVیJ*+ =$ DM#v 5sSGAOE*\4 kqJM6qd7k9ֻ7D/ czj(A LNuk>5Z8 ݳ*W˝aPY w7^J4] 3ޘ*Koԅ I v܇#Һ>d5 +,H JHrW*#Bݣ _\5/B%WE&6ê:2Qj5B}J%ØUJ -\.fr&VEʡU ڦ PL&!j}lnpʈ;'nتλ7\@jG9MB؊CcUhJsU&Z0;i \c0 j=S^*a 0TpToj.^ i.몃2Jk8 :ʵ![yRU-LZ KjF6p׀":5|r'i $g ATȀ/ H q.KH\ c2cXZK\{:k40R7 K8Y+k3>Zfд= PKQoOf  34!qlKm +7n68'X%l(HN(k1Ժ*%W-C(of."bcz/8V& /kIT1ӌ[2ci-kx*Ib4k:T}'0KD 7+|:@B")ìAQZmkA9[SACK5p}C+ z ն@Cku1@&,wHп%vJ+*z]*K ڳoIyLѣrN˷x&]NKSJRPHQKJr ~@R :Sk?TNBS˟ A*Qkq< Wk^eWK[- WR+h<kKVT7у[ 7@|zcYK xH\k]QKXW+Clp9b/;f  f|g; +Xng j)SJj ږ1k*|1kkKE @mNwKmُ BMmkP l+IHx\@j~}X:gqGHܻsK+tq v+MLy˙o_az ?1V {tyq{ $f{ [qk x0Q Pֹ D] P|僌 ELN)"oG- .׎KxwHLkyK7Ea[ m,c^ (ZU 43m5( "sQI˾r\k;O;D \7F+ kh:* J2] 2 _~P. '4eL:-kN.ƨBGK[TF]֐ s)*'gG ߭K& Ǭkh V/5kNV G, 5I" ck]yJ]]Wj )k ;?Tk@mbmO-b0)hF~6k++ -Z+xi|[˷ qz# · b;k haDEhka nDdkEOXyjm)Nj` 1ȫQ ˿O0K /N sy9Në م 6;ngE]^+1 cجЫ*8  ,i< hQ c^dKn k 0Ջ2g+aa+!z3jً(.W,ګ<4CK!< a <2 /ܫe" \-kygK)V, Yrg+m!&ѩ8G  ʒˑپde0`K){˿LQ3ɲ8&{2-L}HLI = xc }b"J u #] n{"쳤F?y%Z.&lH)WQ& -&lݭ K)-= T %0*,ť޿v* 4[Әc^, } g- 3%IT%1]/!M.,08  3,a)%Wm6U)y k{i7̢c&=6 ,;tƒ<+OH= Ǐ)=쵅& |@> "T GK>l]2\-4Al+ RA G Cx RC,\E[gELߋn)kHl?Hi uJJfeG&H4V2JV5{M bN,Ok \jO V)"JM ^Ml*m cR,/ EHSLITuĊŸU -F _Vr+WLX 5#X!'VLwK Z XTly$V̩^]J?3A^̶$Aru_κ%_l*)I]f,DzvNf  IhLyqH[ iB ࠓg SO*\*h 8z(p6ll + Ͼk0F{m u&k/pL G(Wq:sqOՏp`WRq eDRh,ѵUbLvK>w ,Q\cw,}TdIxPo YvL!SyL&; $)| Drݗ|5g~~L^'O B*b~ G;l M?LشPn{~ 5ِR͎ ۞y&| i!W } q%ǰD?Mĉlm!ς.'e,ޚ7Ն׃U,rQe<{UFc, >?~̢ $gatL 8Т̗̓}'RL} Ez.Uoˀ O " l/,X*_eilǔ"؜lw>4,5fȞ9B Gܠ|sHJ-'읬ՕQl̚퍦^RåbbuΩU6TbǧlX~ULL l2E} HkALI7nL 6Q^N'5pAuJjF?yCl% ͥC,dIf -ksC, }ȕ_ .*ͺˡ޻ $^s % fX, 2l0,q9X ˌj \ | .ь(|B, 2Rj,_8 j,OF:l;T2~جn\ iN kLP@I ٌ)D*'G ?[̮;S,~3g଎wFUS5+& 'Y.b մ-,q8ˌ̀d~ 笗/+(lPlDKxП팋Q1 |m%o# mxwt]LpeLFNLRn9t1# ApL(^ ilD-= J#c3^<0bx09klha T<GF@_h^͐ c#\ ӢM7s'!o' 3Km.A(-EKtJt m,O [ M~SnrSMK\m}$[o>m* f͍v -pM7ɉSb{Yz_1 NM˰VM)1eͿ G_Ot$5m]| a@<&mWז' ݬ򗗤+Ms*:J./Uq{1- Q4;ȭ)z5 `YG6'isK4hk 792; JM>?6 2?M&c@MA-e OgA-mFCm/ ZMEmi>H {blIyX gJm|*ȦLJa\D-L- :]H-rH-OM7C+UP{Q 8[f6BR yɋP=CKrR-q. PoV(d/WMs>X$JX- (;XTy [pf^ •[]UwZMNC\ͅs_-Ҙ>!cMA3cuMAem# CGfmλ11|g- رghm8ᮙix%i LN&ha䖪?l8xA(fem` ;Vfl-I2n w\p͈s|Iq}Zr@31q ls Wl͐k*v-e`(pmehPxO w%sMY.fwM{t^{-%|{ @́~-Am^7kF;'~88lmԴ 7--Iú_mm?M] o\r. c +%zmX} ^_Ǝʴ ~0m\;ͼeC3$\J @y  *+2A4ퟟcj&]OvV$ x ML 5Cm9g :pXnM&11- SI-eiQўM5}J@ %!{' Bџm453 Ȥ-4LTmj%r[X?5MCQl 2xdHOԩ-fW--ZsP:aErmkҬm9: 3,M]Y @lm7>̢a g-ԍiIn<ֿmruEƪ:m>Paŀ f0 XnmmjoɍU-G MWFgr j-j͍,7 ;V\ ݈}w]TT%2-Jx ST eGM|2M)E\ moZ$ͪ .mP=sZ `-F?nqv㊰aQ WO NN9"% qh*.T`$LO w1tαoO .p Iuy N<3 PqJ v.9 +#P nX  @[ XKw ~B''lXh n\ N,MGjp. \ɺN.~[>d% #.u4 d:NHt*.G#_ >$n͊x$8\b&nW0V\}'.e}J>|_(.Zao+).vSZS`,.Y] (-#|-!: U0g}d1n+, #1B] <3 W4=Z 4>60HD37,ϸAQ8~-[9.| {s:.@8*:.Mt ~t<F/ ] = KBU9X]?O$R?N~` ,bK8N'e6BnW e@p@MY/o?Wae9o8~ieFf Bϫ]b(I\%J4"3Kw OLJT= o MaK 6K.q{NF^gPf 5!AP]{HR|9ykSHh~sT2: gU.}YBM}MnXKirɚY.p -![%Ю\T|)a>]~˦^x8 _ L/_`uQ_)΂Xt]NGVdrJe FYfsJ [8(ys.~{_\k)ދvp(nIw3 ]x.?Iyt|Kx_?b{WpAk|e!}&~h| (HNZ~LWo!l㺂N<CrKe[_xοEs/'~.Cb>?zN ;Q/,"όNˤ2[i wV/~s-ӿUN;v bW->{GCȻNj d)b3p;. O 6TT .ȵW Nk *7Nۏ|q:O.(ʽn-NY=w+. !"^5NO&y$n!䇋N3" 5nnon3 \ZOnf ,vϮ1'QT!pѮQ MmΎO _nvx)U/t*ب֎O={MDzd|%{ -&W8KTe ێAyۮP[!N N,N Y!C؎,A,}@NUŻ΁j\4> |gY?Kְ.l$G=%kN/ nv3>_+뎑쮗3If^ja Hɹ.۲RFp<'eXn 9,Ѝη$@s'Ԏu`\:,]ѹSu1=n\mAb(B}6n2.Vk(a7 -fvF?j&ܘX*Az+kkaZ_/3oau{rOXjhCŁ .Y s呵 #|*6 UJ\zn eoJ  ?e2fȊ0/Acm0r9osט1{o]?-6Bt@O->'#k 쩐tE|'*iܗLu!O"xU"w5%P'*E#^(WCQd*V=*ww#s~G,ORY -B)ɫ*|wax/\u}m60/ *0/o˜42JZ52/<(h0-z-W0"10OnPJ7I7 58:9oƣ }p*9cZ2?Qh;/[= t< p = 0+>}x3?9ny;O0`}KE.e掎oŊy#U _oL*1/DULL! Ƭ/oYk5& YϘ CiZ!E|ϯ2|ċc܆og7 X@RKOH!Aӟy p3s//#C5:Op7r6xn#]v/t\+joFq h|!@|hSp$9Cdg úo@|)Q}/{gz1 ұt~d}F&Tm-8Ou|/;j6nw/i\#D  虋PN o:ǏEK7~XHn̯DG7O)t Ϗ*Poiw`֯77 !"oTIMK/Y$Fs%z Aݛ&o[Ylm֕/C^)*ӭrB#v߯qeB O>^.;|.ip~+M(obZvO W{X}Ot? $¾nwa,eOӵhv= &Yo`vO!J'8jpcRZpQ&O07JoY  r6K) $if 'ZJf*TpxE%v>e{|ou$@'S bIk>A 0kK.|-BRC T PA _c /u * PIJ-+(>p-Yp3+D0 wh3D0l]QPLnhVpwQ96Ksg;r`{pP{ld0*Tw\2kܺ `oP#Pnp((x7 !rep/9# my$T $[p%uWɨ B&p_“'P\ mɧ0 AF)H SLiAi"P`.08/@,#:tU,Pa,Я>/ 3)30X 5us6"zW=T60a$k!(*9[Ɔ:V >0#N\>Уm@HIBH>eEP6* F KpF`K oMO!(NJabN,x-3SPpK3n&QF\-Lw))TS3:T08 #U0f ,(VH.Y6_ qVZe51Yp\9\D-^V:qO^ ~`PձJZPBبbН \Wc C_ұcpu @a9e `Bfpo$FCDd0 T @gqn9jT&jts.l3f|}sm0pL暈m$>GopjcVspZѾqoQbpp*} $o#Wq;%1!up{v&SreZvPlvTQ7OHxt<^G4yf`{EvJq:p0mh ~𰅜%:! ^lp8B:tX„0Ђ(Kk4:3YrBWP "1u0}5/й{_P;SЌpp$Wjcҋ+3% x2 {x+СhzP#rHoVp}E K!+ʟE.YXӶGe0b8HIm 9,dR\JPʳ׃>SӜ=gix?