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.IsCpp() && Parent &&
315 Parent->is(TT_TemplateCloser) &&
316 (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
317 Contexts.back().InTemplateArgument);
319 bool StartsObjCMethodExpr =
320 !CppArrayTemplates && Style.IsCpp() &&
321 Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) &&
322 CurrentToken->isNot(tok::l_brace) &&
324 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
325 tok::kw_return, tok::kw_throw) ||
326 Parent->isUnaryOperator() ||
327 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
328 getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown);
329 bool ColonFound = false;
331 unsigned BindingIncrease = 1;
332 if (Left->is(TT_Unknown)) {
333 if (StartsObjCMethodExpr) {
334 Left->Type = TT_ObjCMethodExpr;
335 } else if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
336 Contexts.back().ContextKind == tok::l_brace &&
337 Parent->isOneOf(tok::l_brace, tok::comma)) {
338 Left->Type = TT_JsComputedPropertyName;
339 } else if (CurrentToken->is(tok::r_square) && Parent &&
340 Parent->is(TT_TemplateCloser)) {
341 Left->Type = TT_ArraySubscriptLSquare;
342 } else if (Style.Language == FormatStyle::LK_Proto ||
343 (!CppArrayTemplates && Parent &&
344 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
345 tok::comma, tok::l_paren, tok::l_square,
346 tok::question, tok::colon, tok::kw_return,
347 // Should only be relevant to JavaScript:
349 Left->Type = TT_ArrayInitializerLSquare;
351 BindingIncrease = 10;
352 Left->Type = TT_ArraySubscriptLSquare;
356 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
357 Contexts.back().IsExpression = true;
358 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
360 while (CurrentToken) {
361 if (CurrentToken->is(tok::r_square)) {
362 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_paren) &&
363 Left->is(TT_ObjCMethodExpr)) {
364 // An ObjC method call is rarely followed by an open parenthesis.
365 // FIXME: Do we incorrectly label ":" with this?
366 StartsObjCMethodExpr = false;
367 Left->Type = TT_Unknown;
369 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
370 CurrentToken->Type = TT_ObjCMethodExpr;
371 // determineStarAmpUsage() thinks that '*' '[' is allocating an
372 // array of pointers, but if '[' starts a selector then '*' is a
374 if (Parent && Parent->is(TT_PointerOrReference))
375 Parent->Type = TT_BinaryOperator;
377 Left->MatchingParen = CurrentToken;
378 CurrentToken->MatchingParen = Left;
379 if (Contexts.back().FirstObjCSelectorName) {
380 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
381 Contexts.back().LongestObjCSelectorName;
382 if (Left->BlockParameterCount > 1)
383 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
388 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
390 if (CurrentToken->is(tok::colon)) {
391 if (Left->is(TT_ArraySubscriptLSquare)) {
392 Left->Type = TT_ObjCMethodExpr;
393 StartsObjCMethodExpr = true;
394 Contexts.back().ColonIsObjCMethodExpr = true;
395 if (Parent && Parent->is(tok::r_paren))
396 Parent->Type = TT_CastRParen;
400 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
402 Left->Type = TT_ArrayInitializerLSquare;
403 FormatToken *Tok = CurrentToken;
406 updateParameterCount(Left, Tok);
413 FormatToken *Left = CurrentToken->Previous;
414 Left->ParentBracket = Contexts.back().ContextKind;
416 if (Contexts.back().CaretFound)
417 Left->Type = TT_ObjCBlockLBrace;
418 Contexts.back().CaretFound = false;
420 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
421 Contexts.back().ColonIsDictLiteral = true;
422 if (Left->BlockKind == BK_BracedInit)
423 Contexts.back().IsExpression = true;
425 while (CurrentToken) {
426 if (CurrentToken->is(tok::r_brace)) {
427 Left->MatchingParen = CurrentToken;
428 CurrentToken->MatchingParen = Left;
432 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
434 updateParameterCount(Left, CurrentToken);
435 if (CurrentToken->isOneOf(tok::colon, tok::l_brace)) {
436 FormatToken *Previous = CurrentToken->getPreviousNonComment();
437 if (((CurrentToken->is(tok::colon) &&
438 (!Contexts.back().ColonIsDictLiteral || !Style.IsCpp())) ||
439 Style.Language == FormatStyle::LK_Proto) &&
440 (Previous->Tok.getIdentifierInfo() ||
441 Previous->is(tok::string_literal)))
442 Previous->Type = TT_SelectorName;
443 if (CurrentToken->is(tok::colon) ||
444 Style.Language == FormatStyle::LK_JavaScript)
445 Left->Type = TT_DictLiteral;
447 if (CurrentToken->is(tok::comma) &&
448 Style.Language == FormatStyle::LK_JavaScript)
449 Left->Type = TT_DictLiteral;
457 void updateParameterCount(FormatToken *Left, FormatToken *Current) {
458 if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block)
459 ++Left->BlockParameterCount;
460 if (Current->is(tok::comma)) {
461 ++Left->ParameterCount;
463 Left->Role.reset(new CommaSeparatedList(Style));
464 Left->Role->CommaFound(Current);
465 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
466 Left->ParameterCount = 1;
470 bool parseConditional() {
471 while (CurrentToken) {
472 if (CurrentToken->is(tok::colon)) {
473 CurrentToken->Type = TT_ConditionalExpr;
483 bool parseTemplateDeclaration() {
484 if (CurrentToken && CurrentToken->is(tok::less)) {
485 CurrentToken->Type = TT_TemplateOpener;
490 CurrentToken->Previous->ClosesTemplateDeclaration = true;
496 bool consumeToken() {
497 FormatToken *Tok = CurrentToken;
499 switch (Tok->Tok.getKind()) {
502 if (!Tok->Previous && Line.MustBeDeclaration)
503 Tok->Type = TT_ObjCMethodSpecifier;
508 // Colons from ?: are handled in parseConditional().
509 if (Style.Language == FormatStyle::LK_JavaScript) {
510 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
511 (Contexts.size() == 1 && // switch/case labels
512 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
513 Contexts.back().ContextKind == tok::l_paren || // function params
514 Contexts.back().ContextKind == tok::l_square || // array type
515 (Contexts.size() == 1 &&
516 Line.MustBeDeclaration)) { // method/property declaration
517 Tok->Type = TT_JsTypeColon;
521 if (Contexts.back().ColonIsDictLiteral ||
522 Style.Language == FormatStyle::LK_Proto) {
523 Tok->Type = TT_DictLiteral;
524 } else if (Contexts.back().ColonIsObjCMethodExpr ||
525 Line.startsWith(TT_ObjCMethodSpecifier)) {
526 Tok->Type = TT_ObjCMethodExpr;
527 const FormatToken *BeforePrevious = Tok->Previous->Previous;
528 if (!BeforePrevious ||
529 !(BeforePrevious->is(TT_CastRParen) ||
530 (BeforePrevious->is(TT_ObjCMethodExpr) &&
531 BeforePrevious->is(tok::colon))) ||
532 BeforePrevious->is(tok::r_square) ||
533 Contexts.back().LongestObjCSelectorName == 0) {
534 Tok->Previous->Type = TT_SelectorName;
535 if (Tok->Previous->ColumnWidth >
536 Contexts.back().LongestObjCSelectorName)
537 Contexts.back().LongestObjCSelectorName =
538 Tok->Previous->ColumnWidth;
539 if (!Contexts.back().FirstObjCSelectorName)
540 Contexts.back().FirstObjCSelectorName = Tok->Previous;
542 } else if (Contexts.back().ColonIsForRangeExpr) {
543 Tok->Type = TT_RangeBasedForLoopColon;
544 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
545 Tok->Type = TT_BitFieldColon;
546 } else if (Contexts.size() == 1 &&
547 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) {
548 if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren,
550 Tok->Type = TT_CtorInitializerColon;
552 Tok->Type = TT_InheritanceColon;
553 } else if (Tok->Previous->is(tok::identifier) && Tok->Next &&
554 Tok->Next->isOneOf(tok::r_paren, tok::comma)) {
555 // This handles a special macro in ObjC code where selectors including
556 // the colon are passed as macro arguments.
557 Tok->Type = TT_ObjCMethodExpr;
558 } else if (Contexts.back().ContextKind == tok::l_paren) {
559 Tok->Type = TT_InlineASMColon;
564 // | and & in declarations/type expressions represent union and
565 // intersection types, respectively.
566 if (Style.Language == FormatStyle::LK_JavaScript &&
567 !Contexts.back().IsExpression)
568 Tok->Type = TT_JsTypeOperator;
572 if (CurrentToken && CurrentToken->is(tok::l_paren)) {
574 if (!parseParens(/*LookForDecls=*/true))
579 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Previous &&
580 Tok->Previous->is(tok::period))
582 Contexts.back().ColonIsForRangeExpr = true;
588 // When faced with 'operator()()', the kw_operator handler incorrectly
589 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
590 // the first two parens OverloadedOperators and the second l_paren an
591 // OverloadedOperatorLParen.
