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().InInheritanceList)
680 Tok->Type = TT_InheritanceComma;
681 else if (Contexts.back().FirstStartOfName &&
682 (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
683 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
684 Line.IsMultiVariableDeclStmt = true;
686 if (Contexts.back().IsForEachMacro)
687 Contexts.back().IsExpression = true;
689 case tok::identifier:
690 if (Tok->isOneOf(Keywords.kw___has_include,
691 Keywords.kw___has_include_next)) {
701 void parseIncludeDirective() {
702 if (CurrentToken && CurrentToken->is(tok::less)) {
704 while (CurrentToken) {
705 if (CurrentToken->isNot(tok::comment) || CurrentToken->Next)
706 CurrentToken->Type = TT_ImplicitStringLiteral;
712 void parseWarningOrError() {
714 // We still want to format the whitespace left of the first token of the
717 while (CurrentToken) {
718 CurrentToken->Type = TT_ImplicitStringLiteral;
724 next(); // Consume "pragma".
726 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
727 bool IsMark = CurrentToken->is(Keywords.kw_mark);
728 next(); // Consume "mark".
729 next(); // Consume first token (so we fix leading whitespace).
730 while (CurrentToken) {
731 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
732 CurrentToken->Type = TT_ImplicitStringLiteral;
738 void parseHasInclude() {
739 if (!CurrentToken || !CurrentToken->is(tok::l_paren))
742 parseIncludeDirective();
746 LineType parsePreprocessorDirective() {
747 bool IsFirstToken = CurrentToken->IsFirst;
748 LineType Type = LT_PreprocessorDirective;
753 if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
754 // JavaScript files can contain shebang lines of the form:
755 // #!/usr/bin/env node
756 // Treat these like C++ #include directives.
757 while (CurrentToken) {
758 // Tokens cannot be comments here.
759 CurrentToken->Type = TT_ImplicitStringLiteral;
762 return LT_ImportStatement;
765 if (CurrentToken->Tok.is(tok::numeric_constant)) {
766 CurrentToken->SpacesRequiredBefore = 1;
769 // Hashes in the middle of a line can lead to any strange token
771 if (!CurrentToken->Tok.getIdentifierInfo())
773 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
774 case tok::pp_include:
775 case tok::pp_include_next:
778 parseIncludeDirective();
779 Type = LT_ImportStatement;
782 case tok::pp_warning:
783 parseWarningOrError();
790 Contexts.back().IsExpression = true;
796 while (CurrentToken) {
797 FormatToken *Tok = CurrentToken;
799 if (Tok->is(tok::l_paren))
801 else if (Tok->isOneOf(Keywords.kw___has_include,
802 Keywords.kw___has_include_next))
809 LineType parseLine() {
810 NonTemplateLess.clear();
811 if (CurrentToken->is(tok::hash))
812 return parsePreprocessorDirective();
814 // Directly allow to 'import <string-literal>' to support protocol buffer
815 // definitions (code.google.com/p/protobuf) or missing "#" (either way we
816 // should not break the line).
817 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
818 if ((Style.Language == FormatStyle::LK_Java &&
819 CurrentToken->is(Keywords.kw_package)) ||
820 (Info && Info->getPPKeywordID() == tok::pp_import &&
821 CurrentToken->Next &&
822 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
825 parseIncludeDirective();
826 return LT_ImportStatement;
829 // If this line starts and ends in '<' and '>', respectively, it is likely
830 // part of "#define <a/b.h>".
831 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
832 parseIncludeDirective();
833 return LT_ImportStatement;
836 // In .proto files, top-level options are very similar to import statements
837 // and should not be line-wrapped.
838 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
839 CurrentToken->is(Keywords.kw_option)) {
841 if (CurrentToken && CurrentToken->is(tok::identifier))
842 return LT_ImportStatement;
845 bool KeywordVirtualFound = false;
846 bool ImportStatement = false;
848 // import {...} from '...';
849 if (Style.Language == FormatStyle::LK_JavaScript &&
850 CurrentToken->is(Keywords.kw_import))
851 ImportStatement = true;
853 while (CurrentToken) {
854 if (CurrentToken->is(tok::kw_virtual))
855 KeywordVirtualFound = true;
856 if (Style.Language == FormatStyle::LK_JavaScript) {
857 // export {...} from '...';
858 // An export followed by "from 'some string';" is a re-export from
859 // another module identified by a URI and is treated as a
860 // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
861 // Just "export {...};" or "export class ..." should not be treated as
862 // an import in this sense.
863 if (Line.First->is(tok::kw_export) &&
864 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
865 CurrentToken->Next->isStringLiteral())
866 ImportStatement = true;
867 if (isClosureImportStatement(*CurrentToken))
868 ImportStatement = true;
873 if (KeywordVirtualFound)
874 return LT_VirtualFunctionDecl;
876 return LT_ImportStatement;
878 if (Line.startsWith(TT_ObjCMethodSpecifier)) {
879 if (Contexts.back().FirstObjCSelectorName)
880 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
881 Contexts.back().LongestObjCSelectorName;
882 return LT_ObjCMethodDecl;
889 bool isClosureImportStatement(const FormatToken &Tok) {
890 // FIXME: Closure-library specific stuff should not be hard-coded but be
892 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
893 Tok.Next->Next && (Tok.Next->Next->TokenText == "module" ||
894 Tok.Next->Next->TokenText == "provide" ||
895 Tok.Next->Next->TokenText == "require" ||
896 Tok.Next->Next->TokenText == "setTestOnly" ||
897 Tok.Next->Next->TokenText == "forwardDeclare") &&
898 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
901 void resetTokenMetadata(FormatToken *Token) {
905 // Reset token type in case we have already looked at it and then
906 // recovered from an error (e.g. failure to find the matching >).
907 if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro,
908 TT_FunctionLBrace, TT_ImplicitStringLiteral,
909 TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow,
910 TT_OverloadedOperator, TT_RegexLiteral,
911 TT_TemplateString, TT_ObjCStringLiteral))
912 CurrentToken->Type = TT_Unknown;
913 CurrentToken->Role.reset();
914 CurrentToken->MatchingParen = nullptr;
915 CurrentToken->FakeLParens.clear();
916 CurrentToken->FakeRParens = 0;
921 CurrentToken->NestingLevel = Contexts.size() - 1;
922 CurrentToken->BindingStrength = Contexts.back().BindingStrength;
923 modifyContext(*CurrentToken);
924 determineTokenType(*CurrentToken);
925 CurrentToken = CurrentToken->Next;
928 resetTokenMetadata(CurrentToken);
931 /// \brief A struct to hold information valid in a specific context, e.g.
932 /// a pair of parenthesis.
934 Context(tok::TokenKind ContextKind, unsigned BindingStrength,
936 : ContextKind(ContextKind), BindingStrength(BindingStrength),
937 IsExpression(IsExpression) {}
939 tok::TokenKind ContextKind;
940 unsigned BindingStrength;
942 unsigned LongestObjCSelectorName = 0;
943 bool ColonIsForRangeExpr = false;
944 bool ColonIsDictLiteral = false;
945 bool ColonIsObjCMethodExpr = false;
946 FormatToken *FirstObjCSelectorName = nullptr;
947 FormatToken *FirstStartOfName = nullptr;
948 bool CanBeExpression = true;
949 bool InTemplateArgument = false;
950 bool InCtorInitializer = false;
951 bool InInheritanceList = false;
952 bool CaretFound = false;
953 bool IsForEachMacro = false;
956 /// \brief Puts a new \c Context onto the stack \c Contexts for the lifetime
957 /// of each instance.
958 struct ScopedContextCreator {
961 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
964 P.Contexts.push_back(Context(ContextKind,
965 P.Contexts.back().BindingStrength + Increase,
966 P.Contexts.back().IsExpression));
969 ~ScopedContextCreator() { P.Contexts.pop_back(); }
972 void modifyContext(const FormatToken &Current) {
973 if (Current.getPrecedence() == prec::Assignment &&
974 !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
975 // Type aliases use `type X = ...;` in TypeScript.
