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_LIB_FORMAT_UNWRAPPEDLINEPARSER_H
17 #define LLVM_CLANG_LIB_FORMAT_UNWRAPPEDLINEPARSER_H
18
19 #include "FormatToken.h"
20 #include "clang/Basic/IdentifierTable.h"
21 #include "clang/Format/Format.h"
22 #include <list>
23 #include <stack>
24
25 namespace clang {
26 namespace format {
27
28 struct UnwrappedLineNode;
29
30 /// \brief An unwrapped line is a sequence of \c Token, that we would like to
31 /// put on a single line if there was no column limit.
32 ///
33 /// This is used as a main interface between the \c UnwrappedLineParser and the
34 /// \c UnwrappedLineFormatter. The key property is that changing the formatting
35 /// within an unwrapped line does not affect any other unwrapped lines.
36 struct UnwrappedLine {
37 UnwrappedLine();
38
39 // FIXME: Don't use std::list here.
40 /// \brief The \c Tokens comprising this \c UnwrappedLine.
41 std::list<UnwrappedLineNode> Tokens;
42
43 /// \brief The indent level of the \c UnwrappedLine.
44 unsigned Level;
45
46 /// \brief Whether this \c UnwrappedLine is part of a preprocessor directive.
47 bool InPPDirective;
48
49 bool MustBeDeclaration;
50 };
51
52 class UnwrappedLineConsumer {
53 public:
~UnwrappedLineConsumer()54 virtual ~UnwrappedLineConsumer() {}
55 virtual void consumeUnwrappedLine(const UnwrappedLine &Line) = 0;
56 virtual void finishRun() = 0;
57 };
58
59 class FormatTokenSource;
60
61 class UnwrappedLineParser {
62 public:
63 UnwrappedLineParser(const FormatStyle &Style,
64 const AdditionalKeywords &Keywords,
65 ArrayRef<FormatToken *> Tokens,
66 UnwrappedLineConsumer &Callback);
67
68 void parse();
69
70 private:
71 void reset();
72 void parseFile();
73 void parseLevel(bool HasOpeningBrace);
74 void parseBlock(bool MustBeDeclaration, bool AddLevel = true,
75 bool MunchSemi = true);
76 void parseChildBlock();
77 void parsePPDirective();
78 void parsePPDefine();
79 void parsePPIf(bool IfDef);
80 void parsePPElIf();
81 void parsePPElse();
82 void parsePPEndIf();
83 void parsePPUnknown();
84 void parseStructuralElement();
85 bool tryToParseBracedList();
86 bool parseBracedList(bool ContinueOnSemicolons = false);
87 void parseParens();
88 void parseSquare();
89 void parseIfThenElse();
90 void parseTryCatch();
91 void parseForOrWhileLoop();
92 void parseDoWhile();
93 void parseLabel();
94 void parseCaseLabel();
95 void parseSwitch();
96 void parseNamespace();
97 void parseNew();
98 void parseAccessSpecifier();
99 void parseEnum();
100 void parseJavaEnumBody();
101 void parseRecord();
102 void parseObjCProtocolList();
103 void parseObjCUntilAtEnd();
104 void parseObjCInterfaceOrImplementation();
105 void parseObjCProtocol();
106 void parseJavaScriptEs6ImportExport();
107 bool tryToParseLambda();
108 bool tryToParseLambdaIntroducer();
109 void tryToParseJSFunction();
110 void addUnwrappedLine();
111 bool eof() const;
112 void nextToken();
113 void readToken();
114 void flushComments(bool NewlineBeforeNext);
115 void pushToken(FormatToken *Tok);
116 void calculateBraceTypes(bool ExpectClassBody = false);
117
118 // Marks a conditional compilation edge (for example, an '#if', '#ifdef',
119 // '#else' or merge conflict marker). If 'Unreachable' is true, assumes
120 // this branch either cannot be taken (for example '#if false'), or should
121 // not be taken in this round.
122 void conditionalCompilationCondition(bool Unreachable);
123 void conditionalCompilationStart(bool Unreachable);
124 void conditionalCompilationAlternative();
125 void conditionalCompilationEnd();
126
127 bool isOnNewLine(const FormatToken &FormatTok);
128
129 // FIXME: We are constantly running into bugs where Line.Level is incorrectly
130 // subtracted from beyond 0. Introduce a method to subtract from Line.Level
131 // and use that everywhere in the Parser.
