1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 ///
9 /// \file
10 /// This file implements a token annotator, i.e. creates
11 /// \c AnnotatedTokens out of \c FormatTokens with required extra information.
12 ///
13 //===----------------------------------------------------------------------===//
14
15 #include "TokenAnnotator.h"
16 #include "FormatToken.h"
17 #include "clang/Basic/SourceManager.h"
18 #include "clang/Basic/TokenKinds.h"
19 #include "llvm/ADT/SmallPtrSet.h"
20 #include "llvm/Support/Debug.h"
21
22 #define DEBUG_TYPE "format-token-annotator"
23
24 namespace clang {
25 namespace format {
26
27 namespace {
28
29 /// Returns \c true if the line starts with a token that can start a statement
30 /// with an initializer.
startsWithInitStatement(const AnnotatedLine & Line)31 static bool startsWithInitStatement(const AnnotatedLine &Line) {
32 return Line.startsWith(tok::kw_for) || Line.startsWith(tok::kw_if) ||
33 Line.startsWith(tok::kw_switch);
34 }
35
36 /// Returns \c true if the token can be used as an identifier in
37 /// an Objective-C \c \@selector, \c false otherwise.
38 ///
39 /// Because getFormattingLangOpts() always lexes source code as
40 /// Objective-C++, C++ keywords like \c new and \c delete are
41 /// lexed as tok::kw_*, not tok::identifier, even for Objective-C.
42 ///
43 /// For Objective-C and Objective-C++, both identifiers and keywords
44 /// are valid inside @selector(...) (or a macro which
45 /// invokes @selector(...)). So, we allow treat any identifier or
46 /// keyword as a potential Objective-C selector component.
canBeObjCSelectorComponent(const FormatToken & Tok)47 static bool canBeObjCSelectorComponent(const FormatToken &Tok) {
48 return Tok.Tok.getIdentifierInfo() != nullptr;
49 }
50
51 /// With `Left` being '(', check if we're at either `[...](` or
52 /// `[...]<...>(`, where the [ opens a lambda capture list.
isLambdaParameterList(const FormatToken * Left)53 static bool isLambdaParameterList(const FormatToken *Left) {
54 // Skip <...> if present.
55 if (Left->Previous && Left->Previous->is(tok::greater) &&
56 Left->Previous->MatchingParen &&
57 Left->Previous->MatchingParen->is(TT_TemplateOpener)) {
58 Left = Left->Previous->MatchingParen;
59 }
60
61 // Check for `[...]`.
62 return Left->Previous && Left->Previous->is(tok::r_square) &&
63 Left->Previous->MatchingParen &&
64 Left->Previous->MatchingParen->is(TT_LambdaLSquare);
65 }
66
67 /// Returns \c true if the token is followed by a boolean condition, \c false
68 /// otherwise.
isKeywordWithCondition(const FormatToken & Tok)69 static bool isKeywordWithCondition(const FormatToken &Tok) {
70 return Tok.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, tok::kw_switch,
71 tok::kw_constexpr, tok::kw_catch);
72 }
73
74 /// A parser that gathers additional information about tokens.
75 ///
76 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
77 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
78 /// into template parameter lists.
79 class AnnotatingParser {
80 public:
AnnotatingParser(const FormatStyle & Style,AnnotatedLine & Line,const AdditionalKeywords & Keywords)81 AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
82 const AdditionalKeywords &Keywords)
83 : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
84 Keywords(Keywords) {
85 Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
86 resetTokenMetadata();
87 }
88
89 private:
parseAngle()90 bool parseAngle() {
91 if (!CurrentToken || !CurrentToken->Previous)
92 return false;
93 if (NonTemplateLess.count(CurrentToken->Previous))
94 return false;
95
96 const FormatToken &Previous = *CurrentToken->Previous; // The '<'.
97 if (Previous.Previous) {
98 if (Previous.Previous->Tok.isLiteral())
99 return false;
100 if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
101 (!Previous.Previous->MatchingParen ||
102 !Previous.Previous->MatchingParen->is(
103 TT_OverloadedOperatorLParen))) {
104 return false;
105 }
106 }
107
108 FormatToken *Left = CurrentToken->Previous;
109 Left->ParentBracket = Contexts.back().ContextKind;
110 ScopedContextCreator ContextCreator(*this, tok::less, 12);
111
112 // If this angle is in the context of an expression, we need to be more
113 // hesitant to detect it as opening template parameters.