976 !(Style.Language == FormatStyle::LK_JavaScript &&
977 Line.startsWith(Keywords.kw_type, tok::identifier)) &&
978 (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
979 Contexts.back().IsExpression = true;
980 if (!Line.startsWith(TT_UnaryOperator)) {
981 for (FormatToken *Previous = Current.Previous;
982 Previous && Previous->Previous &&
983 !Previous->Previous->isOneOf(tok::comma, tok::semi);
984 Previous = Previous->Previous) {
985 if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
986 Previous = Previous->MatchingParen;
990 if (Previous->opensScope())
992 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
993 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
994 Previous->Previous && Previous->Previous->isNot(tok::equal))
995 Previous->Type = TT_PointerOrReference;
998 } else if (Current.is(tok::lessless) &&
999 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1000 Contexts.back().IsExpression = true;
1001 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1002 Contexts.back().IsExpression = true;
1003 } else if (Current.is(TT_TrailingReturnArrow)) {
1004 Contexts.back().IsExpression = false;
1005 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1006 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1007 } else if (Current.Previous &&
1008 Current.Previous->is(TT_CtorInitializerColon)) {
1009 Contexts.back().IsExpression = true;
1010 Contexts.back().InCtorInitializer = true;
1011 } else if (Current.Previous &&
1012 Current.Previous->is(TT_InheritanceColon)) {
1013 Contexts.back().InInheritanceList = true;
1014 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1015 for (FormatToken *Previous = Current.Previous;
1016 Previous && Previous->isOneOf(tok::star, tok::amp);
1017 Previous = Previous->Previous)
1018 Previous->Type = TT_PointerOrReference;
1019 if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1020 Contexts.back().IsExpression = false;
1021 } else if (Current.is(tok::kw_new)) {
1022 Contexts.back().CanBeExpression = false;
1023 } else if (Current.isOneOf(tok::semi, tok::exclaim)) {
1024 // This should be the condition or increment in a for-loop.
1025 Contexts.back().IsExpression = true;
1029 void determineTokenType(FormatToken &Current) {
1030 if (!Current.is(TT_Unknown))
1031 // The token type is already known.
1034 if (Style.Language == FormatStyle::LK_JavaScript) {
1035 if (Current.is(tok::exclaim)) {
1036 if (Current.Previous &&
1037 (Current.Previous->isOneOf(tok::identifier, tok::r_paren,
1038 tok::r_square, tok::r_brace) ||
1039 Current.Previous->Tok.isLiteral())) {
1040 Current.Type = TT_JsNonNullAssertion;
1044 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1045 Current.Type = TT_JsNonNullAssertion;
1051 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1052 // function declaration have been found. In this case, 'Current' is a
1053 // trailing token of this declaration and thus cannot be a name.
1054 if (Current.is(Keywords.kw_instanceof)) {
1055 Current.Type = TT_BinaryOperator;
1056 } else if (isStartOfName(Current) &&
1057 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1058 Contexts.back().FirstStartOfName = &Current;
1059 Current.Type = TT_StartOfName;
1060 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1062 } else if (Current.is(tok::arrow) &&
1063 Style.Language == FormatStyle::LK_Java) {
1064 Current.Type = TT_LambdaArrow;
1065 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1066 Current.NestingLevel == 0) {
1067 Current.Type = TT_TrailingReturnArrow;
1068 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1070 determineStarAmpUsage(Current, Contexts.back().CanBeExpression &&
1071 Contexts.back().IsExpression,
1072 Contexts.back().InTemplateArgument);
1073 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1074 Current.Type = determinePlusMinusCaretUsage(Current);
1075 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1076 Contexts.back().CaretFound = true;
1077 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1078 Current.Type = determineIncrementUsage(Current);
1079 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1080 Current.Type = TT_UnaryOperator;
1081 } else if (Current.is(tok::question)) {
1082 if (Style.Language == FormatStyle::LK_JavaScript &&
1083 Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1084 // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1085 // on the interface, not a ternary expression.
1086 Current.Type = TT_JsTypeOptionalQuestion;
1088 Current.Type = TT_ConditionalExpr;
1090 } else if (Current.isBinaryOperator() &&
1091 (!Current.Previous || Current.Previous->isNot(tok::l_square))) {
1092 Current.Type = TT_BinaryOperator;
1093 } else if (Current.is(tok::comment)) {
1094 if (Current.TokenText.startswith("/*")) {
1095 if (Current.TokenText.endswith("*/"))
1096 Current.Type = TT_BlockComment;
1098 // The lexer has for some reason determined a comment here. But we
1099 // cannot really handle it, if it isn't properly terminated.
1100 Current.Tok.setKind(tok::unknown);
1102 Current.Type = TT_LineComment;
1104 } else if (Current.is(tok::r_paren)) {
1105 if (rParenEndsCast(Current))
1106 Current.Type = TT_CastRParen;
1107 if (Current.MatchingParen && Current.Next &&
1108 !Current.Next->isBinaryOperator() &&
1109 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1110 tok::comma, tok::period, tok::arrow,
1112 if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1113 // Make sure this isn't the return type of an Obj-C block declaration
1114 if (AfterParen->Tok.isNot(tok::caret)) {
1115 if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1116 if (BeforeParen->is(tok::identifier) &&
1117 BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1118 (!BeforeParen->Previous ||
1119 BeforeParen->Previous->ClosesTemplateDeclaration))
1120 Current.Type = TT_FunctionAnnotationRParen;
1123 } else if (Current.is(tok::at) && Current.Next &&
1124 Style.Language != FormatStyle::LK_JavaScript &&
1125 Style.Language != FormatStyle::LK_Java) {
1126 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1127 // marks declarations and properties that need special formatting.
1128 switch (Current.Next->Tok.getObjCKeywordID()) {
1129 case tok::objc_interface:
1130 case tok::objc_implementation:
1131 case tok::objc_protocol:
1132 Current.Type = TT_ObjCDecl;
1134 case tok::objc_property:
1135 Current.Type = TT_ObjCProperty;
1140 } else if (Current.is(tok::period)) {
1141 FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1142 if (PreviousNoComment &&
1143 PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1144 Current.Type = TT_DesignatedInitializerPeriod;
1145 else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1146 Current.Previous->isOneOf(TT_JavaAnnotation,
1147 TT_LeadingJavaAnnotation)) {
1148 Current.Type = Current.Previous->Type;
1150 } else if (Current.isOneOf(tok::identifier, tok::kw_const) &&
1152 !Current.Previous->isOneOf(tok::equal, tok::at) &&
1153 Line.MightBeFunctionDecl && Contexts.size() == 1) {
1154 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1155 // function declaration have been found.
1156 Current.Type = TT_TrailingAnnotation;
1157 } else if ((Style.Language == FormatStyle::LK_Java ||
1158 Style.Language == FormatStyle::LK_JavaScript) &&
1160 if (Current.Previous->is(tok::at) &&
1161 Current.isNot(Keywords.kw_interface)) {
1162 const FormatToken &AtToken = *Current.Previous;
1163 const FormatToken *Previous = AtToken.getPreviousNonComment();
1164 if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1165 Current.Type = TT_LeadingJavaAnnotation;
1167 Current.Type = TT_JavaAnnotation;
1168 } else if (Current.Previous->is(tok::period) &&
1169 Current.Previous->isOneOf(TT_JavaAnnotation,
1170 TT_LeadingJavaAnnotation)) {
1171 Current.Type = Current.Previous->Type;
1176 /// \brief Take a guess at whether \p Tok starts a name of a function or
1177 /// variable declaration.
1179 /// This is a heuristic based on whether \p Tok is an identifier following
1180 /// something that is likely a type.
1181 bool isStartOfName(const FormatToken &Tok) {
1182 if (Tok.isNot(tok::identifier) || !Tok.Previous)
1185 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1188 if (Style.Language == FormatStyle::LK_JavaScript &&
1189 Tok.Previous->is(Keywords.kw_in))
1192 // Skip "const" as it does not have an influence on whether this is a name.