593 Tok->Previous->is(tok::r_paren) &&
594 Tok->Previous->MatchingParen &&
595 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
596 Tok->Previous->Type = TT_OverloadedOperator;
597 Tok->Previous->MatchingParen->Type = TT_OverloadedOperator;
598 Tok->Type = TT_OverloadedOperatorLParen;
603 if (Line.MustBeDeclaration && Contexts.size() == 1 &&
604 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
606 !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute,
607 TT_LeadingJavaAnnotation)))
608 Line.MightBeFunctionDecl = true;
620 Tok->Type = TT_TemplateOpener;
622 Tok->Type = TT_BinaryOperator;
623 NonTemplateLess.insert(Tok);
632 // Lines can start with '}'.
637 Tok->Type = TT_BinaryOperator;
639 case tok::kw_operator:
640 while (CurrentToken &&
641 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
642 if (CurrentToken->isOneOf(tok::star, tok::amp))
643 CurrentToken->Type = TT_PointerOrReference;
646 CurrentToken->Previous->isOneOf(TT_BinaryOperator, tok::comma))
647 CurrentToken->Previous->Type = TT_OverloadedOperator;
650 CurrentToken->Type = TT_OverloadedOperatorLParen;
651 if (CurrentToken->Previous->is(TT_BinaryOperator))
652 CurrentToken->Previous->Type = TT_OverloadedOperator;
656 if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next &&
657 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
659 // Question marks before semicolons, colons, etc. indicate optional
660 // types (fields, parameters), e.g.
661 // function(x?: string, y?) {...}
663 Tok->Type = TT_JsTypeOptionalQuestion;
666 // Declarations cannot be conditional expressions, this can only be part
667 // of a type declaration.
668 if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
669 Style.Language == FormatStyle::LK_JavaScript)
673 case tok::kw_template:
674 parseTemplateDeclaration();
677 if (Contexts.back().InCtorInitializer)
678 Tok->Type = TT_CtorInitializerComma;
679 else if (Contexts.back().FirstStartOfName &&
680 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
681 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
682 Line.IsMultiVariableDeclStmt = true;
684 if (Contexts.back().IsForEachMacro)
685 Contexts.back().IsExpression = true;
687 case tok::identifier:
688 if (Tok->isOneOf(Keywords.kw___has_include,
689 Keywords.kw___has_include_next)) {
699 void parseIncludeDirective() {
700 if (CurrentToken && CurrentToken->is(tok::less)) {
702 while (CurrentToken) {
703 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next)
704 CurrentToken->Type = TT_ImplicitStringLiteral;
710 void parseWarningOrError() {
712 // We still want to format the whitespace left of the first token of the
715 while (CurrentToken) {
716 CurrentToken->Type = TT_ImplicitStringLiteral;
722 next(); // Consume "pragma".
724 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
725 bool IsMark = CurrentToken->is(Keywords.kw_mark);
726 next(); // Consume "mark".
727 next(); // Consume first token (so we fix leading whitespace).
728 while (CurrentToken) {
729 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
730 CurrentToken->Type = TT_ImplicitStringLiteral;
736 void parseHasInclude() {
737 if (!CurrentToken || !CurrentToken->is(tok::l_paren))
740 parseIncludeDirective();
744 LineType parsePreprocessorDirective() {
745 bool IsFirstToken = CurrentToken->IsFirst;
746 LineType Type = LT_PreprocessorDirective;
751 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
752 // JavaScript files can contain shebang lines of the form:
753 // #!/usr/bin/env node
754 // Treat these like C++ #include directives.
755 while (CurrentToken) {
756 // Tokens cannot be comments here.
757 CurrentToken->Type = TT_ImplicitStringLiteral;
760 return LT_ImportStatement;
763 if (CurrentToken->Tok.is(tok::numeric_constant)) {
764 CurrentToken->SpacesRequiredBefore = 1;
767 // Hashes in the middle of a line can lead to any strange token
769 if (!CurrentToken->Tok.getIdentifierInfo())
771 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
772 case tok::pp_include:
773 case tok::pp_include_next:
776 parseIncludeDirective();
777 Type = LT_ImportStatement;
780 case tok::pp_warning:
781 parseWarningOrError();
788 Contexts.back().IsExpression = true;
794 while (CurrentToken) {
795 FormatToken *Tok = CurrentToken;
797 if (Tok->isOneOf(Keywords.kw___has_include,
798 Keywords.kw___has_include_next)) {
806 LineType parseLine() {
807 NonTemplateLess.clear();
808 if (CurrentToken->is(tok::hash))
809 return parsePreprocessorDirective();
811 // Directly allow to 'import <string-literal>' to support protocol buffer
812 // definitions (code.google.com/p/protobuf) or missing "#" (either way we
813 // should not break the line).
814 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
815 if ((Style.Language == FormatStyle::LK_Java &&
816 CurrentToken->is(Keywords.kw_package)) ||
817 (Info && Info->getPPKeywordID() == tok::pp_import &&
818 CurrentToken->Next &&
819 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
822 parseIncludeDirective();
823 return LT_ImportStatement;
826 // If this line starts and ends in '<' and '>', respectively, it is likely
827 // part of "#define <a/b.h>".
828 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
829 parseIncludeDirective();
830 return LT_ImportStatement;
833 // In .proto files, top-level options are very similar to import statements
834 // and should not be line-wrapped.
835 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
836 CurrentToken->is(Keywords.kw_option)) {
838 if (CurrentToken && CurrentToken->is(tok::identifier))
839 return LT_ImportStatement;
842 bool KeywordVirtualFound = false;
843 bool ImportStatement = false;
845 // import {...} from '...';
846 if (Style.Language == FormatStyle::LK_JavaScript &&
847 CurrentToken->is(Keywords.kw_import))
848 ImportStatement = true;
850 while (CurrentToken) {
851 if (CurrentToken->is(tok::kw_virtual))
852 KeywordVirtualFound = true;
853 if (Style.Language == FormatStyle::LK_JavaScript) {
854 // export {...} from '...';
855 // An export followed by "from 'some string';" is a re-export from
856 // another module identified by a URI and is treated as a
857 // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
858 // Just "export {...};" or "export class ..." should not be treated as
859 // an import in this sense.
860 if (Line.First->is(tok::kw_export) &&
861 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
862 CurrentToken->Next->isStringLiteral())
863 ImportStatement = true;
864 if (isClosureImportStatement(*CurrentToken))
865 ImportStatement = true;
870 if (KeywordVirtualFound)
871 return LT_VirtualFunctionDecl;
873 return LT_ImportStatement;
875 if (Line.startsWith(TT_ObjCMethodSpecifier)) {
876 if (Contexts.back().FirstObjCSelectorName)
877 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
878 Contexts.back().LongestObjCSelectorName;
879 return LT_ObjCMethodDecl;
886 bool isClosureImportStatement(const FormatToken &Tok) {
887 // FIXME: Closure-library specific stuff should not be hard-coded but be
889 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
890 Tok.Next->Next && (Tok.Next->Next->TokenText == "module" ||
891 Tok.Next->Next->TokenText == "provide" ||
892 Tok.Next->Next->TokenText == "require" ||
893 Tok.Next->Next->TokenText == "setTestOnly" ||
894 Tok.Next->Next->TokenText == "forwardDeclare") &&
895 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
898 void resetTokenMetadata(FormatToken *Token) {
902 // Reset token type in case we have already looked at it and then
903 // recovered from an error (e.g. failure to find the matching >).
904 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro,
905 TT_FunctionLBrace, TT_ImplicitStringLiteral,
906 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow,
907 TT_OverloadedOperator, TT_RegexLiteral,
909 CurrentToken->Type = TT_Unknown;
910 CurrentToken->Role.reset();
911 CurrentToken->MatchingParen = nullptr;
912 CurrentToken->FakeLParens.clear();
913 CurrentToken->FakeRParens = 0;
918 CurrentToken->NestingLevel = Contexts.size() - 1;
919 CurrentToken->BindingStrength = Contexts.back().BindingStrength;
920 modifyContext(*CurrentToken);
921 determineTokenType(*CurrentToken);
922 CurrentToken = CurrentToken->Next;
925 resetTokenMetadata(CurrentToken);
928 /// \brief A struct to hold information valid in a specific context, e.g.
929 /// a pair of parenthesis.
931 Context(tok::TokenKind ContextKind, unsigned BindingStrength,
933 : ContextKind(ContextKind), BindingStrength(BindingStrength),
934 IsExpression(IsExpression) {}
936 tok::TokenKind ContextKind;
937 unsigned BindingStrength;
939 unsigned LongestObjCSelectorName = 0;
940 bool ColonIsForRangeExpr = false;
941 bool ColonIsDictLiteral = false;
942 bool ColonIsObjCMethodExpr = false;
943 FormatToken *FirstObjCSelectorName = nullptr;
944 FormatToken *FirstStartOfName = nullptr;
945 bool CanBeExpression = true;
946 bool InTemplateArgument = false;
947 bool InCtorInitializer = false;
948 bool CaretFound = false;
949 bool IsForEachMacro = false;
952 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime
953 /// of each instance.
