1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===//
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 implements a token annotator, i.e. creates
12 /// \c AnnotatedTokens out of \c FormatTokens with required extra information.
14 //===----------------------------------------------------------------------===//
16 #include "TokenAnnotator.h"
17 #include "clang/Basic/SourceManager.h"
18 #include "llvm/ADT/SmallPtrSet.h"
19 #include "llvm/Support/Debug.h"
21 #define DEBUG_TYPE "format-token-annotator"
28 /// \brief A parser that gathers additional information about tokens.
30 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
31 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
32 /// into template parameter lists.
33 class AnnotatingParser {
35 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
36 const AdditionalKeywords &Keywords)
37 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
39 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
40 resetTokenMetadata(CurrentToken);
45 if (!CurrentToken || !CurrentToken->Previous)
47 if (NonTemplateLess.count(CurrentToken->Previous))
50 const FormatToken& Previous = *CurrentToken->Previous;
51 if (Previous.Previous) {
52 if (Previous.Previous->Tok.isLiteral())
54 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
55 (!Previous.Previous->MatchingParen ||
56 !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen)))
60 FormatToken *Left = CurrentToken->Previous;
61 Left->ParentBracket = Contexts.back().ContextKind;
62 ScopedContextCreator ContextCreator(*this, tok::less, 12);
64 // If this angle is in the context of an expression, we need to be more
65 // hesitant to detect it as opening template parameters.
66 bool InExprContext = Contexts.back().IsExpression;
68 Contexts.back().IsExpression = false;
69 // If there's a template keyword before the opening angle bracket, this is a
70 // template parameter, not an argument.
71 Contexts.back().InTemplateArgument =
72 Left->Previous && Left->Previous->Tok.isNot(tok::kw_template);
74 if (Style.Language == FormatStyle::LK_Java &&
75 CurrentToken->is(tok::question))
78 while (CurrentToken) {
79 if (CurrentToken->is(tok::greater)) {
80 Left->MatchingParen = CurrentToken;
81 CurrentToken->MatchingParen = Left;
82 CurrentToken->Type = TT_TemplateCloser;
86 if (CurrentToken->is(tok::question) &&
87 Style.Language == FormatStyle::LK_Java) {
91 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
92 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext))
94 // If a && or || is found and interpreted as a binary operator, this set
95 // of angles is likely part of something like "a < b && c > d". If the
96 // angles are inside an expression, the ||/&& might also be a binary
97 // operator that was misinterpreted because we are parsing template
99 // FIXME: This is getting out of hand, write a decent parser.
100 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
101 CurrentToken->Previous->is(TT_BinaryOperator) &&
102 Contexts[Contexts.size() - 2].IsExpression &&
103 !Line.startsWith(tok::kw_template))
105 updateParameterCount(Left, CurrentToken);
112 bool parseParens(bool LookForDecls = false) {
115 FormatToken *Left = CurrentToken->Previous;
116 Left->ParentBracket = Contexts.back().ContextKind;
117 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
119 // FIXME: This is a bit of a hack. Do better.
120 Contexts.back().ColonIsForRangeExpr =
121 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
123 bool StartsObjCMethodExpr = false;
124 if (CurrentToken->is(tok::caret)) {
125 // (^ can start a block type.
126 Left->Type = TT_ObjCBlockLParen;
127 } else if (FormatToken *MaybeSel = Left->Previous) {
128 // @selector( starts a selector.
129 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
130 MaybeSel->Previous->is(tok::at)) {
131 StartsObjCMethodExpr = true;
135 if (Left->is(TT_OverloadedOperatorLParen)) {
136 Contexts.back().IsExpression = false;
137 } else if (Style.Language == FormatStyle::LK_JavaScript &&
138 Line.startsWith(Keywords.kw_type, tok::identifier)) {
140 Contexts.back().IsExpression = false;
141 } else if (Left->Previous &&
142 (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype,
143 tok::kw_if, tok::kw_while, tok::l_paren,
145 Left->Previous->is(TT_BinaryOperator))) {
146 // static_assert, if and while usually contain expressions.
147 Contexts.back().IsExpression = true;
148 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
149 (Left->Previous->is(Keywords.kw_function) ||
150 (Left->Previous->endsSequence(tok::identifier,
151 Keywords.kw_function)))) {
152 // function(...) or function f(...)
153 Contexts.back().IsExpression = false;
154 } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
155 Left->Previous->is(TT_JsTypeColon)) {
156 // let x: (SomeType);
157 Contexts.back().IsExpression = false;
158 } else if (Left->Previous && Left->Previous->is(tok::r_square) &&
159 Left->Previous->MatchingParen &&
160 Left->Previous->MatchingParen->is(TT_LambdaLSquare)) {
161 // This is a parameter list of a lambda expression.
162 Contexts.back().IsExpression = false;
163 } else if (Line.InPPDirective &&
164 (!Left->Previous || !Left->Previous->is(tok::identifier))) {
165 Contexts.back().IsExpression = true;
166 } else if (Contexts[Contexts.size() - 2].CaretFound) {
167 // This is the parameter list of an ObjC block.
168 Contexts.back().IsExpression = false;
169 } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) {
170 Left->Type = TT_AttributeParen;
171 } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) {
172 // The first argument to a foreach macro is a declaration.
173 Contexts.back().IsForEachMacro = true;
174 Contexts.back().IsExpression = false;
175 } else if (Left->Previous && Left->Previous->MatchingParen &&
176 Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
177 Contexts.back().IsExpression = false;
178 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
180 Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch);
181 Contexts.back().IsExpression = !IsForOrCatch;
184 if (StartsObjCMethodExpr) {
185 Contexts.back().ColonIsObjCMethodExpr = true;
186 Left->Type = TT_ObjCMethodExpr;
189 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
190 bool ProbablyFunctionType = CurrentToken->isOneOf(tok::star, tok::amp);
191 bool HasMultipleLines = false;
192 bool HasMultipleParametersOnALine = false;
193 bool MightBeObjCForRangeLoop =
194 Left->Previous && Left->Previous->is(tok::kw_for);
195 while (CurrentToken) {
196 // LookForDecls is set when "if (" has been seen. Check for
197 // 'identifier' '*' 'identifier' followed by not '=' -- this
198 // '*' has to be a binary operator but determineStarAmpUsage() will
199 // categorize it as an unary operator, so set the right type here.
200 if (LookForDecls && CurrentToken->Next) {
201 FormatToken *Prev = CurrentToken->getPreviousNonComment();
203 FormatToken *PrevPrev = Prev->getPreviousNonComment();
204 FormatToken *Next = CurrentToken->Next;
205 if (PrevPrev && PrevPrev->is(tok::identifier) &&
206 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
207 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
208 Prev->Type = TT_BinaryOperator;
209 LookForDecls = false;
214 if (CurrentToken->Previous->is(TT_PointerOrReference) &&
215 CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
217 ProbablyFunctionType = true;
218 if (CurrentToken->is(tok::comma))
219 MightBeFunctionType = false;
220 if (CurrentToken->Previous->is(TT_BinaryOperator))
221 Contexts.back().IsExpression = true;
222 if (CurrentToken->is(tok::r_paren)) {
223 if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next &&
224 (CurrentToken->Next->is(tok::l_paren) ||
225 (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration)))
226 Left->Type = TT_FunctionTypeLParen;
227 Left->MatchingParen = CurrentToken;
228 CurrentToken->MatchingParen = Left;
230 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
231 Left->Previous && Left->Previous->is(tok::l_paren)) {
232 // Detect the case where macros are used to generate lambdas or
233 // function bodies, e.g.:
234 // auto my_lambda = MARCO((Type *type, int i) { .. body .. });
235 for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) {
236 if (Tok->is(TT_BinaryOperator) &&
237 Tok->isOneOf(tok::star, tok::amp, tok::ampamp))
238 Tok->Type = TT_PointerOrReference;
242 if (StartsObjCMethodExpr) {
243 CurrentToken->Type = TT_ObjCMethodExpr;
244 if (Contexts.back().FirstObjCSelectorName) {
245 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
246 Contexts.back().LongestObjCSelectorName;
250 if (Left->is(TT_AttributeParen))
251 CurrentToken->Type = TT_AttributeParen;
252 if (Left->Previous && Left->Previous->is(TT_JavaAnnotation))
253 CurrentToken->Type = TT_JavaAnnotation;
254 if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation))
255 CurrentToken->Type = TT_LeadingJavaAnnotation;
257 if (!HasMultipleLines)
258 Left->PackingKind = PPK_Inconclusive;
259 else if (HasMultipleParametersOnALine)
260 Left->PackingKind = PPK_BinPacked;
262 Left->PackingKind = PPK_OnePerLine;
267 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
270 if (CurrentToken->is(tok::l_brace))
271 Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen
272 if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
273 !CurrentToken->Next->HasUnescapedNewline &&
274 !CurrentToken->Next->isTrailingComment())
275 HasMultipleParametersOnALine = true;
276 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
277 CurrentToken->Previous->isSimpleTypeSpecifier()) &&
278 !CurrentToken->is(tok::l_brace))
279 Contexts.back().IsExpression = false;
280 if (CurrentToken->isOneOf(tok::semi, tok::colon))
281 MightBeObjCForRangeLoop = false;
282 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in))
283 CurrentToken->Type = TT_ObjCForIn;
284 // When we discover a 'new', we set CanBeExpression to 'false' in order to
285 // parse the type correctly. Reset that after a comma.
286 if (CurrentToken->is(tok::comma))
287 Contexts.back().CanBeExpression = true;
289 FormatToken *Tok = CurrentToken;
292 updateParameterCount(Left, Tok);
293 if (CurrentToken && CurrentToken->HasUnescapedNewline)
294 HasMultipleLines = true;
303 // A '[' could be an index subscript (after an identifier or after
304 // ')' or ']'), it could be the start of an Objective-C method
305 // expression, or it could the start of an Objective-C array literal.
306 FormatToken *Left = CurrentToken->Previous;
307 Left->ParentBracket = Contexts.back().ContextKind;
308 FormatToken *Parent = Left->getPreviousNonComment();
310 // Cases where '>' is followed by '['.
311 // In C++, this can happen either in array of templates (foo<int>[10])
312 // or when array is a nested template type (unique_ptr<type1<type2>[]>).