^e ^HJ +hv'A2 kBgIB՜L0?pXRp\o{0wevSʂL ooZ0KOG(18 Qwg' j\qyPn ,up Q7, T c8q5\cZ#lě2gXsћ"r~d!:5EqрoQ=:ɠ/ ز o]w2fDq: _Q]l 8\i#3$H/ % /}%QZ L'1%$֧(})(4ճ)-]s}#1 V+3M-v,V.Ѥi@/Q~XV01VG IO9111Q0Qc# 63QE}M3q.5qhKt72e1h8qTIDe81sŜXv: Č9 1m<1b. 4u8E`=J{hZ1;(N1B L\J vC fE1rWs )cq6/(b" 1mqאь CKѲB$!:,QijO1dJ {1c {qaW=cQt?*q.|j&Ss8hQ$폥qU@qHdQ}ynH5Ԩzn y/u܋9Dx'_ az)sªھ$0q eq(q)HvC`xNqH&!!1&{% Х7tQ ;15l1v<%1&wQbquYǯ7Qx8Q`5Pk9h&r-RƱGP ,}DZ+ }7q` 8pɱ}˪9Lȑ0;D%NϮ<ͱC$M8ہVSbs<\qQ֋ Qe; o *hvmm x!*1! q=r,\1 )kQʂ5Q]®(q8p),QJ PڱcS AWѯ*V: q ZqINhQ:-̫$q4OlD* k2RD YBp ꑝ4E/OQN61_9uA d!+򱖢!T1{y~QPj$EvɎ11m 1 Ag8j rU'`*Aoh1S0TR2hʇ!ı7 ;$>d3 2I1t <AJ Ru5[ o FKU $ / ~ոү7o% J&"- dr'0XHu%r ̴rZ\r94%watBՠ hXhmlH^+D*^D Uy#\"drl  5R*'r O6~$) h?k#Ca@$rHުbl$0>^IMFeM|*ܹ/\|-.&] ^:-a /ҴB0Ru\n1rXTP/) }/2Bp 42幪3d d:46R; @T-#i$B7{N4+:R=L#B*A;1o$;2ZڡD\?ˍ>@ظFi@RTޙDr]ײ}D1 OiPRGgfK2esLrט>Mҝ N; a#O-on?=^Pj{cNrR˷-gRҲ8; %Qr TETp.TsWgA&UQ V~U XrU_{TS^hk[qʐZC|+Xx|$Ͱ[Rx7E [_m0ģ_ɭwa2 5\W2JWXrUВd>Q@cw drvf~‰OĽfr@&gi2 iV?i;'6l `cLmR9 ro[##pr{ڀp2n!Et[ޝvR:.Jwy Ix4Bz?'{2td-*|阄 =<2 ݁i<*; #.M;~&4KrREDo; c z P9 B 2a62E@JT|<iRO.U6Ĵd{bҏ:K'e:ǐtdd@$rJ : ?M?q<$8 ~R{ښڄ ѬҳG+p ?֩R0)֗Ö|㏛RI-ri]1zަ{LuxçR NR*plM[prYUT˫2[5å=! rc 㮒 a]*ҧe s,$rSap+byI[2 7_-nX _ARd\rF޵y8 ߹r 823Jor9 I+Ƚ>zF 8_qH8 j+dA{ 2l *ҝ2$A2ن y*2 |qDtZpm%W]sfɒ"l' ʲƓt uL3)X:˶)Z\)3YphŖ)SWME)*3`v-R {.3`N\/\Z/sh>s0B8a1z13:R!1ӸU?CIG2ӋR|Xw58z06fȸP7}Ka6s̻46hK,1:K[ ;3m? `5;;>~,a<3;I@CqeٝA$0M]BSW6zAn.>rD`VFBE9hEgbEo D6 ] HW;S*I"3lUIλ J ykJ:R A\;JS3N؝ϬKJ r1rI26JȡMs KScTKIS.sUEC1Jݲ7d~Bꭗ'WE6EiWĄHY(-Z3ܣ[S~yu:\S"uG[s۱M\3 %-_ V>Yt]a3&^byb ^cEHrccJUeĭQS"6̿UgD CJi\mis`|\lwm^y^ϒmJu+WnSͫgapSAsprdcpbrd O2sŻx5sSQ{pSFcV,w6g xDOvxs'%<z0`v{ SL]|' }szӺ8t-3SBT\_ӥVb|3Gk|aMLPc)[pH$i=H{$f:D~b3$c =ۑgSnR= /~13[Zۺd3OA]}a^?0(S᷐3!AP3- :~B]cKR$39xĢ#pU$ 70S}d[mÓsvG M8 FLkfsz 4~ S譚 adn^ ׻1 i@# >\ S2L8"[d¥L C`13y6{ndQ?HS%sjhlxȫP)SEX 3= pS$rsVikg82& S M%R#n-.3"S,ISډ\G>$1mSs*{iN뺳+ҽ?5e35^p+fsw*<ij`Ye|SW~*3.8ȓ{8 C ɓVBYʳyQ7SʳqzJ޴w9?2͓U4iS3 ɧs"r)_cD3L2b)ܶ3.Yt_( 􏒐(!qHDL T? !tY QT,"T~j@ ^>#t% ?~$TQ.Ґ%%6]&tefYHs(aQ{}|_)t&ڞ#t9)Mb+Ե8_,F rt:-t82-tv_w|/t5 )m@0>1Xړc%-D,734mq)ȓ45y6k։pd8QLy9 l؇C?T_eՁ@tOb`YA46ttB;p A(BԺDTÖq9CE DmF~6FkOvG4- CTG< uH?+:Htԙ+gCaK{MTO aIBt28OSQʨ P4N羨ADB7;BV@A SԟVNaqTtat$ T4Q&WVte}:EV8)bX[XYt3;ʨY4P U4& l\GJ]ֺ]-\6 Yy_V\ 6$1]psWNa^eCZ ` ]todT` eԴO e-gqduY 󋕁hҥfj}?AjFg* jS^ kF}k]YmTqYS;fG͝-^t ,zp8uSh a_YbrtO+iuU]4D8pjc^8rxːwTwB1xtkN`*J otYpem59}\ŕ~s^~K VMp~TEe hH9z"G, M8 &gP"`}益4d+Qu  ٍ݇ SԏPM7;4߻txf;@ tۭ .7~zwWa~ ,ǔjS IBzǴ4cW .ԟu ̆ >7R4; ̴0hfʹ9 dʹєQjF]ϔTmT*4t:geTֳLӔAW8=7pմ CRԻѨzEtQT״bxشQ+1|eTcTOTr ƪ~ݔboK ߔ)\2HޔlN*jBխ%WbAT9t4 4H3UH,TFv4ɻ L\οAH 3ԾX,ep/ iP$V50t=ώvl1Y&3T5 z5ԢQvM ,M#NP"ڱM49qWTi24 ~dԩV:A 4( q5#<x3W{9!?5q1O0À" lj)F0R +aUv  u Cf ^ k 5 dm !Rw բ mpcH Umd˅k5c<uos- U % XU %L*2_D) u'ZaJ2<0{ c!5ADڴgO` j P^Xu b085)= upRd,wc!խqj"UmPvz7#)9%9J=D&Krs.'=2?<m(5+q$9c)UO %5e B88%u[%u;m-UyR.U Ջ/ڴ/n'1̌2 0u: mx.9 HQu( @M5vϩS $]OOm>SUvFNTVS?Vؽ *RuX3 Y$62YxKy[0 mdliA%ְ5J ipU1lbb?KHR5O -JvjJ.'i'dνS95 z 5} BUl'S X51i%?U5U5v'z޻LIb5\e$5x8K T~}XtR,15 }5`&8bUp"H5j;HCbH n G]RHXM@f5|  '_U?#Qj1$\)NضŹ5u*U7LEU_FuO־#D3 H1r5F O up`&8w(P1õ7p[BY<*urդU ~:-p.M760!?u{1Mܵ*LBg412-6eHd.!5# 0Ow#ƩQY+R$:F~jobz&9uU ӝ7 ܱLgxL0YzTmO鴦5ѻn,t3u1\M we "ܟu9l Jc!UFgl~KW5 o?I~b~ z׾sy5{SC~"E=VU!c v_^= Flv`ԉ26[U`['rw6P2v.Po >/t'bÉ6p'|x$6 e^#ֈe2' #vn %9]i_4i- v bFGcn!x"ע" tF#6{ WEI@>%'1Z&ֺ.Zp'6 (_s1* ۱%!+6y"%B+6^]5e,UIv[ǟ!$ԟ>/vm 3)-V@oW'/Dvf2hSi2h46~,,n6g 2.7®7 Y9{$X;aH=ys >v1 *;=6C/N@3ŀc*@VP3 uB+ wth] uq t~; w65*;L28xpXNtv=a1Wt4MMOma_FJq|.'~ֆ %Gohvr Sz=vm- *vK$7 >cVRYE_g[֙fCk{֌R pٵ_`fvk[JzH>][?TVHS:V hAVJ oCB WzvXV WVS` 0/V"3  ,J d}|f2ߋ2^EHvf^SV_ ݖVZޖVõyտ|ln>i(-@} $ d}7|4H͔֟r\|E\;tQxv0V#3V`.%jLIǭV{ ~M%S6aƨ6ݑ q3I_֦aP*=| K/jv?싏Vn&8< }9˙h!ֈ$_7pʲ0@7 ,V'+v2↕j oN0vMLNɧ]AV0v j6JJv7v9#6ʶ # uv4 T\S.V^:-U>6IBY~Ɩ7$k[+*DVی Hym: MoVDeVeZc*gb͖wQU$ j+1EbR$ Z[ݞ6m1ձ&іx?Ӗ[XaҶ w֏$" }U6M vqVnJB3V)8cHFv0 Ȳ6+T]٩c>@.AV Җ,v?^ѩږ?eiV\UDS6پw$ jV @E؃5?,0dYu &{}`3O M/N 9V RV$F,RM}s6JE%Gvyep6]XMO5c 26>ɼvڼ +Q ku6 qq]Yxv\U fFvf?^ё T $~s=Ve-. @5<&˶(p3TsEѶq#cuW_EZ7})>tVvv7 9Vn(OWcii wJ w#S8= n7:k z&QgњxVn> WrUh %`ܫ Ө]׮$SD.| } 5yK zWZY ĴD uqA%X[ C{:!TS4l5NslW] ܏77 9 _؇AN !7ub_"׿ 9W_#_skh$ɲ)%0}Fu7 ^޹'7Ǽ J(WИ)('אzt+_)ъz*-wEiyN.eJ#'fn)Y0V4 &E0w0K.whz#<3G=s 즹34JA'w_i_)ѽrǚ&7epNεv&Zv"_hk1!7bD&湔 I>7i ^?w;g`R@Wh 7`?"Pd?