114 bool InExprContext = Contexts.back().IsExpression;
115
116 Contexts.back().IsExpression = false;
117 // If there's a template keyword before the opening angle bracket, this is a
118 // template parameter, not an argument.
119 if (Left->Previous && Left->Previous->isNot(tok::kw_template))
120 Contexts.back().ContextType = Context::TemplateArgument;
121
122 if (Style.Language == FormatStyle::LK_Java &&
123 CurrentToken->is(tok::question)) {
124 next();
125 }
126
127 while (CurrentToken) {
128 if (CurrentToken->is(tok::greater)) {
129 // Try to do a better job at looking for ">>" within the condition of
130 // a statement. Conservatively insert spaces between consecutive ">"
131 // tokens to prevent splitting right bitshift operators and potentially
132 // altering program semantics. This check is overly conservative and
133 // will prevent spaces from being inserted in select nested template
134 // parameter cases, but should not alter program semantics.
135 if (CurrentToken->Next && CurrentToken->Next->is(tok::greater) &&
136 Left->ParentBracket != tok::less &&
137 (isKeywordWithCondition(*Line.First) ||
138 CurrentToken->getStartOfNonWhitespace() ==
139 CurrentToken->Next->getStartOfNonWhitespace().getLocWithOffset(
140 -1))) {
141 return false;
142 }
143 Left->MatchingParen = CurrentToken;
144 CurrentToken->MatchingParen = Left;
145 // In TT_Proto, we must distignuish between:
146 // map<key, value>
147 // msg < item: data >
148 // msg: < item: data >
149 // In TT_TextProto, map<key, value> does not occur.
150 if (Style.Language == FormatStyle::LK_TextProto ||
151 (Style.Language == FormatStyle::LK_Proto && Left->Previous &&
152 Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
153 CurrentToken->setType(TT_DictLiteral);
154 } else {
155 CurrentToken->setType(TT_TemplateCloser);
156 }
157 next();
158 return true;
159 }
160 if (CurrentToken->is(tok::question) &&
161 Style.Language == FormatStyle::LK_Java) {
162 next();
163 continue;
164 }
165 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
166 (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext &&
167 !Style.isCSharp() && Style.Language != FormatStyle::LK_Proto &&
168 Style.Language != FormatStyle::LK_TextProto)) {
169 return false;
170 }
171 // If a && or || is found and interpreted as a binary operator, this set
172 // of angles is likely part of something like "a < b && c > d". If the
173 // angles are inside an expression, the ||/&& might also be a binary
174 // operator that was misinterpreted because we are parsing template
175 // parameters.
176 // FIXME: This is getting out of hand, write a decent parser.
177 if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
178 CurrentToken->Previous->is(TT_BinaryOperator) &&
179 Contexts[Contexts.size() - 2].IsExpression &&
180 !Line.startsWith(tok::kw_template)) {
181 return false;
182 }
183 updateParameterCount(Left, CurrentToken);
184 if (Style.Language == FormatStyle::LK_Proto) {
185 if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) {
186 if (CurrentToken->is(tok::colon) ||
187 (CurrentToken->isOneOf(tok::l_brace, tok::less) &&
188 Previous->isNot(tok::colon))) {
189 Previous->setType(TT_SelectorName);
190 }
191 }
192 }
193 if (!consumeToken())
194 return false;
195 }
196 return false;
197 }
198
parseUntouchableParens()199 bool parseUntouchableParens() {
200 while (CurrentToken) {
201 CurrentToken->Finalized = true;
202 switch (CurrentToken->Tok.getKind()) {
203 case tok::l_paren:
204 next();
205 if (!parseUntouchableParens())
206 return false;
207 continue;
208 case tok::r_paren:
209 next();
210 return true;
211 default:
212 // no-op
213 break;
214 }
215 next();
216 }
217 return false;
218 }
219
parseParens(bool LookForDecls=false)220 bool parseParens(bool LookForDecls = false) {
221 if (!CurrentToken)
222 return false;
223 assert(CurrentToken->Previous && "Unknown previous token");
224 FormatToken &OpeningParen = *CurrentToken->Previous;
225 assert(OpeningParen.is(tok::l_paren));
226 FormatToken *PrevNonComment = OpeningParen.getPreviousNonComment();
227 OpeningParen.ParentBracket = Contexts.back().ContextKind;
228 ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
229
230 // FIXME: This is a bit of a hack. Do better.