313 bool CppArrayTemplates =
314 Style.Language == FormatStyle::LK_Cpp && Parent &&
315 Parent->is(TT_TemplateCloser) &&
316 (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
317 Contexts.back().InTemplateArgument);
319 bool StartsObjCMethodExpr =
320 !CppArrayTemplates && (Style.Language == FormatStyle::LK_Cpp ||
321 Style.Language == FormatStyle::LK_ObjC) &&
322 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) &&
323 CurrentToken->isNot(tok::l_brace) &&
325 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
326 tok::kw_return, tok::kw_throw) ||
327 Parent->isUnaryOperator() ||
328 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
329 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown);
330 bool ColonFound = false;
332 unsigned BindingIncrease = 1;
333 if (Left->is(TT_Unknown)) {
334 if (StartsObjCMethodExpr) {
335 Left->Type = TT_ObjCMethodExpr;
336 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
337 Contexts.back().ContextKind == tok::l_brace &&
338 Parent->isOneOf(tok::l_brace, tok::comma)) {
339 Left->Type = TT_JsComputedPropertyName;
340 } else if (CurrentToken->is(tok::r_square) && Parent &&
341 Parent->is(TT_TemplateCloser)) {
342 Left->Type = TT_ArraySubscriptLSquare;
343 } else if (Style.Language == FormatStyle::LK_Proto ||
344 (!CppArrayTemplates && Parent &&
345 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
346 tok::comma, tok::l_paren, tok::l_square,
347 tok::question, tok::colon, tok::kw_return,
348 // Should only be relevant to JavaScript:
350 Left->Type = TT_ArrayInitializerLSquare;
352 BindingIncrease = 10;
353 Left->Type = TT_ArraySubscriptLSquare;
357 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
358 Contexts.back().IsExpression = true;
359 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
361 while (CurrentToken) {
362 if (CurrentToken->is(tok::r_square)) {
363 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) &&
364 Left->is(TT_ObjCMethodExpr)) {
365 // An ObjC method call is rarely followed by an open parenthesis.
366 // FIXME: Do we incorrectly label ":" with this?
367 StartsObjCMethodExpr = false;
368 Left->Type = TT_Unknown;
370 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
371 CurrentToken->Type = TT_ObjCMethodExpr;
372 // determineStarAmpUsage() thinks that '*' '[' is allocating an
373 // array of pointers, but if '[' starts a selector then '*' is a
375 if (Parent && Parent->is(TT_PointerOrReference))
376 Parent->Type = TT_BinaryOperator;
378 Left->MatchingParen = CurrentToken;
379 CurrentToken->MatchingParen = Left;
380 if (Contexts.back().FirstObjCSelectorName) {
381 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
382 Contexts.back().LongestObjCSelectorName;
383 if (Left->BlockParameterCount > 1)
384 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
389 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
391 if (CurrentToken->is(tok::colon)) {
392 if (Left->is(TT_ArraySubscriptLSquare)) {
393 Left->Type = TT_ObjCMethodExpr;
394 StartsObjCMethodExpr = true;
395 Contexts.back().ColonIsObjCMethodExpr = true;
396 if (Parent && Parent->is(tok::r_paren))
397 Parent->Type = TT_CastRParen;
401 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
403 Left->Type = TT_ArrayInitializerLSquare;
404 FormatToken *Tok = CurrentToken;
407 updateParameterCount(Left, Tok);
414 FormatToken *Left = CurrentToken->Previous;
415 Left->ParentBracket = Contexts.back().ContextKind;
417 if (Contexts.back().CaretFound)
418 Left->Type = TT_ObjCBlockLBrace;
419 Contexts.back().CaretFound = false;
421 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
422 Contexts.back().ColonIsDictLiteral = true;
423 if (Left->BlockKind == BK_BracedInit)
424 Contexts.back().IsExpression = true;
426 while (CurrentToken) {
427 if (CurrentToken->is(tok::r_brace)) {
428 Left->MatchingParen = CurrentToken;
429 CurrentToken->MatchingParen = Left;
433 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
435 updateParameterCount(Left, CurrentToken);
436 if (CurrentToken->isOneOf(tok::colon, tok::l_brace)) {
437 FormatToken *Previous = CurrentToken->getPreviousNonComment();
438 if (((CurrentToken->is(tok::colon) &&
439 (!Contexts.back().ColonIsDictLiteral ||
440 (Style.Language != FormatStyle::LK_Cpp &&
441 Style.Language != FormatStyle::LK_ObjC))) ||
442 Style.Language == FormatStyle::LK_Proto) &&
443 (Previous->Tok.getIdentifierInfo() ||
444 Previous->is(tok::string_literal)))
445 Previous->Type = TT_SelectorName;
446 if (CurrentToken->is(tok::colon) ||
447 Style.Language == FormatStyle::LK_JavaScript)
448 Left->Type = TT_DictLiteral;
450 if (CurrentToken->is(tok::comma) &&
451 Style.Language == FormatStyle::LK_JavaScript)
452 Left->Type = TT_DictLiteral;
460 void updateParameterCount(FormatToken *Left, FormatToken *Current) {
461 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block)
462 ++Left->BlockParameterCount;
463 if (Current->is(tok::comma)) {
464 ++Left->ParameterCount;
466 Left->Role.reset(new CommaSeparatedList(Style));
467 Left->Role->CommaFound(Current);
468 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
469 Left->ParameterCount = 1;
473 bool parseConditional() {
474 while (CurrentToken) {
475 if (CurrentToken->is(tok::colon)) {
476 CurrentToken->Type = TT_ConditionalExpr;
486 bool parseTemplateDeclaration() {
487 if (CurrentToken && CurrentToken->is(tok::less)) {
488 CurrentToken->Type = TT_TemplateOpener;
493 CurrentToken->Previous->ClosesTemplateDeclaration = true;
499 bool consumeToken() {
500 FormatToken *Tok = CurrentToken;
502 switch (Tok->Tok.getKind()) {
505 if (!Tok->Previous && Line.MustBeDeclaration)
506 Tok->Type = TT_ObjCMethodSpecifier;
511 // Colons from ?: are handled in parseConditional().
512 if (Style.Language == FormatStyle::LK_JavaScript) {
513 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
514 (Contexts.size() == 1 && // switch/case labels
515 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
516 Contexts.back().ContextKind == tok::l_paren || // function params
517 Contexts.back().ContextKind == tok::l_square || // array type
518 (Contexts.size() == 1 &&
519 Line.MustBeDeclaration)) { // method/property declaration
520 Tok->Type = TT_JsTypeColon;
524 if (Contexts.back().ColonIsDictLiteral ||
525 Style.Language == FormatStyle::LK_Proto) {
526 Tok->Type = TT_DictLiteral;
527 } else if (Contexts.back().ColonIsObjCMethodExpr ||
528 Line.startsWith(TT_ObjCMethodSpecifier)) {
529 Tok->Type = TT_ObjCMethodExpr;
530 const FormatToken *BeforePrevious = Tok->Previous->Previous;
531 if (!BeforePrevious ||
532 !(BeforePrevious->is(TT_CastRParen) ||
533 (BeforePrevious->is(TT_ObjCMethodExpr) &&
534 BeforePrevious->is(tok::colon))) ||
535 BeforePrevious->is(tok::r_square) ||
536 Contexts.back().LongestObjCSelectorName == 0) {
537 Tok->Previous->Type = TT_SelectorName;
538 if (Tok->Previous->ColumnWidth >
539 Contexts.back().LongestObjCSelectorName)
540 Contexts.back().LongestObjCSelectorName =
541 Tok->Previous->ColumnWidth;
542 if (!Contexts.back().FirstObjCSelectorName)
543 Contexts.back().FirstObjCSelectorName = Tok->Previous;
545 } else if (Contexts.back().ColonIsForRangeExpr) {
546 Tok->Type = TT_RangeBasedForLoopColon;
547 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
548 Tok->Type = TT_BitFieldColon;
549 } else if (Contexts.size() == 1 &&
550 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) {
551 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren,
553 Tok->Type = TT_CtorInitializerColon;
555 Tok->Type = TT_InheritanceColon;
556 } else if (Tok->Previous->is(tok::identifier) && Tok->Next &&
557 Tok->Next->isOneOf(tok::r_paren, tok::comma)) {
558 // This handles a special macro in ObjC code where selectors including
559 // the colon are passed as macro arguments.
560 Tok->Type = TT_ObjCMethodExpr;
561 } else if (Contexts.back().ContextKind == tok::l_paren) {
562 Tok->Type = TT_InlineASMColon;
567 // | and & in declarations/type expressions represent union and
568 // intersection types, respectively.
569 if (Style.Language == FormatStyle::LK_JavaScript &&
570 !Contexts.back().IsExpression)
571 Tok->Type = TT_JsTypeOperator;
575 if (CurrentToken && CurrentToken->is(tok::l_paren)) {
577 if (!parseParens(/*LookForDecls=*/true))
582 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Previous &&
583 Tok->Previous->is(tok::period))
585 Contexts.back().ColonIsForRangeExpr = true;
591 // When faced with 'operator()()', the kw_operator handler incorrectly
592 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
593 // the first two parens OverloadedOperators and the second l_paren an
594 // OverloadedOperatorLParen.
596 Tok->Previous->is(tok::r_paren) &&
597 Tok->Previous->MatchingParen &&
598 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
599 Tok->Previous->Type = TT_OverloadedOperator;
600 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator;
601 Tok->Type = TT_OverloadedOperatorLParen;
606 if (Line.MustBeDeclaration && Contexts.size() == 1 &&
607 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
609 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute,
610 TT_LeadingJavaAnnotation)))
611 Line.MightBeFunctionDecl = true;
623 Tok->Type = TT_TemplateOpener;
625 Tok->Type = TT_BinaryOperator;
626 NonTemplateLess.insert(Tok);
635 // Lines can start with '}'.
640 Tok->Type = TT_BinaryOperator;
642 case tok::kw_operator:
643 while (CurrentToken &&
644 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
645 if (CurrentToken->isOneOf(tok::star, tok::amp))
646 CurrentToken->Type = TT_PointerOrReference;
649 CurrentToken->Previous->isOneOf(TT_BinaryOperator, tok::comma))
650 CurrentToken->Previous->Type = TT_OverloadedOperator;
653 CurrentToken->Type = TT_OverloadedOperatorLParen;
654 if (CurrentToken->Previous->is(TT_BinaryOperator))
655 CurrentToken->Previous->Type = TT_OverloadedOperator;
659 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next &&
660 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
662 // Question marks before semicolons, colons, etc. indicate optional
663 // types (fields, parameters), e.g.