/5 CWjkD#kw>3XFw}G7w/ HW. 8?8"I׶[GĔEkLC er~M(ݐNw"]_WmQިVbQU+sXuV"$ZtXQr[YwkLΔY)t\K,YU]-(+.Bl_f=_WI;aw{xNAa? +Dbwq GdWCeeqovbV&g>s~f5kiwy@ɊcטY`73W$L"א{hi.nwZ "o6 |p "Sq8( ]A:fp4> ->$r7 Vq]u )(}vAO^Xue]dzOܲxw:{? vקIe~2rwr|Х'v׳#+<|׎ m}š4}W]jԈӁ>팀 #ρ˅l6WF | eʂ4Ϭqa.?2F LˆW /Zew);t TwTݑpI"WKs`*?FX,7xzÇg'3+^zK2˗LVg@izY_V4ѕ7WvnbQ'wH_ !rUl1:j ; f!i>=ZBHVwƣ7 wr~ך6קėP[=^i T馗 z}w6vNw5'{{C2gwz )#rAQ b17,\WTcEKT4}9.U` pM⳷=@Wx EX7K9 ǺvR]e Bؼ Uk/=ˆ=RgE7qƿ7ɺ0+j 7[d- SrdfėD $ Ɨ d*7= W_wH%dwd Ar5GBNuW,ٸ@kWkųx85K7muѷyS Gї͹ eոw;!7ůC7lt_N7nڳS=ٗF^QVF׺< 02:;Wh WcAnQ q0Wh _)pͧs}dK6Zr/{To=Sz()_ e#ȶ x! xh L@ d kXox_]? x77& sXmMZPE-KNxk,c{L8V{"7=j3}[l {8X"Jx{9 =kѳ~" qXڮ{ХLtWJ𱐌![W!BZ!kI FBX#Al^=@$8P"(+%8OY<>j%8zp(쇉}jY6 sؕ&rCwx=t*(Jjj93L,-XPco-xd *)V0ؖP1@EU!101>}{5>X(5J7x7R5xho:::*'@78-$>xyWK֯T?xu?ɦ=BVCCSCGD ?i:EXw)Ehz5ЧdDXz&H͔ }qKe}fOʖ0DPx bQƿvۧRB &Pcz8 RDK;U' cVlwmͬ#W()YU>m Zx² P[xo {"\c Õ]<_2PI]x z!Z#]|`xI-\bx5915bxft[cXB ǺQ/aX$8NJfO]ӚgX^7phxQ}i.6jx9 cgkM(g7 &=@;m)CInWS"o'_ *k_8:\rķqaގcf}aMs8/ k`nsjjuL݆h8 a>.)?W7UO ACi1 Rx{F :XR *JrDxO 'qQa3Ac *낦x3Uy#XBZRYYDRȰe|Hj٦8,o `˜8%=nb .d?xl>w Y~8؊<7Rc8de 8LPdNuҽ6 7U,(ExMȚMcXNY V<8 =8uyXn<.M,Q`L<ɘjkl XMmpߤx>Ř Ϙa z@ݿ\w.Xwچ=Fu2\Ri̸'g[՘ UyR{ո5kA#ux"u,F׸Ka:Qn >8=X l 8k>jn Xd a:o"rno UO/=r 7Sm7%cS՚gxeMj(|^"t9x;96*8x yП8O :xϒpPe_ŭ :44%D<֝D)m!}-~a ΂xZ+:A,X9:K3%mF8E8، a?~4^5qFI8K VKQI̽8@w2F-jWe 18V+ Rى6L#|4k~,K:_v }R <ؘA g` Y No 9EH2X YynN 9d4. :Z9y"NH@9'˘kyǝ\I䵿4DN ?YCM$B.""Y[  "o`@0%ٟ#n'& 2t&Y@X^%JԹ)Y_ 8S)> +Yw 0,w ~5) C$.Hcu.=~{0F.y-&;*~29OR4ٌ s59F'`ܸi67fao@6) W4DjD\9yGa*u[])`Y 5Z[ckXd8+AeO 7dDdg% 't$ Au%z>"Clڠ<} ^-:ec Z$eJ^ /  zdc:V  :L z>b* ZoŔTwz LKXl ft}ĝƷz&f&_Ock:}8 HhdA^:AHY CTH [Z݆ rv4eZQ =IjV:~?"ošd8"#*4"zqG~%z}/Z"D !kK( M w,X^(ZeO-lکꃒ- V 8},F7/Z^ $ {9/+r1ZY(#(erv13ŞX39q'5 08LO6ھU;ۛ7N%O:ש:Z<yK~Dz7EZ. [YFZTdJZKz/LdbqKqsdN>~NZbfNZ쀮NT*: t3UP\ !!T ; `ҸWw PU:}-vjYZ#|>TY b Ϧ\n"\o %dV\ 8_e xbgmFcRuHcz7bbFgڤ_h|<śhڏ@^dj~] fnkZy aj:imz<ftm[yWlћjn'3%4n+X6rnTSg6lVd |mvɡw< ZcwZ8~y eлfyڶbk_|z1 [_}zTxgڀZڵlт?~Ʉ{_^@X@ {j $'*: vMw0rzz4C2 ϕV!3 Fd_8Zj, ť4p vvZ~b Ñ ln%p I;8  Z:0VgZItЏ镚L^ טz48EǙZ[1 Ioᘚp ~5Zv5-Ӱ\.߁:s0ĊM %/V^l ӻ);.):;jC8Qփ:z$ZD)fAYNS8-7Zu 8| )@JDȡ!K]7iz[_[neZIrC xKͺ/+:pdl#(ZGg~պ5\&ך6A4Oc$*:vV ;qٺt9P-|/jIZ+mgh|6 d:ۗ੅_xIwޣ-Kn B:ǰD#^ɹrf\麮# j~i:.A ڲo:< stzŸ 9y?r K38Gm .[z{{˨,X,r9= ;M?&Z/ Zw )=D=?PT{Դ ON neZ𚧐?ZSw0Zt0 [7g.22RL-;{sHBۣ^) ڮٶѢ"Z8muE?|, +ʰ ;MX ] DG^ li/ N kt;Dڥ;gP>0'%I kG[G^y1;r=Ӧ[˞[E ny̵?=i' iQ+;|yfs  " JSe Eޖs{>+TM1xy]"[O m[A@m$Jk:WtS]^{2)ڴb(7dF< [k*;^ k@w%+{\1<Ѹ,4P(+ VV.MHS.jiQC*0Ѵ)±nɆ3;%a9{ZD:sF/t9,Y< x=;J?[s@M CA0O3 A;Ec?nPDnqV E__[40D,%G tA{CjCK?;PA#C] LG0KMF$GuL[-LO۴ DP[@ZMj FMX T1p{XTKLE DXX;N62Y[8IZf[eV xD[: v,Qb{5XZfb;|Nbd;)68'e[O uhiy rh 7UQk{k+2RYn߇/o[s:zn۲0!n`n;OTlStQv;2T}ڜWvo)}x[,Hy`/k~zωM {[^q ;cz)6}jثmJ;A|!6 cYq'!)xH|=jj{t0|12s;x'/<0 c"B}[_AKXː;]mQt ؓ n%ܑӠ{gt UҜ ] 6.;i){ 6P[ ~ )e$}&tƍ;o_ߥ53Ίx[m7t-Le ;3[4V!phHl`E+"[ VA]me` "r[uf{Qhm$_6LN{+%옭uw^[:ДH/m\;xA0x[?ʻsL(>n!؎F:sR!ےlD# H[]T֙i0b}[V1${lS ƛ!pd;Rm lޏ"vۧ$6ОY;C*[`y#;:Ȟ;R&kuT>`[6-[F Ep;ـwbӛnӻ:t J5@;}O+ɑOܛNO ǁݛIya@T  C;[,0H#`;4 ]^ݻkѢ;Mڒ Vb*Bj;! t~[x KF; 2V컂0)E;Ӷ0{&<u`. ů& I$b _%{;u9*+Y#{%rƪ)ZT[ (5B۪m ";A -kDk}d|QfΚ|< C|ܲ.I| w# ߬Ѡ2&y.z\M$$#^qp LJ6h \rM_ 0 =~rHK <{;N\ո yJVbʅ fx;>ܲ5(2v\u/V\aPtݼB#ƺ5gF@Hvۜ~~uywYˋWu<`\o|Q\Sn|W \|I=mKv<4o :~imC9 ?}~|_v1oLMP+1@ރ-"ɽ 63ݩibo4]J~'0Rd5=OU5DDH>]Ɍ[.]M8dk9=7^ p4}=fX ;RrM"f;ݍl&<= y>]C*AYT? :Ӕ.us)<6rB}͗X CS;QmD|nE]2 v`F]RaelF. XlB6=;B}XhQJ!(KTJ0X} BÂfAb jN=}WO|dO=8 (|Q.e+R[8S%<@=h'Uw `}V}VC2"W ^QPWB)Y,ܦiR7e&[u!iV]s]) hNGG]]pz7]]| k<ſ#Մ/}(t dT I}I@Rn,]G^f7ɺhݣq%rk=RgԄowp@qJ? 6"OSwu]_Nw]\w}5^Տ(y=P. my+*B_G}=a$R~=)a<ݗ2 }vD\5݁oU! 8="a'U}= 2T;f 6eP)0I `Wȁ=vT քヽ>)j}}ʮz}f15Џ=(]]ZQ:5-(c T}="oz<bA^ls]Dw] ]ݔҲP(e=S=vA 3KR[Dƙ1`]\(}ve* د.9=&*{'}f2E(u%J=ɓ  RlaE C%1 ȿwsղ 7=PTb/6z E hF+ظ]+]rŠ伝_JV݂tڽi4]G $Q9u4ϫMW w]8cR]0xY˽F=f Xwq].FMuU=3/"}ǻ~=*`]} <-]; _)i֝Ze9(%=B^.ss lq`{q=#0>]UZ]ߝu I F5JA,9,|%=H[s/9ĕ)iY]x|}= v.  ;X=%4I `<}NhjڃC=9<f3n]jD%ݩS]AY5݈^i=, h;}l`U/Tp:*n SB5=٥ od!YzTcL -]JyA* Zgc6H DRxf={}NR rb^NPPZCaBީ3ہ~5 U~Q[VPձ 킽 rnyLT4H6>y^y_K=q dF^~Xdg T^& I%?Z);^zugWrN N_+k]7 r tk/!Z֏ ^bCb#8 6\&^3b@kl({^ v : (^ \*҇_,~EuN6H->509.>ݤ /<R%0WS[R.>حu1>E~Ez3~NlZ,6>I_7^Vzذ8~&c9޽ 148^h;.8:Tgz<:~=>:mJK6~$'AJ!ZlAy  }C`|cC~|;ᙶE^qRE"5G_YDGy ZJ~%CJ~3YPL~M8N|M 9[N8'șMrt}P}z8P^P;JpQރ`"Q>1RH" :WO^R^1XV.SV^DjGVܩV^ SXYi iLZ~IZ7p[>`sx S`"( Aa~VT ;-!*`L ucD:٘dw Ne^.