xref: /freebsd/contrib/llvm-project/clang/lib/Format/FormatTokenLexer.h (revision 0fca6ea1d4eea4c934cfff25ac9ee8ad6fe95583)
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 tryMergeNSStringLiteral();
52   bool tryMergeJSPrivateIdentifier();
53   bool tryMergeCSharpStringLiteral();
54   bool tryMergeCSharpKeywordVariables();
55   bool tryMergeNullishCoalescingEqual();
56   bool tryTransformCSharpForEach();
57   bool tryMergeForEach();
58   bool tryTransformTryUsageForC();
59 
60   // Merge the most recently lexed tokens into a single token if their kinds are
61   // correct.
62   bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType);
63   // Merge without checking their kinds.
64   bool tryMergeTokens(size_t Count, TokenType NewType);
65   // Merge if their kinds match any one of Kinds.
66   bool tryMergeTokensAny(ArrayRef<ArrayRef<tok::TokenKind>> Kinds,
67                          TokenType NewType);
68 
69   // Returns \c true if \p Tok can only be followed by an operand in JavaScript.
70   bool precedesOperand(FormatToken *Tok);
71 
72   bool canPrecedeRegexLiteral(FormatToken *Prev);
73 
74   // Tries to parse a JavaScript Regex literal starting at the current token,
75   // if that begins with a slash and is in a location where JavaScript allows
76   // regex literals. Changes the current token to a regex literal and updates
77   // its text if successful.
78   void tryParseJSRegexLiteral();
79 
80   // Handles JavaScript template strings.
81   //
82   // JavaScript template strings use backticks ('`') as delimiters, and allow
83   // embedding expressions nested in ${expr-here}. Template strings can be
84   // nested recursively, i.e. expressions can contain template strings in turn.
85   //
86   // The code below parses starting from a backtick, up to a closing backtick or
87   // an opening ${. It also maintains a stack of lexing contexts to handle
88   // nested template parts by balancing curly braces.
89   void handleTemplateStrings();
90 
91   void handleCSharpVerbatimAndInterpolatedStrings();
92 
93   // Handles TableGen multiline strings. It has the form [{ ... }].
94   void handleTableGenMultilineString();
95   // Handles TableGen numeric like identifiers.
96   // They have a forms of [0-9]*[_a-zA-Z]([_a-zA-Z0-9]*). But limited to the
97   // case it is not lexed as an integer.
98   void handleTableGenNumericLikeIdentifier();
99 
100   void tryParsePythonComment();
101 
102   bool tryMerge_TMacro();
103 
104   bool tryMergeConflictMarkers();
105 
106   void truncateToken(size_t NewLen);
107 
108   FormatToken *getStashedToken();
109 
110   FormatToken *getNextToken();
111 
112   FormatToken *FormatTok;
113   bool IsFirstToken;
114   std::stack<LexerState> StateStack;
115   unsigned Column;
116   unsigned TrailingWhitespace;
117   std::unique_ptr<Lexer> Lex;
118   LangOptions LangOpts;
119   const SourceManager &SourceMgr;
120   FileID ID;
121   const FormatStyle &Style;
122   IdentifierTable &IdentTable;
123   AdditionalKeywords Keywords;
124   encoding::Encoding Encoding;
125   llvm::SpecificBumpPtrAllocator<FormatToken> &Allocator;
126   // Index (in 'Tokens') of the last token that starts a new line.
127   unsigned FirstInLineIndex;
128   SmallVector<FormatToken *, 16> Tokens;
129 
130   llvm::SmallMapVector<IdentifierInfo *, TokenType, 8> Macros;
131 
132   llvm::SmallPtrSet<IdentifierInfo *, 8> TypeNames;
133 
134   bool FormattingDisabled;
135 
136   llvm::Regex MacroBlockBeginRegex;
137   llvm::Regex MacroBlockEndRegex;
138 
139   // Targets that may appear inside a C# attribute.
140   static const llvm::StringSet<> CSharpAttributeTargets;
141 
142   /// Handle Verilog-specific tokens.
143   bool readRawTokenVerilogSpecific(Token &Tok);
144 
145   void readRawToken(FormatToken &Tok);
146 
147   void resetLexer(unsigned Offset);
148 };
149 
150 } // namespace format
151 } // namespace clang
152 
153 #endif
154