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/ADT/MapVector.h"
24 #include "llvm/ADT/StringSet.h"
25 #include "llvm/Support/Regex.h"
26 
27 #include <stack>
28 
29 namespace clang {
30 namespace format {
31 
32 enum LexerState {
33   NORMAL,
34   TEMPLATE_STRING,
35   TOKEN_STASHED,
36 };
37 
38 class FormatTokenLexer {
39 public:
40   FormatTokenLexer(const SourceManager &SourceMgr, FileID ID, unsigned Column,
41                    const FormatStyle &Style, encoding::Encoding Encoding);
42 
43   ArrayRef<FormatToken *> lex();
44 
45   const AdditionalKeywords &getKeywords() { return Keywords; }
46 
47 private:
48   void tryMergePreviousTokens();
49 
50   bool tryMergeLessLess();
51   bool tryMergeNSStringLiteral();
52   bool tryMergeJSPrivateIdentifier();
53   bool tryMergeCSharpStringLiteral();
54   bool tryMergeCSharpKeywordVariables();
55   bool tryMergeCSharpNullConditionals();
56   bool tryMergeCSharpDoubleQuestion();
57   bool tryTransformCSharpForEach();
58   bool tryMergeCSharpAttributeAndTarget();
59 
60   bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType);
61 
62   // Returns \c true if \p Tok can only be followed by an operand in JavaScript.
63   bool precedesOperand(FormatToken *Tok);
64 
65   bool canPrecedeRegexLiteral(FormatToken *Prev);
66 
67   // Tries to parse a JavaScript Regex literal starting at the current token,
68   // if that begins with a slash and is in a location where JavaScript allows
69   // regex literals. Changes the current token to a regex literal and updates
70   // its text if successful.
71   void tryParseJSRegexLiteral();
72 
73   // Handles JavaScript template strings.
74   //
75   // JavaScript template strings use backticks ('`') as delimiters, and allow
76   // embedding expressions nested in ${expr-here}. Template strings can be
77   // nested recursively, i.e. expressions can contain template strings in turn.
78   //
79   // The code below parses starting from a backtick, up to a closing backtick or
80   // an opening ${. It also maintains a stack of lexing contexts to handle
81   // nested template parts by balancing curly braces.
82   void handleTemplateStrings();
83 
84   void handleCSharpVerbatimAndInterpolatedStrings();
85 
86   void tryParsePythonComment();
87 
88   bool tryMerge_TMacro();
89 
90   bool tryMergeConflictMarkers();
91 
92   FormatToken *getStashedToken();
93 
94   FormatToken *getNextToken();
95 
96   FormatToken *FormatTok;
97   bool IsFirstToken;
98   std::stack<LexerState> StateStack;
99   unsigned Column;
100   unsigned TrailingWhitespace;
101   std::unique_ptr<Lexer> Lex;
102   const SourceManager &SourceMgr;
103   FileID ID;
104   const FormatStyle &Style;
105   IdentifierTable IdentTable;
106   AdditionalKeywords Keywords;
107   encoding::Encoding Encoding;
108   llvm::SpecificBumpPtrAllocator<FormatToken> Allocator;
109   // Index (in 'Tokens') of the last token that starts a new line.
110   unsigned FirstInLineIndex;
111   SmallVector<FormatToken *, 16> Tokens;
112 
113   llvm::SmallMapVector<IdentifierInfo *, TokenType, 8> Macros;
114 
115   bool FormattingDisabled;
116 
117   llvm::Regex MacroBlockBeginRegex;
118   llvm::Regex MacroBlockEndRegex;
119 
120   // Targets that may appear inside a C# attribute.
121   static const llvm::StringSet<> CSharpAttributeTargets;
122 
123   void readRawToken(FormatToken &Tok);
124 
125   void resetLexer(unsigned Offset);
126 };
127 
128 } // namespace format
129 } // namespace clang
130 
131 #endif
132