664 // function(x?: string, y?) {...}
666 Tok->Type = TT_JsTypeOptionalQuestion;
669 // Declarations cannot be conditional expressions, this can only be part
670 // of a type declaration.
671 if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
672 Style.Language == FormatStyle::LK_JavaScript)
676 case tok::kw_template:
677 parseTemplateDeclaration();
680 if (Contexts.back().InCtorInitializer)
681 Tok->Type = TT_CtorInitializerComma;
682 else if (Contexts.back().FirstStartOfName &&
683 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
684 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
685 Line.IsMultiVariableDeclStmt = true;
687 if (Contexts.back().IsForEachMacro)
688 Contexts.back().IsExpression = true;
690 case tok::identifier:
691 if (Tok->isOneOf(Keywords.kw___has_include,
692 Keywords.kw___has_include_next)) {
702 void parseIncludeDirective() {
703 if (CurrentToken && CurrentToken->is(tok::less)) {
705 while (CurrentToken) {
706 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next)
707 CurrentToken->Type = TT_ImplicitStringLiteral;
713 void parseWarningOrError() {
715 // We still want to format the whitespace left of the first token of the
718 while (CurrentToken) {
719 CurrentToken->Type = TT_ImplicitStringLiteral;
725 next(); // Consume "pragma".
727 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
728 bool IsMark = CurrentToken->is(Keywords.kw_mark);
729 next(); // Consume "mark".
730 next(); // Consume first token (so we fix leading whitespace).
731 while (CurrentToken) {
732 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
733 CurrentToken->Type = TT_ImplicitStringLiteral;
739 void parseHasInclude() {
740 if (!CurrentToken || !CurrentToken->is(tok::l_paren))
743 parseIncludeDirective();
747 LineType parsePreprocessorDirective() {
748 bool IsFirstToken = CurrentToken->IsFirst;
749 LineType Type = LT_PreprocessorDirective;
754 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
755 // JavaScript files can contain shebang lines of the form:
756 // #!/usr/bin/env node
757 // Treat these like C++ #include directives.
758 while (CurrentToken) {
759 // Tokens cannot be comments here.
760 CurrentToken->Type = TT_ImplicitStringLiteral;
763 return LT_ImportStatement;
766 if (CurrentToken->Tok.is(tok::numeric_constant)) {
767 CurrentToken->SpacesRequiredBefore = 1;
770 // Hashes in the middle of a line can lead to any strange token
772 if (!CurrentToken->Tok.getIdentifierInfo())
774 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
775 case tok::pp_include:
776 case tok::pp_include_next:
779 parseIncludeDirective();
780 Type = LT_ImportStatement;
783 case tok::pp_warning:
784 parseWarningOrError();
791 Contexts.back().IsExpression = true;
797 while (CurrentToken) {
798 FormatToken *Tok = CurrentToken;
800 if (Tok->isOneOf(Keywords.kw___has_include,
801 Keywords.kw___has_include_next)) {
809 LineType parseLine() {
810 NonTemplateLess.clear();
811 if (CurrentToken->is(tok::hash))
812 return parsePreprocessorDirective();
814 // Directly allow to 'import <string-literal>' to support protocol buffer
815 // definitions (code.google.com/p/protobuf) or missing "#" (either way we
816 // should not break the line).
817 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
818 if ((Style.Language == FormatStyle::LK_Java &&
819 CurrentToken->is(Keywords.kw_package)) ||
820 (Info && Info->getPPKeywordID() == tok::pp_import &&
821 CurrentToken->Next &&
822 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
825 parseIncludeDirective();
826 return LT_ImportStatement;
829 // If this line starts and ends in '<' and '>', respectively, it is likely
830 // part of "#define <a/b.h>".
831 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
832 parseIncludeDirective();
833 return LT_ImportStatement;
836 // In .proto files, top-level options are very similar to import statements
837 // and should not be line-wrapped.
838 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
839 CurrentToken->is(Keywords.kw_option)) {
841 if (CurrentToken && CurrentToken->is(tok::identifier))
842 return LT_ImportStatement;
845 bool KeywordVirtualFound = false;
846 bool ImportStatement = false;
848 // import {...} from '...';
849 if (Style.Language == FormatStyle::LK_JavaScript &&
850 CurrentToken->is(Keywords.kw_import))
851 ImportStatement = true;
853 while (CurrentToken) {
854 if (CurrentToken->is(tok::kw_virtual))
855 KeywordVirtualFound = true;
856 if (Style.Language == FormatStyle::LK_JavaScript) {
857 // export {...} from '...';
858 // An export followed by "from 'some string';" is a re-export from
859 // another module identified by a URI and is treated as a
860 // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
861 // Just "export {...};" or "export class ..." should not be treated as
862 // an import in this sense.
863 if (Line.First->is(tok::kw_export) &&
864 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
865 CurrentToken->Next->isStringLiteral())
866 ImportStatement = true;
867 if (isClosureImportStatement(*CurrentToken))
868 ImportStatement = true;
873 if (KeywordVirtualFound)
874 return LT_VirtualFunctionDecl;
876 return LT_ImportStatement;
878 if (Line.startsWith(TT_ObjCMethodSpecifier)) {
879 if (Contexts.back().FirstObjCSelectorName)
880 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
881 Contexts.back().LongestObjCSelectorName;
882 return LT_ObjCMethodDecl;
889 bool isClosureImportStatement(const FormatToken &Tok) {
890 // FIXME: Closure-library specific stuff should not be hard-coded but be
892 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
893 Tok.Next->Next && (Tok.Next->Next->TokenText == "module" ||
894 Tok.Next->Next->TokenText == "provide" ||
895 Tok.Next->Next->TokenText == "require" ||
896 Tok.Next->Next->TokenText == "setTestOnly" ||
897 Tok.Next->Next->TokenText == "forwardDeclare") &&
898 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
901 void resetTokenMetadata(FormatToken *Token) {
905 // Reset token type in case we have already looked at it and then
906 // recovered from an error (e.g. failure to find the matching >).
907 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro,
908 TT_FunctionLBrace, TT_ImplicitStringLiteral,
909 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow,
910 TT_OverloadedOperator, TT_RegexLiteral,
912 CurrentToken->Type = TT_Unknown;
913 CurrentToken->Role.reset();
914 CurrentToken->MatchingParen = nullptr;
915 CurrentToken->FakeLParens.clear();
916 CurrentToken->FakeRParens = 0;
921 CurrentToken->NestingLevel = Contexts.size() - 1;
922 CurrentToken->BindingStrength = Contexts.back().BindingStrength;
923 modifyContext(*CurrentToken);
924 determineTokenType(*CurrentToken);
925 CurrentToken = CurrentToken->Next;
928 resetTokenMetadata(CurrentToken);
931 /// \brief A struct to hold information valid in a specific context, e.g.
932 /// a pair of parenthesis.
934 Context(tok::TokenKind ContextKind, unsigned BindingStrength,
936 : ContextKind(ContextKind), BindingStrength(BindingStrength),
937 IsExpression(IsExpression) {}
939 tok::TokenKind ContextKind;
940 unsigned BindingStrength;
942 unsigned LongestObjCSelectorName = 0;
943 bool ColonIsForRangeExpr = false;
944 bool ColonIsDictLiteral = false;
945 bool ColonIsObjCMethodExpr = false;
946 FormatToken *FirstObjCSelectorName = nullptr;
947 FormatToken *FirstStartOfName = nullptr;
948 bool CanBeExpression = true;
949 bool InTemplateArgument = false;
950 bool InCtorInitializer = false;
951 bool CaretFound = false;
952 bool IsForEachMacro = false;
955 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime
956 /// of each instance.
957 struct ScopedContextCreator {
960 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
963 P.Contexts.push_back(Context(ContextKind,
964 P.Contexts.back().BindingStrength + Increase,
965 P.Contexts.back().IsExpression));
968 ~ScopedContextCreator() { P.Contexts.pop_back(); }
971 void modifyContext(const FormatToken &Current) {
972 if (Current.getPrecedence() == prec::Assignment &&
973 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
974 // Type aliases use `type X = ...;` in TypeScript.
975 !(Style.Language == FormatStyle::LK_JavaScript &&
976 Line.startsWith(Keywords.kw_type, tok::identifier)) &&
977 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
978 Contexts.back().IsExpression = true;
979 if (!Line.startsWith(TT_UnaryOperator)) {
980 for (FormatToken *Previous = Current.Previous;
981 Previous && Previous->Previous &&
982 !Previous->Previous->isOneOf(tok::comma, tok::semi);
983 Previous = Previous->Previous) {
984 if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
985 Previous = Previous->MatchingParen;
989 if (Previous->opensScope())
991 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
992 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
993 Previous->Previous && Previous->Previous->isNot(tok::equal))
994 Previous->Type = TT_PointerOrReference;
997 } else if (Current.is(tok::lessless) &&
998 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
999 Contexts.back().IsExpression = true;
1000 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1001 Contexts.back().IsExpression = true;
1002 } else if (Current.is(TT_TrailingReturnArrow)) {
1003 Contexts.back().IsExpression = false;
1004 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1005 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1006 } else if (Current.Previous &&
1007 Current.Previous->is(TT_CtorInitializerColon)) {
1008 Contexts.back().IsExpression = true;
1009 Contexts.back().InCtorInitializer = true;
1010 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1011 for (FormatToken *Previous = Current.Previous;
1012 Previous && Previous->isOneOf(tok::star, tok::amp);
1013 Previous = Previous->Previous)
1014 Previous->Type = TT_PointerOrReference;
1015 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1016 Contexts.back().IsExpression = false;
1017 } else if (Current.is(tok::kw_new)) {
1018 Contexts.back().CanBeExpression = false;
1019 } else if (Current.isOneOf(tok::semi, tok::exclaim)) {
1020 // This should be the condition or increment in a for-loop.