#lckg~Jg>witjx l>NJmi;{,nS,:o~6 2p^=J'Kr"psX(߸75t޾Yxte v>A#s^sfcx= ay. #x |~ ~rjNt>(}0~i =BSv)SƃOK5?aD oԥN>I B ~[KWF<ۍ+pd ?]^RQ ՘!jĖvg sޏRNMNS+򏕞j3!O]bΖ<6۱C͙~w"-~>G!@M~0g U҂ +骠~B Ft'dTӽu娖7gݷǥ^*$/פeZK( ~x>ep&wql2W9}~q) ?񷧞(ەw`~/PЭ~ J' z#^1B>Fɱ Pzz|=bQ^!> mx‹d ֲ~Hf Ds^ ~qk p" ޻G Y5ϒ¼mͽBq >~;̓.~-xI~W<M-ľpSREҹ1`G(фHEɞcq!01v/ɾ7.޼ҍ` @ +k;Ut]0%sH)Zi_"7&#V8O~WR. 2d־1Er3gB;|BP20 =,Qpb^SP ܾmrXJf ~8 twzmQgmio M!wzB82T~FN(#p*{!<^7?^~YFhiYx>W8YtN^^ = J;zNrVލ 3PA~! ~β J+-MޡX +P aEQ@z.?fmRZN_D"_"k_ghX<O\{?MNeh} _ k ?"|ߩ0N{k `;wVj-lyo?n wo["?Ђ L\#?E|$*# E|پ&r"'Wm(Yp(? `l /'?Bz>{&_ } aQ'?#,0E,__H,?N~J._ ^/_YQU/?SOsc/?;zt+1o 2*($5H>$5t ؋C6sm(Xf511 q]Z9Y+ -߂ DNCiwAD]E[uF?I X0ZDoH|=LN3v)MQʲ.M!n*R_-sS_ү jMT_shU?$E KRx@ }TL .|Xi\3S>{k+Zp{zƂZD OE\yU?<׆G\\qK^ Yg /?c_4 h_З c[?_4)Uc" ' ,Y eJm|fXfiX? iG(~l_yZ.̇m^ 6-nJ:Gtm { hlp?ǭ dq: o[qpEYBYs_QsR Y]s_NoCo_Z2Kw/,ƺs|`v)KWz I3{?8,|?D; YAxFl_ I6iSul:l!P+xut;>4 ; Yp_ ̩غJ.r^=ㇿ^oMo j5Sy p[_On=x?O5#[;Gi?(M$?ӳodޯ;uz^[ 4<ܻAvo#a>($B58 q$P Om ?bafQ?`._Ag"8?v0`y& C]ƿUq9?j"Ɵ Y_Y >' }W _OSNP_rgu=PF LLoBߣ ̷͟Ka좨+x_bin?ک|~߭$$ǿP='ݍt52_ +ވv׿nO`XSg*~u_x @q _v;Tg?w^V?\RۿbKsnߪ, "ۿa2Y Rߝd_:+ Ye_r4e̪֩'Ge=w_ 7_/6뿐6q6}AT.Udr?T_+ɀ_HAg [_ 7Fj:LƟxjdN2Sh]_!1n߀tqs.xJ <5x f_ i?:6Ã)>сF%Z) !FJ{`q5!od=)$3z h}0VY<[aJ A>-U*VgOZxe. |fLEu Cg4,K`"`-W(N#5{6D\i)[ OEFWQXNI s :-XRGrJ~8\ucLq;zbG\t9M#-u ZM?wqr [iygAUp3;#47z\ p.~jYw*'Jx?0N,Lp 0Spf@wAD-7O,1=qLS8Xp;: c6-J1G1b4?N(.d' _CnYb[ {[IXxo#-{5;V~0EK,FnG[ %ybIt0V xb\#Z zZ,%`* TDrE&)kuDAG`KB6{22I&L;P?rN4_?`> hXW^DQ{5 Zrf\!nv~ Db:#.}2NpqY-FHN#'iuav #]^|.!Os-!f4*!n=/Wm|!lrF)IX;Zd5G;+\b+*W3:)8JW/$9A5J1MuM7*ig0%@". G.]mP2/CANs|Oiu1d+(`I*eOP^ K (9;&O/<BzUilx"r. qG}^zoYL9::4&78:v( )=/,(kCJV:{~!jUR|S:9#G sTT !9;v"~">2>=?RQ*be~-" d/.F*;TRuv.O^*D@3R2kkD0O?/~?9f{Wm}CBWcAC FN2%h<Z>^UFLwu"]n;mp9Rq[va|Wq 7u6+CAXQ Mh9DjJbDH[N X4{E1a\t;Eg&75-EF_cXdS6nJa-puH;Oz*FO)>KVYMkJwla)L%qq9| !'[E*Av`}QR I{u 4r$)LoKbqlwmgW i&)KN'8 8 BOf ">IXqTFkV e;Fm0-[q D t2l47OKb H.SZcB?uMoK N&L*5fvv\qQwMz84**bLHHJ?C8N?{2F}Cs Vbb\/SlJb'O;hkO5XYQk9 19PCNig7:m@SdssO_Xd{m'h;|3k m*U0!]fmFXF5&G8mq>:I:BdJR/OFZM@ D:Su=]UuF@5C1<%..He)cSws$1]!T2nzYt".m2PeEhiYL" & Mo' " CE^BRl^$|`}7J5C3mz<65ua"9(vKP}/uGw x^x AN9p0W2s 7y-KIv`$!=] Wy%Q ](C>F!4#>;E0C#p]W(+@m+lO`KAS~I-os5F7:*(T wm64 +4*l$Y 0fc`Su A2E)(`],N,@XDqu@=QK -.3K}0mzt)RJ]Xzrxn +m7m{}zrQA^ujuIz$%D#!v"=JbcE"U|/wz@iB8/W/: SQv-6iDz%upb-tk$)BMCmKSsmH,K@.hckT~dZJZO;KS.R|>6sWoo%DpK1.?~FhgJuCCCxHB57v(a_"mHr`!q mV GJO)?lmLkO9TlI [[yr?b72z"Mmepa'jd :_Q{h @% T(1 I K15hp,pCZO ,(fXdq[JdSQT/f3IH#"o15cG _Q*wU+3~U[MX(qZ` *wQ>Cjw2P|f63?{80E"E$(g-H3";z,E*x'8$qA{Mj"@rS*#*wC[P6O E[C707y'3-F(*k"wg+Q%?z)jZQ9(moY a/NB1q s &Hp# iX`^m68:_9?{zylM1Fp+vLy>)[o 9vcep!f yM4i^PE bRfA7Oc 8bx`--Jh:R_=mht%} X'[*(!ia5|"JG($qJ TRGRS-CkPhYfdSn0tC9O'S4B.zzM!*kS`v'^<|E 4mh 4"Z" 0u'g&QYcBJvZaE4hI'/ce:B_iN<rbgFj'tC9'q2h%*`~H@p#FZQG~S^XQeJ}A]1Qs e}YYM^@JYIoGE_RT\oF TbJP@,qyL%iN[Kw8F>]CE.|(_|\Dz;\rSl2ILV%(XjW y27cu /AsV/z97"pI?%M8dV&h\:>n| }04npVkaG[=<>Wu$=r[I8]\pY3._vL[Qc]wA} 0kp.9 hxcw4 N=Ty6k?Dj$Ly9m[ZLpOHZ&TlK>nYO\ !s |9$vf6eW]Qj$:(3W5>XH#,+D_Xp'ljiEim _OoFEY;. Ow!Hmgwd&#IMB[{&+?3U.A;*o_/J=>4p>9g>< mg#QN9$%z3rvkw!LIKyz++;?H>.as f@c-8z70`hR'!I)"\Aa}Pr6hh84)`.Qch_7sS tQi2 kM[N!1#:Epj'P JD[E*i &5E_ a |<)+fNo?+#6S k]P r,:ai;Q8Ss|4[|GfNC @l7"$j.](0*=t:^Cr054\q-B?EpUXSq+A'V\NmHT7"B?'h3C)uwKI [H6\2Y)3/% c$ c 2aAlG\oK/Sgy|U^B6vi=h5})h4^ A# zG}=vQq+8 [] u% xg5M*<=Ea`rjW=w?J}H@ |CA{ 0eK@Lkq6dSauC1H$eUr""yq\Qz;_v y5 nF4CpZU 7v%{kBj4/P6 y4}g/o\RU?c\2S&pOFF\@kLG`LYE?eY8epxqs8uv<6qsG(saw[E[SXCc mZ3?>8Xu:,FRw^e*5~wM6Qq%s#MXU:FKc 7X`,9)*wFG(& 5%Mk.6jt-jNZp{V=0D1MLUAhjmKY+UjiDZ(>97Fr\ Z^]Tgqv8FqNq_`X~4]bGV]C=8ayl)kA{NR619EM= an(Cq _^w98X2-uyxRz/I*/d!Ns`u\>DfLE U#lVEP>o+XUV*;G (iH]gk.s Y^K k[xKq? ?)ue*U qAO%'f>b kc^_j >85H , 'leY_"^Ck!FuM%szO{edf'%%r`)~s OD=!bk RTU p&3y=V]wqnE%FDjXf*/jvj*(B^2dqdY_}<Z"~9<~4x?kQg j]9=wU) !n0% )|%!+0eN>[m/jZD>J"X0[0Hrh%_+eCy&}v{eFUT/$AF6hk;"a`-JQ_cASe)^Pl ZH X"&xS^v]..zZxVVN 89lL|>IAl1M)ah*UEOsD L@) mG{Ctd cT_/E"z3{rP /h8D\htN yRq-KlyT~}oq9(zB:{`^|8y\+95 ,A=-SZSrK.=$fj R.i Bf>b0KwJ[9,q@GS/aP$GzyR+ 9RK4`@ VlLG|t2p}TdHm J2#l xQ<\LWt\_(GO(kBnRWVB#?&7r`D0[7]1W2I_j'(+iIoN?Kpg&"=56%oEWH% S56(,=CtUeVcr=gm;uhai1l ]Okb[eAO)w.>ZcR6y3}lk\]|s%#m?$A_*|5gXm>wTCK*! d['( -v2 71,`V(yd'{ r&5&oLQl;mF*5jv7,57D)f&Ut<'Ibp"0Bl8I}`;2IF5i~ qB'L<HSK SyNZ&!Zg^%lQut1?