954 struct ScopedContextCreator {
957 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
960 P.Contexts.push_back(Context(ContextKind,
961 P.Contexts.back().BindingStrength + Increase,
962 P.Contexts.back().IsExpression));
965 ~ScopedContextCreator() { P.Contexts.pop_back(); }
968 void modifyContext(const FormatToken &Current) {
969 if (Current.getPrecedence() == prec::Assignment &&
970 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
971 // Type aliases use `type X = ...;` in TypeScript.
972 !(Style.Language == FormatStyle::LK_JavaScript &&
973 Line.startsWith(Keywords.kw_type, tok::identifier)) &&
974 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
975 Contexts.back().IsExpression = true;
976 if (!Line.startsWith(TT_UnaryOperator)) {
977 for (FormatToken *Previous = Current.Previous;
978 Previous && Previous->Previous &&
979 !Previous->Previous->isOneOf(tok::comma, tok::semi);
980 Previous = Previous->Previous) {
981 if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
982 Previous = Previous->MatchingParen;
986 if (Previous->opensScope())
988 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
989 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
990 Previous->Previous && Previous->Previous->isNot(tok::equal))
991 Previous->Type = TT_PointerOrReference;
994 } else if (Current.is(tok::lessless) &&
995 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
996 Contexts.back().IsExpression = true;
997 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
998 Contexts.back().IsExpression = true;
999 } else if (Current.is(TT_TrailingReturnArrow)) {
1000 Contexts.back().IsExpression = false;
1001 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1002 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1003 } else if (Current.Previous &&
1004 Current.Previous->is(TT_CtorInitializerColon)) {
1005 Contexts.back().IsExpression = true;
1006 Contexts.back().InCtorInitializer = true;
1007 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1008 for (FormatToken *Previous = Current.Previous;
1009 Previous && Previous->isOneOf(tok::star, tok::amp);
1010 Previous = Previous->Previous)
1011 Previous->Type = TT_PointerOrReference;
1012 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1013 Contexts.back().IsExpression = false;
1014 } else if (Current.is(tok::kw_new)) {
1015 Contexts.back().CanBeExpression = false;
1016 } else if (Current.isOneOf(tok::semi, tok::exclaim)) {
1017 // This should be the condition or increment in a for-loop.
1018 Contexts.back().IsExpression = true;
1022 void determineTokenType(FormatToken &Current) {
1023 if (!Current.is(TT_Unknown))
1024 // The token type is already known.
1027 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1028 // function declaration have been found. In this case, 'Current' is a
1029 // trailing token of this declaration and thus cannot be a name.
1030 if (Current.is(Keywords.kw_instanceof)) {
1031 Current.Type = TT_BinaryOperator;
1032 } else if (isStartOfName(Current) &&
1033 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1034 Contexts.back().FirstStartOfName = &Current;
1035 Current.Type = TT_StartOfName;
1036 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1038 } else if (Current.is(tok::arrow) &&
1039 Style.Language == FormatStyle::LK_Java) {
1040 Current.Type = TT_LambdaArrow;
1041 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1042 Current.NestingLevel == 0) {
1043 Current.Type = TT_TrailingReturnArrow;
1044 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1046 determineStarAmpUsage(Current, Contexts.back().CanBeExpression &&
1047 Contexts.back().IsExpression,
1048 Contexts.back().InTemplateArgument);
1049 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1050 Current.Type = determinePlusMinusCaretUsage(Current);
1051 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1052 Contexts.back().CaretFound = true;
1053 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1054 Current.Type = determineIncrementUsage(Current);
1055 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1056 Current.Type = TT_UnaryOperator;
1057 } else if (Current.is(tok::question)) {
1058 if (Style.Language == FormatStyle::LK_JavaScript &&
1059 Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1060 // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1061 // on the interface, not a ternary expression.
1062 Current.Type = TT_JsTypeOptionalQuestion;
1064 Current.Type = TT_ConditionalExpr;
1066 } else if (Current.isBinaryOperator() &&
1067 (!Current.Previous || Current.Previous->isNot(tok::l_square))) {
1068 Current.Type = TT_BinaryOperator;
1069 } else if (Current.is(tok::comment)) {
1070 if (Current.TokenText.startswith("/*")) {
1071 if (Current.TokenText.endswith("*/"))
1072 Current.Type = TT_BlockComment;
1074 // The lexer has for some reason determined a comment here. But we
1075 // cannot really handle it, if it isn't properly terminated.
1076 Current.Tok.setKind(tok::unknown);
1078 Current.Type = TT_LineComment;
1080 } else if (Current.is(tok::r_paren)) {
1081 if (rParenEndsCast(Current))
1082 Current.Type = TT_CastRParen;
1083 if (Current.MatchingParen && Current.Next &&
1084 !Current.Next->isBinaryOperator() &&
1085 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1086 tok::comma, tok::period, tok::arrow,
1088 if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1089 // Make sure this isn't the return type of an Obj-C block declaration
1090 if (AfterParen->Tok.isNot(tok::caret)) {
1091 if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1092 if (BeforeParen->is(tok::identifier) &&
1093 BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1094 (!BeforeParen->Previous ||
1095 BeforeParen->Previous->ClosesTemplateDeclaration))
1096 Current.Type = TT_FunctionAnnotationRParen;
1099 } else if (Current.is(tok::at) && Current.Next) {
1100 if (Current.Next->isStringLiteral()) {
1101 Current.Type = TT_ObjCStringLiteral;
1103 switch (Current.Next->Tok.getObjCKeywordID()) {
1104 case tok::objc_interface:
1105 case tok::objc_implementation:
1106 case tok::objc_protocol:
1107 Current.Type = TT_ObjCDecl;
1109 case tok::objc_property:
1110 Current.Type = TT_ObjCProperty;
1116 } else if (Current.is(tok::period)) {
1117 FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1118 if (PreviousNoComment &&
1119 PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1120 Current.Type = TT_DesignatedInitializerPeriod;
1121 else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1122 Current.Previous->isOneOf(TT_JavaAnnotation,
1123 TT_LeadingJavaAnnotation)) {
1124 Current.Type = Current.Previous->Type;
1126 } else if (Current.isOneOf(tok::identifier, tok::kw_const) &&
1128 !Current.Previous->isOneOf(tok::equal, tok::at) &&
1129 Line.MightBeFunctionDecl && Contexts.size() == 1) {
1130 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1131 // function declaration have been found.
1132 Current.Type = TT_TrailingAnnotation;
1133 } else if ((Style.Language == FormatStyle::LK_Java ||
1134 Style.Language == FormatStyle::LK_JavaScript) &&
1136 if (Current.Previous->is(tok::at) &&
1137 Current.isNot(Keywords.kw_interface)) {
1138 const FormatToken &AtToken = *Current.Previous;
1139 const FormatToken *Previous = AtToken.getPreviousNonComment();
1140 if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1141 Current.Type = TT_LeadingJavaAnnotation;
1143 Current.Type = TT_JavaAnnotation;
1144 } else if (Current.Previous->is(tok::period) &&
1145 Current.Previous->isOneOf(TT_JavaAnnotation,
1146 TT_LeadingJavaAnnotation)) {
1147 Current.Type = Current.Previous->Type;
1152 /// \brief Take a guess at whether \p Tok starts a name of a function or
1153 /// variable declaration.
1155 /// This is a heuristic based on whether \p Tok is an identifier following
1156 /// something that is likely a type.
1157 bool isStartOfName(const FormatToken &Tok) {
1158 if (Tok.isNot(tok::identifier) || !Tok.Previous)
1161 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1164 if (Style.Language == FormatStyle::LK_JavaScript &&
1165 Tok.Previous->is(Keywords.kw_in))
1168 // Skip "const" as it does not have an influence on whether this is a name.
1169 FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1170 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1171 PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1173 if (!PreviousNotConst)
1176 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1177 PreviousNotConst->Previous &&
1178 PreviousNotConst->Previous->is(tok::hash);
1180 if (PreviousNotConst->is(TT_TemplateCloser))
1181 return PreviousNotConst && PreviousNotConst->MatchingParen &&
1182 PreviousNotConst->MatchingParen->Previous &&
1183 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1184 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1186 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen &&
1187 PreviousNotConst->MatchingParen->Previous &&
1188 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype))
1191 return (!IsPPKeyword &&
1192 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1193 PreviousNotConst->is(TT_PointerOrReference) ||
1194 PreviousNotConst->isSimpleTypeSpecifier();
1197 /// \brief Determine whether ')' is ending a cast.
1198 bool rParenEndsCast(const FormatToken &Tok) {
1199 // C-style casts are only used in C++ and Java.
1200 if (!Style.IsCpp() && Style.Language != FormatStyle::LK_Java)
1203 // Empty parens aren't casts and there are no casts at the end of the line.
1204 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1207 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1209 // If there is a closing parenthesis left of the current parentheses,
1210 // look past it as these might be chained casts.