1193 FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1194 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1195 PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1197 if (!PreviousNotConst)
1200 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1201 PreviousNotConst->Previous &&
1202 PreviousNotConst->Previous->is(tok::hash);
1204 if (PreviousNotConst->is(TT_TemplateCloser))
1205 return PreviousNotConst && PreviousNotConst->MatchingParen &&
1206 PreviousNotConst->MatchingParen->Previous &&
1207 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1208 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1210 if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen &&
1211 PreviousNotConst->MatchingParen->Previous &&
1212 PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype))
1215 return (!IsPPKeyword &&
1216 PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1217 PreviousNotConst->is(TT_PointerOrReference) ||
1218 PreviousNotConst->isSimpleTypeSpecifier();
1221 /// \brief Determine whether ')' is ending a cast.
1222 bool rParenEndsCast(const FormatToken &Tok) {
1223 // C-style casts are only used in C++ and Java.
1224 if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java)
1227 // Empty parens aren't casts and there are no casts at the end of the line.
1228 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1231 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1233 // If there is a closing parenthesis left of the current parentheses,
1234 // look past it as these might be chained casts.
1235 if (LeftOfParens->is(tok::r_paren)) {
1236 if (!LeftOfParens->MatchingParen ||
1237 !LeftOfParens->MatchingParen->Previous)
1239 LeftOfParens = LeftOfParens->MatchingParen->Previous;
1242 // If there is an identifier (or with a few exceptions a keyword) right
1243 // before the parentheses, this is unlikely to be a cast.
1244 if (LeftOfParens->Tok.getIdentifierInfo() &&
1245 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1249 // Certain other tokens right before the parentheses are also signals that
1250 // this cannot be a cast.
1251 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1252 TT_TemplateCloser, tok::ellipsis))
1256 if (Tok.Next->is(tok::question))
1259 // As Java has no function types, a "(" after the ")" likely means that this
1261 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1264 // If a (non-string) literal follows, this is likely a cast.
1265 if (Tok.Next->isNot(tok::string_literal) &&
1266 (Tok.Next->Tok.isLiteral() ||
1267 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1270 // Heuristically try to determine whether the parentheses contain a type.
1271 bool ParensAreType =
1273 Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) ||
1274 Tok.Previous->isSimpleTypeSpecifier();
1275 bool ParensCouldEndDecl =
1276 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1277 if (ParensAreType && !ParensCouldEndDecl)
1280 // At this point, we heuristically assume that there are no casts at the
1281 // start of the line. We assume that we have found most cases where there
1282 // are by the logic above, e.g. "(void)x;".
1286 // Certain token types inside the parentheses mean that this can't be a
1288 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1289 Token = Token->Next)
1290 if (Token->is(TT_BinaryOperator))
1293 // If the following token is an identifier or 'this', this is a cast. All
1294 // cases where this can be something else are handled above.
1295 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1298 if (!Tok.Next->Next)
1301 // If the next token after the parenthesis is a unary operator, assume
1302 // that this is cast, unless there are unexpected tokens inside the
1305 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1306 if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1307 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1309 // Search for unexpected tokens.
1310 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1311 Prev = Prev->Previous) {
1312 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1318 /// \brief Return the type of the given token assuming it is * or &.
1319 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1320 bool InTemplateArgument) {
1321 if (Style.Language == FormatStyle::LK_JavaScript)
1322 return TT_BinaryOperator;
1324 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1326 return TT_UnaryOperator;
1328 const FormatToken *NextToken = Tok.getNextNonComment();
1330 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) ||
1331 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1332 return TT_PointerOrReference;
1334 if (PrevToken->is(tok::coloncolon))
1335 return TT_PointerOrReference;
1337 if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1338 tok::comma, tok::semi, tok::kw_return, tok::colon,
1339 tok::equal, tok::kw_delete, tok::kw_sizeof) ||
1340 PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1341 TT_UnaryOperator, TT_CastRParen))
1342 return TT_UnaryOperator;
1344 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1345 return TT_PointerOrReference;
1346 if (NextToken->is(tok::kw_operator) && !IsExpression)
1347 return TT_PointerOrReference;
1348 if (NextToken->isOneOf(tok::comma, tok::semi))
1349 return TT_PointerOrReference;
1351 if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen &&
1352 PrevToken->MatchingParen->Previous &&
1353 PrevToken->MatchingParen->Previous->isOneOf(tok::kw_typeof,
1355 return TT_PointerOrReference;
1357 if (PrevToken->Tok.isLiteral() ||
1358 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1359 tok::kw_false, tok::r_brace) ||
1360 NextToken->Tok.isLiteral() ||
1361 NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1362 NextToken->isUnaryOperator() ||
1363 // If we know we're in a template argument, there are no named
1364 // declarations. Thus, having an identifier on the right-hand side
1365 // indicates a binary operator.
1366 (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1367 return TT_BinaryOperator;
1369 // "&&(" is quite unlikely to be two successive unary "&".
1370 if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren))
1371 return TT_BinaryOperator;
1373 // This catches some cases where evaluation order is used as control flow:
1375 const FormatToken *NextNextToken = NextToken->getNextNonComment();
1376 if (NextNextToken && NextNextToken->is(tok::arrow))
1377 return TT_BinaryOperator;
1379 // It is very unlikely that we are going to find a pointer or reference type
1380 // definition on the RHS of an assignment.
1381 if (IsExpression && !Contexts.back().CaretFound)
1382 return TT_BinaryOperator;
1384 return TT_PointerOrReference;
1387 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
1388 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1390 return TT_UnaryOperator;
1392 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator) &&
1393 !PrevToken->is(tok::exclaim))
1394 // There aren't any trailing unary operators except for TypeScript's
1395 // non-null operator (!). Thus, this must be squence of leading operators.
1396 return TT_UnaryOperator;
1398 // Use heuristics to recognize unary operators.
1399 if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
1400 tok::question, tok::colon, tok::kw_return,
1401 tok::kw_case, tok::at, tok::l_brace))
1402 return TT_UnaryOperator;
1404 // There can't be two consecutive binary operators.
1405 if (PrevToken->is(TT_BinaryOperator))
1406 return TT_UnaryOperator;
1408 // Fall back to marking the token as binary operator.
1409 return TT_BinaryOperator;
1412 /// \brief Determine whether ++/-- are pre- or post-increments/-decrements.
1413 TokenType determineIncrementUsage(const FormatToken &Tok) {
1414 const FormatToken *PrevToken = Tok.getPreviousNonComment();
1415 if (!PrevToken || PrevToken->is(TT_CastRParen))
1416 return TT_UnaryOperator;
1417 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
1418 return TT_TrailingUnaryOperator;
1420 return TT_UnaryOperator;
1423 SmallVector<Context, 8> Contexts;
1425 const FormatStyle &Style;
1426 AnnotatedLine &Line;
1427 FormatToken *CurrentToken;
1429 const AdditionalKeywords &Keywords;
1431 // Set of "<" tokens that do not open a template parameter list. If parseAngle
1432 // determines that a specific token can't be a template opener, it will make
1433 // same decision irrespective of the decisions for tokens leading up to it.
1434 // Store this information to prevent this from causing exponential runtime.
1435 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
1438 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
1439 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
1441 /// \brief Parses binary expressions by inserting fake parenthesis based on
1442 /// operator precedence.
1443 class ExpressionParser {
1445 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
1446 AnnotatedLine &Line)
1447 : Style(Style), Keywords(Keywords), Current(Line.First) {}
1449 /// \brief Parse expressions with the given operatore precedence.
1450 void parse(int Precedence = 0) {
1451 // Skip 'return' and ObjC selector colons as they are not part of a binary
1453 while (Current && (Current->is(tok::kw_return) ||
1454 (Current->is(tok::colon) &&
1455 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
1458 if (!Current || Precedence > PrecedenceArrowAndPeriod)
1461 // Conditional expressions need to be parsed separately for proper nesting.
