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 "clang/Basic/SourceLocation.h"
21 #include "clang/Basic/SourceManager.h"
22 #include "clang/Format/Format.h"
23 #include "llvm/Support/Regex.h"
24 #include "llvm/ADT/MapVector.h"
25 
26 #include <stack>
27 
28 namespace clang {
29 namespace format {
30 
31 enum LexerState {
32   NORMAL,
33   TEMPLATE_STRING,
34   TOKEN_STASHED,
35 };
36 
37 class FormatTokenLexer {
38 public:
39   FormatTokenLexer(const SourceManager &SourceMgr, FileID ID, unsigned Column,
40                    const FormatStyle &Style, encoding::Encoding Encoding);
41 
42   ArrayRef<FormatToken *> lex();
43 
44   const AdditionalKeywords &getKeywords() { return Keywords; }
45 
46 private:
47   void tryMergePreviousTokens();
48 
49   bool tryMergeLessLess();
50   bool tryMergeNSStringLiteral();
51 
52   bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType);
53 
54   // Returns \c true if \p Tok can only be followed by an operand in JavaScript.
55   bool precedesOperand(FormatToken *Tok);
56 
57   bool canPrecedeRegexLiteral(FormatToken *Prev);
58 
59   // Tries to parse a JavaScript Regex literal starting at the current token,
60   // if that begins with a slash and is in a location where JavaScript allows
61   // regex literals. Changes the current token to a regex literal and updates
62   // its text if successful.
63   void tryParseJSRegexLiteral();
64 
65   // Handles JavaScript template strings.
66   //
67   // JavaScript template strings use backticks ('`') as delimiters, and allow
68   // embedding expressions nested in ${expr-here}. Template strings can be
69   // nested recursively, i.e. expressions can contain template strings in turn.
70   //
71   // The code below parses starting from a backtick, up to a closing backtick or
72   // an opening ${. It also maintains a stack of lexing contexts to handle
73   // nested template parts by balancing curly braces.
74   void handleTemplateStrings();
75 
76   void tryParsePythonComment();
77 
78   bool tryMerge_TMacro();
79 
80   bool tryMergeConflictMarkers();
81 
82   FormatToken *getStashedToken();
83 
84   FormatToken *getNextToken();
85 
86   FormatToken *FormatTok;
87   bool IsFirstToken;
88   std::stack<LexerState> StateStack;
89   unsigned Column;
90   unsigned TrailingWhitespace;
91   std::unique_ptr<Lexer> Lex;
92   const SourceManager &SourceMgr;
93   FileID ID;
94   const FormatStyle &Style;
95   IdentifierTable IdentTable;
96   AdditionalKeywords Keywords;
97   encoding::Encoding Encoding;
98   llvm::SpecificBumpPtrAllocator<FormatToken> Allocator;
99   // Index (in 'Tokens') of the last token that starts a new line.
100   unsigned FirstInLineIndex;
101   SmallVector<FormatToken *, 16> Tokens;
102 
103   llvm::SmallMapVector<IdentifierInfo *, TokenType, 8> Macros;
104 
105   bool FormattingDisabled;
106 
107   llvm::Regex MacroBlockBeginRegex;
108   llvm::Regex MacroBlockEndRegex;
109 
110   void readRawToken(FormatToken &Tok);
111 
112   void resetLexer(unsigned Offset);
113 };
114 
115 } // namespace format
116 } // namespace clang
117 
118 #endif
119