1211 if (LeftOfParens->is(tok::r_paren)) {
1212 if (!LeftOfParens->MatchingParen ||
1213 !LeftOfParens->MatchingParen->Previous)
1215 LeftOfParens = LeftOfParens->MatchingParen->Previous;
1218 // If there is an identifier (or with a few exceptions a keyword) right
1219 // before the parentheses, this is unlikely to be a cast.
1220 if (LeftOfParens->Tok.getIdentifierInfo() &&
1221 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1225 // Certain other tokens right before the parentheses are also signals that
1226 // this cannot be a cast.
1227 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1228 TT_TemplateCloser, tok::ellipsis))
1232 if (Tok.Next->is(tok::question))
1235 // As Java has no function types, a "(" after the ")" likely means that this
1237 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1240 // If a (non-string) literal follows, this is likely a cast.
1241 if (Tok.Next->isNot(tok::string_literal) &&
1242 (Tok.Next->Tok.isLiteral() ||
1243 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1246 // Heuristically try to determine whether the parentheses contain a type.
1247 bool ParensAreType =
1249 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) ||
1250 Tok.Previous->isSimpleTypeSpecifier();
1251 bool ParensCouldEndDecl =
1252 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1253 if (ParensAreType && !ParensCouldEndDecl)
1256 // At this point, we heuristically assume that there are no casts at the
1257 // start of the line. We assume that we have found most cases where there
1258 // are by the logic above, e.g. "(void)x;".
1262 // Certain token types inside the parentheses mean that this can't be a
1264 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1265 Token = Token->Next)
1266 if (Token->is(TT_BinaryOperator))
1269 // If the following token is an identifier or 'this', this is a cast. All
1270 // cases where this can be something else are handled above.
1271 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1274 if (!Tok.Next->Next)
1277 // If the next token after the parenthesis is a unary operator, assume
1278 // that this is cast, unless there are unexpected tokens inside the
1281 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1282 if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1283 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1285 // Search for unexpected tokens.
1286 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1287 Prev = Prev->Previous) {
1288 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1294 /// \brief Return the type of the given token assuming it is * or &.
1295 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1296 bool InTemplateArgument) {
1297 if (Style.Language == FormatStyle::LK_JavaScript)
1298 return TT_BinaryOperator;
1300 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1302 return TT_UnaryOperator;
1304 const FormatToken *NextToken = Tok.getNextNonComment();
1306 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) ||
1307 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1308 return TT_PointerOrReference;
1310 if (PrevToken->is(tok::coloncolon))
1311 return TT_PointerOrReference;
1313 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1314 tok::comma, tok::semi, tok::kw_return, tok::colon,
1315 tok::equal, tok::kw_delete, tok::kw_sizeof) ||
1316 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1317 TT_UnaryOperator, TT_CastRParen))
1318 return TT_UnaryOperator;
1320 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1321 return TT_PointerOrReference;
1322 if (NextToken->is(tok::kw_operator) && !IsExpression)
1323 return TT_PointerOrReference;
1324 if (NextToken->isOneOf(tok::comma, tok::semi))
1325 return TT_PointerOrReference;
1327 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen &&
1328 PrevToken->MatchingParen->Previous &&
1329 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof,
1331 return TT_PointerOrReference;
1333 if (PrevToken->Tok.isLiteral() ||
1334 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1335 tok::kw_false, tok::r_brace) ||
1336 NextToken->Tok.isLiteral() ||
1337 NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1338 NextToken->isUnaryOperator() ||
1339 // If we know we're in a template argument, there are no named
1340 // declarations. Thus, having an identifier on the right-hand side
1341 // indicates a binary operator.
1342 (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1343 return TT_BinaryOperator;
1345 // "&&(" is quite unlikely to be two successive unary "&".
1346 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren))
1347 return TT_BinaryOperator;
1349 // This catches some cases where evaluation order is used as control flow:
1351 const FormatToken *NextNextToken = NextToken->getNextNonComment();
1352 if (NextNextToken && NextNextToken->is(tok::arrow))
1353 return TT_BinaryOperator;
1355 // It is very unlikely that we are going to find a pointer or reference type
1356 // definition on the RHS of an assignment.
1357 if (IsExpression && !Contexts.back().CaretFound)
1358 return TT_BinaryOperator;
1360 return TT_PointerOrReference;
1363 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
1364 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1366 return TT_UnaryOperator;
1368 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator) &&
1369 !PrevToken->is(tok::exclaim))
1370 // There aren't any trailing unary operators except for TypeScript's
1371 // non-null operator (!). Thus, this must be squence of leading operators.
1372 return TT_UnaryOperator;
1374 // Use heuristics to recognize unary operators.
1375 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
1376 tok::question, tok::colon, tok::kw_return,
1377 tok::kw_case, tok::at, tok::l_brace))
1378 return TT_UnaryOperator;
1380 // There can't be two consecutive binary operators.
1381 if (PrevToken->is(TT_BinaryOperator))
1382 return TT_UnaryOperator;
1384 // Fall back to marking the token as binary operator.
1385 return TT_BinaryOperator;
1388 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements.
1389 TokenType determineIncrementUsage(const FormatToken &Tok) {
1390 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1391 if (!PrevToken || PrevToken->is(TT_CastRParen))
1392 return TT_UnaryOperator;
1393 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
1394 return TT_TrailingUnaryOperator;
1396 return TT_UnaryOperator;
1399 SmallVector<Context, 8> Contexts;
1401 const FormatStyle &Style;
1402 AnnotatedLine &Line;
1403 FormatToken *CurrentToken;
1405 const AdditionalKeywords &Keywords;
1407 // Set of "<" tokens that do not open a template parameter list. If parseAngle
1408 // determines that a specific token can't be a template opener, it will make
1409 // same decision irrespective of the decisions for tokens leading up to it.
1410 // Store this information to prevent this from causing exponential runtime.
1411 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
1414 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
1415 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
1417 /// \brief Parses binary expressions by inserting fake parenthesis based on
1418 /// operator precedence.
1419 class ExpressionParser {
1421 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
1422 AnnotatedLine &Line)
1423 : Style(Style), Keywords(Keywords), Current(Line.First) {}
1425 /// \brief Parse expressions with the given operatore precedence.
1426 void parse(int Precedence = 0) {
1427 // Skip 'return' and ObjC selector colons as they are not part of a binary
1429 while (Current && (Current->is(tok::kw_return) ||
1430 (Current->is(tok::colon) &&
1431 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
1434 if (!Current || Precedence > PrecedenceArrowAndPeriod)
1437 // Conditional expressions need to be parsed separately for proper nesting.
1438 if (Precedence == prec::Conditional) {
1439 parseConditionalExpr();
1443 // Parse unary operators, which all have a higher precedence than binary
1445 if (Precedence == PrecedenceUnaryOperator) {
1446 parseUnaryOperator();
1450 FormatToken *Start = Current;
1451 FormatToken *LatestOperator = nullptr;
1452 unsigned OperatorIndex = 0;
1455 // Consume operators with higher precedence.
1456 parse(Precedence + 1);
1458 int CurrentPrecedence = getCurrentPrecedence();
1460 if (Current && Current->is(TT_SelectorName) &&
1461 Precedence == CurrentPrecedence) {
1463 addFakeParenthesis(Start, prec::Level(Precedence));
1467 // At the end of the line or when an operator with higher precedence is
1468 // found, insert fake parenthesis and return.
1470 (Current->closesScope() &&
1471 (Current->MatchingParen || Current->is(TT_TemplateString))) ||
1472 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
1473 (CurrentPrecedence == prec::Conditional &&
1474 Precedence == prec::Assignment && Current->is(tok::colon))) {
1478 // Consume scopes: (), [], <> and {}
1479 if (Current->opensScope()) {
1480 // In fragment of a JavaScript template string can look like '}..${' and
1481 // thus close a scope and open a new one at the same time.
1482 while (Current && (!Current->closesScope() || Current->opensScope())) {
1489 if (CurrentPrecedence == Precedence) {
1491 LatestOperator->NextOperator = Current;
1492 LatestOperator = Current;
1493 Current->OperatorIndex = OperatorIndex;
1496 next(/*SkipPastLeadingComments=*/Precedence > 0);
1500 if (LatestOperator && (Current || Precedence > 0)) {
1501 // LatestOperator->LastOperator = true;
1502 if (Precedence == PrecedenceArrowAndPeriod) {
1503 // Call expressions don't have a binary operator precedence.
1504 addFakeParenthesis(Start, prec::Unknown);
1506 addFakeParenthesis(Start, prec::Level(Precedence));
1512 /// \brief Gets the precedence (+1) of the given token for binary operators
1513 /// and other tokens that we treat like binary operators.