1462 if (Precedence == prec::Conditional) {
1463 parseConditionalExpr();
1467 // Parse unary operators, which all have a higher precedence than binary
1469 if (Precedence == PrecedenceUnaryOperator) {
1470 parseUnaryOperator();
1474 FormatToken *Start = Current;
1475 FormatToken *LatestOperator = nullptr;
1476 unsigned OperatorIndex = 0;
1479 // Consume operators with higher precedence.
1480 parse(Precedence + 1);
1482 int CurrentPrecedence = getCurrentPrecedence();
1484 if (Current && Current->is(TT_SelectorName) &&
1485 Precedence == CurrentPrecedence) {
1487 addFakeParenthesis(Start, prec::Level(Precedence));
1491 // At the end of the line or when an operator with higher precedence is
1492 // found, insert fake parenthesis and return.
1494 (Current->closesScope() &&
1495 (Current->MatchingParen || Current->is(TT_TemplateString))) ||
1496 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
1497 (CurrentPrecedence == prec::Conditional &&
1498 Precedence == prec::Assignment && Current->is(tok::colon))) {
1502 // Consume scopes: (), [], <> and {}
1503 if (Current->opensScope()) {
1504 // In fragment of a JavaScript template string can look like '}..${' and
1505 // thus close a scope and open a new one at the same time.
1506 while (Current && (!Current->closesScope() || Current->opensScope())) {
1513 if (CurrentPrecedence == Precedence) {
1515 LatestOperator->NextOperator = Current;
1516 LatestOperator = Current;
1517 Current->OperatorIndex = OperatorIndex;
1520 next(/*SkipPastLeadingComments=*/Precedence > 0);
1524 if (LatestOperator && (Current || Precedence > 0)) {
1525 // LatestOperator->LastOperator = true;
1526 if (Precedence == PrecedenceArrowAndPeriod) {
1527 // Call expressions don't have a binary operator precedence.
1528 addFakeParenthesis(Start, prec::Unknown);
1530 addFakeParenthesis(Start, prec::Level(Precedence));
1536 /// \brief Gets the precedence (+1) of the given token for binary operators
1537 /// and other tokens that we treat like binary operators.
1538 int getCurrentPrecedence() {
1540 const FormatToken *NextNonComment = Current->getNextNonComment();
1541 if (Current->is(TT_ConditionalExpr))
1542 return prec::Conditional;
1543 if (NextNonComment && NextNonComment->is(tok::colon) &&
1544 NextNonComment->is(TT_DictLiteral))
1545 return prec::Assignment;
1546 if (Current->is(TT_JsComputedPropertyName))
1547 return prec::Assignment;
1548 if (Current->is(TT_LambdaArrow))
1550 if (Current->is(TT_JsFatArrow))
1551 return prec::Assignment;
1552 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
1553 (Current->is(tok::comment) && NextNonComment &&
1554 NextNonComment->is(TT_SelectorName)))
1556 if (Current->is(TT_RangeBasedForLoopColon))
1558 if ((Style.Language == FormatStyle::LK_Java ||
1559 Style.Language == FormatStyle::LK_JavaScript) &&
1560 Current->is(Keywords.kw_instanceof))
1561 return prec::Relational;
1562 if (Style.Language == FormatStyle::LK_JavaScript &&
1563 Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
1564 return prec::Relational;
1565 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
1566 return Current->getPrecedence();
1567 if (Current->isOneOf(tok::period, tok::arrow))
1568 return PrecedenceArrowAndPeriod;
1569 if ((Style.Language == FormatStyle::LK_Java ||
1570 Style.Language == FormatStyle::LK_JavaScript) &&
1571 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
1572 Keywords.kw_throws))
1578 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
1579 Start->FakeLParens.push_back(Precedence);
1580 if (Precedence > prec::Unknown)
1581 Start->StartsBinaryExpression = true;
1583 FormatToken *Previous = Current->Previous;
1584 while (Previous->is(tok::comment) && Previous->Previous)
1585 Previous = Previous->Previous;
1586 ++Previous->FakeRParens;
1587 if (Precedence > prec::Unknown)
1588 Previous->EndsBinaryExpression = true;
1592 /// \brief Parse unary operator expressions and surround them with fake
1593 /// parentheses if appropriate.
1594 void parseUnaryOperator() {
1595 if (!Current || Current->isNot(TT_UnaryOperator)) {
1596 parse(PrecedenceArrowAndPeriod);
1600 FormatToken *Start = Current;
1602 parseUnaryOperator();
1604 // The actual precedence doesn't matter.
1605 addFakeParenthesis(Start, prec::Unknown);
1608 void parseConditionalExpr() {
1609 while (Current && Current->isTrailingComment()) {
1612 FormatToken *Start = Current;
1613 parse(prec::LogicalOr);
1614 if (!Current || !Current->is(tok::question))
1617 parse(prec::Assignment);
1618 if (!Current || Current->isNot(TT_ConditionalExpr))
1621 parse(prec::Assignment);
1622 addFakeParenthesis(Start, prec::Conditional);
1625 void next(bool SkipPastLeadingComments = true) {
1627 Current = Current->Next;
1629 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
1630 Current->isTrailingComment())
1631 Current = Current->Next;
1634 const FormatStyle &Style;
1635 const AdditionalKeywords &Keywords;
1636 FormatToken *Current;
1639 } // end anonymous namespace
1641 void TokenAnnotator::setCommentLineLevels(
1642 SmallVectorImpl<AnnotatedLine *> &Lines) {
1643 const AnnotatedLine *NextNonCommentLine = nullptr;
1644 for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
1647 bool CommentLine = (*I)->First;
1648 for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
1649 if (!Tok->is(tok::comment)) {
1650 CommentLine = false;
1654 if (NextNonCommentLine && CommentLine)
1655 (*I)->Level = NextNonCommentLine->Level;
1657 NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
1659 setCommentLineLevels((*I)->Children);
1663 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
1664 unsigned Result = 0;
1665 for (const auto* Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
1666 Result = std::max(Result, Tok->NestingLevel);
1670 void TokenAnnotator::annotate(AnnotatedLine &Line) {
1671 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1672 E = Line.Children.end();
1676 AnnotatingParser Parser(Style, Line, Keywords);
1677 Line.Type = Parser.parseLine();
1679 // With very deep nesting, ExpressionParser uses lots of stack and the
1680 // formatting algorithm is very slow. We're not going to do a good job here
1681 // anyway - it's probably generated code being formatted by mistake.
1682 // Just skip the whole line.
1683 if (maxNestingDepth(Line) > 50)
1684 Line.Type = LT_Invalid;
1686 if (Line.Type == LT_Invalid)
1689 ExpressionParser ExprParser(Style, Keywords, Line);
1692 if (Line.startsWith(TT_ObjCMethodSpecifier))
1693 Line.Type = LT_ObjCMethodDecl;
1694 else if (Line.startsWith(TT_ObjCDecl))
1695 Line.Type = LT_ObjCDecl;
1696 else if (Line.startsWith(TT_ObjCProperty))
1697 Line.Type = LT_ObjCProperty;
1699 Line.First->SpacesRequiredBefore = 1;
1700 Line.First->CanBreakBefore = Line.First->MustBreakBefore;
1703 // This function heuristically determines whether 'Current' starts the name of a
1704 // function declaration.
1705 static bool isFunctionDeclarationName(const FormatToken &Current,
1706 const AnnotatedLine &Line) {
1707 auto skipOperatorName = [](const FormatToken* Next) -> const FormatToken* {
1708 for (; Next; Next = Next->Next) {
1709 if (Next->is(TT_OverloadedOperatorLParen))
1711 if (Next->is(TT_OverloadedOperator))
1713 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
1714 // For 'new[]' and 'delete[]'.
1715 if (Next->Next && Next->Next->is(tok::l_square) &&
1716 Next->Next->Next && Next->Next->Next->is(tok::r_square))
1717 Next = Next->Next->Next;
1726 // Find parentheses of parameter list.