231 Contexts.back().ColonIsForRangeExpr =
232 Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
233
234 if (OpeningParen.Previous &&
235 OpeningParen.Previous->is(TT_UntouchableMacroFunc)) {
236 OpeningParen.Finalized = true;
237 return parseUntouchableParens();
238 }
239
240 bool StartsObjCMethodExpr = false;
241 if (FormatToken *MaybeSel = OpeningParen.Previous) {
242 // @selector( starts a selector.
243 if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
244 MaybeSel->Previous->is(tok::at)) {
245 StartsObjCMethodExpr = true;
246 }
247 }
248
249 if (OpeningParen.is(TT_OverloadedOperatorLParen)) {
250 // Find the previous kw_operator token.
251 FormatToken *Prev = &OpeningParen;
252 while (!Prev->is(tok::kw_operator)) {
253 Prev = Prev->Previous;
254 assert(Prev && "Expect a kw_operator prior to the OperatorLParen!");
255 }
256
257 // If faced with "a.operator*(argument)" or "a->operator*(argument)",
258 // i.e. the operator is called as a member function,
259 // then the argument must be an expression.
260 bool OperatorCalledAsMemberFunction =
261 Prev->Previous && Prev->Previous->isOneOf(tok::period, tok::arrow);
262 Contexts.back().IsExpression = OperatorCalledAsMemberFunction;
263 } else if (Style.isJavaScript() &&
264 (Line.startsWith(Keywords.kw_type, tok::identifier) ||
265 Line.startsWith(tok::kw_export, Keywords.kw_type,
266 tok::identifier))) {
267 // type X = (...);
268 // export type X = (...);
269 Contexts.back().IsExpression = false;
270 } else if (OpeningParen.Previous &&
271 (OpeningParen.Previous->isOneOf(tok::kw_static_assert,
272 tok::kw_while, tok::l_paren,
273 tok::comma, TT_BinaryOperator) ||
274 OpeningParen.Previous->isIf())) {
275 // static_assert, if and while usually contain expressions.
276 Contexts.back().IsExpression = true;
277 } else if (Style.isJavaScript() && OpeningParen.Previous &&
278 (OpeningParen.Previous->is(Keywords.kw_function) ||
279 (OpeningParen.Previous->endsSequence(tok::identifier,
280 Keywords.kw_function)))) {
281 // function(...) or function f(...)
282 Contexts.back().IsExpression = false;
283 } else if (Style.isJavaScript() && OpeningParen.Previous &&
284 OpeningParen.Previous->is(TT_JsTypeColon)) {
285 // let x: (SomeType);
286 Contexts.back().IsExpression = false;
287 } else if (isLambdaParameterList(&OpeningParen)) {
288 // This is a parameter list of a lambda expression.
289 Contexts.back().IsExpression = false;
290 } else if (Line.InPPDirective &&
291 (!OpeningParen.Previous ||
292 !OpeningParen.Previous->is(tok::identifier))) {
293 Contexts.back().IsExpression = true;
294 } else if (Contexts[Contexts.size() - 2].CaretFound) {
295 // This is the parameter list of an ObjC block.
296 Contexts.back().IsExpression = false;
297 } else if (OpeningParen.Previous &&
298 OpeningParen.Previous->is(TT_ForEachMacro)) {
299 // The first argument to a foreach macro is a declaration.
300 Contexts.back().ContextType = Context::ForEachMacro;
301 Contexts.back().IsExpression = false;
302 } else if (OpeningParen.Previous && OpeningParen.Previous->MatchingParen &&
303 OpeningParen.Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
304 Contexts.back().IsExpression = false;
305 } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
306 bool IsForOrCatch =
307 OpeningParen.Previous &&
308 OpeningParen.Previous->isOneOf(tok::kw_for, tok::kw_catch);
309 Contexts.back().IsExpression = !IsForOrCatch;
310 }
311
312 // Infer the role of the l_paren based on the previous token if we haven't
313 // detected one one yet.