132 std::unique_ptr<UnwrappedLine> Line;
133
134 // Comments are sorted into unwrapped lines by whether they are in the same
135 // line as the previous token, or not. If not, they belong to the next token.
136 // Since the next token might already be in a new unwrapped line, we need to
137 // store the comments belonging to that token.
138 SmallVector<FormatToken *, 1> CommentsBeforeNextToken;
139 FormatToken *FormatTok;
140 bool MustBreakBeforeNextToken;
141
142 // The parsed lines. Only added to through \c CurrentLines.
143 SmallVector<UnwrappedLine, 8> Lines;
144
145 // Preprocessor directives are parsed out-of-order from other unwrapped lines.
146 // Thus, we need to keep a list of preprocessor directives to be reported
147 // after an unwarpped line that has been started was finished.
148 SmallVector<UnwrappedLine, 4> PreprocessorDirectives;
149
150 // New unwrapped lines are added via CurrentLines.
151 // Usually points to \c &Lines. While parsing a preprocessor directive when
152 // there is an unfinished previous unwrapped line, will point to
153 // \c &PreprocessorDirectives.
154 SmallVectorImpl<UnwrappedLine> *CurrentLines;
155
156 // We store for each line whether it must be a declaration depending on
157 // whether we are in a compound statement or not.
158 std::vector<bool> DeclarationScopeStack;
159
160 const FormatStyle &Style;
161 const AdditionalKeywords &Keywords;
162
163 FormatTokenSource *Tokens;
164 UnwrappedLineConsumer &Callback;
165
166 // FIXME: This is a temporary measure until we have reworked the ownership
167 // of the format tokens. The goal is to have the actual tokens created and
168 // owned outside of and handed into the UnwrappedLineParser.
169 ArrayRef<FormatToken *> AllTokens;
170
171 // Represents preprocessor branch type, so we can find matching
172 // #if/#else/#endif directives.
173 enum PPBranchKind {
174 PP_Conditional, // Any #if, #ifdef, #ifndef, #elif, block outside #if 0
175 PP_Unreachable // #if 0 or a conditional preprocessor block inside #if 0
176 };
177
178 // Keeps a stack of currently active preprocessor branching directives.
179 SmallVector<PPBranchKind, 16> PPStack;
180
181 // The \c UnwrappedLineParser re-parses the code for each combination
182 // of preprocessor branches that can be taken.
183 // To that end, we take the same branch (#if, #else, or one of the #elif
184 // branches) for each nesting level of preprocessor branches.
185 // \c PPBranchLevel stores the current nesting level of preprocessor
186 // branches during one pass over the code.
187 int PPBranchLevel;
188
189 // Contains the current branch (#if, #else or one of the #elif branches)
190 // for each nesting level.
191 SmallVector<int, 8> PPLevelBranchIndex;
192
193 // Contains the maximum number of branches at each nesting level.
194 SmallVector<int, 8> PPLevelBranchCount;
195
196 // Contains the number of branches per nesting level we are currently
197 // in while parsing a preprocessor branch sequence.
198 // This is used to update PPLevelBranchCount at the end of a branch
199 // sequence.
200 std::stack<int> PPChainBranchIndex;
201
202 friend class ScopedLineState;
203 friend class CompoundStatementIndenter;
204 };
205
206 struct UnwrappedLineNode {
UnwrappedLineNodeUnwrappedLineNode207 UnwrappedLineNode() : Tok(nullptr) {}
UnwrappedLineNodeUnwrappedLineNode208 UnwrappedLineNode(FormatToken *Tok) : Tok(Tok) {}
209
210 FormatToken *Tok;
211 SmallVector<UnwrappedLine, 0> Children;
212 };
213
UnwrappedLine()214 inline UnwrappedLine::UnwrappedLine()
215 : Level(0), InPPDirective(false), MustBeDeclaration(false) {}
216
217 } // end namespace format
218 } // end namespace clang
219
220 #endif
221