#'##L|KT]9EP;^Q^| pzwy@hu*y%k^|AJy.k0a?R!:S%HVq?_8~=;ex^]`r$u3A#/)4{;Ayh_Yw2bE$nv{sCb]?~5@Jzh#{Ob*qz1b),3]6^b|D2z`~ \C>1E 2IatzY>].SQccMn|*2jt1*Rmi/ ;Y![?| M:{]-^@q}}eti5weM + fW+S~c*b"1tCh0Ly^[",%./9\*F2F}wx1; Y>KQo7XA9nXHFOmnMpq;"'H):oA=tm["_&#~CWtQ$ e"Z|u 9usMp+={b/.}&5 X|-RCqH 'TV_B!(IyOV,p&~RXZ'<{VCndv*fO7_D/l mc>i{d/.:P:.zd xTN2+4iH6cg3(D_ Wu elrw]A@ 6o!c^{Tri@U:>H'< lsm <P[G{1cH]T[[?{BztC{I<RNqsMg:J 9* GW,7e=Xvj vF@:'NOrTh~`?=(BHz~4G-\g>Fi\- V)TX7Eo9cu#+1E%`EMjFEFMPnOaLz_z*vKJ/8Lga|%v MT|0"y]O2 `~ae3 -L9ti\t .XI) cT(jNmd/t^K:Z##cr}2H9=: ~$J>RM4jMI Dg6xl1UH]O]Ajxh OA<<Lj{H9InTkGq1=aC:nvQl"6+PptD6x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sYcK~[_/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18_/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/mod.rs !?CC?PP"6%9+*MIJHONKNNOLLLOOK ;:>EE?FF"6%9+IB/(NLKHMNL@DGOILOM0LE PHNHIJJOKOMN8M+MPLLJN-OPLPO-P7JNNNFMLM,NO-bHFLHGI K29?MONNOILK"MLKNPO%@OON"KNHG:IMLLN/, <  Cf8&##,Gb/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/sparse.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18b/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/sparse.rs _gc[{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ɚ̠qc,N{}e/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/automaton.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18e/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/automaton.rs qTs %  *:PNNKNLPPNJIJINBEDLKNBJIILNNJNN">IDLLNO+LPNA-PLOIKJIKM.ILLNKPMLHFFGILKLP! EN,N KK LM9MNIEKNK!O* >0)L*I/. P>+,3 BMMK<OP?O>$MN0LKKPNNNPNNKNK!PE >0)L* MI/. LOO++,3 ;MM7HOK=OL8GIJFHN*MMM0 &OM6OM,HKN0'?<+#$E/:7 OM6OM,HKN0'?<+*/:7 MOKOKPPNP.OPO#@LN&MHJL& * M8DDEJEEEF3 J JJGKNGOMLK#NP<MKKN' %* 5POLOM'OKNI9213:99/9KLH23<NJIPF/+21$ LLK:/521&MGK71521&MP.B521%81%91%3 5OOMN=PMNOOLHKJLHMPMP2ONMPPLJGJ?c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rg/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/determinize.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18g/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/determinize.rs V(VƮ,a2@. # #B*' >#+P)2$"* NPOO#"""-JLN3 NMKJ%*"5#:"= 8J>KLM11C OG: PML% PG#- O#ONLPPO>I>- "IPOPHGKLLMJPP"JNL&KHKKNL: M(KOKNJJJKJ$ILPNO'.NJ8OA;*MP1;< M LKIJMP3CFO,/4,PNOJ.NDNH2 ;#!%# PP%N$#H?24- $O?9C LNNE$KLP&JOF/L6/$= &I3*"3+OO@4@ 2A BOMLOMKMP#0-N LNNHN&H*")98(G NMMLIE PNILOOP/@D$ =EAE( ?ACC&-LA(NC(NC(M&-( JKL)P5/MGMOK6" #,)$ 6+  &L5OO?MNP"/GIM-+ 7KONP"'.*/)+( (N$91&):1H B-H PO(J3;-& LI>@$- OK& 7KI"70,qxw/,!Jd/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/remapper.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18d/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/remapper.rs /A|iwG-qMM&KNNNOK9DNOPMPO!M0+"MP KLKO.ENPMJN;PN0;KO(NKO7JGL7LMIOON-:LM<9;O!MJ(! !MI;=GONNM LMLO LPPE'$5("?"* >CNLPLNLIP HN3-/&-5HOK6 B#'& %$ @2 @'  ?'! %GP @- @" ԁ=e 5a[/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/lib.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18[/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/lib.rs Ԇ5J x[{IKJFOONL*M M)KJ PIL MLKMNJN ONGIOMI:KPJKAMJM1-MJ/H M/KNOOMN.844; 5$;;<<*A'+*ONM,63D"A''&*/#$+PN/,$D"H'"3*/0+!ONP-/ @8+9((!!0+NLFLNOMNP7KMLMLJM-LNMMPOKOLMPNK@JPDLMOLNOJON  N-CPMNK?N,IK ILNKMJMM&PPGMILM<I.%^QSDiNINNNKHMP iLNO<LIMNMKMOJMPNMMLMNHPML)KMN /BDOOKGNOLHHMO<*NJOKOKLMLJ&OLIHPKJHOPNPK*MNLONJMDFLJ8PPNOOFLKOLD J>0OI@HLOK MK]K-JL67pkD-LIGKJRHMQJLPNJN)GGMKOOOML)4NEM4J-=CK$ MN+ *JPPM O +IIPOLOPL@LN ,17.(PI-FKJFPKN!! I#',  = ! .#JHM'(8Dh/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18h/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/mod.rs c.3cjmQGENOOEPT'HPO#?5FPLO L%GPLNO JFMPJOKN@LNMMNMN="    M ~|2bX}͙3Fl/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/builder.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18l/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/builder.rs vR'PAh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mfJGRMq/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/literal_trie.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18q/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/literal_trie.rs 9p{ G3:NLJ PCLMOMOLP ONKKMNMNOO$IPJOMNP?MJ?MNLN*5OIO6MPPLHDLON7 NPJM!O-%7A-%6,JL$K&,P4 'KMM'7?)3IC4N6D3< <O+G KMKLPKOONL5( =439'%#9'%PNN"-#:ML1&8@F0(N3LK>#..IIMB5 ;&ML/D  <NK912OO?M$B("($1073 (G&BA GMOKNOPJ78!NLA4JMLMDLO('JPA,),,(EHKMPNKHLF$KLGLN)CPPLL"! O'MENOOOD 504K0 $LMP.