1727 const FormatToken *Next = Current.Next;
1728 if (Current.is(tok::kw_operator)) {
1729 if (Current.Previous && Current.Previous->is(tok::coloncolon))
1731 Next = skipOperatorName(Next);
1733 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
1735 for (; Next; Next = Next->Next) {
1736 if (Next->is(TT_TemplateOpener)) {
1737 Next = Next->MatchingParen;
1738 } else if (Next->is(tok::coloncolon)) {
1742 if (Next->is(tok::kw_operator)) {
1743 Next = skipOperatorName(Next->Next);
1746 if (!Next->is(tok::identifier))
1748 } else if (Next->is(tok::l_paren)) {
1756 // Check whether parameter list can belong to a function declaration.
1757 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
1759 // If the lines ends with "{", this is likely an function definition.
1760 if (Line.Last->is(tok::l_brace))
1762 if (Next->Next == Next->MatchingParen)
1763 return true; // Empty parentheses.
1764 // If there is an &/&& after the r_paren, this is likely a function.
1765 if (Next->MatchingParen->Next &&
1766 Next->MatchingParen->Next->is(TT_PointerOrReference))
1768 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
1770 if (Tok->is(tok::l_paren) && Tok->MatchingParen) {
1771 Tok = Tok->MatchingParen;
1774 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
1775 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
1777 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
1778 Tok->Tok.isLiteral())
1784 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
1785 assert(Line.MightBeFunctionDecl);
1787 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
1788 Style.AlwaysBreakAfterReturnType ==
1789 FormatStyle::RTBS_TopLevelDefinitions) &&
1793 switch (Style.AlwaysBreakAfterReturnType) {
1794 case FormatStyle::RTBS_None:
1796 case FormatStyle::RTBS_All:
1797 case FormatStyle::RTBS_TopLevel:
1799 case FormatStyle::RTBS_AllDefinitions:
1800 case FormatStyle::RTBS_TopLevelDefinitions:
1801 return Line.mightBeFunctionDefinition();
1807 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
1808 for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1809 E = Line.Children.end();
1811 calculateFormattingInformation(**I);
1814 Line.First->TotalLength =
1815 Line.First->IsMultiline ? Style.ColumnLimit : Line.First->ColumnWidth;
1816 FormatToken *Current = Line.First->Next;
1817 bool InFunctionDecl = Line.MightBeFunctionDecl;
1819 if (isFunctionDeclarationName(*Current, Line))
1820 Current->Type = TT_FunctionDeclarationName;
1821 if (Current->is(TT_LineComment)) {
1822 if (Current->Previous->BlockKind == BK_BracedInit &&
1823 Current->Previous->opensScope())
1824 Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1;
1826 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
1828 // If we find a trailing comment, iterate backwards to determine whether
1829 // it seems to relate to a specific parameter. If so, break before that
1830 // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
1831 // to the previous line in:
1835 if (!Current->HasUnescapedNewline) {
1836 for (FormatToken *Parameter = Current->Previous; Parameter;
1837 Parameter = Parameter->Previous) {
1838 if (Parameter->isOneOf(tok::comment, tok::r_brace))
1840 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
1841 if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
1842 Parameter->HasUnescapedNewline)
1843 Parameter->MustBreakBefore = true;
1848 } else if (Current->SpacesRequiredBefore == 0 &&
1849 spaceRequiredBefore(Line, *Current)) {
1850 Current->SpacesRequiredBefore = 1;
1853 Current->MustBreakBefore =
1854 Current->MustBreakBefore || mustBreakBefore(Line, *Current);
1856 if (!Current->MustBreakBefore && InFunctionDecl &&
1857 Current->is(TT_FunctionDeclarationName))
1858 Current->MustBreakBefore = mustBreakForReturnType(Line);
1860 Current->CanBreakBefore =
1861 Current->MustBreakBefore || canBreakBefore(Line, *Current);
1862 unsigned ChildSize = 0;
1863 if (Current->Previous->Children.size() == 1) {
1864 FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
1865 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
1866 : LastOfChild.TotalLength + 1;
1868 const FormatToken *Prev = Current->Previous;
1869 if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
1870 (Prev->Children.size() == 1 &&
1871 Prev->Children[0]->First->MustBreakBefore) ||
1872 Current->IsMultiline)
1873 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
1875 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
1876 ChildSize + Current->SpacesRequiredBefore;
1878 if (Current->is(TT_CtorInitializerColon))
1879 InFunctionDecl = false;
1881 // FIXME: Only calculate this if CanBreakBefore is true once static
1882 // initializers etc. are sorted out.
1883 // FIXME: Move magic numbers to a better place.
1884 Current->SplitPenalty = 20 * Current->BindingStrength +
1885 splitPenalty(Line, *Current, InFunctionDecl);
1887 Current = Current->Next;
1890 calculateUnbreakableTailLengths(Line);
1891 unsigned IndentLevel = Line.Level;
1892 for (Current = Line.First; Current != nullptr; Current = Current->Next) {
1894 Current->Role->precomputeFormattingInfos(Current);
1895 if (Current->MatchingParen &&
1896 Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
1897 assert(IndentLevel > 0);
1900 Current->IndentLevel = IndentLevel;
1901 if (Current->opensBlockOrBlockTypeList(Style))
1905 DEBUG({ printDebugInfo(Line); });
1908 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
1909 unsigned UnbreakableTailLength = 0;
1910 FormatToken *Current = Line.Last;
1912 Current->UnbreakableTailLength = UnbreakableTailLength;
1913 if (Current->CanBreakBefore ||
1914 Current->isOneOf(tok::comment, tok::string_literal)) {
1915 UnbreakableTailLength = 0;
1917 UnbreakableTailLength +=
1918 Current->ColumnWidth + Current->SpacesRequiredBefore;
1920 Current = Current->Previous;
1924 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
1925 const FormatToken &Tok,
1926 bool InFunctionDecl) {
1927 const FormatToken &Left = *Tok.Previous;
1928 const FormatToken &Right = Tok;
1930 if (Left.is(tok::semi))
1933 if (Style.Language == FormatStyle::LK_Java) {
1934 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
1936 if (Right.is(Keywords.kw_implements))
1938 if (Left.is(tok::comma) && Left.NestingLevel == 0)
1940 } else if (Style.Language == FormatStyle::LK_JavaScript) {
1941 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
1943 if (Left.is(TT_JsTypeColon))
1945 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
1946 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
1950 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
1952 if (Right.is(tok::l_square)) {
1953 if (Style.Language == FormatStyle::LK_Proto)
1955 if (Left.is(tok::r_square))
1957 // Slightly prefer formatting local lambda definitions like functions.
1958 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
1960 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
1961 TT_ArrayInitializerLSquare))
1965 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
1966 Right.is(tok::kw_operator)) {
1967 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
1969 if (Left.is(TT_StartOfName))
1971 if (InFunctionDecl && Right.NestingLevel == 0)
1972 return Style.PenaltyReturnTypeOnItsOwnLine;
1975 if (Right.is(TT_PointerOrReference))
1977 if (Right.is(TT_LambdaArrow))
1979 if (Left.is(tok::equal) && Right.is(tok::l_brace))
1981 if (Left.is(TT_CastRParen))
1983 if (Left.is(tok::coloncolon) ||
1984 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
1986 if (Left.isOneOf(tok::kw_class, tok::kw_struct))
1988 if (Left.is(tok::comment))
1991 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon))
1994 if (Right.isMemberAccess()) {
1995 // Breaking before the "./->" of a chained call/member access is reasonably
1996 // cheap, as formatting those with one call per line is generally
1997 // desirable. In particular, it should be cheaper to break before the call
1998 // than it is to break inside a call's parameters, which could lead to weird
1999 // "hanging" indents. The exception is the very last "./->" to support this
2000 // frequent pattern:
2002 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
2005 // which might otherwise be blown up onto many lines. Here, clang-format
2006 // won't produce "hanging" indents anyway as there is no other trailing
2009 // Also apply higher penalty is not a call as that might lead to a wrapping
2013 // .aaaaaaaaa.bbbbbbbb(cccccccc);
2014 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
2019 if (Right.is(TT_TrailingAnnotation) &&
2020 (!Right.Next || Right.Next->isNot(tok::l_paren))) {
2021 // Moving trailing annotations to the next line is fine for ObjC method
2023 if (Line.startsWith(TT_ObjCMethodSpecifier))
2025 // Generally, breaking before a trailing annotation is bad unless it is
2026 // function-like. It seems to be especially preferable to keep standard
2027 // annotations (i.e. "const", "final" and "override") on the same line.