1021 Contexts.back().IsExpression = true;
1025 void determineTokenType(FormatToken &Current) {
1026 if (!Current.is(TT_Unknown))
1027 // The token type is already known.
1030 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1031 // function declaration have been found. In this case, 'Current' is a
1032 // trailing token of this declaration and thus cannot be a name.
1033 if (Current.is(Keywords.kw_instanceof)) {
1034 Current.Type = TT_BinaryOperator;
1035 } else if (isStartOfName(Current) &&
1036 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1037 Contexts.back().FirstStartOfName = &Current;
1038 Current.Type = TT_StartOfName;
1039 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1041 } else if (Current.is(tok::arrow) &&
1042 Style.Language == FormatStyle::LK_Java) {
1043 Current.Type = TT_LambdaArrow;
1044 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1045 Current.NestingLevel == 0) {
1046 Current.Type = TT_TrailingReturnArrow;
1047 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1049 determineStarAmpUsage(Current, Contexts.back().CanBeExpression &&
1050 Contexts.back().IsExpression,
1051 Contexts.back().InTemplateArgument);
1052 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1053 Current.Type = determinePlusMinusCaretUsage(Current);
1054 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1055 Contexts.back().CaretFound = true;
1056 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1057 Current.Type = determineIncrementUsage(Current);
1058 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1059 Current.Type = TT_UnaryOperator;
1060 } else if (Current.is(tok::question)) {
1061 if (Style.Language == FormatStyle::LK_JavaScript &&
1062 Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1063 // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1064 // on the interface, not a ternary expression.
1065 Current.Type = TT_JsTypeOptionalQuestion;
1067 Current.Type = TT_ConditionalExpr;
1069 } else if (Current.isBinaryOperator() &&
1070 (!Current.Previous || Current.Previous->isNot(tok::l_square))) {
1071 Current.Type = TT_BinaryOperator;
1072 } else if (Current.is(tok::comment)) {
1073 if (Current.TokenText.startswith("/*")) {
1074 if (Current.TokenText.endswith("*/"))
1075 Current.Type = TT_BlockComment;
1077 // The lexer has for some reason determined a comment here. But we
1078 // cannot really handle it, if it isn't properly terminated.
1079 Current.Tok.setKind(tok::unknown);
1081 Current.Type = TT_LineComment;
1083 } else if (Current.is(tok::r_paren)) {
1084 if (rParenEndsCast(Current))
1085 Current.Type = TT_CastRParen;
1086 if (Current.MatchingParen && Current.Next &&
1087 !Current.Next->isBinaryOperator() &&
1088 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1089 tok::period, tok::arrow, tok::coloncolon))
1090 if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1091 // Make sure this isn't the return type of an Obj-C block declaration
1092 if (AfterParen->Tok.isNot(tok::caret)) {
1093 if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1094 if (BeforeParen->is(tok::identifier) &&
1095 BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1096 (!BeforeParen->Previous ||
1097 BeforeParen->Previous->ClosesTemplateDeclaration))
1098 Current.Type = TT_FunctionAnnotationRParen;
1101 } else if (Current.is(tok::at) && Current.Next) {
1102 if (Current.Next->isStringLiteral()) {
1103 Current.Type = TT_ObjCStringLiteral;
1105 switch (Current.Next->Tok.getObjCKeywordID()) {
1106 case tok::objc_interface:
1107 case tok::objc_implementation:
1108 case tok::objc_protocol:
1109 Current.Type = TT_ObjCDecl;
1111 case tok::objc_property:
1112 Current.Type = TT_ObjCProperty;
1118 } else if (Current.is(tok::period)) {
1119 FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1120 if (PreviousNoComment &&
1121 PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1122 Current.Type = TT_DesignatedInitializerPeriod;
1123 else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1124 Current.Previous->isOneOf(TT_JavaAnnotation,
1125 TT_LeadingJavaAnnotation)) {
1126 Current.Type = Current.Previous->Type;
1128 } else if (Current.isOneOf(tok::identifier, tok::kw_const) &&
1130 !Current.Previous->isOneOf(tok::equal, tok::at) &&
1131 Line.MightBeFunctionDecl && Contexts.size() == 1) {
1132 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1133 // function declaration have been found.
1134 Current.Type = TT_TrailingAnnotation;
1135 } else if ((Style.Language == FormatStyle::LK_Java ||
1136 Style.Language == FormatStyle::LK_JavaScript) &&
1138 if (Current.Previous->is(tok::at) &&
1139 Current.isNot(Keywords.kw_interface)) {
1140 const FormatToken &AtToken = *Current.Previous;
1141 const FormatToken *Previous = AtToken.getPreviousNonComment();
1142 if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1143 Current.Type = TT_LeadingJavaAnnotation;
1145 Current.Type = TT_JavaAnnotation;
1146 } else if (Current.Previous->is(tok::period) &&
1147 Current.Previous->isOneOf(TT_JavaAnnotation,
1148 TT_LeadingJavaAnnotation)) {
1149 Current.Type = Current.Previous->Type;
1154 /// \brief Take a guess at whether \p Tok starts a name of a function or
1155 /// variable declaration.
1157 /// This is a heuristic based on whether \p Tok is an identifier following
1158 /// something that is likely a type.
1159 bool isStartOfName(const FormatToken &Tok) {
1160 if (Tok.isNot(tok::identifier) || !Tok.Previous)
1163 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof))
1165 if (Style.Language == FormatStyle::LK_JavaScript &&
1166 Tok.Previous->is(Keywords.kw_in))
1169 // Skip "const" as it does not have an influence on whether this is a name.
1170 FormatToken *PreviousNotConst = Tok.Previous;
1171 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1172 PreviousNotConst = PreviousNotConst->Previous;
1174 if (!PreviousNotConst)
1177 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1178 PreviousNotConst->Previous &&
1179 PreviousNotConst->Previous->is(tok::hash);
1181 if (PreviousNotConst->is(TT_TemplateCloser))
1182 return PreviousNotConst && PreviousNotConst->MatchingParen &&
1183 PreviousNotConst->MatchingParen->Previous &&
1184 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1185 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1187 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen &&
1188 PreviousNotConst->MatchingParen->Previous &&
1189 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype))
1192 return (!IsPPKeyword &&
1193 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1194 PreviousNotConst->is(TT_PointerOrReference) ||
1195 PreviousNotConst->isSimpleTypeSpecifier();
1198 /// \brief Determine whether ')' is ending a cast.
1199 bool rParenEndsCast(const FormatToken &Tok) {
1200 // C-style casts are only used in C++ and Java.
1201 if (Style.Language != FormatStyle::LK_Cpp &&
1202 Style.Language != FormatStyle::LK_ObjC &&
1203 Style.Language != FormatStyle::LK_Java)
1206 // Empty parens aren't casts and there are no casts at the end of the line.
1207 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1210 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1212 // If there is a closing parenthesis left of the current parentheses,
1213 // look past it as these might be chained casts.
1214 if (LeftOfParens->is(tok::r_paren)) {
1215 if (!LeftOfParens->MatchingParen ||
1216 !LeftOfParens->MatchingParen->Previous)
1218 LeftOfParens = LeftOfParens->MatchingParen->Previous;
1221 // If there is an identifier (or with a few exceptions a keyword) right
1222 // before the parentheses, this is unlikely to be a cast.
1223 if (LeftOfParens->Tok.getIdentifierInfo() &&
1224 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1228 // Certain other tokens right before the parentheses are also signals that
1229 // this cannot be a cast.
1230 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1231 TT_TemplateCloser, tok::ellipsis))
1235 if (Tok.Next->is(tok::question))
1238 // As Java has no function types, a "(" after the ")" likely means that this
1240 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1243 // If a (non-string) literal follows, this is likely a cast.
1244 if (Tok.Next->isNot(tok::string_literal) &&
1245 (Tok.Next->Tok.isLiteral() ||
1246 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1249 // Heuristically try to determine whether the parentheses contain a type.
1250 bool ParensAreType =
1252 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) ||
1253 Tok.Previous->isSimpleTypeSpecifier();
1254 bool ParensCouldEndDecl =
1255 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1256 if (ParensAreType && !ParensCouldEndDecl)
1259 // At this point, we heuristically assume that there are no casts at the
1260 // start of the line. We assume that we have found most cases where there
1261 // are by the logic above, e.g. "(void)x;".
1265 // Certain token types inside the parentheses mean that this can't be a
1267 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1268 Token = Token->Next)
1269 if (Token->is(TT_BinaryOperator))
1272 // If the following token is an identifier or 'this', this is a cast. All
1273 // cases where this can be something else are handled above.
1274 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1277 if (!Tok.Next->Next)
1280 // If the next token after the parenthesis is a unary operator, assume
1281 // that this is cast, unless there are unexpected tokens inside the
1284 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1285 if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1286 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1288 // Search for unexpected tokens.
1289 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1290 Prev = Prev->Previous) {
1291 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1297 /// \brief Return the type of the given token assuming it is * or &.
1298 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1299 bool InTemplateArgument) {
1300 if (Style.Language == FormatStyle::LK_JavaScript)
1301 return TT_BinaryOperator;
1303 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1305 return TT_UnaryOperator;
1307 const FormatToken *NextToken = Tok.getNextNonComment();
1308 if (!NextToken || NextToken->isOneOf(tok::arrow, tok::equal) ||
1309 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1310 return TT_PointerOrReference;
1312 if (PrevToken->is(tok::coloncolon))
1313 return TT_PointerOrReference;
1315 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1316 tok::comma, tok::semi, tok::kw_return, tok::colon,
1317 tok::equal, tok::kw_delete, tok::kw_sizeof) ||
1318 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1319 TT_UnaryOperator, TT_CastRParen))
1320 return TT_UnaryOperator;
1322 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1323 return TT_PointerOrReference;
1324 if (NextToken->is(tok::kw_operator) && !IsExpression)
1325 return TT_PointerOrReference;
1326 if (NextToken->isOneOf(tok::comma, tok::semi))
1327 return TT_PointerOrReference;
1329 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen &&
1330 PrevToken->MatchingParen->Previous &&
1331 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof,
1333 return TT_PointerOrReference;
1335 if (PrevToken->Tok.isLiteral() ||
1336 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1337 tok::kw_false, tok::r_brace) ||
1338 NextToken->Tok.isLiteral() ||
1339 NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1340 NextToken->isUnaryOperator() ||
1341 // If we know we're in a template argument, there are no named
1342 // declarations. Thus, having an identifier on the right-hand side
1343 // indicates a binary operator.