-(/ <./#L,6"G6.5%%"&% OOL#3&,"5:8 2 4'G7!    /&!#))"#*' # /$%%))##"+P( #_jgheT>K8Kh/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/nfa.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18h/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/nfa.rs >Ǐɀ8+w۔۔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}z%]Eo/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/range_trie.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18o/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/range_trie.rs 擹DF&? &&}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r uJ?e/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/alphabet.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18e/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/alphabet.rs 4?~!p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ԻTlle/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/captures.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18e/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/captures.rs rcVS4r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tͨ4#h/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/interpolate.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18h/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/interpolate.rs bS3ءSn>OLMNHA OLGMO>OMNHNINKOPO7"NOL'!PMJ7MMON!F+!%  $0:$511 #I- 0C&#1 3."7$ O JIEMMON!F+/  *1:$*:1 #>- 08& 1 3."7$ (M>B-1"-%%$(K0AONO@"* 0IMOP5-*%' PJ>M653PLL;*%' L(@ 0+'B 5K /B 02**,700/"## .101/113335597!,%0%&:1 7!,$09&:1 (S;2:1          &  &  &  &  &  & JML$ & aaaabbbbbbcccccc{\fga/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/lazy.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/lazy.rs 'fs.„ =NF@KJP,PKJLNLMJL"#NKL&WLLFP095NK!0&#<HMK;LN0*8>;  -*LKKLNHM#NOQN G$KOP1MN&L1GEP;=$ 'IJM').(9IKL35'"#'P<5OJ% NO*'9PN# B?H %, MP5   ,,8E!KL ###4NMEONN%#!$KOP1GIK!EJ@;718 +HLO<)MMO5JNIP5[JN,:%%&'OOKNB(LNLLFNO!*?@88>NO!LJ).MJ)L-! NN'GO B E?#E/- %:LD< NKPMM#"MMMB  !!6< $$&)NGKPOGLGC (8)!D(JI% bqJrf|a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/look.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/look.rs 1~)'&Im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Ջ͏̑БЕۖėȗژޘ™ޞءܡݥɩͩׯֱڱԵλһʀ΀ǂdžˌ֍ÎՏُɖȞàǠ֡ڡզާʩΩȭԮǰ˰ڱޱƴǸžпع%Ý+JB{.a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/pool.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/pool.rs "7I>!XUD#ӾӾ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ͱ[]:n/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/determinize/state.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18n/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/determinize/state.rs 6 7̵®q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` Kec/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/memchr.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18c/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/memchr.rs /ЁҼ36T9X]NLD+9E%9N*9.9F&9O+9/09E/9N:9E9F09O;9F? Tǯc/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/search.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18c/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/search.rs w ԋ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h+*PL0'A36J"5" 5 '  :,MP;='9?L>=)2;O9=29AOK=6@F-O!'H N<LJ"JN LBLJNMF@JBKLM%n lNMPP9))G=F!3B9BMLLN2.I0$  OPOLN;NM& 9*. G?( !?K( !?K !+#(+ 3&8ps):m/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/unicode_data/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18m/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/unicode_data/mod.rs +Q0 H9J F' L(IGLK";;}#Gea/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/regex.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/regex.rs cq={7Jz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kCueƶ؊a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/accel.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/accel.rs }#iD N PNNKNPD22*2**MKLM*LNLMLPJNLO LMPHPL=*04 <BFNNPN=1>MONML .PL5"6>E74NO 56>E74KH>PN+J<-#-NDL*>M3,KIN;"!N@LJN'P!D5HFII'O) -5 L27/KP")6 I&5A9710M&+M2).. CJ+E*PPMJOJC9/5 %4.