13af6ab5fSopenharmony_ci/** 23af6ab5fSopenharmony_ci * Copyright (c) 2021-2022 Huawei Device Co., Ltd. 33af6ab5fSopenharmony_ci * Licensed under the Apache License, Version 2.0 (the "License"); 43af6ab5fSopenharmony_ci * you may not use this file except in compliance with the License. 53af6ab5fSopenharmony_ci * You may obtain a copy of the License at 63af6ab5fSopenharmony_ci * 73af6ab5fSopenharmony_ci * http://www.apache.org/licenses/LICENSE-2.0 83af6ab5fSopenharmony_ci * 93af6ab5fSopenharmony_ci * Unless required by applicable law or agreed to in writing, software 103af6ab5fSopenharmony_ci * distributed under the License is distributed on an "AS IS" BASIS, 113af6ab5fSopenharmony_ci * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 123af6ab5fSopenharmony_ci * See the License for the specific language governing permissions and 133af6ab5fSopenharmony_ci * limitations under the License. 143af6ab5fSopenharmony_ci */ 153af6ab5fSopenharmony_ci 163af6ab5fSopenharmony_ci#ifndef ES2PANDA_PARSER_CORE_REGEXP_H 173af6ab5fSopenharmony_ci#define ES2PANDA_PARSER_CORE_REGEXP_H 183af6ab5fSopenharmony_ci 193af6ab5fSopenharmony_ci#include <util/enumbitops.h> 203af6ab5fSopenharmony_ci#include <util/ustring.h> 213af6ab5fSopenharmony_ci 223af6ab5fSopenharmony_ci#include <unordered_set> 233af6ab5fSopenharmony_ci 243af6ab5fSopenharmony_cinamespace panda::es2panda::lexer { 253af6ab5fSopenharmony_ci 263af6ab5fSopenharmony_cienum class RegExpFlags : uint8_t { 273af6ab5fSopenharmony_ci EMPTY = 0, 283af6ab5fSopenharmony_ci GLOBAL = 1 << 0, 293af6ab5fSopenharmony_ci IGNORE_CASE = 1 << 1, 303af6ab5fSopenharmony_ci MULTILINE = 1 << 2, 313af6ab5fSopenharmony_ci STICKY = 1 << 3, 323af6ab5fSopenharmony_ci UNICODE = 1 << 4, 333af6ab5fSopenharmony_ci DOTALL = 1 << 5, 343af6ab5fSopenharmony_ci HAS_INDICES = 1 << 6, 353af6ab5fSopenharmony_ci}; 363af6ab5fSopenharmony_ci 373af6ab5fSopenharmony_ciDEFINE_BITOPS(RegExpFlags) 383af6ab5fSopenharmony_ci 393af6ab5fSopenharmony_ciclass RegExpError : std::exception { 403af6ab5fSopenharmony_cipublic: 413af6ab5fSopenharmony_ci explicit RegExpError(const std::string_view &m); 423af6ab5fSopenharmony_ci std::string message; 433af6ab5fSopenharmony_ci}; 443af6ab5fSopenharmony_ci 453af6ab5fSopenharmony_cistruct RegExp { 463af6ab5fSopenharmony_ci RegExp(util::StringView p, util::StringView f, RegExpFlags reFlags); 473af6ab5fSopenharmony_ci 483af6ab5fSopenharmony_ci util::StringView patternStr; 493af6ab5fSopenharmony_ci util::StringView flagsStr; 503af6ab5fSopenharmony_ci RegExpFlags flags; 513af6ab5fSopenharmony_ci}; 523af6ab5fSopenharmony_ci 533af6ab5fSopenharmony_ciclass RegExpParser { 543af6ab5fSopenharmony_cipublic: 553af6ab5fSopenharmony_ci explicit RegExpParser(const RegExp &re, ArenaAllocator *allocator); 563af6ab5fSopenharmony_ci void ParsePattern(); 573af6ab5fSopenharmony_ci 583af6ab5fSopenharmony_ciprivate: 593af6ab5fSopenharmony_ci void ParseDisjunction(); 603af6ab5fSopenharmony_ci void ParseAlternatives(); 613af6ab5fSopenharmony_ci void ParseAlternative(); 623af6ab5fSopenharmony_ci 633af6ab5fSopenharmony_ci void ParseNonCapturingGroup(); 643af6ab5fSopenharmony_ci void ParseNamedCapturingGroup(); 653af6ab5fSopenharmony_ci void ParseCapturingGroup(); 663af6ab5fSopenharmony_ci 673af6ab5fSopenharmony_ci void ParseAssertion(); 683af6ab5fSopenharmony_ci char32_t ParseClassAtom(); 693af6ab5fSopenharmony_ci void ParseCharacterClass(); 703af6ab5fSopenharmony_ci void ParseAtomEscape(); 713af6ab5fSopenharmony_ci 723af6ab5fSopenharmony_ci uint32_t ParseControlEscape(); 733af6ab5fSopenharmony_ci uint32_t ParseDecimalEscape(); 743af6ab5fSopenharmony_ci uint32_t ParseLegacyOctalEscape(); 753af6ab5fSopenharmony_ci uint32_t ParseHexEscape(); 763af6ab5fSopenharmony_ci uint32_t ParseUnicodeDigits(); 773af6ab5fSopenharmony_ci uint32_t ParseUnicodeEscape(); 783af6ab5fSopenharmony_ci 793af6ab5fSopenharmony_ci void ParseUnicodePropertyEscape(); 803af6ab5fSopenharmony_ci void ParseNamedBackreference(); 813af6ab5fSopenharmony_ci 823af6ab5fSopenharmony_ci void ParseQuantifier(); 833af6ab5fSopenharmony_ci bool ParseBracedQuantifier(); 843af6ab5fSopenharmony_ci 853af6ab5fSopenharmony_ci bool IsSyntaxCharacter(char32_t cp) const; 863af6ab5fSopenharmony_ci bool ParsePatternCharacter(); 873af6ab5fSopenharmony_ci 883af6ab5fSopenharmony_ci util::StringView ParseIdent(); 893af6ab5fSopenharmony_ci 903af6ab5fSopenharmony_ci bool Unicode() const; 913af6ab5fSopenharmony_ci 923af6ab5fSopenharmony_ci char32_t Peek() const; 933af6ab5fSopenharmony_ci char32_t Next(); 943af6ab5fSopenharmony_ci void Advance(); 953af6ab5fSopenharmony_ci bool Eos() const; 963af6ab5fSopenharmony_ci void ValidateNamedGroupReferences(); 973af6ab5fSopenharmony_ci 983af6ab5fSopenharmony_ci RegExp re_; 993af6ab5fSopenharmony_ci ArenaAllocator *allocator_ {}; 1003af6ab5fSopenharmony_ci util::StringView::Iterator iter_; 1013af6ab5fSopenharmony_ci uint32_t capturingGroupCount_; 1023af6ab5fSopenharmony_ci std::unordered_set<util::StringView> groupNames_; 1033af6ab5fSopenharmony_ci std::unordered_set<util::StringView> namedGroupReferences_; 1043af6ab5fSopenharmony_ci}; 1053af6ab5fSopenharmony_ci 1063af6ab5fSopenharmony_ci} // namespace panda::es2panda::lexer 1073af6ab5fSopenharmony_ci 1083af6ab5fSopenharmony_ci#endif 109