1344 (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1345 return TT_BinaryOperator;
1347 // "&&(" is quite unlikely to be two successive unary "&".
1348 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren))
1349 return TT_BinaryOperator;
1351 // This catches some cases where evaluation order is used as control flow:
1353 const FormatToken *NextNextToken = NextToken->getNextNonComment();
1354 if (NextNextToken && NextNextToken->is(tok::arrow))
1355 return TT_BinaryOperator;
1357 // It is very unlikely that we are going to find a pointer or reference type
1358 // definition on the RHS of an assignment.
1359 if (IsExpression && !Contexts.back().CaretFound)
1360 return TT_BinaryOperator;
1362 return TT_PointerOrReference;
1365 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
1366 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1368 return TT_UnaryOperator;
1370 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator) &&
1371 !PrevToken->is(tok::exclaim))
1372 // There aren't any trailing unary operators except for TypeScript's
1373 // non-null operator (!). Thus, this must be squence of leading operators.
1374 return TT_UnaryOperator;
1376 // Use heuristics to recognize unary operators.
1377 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
1378 tok::question, tok::colon, tok::kw_return,
1379 tok::kw_case, tok::at, tok::l_brace))
1380 return TT_UnaryOperator;
1382 // There can't be two consecutive binary operators.
1383 if (PrevToken->is(TT_BinaryOperator))
1384 return TT_UnaryOperator;
1386 // Fall back to marking the token as binary operator.
1387 return TT_BinaryOperator;
1390 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements.
1391 TokenType determineIncrementUsage(const FormatToken &Tok) {
1392 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1393 if (!PrevToken || PrevToken->is(TT_CastRParen))
1394 return TT_UnaryOperator;
1395 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
1396 return TT_TrailingUnaryOperator;
1398 return TT_UnaryOperator;
1401 SmallVector<Context, 8> Contexts;
1403 const FormatStyle &Style;
1404 AnnotatedLine &Line;
1405 FormatToken *CurrentToken;
1407 const AdditionalKeywords &Keywords;
1409 // Set of "<" tokens that do not open a template parameter list. If parseAngle
1410 // determines that a specific token can't be a template opener, it will make
1411 // same decision irrespective of the decisions for tokens leading up to it.
1412 // Store this information to prevent this from causing exponential runtime.
1413 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
1416 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
1417 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
1419 /// \brief Parses binary expressions by inserting fake parenthesis based on
1420 /// operator precedence.
1421 class ExpressionParser {
1423 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
1424 AnnotatedLine &Line)
1425 : Style(Style), Keywords(Keywords), Current(Line.First) {}
1427 /// \brief Parse expressions with the given operatore precedence.
1428 void parse(int Precedence = 0) {
1429 // Skip 'return' and ObjC selector colons as they are not part of a binary
1431 while (Current && (Current->is(tok::kw_return) ||
1432 (Current->is(tok::colon) &&
1433 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
1436 if (!Current || Precedence > PrecedenceArrowAndPeriod)
1439 // Conditional expressions need to be parsed separately for proper nesting.
1440 if (Precedence == prec::Conditional) {
1441 parseConditionalExpr();
1445 // Parse unary operators, which all have a higher precedence than binary
1447 if (Precedence == PrecedenceUnaryOperator) {
1448 parseUnaryOperator();
1452 FormatToken *Start = Current;
1453 FormatToken *LatestOperator = nullptr;
1454 unsigned OperatorIndex = 0;
1457 // Consume operators with higher precedence.
1458 parse(Precedence + 1);
1460 int CurrentPrecedence = getCurrentPrecedence();
1462 if (Current && Current->is(TT_SelectorName) &&
1463 Precedence == CurrentPrecedence) {
1465 addFakeParenthesis(Start, prec::Level(Precedence));
1469 // At the end of the line or when an operator with higher precedence is
1470 // found, insert fake parenthesis and return.
1472 (Current->closesScope() &&
1473 (Current->MatchingParen || Current->is(TT_TemplateString))) ||
1474 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
1475 (CurrentPrecedence == prec::Conditional &&
1476 Precedence == prec::Assignment && Current->is(tok::colon))) {
1480 // Consume scopes: (), [], <> and {}
1481 if (Current->opensScope()) {
1482 // In fragment of a JavaScript template string can look like '}..${' and
1483 // thus close a scope and open a new one at the same time.
1484 while (Current && (!Current->closesScope() || Current->opensScope())) {
1491 if (CurrentPrecedence == Precedence) {
1493 LatestOperator->NextOperator = Current;
1494 LatestOperator = Current;
1495 Current->OperatorIndex = OperatorIndex;
1498 next(/*SkipPastLeadingComments=*/Precedence > 0);
1502 if (LatestOperator && (Current || Precedence > 0)) {
1503 // LatestOperator->LastOperator = true;
1504 if (Precedence == PrecedenceArrowAndPeriod) {
1505 // Call expressions don't have a binary operator precedence.
1506 addFakeParenthesis(Start, prec::Unknown);
1508 addFakeParenthesis(Start, prec::Level(Precedence));
1514 /// \brief Gets the precedence (+1) of the given token for binary operators
1515 /// and other tokens that we treat like binary operators.
1516 int getCurrentPrecedence() {
1518 const FormatToken *NextNonComment = Current->getNextNonComment();
1519 if (Current->is(TT_ConditionalExpr))
1520 return prec::Conditional;
1521 if (NextNonComment && NextNonComment->is(tok::colon) &&
1522 NextNonComment->is(TT_DictLiteral))
1524 if (Current->is(TT_LambdaArrow))
1526 if (Current->is(TT_JsFatArrow))
1527 return prec::Assignment;
1528 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName,
1529 TT_JsComputedPropertyName) ||
1530 (Current->is(tok::comment) && NextNonComment &&
1531 NextNonComment->is(TT_SelectorName)))
1533 if (Current->is(TT_RangeBasedForLoopColon))
1535 if ((Style.Language == FormatStyle::LK_Java ||
1536 Style.Language == FormatStyle::LK_JavaScript) &&
1537 Current->is(Keywords.kw_instanceof))
1538 return prec::Relational;
1539 if (Style.Language == FormatStyle::LK_JavaScript &&
1540 Current->is(Keywords.kw_in))
1541 return prec::Relational;
1542 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
1543 return Current->getPrecedence();
1544 if (Current->isOneOf(tok::period, tok::arrow))
1545 return PrecedenceArrowAndPeriod;
1546 if ((Style.Language == FormatStyle::LK_Java ||
1547 Style.Language == FormatStyle::LK_JavaScript) &&
1548 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
1549 Keywords.kw_throws))
1555 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
1556 Start->FakeLParens.push_back(Precedence);
1557 if (Precedence > prec::Unknown)
1558 Start->StartsBinaryExpression = true;
1560 FormatToken *Previous = Current->Previous;
1561 while (Previous->is(tok::comment) && Previous->Previous)
1562 Previous = Previous->Previous;
1563 ++Previous->FakeRParens;
1564 if (Precedence > prec::Unknown)
1565 Previous->EndsBinaryExpression = true;
1569 /// \brief Parse unary operator expressions and surround them with fake
1570 /// parentheses if appropriate.
1571 void parseUnaryOperator() {
1572 if (!Current || Current->isNot(TT_UnaryOperator)) {
1573 parse(PrecedenceArrowAndPeriod);
1577 FormatToken *Start = Current;
1579 parseUnaryOperator();
1581 // The actual precedence doesn't matter.
1582 addFakeParenthesis(Start, prec::Unknown);
1585 void parseConditionalExpr() {
1586 while (Current && Current->isTrailingComment()) {
1589 FormatToken *Start = Current;
1590 parse(prec::LogicalOr);
1591 if (!Current || !Current->is(tok::question))
1594 parse(prec::Assignment);
1595 if (!Current || Current->isNot(TT_ConditionalExpr))
1598 parse(prec::Assignment);
1599 addFakeParenthesis(Start, prec::Conditional);
1602 void next(bool SkipPastLeadingComments = true) {
1604 Current = Current->Next;
1606 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
1607 Current->isTrailingComment())
1608 Current = Current->Next;
1611 const FormatStyle &Style;
1612 const AdditionalKeywords &Keywords;
1613 FormatToken *Current;
1616 } // end anonymous namespace
1618 void TokenAnnotator::setCommentLineLevels(
1619 SmallVectorImpl<AnnotatedLine *> &Lines) {
1620 const AnnotatedLine *NextNonCommentLine = nullptr;
1621 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
1624 bool CommentLine = (*I)->First;
1625 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
1626 if (!Tok->is(tok::comment)) {
1627 CommentLine = false;
1631 if (NextNonCommentLine && CommentLine)
1632 (*I)->Level = NextNonCommentLine->Level;
1634 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
1636 setCommentLineLevels((*I)->Children);
1640 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
1641 unsigned Result = 0;
1642 for (const auto* Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
1643 Result = std::max(Result, Tok->NestingLevel);
1647 void TokenAnnotator::annotate(AnnotatedLine &Line) {
1648 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1649 E = Line.Children.end();
1653 AnnotatingParser Parser(Style, Line, Keywords);
1654 Line.Type = Parser.parseLine();
1656 // With very deep nesting, ExpressionParser uses lots of stack and the
1657 // formatting algorithm is very slow. We're not going to do a good job here
1658 // anyway - it's probably generated code being formatted by mistake.
1659 // Just skip the whole line.
1660 if (maxNestingDepth(Line) > 50)
1661 Line.Type = LT_Invalid;
1663 if (Line.Type == LT_Invalid)
1666 ExpressionParser ExprParser(Style, Keywords, Line);
1669 if (Line.startsWith(TT_ObjCMethodSpecifier))
1670 Line.Type = LT_ObjCMethodDecl;
1671 else if (Line.startsWith(TT_ObjCDecl))
1672 Line.Type = LT_ObjCDecl;
1673 else if (Line.startsWith(TT_ObjCProperty))
1674 Line.Type = LT_ObjCProperty;
1676 Line.First->SpacesRequiredBefore = 1;
1677 Line.First->CanBreakBefore = Line.First->MustBreakBefore;
1680 // This function heuristically determines whether 'Current' starts the name of a
1681 // function declaration.