1514 int getCurrentPrecedence() {
1516 const FormatToken *NextNonComment = Current->getNextNonComment();
1517 if (Current->is(TT_ConditionalExpr))
1518 return prec::Conditional;
1519 if (NextNonComment && NextNonComment->is(tok::colon) &&
1520 NextNonComment->is(TT_DictLiteral))
1521 return prec::Assignment;
1522 if (Current->is(TT_JsComputedPropertyName))
1523 return prec::Assignment;
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 (Current->is(tok::comment) && NextNonComment &&
1530 NextNonComment->is(TT_SelectorName)))
1532 if (Current->is(TT_RangeBasedForLoopColon))
1534 if ((Style.Language == FormatStyle::LK_Java ||
1535 Style.Language == FormatStyle::LK_JavaScript) &&
1536 Current->is(Keywords.kw_instanceof))
1537 return prec::Relational;
1538 if (Style.Language == FormatStyle::LK_JavaScript &&
1539 Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
1540 return prec::Relational;
1541 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
1542 return Current->getPrecedence();
1543 if (Current->isOneOf(tok::period, tok::arrow))
1544 return PrecedenceArrowAndPeriod;
1545 if ((Style.Language == FormatStyle::LK_Java ||
1546 Style.Language == FormatStyle::LK_JavaScript) &&
1547 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
1548 Keywords.kw_throws))
1554 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
1555 Start->FakeLParens.push_back(Precedence);
1556 if (Precedence > prec::Unknown)
1557 Start->StartsBinaryExpression = true;
1559 FormatToken *Previous = Current->Previous;
1560 while (Previous->is(tok::comment) && Previous->Previous)
1561 Previous = Previous->Previous;
1562 ++Previous->FakeRParens;
1563 if (Precedence > prec::Unknown)
1564 Previous->EndsBinaryExpression = true;
1568 /// \brief Parse unary operator expressions and surround them with fake
1569 /// parentheses if appropriate.
1570 void parseUnaryOperator() {
1571 if (!Current || Current->isNot(TT_UnaryOperator)) {
1572 parse(PrecedenceArrowAndPeriod);
1576 FormatToken *Start = Current;
1578 parseUnaryOperator();
1580 // The actual precedence doesn't matter.
1581 addFakeParenthesis(Start, prec::Unknown);
1584 void parseConditionalExpr() {
1585 while (Current && Current->isTrailingComment()) {
1588 FormatToken *Start = Current;
1589 parse(prec::LogicalOr);
1590 if (!Current || !Current->is(tok::question))
1593 parse(prec::Assignment);
1594 if (!Current || Current->isNot(TT_ConditionalExpr))
1597 parse(prec::Assignment);
1598 addFakeParenthesis(Start, prec::Conditional);
1601 void next(bool SkipPastLeadingComments = true) {
1603 Current = Current->Next;
1605 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
1606 Current->isTrailingComment())
1607 Current = Current->Next;
1610 const FormatStyle &Style;
1611 const AdditionalKeywords &Keywords;
1612 FormatToken *Current;
1615 } // end anonymous namespace
1617 void TokenAnnotator::setCommentLineLevels(
1618 SmallVectorImpl<AnnotatedLine *> &Lines) {
1619 const AnnotatedLine *NextNonCommentLine = nullptr;
1620 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
1623 bool CommentLine = (*I)->First;
1624 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
1625 if (!Tok->is(tok::comment)) {
1626 CommentLine = false;
1630 if (NextNonCommentLine && CommentLine)
1631 (*I)->Level = NextNonCommentLine->Level;
1633 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
1635 setCommentLineLevels((*I)->Children);
1639 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
1640 unsigned Result = 0;
1641 for (const auto* Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
1642 Result = std::max(Result, Tok->NestingLevel);
1646 void TokenAnnotator::annotate(AnnotatedLine &Line) {
1647 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1648 E = Line.Children.end();
1652 AnnotatingParser Parser(Style, Line, Keywords);
1653 Line.Type = Parser.parseLine();
1655 // With very deep nesting, ExpressionParser uses lots of stack and the
1656 // formatting algorithm is very slow. We're not going to do a good job here
1657 // anyway - it's probably generated code being formatted by mistake.
1658 // Just skip the whole line.
1659 if (maxNestingDepth(Line) > 50)
1660 Line.Type = LT_Invalid;
1662 if (Line.Type == LT_Invalid)
1665 ExpressionParser ExprParser(Style, Keywords, Line);
1668 if (Line.startsWith(TT_ObjCMethodSpecifier))
1669 Line.Type = LT_ObjCMethodDecl;
1670 else if (Line.startsWith(TT_ObjCDecl))
1671 Line.Type = LT_ObjCDecl;
1672 else if (Line.startsWith(TT_ObjCProperty))
1673 Line.Type = LT_ObjCProperty;
1675 Line.First->SpacesRequiredBefore = 1;
1676 Line.First->CanBreakBefore = Line.First->MustBreakBefore;
1679 // This function heuristically determines whether 'Current' starts the name of a
1680 // function declaration.
1681 static bool isFunctionDeclarationName(const FormatToken &Current,
1682 const AnnotatedLine &Line) {
1683 auto skipOperatorName = [](const FormatToken* Next) -> const FormatToken* {
1684 for (; Next; Next = Next->Next) {
1685 if (Next->is(TT_OverloadedOperatorLParen))
1687 if (Next->is(TT_OverloadedOperator))
1689 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
1690 // For 'new[]' and 'delete[]'.
1691 if (Next->Next && Next->Next->is(tok::l_square) &&
1692 Next->Next->Next && Next->Next->Next->is(tok::r_square))
1693 Next = Next->Next->Next;
1702 // Find parentheses of parameter list.
1703 const FormatToken *Next = Current.Next;
1704 if (Current.is(tok::kw_operator)) {
1705 if (Current.Previous && Current.Previous->is(tok::coloncolon))
1707 Next = skipOperatorName(Next);
1709 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
1711 for (; Next; Next = Next->Next) {
1712 if (Next->is(TT_TemplateOpener)) {
1713 Next = Next->MatchingParen;
1714 } else if (Next->is(tok::coloncolon)) {
1718 if (Next->is(tok::kw_operator)) {
1719 Next = skipOperatorName(Next->Next);
1722 if (!Next->is(tok::identifier))
1724 } else if (Next->is(tok::l_paren)) {
1732 // Check whether parameter list can belong to a function declaration.
1733 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
1735 // If the lines ends with "{", this is likely an function definition.
1736 if (Line.Last->is(tok::l_brace))
1738 if (Next->Next == Next->MatchingParen)
1739 return true; // Empty parentheses.
1740 // If there is an &/&& after the r_paren, this is likely a function.
1741 if (Next->MatchingParen->Next &&
1742 Next->MatchingParen->Next->is(TT_PointerOrReference))
1744 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
1746 if (Tok->is(tok::l_paren) && Tok->MatchingParen) {
1747 Tok = Tok->MatchingParen;
1750 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
1751 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
1753 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
1754 Tok->Tok.isLiteral())
1760 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
1761 assert(Line.MightBeFunctionDecl);
1763 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
1764 Style.AlwaysBreakAfterReturnType ==
1765 FormatStyle::RTBS_TopLevelDefinitions) &&
1769 switch (Style.AlwaysBreakAfterReturnType) {
1770 case FormatStyle::RTBS_None:
1772 case FormatStyle::RTBS_All:
1773 case FormatStyle::RTBS_TopLevel:
1775 case FormatStyle::RTBS_AllDefinitions:
1776 case FormatStyle::RTBS_TopLevelDefinitions:
1777 return Line.mightBeFunctionDefinition();
1783 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
1784 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1785 E = Line.Children.end();
1787 calculateFormattingInformation(**I);
1790 Line.First->TotalLength =
1791 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth;
1792 FormatToken *Current = Line.First->Next;
1793 bool InFunctionDecl = Line.MightBeFunctionDecl;
1795 if (isFunctionDeclarationName(*Current, Line))
1796 Current->Type = TT_FunctionDeclarationName;
1797 if (Current->is(TT_LineComment)) {
1798 if (Current->Previous->BlockKind == BK_BracedInit &&
1799 Current->Previous->opensScope())
1800 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1;
1802 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
1804 // If we find a trailing comment, iterate backwards to determine whether
1805 // it seems to relate to a specific parameter. If so, break before that
1806 // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
1807 // to the previous line in:
1811 if (!Current->HasUnescapedNewline) {
1812 for (FormatToken *Parameter = Current->Previous; Parameter;
1813 Parameter = Parameter->Previous) {
1814 if (Parameter->isOneOf(tok::comment, tok::r_brace))
1816 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
1817 if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
1818 Parameter->HasUnescapedNewline)
1819 Parameter->MustBreakBefore = true;
1824 } else if (Current->SpacesRequiredBefore == 0 &&
1825 spaceRequiredBefore(Line, *Current)) {
1826 Current->SpacesRequiredBefore = 1;
1829 Current->MustBreakBefore =
1830 Current->MustBreakBefore || mustBreakBefore(Line, *Current);
1832 if (!Current->MustBreakBefore && InFunctionDecl &&
1833 Current->is(TT_FunctionDeclarationName))
1834 Current->MustBreakBefore = mustBreakForReturnType(Line);
1836 Current->CanBreakBefore =
1837 Current->MustBreakBefore || canBreakBefore(Line, *Current);
1838 unsigned ChildSize = 0;
1839 if (Current->Previous->Children.size() == 1) {
1840 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
1841 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
1842 : LastOfChild.TotalLength + 1;
1844 const FormatToken *Prev = Current->Previous;
1845 if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
1846 (Prev->Children.size() == 1 &&
1847 Prev->Children[0]->First->MustBreakBefore) ||
1848 Current->IsMultiline)
1849 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
1851 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
1852 ChildSize + Current->SpacesRequiredBefore;
1854 if (Current->is(TT_CtorInitializerColon))
1855 InFunctionDecl = false;
1857 // FIXME: Only calculate this if CanBreakBefore is true once static
1858 // initializers etc. are sorted out.