314 if (PrevNonComment && OpeningParen.is(TT_Unknown)) {
315 if (PrevNonComment->is(tok::kw___attribute)) {
316 OpeningParen.setType(TT_AttributeParen);
317 } else if (PrevNonComment->isOneOf(TT_TypenameMacro, tok::kw_decltype,
318 tok::kw_typeof, tok::kw__Atomic,
319 tok::kw___underlying_type)) {
320 OpeningParen.setType(TT_TypeDeclarationParen);
321 // decltype() and typeof() usually contain expressions.
322 if (PrevNonComment->isOneOf(tok::kw_decltype, tok::kw_typeof))
323 Contexts.back().IsExpression = true;
324 }
325 }
326
327 if (StartsObjCMethodExpr) {
328 Contexts.back().ColonIsObjCMethodExpr = true;
329 OpeningParen.setType(TT_ObjCMethodExpr);
330 }
331
332 // MightBeFunctionType and ProbablyFunctionType are used for
333 // function pointer and reference types as well as Objective-C
334 // block types:
335 //
336 // void (*FunctionPointer)(void);
337 // void (&FunctionReference)(void);
338 // void (&&FunctionReference)(void);
339 // void (^ObjCBlock)(void);
340 bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
341 bool ProbablyFunctionType =
342 CurrentToken->isOneOf(tok::star, tok::amp, tok::ampamp, tok::caret);
343 bool HasMultipleLines = false;
344 bool HasMultipleParametersOnALine = false;
345 bool MightBeObjCForRangeLoop =
346 OpeningParen.Previous && OpeningParen.Previous->is(tok::kw_for);
347 FormatToken *PossibleObjCForInToken = nullptr;
348 while (CurrentToken) {
349 // LookForDecls is set when "if (" has been seen. Check for
350 // 'identifier' '*' 'identifier' followed by not '=' -- this
351 // '*' has to be a binary operator but determineStarAmpUsage() will
352 // categorize it as an unary operator, so set the right type here.
353 if (LookForDecls && CurrentToken->Next) {
354 FormatToken *Prev = CurrentToken->getPreviousNonComment();
355 if (Prev) {
356 FormatToken *PrevPrev = Prev->getPreviousNonComment();
357 FormatToken *Next = CurrentToken->Next;
358 if (PrevPrev && PrevPrev->is(tok::identifier) &&
359 Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
360 CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
361 Prev->setType(TT_BinaryOperator);
362 LookForDecls = false;
363 }
364 }
365 }
366
367 if (CurrentToken->Previous->is(TT_PointerOrReference) &&
368 CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
369 tok::coloncolon)) {
370 ProbablyFunctionType = true;
371 }
372 if (CurrentToken->is(tok::comma))
373 MightBeFunctionType = false;
374 if (CurrentToken->Previous->is(TT_BinaryOperator))
375 Contexts.back().IsExpression = true;
376 if (CurrentToken->is(tok::r_paren)) {
377 if (OpeningParen.isNot(TT_CppCastLParen) && MightBeFunctionType &&
378 ProbablyFunctionType && CurrentToken->Next &&
379 (CurrentToken->Next->is(tok::l_paren) ||
380 (CurrentToken->Next->is(tok::l_square) &&
381 Line.MustBeDeclaration))) {
382 OpeningParen.setType(OpeningParen.Next->is(tok::caret)
383 ? TT_ObjCBlockLParen
384 : TT_FunctionTypeLParen);
385 }
386 OpeningParen.MatchingParen = CurrentToken;
387 CurrentToken->MatchingParen = &OpeningParen;
388
389 if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
390 OpeningParen.Previous && OpeningParen.Previous->is(tok::l_paren)) {
391 // Detect the case where macros are used to generate lambdas or
392 // function bodies, e.g.:
393 // auto my_lambda = MACRO((Type *type, int i) { .. body .. });
394 for (FormatToken *Tok = &OpeningParen; Tok != CurrentToken;
395 Tok = Tok->Next) {
396 if (Tok->is(TT_BinaryOperator) &&
397 Tok->isOneOf(tok::star, tok::amp, tok::ampamp)) {
398 Tok->setType(TT_PointerOrReference);
399 }
400 }
401 }
402
403 if (StartsObjCMethodExpr) {