1682 static bool isFunctionDeclarationName(const FormatToken &Current,
1683 const AnnotatedLine &Line) {
1684 auto skipOperatorName = [](const FormatToken* Next) -> const FormatToken* {
1685 for (; Next; Next = Next->Next) {
1686 if (Next->is(TT_OverloadedOperatorLParen))
1688 if (Next->is(TT_OverloadedOperator))
1690 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
1691 // For 'new[]' and 'delete[]'.
1692 if (Next->Next && Next->Next->is(tok::l_square) &&
1693 Next->Next->Next && Next->Next->Next->is(tok::r_square))
1694 Next = Next->Next->Next;
1703 // Find parentheses of parameter list.
1704 const FormatToken *Next = Current.Next;
1705 if (Current.is(tok::kw_operator)) {
1706 if (Current.Previous && Current.Previous->is(tok::coloncolon))
1708 Next = skipOperatorName(Next);
1710 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
1712 for (; Next; Next = Next->Next) {
1713 if (Next->is(TT_TemplateOpener)) {
1714 Next = Next->MatchingParen;
1715 } else if (Next->is(tok::coloncolon)) {
1719 if (Next->is(tok::kw_operator)) {
1720 Next = skipOperatorName(Next->Next);
1723 if (!Next->is(tok::identifier))
1725 } else if (Next->is(tok::l_paren)) {
1733 // Check whether parameter list can belong to a function declaration.
1734 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
1736 // If the lines ends with "{", this is likely an function definition.
1737 if (Line.Last->is(tok::l_brace))
1739 if (Next->Next == Next->MatchingParen)
1740 return true; // Empty parentheses.
1741 // If there is an &/&& after the r_paren, this is likely a function.
1742 if (Next->MatchingParen->Next &&
1743 Next->MatchingParen->Next->is(TT_PointerOrReference))
1745 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
1747 if (Tok->is(tok::l_paren) && Tok->MatchingParen) {
1748 Tok = Tok->MatchingParen;
1751 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
1752 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
1754 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
1755 Tok->Tok.isLiteral())
1761 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
1762 assert(Line.MightBeFunctionDecl);
1764 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
1765 Style.AlwaysBreakAfterReturnType ==
1766 FormatStyle::RTBS_TopLevelDefinitions) &&
1770 switch (Style.AlwaysBreakAfterReturnType) {
1771 case FormatStyle::RTBS_None:
1773 case FormatStyle::RTBS_All:
1774 case FormatStyle::RTBS_TopLevel:
1776 case FormatStyle::RTBS_AllDefinitions:
1777 case FormatStyle::RTBS_TopLevelDefinitions:
1778 return Line.mightBeFunctionDefinition();
1784 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
1785 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1786 E = Line.Children.end();
1788 calculateFormattingInformation(**I);
1791 Line.First->TotalLength =
1792 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth;
1793 FormatToken *Current = Line.First->Next;
1794 bool InFunctionDecl = Line.MightBeFunctionDecl;
1796 if (isFunctionDeclarationName(*Current, Line))
1797 Current->Type = TT_FunctionDeclarationName;
1798 if (Current->is(TT_LineComment)) {
1799 if (Current->Previous->BlockKind == BK_BracedInit &&
1800 Current->Previous->opensScope())
1801 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1;
1803 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
1805 // If we find a trailing comment, iterate backwards to determine whether
1806 // it seems to relate to a specific parameter. If so, break before that
1807 // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
1808 // to the previous line in:
1812 if (!Current->HasUnescapedNewline) {
1813 for (FormatToken *Parameter = Current->Previous; Parameter;
1814 Parameter = Parameter->Previous) {
1815 if (Parameter->isOneOf(tok::comment, tok::r_brace))
1817 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
1818 if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
1819 Parameter->HasUnescapedNewline)
1820 Parameter->MustBreakBefore = true;
1825 } else if (Current->SpacesRequiredBefore == 0 &&
1826 spaceRequiredBefore(Line, *Current)) {
1827 Current->SpacesRequiredBefore = 1;
1830 Current->MustBreakBefore =
1831 Current->MustBreakBefore || mustBreakBefore(Line, *Current);
1833 if (!Current->MustBreakBefore && InFunctionDecl &&
1834 Current->is(TT_FunctionDeclarationName))
1835 Current->MustBreakBefore = mustBreakForReturnType(Line);
1837 Current->CanBreakBefore =
1838 Current->MustBreakBefore || canBreakBefore(Line, *Current);
1839 unsigned ChildSize = 0;
1840 if (Current->Previous->Children.size() == 1) {
1841 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
1842 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
1843 : LastOfChild.TotalLength + 1;
1845 const FormatToken *Prev = Current->Previous;
1846 if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
1847 (Prev->Children.size() == 1 &&
1848 Prev->Children[0]->First->MustBreakBefore) ||
1849 Current->IsMultiline)
1850 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
1852 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
1853 ChildSize + Current->SpacesRequiredBefore;
1855 if (Current->is(TT_CtorInitializerColon))
1856 InFunctionDecl = false;
1858 // FIXME: Only calculate this if CanBreakBefore is true once static
1859 // initializers etc. are sorted out.
1860 // FIXME: Move magic numbers to a better place.
1861 Current->SplitPenalty = 20 * Current->BindingStrength +
1862 splitPenalty(Line, *Current, InFunctionDecl);
1864 Current = Current->Next;
1867 calculateUnbreakableTailLengths(Line);
1868 unsigned IndentLevel = Line.Level;
1869 for (Current = Line.First; Current != nullptr; Current = Current->Next) {
1871 Current->Role->precomputeFormattingInfos(Current);
1872 if (Current->MatchingParen &&
1873 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
1874 assert(IndentLevel > 0);
1877 Current->IndentLevel = IndentLevel;
1878 if (Current->opensBlockOrBlockTypeList(Style))
1882 DEBUG({ printDebugInfo(Line); });
1885 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
1886 unsigned UnbreakableTailLength = 0;
1887 FormatToken *Current = Line.Last;
1889 Current->UnbreakableTailLength = UnbreakableTailLength;
1890 if (Current->CanBreakBefore ||
1891 Current->isOneOf(tok::comment, tok::string_literal)) {
1892 UnbreakableTailLength = 0;
1894 UnbreakableTailLength +=
1895 Current->ColumnWidth + Current->SpacesRequiredBefore;
1897 Current = Current->Previous;
1901 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
1902 const FormatToken &Tok,
1903 bool InFunctionDecl) {
1904 const FormatToken &Left = *Tok.Previous;
1905 const FormatToken &Right = Tok;
1907 if (Left.is(tok::semi))
1910 if (Style.Language == FormatStyle::LK_Java) {
1911 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
1913 if (Right.is(Keywords.kw_implements))
1915 if (Left.is(tok::comma) && Left.NestingLevel == 0)
1917 } else if (Style.Language == FormatStyle::LK_JavaScript) {
1918 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
1920 if (Left.is(TT_JsTypeColon))
1922 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
1923 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
1927 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
1929 if (Right.is(tok::l_square)) {
1930 if (Style.Language == FormatStyle::LK_Proto)
1932 if (Left.is(tok::r_square))
1934 // Slightly prefer formatting local lambda definitions like functions.
1935 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
1937 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
1938 TT_ArrayInitializerLSquare))
1942 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
1943 Right.is(tok::kw_operator)) {
1944 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
1946 if (Left.is(TT_StartOfName))
1948 if (InFunctionDecl && Right.NestingLevel == 0)
1949 return Style.PenaltyReturnTypeOnItsOwnLine;
1952 if (Right.is(TT_PointerOrReference))
1954 if (Right.is(TT_LambdaArrow))
1956 if (Left.is(tok::equal) && Right.is(tok::l_brace))
1958 if (Left.is(TT_CastRParen))
1960 if (Left.is(tok::coloncolon) ||
1961 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
1963 if (Left.isOneOf(tok::kw_class, tok::kw_struct))
1965 if (Left.is(tok::comment))
1968 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon))
1971 if (Right.isMemberAccess()) {
1972 // Breaking before the "./->" of a chained call/member access is reasonably
1973 // cheap, as formatting those with one call per line is generally
1974 // desirable. In particular, it should be cheaper to break before the call
1975 // than it is to break inside a call's parameters, which could lead to weird
1976 // "hanging" indents. The exception is the very last "./->" to support this
1977 // frequent pattern:
1979 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
1982 // which might otherwise be blown up onto many lines. Here, clang-format
1983 // won't produce "hanging" indents anyway as there is no other trailing
1986 // Also apply higher penalty is not a call as that might lead to a wrapping
1990 // .aaaaaaaaa.bbbbbbbb(cccccccc);
1991 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
1996 if (Right.is(TT_TrailingAnnotation) &&
1997 (!Right.Next || Right.Next->isNot(tok::l_paren))) {
1998 // Moving trailing annotations to the next line is fine for ObjC method
2000 if (Line.startsWith(TT_ObjCMethodSpecifier))
2002 // Generally, breaking before a trailing annotation is bad unless it is
2003 // function-like. It seems to be especially preferable to keep standard
2004 // annotations (i.e. "const", "final" and "override") on the same line.
2005 // Use a slightly higher penalty after ")" so that annotations like
2006 // "const override" are kept together.
2007 bool is_short_annotation = Right.TokenText.size() < 10;
2008 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2011 // In for-loops, prefer breaking at ',' and ';'.
2012 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2015 // In Objective-C method expressions, prefer breaking before "param:" over
2016 // breaking after it.
2017 if (Right.is(TT_SelectorName))
2019 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2020 return Line.MightBeFunctionDecl ? 50 : 500;
2022 if (Left.is(tok::l_paren) && InFunctionDecl &&
2023 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2025 if (Left.is(tok::l_paren) && Left.Previous &&
2026 Left.Previous->isOneOf(tok::kw_if, tok::kw_for))
2028 if (Left.is(tok::equal) && InFunctionDecl)
2030 if (Right.is(tok::r_brace))
2032 if (Left.is(TT_TemplateOpener))
2034 if (Left.opensScope()) {
2035 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
2037 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2040 if (Left.is(TT_JavaAnnotation))
2043 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2044 Left.Previous->isLabelString() &&
2045 (Left.NextOperator || Left.OperatorIndex != 0))
2047 if (Right.is(tok::plus) && Left.isLabelString() &&
2048 (Right.NextOperator || Right.OperatorIndex != 0))
2050 if (Left.is(tok::comma))
2052 if (Right.is(tok::lessless) && Left.isLabelString() &&
2053 (Right.NextOperator || Right.OperatorIndex != 1))
2055 if (Right.is(tok::lessless)) {
2056 // Breaking at a << is really cheap.