1859 // FIXME: Move magic numbers to a better place.
1860 Current->SplitPenalty = 20 * Current->BindingStrength +
1861 splitPenalty(Line, *Current, InFunctionDecl);
1863 Current = Current->Next;
1866 calculateUnbreakableTailLengths(Line);
1867 unsigned IndentLevel = Line.Level;
1868 for (Current = Line.First; Current != nullptr; Current = Current->Next) {
1870 Current->Role->precomputeFormattingInfos(Current);
1871 if (Current->MatchingParen &&
1872 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
1873 assert(IndentLevel > 0);
1876 Current->IndentLevel = IndentLevel;
1877 if (Current->opensBlockOrBlockTypeList(Style))
1881 DEBUG({ printDebugInfo(Line); });
1884 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
1885 unsigned UnbreakableTailLength = 0;
1886 FormatToken *Current = Line.Last;
1888 Current->UnbreakableTailLength = UnbreakableTailLength;
1889 if (Current->CanBreakBefore ||
1890 Current->isOneOf(tok::comment, tok::string_literal)) {
1891 UnbreakableTailLength = 0;
1893 UnbreakableTailLength +=
1894 Current->ColumnWidth + Current->SpacesRequiredBefore;
1896 Current = Current->Previous;
1900 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
1901 const FormatToken &Tok,
1902 bool InFunctionDecl) {
1903 const FormatToken &Left = *Tok.Previous;
1904 const FormatToken &Right = Tok;
1906 if (Left.is(tok::semi))
1909 if (Style.Language == FormatStyle::LK_Java) {
1910 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
1912 if (Right.is(Keywords.kw_implements))
1914 if (Left.is(tok::comma) && Left.NestingLevel == 0)
1916 } else if (Style.Language == FormatStyle::LK_JavaScript) {
1917 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
1919 if (Left.is(TT_JsTypeColon))
1921 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
1922 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
1926 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
1928 if (Right.is(tok::l_square)) {
1929 if (Style.Language == FormatStyle::LK_Proto)
1931 if (Left.is(tok::r_square))
1933 // Slightly prefer formatting local lambda definitions like functions.
1934 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
1936 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
1937 TT_ArrayInitializerLSquare))
1941 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
1942 Right.is(tok::kw_operator)) {
1943 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
1945 if (Left.is(TT_StartOfName))
1947 if (InFunctionDecl && Right.NestingLevel == 0)
1948 return Style.PenaltyReturnTypeOnItsOwnLine;
1951 if (Right.is(TT_PointerOrReference))
1953 if (Right.is(TT_LambdaArrow))
1955 if (Left.is(tok::equal) && Right.is(tok::l_brace))
1957 if (Left.is(TT_CastRParen))
1959 if (Left.is(tok::coloncolon) ||
1960 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
1962 if (Left.isOneOf(tok::kw_class, tok::kw_struct))
1964 if (Left.is(tok::comment))
1967 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon))
1970 if (Right.isMemberAccess()) {
1971 // Breaking before the "./->" of a chained call/member access is reasonably
1972 // cheap, as formatting those with one call per line is generally
1973 // desirable. In particular, it should be cheaper to break before the call
1974 // than it is to break inside a call's parameters, which could lead to weird
1975 // "hanging" indents. The exception is the very last "./->" to support this
1976 // frequent pattern:
1978 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
1981 // which might otherwise be blown up onto many lines. Here, clang-format
1982 // won't produce "hanging" indents anyway as there is no other trailing
1985 // Also apply higher penalty is not a call as that might lead to a wrapping
1989 // .aaaaaaaaa.bbbbbbbb(cccccccc);
1990 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
1995 if (Right.is(TT_TrailingAnnotation) &&
1996 (!Right.Next || Right.Next->isNot(tok::l_paren))) {
1997 // Moving trailing annotations to the next line is fine for ObjC method
1999 if (Line.startsWith(TT_ObjCMethodSpecifier))
2001 // Generally, breaking before a trailing annotation is bad unless it is
2002 // function-like. It seems to be especially preferable to keep standard
2003 // annotations (i.e. "const", "final" and "override") on the same line.
2004 // Use a slightly higher penalty after ")" so that annotations like
2005 // "const override" are kept together.
2006 bool is_short_annotation = Right.TokenText.size() < 10;
2007 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2010 // In for-loops, prefer breaking at ',' and ';'.
2011 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2014 // In Objective-C method expressions, prefer breaking before "param:" over
2015 // breaking after it.
2016 if (Right.is(TT_SelectorName))
2018 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2019 return Line.MightBeFunctionDecl ? 50 : 500;
2021 if (Left.is(tok::l_paren) && InFunctionDecl &&
2022 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2024 if (Left.is(tok::l_paren) && Left.Previous &&
2025 Left.Previous->isOneOf(tok::kw_if, tok::kw_for))
2027 if (Left.is(tok::equal) && InFunctionDecl)
2029 if (Right.is(tok::r_brace))
2031 if (Left.is(TT_TemplateOpener))
2033 if (Left.opensScope()) {
2034 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
2036 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2039 if (Left.is(TT_JavaAnnotation))
2042 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2043 Left.Previous->isLabelString() &&
2044 (Left.NextOperator || Left.OperatorIndex != 0))
2046 if (Right.is(tok::plus) && Left.isLabelString() &&
2047 (Right.NextOperator || Right.OperatorIndex != 0))
2049 if (Left.is(tok::comma))
2051 if (Right.is(tok::lessless) && Left.isLabelString() &&
2052 (Right.NextOperator || Right.OperatorIndex != 1))
2054 if (Right.is(tok::lessless)) {
2055 // Breaking at a << is really cheap.
2056 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2057 // Slightly prefer to break before the first one in log-like statements.
2061 if (Left.is(TT_ConditionalExpr))
2062 return prec::Conditional;
2063 prec::Level Level = Left.getPrecedence();
2064 if (Level != prec::Unknown)
2066 Level = Right.getPrecedence();
2067 if (Level != prec::Unknown)
2073 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2074 const FormatToken &Left,
2075 const FormatToken &Right) {
2076 if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2078 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2079 Left.Tok.getObjCKeywordID() == tok::objc_property)
2081 if (Right.is(tok::hashhash))
2082 return Left.is(tok::hash);
2083 if (Left.isOneOf(tok::hashhash, tok::hash))
2084 return Right.is(tok::hash);
2085 if (Left.is(tok::l_paren) && Right.is(tok::r_paren))
2086 return Style.SpaceInEmptyParentheses;
2087 if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2088 return (Right.is(TT_CastRParen) ||
2089 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2090 ? Style.SpacesInCStyleCastParentheses
2091 : Style.SpacesInParentheses;
2092 if (Right.isOneOf(tok::semi, tok::comma))
2094 if (Right.is(tok::less) &&
2095 Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList)
2097 if (Right.is(tok::less) && Left.is(tok::kw_template))
2098 return Style.SpaceAfterTemplateKeyword;
2099 if (Left.isOneOf(tok::exclaim, tok::tilde))
2101 if (Left.is(tok::at) &&
2102 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2103 tok::numeric_constant, tok::l_paren, tok::l_brace,
2104 tok::kw_true, tok::kw_false))
2106 if (Left.is(tok::colon))
2107 return !Left.is(TT_ObjCMethodExpr);
2108 if (Left.is(tok::coloncolon))
2110 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less))
2112 if (Right.is(tok::ellipsis))
2113 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2114 Left.Previous->is(tok::kw_case));
2115 if (Left.is(tok::l_square) && Right.is(tok::amp))
2117 if (Right.is(TT_PointerOrReference))
2118 return (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) ||
2119 (Left.Tok.isLiteral() || (Left.is(tok::kw_const) && Left.Previous &&
2120 Left.Previous->is(tok::r_paren)) ||
2121 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2122 (Style.PointerAlignment != FormatStyle::PAS_Left ||
2123 (Line.IsMultiVariableDeclStmt &&
2124 (Left.NestingLevel == 0 ||
2125 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2126 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2127 (!Left.is(TT_PointerOrReference) ||
2128 (Style.PointerAlignment != FormatStyle::PAS_Right &&
2129 !Line.IsMultiVariableDeclStmt)))
2131 if (Left.is(TT_PointerOrReference))
2132 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2133 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2134 !Right.is(TT_StartOfName)) ||
2135 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) ||
2136 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2138 (Style.PointerAlignment != FormatStyle::PAS_Right &&
2139 !Line.IsMultiVariableDeclStmt) &&
2141 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon));
2142 if (Right.is(tok::star) && Left.is(tok::l_paren))
2144 if (Left.is(tok::l_square))
2145 return (Left.is(TT_ArrayInitializerLSquare) &&
2146 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) ||
2147 (Left.is(TT_ArraySubscriptLSquare) && Style.SpacesInSquareBrackets &&
2148 Right.isNot(tok::r_square));
2149 if (Right.is(tok::r_square))
2150 return Right.MatchingParen &&
2151 ((Style.SpacesInContainerLiterals &&
2152 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) ||
2153 (Style.SpacesInSquareBrackets &&
2154 Right.MatchingParen->is(TT_ArraySubscriptLSquare)));
2155 if (Right.is(tok::l_square) &&
2156 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare) &&
2157 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral))
2159 if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2160 return !Left.Children.empty(); // No spaces in "{}".
