Martin Probst | c4a0dd4 | 2016-05-20 11:24:24 +0000 | [diff] [blame^] | 1 | //===--- FormatTokenLexer.h - Format C++ code ----------------*- C++ ----*-===// |
| 2 | // |
| 3 | // The LLVM Compiler Infrastructure |
| 4 | // |
| 5 | // This file is distributed under the University of Illinois Open Source |
| 6 | // License. See LICENSE.TXT for details. |
| 7 | // |
| 8 | //===----------------------------------------------------------------------===// |
| 9 | /// |
| 10 | /// \file |
| 11 | /// \brief This file contains FormatTokenLexer, which tokenizes a source file |
| 12 | /// into a token stream suitable for ClangFormat. |
| 13 | /// |
| 14 | //===----------------------------------------------------------------------===// |
| 15 | |
| 16 | #ifndef LLVM_CLANG_LIB_FORMAT_FORMATTOKENLEXER_H |
| 17 | #define LLVM_CLANG_LIB_FORMAT_FORMATTOKENLEXER_H |
| 18 | |
| 19 | #include "Encoding.h" |
| 20 | #include "FormatToken.h" |
| 21 | #include "clang/Basic/SourceLocation.h" |
| 22 | #include "clang/Basic/SourceManager.h" |
| 23 | #include "clang/Format/Format.h" |
| 24 | #include "llvm/Support/Regex.h" |
| 25 | |
| 26 | namespace clang { |
| 27 | namespace format { |
| 28 | |
| 29 | class FormatTokenLexer { |
| 30 | public: |
| 31 | FormatTokenLexer(const SourceManager &SourceMgr, FileID ID, |
| 32 | const FormatStyle &Style, encoding::Encoding Encoding); |
| 33 | |
| 34 | ArrayRef<FormatToken *> lex(); |
| 35 | |
| 36 | const AdditionalKeywords &getKeywords() { return Keywords; } |
| 37 | |
| 38 | private: |
| 39 | void tryMergePreviousTokens(); |
| 40 | |
| 41 | bool tryMergeLessLess(); |
| 42 | |
| 43 | bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType); |
| 44 | |
| 45 | // Returns \c true if \p Tok can only be followed by an operand in JavaScript. |
| 46 | bool precedesOperand(FormatToken *Tok); |
| 47 | |
| 48 | bool canPrecedeRegexLiteral(FormatToken *Prev); |
| 49 | |
| 50 | // Tries to parse a JavaScript Regex literal starting at the current token, |
| 51 | // if that begins with a slash and is in a location where JavaScript allows |
| 52 | // regex literals. Changes the current token to a regex literal and updates |
| 53 | // its text if successful. |
| 54 | void tryParseJSRegexLiteral(); |
| 55 | |
| 56 | void tryParseTemplateString(); |
| 57 | |
| 58 | bool tryMerge_TMacro(); |
| 59 | |
| 60 | bool tryMergeConflictMarkers(); |
| 61 | |
| 62 | FormatToken *getStashedToken(); |
| 63 | |
| 64 | FormatToken *getNextToken(); |
| 65 | |
| 66 | FormatToken *FormatTok; |
| 67 | bool IsFirstToken; |
| 68 | bool GreaterStashed, LessStashed; |
| 69 | unsigned Column; |
| 70 | unsigned TrailingWhitespace; |
| 71 | std::unique_ptr<Lexer> Lex; |
| 72 | const SourceManager &SourceMgr; |
| 73 | FileID ID; |
| 74 | const FormatStyle &Style; |
| 75 | IdentifierTable IdentTable; |
| 76 | AdditionalKeywords Keywords; |
| 77 | encoding::Encoding Encoding; |
| 78 | llvm::SpecificBumpPtrAllocator<FormatToken> Allocator; |
| 79 | // Index (in 'Tokens') of the last token that starts a new line. |
| 80 | unsigned FirstInLineIndex; |
| 81 | SmallVector<FormatToken *, 16> Tokens; |
| 82 | SmallVector<IdentifierInfo *, 8> ForEachMacros; |
| 83 | |
| 84 | bool FormattingDisabled; |
| 85 | |
| 86 | llvm::Regex MacroBlockBeginRegex; |
| 87 | llvm::Regex MacroBlockEndRegex; |
| 88 | |
| 89 | void readRawToken(FormatToken &Tok); |
| 90 | |
| 91 | void resetLexer(unsigned Offset); |
| 92 | }; |
| 93 | |
| 94 | } // namespace format |
| 95 | } // namespace clang |
| 96 | |
| 97 | #endif |