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/LangOptions.h"
21 #include "clang/Basic/SourceLocation.h"
22 #include "clang/Basic/SourceManager.h"
23 #include "clang/Format/Format.h"
24 #include "llvm/ADT/MapVector.h"
25 #include "llvm/ADT/SmallPtrSet.h"
26 #include "llvm/ADT/StringSet.h"
27 #include "llvm/Support/Regex.h"
28 
29 #include <stack>
30 
31 namespace clang {
32 namespace format {
33 
34 enum LexerState {
35   NORMAL,
36   TEMPLATE_STRING,
37   TOKEN_STASHED,
38 };
39 
40 class FormatTokenLexer {
41 public:
42   FormatTokenLexer(const SourceManager &SourceMgr, FileID ID, unsigned Column,
43                    const FormatStyle &Style, encoding::Encoding Encoding,
44                    llvm::SpecificBumpPtrAllocator<FormatToken> &Allocator,
45                    IdentifierTable &IdentTable);
46 
47   ArrayRef<FormatToken *> lex();
48 
getKeywords()49   const AdditionalKeywords &getKeywords() { return Keywords; }
50 
51 private:
52   void tryMergePreviousTokens();
53 
54   bool tryMergeLessLess();
55   bool tryMergeGreaterGreater();
56   bool tryMergeNSStringLiteral();
57   bool tryMergeJSPrivateIdentifier();
58   bool tryMergeCSharpStringLiteral();
59   bool tryMergeCSharpKeywordVariables();
60   bool tryMergeNullishCoalescingEqual();
61   bool tryTransformCSharpForEach();
62   bool tryMergeForEach();
63   bool tryTransformTryUsageForC();
64 
65   // Merge the most recently lexed tokens into a single token if their kinds are
66   // correct.
67   bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType);
68   // Merge without checking their kinds.
69   bool tryMergeTokens(size_t Count, TokenType NewType);
70   // Merge if their kinds match any one of Kinds.
71   bool tryMergeTokensAny(ArrayRef<ArrayRef<tok::TokenKind>> Kinds,
72                          TokenType NewType);
73 
74   // Returns \c true if \p Tok can only be followed by an operand in JavaScript.
75   bool precedesOperand(FormatToken *Tok);
76 
77   bool canPrecedeRegexLiteral(FormatToken *Prev);
78 
79   // Tries to parse a JavaScript Regex literal starting at the current token,
80   // if that begins with a slash and is in a location where JavaScript allows
81   // regex literals. Changes the current token to a regex literal and updates
82   // its text if successful.
83   void tryParseJSRegexLiteral();
84 
85   // Handles JavaScript template strings.
86   //
87   // JavaScript template strings use backticks ('`') as delimiters, and allow
88   // embedding expressions nested in ${expr-here}. Template strings can be
89   // nested recursively, i.e. expressions can contain template strings in turn.
90   //
91   // The code below parses starting from a backtick, up to a closing backtick or
92   // an opening ${. It also maintains a stack of lexing contexts to handle
93   // nested template parts by balancing curly braces.
94   void handleTemplateStrings();
95 
96   void handleCSharpVerbatimAndInterpolatedStrings();
97 
98   // Handles TableGen multiline strings. It has the form [{ ... }].
99   void handleTableGenMultilineString();
100   // Handles TableGen numeric like identifiers.
101   // They have a forms of [0-9]*[_a-zA-Z]([_a-zA-Z0-9]*). But limited to the
102   // case it is not lexed as an integer.
103   void handleTableGenNumericLikeIdentifier();
104 
105   void tryParsePythonComment();
106 
107   bool tryMerge_TMacro();
108 
109   bool tryMergeConflictMarkers();
110 
111   void truncateToken(size_t NewLen);
112 
113   FormatToken *getStashedToken();
114 
115   FormatToken *getNextToken();
116 
117   FormatToken *FormatTok;
118   bool IsFirstToken;
119   std::stack<LexerState> StateStack;
120   unsigned Column;
121   unsigned TrailingWhitespace;
122   std::unique_ptr<Lexer> Lex;
123   LangOptions LangOpts;
124   const SourceManager &SourceMgr;
125   FileID ID;
126   const FormatStyle &Style;
127   IdentifierTable &IdentTable;
128   AdditionalKeywords Keywords;
129   encoding::Encoding Encoding;
130   llvm::SpecificBumpPtrAllocator<FormatToken> &Allocator;
131   // Index (in 'Tokens') of the last token that starts a new line.
132   unsigned FirstInLineIndex;
133   SmallVector<FormatToken *, 16> Tokens;
134 
135   llvm::SmallMapVector<IdentifierInfo *, TokenType, 8> Macros;
136 
137   llvm::SmallPtrSet<IdentifierInfo *, 8> TypeNames;
138 
139   bool FormattingDisabled;
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