2028 // Use a slightly higher penalty after ")" so that annotations like
2029 // "const override" are kept together.
2030 bool is_short_annotation = Right.TokenText.size() < 10;
2031 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2034 // In for-loops, prefer breaking at ',' and ';'.
2035 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2038 // In Objective-C method expressions, prefer breaking before "param:" over
2039 // breaking after it.
2040 if (Right.is(TT_SelectorName))
2042 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2043 return Line.MightBeFunctionDecl ? 50 : 500;
2045 if (Left.is(tok::l_paren) && InFunctionDecl &&
2046 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2048 if (Left.is(tok::l_paren) && Left.Previous &&
2049 Left.Previous->isOneOf(tok::kw_if, tok::kw_for))
2051 if (Left.is(tok::equal) && InFunctionDecl)
2053 if (Right.is(tok::r_brace))
2055 if (Left.is(TT_TemplateOpener))
2057 if (Left.opensScope()) {
2058 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
2060 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2063 if (Left.is(TT_JavaAnnotation))
2066 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2067 Left.Previous->isLabelString() &&
2068 (Left.NextOperator || Left.OperatorIndex != 0))
2070 if (Right.is(tok::plus) && Left.isLabelString() &&
2071 (Right.NextOperator || Right.OperatorIndex != 0))
2073 if (Left.is(tok::comma))
2075 if (Right.is(tok::lessless) && Left.isLabelString() &&
2076 (Right.NextOperator || Right.OperatorIndex != 1))
2078 if (Right.is(tok::lessless)) {
2079 // Breaking at a << is really cheap.
2080 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2081 // Slightly prefer to break before the first one in log-like statements.
2085 if (Left.is(TT_ConditionalExpr))
2086 return prec::Conditional;
2087 prec::Level Level = Left.getPrecedence();
2088 if (Level != prec::Unknown)
2090 Level = Right.getPrecedence();
2091 if (Level != prec::Unknown)
2097 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2098 const FormatToken &Left,
2099 const FormatToken &Right) {
2100 if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2102 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2103 Left.Tok.getObjCKeywordID() == tok::objc_property)
2105 if (Right.is(tok::hashhash))
2106 return Left.is(tok::hash);
2107 if (Left.isOneOf(tok::hashhash, tok::hash))
2108 return Right.is(tok::hash);
2109 if (Left.is(tok::l_paren) && Right.is(tok::r_paren))
2110 return Style.SpaceInEmptyParentheses;
2111 if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2112 return (Right.is(TT_CastRParen) ||
2113 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2114 ? Style.SpacesInCStyleCastParentheses
2115 : Style.SpacesInParentheses;
2116 if (Right.isOneOf(tok::semi, tok::comma))
2118 if (Right.is(tok::less) &&
2119 Line.Type == LT_ObjCDecl && Style.ObjCSpaceBeforeProtocolList)
2121 if (Right.is(tok::less) && Left.is(tok::kw_template))
2122 return Style.SpaceAfterTemplateKeyword;
2123 if (Left.isOneOf(tok::exclaim, tok::tilde))
2125 if (Left.is(tok::at) &&
2126 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2127 tok::numeric_constant, tok::l_paren, tok::l_brace,
2128 tok::kw_true, tok::kw_false))
2130 if (Left.is(tok::colon))
2131 return !Left.is(TT_ObjCMethodExpr);
2132 if (Left.is(tok::coloncolon))
2134 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less))
2136 if (Right.is(tok::ellipsis))
2137 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2138 Left.Previous->is(tok::kw_case));
2139 if (Left.is(tok::l_square) && Right.is(tok::amp))
2141 if (Right.is(TT_PointerOrReference))
2142 return (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) ||
2143 (Left.Tok.isLiteral() || (Left.is(tok::kw_const) && Left.Previous &&
2144 Left.Previous->is(tok::r_paren)) ||
2145 (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2146 (Style.PointerAlignment != FormatStyle::PAS_Left ||
2147 (Line.IsMultiVariableDeclStmt &&
2148 (Left.NestingLevel == 0 ||
2149 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2150 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2151 (!Left.is(TT_PointerOrReference) ||
2152 (Style.PointerAlignment != FormatStyle::PAS_Right &&
2153 !Line.IsMultiVariableDeclStmt)))
2155 if (Left.is(TT_PointerOrReference))
2156 return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2157 (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2158 !Right.is(TT_StartOfName)) ||
2159 (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) ||
2160 (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2162 (Style.PointerAlignment != FormatStyle::PAS_Right &&
2163 !Line.IsMultiVariableDeclStmt) &&
2165 !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon));
2166 if (Right.is(tok::star) && Left.is(tok::l_paren))
2168 if (Left.is(tok::l_square))
2169 return (Left.is(TT_ArrayInitializerLSquare) &&
2170 Style.SpacesInContainerLiterals && Right.isNot(tok::r_square)) ||
2171 (Left.is(TT_ArraySubscriptLSquare) && Style.SpacesInSquareBrackets &&
2172 Right.isNot(tok::r_square));
2173 if (Right.is(tok::r_square))
2174 return Right.MatchingParen &&
2175 ((Style.SpacesInContainerLiterals &&
2176 Right.MatchingParen->is(TT_ArrayInitializerLSquare)) ||
2177 (Style.SpacesInSquareBrackets &&
2178 Right.MatchingParen->is(TT_ArraySubscriptLSquare)));
2179 if (Right.is(tok::l_square) &&
2180 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare) &&
2181 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral))
2183 if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2184 return !Left.Children.empty(); // No spaces in "{}".
2185 if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) ||
2186 (Right.is(tok::r_brace) && Right.MatchingParen &&
2187 Right.MatchingParen->BlockKind != BK_Block))
2188 return !Style.Cpp11BracedListStyle;
2189 if (Left.is(TT_BlockComment))
2190 return !Left.TokenText.endswith("=*/");
2191 if (Right.is(tok::l_paren)) {
2192 if (Left.is(tok::r_paren) && Left.is(TT_AttributeParen))
2194 return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
2195 (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
2196 (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while,
2197 tok::kw_switch, tok::kw_case, TT_ForEachMacro,
2199 (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
2200 tok::kw_new, tok::kw_delete) &&
2201 (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
2202 (Style.SpaceBeforeParens == FormatStyle::SBPO_Always &&
2203 (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
2204 Left.is(tok::r_paren)) &&
2205 Line.Type != LT_PreprocessorDirective);
2207 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
2209 if (Right.is(TT_UnaryOperator))
2210 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
2211 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
2212 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
2214 Left.isSimpleTypeSpecifier()) &&
2215 Right.is(tok::l_brace) && Right.getNextNonComment() &&
2216 Right.BlockKind != BK_Block)
2218 if (Left.is(tok::period) || Right.is(tok::period))
2220 if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
2222 if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
2223 Left.MatchingParen->Previous &&
2224 Left.MatchingParen->Previous->is(tok::period))
2225 // A.<B>DoSomething();
2227 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
2232 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
2233 const FormatToken &Right) {
2234 const FormatToken &Left = *Right.Previous;
2235 if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
2236 return true; // Never ever merge two identifiers.