2057 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2058 // Slightly prefer to break before the first one in log-like statements.
2062 if (Left.is(TT_ConditionalExpr))
2063 return prec::Conditional;
2064 prec::Level Level = Left.getPrecedence();
2065 if (Level != prec::Unknown)
2067 Level = Right.getPrecedence();
2068 if (Level != prec::Unknown)
2074 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2075 const FormatToken &Left,
2076 const FormatToken &Right) {
2077 if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2079 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2080 Left.Tok.getObjCKeywordID() == tok::objc_property)
2082 if (Right.is(tok::hashhash))
2083 return Left.is(tok::hash);
2084 if (Left.isOneOf(tok::hashhash, tok::hash))
2085 return Right.is(tok::hash);
2086 if (Left.is(tok::l_paren) && Right.is(tok::r_paren))
2087 return Style.SpaceInEmptyParentheses;
2088 if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2089 return (Right.is(TT_CastRParen) ||
2090 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2091 ? Style.SpacesInCStyleCastParentheses
2092 : Style.SpacesInParentheses;
2093 if (Right.isOneOf(tok::semi, tok::comma))
2095 if (Right.is(tok::less) &&
2096 Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList)
2098 if (Right.is(tok::less) && Left.is(tok::kw_template))
2099 return Style.SpaceAfterTemplateKeyword;
2100 if (Left.isOneOf(tok::exclaim, tok::tilde))
2102 if (Left.is(tok::at) &&
2103 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2104 tok::numeric_constant, tok::l_paren, tok::l_brace,
2105 tok::kw_true, tok::kw_false))
2107 if (Left.is(tok::colon))
2108 return !Left.is(TT_ObjCMethodExpr);
2109 if (Left.is(tok::coloncolon))
2111 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less))
2113 if (Right.is(tok::ellipsis))
2114 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2115 Left.Previous->is(tok::kw_case));
2116 if (Left.is(tok::l_square) && Right.is(tok::amp))
2118 if (Right.is(TT_PointerOrReference))
2119 return (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) ||
2120 (Left.Tok.isLiteral() || (Left.is(tok::kw_const) && Left.Previous &&
2121 Left.Previous->is(tok::r_paren)) ||
2122 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2123 (Style.PointerAlignment != FormatStyle::PAS_Left ||
2124 (Line.IsMultiVariableDeclStmt &&
2125 (Left.NestingLevel == 0 ||
2126 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2127 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2128 (!Left.is(TT_PointerOrReference) ||
2129 (Style.PointerAlignment != FormatStyle::PAS_Right &&
2130 !Line.IsMultiVariableDeclStmt)))
2132 if (Left.is(TT_PointerOrReference))
2133 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2134 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2135 !Right.is(TT_StartOfName)) ||
2136 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) ||
2137 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2139 (Style.PointerAlignment != FormatStyle::PAS_Right &&
2140 !Line.IsMultiVariableDeclStmt) &&
2142 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon));
2143 if (Right.is(tok::star) && Left.is(tok::l_paren))
2145 if (Left.is(tok::l_square))
2146 return (Left.is(TT_ArrayInitializerLSquare) &&
2147 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) ||
2148 (Left.is(TT_ArraySubscriptLSquare) && Style.SpacesInSquareBrackets &&
2149 Right.isNot(tok::r_square));
2150 if (Right.is(tok::r_square))
2151 return Right.MatchingParen &&
2152 ((Style.SpacesInContainerLiterals &&
2153 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) ||
2154 (Style.SpacesInSquareBrackets &&
2155 Right.MatchingParen->is(TT_ArraySubscriptLSquare)));
2156 if (Right.is(tok::l_square) &&
2157 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare) &&
2158 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral))
2160 if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2161 return !Left.Children.empty(); // No spaces in "{}".
2162 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) ||
2163 (Right.is(tok::r_brace) && Right.MatchingParen &&
2164 Right.MatchingParen->BlockKind != BK_Block))
2165 return !Style.Cpp11BracedListStyle;
2166 if (Left.is(TT_BlockComment))
2167 return !Left.TokenText.endswith("=*/");
2168 if (Right.is(tok::l_paren)) {
2169 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen))
2171 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
2172 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
2173 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while,
2174 tok::kw_switch, tok::kw_case, TT_ForEachMacro,
2176 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
2177 tok::kw_new, tok::kw_delete) &&
2178 (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
2179 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always &&
2180 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
2181 Left.is(tok::r_paren)) &&
2182 Line.Type != LT_PreprocessorDirective);
2184 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
2186 if (Right.is(TT_UnaryOperator))
2187 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
2188 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
2189 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
2191 Left.isSimpleTypeSpecifier()) &&
2192 Right.is(tok::l_brace) && Right.getNextNonComment() &&
2193 Right.BlockKind != BK_Block)
2195 if (Left.is(tok::period) || Right.is(tok::period))
2197 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
2199 if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
2200 Left.MatchingParen->Previous &&
2201 Left.MatchingParen->Previous->is(tok::period))
2202 // A.<B>DoSomething();
2204 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
2209 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
2210 const FormatToken &Right) {
2211 const FormatToken &Left = *Right.Previous;
2212 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
2213 return true; // Never ever merge two identifiers.
2214 if (Style.Language == FormatStyle::LK_Cpp) {
2215 if (Left.is(tok::kw_operator))
2216 return Right.is(tok::coloncolon);
2217 } else if (Style.Language == FormatStyle::LK_Proto) {
2218 if (Right.is(tok::period) &&
2219 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
2220 Keywords.kw_repeated, Keywords.kw_extend))
2222 if (Right.is(tok::l_paren) &&
2223 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
2225 } else if (Style.Language == FormatStyle::LK_JavaScript) {
2226 if (Left.is(TT_JsFatArrow))
2228 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2229 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2231 if (Left.is(tok::identifier) && Right.is(TT_TemplateString))
2233 if (Right.is(tok::star) &&
2234 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2236 if (Right.isOneOf(tok::l_brace, tok::l_square) &&
2237 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2239 // JS methods can use some keywords as names (e.g. `delete()`).
2240 if (Right.is(tok::l_paren) && Line.MustBeDeclaration &&
2241 Left.Tok.getIdentifierInfo())
2243 if (Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
2244 Keywords.kw_of, tok::kw_const) &&
2245 (!Left.Previous || !Left.Previous->is(tok::period)))
2247 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
2248 Left.Previous->is(tok::period) && Right.is(tok::l_paren))
2250 if (Left.is(Keywords.kw_as) &&
2251 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
2253 if (Left.is(tok::kw_default) && Left.Previous &&
2254 Left.Previous->is(tok::kw_export))
2256 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
2258 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
2260 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
2262 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
2263 Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
2265 if (Left.is(tok::ellipsis))
2267 if (Left.is(TT_TemplateCloser) &&
2268 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
2269 Keywords.kw_implements, Keywords.kw_extends))
2270 // Type assertions ('<type>expr') are not followed by whitespace. Other
2271 // locations that should have whitespace following are identified by the
2272 // above set of follower tokens.
2274 // Postfix non-null assertion operator, as in `foo!.bar()`.
2275 if (Right.is(tok::exclaim) && (Left.isOneOf(tok::identifier, tok::r_paren,
2276 tok::r_square, tok::r_brace) ||
2277 Left.Tok.isLiteral()))
2279 if (Left.is(tok::exclaim) && Right.is(Keywords.kw_as))
2280 return true; // "x! as string"
2281 } else if (Style.Language == FormatStyle::LK_Java) {
2282 if (Left.is(tok::r_square) && Right.is(tok::l_brace))
2284 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
2285 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
2286 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
2287 tok::kw_protected) ||
2288 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
2289 Keywords.kw_native)) &&
2290 Right.is(TT_TemplateOpener))
2293 if (Left.is(TT_ImplicitStringLiteral))
2294 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2295 if (Line.Type == LT_ObjCMethodDecl) {
2296 if (Left.is(TT_ObjCMethodSpecifier))
2298 if (Left.is(tok::r_paren) && Right.is(tok::identifier))
2299 // Don't space between ')' and <id>
2302 if (Line.Type == LT_ObjCProperty &&
2303 (Right.is(tok::equal) || Left.is(tok::equal)))
2306 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
2307 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
2309 if (Right.is(TT_OverloadedOperatorLParen))
2310 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2311 if (Left.is(tok::comma))
2313 if (Right.is(tok::comma))
2315 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen))
2317 if (Right.is(tok::colon)) {
2318 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
2319 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
2321 if (Right.is(TT_ObjCMethodExpr))
2323 if (Left.is(tok::question))
2325 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
2327 if (Right.is(TT_DictLiteral))
2328 return Style.SpacesInContainerLiterals;
2331 if (Left.is(TT_UnaryOperator))
2332 return Right.is(TT_BinaryOperator);
2334 // If the next token is a binary operator or a selector name, we have
2335 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
2336 if (Left.is(TT_CastRParen))
2337 return Style.SpaceAfterCStyleCast ||
2338 Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
2340 if (Left.is(tok::greater) && Right.is(tok::greater))
2341 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
2342 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
2343 if (Right.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
2344 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar))
2346 if (!Style.SpaceBeforeAssignmentOperators &&
2347 Right.getPrecedence() == prec::Assignment)
2349 if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
2350 // Generally don't remove existing spaces between an identifier and "::".
2351 // The identifier might actually be a macro name such as ALWAYS_INLINE. If
2352 // this turns out to be too lenient, add analysis of the identifier itself.
