1 //===--- UnwrappedLineParser.h - Format C++ code ----------------*- C++ -*-===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
11 /// \brief This file contains the declaration of the UnwrappedLineParser,
12 /// which turns a stream of tokens into UnwrappedLines.
14 //===----------------------------------------------------------------------===//
16 #ifndef LLVM_CLANG_LIB_FORMAT_UNWRAPPEDLINEPARSER_H
17 #define LLVM_CLANG_LIB_FORMAT_UNWRAPPEDLINEPARSER_H
19 #include "FormatToken.h"
20 #include "clang/Basic/IdentifierTable.h"
21 #include "clang/Format/Format.h"
22 #include "llvm/Support/Regex.h"
29 struct UnwrappedLineNode;
31 /// \brief An unwrapped line is a sequence of \c Token, that we would like to
32 /// put on a single line if there was no column limit.
34 /// This is used as a main interface between the \c UnwrappedLineParser and the
35 /// \c UnwrappedLineFormatter. The key property is that changing the formatting
36 /// within an unwrapped line does not affect any other unwrapped lines.
37 struct UnwrappedLine {
40 // FIXME: Don't use std::list here.
41 /// \brief The \c Tokens comprising this \c UnwrappedLine.
42 std::list<UnwrappedLineNode> Tokens;
44 /// \brief The indent level of the \c UnwrappedLine.
47 /// \brief Whether this \c UnwrappedLine is part of a preprocessor directive.
50 bool MustBeDeclaration;
52 /// \brief If this \c UnwrappedLine closes a block in a sequence of lines,
53 /// \c MatchingOpeningBlockLineIndex stores the index of the corresponding
54 /// opening line. Otherwise, \c MatchingOpeningBlockLineIndex must be
56 size_t MatchingOpeningBlockLineIndex;
58 static const size_t kInvalidIndex = -1;
61 class UnwrappedLineConsumer {
63 virtual ~UnwrappedLineConsumer() {}
64 virtual void consumeUnwrappedLine(const UnwrappedLine &Line) = 0;
65 virtual void finishRun() = 0;
68 class FormatTokenSource;
70 class UnwrappedLineParser {
72 UnwrappedLineParser(const FormatStyle &Style,
73 const AdditionalKeywords &Keywords,
74 ArrayRef<FormatToken *> Tokens,
75 UnwrappedLineConsumer &Callback);
82 void parseLevel(bool HasOpeningBrace);
83 void parseBlock(bool MustBeDeclaration, bool AddLevel = true,
84 bool MunchSemi = true);
85 void parseChildBlock();
86 void parsePPDirective();
88 void parsePPIf(bool IfDef);
92 void parsePPUnknown();
93 void readTokenWithJavaScriptASI();
94 void parseStructuralElement();
95 bool tryToParseBracedList();
96 bool parseBracedList(bool ContinueOnSemicolons = false,
97 tok::TokenKind ClosingBraceKind = tok::r_brace);
100 void parseIfThenElse();
101 void parseTryCatch();
102 void parseForOrWhileLoop();
105 void parseCaseLabel();
107 void parseNamespace();
109 void parseAccessSpecifier();
111 void parseJavaEnumBody();
112 // Parses a record (aka class) as a top level element. If ParseAsExpr is true,
113 // parses the record as a child block, i.e. if the class declaration is an
115 void parseRecord(bool ParseAsExpr = false);
116 void parseObjCProtocolList();
117 void parseObjCUntilAtEnd();
118 void parseObjCInterfaceOrImplementation();
119 void parseObjCProtocol();
120 void parseJavaScriptEs6ImportExport();
121 bool tryToParseLambda();
122 bool tryToParseLambdaIntroducer();
123 void tryToParseJSFunction();
124 void addUnwrappedLine();
127 const FormatToken *getPreviousToken();
130 // Decides which comment tokens should be added to the current line and which
131 // should be added as comments before the next token.