404 CurrentToken->setType(TT_ObjCMethodExpr);
405 if (Contexts.back().FirstObjCSelectorName) {
406 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
407 Contexts.back().LongestObjCSelectorName;
408 }
409 }
410
411 if (OpeningParen.is(TT_AttributeParen))
412 CurrentToken->setType(TT_AttributeParen);
413 if (OpeningParen.is(TT_TypeDeclarationParen))
414 CurrentToken->setType(TT_TypeDeclarationParen);
415 if (OpeningParen.Previous &&
416 OpeningParen.Previous->is(TT_JavaAnnotation)) {
417 CurrentToken->setType(TT_JavaAnnotation);
418 }
419 if (OpeningParen.Previous &&
420 OpeningParen.Previous->is(TT_LeadingJavaAnnotation)) {
421 CurrentToken->setType(TT_LeadingJavaAnnotation);
422 }
423 if (OpeningParen.Previous &&
424 OpeningParen.Previous->is(TT_AttributeSquare)) {
425 CurrentToken->setType(TT_AttributeSquare);
426 }
427
428 if (!HasMultipleLines)
429 OpeningParen.setPackingKind(PPK_Inconclusive);
430 else if (HasMultipleParametersOnALine)
431 OpeningParen.setPackingKind(PPK_BinPacked);
432 else
433 OpeningParen.setPackingKind(PPK_OnePerLine);
434
435 next();
436 return true;
437 }
438 if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
439 return false;
440
441 if (CurrentToken->is(tok::l_brace) && OpeningParen.is(TT_ObjCBlockLParen))
442 OpeningParen.setType(TT_Unknown);
443 if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
444 !CurrentToken->Next->HasUnescapedNewline &&
445 !CurrentToken->Next->isTrailingComment()) {
446 HasMultipleParametersOnALine = true;
447 }
448 bool ProbablyFunctionTypeLParen =
449 (CurrentToken->is(tok::l_paren) && CurrentToken->Next &&
450 CurrentToken->Next->isOneOf(tok::star, tok::amp, tok::caret));
451 if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
452 CurrentToken->Previous->isSimpleTypeSpecifier()) &&
453 !(CurrentToken->is(tok::l_brace) ||
454 (CurrentToken->is(tok::l_paren) && !ProbablyFunctionTypeLParen))) {
455 Contexts.back().IsExpression = false;
456 }
457 if (CurrentToken->isOneOf(tok::semi, tok::colon)) {
458 MightBeObjCForRangeLoop = false;
459 if (PossibleObjCForInToken) {
460 PossibleObjCForInToken->setType(TT_Unknown);
461 PossibleObjCForInToken = nullptr;
462 }
463 }
464 if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) {
465 PossibleObjCForInToken = CurrentToken;
466 PossibleObjCForInToken->setType(TT_ObjCForIn);
467 }
468 // When we discover a 'new', we set CanBeExpression to 'false' in order to
469 // parse the type correctly. Reset that after a comma.
470 if (CurrentToken->is(tok::comma))
471 Contexts.back().CanBeExpression = true;
472
473 FormatToken *Tok = CurrentToken;
474 if (!consumeToken())
475 return false;
476 updateParameterCount(&OpeningParen, Tok);
477 if (CurrentToken && CurrentToken->HasUnescapedNewline)
478 HasMultipleLines = true;
479 }
480 return false;
481 }
482
isCSharpAttributeSpecifier(const FormatToken & Tok)483 bool isCSharpAttributeSpecifier(const FormatToken &Tok) {
484 if (!Style.isCSharp())
485 return false;
486
487 // `identifier[i]` is not an attribute.
488 if (Tok.Previous && Tok.Previous->is(tok::identifier))
489 return false;
490
491 // Chains of [] in `identifier[i][j][k]` are not attributes.
492 if (Tok.Previous && Tok.Previous->is(tok::r_square)) {
493 auto *MatchingParen = Tok.Previous->MatchingParen;
494 if (!MatchingParen || MatchingParen->is(TT_ArraySubscriptLSquare))
495 return false;
496 }
497
498 const FormatToken *AttrTok = Tok.Next;
499 if (!AttrTok)
500 return false;
501
502 // Just an empty declaration e.g. string [].
503 if (AttrTok->is(tok::r_square))
504 return false;
505
506 // Move along the tokens inbetween the '[' and ']' e.g. [STAThread].