2161 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) ||
2162 (Right.is(tok::r_brace) && Right.MatchingParen &&
2163 Right.MatchingParen->BlockKind != BK_Block))
2164 return !Style.Cpp11BracedListStyle;
2165 if (Left.is(TT_BlockComment))
2166 return !Left.TokenText.endswith("=*/");
2167 if (Right.is(tok::l_paren)) {
2168 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen))
2170 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
2171 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
2172 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while,
2173 tok::kw_switch, tok::kw_case, TT_ForEachMacro,
2175 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
2176 tok::kw_new, tok::kw_delete) &&
2177 (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
2178 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always &&
2179 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
2180 Left.is(tok::r_paren)) &&
2181 Line.Type != LT_PreprocessorDirective);
2183 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
2185 if (Right.is(TT_UnaryOperator))
2186 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
2187 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
2188 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
2190 Left.isSimpleTypeSpecifier()) &&
2191 Right.is(tok::l_brace) && Right.getNextNonComment() &&
2192 Right.BlockKind != BK_Block)
2194 if (Left.is(tok::period) || Right.is(tok::period))
2196 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
2198 if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
2199 Left.MatchingParen->Previous &&
2200 Left.MatchingParen->Previous->is(tok::period))
2201 // A.<B>DoSomething();
2203 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
2208 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
2209 const FormatToken &Right) {
2210 const FormatToken &Left = *Right.Previous;
2211 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
2212 return true; // Never ever merge two identifiers.
2213 if (Style.IsCpp()) {
2214 if (Left.is(tok::kw_operator))
2215 return Right.is(tok::coloncolon);
2216 } else if (Style.Language == FormatStyle::LK_Proto) {
2217 if (Right.is(tok::period) &&
2218 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
2219 Keywords.kw_repeated, Keywords.kw_extend))
2221 if (Right.is(tok::l_paren) &&
2222 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
2224 } else if (Style.Language == FormatStyle::LK_JavaScript) {
2225 if (Left.is(TT_JsFatArrow))
2227 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
2228 Right.MatchingParen) {
2229 const FormatToken *Next = Right.MatchingParen->getNextNonComment();
2230 // An async arrow function, for example: `x = async () => foo();`,
2231 // as opposed to calling a function called async: `x = async();`
2232 if (Next && Next->is(TT_JsFatArrow))
2235 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2236 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2238 if (Left.is(tok::identifier) && Right.is(TT_TemplateString))
2240 if (Right.is(tok::star) &&
2241 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2243 if (Right.isOneOf(tok::l_brace, tok::l_square) &&
2244 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2246 // JS methods can use some keywords as names (e.g. `delete()`).
2247 if (Right.is(tok::l_paren) && Line.MustBeDeclaration &&
2248 Left.Tok.getIdentifierInfo())
2250 if (Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
2251 Keywords.kw_of, tok::kw_const) &&
2252 (!Left.Previous || !Left.Previous->is(tok::period)))
2254 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
2255 Left.Previous->is(tok::period) && Right.is(tok::l_paren))
2257 if (Left.is(Keywords.kw_as) &&
2258 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
2260 if (Left.is(tok::kw_default) && Left.Previous &&
2261 Left.Previous->is(tok::kw_export))
2263 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
2265 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
2267 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
2269 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
2270 Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
2272 if (Left.is(tok::ellipsis))
2274 if (Left.is(TT_TemplateCloser) &&
2275 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
2276 Keywords.kw_implements, Keywords.kw_extends))
2277 // Type assertions ('<type>expr') are not followed by whitespace. Other
2278 // locations that should have whitespace following are identified by the
2279 // above set of follower tokens.
2281 // Postfix non-null assertion operator, as in `foo!.bar()`.
2282 if (Right.is(tok::exclaim) && (Left.isOneOf(tok::identifier, tok::r_paren,
2283 tok::r_square, tok::r_brace) ||
2284 Left.Tok.isLiteral()))
2286 if (Left.is(tok::exclaim) && Right.is(Keywords.kw_as))
2287 return true; // "x! as string"
2288 } else if (Style.Language == FormatStyle::LK_Java) {
2289 if (Left.is(tok::r_square) && Right.is(tok::l_brace))
2291 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
2292 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
2293 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
2294 tok::kw_protected) ||
2295 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
2296 Keywords.kw_native)) &&
2297 Right.is(TT_TemplateOpener))
2300 if (Left.is(TT_ImplicitStringLiteral))
2301 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2302 if (Line.Type == LT_ObjCMethodDecl) {
2303 if (Left.is(TT_ObjCMethodSpecifier))
2305 if (Left.is(tok::r_paren) && Right.is(tok::identifier))
2306 // Don't space between ')' and <id>
2309 if (Line.Type == LT_ObjCProperty &&
2310 (Right.is(tok::equal) || Left.is(tok::equal)))
2313 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
2314 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
2316 if (Right.is(TT_OverloadedOperatorLParen))
2317 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2318 if (Left.is(tok::comma))
2320 if (Right.is(tok::comma))
2322 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen))
2324 if (Right.is(tok::colon)) {
2325 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
2326 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
2328 if (Right.is(TT_ObjCMethodExpr))
2330 if (Left.is(tok::question))
2332 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
2334 if (Right.is(TT_DictLiteral))
2335 return Style.SpacesInContainerLiterals;
2338 if (Left.is(TT_UnaryOperator))
2339 return Right.is(TT_BinaryOperator);
2341 // If the next token is a binary operator or a selector name, we have
2342 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
2343 if (Left.is(TT_CastRParen))
2344 return Style.SpaceAfterCStyleCast ||
2345 Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
2347 if (Left.is(tok::greater) && Right.is(tok::greater))
2348 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
2349 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
2350 if (Right.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
2351 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar))
2353 if (!Style.SpaceBeforeAssignmentOperators &&
2354 Right.getPrecedence() == prec::Assignment)
2356 if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
2357 // Generally don't remove existing spaces between an identifier and "::".
2358 // The identifier might actually be a macro name such as ALWAYS_INLINE. If
2359 // this turns out to be too lenient, add analysis of the identifier itself.
2360 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2361 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment))
2362 return (Left.is(TT_TemplateOpener) &&
2363 Style.Standard == FormatStyle::LS_Cpp03) ||
2364 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
2365 TT_TemplateCloser, TT_TemplateOpener));
2366 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
2367 return Style.SpacesInAngles;
2368 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
2369 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
2370 !Right.is(tok::r_paren)))
2372 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
2373 Right.isNot(TT_FunctionTypeLParen))
2374 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2375 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
2376 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
2378 if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
2379 Line.startsWith(tok::hash))
2381 if (Right.is(TT_TrailingUnaryOperator))
2383 if (Left.is(TT_RegexLiteral))
2385 return spaceRequiredBetween(Line, Left, Right);
2388 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
2389 static bool isAllmanBrace(const FormatToken &Tok) {
2390 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block &&
2391 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral);
2394 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
2395 const FormatToken &Right) {
2396 const FormatToken &Left = *Right.Previous;
2397 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
2400 if (Style.Language == FormatStyle::LK_JavaScript) {
2401 // FIXME: This might apply to other languages and token kinds.
2402 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
2403 Left.Previous->is(tok::string_literal))
2405 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
2406 Left.Previous && Left.Previous->is(tok::equal) &&
2407 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
2409 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
2411 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
2412 // Object literals on the top level of a file are treated as "enum-style".
2413 // Each key/value pair is put on a separate line, instead of bin-packing.
2415 if (Left.is(tok::l_brace) && Line.Level == 0 &&
2416 (Line.startsWith(tok::kw_enum) ||
2417 Line.startsWith(tok::kw_export, tok::kw_enum)))
2418 // JavaScript top-level enum key/value pairs are put on separate lines
2419 // instead of bin-packing.
2421 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
2422 !Left.Children.empty())
2423 // Support AllowShortFunctionsOnASingleLine for JavaScript.