2237 if (Style.isCpp()) {
2238 if (Left.is(tok::kw_operator))
2239 return Right.is(tok::coloncolon);
2240 } else if (Style.Language == FormatStyle::LK_Proto) {
2241 if (Right.is(tok::period) &&
2242 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
2243 Keywords.kw_repeated, Keywords.kw_extend))
2245 if (Right.is(tok::l_paren) &&
2246 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
2248 } else if (Style.Language == FormatStyle::LK_JavaScript) {
2249 if (Left.is(TT_JsFatArrow))
2251 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
2252 Right.MatchingParen) {
2253 const FormatToken *Next = Right.MatchingParen->getNextNonComment();
2254 // An async arrow function, for example: `x = async () => foo();`,
2255 // as opposed to calling a function called async: `x = async();`
2256 if (Next && Next->is(TT_JsFatArrow))
2259 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2260 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2262 if (Left.is(tok::identifier) && Right.is(TT_TemplateString))
2264 if (Right.is(tok::star) &&
2265 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2267 if (Right.isOneOf(tok::l_brace, tok::l_square) &&
2268 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2270 // JS methods can use some keywords as names (e.g. `delete()`).
2271 if (Right.is(tok::l_paren) && Line.MustBeDeclaration &&
2272 Left.Tok.getIdentifierInfo())
2274 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
2276 // "of" is only a keyword if it appears after another identifier
2277 // (e.g. as "const x of y" in a for loop).
2278 (Left.is(Keywords.kw_of) && Left.Previous &&
2279 Left.Previous->Tok.getIdentifierInfo())) &&
2280 (!Left.Previous || !Left.Previous->is(tok::period)))
2282 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
2283 Left.Previous->is(tok::period) && Right.is(tok::l_paren))
2285 if (Left.is(Keywords.kw_as) &&
2286 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
2288 if (Left.is(tok::kw_default) && Left.Previous &&
2289 Left.Previous->is(tok::kw_export))
2291 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
2293 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
2295 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
2297 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
2298 Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
2300 if (Left.is(tok::ellipsis))
2302 if (Left.is(TT_TemplateCloser) &&
2303 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
2304 Keywords.kw_implements, Keywords.kw_extends))
2305 // Type assertions ('<type>expr') are not followed by whitespace. Other
2306 // locations that should have whitespace following are identified by the
2307 // above set of follower tokens.
2309 if (Right.is(TT_JsNonNullAssertion))
2311 if (Left.is(TT_JsNonNullAssertion) && Right.is(Keywords.kw_as))
2312 return true; // "x! as string"
2313 } else if (Style.Language == FormatStyle::LK_Java) {
2314 if (Left.is(tok::r_square) && Right.is(tok::l_brace))
2316 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
2317 return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
2318 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
2319 tok::kw_protected) ||
2320 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
2321 Keywords.kw_native)) &&
2322 Right.is(TT_TemplateOpener))
2325 if (Left.is(TT_ImplicitStringLiteral))
2326 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2327 if (Line.Type == LT_ObjCMethodDecl) {
2328 if (Left.is(TT_ObjCMethodSpecifier))
2330 if (Left.is(tok::r_paren) && Right.is(tok::identifier))
2331 // Don't space between ')' and <id>
2334 if (Line.Type == LT_ObjCProperty &&
2335 (Right.is(tok::equal) || Left.is(tok::equal)))
2338 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
2339 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
2341 if (Right.is(TT_OverloadedOperatorLParen))
2342 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2343 if (Left.is(tok::comma))
2345 if (Right.is(tok::comma))
2347 if (Right.isOneOf(TT_CtorInitializerColon, TT_ObjCBlockLParen))
2349 if (Right.is(tok::colon)) {
2350 if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
2351 !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
2353 if (Right.is(TT_ObjCMethodExpr))
2355 if (Left.is(tok::question))
2357 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
2359 if (Right.is(TT_DictLiteral))
2360 return Style.SpacesInContainerLiterals;
2363 if (Left.is(TT_UnaryOperator))
2364 return Right.is(TT_BinaryOperator);
2366 // If the next token is a binary operator or a selector name, we have
2367 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
2368 if (Left.is(TT_CastRParen))
2369 return Style.SpaceAfterCStyleCast ||
2370 Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
2372 if (Left.is(tok::greater) && Right.is(tok::greater))
2373 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
2374 (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
2375 if (Right.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
2376 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar))
2378 if (!Style.SpaceBeforeAssignmentOperators &&
2379 Right.getPrecedence() == prec::Assignment)
2381 if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
2382 // Generally don't remove existing spaces between an identifier and "::".
2383 // The identifier might actually be a macro name such as ALWAYS_INLINE. If
2384 // this turns out to be too lenient, add analysis of the identifier itself.
2385 return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2386 if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment))
2387 return (Left.is(TT_TemplateOpener) &&
2388 Style.Standard == FormatStyle::LS_Cpp03) ||
2389 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
2390 tok::kw___super, TT_TemplateCloser, TT_TemplateOpener));
2391 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
2392 return Style.SpacesInAngles;
2393 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
2394 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
2395 !Right.is(tok::r_paren)))
2397 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
2398 Right.isNot(TT_FunctionTypeLParen))
2399 return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2400 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
2401 Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
2403 if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
2404 Line.startsWith(tok::hash))
2406 if (Right.is(TT_TrailingUnaryOperator))
2408 if (Left.is(TT_RegexLiteral))
2410 return spaceRequiredBetween(Line, Left, Right);
2413 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
2414 static bool isAllmanBrace(const FormatToken &Tok) {
2415 return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block &&
2416 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral);
2419 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
2420 const FormatToken &Right) {
2421 const FormatToken &Left = *Right.Previous;
2422 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
2425 if (Style.Language == FormatStyle::LK_JavaScript) {
2426 // FIXME: This might apply to other languages and token kinds.
2427 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
2428 Left.Previous->is(tok::string_literal))
2430 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
2431 Left.Previous && Left.Previous->is(tok::equal) &&
2432 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
2434 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
2436 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
2437 // Object literals on the top level of a file are treated as "enum-style".
2438 // Each key/value pair is put on a separate line, instead of bin-packing.
2440 if (Left.is(tok::l_brace) && Line.Level == 0 &&
2441 (Line.startsWith(tok::kw_enum) ||
2442 Line.startsWith(tok::kw_export, tok::kw_enum)))
2443 // JavaScript top-level enum key/value pairs are put on separate lines
2444 // instead of bin-packing.
2446 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
2447 !Left.Children.empty())
2448 // Support AllowShortFunctionsOnASingleLine for JavaScript.
2449 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
2450 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
2451 (Left.NestingLevel == 0 && Line.Level == 0 &&
2452 Style.AllowShortFunctionsOnASingleLine ==
2453 FormatStyle::SFS_Inline);
2454 } else if (Style.Language == FormatStyle::LK_Java) {
2455 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
2456 Right.Next->is(tok::string_literal))
2458 } else if (Style.Language == FormatStyle::LK_Cpp ||
2459 Style.Language == FormatStyle::LK_ObjC ||
2460 Style.Language == FormatStyle::LK_Proto) {
2461 if (Left.isStringLiteral() && Right.isStringLiteral())
2465 // If the last token before a '}' is a comma or a trailing comment, the
2466 // intention is to insert a line break after it in order to make shuffling
2467 // around entries easier.
2468 const FormatToken *BeforeClosingBrace = nullptr;
2469 if (Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) &&
2470 Left.BlockKind != BK_Block && Left.MatchingParen)
2471 BeforeClosingBrace = Left.MatchingParen->Previous;
2472 else if (Right.MatchingParen &&
2473 Right.MatchingParen->isOneOf(tok::l_brace,
2474 TT_ArrayInitializerLSquare))
2475 BeforeClosingBrace = &Left;
2476 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
2477 BeforeClosingBrace->isTrailingComment()))
2480 if (Right.is(tok::comment))
2481 return Left.BlockKind != BK_BracedInit &&
2482 Left.isNot(TT_CtorInitializerColon) &&
2483 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
2484 if (Left.isTrailingComment())
2486 if (Right.Previous->IsUnterminatedLiteral)
2488 if (Right.is(tok::lessless) && Right.Next &&
2489 Right.Previous->is(tok::string_literal) &&
2490 Right.Next->is(tok::string_literal))
2492 if (Right.Previous->ClosesTemplateDeclaration &&
2493 Right.Previous->MatchingParen &&
2494 Right.Previous->MatchingParen->NestingLevel == 0 &&
2495 Style.AlwaysBreakTemplateDeclarations)
2497 if ((Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) &&
2498 Style.BreakConstructorInitializersBeforeComma &&
2499 !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
2501 // Break only if we have multiple inheritance.