-B!I(:I4E/ 4B?>**OGOI )*$C" J ECP8#GP-=J/ A*;G' -?*.K NPJGIJMOL6KKMKA @"(OI1MBN!?DG02D /7MK7NOL0J". MN9 "11 + 8!#L"%2H!'O"+A?"-"=JJ"G%#; HBD9j0gd/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/minimize.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18d/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/minimize.rs <5|/ƦX&}q%; . *CCHDIIN PLONKNNOPOPK:LMONGJ"",(?$%;3-KLNHLFIL 8$@C=*?*64 L...#P<36?NLPOMO )4O-C"K-C"!H..C%12"46%6I" NPQJNHB$@ MKL5@)CL  :'OLOHO?NI3-A GGKO81LKMPM:=&% LMIC(@*. PNM42LNNILO:.*6*+I4,,,, +*+I4$,,,,'& *>3!J$ !DJP$P-.$0'8A!2&6)* &@ O# = $3M 8%(!&'' !'?52. D1 DOI'#"'#"'#"'#" @1( DOI'#"'%"'#"'%"  ;÷V^ՍLtb/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/search.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18b/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/search.rs 'hBNMsd, 2 4( -6 N"01 "12 5(  -5M) JP%"$=1 !0PPOP<1$ !&? MJIINMMP1 NP"!09 -*:N  5# !3+&6<,> )&65  J4M5- ##d1c/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/special.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18c/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/special.rs WOGmCf{9  @5LM MLKNMNJMMJKKKK#KLJLLKOLCFMNIG3NJNIG LIINN4PKN?PONL4(/HHKMLONO57("K$!0ML$IG)K6-I6K6I2JNM/5PP3NHKMNIMK+ >>>>>>>>>>7NFPP!1#0#7#6#1#0#PL"% N"N (,,,,,, LKJM-O&6LHB- .3::::::  3JDF>= >= >= >= >= >= .-D -D -D F>B =B =B PD PD OD A%< '> '> '> K&(ONM>5M NLKO')8-O <@BBBBBB!9 H A1L<"'F LMP&EOL";KKAH>H>6I?PJ?PI?PN=M G+2"=M G*G*7OC3{K!a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/start.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/dfa/start.rs Y\Tȥbv99:N JIFNKJOO6-7 MPPM,M8MH,20L ')* H "(& 192A90?PX́sùe@g$j/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/error.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18j/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/error.rs +dRH% 66&JK2HJPMJKJ OMH5!PKL@(OL&*HC1OM/L,MM).HK)LOJ/0H (C9J;L7B&N@$LDIDK9A(E':< )K'K-7)@H:6F<6COE:.';F! B0 X]rtU;KU,h/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/map.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18h/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/thompson/map.rs ܐRԐ<; 1NLMP%+ O JMM LKMNF8O/P=0!KL7G?M$MIHE3<J8!H9#L /7+V*09<:1NLO%+  JMM PKxHlqM(?Qc/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/escape.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18c/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/escape.rs ˴ a7id(PFT6GMNM=MN&GIN% N#L'2 F>NJ%,2GL 7$1)-+B "<,9 [zި\.ma/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/iter.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/iter.rs Yl*3WgtNLKNM/K9%@LMK*;LKPLK9IIO3KOPOJM%C7NK)NPPLHMJNOONKNPOKK#AOMOMOKIK:JKJLNPLCM KNGM:'KMLHK>'5()'<31 LB/@FNP,#@LPL%31;IN52O&JPNO> " ;+D+5Q#5Q#5Q#Q#3 P " #+(+4Q#4Q#4Q#Q#3 F G. OD+4'5! M%#5'5! #5'5! #'5! #3 = C) 7H L&P / G1% 5M) *0OK+ C1% BH) '-O5LHJF$ " ;+DK1 9)9)9))3 &# G0O0LHFF "" =+DF- 8)9)9))3 J C,O:LHGHONKKKJ?OM +" @ K;**643D-4E,3C*&3 " G3KLKO // GNKIG OKKJKOMPP1MJIOPILLNI *+ CJMN(E:7KOF'(7ON38;HL=23C/B?=K-#+PM (E:7KOI5>%;HL=20C. :9K)#+KP(E:7KOD16!;HL=2,?* 7H IM(E:7KOB$%7ON350C.AK#(%-( :K*#'+LL (E:7KOJ-8-C- ;H iWbĬo!s/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/aho_corasick.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18s/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/aho_corasick.rs 'M=00 94#'=  8 LIMMKMMNL(-?;NPMN543=>%+;LOKJHL5MOKJO!"!$N!% "%B=  8 OB D=  8 ;7-B &=  8 #  =  8 JOLON@LOLMLO( $b )('=  8 (+.'!4 %BN(! D,$B &  tԌ] km/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/memchr.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18m/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/memchr.rs S;8AT# b%%'8  3 $01 %B8  3 >,%$ D,B &  #'8  3 $@--" %B8  3 G,%$ D,(B &  ''8  3 $@---& %B8  3 O,%$ D,7B &  J(zo'Jm/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/memmem.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18m/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/memmem.rs :BVw9\JE-'N  I $.K$ %BN  I 8,>$ DN  I /4F &N  I '  N  I  )> 57-wl/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/teddy.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18l/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/teddy.rs ͞8ɉ;Cj669#OIJK#4.PNMPOLO!4)8PO<4 '=  8 CIN+IIMNGB?D<M=/!