1 //===--- FormatTokenLexer.h - Format C++ code ----------------*- C++ ----*-===// 2 // 3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4 // See https://llvm.org/LICENSE.txt for license information. 5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6 // 7 //===----------------------------------------------------------------------===// 8 /// 9 /// \file 10 /// This file contains FormatTokenLexer, which tokenizes a source file 11 /// into a token stream suitable for ClangFormat. 12 /// 13 //===----------------------------------------------------------------------===// 14 15 #ifndef LLVM_CLANG_LIB_FORMAT_FORMATTOKENLEXER_H 16 #define LLVM_CLANG_LIB_FORMAT_FORMATTOKENLEXER_H 17 18 #include "Encoding.h" 19 #include "FormatToken.h" 20 #include "llvm/ADT/MapVector.h" 21 #include "llvm/ADT/SmallPtrSet.h" 22 #include "llvm/ADT/StringSet.h" 23 24 #include <stack> 25 26 namespace clang { 27 namespace format { 28 29 enum LexerState { 30 NORMAL, 31 TEMPLATE_STRING, 32 TOKEN_STASHED, 33 }; 34 35 class FormatTokenLexer { 36 public: 37 FormatTokenLexer(const SourceManager &SourceMgr, FileID ID, unsigned Column, 38 const FormatStyle &Style, encoding::Encoding Encoding, 39 llvm::SpecificBumpPtrAllocator<FormatToken> &Allocator, 40 IdentifierTable &IdentTable); 41 42 ArrayRef<FormatToken *> lex(); 43 getKeywords()44 const AdditionalKeywords &getKeywords() { return Keywords; } 45 46 private: 47 void tryMergePreviousTokens(); 48 49 bool tryMergeLessLess(); 50 bool tryMergeGreaterGreater(); 51 bool tryMergeUserDefinedLiteral(); 52 bool tryMergeNSStringLiteral(); 53 bool tryMergeJSPrivateIdentifier(); 54 bool tryMergeCSharpStringLiteral(); 55 bool tryMergeCSharpKeywordVariables(); 56 bool tryMergeNullishCoalescingEqual(); 57 bool tryTransformCSharpForEach(); 58 bool tryMergeForEach(); 59 bool tryTransformTryUsageForC(); 60 61 // Merge the most recently lexed tokens into a single token if their kinds are 62 // correct. 63 bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType); 64 // Merge without checking their kinds. 65 bool tryMergeTokens(size_t Count, TokenType NewType); 66 // Merge if their kinds match any one of Kinds. 67 bool tryMergeTokensAny(ArrayRef<ArrayRef<tok::TokenKind>> Kinds, 68 TokenType NewType); 69 70 // Returns \c true if \p Tok can only be followed by an operand in JavaScript. 71 bool precedesOperand(FormatToken *Tok); 72 73 bool canPrecedeRegexLiteral(FormatToken *Prev); 74 75 void tryParseJavaTextBlock(); 76 77 // Tries to parse a JavaScript Regex literal starting at the current token, 78 // if that begins with a slash and is in a location where JavaScript allows 79 // regex literals. Changes the current token to a regex literal and updates 80 // its text if successful. 81 void tryParseJSRegexLiteral(); 82 83 // Handles JavaScript template strings. 84 // 85 // JavaScript template strings use backticks ('`') as delimiters, and allow 86 // embedding expressions nested in ${expr-here}. Template strings can be 87 // nested recursively, i.e. expressions can contain template strings in turn. 88 // 89 // The code below parses starting from a backtick, up to a closing backtick or 90 // an opening ${. It also maintains a stack of lexing contexts to handle 91 // nested template parts by balancing curly braces. 92 void handleTemplateStrings(); 93 94 void handleCSharpVerbatimAndInterpolatedStrings(); 95 96 // Handles TableGen multiline strings. It has the form [{ ... }]. 97 void handleTableGenMultilineString(); 98 // Handles TableGen numeric like identifiers. 99 // They have a forms of [0-9]*[_a-zA-Z]([_a-zA-Z0-9]*). But limited to the 100 // case it is not lexed as an integer. 101 void handleTableGenNumericLikeIdentifier(); 102 103 void tryParsePythonComment(); 104 105 bool tryMerge_TMacro(); 106 107 bool tryMergeConflictMarkers(); 108 109 void truncateToken(size_t NewLen); 110 111 FormatToken *getStashedToken(); 112 113 FormatToken *getNextToken(); 114 115 FormatToken *FormatTok; 116 bool IsFirstToken; 117 std::stack<LexerState> StateStack; 118 unsigned Column; 119 unsigned TrailingWhitespace; 120 std::unique_ptr<Lexer> Lex; 121 LangOptions LangOpts; 122 const SourceManager &SourceMgr; 123 FileID ID; 124 const FormatStyle &Style; 125 IdentifierTable &IdentTable; 126 AdditionalKeywords Keywords; 127 encoding::Encoding Encoding; 128 llvm::SpecificBumpPtrAllocator<FormatToken> &Allocator; 129 // Index (in 'Tokens') of the last token that starts a new line. 130 unsigned FirstInLineIndex; 131 SmallVector<FormatToken *, 16> Tokens; 132 133 llvm::SmallMapVector<IdentifierInfo *, TokenType, 8> Macros; 134 135 llvm::SmallPtrSet<IdentifierInfo *, 8> MacrosSkippedByRemoveParentheses, 136 TemplateNames, TypeNames, VariableTemplates; 137 138 bool FormattingDisabled; 139 llvm::Regex FormatOffRegex; // For one line. 140 141 llvm::Regex MacroBlockBeginRegex; 142 llvm::Regex MacroBlockEndRegex; 143 144 // Targets that may appear inside a C# attribute. 145 static const llvm::StringSet<> CSharpAttributeTargets; 146 147 /// Handle Verilog-specific tokens. 148 bool readRawTokenVerilogSpecific(Token &Tok); 149 150 void readRawToken(FormatToken &Tok); 151 152 void resetLexer(unsigned Offset); 153 }; 154 155 } // namespace format 156 } // namespace clang 157 158 #endif 159