2502 if (Style.BreakBeforeInheritanceComma &&
2503 Right.is(TT_InheritanceComma))
2505 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
2506 // Raw string literals are special wrt. line breaks. The author has made a
2507 // deliberate choice and might have aligned the contents of the string
2508 // literal accordingly. Thus, we try keep existing line breaks.
2509 return Right.NewlinesBefore > 0;
2510 if (Right.Previous->is(tok::l_brace) && Right.NestingLevel == 1 &&
2511 Style.Language == FormatStyle::LK_Proto)
2512 // Don't put enums onto single lines in protocol buffers.
2514 if (Right.is(TT_InlineASMBrace))
2515 return Right.HasUnescapedNewline;
2516 if (isAllmanBrace(Left) || isAllmanBrace(Right))
2517 return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
2518 (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
2519 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
2520 if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine)
2523 if ((Style.Language == FormatStyle::LK_Java ||
2524 Style.Language == FormatStyle::LK_JavaScript) &&
2525 Left.is(TT_LeadingJavaAnnotation) &&
2526 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
2527 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
2533 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
2534 const FormatToken &Right) {
2535 const FormatToken &Left = *Right.Previous;
2537 // Language-specific stuff.
2538 if (Style.Language == FormatStyle::LK_Java) {
2539 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
2540 Keywords.kw_implements))
2542 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
2543 Keywords.kw_implements))
2545 } else if (Style.Language == FormatStyle::LK_JavaScript) {
2546 const FormatToken *NonComment = Right.getPreviousNonComment();
2548 NonComment->isOneOf(
2549 tok::kw_return, tok::kw_continue, tok::kw_break, tok::kw_throw,
2550 Keywords.kw_interface, Keywords.kw_type, tok::kw_static,
2551 tok::kw_public, tok::kw_private, tok::kw_protected,
2552 Keywords.kw_abstract, Keywords.kw_get, Keywords.kw_set))
2553 return false; // Otherwise a semicolon is inserted.
2554 if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
2556 if (Left.is(TT_JsTypeColon))
2558 if (Right.NestingLevel == 0 && Right.is(Keywords.kw_is))
2560 if (Left.is(Keywords.kw_in))
2561 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
2562 if (Right.is(Keywords.kw_in))
2563 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
2564 if (Right.is(Keywords.kw_as))
2565 return false; // must not break before as in 'x as type' casts
2566 if (Left.is(Keywords.kw_as))
2568 if (Left.is(TT_JsNonNullAssertion))
2570 if (Left.is(Keywords.kw_declare) &&
2571 Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
2572 Keywords.kw_function, tok::kw_class, tok::kw_enum,
2573 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
2574 Keywords.kw_let, tok::kw_const))
2575 // See grammar for 'declare' statements at:
2576 // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
2578 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
2579 Right.isOneOf(tok::identifier, tok::string_literal))
2580 return false; // must not break in "module foo { ...}"
2581 if (Right.is(TT_TemplateString) && Right.closesScope())
2583 if (Left.is(TT_TemplateString) && Left.opensScope())
2587 if (Left.is(tok::at))
2589 if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
2591 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
2592 return !Right.is(tok::l_paren);
2593 if (Right.is(TT_PointerOrReference))
2594 return Line.IsMultiVariableDeclStmt ||
2595 (Style.PointerAlignment == FormatStyle::PAS_Right &&
2596 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
2597 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2598 Right.is(tok::kw_operator))
2600 if (Left.is(TT_PointerOrReference))
2602 if (Right.isTrailingComment())
2603 // We rely on MustBreakBefore being set correctly here as we should not
2604 // change the "binding" behavior of a comment.
2605 // The first comment in a braced lists is always interpreted as belonging to
2606 // the first list element. Otherwise, it should be placed outside of the
2608 return Left.BlockKind == BK_BracedInit;
2609 if (Left.is(tok::question) && Right.is(tok::colon))
2611 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
2612 return Style.BreakBeforeTernaryOperators;
2613 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
2614 return !Style.BreakBeforeTernaryOperators;
2615 if (Right.is(TT_InheritanceColon))
2617 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
2618 Left.isNot(TT_SelectorName))
2620 if (Right.is(tok::colon) &&
2621 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
2623 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr))
2625 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
2626 Right.Next->is(TT_ObjCMethodExpr)))
2627 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
2628 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
2630 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
2632 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
2633 TT_OverloadedOperator))
2635 if (Left.is(TT_RangeBasedForLoopColon))
2637 if (Right.is(TT_RangeBasedForLoopColon))
2639 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
2641 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
2642 Left.is(tok::kw_operator))
2644 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
2645 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
2647 if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen))
2649 if (Left.is(tok::l_paren) && Left.Previous &&
2650 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
2652 if (Right.is(TT_ImplicitStringLiteral))
2655 if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
2657 if (Right.is(tok::r_square) && Right.MatchingParen &&
2658 Right.MatchingParen->is(TT_LambdaLSquare))
2661 // We only break before r_brace if there was a corresponding break before
2662 // the l_brace, which is tracked by BreakBeforeClosingBrace.
2663 if (Right.is(tok::r_brace))
2664 return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block;
2666 // Allow breaking after a trailing annotation, e.g. after a method
2668 if (Left.is(TT_TrailingAnnotation))
2669 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
2670 tok::less, tok::coloncolon);
2672 if (Right.is(tok::kw___attribute))
2675 if (Left.is(tok::identifier) && Right.is(tok::string_literal))
2678 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2681 if (Left.is(TT_CtorInitializerComma) &&
2682 Style.BreakConstructorInitializersBeforeComma)
2684 if (Right.is(TT_CtorInitializerComma) &&
2685 Style.BreakConstructorInitializersBeforeComma)
2687 if (Left.is(TT_InheritanceComma) && Style.BreakBeforeInheritanceComma)
2689 if (Right.is(TT_InheritanceComma) && Style.BreakBeforeInheritanceComma)
2691 if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
2692 (Left.is(tok::less) && Right.is(tok::less)))
2694 if (Right.is(TT_BinaryOperator) &&
2695 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
2696 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
2697 Right.getPrecedence() != prec::Assignment))
2699 if (Left.is(TT_ArrayInitializerLSquare))
2701 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
2703 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
2704 !Left.isOneOf(tok::arrowstar, tok::lessless) &&
2705 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
2706 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
2707 Left.getPrecedence() == prec::Assignment))
2709 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
2710 tok::kw_class, tok::kw_struct, tok::comment) ||
2711 Right.isMemberAccess() ||
2712 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
2713 tok::colon, tok::l_square, tok::at) ||
2714 (Left.is(tok::r_paren) &&
2715 Right.isOneOf(tok::identifier, tok::kw_const)) ||
2716 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
2717 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
2720 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
2721 llvm::errs() << "AnnotatedTokens:\n";
2722 const FormatToken *Tok = Line.First;
2724 llvm::errs() << " M=" << Tok->MustBreakBefore
2725 << " C=" << Tok->CanBreakBefore
2726 << " T=" << getTokenTypeName(Tok->Type)
2727 << " S=" << Tok->SpacesRequiredBefore
2728 << " B=" << Tok->BlockParameterCount
2729 << " BK=" << Tok->BlockKind
2730 << " P=" << Tok->SplitPenalty << " Name=" << Tok->Tok.getName()
2731 << " L=" << Tok->TotalLength << " PPK=" << Tok->PackingKind
2733 for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
2734 llvm::errs() << Tok->FakeLParens[i] << "/";
2735 llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
2736 llvm::errs() << " Text='" << Tok->TokenText << "'\n";
2738 assert(Tok == Line.Last);
2741 llvm::errs() << "----\n";
2744 } // namespace format
2745 } // namespace clang