@+?" G %B=  8 I,B D=  8 4;7-"F>B &=  8 4K  =  8 MNOOJLNOJPO3LNLMMO" CEVhípVj/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18j/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/prefilter/mod.rs SpűPTŒu5>MMPN@MNMKNLKOPML!KILI  (',)(1DDEMMPILLHJKIMO@OH:L'$D :'%9/K'$ D :':9/#NIHHLLOP4GMKN 2 #H5P$%@ NP$$@ 3 7!M; O,'  " :2332137)= NKMM 2 8M$*%@ 3 N2OKKPLO:KL( 2 )"! O$*%@ 3 -9PNMMP 2 0M$.%@ 3 F'  " * LOPMP= 2 0M$.65B 62 %B 3 H'  " , M*'  " $ 0,'  " MLMLLMJLM.MMO2$'  "  GKJPB?+$PNMALOPCM%M@5%9B&9D(9& 9 KPJMMON'KMMOK1KMPK7GJJMNMIO$GBDO'H! N GI#;N JMPLO8/. 90/ 90/ 8/. 7.- 90/ =53 I NOO NJM-P"MNG"32.> O &:  N MG"32.> O &:  N Ha 6ɽ9ng/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/primitives.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18g/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/primitives.rs Q8ig3-sK N!LM%KJL KKLMPNLP@KMJMF..P4NJMLO 7>%=&9P6BN# %ON(G&,NKNAOMLKNMOKPKNMJGH%LNLNEJ1KONE PJMPP( G!J L@!( 9M=?F@"JF$J%FNNO=J!N.=J,LH(MM(O7(N;P<&M+M4/,+F 5M+OMCHP+// 2;$2/ 5;$'6#"E9I 8#"E.I 8#"E.> 8%"G0G - ,<?+C COJ'H%..GJ #"#/, *CDH, A#;D8 =H6KP">@O%HGM5L)@8H12 'N(4"L 0 L&0 K)0 =MN,."O3MDJJO3OOFBL3%I,I6  O0*OP 22( 5>, "52( "8>, "*/ &=E &=E &=E (?E /?2 E L K/*H-#  ++OI'&   1*)D,'"6 "2) PLKM ) =,1& J*=,M53! HLKLIO4M(*I;F"AOOKMEM&F PHONO;.9 )%5 )#Y-9ZpNgc/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/syntax.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18c/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/syntax.rs m%5U+A n?:;MMLJH  K9&4J/4,LL &'% H?/N2JK&"<A/J, #LK& &'C& =LDLMJPKMML4 M1@MKEIMKL1 FPN:LNIN5MNLLMNONOJ.8! 0MENPOFMKN-2 3(<16+@5"0%B.6+=2/%**1&8>#5)=3)7) #DJHJJLJHJIFLIILHIKGKHPN&LHFLL?,< IK #2EIJD=)" %2#. /2X 2 `^'`/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/int.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18`/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/int.rs CPΉ#vX 22LLPK M2IPMLJ$PPOLLLNM;OOKOO LLJG@ ! ! !! A &   !! A &   !!! A &  "! = &  ! ? &  ! ? &  ! ? &  LLOPNMMO, ! ?"lg/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/sparse_set.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18g/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/sparse_set.rs `xh \ˑ>>F!3*-MEIO&IOO>OMK  PK7,, B,K;(( %$9N1<;ML.FNP86I405MLK>9JNK6K @+K;,6 89*HO-4(7,M1L#DLO9; !A M!/ ! G9%A4%57J1M&G<=3ME*9,Eq4M ,D/a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/utf8.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18a/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/utf8.rs h}`i}O388"FO MNK5,*L$#' ' ' +HM99MN@5A$,F:06'HL395F%/G$&;L(K 5$.+...O,H5<KN5"KLIPI!:ON5/MMOM%IMN6HFJNLOLONM(5 2P/# Yy+,z̅^/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/macros.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18^/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/macros.rs *d- +N,$  33nM㺥@p]}K_/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18_/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/nfa/mod.rs w nդ7LRNPDNPNKK=MLIJN0NLPNMM FN,PN0HKFLLPNHNS8!f=<~P`/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18`/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/mod.rs B^9 cf09L KI LJ"LM$J;F LKPJ.L=7)OLZ/Фsl/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/determinize/mod.rsP/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18l/root/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/regex-automata-0.4.18/src/util/determinize/mod.rs *G3LJ@KFLN&JLEDM GNH7OKINJHILM-D  - LONNMAMM2MMOGKPO)LOO)NPOK(&IP IKM$B)KGKHJ(%OJPNO*'L/3A2A&'), 77 .3 ; 9"/2M ""03 :",/A".1 IOON@ ENN;)* 0!'" O4FN15LI!BLJLMO1KDLOD'24 (#*2$+47IDIIKHHDA&LKNJMOP&KLNOKN":=:.%$-+5@%-+3?%-+ HK.OLOPPMNMNMOL2"G( 5O ( .:MIMMN DI8LOOO/"+OLK4NE!#210(::3?3')NP0@A&; MONNPPOOLMP'JL##3202/242A*7MLOEOJOLHFEPLMOLKMMOJOMMMJMLBNMKLLNNM3//'33)KNNMOPMOPMPMLL1OJI;MJPO'2NH2OPP'2/JJJIJD2 FD.(5L!)' =& )/:< ), 4H0/G)/:< 404N4PEJ)/:< 0P0EJ)/:< )0JNON6)203>@ I{(HYI)1_)a)m)x){)$~)))Y)))>)Χ)))Ѽ))))3)?)=))"))`)))****j * *o**1*f*0*A**#*.'*)*+*?.*U0*1*B3*4*x86_64-unknown-linux-gnu{QwSZP regex_automata-c7b3f1c1cc1ae10f0! (` &6&  99pp&u&&&&&&&%&6&66r6 &&r&r&r&&r&rr%&o&r$r&lr&r&r$$ll&3 b&Cr&&r6rm&&rr9Cp $"l&rOKK3 rust-end-file