1 //===--- UnwrappedLineParser.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 the declaration of the UnwrappedLineParser,
12 /// which turns a stream of tokens into UnwrappedLines.
13 ///
14 //===----------------------------------------------------------------------===//
15 
16 #ifndef LLVM_CLANG_FORMAT_UNWRAPPED_LINE_PARSER_H
17 #define LLVM_CLANG_FORMAT_UNWRAPPED_LINE_PARSER_H
18 
19 #include "clang/Basic/IdentifierTable.h"
20 #include "clang/Basic/SourceManager.h"
21 #include "clang/Format/Format.h"
22 #include "clang/Lex/Lexer.h"
23 #include <list>
24 
25 namespace clang {
26 
27 class DiagnosticsEngine;
28 
29 namespace format {
30 
31 /// \brief A wrapper around a \c Token storing information about the
32 /// whitespace characters preceeding it.
33 struct FormatToken {
34   FormatToken()
35       : NewlinesBefore(0), HasUnescapedNewline(false), WhiteSpaceLength(0),
36         LastNewlineOffset(0), TokenLength(0), IsFirst(false),
37         MustBreakBefore(false) {}
38 
39   /// \brief The \c Token.
40   Token Tok;
41 
42   /// \brief The number of newlines immediately before the \c Token.
43   ///
44   /// This can be used to determine what the user wrote in the original code
45   /// and thereby e.g. leave an empty line between two function definitions.
46   unsigned NewlinesBefore;
47 
48   /// \brief Whether there is at least one unescaped newline before the \c
49   /// Token.
50   bool HasUnescapedNewline;
51 
52   /// \brief The location of the start of the whitespace immediately preceeding
53   /// the \c Token.
54   ///
55   /// Used together with \c WhiteSpaceLength to create a \c Replacement.
56   SourceLocation WhiteSpaceStart;
57 
58   /// \brief The length in characters of the whitespace immediately preceeding
59   /// the \c Token.
60   unsigned WhiteSpaceLength;
61 
62   /// \brief The offset just past the last '\n' in this token's leading
63   /// whitespace (relative to \c WhiteSpaceStart). 0 if there is no '\n'.
64   unsigned LastNewlineOffset;
65 
66   /// \brief The length of the non-whitespace parts of the token. This is
67   /// necessary because we need to handle escaped newlines that are stored
68   /// with the token.
69   unsigned TokenLength;
70 
71   /// \brief Indicates that this is the first token.
72   bool IsFirst;
73 
74   /// \brief Whether there must be a line break before this token.
75   ///
76   /// This happens for example when a preprocessor directive ended directly
77   /// before the token.
78   bool MustBreakBefore;
79 };
80 
81 /// \brief An unwrapped line is a sequence of \c Token, that we would like to
82 /// put on a single line if there was no column limit.
83 ///
84 /// This is used as a main interface between the \c UnwrappedLineParser and the
85 /// \c UnwrappedLineFormatter. The key property is that changing the formatting
86 /// within an unwrapped line does not affect any other unwrapped lines.
87 struct UnwrappedLine {
88   UnwrappedLine() : Level(0), InPPDirective(false), MustBeDeclaration(false) {
89   }
90 
91   // FIXME: Don't use std::list here.
92   /// \brief The \c Tokens comprising this \c UnwrappedLine.
93   std::list<FormatToken> Tokens;
94 
95   /// \brief The indent level of the \c UnwrappedLine.
96   unsigned Level;
97 
98   /// \brief Whether this \c UnwrappedLine is part of a preprocessor directive.
99   bool InPPDirective;
100 
101   bool MustBeDeclaration;
102 };
103 
104 class UnwrappedLineConsumer {
105 public:
106   virtual ~UnwrappedLineConsumer() {
107   }
108   virtual void consumeUnwrappedLine(const UnwrappedLine &Line) = 0;
109 };
110 
111 class FormatTokenSource {
112 public:
113   virtual ~FormatTokenSource() {
114   }
115   virtual FormatToken getNextToken() = 0;
116 };
117 
118 class UnwrappedLineParser {
119 public:
120   UnwrappedLineParser(clang::DiagnosticsEngine &Diag, const FormatStyle &Style,
121                       FormatTokenSource &Tokens,
122                       UnwrappedLineConsumer &Callback);
123 
124   /// Returns true in case of a structural error.
125   bool parse();
126 
127 private:
128   bool parseFile();
129   bool parseLevel(bool HasOpeningBrace);
130   bool parseBlock(bool MustBeDeclaration, unsigned AddLevels = 1);
131   void parsePPDirective();
132   void parsePPDefine();
133   void parsePPUnknown();
134   void parseStructuralElement();
135   void parseBracedList();
136   void parseReturn();
137   void parseParens();
138   void parseIfThenElse();
139   void parseForOrWhileLoop();
140   void parseDoWhile();
141   void parseLabel();
142   void parseCaseLabel();
143   void parseSwitch();
144   void parseNamespace();
145   void parseAccessSpecifier();
146   void parseEnum();
147   void parseRecord();
148   void parseObjCProtocolList();
149   void parseObjCUntilAtEnd();
150   void parseObjCInterfaceOrImplementation();
151   void parseObjCProtocol();
152   void addUnwrappedLine();
153   bool eof() const;
154   void nextToken();
155   void readToken();
156   void flushComments(bool NewlineBeforeNext);
157   void pushToken(const FormatToken &Tok);
158 
159   // FIXME: We are constantly running into bugs where Line.Level is incorrectly
160   // subtracted from beyond 0. Introduce a method to subtract from Line.Level
161   // and use that everywhere in the Parser.
162   OwningPtr<UnwrappedLine> Line;
163 
164   // Comments are sorted into unwrapped lines by whether they are in the same
165   // line as the previous token, or not. If not, they belong to the next token.
166   // Since the next token might already be in a new unwrapped line, we need to
167   // store the comments belonging to that token.
168   SmallVector<FormatToken, 1> CommentsBeforeNextToken;
169   FormatToken FormatTok;
170   bool MustBreakBeforeNextToken;
171 
172   // The parsed lines. Only added to through \c CurrentLines.
173   std::vector<UnwrappedLine> Lines;
174 
175   // Preprocessor directives are parsed out-of-order from other unwrapped lines.
176   // Thus, we need to keep a list of preprocessor directives to be reported
177   // after an unwarpped line that has been started was finished.
178   std::vector<UnwrappedLine> PreprocessorDirectives;
179 
180   // New unwrapped lines are added via CurrentLines.
181   // Usually points to \c &Lines. While parsing a preprocessor directive when
182   // there is an unfinished previous unwrapped line, will point to
183   // \c &PreprocessorDirectives.
184   std::vector<UnwrappedLine> *CurrentLines;
185 
186   // We store for each line whether it must be a declaration depending on
187   // whether we are in a compound statement or not.
188   std::vector<bool> DeclarationScopeStack;
189 
190   clang::DiagnosticsEngine &Diag;
191   const FormatStyle &Style;
192   FormatTokenSource *Tokens;
193   UnwrappedLineConsumer &Callback;
194 
195   friend class ScopedLineState;
196 };
197 
198 } // end namespace format
199 } // end namespace clang
200 
201 #endif // LLVM_CLANG_FORMAT_UNWRAPPED_LINE_PARSER_H
202