2424 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
2425 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
2426 (Left.NestingLevel == 0 && Line.Level == 0 &&
2427 Style.AllowShortFunctionsOnASingleLine ==
2428 FormatStyle::SFS_Inline);
2429 } else if (Style.Language == FormatStyle::LK_Java) {
2430 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
2431 Right.Next->is(tok::string_literal))
2433 } else if (Style.Language == FormatStyle::LK_Cpp ||
2434 Style.Language == FormatStyle::LK_ObjC ||
2435 Style.Language == FormatStyle::LK_Proto) {
2436 if (Left.isStringLiteral() &&
2437 (Right.isStringLiteral() || Right.is(TT_ObjCStringLiteral)))
2441 // If the last token before a '}' is a comma or a trailing comment, the
2442 // intention is to insert a line break after it in order to make shuffling
2443 // around entries easier.
2444 const FormatToken *BeforeClosingBrace = nullptr;
2445 if (Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) &&
2446 Left.BlockKind != BK_Block && Left.MatchingParen)
2447 BeforeClosingBrace = Left.MatchingParen->Previous;
2448 else if (Right.MatchingParen &&
2449 Right.MatchingParen->isOneOf(tok::l_brace,
2450 TT_ArrayInitializerLSquare))
2451 BeforeClosingBrace = &Left;
2452 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
2453 BeforeClosingBrace->isTrailingComment()))
2456 if (Right.is(tok::comment))
2457 return Left.BlockKind != BK_BracedInit &&
2458 Left.isNot(TT_CtorInitializerColon) &&
2459 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
2460 if (Left.isTrailingComment())
2462 if (Right.Previous->IsUnterminatedLiteral)
2464 if (Right.is(tok::lessless) && Right.Next &&
2465 Right.Previous->is(tok::string_literal) &&
2466 Right.Next->is(tok::string_literal))
2468 if (Right.Previous->ClosesTemplateDeclaration &&
2469 Right.Previous->MatchingParen &&
2470 Right.Previous->MatchingParen->NestingLevel == 0 &&
2471 Style.AlwaysBreakTemplateDeclarations)
2473 if ((Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) &&
2474 Style.BreakConstructorInitializersBeforeComma &&
2475 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
2477 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
2478 // Raw string literals are special wrt. line breaks. The author has made a
2479 // deliberate choice and might have aligned the contents of the string
2480 // literal accordingly. Thus, we try keep existing line breaks.
2481 return Right.NewlinesBefore > 0;
2482 if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 &&
2483 Style.Language == FormatStyle::LK_Proto)
2484 // Don't put enums onto single lines in protocol buffers.
2486 if (Right.is(TT_InlineASMBrace))
2487 return Right.HasUnescapedNewline;
2488 if (isAllmanBrace(Left) || isAllmanBrace(Right))
2489 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
2490 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
2491 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
2492 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine)
2495 if ((Style.Language == FormatStyle::LK_Java ||
2496 Style.Language == FormatStyle::LK_JavaScript) &&
2497 Left.is(TT_LeadingJavaAnnotation) &&
2498 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
2499 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
2505 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
2506 const FormatToken &Right) {
2507 const FormatToken &Left = *Right.Previous;
2509 // Language-specific stuff.
2510 if (Style.Language == FormatStyle::LK_Java) {
2511 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
2512 Keywords.kw_implements))
2514 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
2515 Keywords.kw_implements))
2517 } else if (Style.Language == FormatStyle::LK_JavaScript) {
2518 const FormatToken *NonComment = Right.getPreviousNonComment();
2519 if (Left.isOneOf(tok::kw_return, tok::kw_continue, tok::kw_break,
2522 NonComment->isOneOf(tok::kw_return, tok::kw_continue, tok::kw_break,
2524 return false; // Otherwise a semicolon is inserted.
2525 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
2527 if (Left.is(TT_JsTypeColon))
2529 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is))
2531 if (Left.is(Keywords.kw_in))
2532 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
2533 if (Right.is(Keywords.kw_in))
2534 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
2535 if (Right.is(Keywords.kw_as))
2536 return false; // must not break before as in 'x as type' casts
2537 if (Left.is(Keywords.kw_as))
2539 if (Left.is(Keywords.kw_declare) &&
2540 Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
2541 Keywords.kw_function, tok::kw_class, tok::kw_enum,
2542 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
2543 Keywords.kw_let, tok::kw_const))
2544 // See grammar for 'declare' statements at:
2545 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
2547 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
2548 Right.isOneOf(tok::identifier, tok::string_literal))
2549 return false; // must not break in "module foo { ...}"
2550 if (Right.is(TT_TemplateString) && Right.closesScope())
2552 if (Left.is(TT_TemplateString) && Left.opensScope())
2556 if (Left.is(tok::at))
2558 if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
2560 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
2561 return !Right.is(tok::l_paren);
2562 if (Right.is(TT_PointerOrReference))
2563 return Line.IsMultiVariableDeclStmt ||
2564 (Style.PointerAlignment == FormatStyle::PAS_Right &&
2565 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
2566 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2567 Right.is(tok::kw_operator))
2569 if (Left.is(TT_PointerOrReference))
2571 if (Right.isTrailingComment())
2572 // We rely on MustBreakBefore being set correctly here as we should not
2573 // change the "binding" behavior of a comment.
2574 // The first comment in a braced lists is always interpreted as belonging to
2575 // the first list element. Otherwise, it should be placed outside of the
2577 return Left.BlockKind == BK_BracedInit;
2578 if (Left.is(tok::question) && Right.is(tok::colon))
2580 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
2581 return Style.BreakBeforeTernaryOperators;
2582 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
2583 return !Style.BreakBeforeTernaryOperators;
2584 if (Right.is(TT_InheritanceColon))
2586 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
2587 Left.isNot(TT_SelectorName))
2589 if (Right.is(tok::colon) &&
2590 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
2592 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr))
2594 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
2595 Right.Next->is(TT_ObjCMethodExpr)))
2596 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
2597 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
2599 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
2601 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
2602 TT_OverloadedOperator))
2604 if (Left.is(TT_RangeBasedForLoopColon))
2606 if (Right.is(TT_RangeBasedForLoopColon))
2608 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
2610 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
2611 Left.is(tok::kw_operator))
2613 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
2614 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
2616 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen))
2618 if (Left.is(tok::l_paren) && Left.Previous &&
2619 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
2621 if (Right.is(TT_ImplicitStringLiteral))
2624 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
2626 if (Right.is(tok::r_square) && Right.MatchingParen &&
2627 Right.MatchingParen->is(TT_LambdaLSquare))
2630 // We only break before r_brace if there was a corresponding break before
2631 // the l_brace, which is tracked by BreakBeforeClosingBrace.
2632 if (Right.is(tok::r_brace))
2633 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block;
2635 // Allow breaking after a trailing annotation, e.g. after a method
2637 if (Left.is(TT_TrailingAnnotation))
2638 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
2639 tok::less, tok::coloncolon);
2641 if (Right.is(tok::kw___attribute))
2644 if (Left.is(tok::identifier) && Right.is(tok::string_literal))
2647 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2650 if (Left.is(TT_CtorInitializerComma) &&
2651 Style.BreakConstructorInitializersBeforeComma)
2653 if (Right.is(TT_CtorInitializerComma) &&
2654 Style.BreakConstructorInitializersBeforeComma)
2656 if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
2657 (Left.is(tok::less) && Right.is(tok::less)))
2659 if (Right.is(TT_BinaryOperator) &&
2660 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
2661 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
2662 Right.getPrecedence() != prec::Assignment))
2664 if (Left.is(TT_ArrayInitializerLSquare))
2666 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
2668 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
2669 !Left.isOneOf(tok::arrowstar, tok::lessless) &&
2670 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
2671 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
2672 Left.getPrecedence() == prec::Assignment))
2674 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
2675 tok::kw_class, tok::kw_struct, tok::comment) ||
2676 Right.isMemberAccess() ||
2677 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
2678 tok::colon, tok::l_square, tok::at) ||
2679 (Left.is(tok::r_paren) &&
2680 Right.isOneOf(tok::identifier, tok::kw_const)) ||
2681 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
2682 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
2685 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
2686 llvm::errs() << "AnnotatedTokens:\n";
2687 const FormatToken *Tok = Line.First;
2689 llvm::errs() << " M=" << Tok->MustBreakBefore
2690 << " C=" << Tok->CanBreakBefore
2691 << " T=" << getTokenTypeName(Tok->Type)
2692 << " S=" << Tok->SpacesRequiredBefore
2693 << " B=" << Tok->BlockParameterCount
2694 << " BK=" << Tok->BlockKind
2695 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName()
2696 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind
2698 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
2699 llvm::errs() << Tok->FakeLParens[i] << "/";
2700 llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
2701 llvm::errs() << " Text='" << Tok->TokenText << "'\n";
2703 assert(Tok == Line.Last);
2706 llvm::errs() << "----\n";
2709 } // namespace format
2710 } // namespace clang