507 while (AttrTok && AttrTok->isNot(tok::r_square))
508 AttrTok = AttrTok->Next;
509
510 if (!AttrTok)
511 return false;
512
513 // Allow an attribute to be the only content of a file.
514 AttrTok = AttrTok->Next;
515 if (!AttrTok)
516 return true;
517
518 // Limit this to being an access modifier that follows.
519 if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
520 tok::comment, tok::kw_class, tok::kw_static,
521 tok::l_square, Keywords.kw_internal)) {
522 return true;
523 }
524
525 // incase its a [XXX] retval func(....
526 if (AttrTok->Next &&
527 AttrTok->Next->startsSequence(tok::identifier, tok::l_paren)) {
528 return true;
529 }
530
531 return false;
532 }
533
isCpp11AttributeSpecifier(const FormatToken & Tok)534 bool isCpp11AttributeSpecifier(const FormatToken &Tok) {
535 if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square))
536 return false;
537 // The first square bracket is part of an ObjC array literal
538 if (Tok.Previous && Tok.Previous->is(tok::at))
539 return false;
540 const FormatToken *AttrTok = Tok.Next->Next;
541 if (!AttrTok)
542 return false;
543 // C++17 '[[using ns: foo, bar(baz, blech)]]'
544 // We assume nobody will name an ObjC variable 'using'.
545 if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon))
546 return true;
547 if (AttrTok->isNot(tok::identifier))
548 return false;
549 while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) {
550 // ObjC message send. We assume nobody will use : in a C++11 attribute
551 // specifier parameter, although this is technically valid:
552 // [[foo(:)]].
553 if (AttrTok->is(tok::colon) ||
554 AttrTok->startsSequence(tok::identifier, tok::identifier) ||
555 AttrTok->startsSequence(tok::r_paren, tok::identifier)) {
556 return false;
557 }
558 if (AttrTok->is(tok::ellipsis))
559 return true;
560 AttrTok = AttrTok->Next;
561 }
562 return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square);
563 }
564
parseSquare()565 bool parseSquare() {
566 if (!CurrentToken)
567 return false;
568
569 // A '[' could be an index subscript (after an identifier or after
570 // ')' or ']'), it could be the start of an Objective-C method
571 // expression, it could the start of an Objective-C array literal,
572 // or it could be a C++ attribute specifier [[foo::bar]].
573 FormatToken *Left = CurrentToken->Previous;
574 Left->ParentBracket = Contexts.back().ContextKind;
575 FormatToken *Parent = Left->getPreviousNonComment();
576
577 // Cases where '>' is followed by '['.
578 // In C++, this can happen either in array of templates (foo<int>[10])
579 // or when array is a nested template type (unique_ptr<type1<type2>[]>).
580 bool CppArrayTemplates =
581 Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) &&
582 (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
583 Contexts.back().ContextType == Context::TemplateArgument);
584
585 bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) ||
586 Contexts.back().InCpp11AttributeSpecifier;
587
588 // Treat C# Attributes [STAThread] much like C++ attributes [[...]].