2353 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2354 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment))
2355 return (Left.is(TT_TemplateOpener) &&
2356 Style.Standard == FormatStyle::LS_Cpp03) ||
2357 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
2358 TT_TemplateCloser, TT_TemplateOpener));
2359 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
2360 return Style.SpacesInAngles;
2361 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
2362 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
2363 !Right.is(tok::r_paren)))
2365 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
2366 Right.isNot(TT_FunctionTypeLParen))
2367 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2368 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
2369 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
2371 if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
2372 Line.startsWith(tok::hash))
2374 if (Right.is(TT_TrailingUnaryOperator))
2376 if (Left.is(TT_RegexLiteral))
2378 return spaceRequiredBetween(Line, Left, Right);
2381 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
2382 static bool isAllmanBrace(const FormatToken &Tok) {
2383 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block &&
2384 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral);
2387 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
2388 const FormatToken &Right) {
2389 const FormatToken &Left = *Right.Previous;
2390 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
2393 if (Style.Language == FormatStyle::LK_JavaScript) {
2394 // FIXME: This might apply to other languages and token kinds.
2395 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
2396 Left.Previous->is(tok::string_literal))
2398 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
2399 Left.Previous && Left.Previous->is(tok::equal) &&
2400 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
2402 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
2404 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
2405 // Object literals on the top level of a file are treated as "enum-style".
2406 // Each key/value pair is put on a separate line, instead of bin-packing.
2408 if (Left.is(tok::l_brace) && Line.Level == 0 &&
2409 (Line.startsWith(tok::kw_enum) ||
2410 Line.startsWith(tok::kw_export, tok::kw_enum)))
2411 // JavaScript top-level enum key/value pairs are put on separate lines
2412 // instead of bin-packing.
2414 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
2415 !Left.Children.empty())
2416 // Support AllowShortFunctionsOnASingleLine for JavaScript.
2417 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
2418 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
2419 (Left.NestingLevel == 0 && Line.Level == 0 &&
2420 Style.AllowShortFunctionsOnASingleLine ==
2421 FormatStyle::SFS_Inline);
2422 } else if (Style.Language == FormatStyle::LK_Java) {
2423 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
2424 Right.Next->is(tok::string_literal))
2426 } else if (Style.Language == FormatStyle::LK_Cpp ||
2427 Style.Language == FormatStyle::LK_ObjC ||
2428 Style.Language == FormatStyle::LK_Proto) {
2429 if (Left.isStringLiteral() &&
2430 (Right.isStringLiteral() || Right.is(TT_ObjCStringLiteral)))
2434 // If the last token before a '}' is a comma or a trailing comment, the
2435 // intention is to insert a line break after it in order to make shuffling
2436 // around entries easier.
2437 const FormatToken *BeforeClosingBrace = nullptr;
2438 if (Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) &&
2439 Left.BlockKind != BK_Block && Left.MatchingParen)
2440 BeforeClosingBrace = Left.MatchingParen->Previous;
2441 else if (Right.MatchingParen &&
2442 Right.MatchingParen->isOneOf(tok::l_brace,
2443 TT_ArrayInitializerLSquare))
2444 BeforeClosingBrace = &Left;
2445 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
2446 BeforeClosingBrace->isTrailingComment()))
2449 if (Right.is(tok::comment))
2450 return Left.BlockKind != BK_BracedInit &&
2451 Left.isNot(TT_CtorInitializerColon) &&
2452 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
2453 if (Left.isTrailingComment())
2455 if (Right.Previous->IsUnterminatedLiteral)
2457 if (Right.is(tok::lessless) && Right.Next &&
2458 Right.Previous->is(tok::string_literal) &&
2459 Right.Next->is(tok::string_literal))
2461 if (Right.Previous->ClosesTemplateDeclaration &&
2462 Right.Previous->MatchingParen &&
2463 Right.Previous->MatchingParen->NestingLevel == 0 &&
2464 Style.AlwaysBreakTemplateDeclarations)
2466 if ((Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) &&
2467 Style.BreakConstructorInitializersBeforeComma &&
2468 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
2470 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
2471 // Raw string literals are special wrt. line breaks. The author has made a
2472 // deliberate choice and might have aligned the contents of the string
2473 // literal accordingly. Thus, we try keep existing line breaks.
2474 return Right.NewlinesBefore > 0;
2475 if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 &&
2476 Style.Language == FormatStyle::LK_Proto)
2477 // Don't put enums onto single lines in protocol buffers.
2479 if (Right.is(TT_InlineASMBrace))
2480 return Right.HasUnescapedNewline;
2481 if (isAllmanBrace(Left) || isAllmanBrace(Right))
2482 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
2483 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
2484 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
2485 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine)
2488 if ((Style.Language == FormatStyle::LK_Java ||
2489 Style.Language == FormatStyle::LK_JavaScript) &&
2490 Left.is(TT_LeadingJavaAnnotation) &&
2491 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
2492 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
2498 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
2499 const FormatToken &Right) {
2500 const FormatToken &Left = *Right.Previous;
2502 // Language-specific stuff.
2503 if (Style.Language == FormatStyle::LK_Java) {
2504 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
2505 Keywords.kw_implements))
2507 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
2508 Keywords.kw_implements))
2510 } else if (Style.Language == FormatStyle::LK_JavaScript) {
2511 const FormatToken *NonComment = Right.getPreviousNonComment();
2512 if (Left.isOneOf(tok::kw_return, tok::kw_continue, tok::kw_break,
2515 NonComment->isOneOf(tok::kw_return, tok::kw_continue, tok::kw_break,
2517 return false; // Otherwise a semicolon is inserted.
2518 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
2520 if (Left.is(TT_JsTypeColon))
2522 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is))
2524 if (Left.is(Keywords.kw_in))
2525 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
2526 if (Right.is(Keywords.kw_in))
2527 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
2528 if (Right.is(Keywords.kw_as))
2529 return false; // must not break before as in 'x as type' casts
2530 if (Left.is(Keywords.kw_declare) &&
2531 Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
2532 Keywords.kw_function, tok::kw_class, tok::kw_enum,
2533 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
2534 Keywords.kw_let, tok::kw_const))
2535 // See grammar for 'declare' statements at:
2536 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
2538 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
2539 Right.isOneOf(tok::identifier, tok::string_literal)) {
2540 return false; // must not break in "module foo { ...}"
2544 if (Left.is(tok::at))
2546 if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
2548 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
2549 return !Right.is(tok::l_paren);
2550 if (Right.is(TT_PointerOrReference))
2551 return Line.IsMultiVariableDeclStmt ||
2552 (Style.PointerAlignment == FormatStyle::PAS_Right &&
2553 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
2554 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2555 Right.is(tok::kw_operator))
2557 if (Left.is(TT_PointerOrReference))
2559 if (Right.isTrailingComment())
2560 // We rely on MustBreakBefore being set correctly here as we should not
2561 // change the "binding" behavior of a comment.
2562 // The first comment in a braced lists is always interpreted as belonging to
2563 // the first list element. Otherwise, it should be placed outside of the
2565 return Left.BlockKind == BK_BracedInit;
2566 if (Left.is(tok::question) && Right.is(tok::colon))
2568 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
2569 return Style.BreakBeforeTernaryOperators;
2570 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
2571 return !Style.BreakBeforeTernaryOperators;
2572 if (Right.is(TT_InheritanceColon))
2574 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
2575 Left.isNot(TT_SelectorName))
2577 if (Right.is(tok::colon) &&
2578 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
2580 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr))
2582 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
2583 Right.Next->is(TT_ObjCMethodExpr)))
2584 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
2585 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
2587 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
2589 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
2590 TT_OverloadedOperator))
2592 if (Left.is(TT_RangeBasedForLoopColon))
2594 if (Right.is(TT_RangeBasedForLoopColon))
2596 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
2598 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
2599 Left.is(tok::kw_operator))
2601 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
2602 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
2604 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen))
2606 if (Left.is(tok::l_paren) && Left.Previous &&
2607 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
2609 if (Right.is(TT_ImplicitStringLiteral))
2612 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
2614 if (Right.is(tok::r_square) && Right.MatchingParen &&
2615 Right.MatchingParen->is(TT_LambdaLSquare))
2618 // We only break before r_brace if there was a corresponding break before
2619 // the l_brace, which is tracked by BreakBeforeClosingBrace.
2620 if (Right.is(tok::r_brace))
2621 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block;
2623 // Allow breaking after a trailing annotation, e.g. after a method
2625 if (Left.is(TT_TrailingAnnotation))
2626 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
2627 tok::less, tok::coloncolon);
2629 if (Right.is(tok::kw___attribute))
2632 if (Left.is(tok::identifier) && Right.is(tok::string_literal))
2635 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2638 if (Left.is(TT_CtorInitializerComma) &&
2639 Style.BreakConstructorInitializersBeforeComma)
2641 if (Right.is(TT_CtorInitializerComma) &&
2642 Style.BreakConstructorInitializersBeforeComma)
2644 if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
2645 (Left.is(tok::less) && Right.is(tok::less)))
2647 if (Right.is(TT_BinaryOperator) &&
2648 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
2649 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
2650 Right.getPrecedence() != prec::Assignment))
2652 if (Left.is(TT_ArrayInitializerLSquare))
2654 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
2656 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
2657 !Left.isOneOf(tok::arrowstar, tok::lessless) &&
2658 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
2659 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
2660 Left.getPrecedence() == prec::Assignment))
2662 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
2663 tok::kw_class, tok::kw_struct, tok::comment) ||
2664 Right.isMemberAccess() ||
2665 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
2666 tok::colon, tok::l_square, tok::at) ||
2667 (Left.is(tok::r_paren) &&
2668 Right.isOneOf(tok::identifier, tok::kw_const)) ||
2669 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
2670 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
2673 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
2674 llvm::errs() << "AnnotatedTokens:\n";
2675 const FormatToken *Tok = Line.First;
2677 llvm::errs() << " M=" << Tok->MustBreakBefore
2678 << " C=" << Tok->CanBreakBefore
2679 << " T=" << getTokenTypeName(Tok->Type)
2680 << " S=" << Tok->SpacesRequiredBefore
2681 << " B=" << Tok->BlockParameterCount
2682 << " BK=" << Tok->BlockKind
2683 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName()
2684 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind
2686 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
2687 llvm::errs() << Tok->FakeLParens[i] << "/";
2688 llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
2689 llvm::errs() << " Text='" << Tok->TokenText << "'\n";
2691 assert(Tok == Line.Last);
2694 llvm::errs() << "----\n";
2697 } // namespace format
2698 } // namespace clang