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 #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,
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 
51   bool tryMergeTokens(ArrayRef<tok::TokenKind> Kinds, TokenType NewType);
52 
53   // Returns \c true if \p Tok can only be followed by an operand in JavaScript.
54   bool precedesOperand(FormatToken *Tok);
55 
56   bool canPrecedeRegexLiteral(FormatToken *Prev);
57 
58   // Tries to parse a JavaScript Regex literal starting at the current token,
59   // if that begins with a slash and is in a location where JavaScript allows
60   // regex literals. Changes the current token to a regex literal and updates
61   // its text if successful.
62   void tryParseJSRegexLiteral();
63 
64   // Handles JavaScript template strings.
65   //
66   // JavaScript template strings use backticks ('`') as delimiters, and allow
67   // embedding expressions nested in ${expr-here}. Template strings can be
68   // nested recursively, i.e. expressions can contain template strings in turn.
69   //
70   // The code below parses starting from a backtick, up to a closing backtick or
71   // an opening ${. It also maintains a stack of lexing contexts to handle
72   // nested template parts by balancing curly braces.
73   void handleTemplateStrings();
74 
75   bool tryMerge_TMacro();
76 
77   bool tryMergeConflictMarkers();
78 
79   FormatToken *getStashedToken();
80 
81   FormatToken *getNextToken();
82 
83   FormatToken *FormatTok;
84   bool IsFirstToken;
85   std::stack<LexerState> StateStack;
86   unsigned Column;
87   unsigned TrailingWhitespace;
88   std::unique_ptr<Lexer> Lex;
89   const SourceManager &SourceMgr;
90   FileID ID;
91   const FormatStyle &Style;
92   IdentifierTable IdentTable;
93   AdditionalKeywords Keywords;
94   encoding::Encoding Encoding;
95   llvm::SpecificBumpPtrAllocator<FormatToken> Allocator;
96   // Index (in 'Tokens') of the last token that starts a new line.
97   unsigned FirstInLineIndex;
98   SmallVector<FormatToken *, 16> Tokens;
99   SmallVector<IdentifierInfo *, 8> ForEachMacros;
100 
101   bool FormattingDisabled;
102 
103   llvm::Regex MacroBlockBeginRegex;
104   llvm::Regex MacroBlockEndRegex;
105 
106   void readRawToken(FormatToken &Tok);
107 
108   void resetLexer(unsigned Offset);
109 };
110 
111 } // namespace format
112 } // namespace clang
113 
114 #endif
115