133 // Comments specifies the sequence of comment tokens to analyze. They get
134 // either pushed to the current line or added to the comments before the next
137 // NextTok specifies the next token. A null pointer NextTok is supported, and
138 // signifies either the absense of a next token, or that the next token
139 // shouldn't be taken into accunt for the analysis.
140 void distributeComments(const SmallVectorImpl<FormatToken *> &Comments,
141 const FormatToken *NextTok);
143 // Adds the comment preceding the next token to unwrapped lines.
144 void flushComments(bool NewlineBeforeNext);
145 void pushToken(FormatToken *Tok);
146 void calculateBraceTypes(bool ExpectClassBody = false);
148 // Marks a conditional compilation edge (for example, an '#if', '#ifdef',
149 // '#else' or merge conflict marker). If 'Unreachable' is true, assumes
150 // this branch either cannot be taken (for example '#if false'), or should
151 // not be taken in this round.
152 void conditionalCompilationCondition(bool Unreachable);
153 void conditionalCompilationStart(bool Unreachable);
154 void conditionalCompilationAlternative();
155 void conditionalCompilationEnd();
157 bool isOnNewLine(const FormatToken &FormatTok);
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 std::unique_ptr<UnwrappedLine> Line;
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;
172 // The parsed lines. Only added to through \c CurrentLines.
173 SmallVector<UnwrappedLine, 8> Lines;
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 SmallVector<UnwrappedLine, 4> PreprocessorDirectives;
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 SmallVectorImpl<UnwrappedLine> *CurrentLines;
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;
190 const FormatStyle &Style;
191 const AdditionalKeywords &Keywords;
193 llvm::Regex CommentPragmasRegex;
195 FormatTokenSource *Tokens;
196 UnwrappedLineConsumer &Callback;
198 // FIXME: This is a temporary measure until we have reworked the ownership
199 // of the format tokens. The goal is to have the actual tokens created and
200 // owned outside of and handed into the UnwrappedLineParser.
201 ArrayRef<FormatToken *> AllTokens;
203 // Represents preprocessor branch type, so we can find matching
204 // #if/#else/#endif directives.
206 PP_Conditional, // Any #if, #ifdef, #ifndef, #elif, block outside #if 0
207 PP_Unreachable // #if 0 or a conditional preprocessor block inside #if 0
210 // Keeps a stack of currently active preprocessor branching directives.
211 SmallVector<PPBranchKind, 16> PPStack;
213 // The \c UnwrappedLineParser re-parses the code for each combination
214 // of preprocessor branches that can be taken.
215 // To that end, we take the same branch (#if, #else, or one of the #elif
216 // branches) for each nesting level of preprocessor branches.
217 // \c PPBranchLevel stores the current nesting level of preprocessor
218 // branches during one pass over the code.
221 // Contains the current branch (#if, #else or one of the #elif branches)
222 // for each nesting level.
223 SmallVector<int, 8> PPLevelBranchIndex;
225 // Contains the maximum number of branches at each nesting level.
226 SmallVector<int, 8> PPLevelBranchCount;
228 // Contains the number of branches per nesting level we are currently
229 // in while parsing a preprocessor branch sequence.
230 // This is used to update PPLevelBranchCount at the end of a branch
232 std::stack<int> PPChainBranchIndex;
234 friend class ScopedLineState;
235 friend class CompoundStatementIndenter;
238 struct UnwrappedLineNode {
239 UnwrappedLineNode() : Tok(nullptr) {}
240 UnwrappedLineNode(FormatToken *Tok) : Tok(Tok) {}
243 SmallVector<UnwrappedLine, 0> Children;
246 inline UnwrappedLine::UnwrappedLine() : Level(0), InPPDirective(false),
247 MustBeDeclaration(false), MatchingOpeningBlockLineIndex(kInvalidIndex) {}
249 } // end namespace format
250 } // end namespace clang