589 bool IsCSharpAttributeSpecifier =
590 isCSharpAttributeSpecifier(*Left) ||
591 Contexts.back().InCSharpAttributeSpecifier;
592
593 bool InsideInlineASM = Line.startsWith(tok::kw_asm);
594 bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style);
595 bool StartsObjCMethodExpr =
596 !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates &&
597 Style.isCpp() && !IsCpp11AttributeSpecifier &&
598 !IsCSharpAttributeSpecifier && Contexts.back().CanBeExpression &&
599 Left->isNot(TT_LambdaLSquare) &&
600 !CurrentToken->isOneOf(tok::l_brace, tok::r_square) &&
601 (!Parent ||
602 Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
603 tok::kw_return, tok::kw_throw) ||
604 Parent->isUnaryOperator() ||
605 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
606 Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
607 (getBinOpPrecedence(Parent->Tok.getKind(), true, true) >
608 prec::Unknown));
609 bool ColonFound = false;
610
611 unsigned BindingIncrease = 1;
612 if (IsCppStructuredBinding) {
613 Left->setType(TT_StructuredBindingLSquare);
614 } else if (Left->is(TT_Unknown)) {
615 if (StartsObjCMethodExpr) {
616 Left->setType(TT_ObjCMethodExpr);
617 } else if (InsideInlineASM) {
618 Left->setType(TT_InlineASMSymbolicNameLSquare);
619 } else if (IsCpp11AttributeSpecifier) {
620 Left->setType(TT_AttributeSquare);
621 } else if (Style.isJavaScript() && Parent &&
622 Contexts.back().ContextKind == tok::l_brace &&
623 Parent->isOneOf(tok::l_brace, tok::comma)) {
624 Left->setType(TT_JsComputedPropertyName);
625 } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace &&
626 Parent && Parent->isOneOf(tok::l_brace, tok::comma)) {
627 Left->setType(TT_DesignatedInitializerLSquare);
628 } else if (IsCSharpAttributeSpecifier) {
629 Left->setType(TT_AttributeSquare);
630 } else if (CurrentToken->is(tok::r_square) && Parent &&
631 Parent->is(TT_TemplateCloser)) {
632 Left->setType(TT_ArraySubscriptLSquare);
633 } else if (Style.Language == FormatStyle::LK_Proto ||
634 Style.Language == FormatStyle::LK_TextProto) {
635 // Square braces in LK_Proto can either be message field attributes:
636 //
637 // optional Aaa aaa = 1 [
638 // (aaa) = aaa
639 // ];
640 //
641 // extensions 123 [
642 // (aaa) = aaa
643 // ];
644 //
645 // or text proto extensions (in options):
646 //
647 // option (Aaa.options) = {
648 // [type.type/type] {
649 // key: value
650 // }
651 // }
652 //
653 // or repeated fields (in options):
654 //
655 // option (Aaa.options) = {
656 // keys: [ 1, 2, 3 ]
657 // }
658 //
659 // In the first and the third case we want to spread the contents inside
660 // the square braces; in the second we want to keep them inline.
661 Left->setType(TT_ArrayInitializerLSquare);
662 if (!Left->endsSequence(tok::l_square, tok::numeric_constant,
663 tok::equal) &&
664 !Left->endsSequence(tok::l_square, tok::numeric_constant,
665 tok::identifier) &&
666 !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) {
667 Left->setType(TT_ProtoExtensionLSquare);
668 BindingIncrease = 10;
669 }
670 } else if (!CppArrayTemplates && Parent &&
671 Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
672 tok::comma, tok::l_paren, tok::l_square,
673 tok::question, tok::colon, tok::kw_return,
674 // Should only be relevant to JavaScript:
675 tok::kw_default)) {
676 Left->setType(TT_ArrayInitializerLSquare);
677 } else {
678 BindingIncrease = 10;
679 Left->setType(TT_ArraySubscriptLSquare);
680 }
681 }
682
683 ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
684 Contexts.back().IsExpression = true;
685 if (Style.isJavaScript() && Parent && Parent->is(TT_JsTypeColon))
686 Contexts.back().IsExpression = false;
687
688 Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
689 Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier;
690 Contexts.back().InCSharpAttributeSpecifier = IsCSharpAttributeSpecifier;
691
692 while (CurrentToken) {
693 if (CurrentToken->is(tok::r_square)) {
694 if (IsCpp11AttributeSpecifier)
695 CurrentToken->setType(TT_AttributeSquare);
696 if (IsCSharpAttributeSpecifier) {
697 CurrentToken->setType(TT_AttributeSquare);
698 } else if (((CurrentToken->Next &&
699 CurrentToken->Next->is(tok::l_paren)) ||
700 (CurrentToken->Previous &&
701 CurrentToken->Previous->Previous == Left)) &&
702 Left->is(TT_ObjCMethodExpr)) {
703 // An ObjC method call is rarely followed by an open parenthesis. It
704 // also can't be composed of just one token, unless it's a macro that
705 // will be expanded to more tokens.
706 // FIXME: Do we incorrectly label ":" with this?
707 StartsObjCMethodExpr = false;
708 Left->setType(TT_Unknown);
709 }
710 if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
711 CurrentToken->setType(TT_ObjCMethodExpr);
712 // If we haven't seen a colon yet, make sure the last identifier
713 // before the r_square is tagged as a selector name component.
714 if (!ColonFound && CurrentToken->Previous &&
715 CurrentToken->Previous->is(TT_Unknown) &&
716 canBeObjCSelectorComponent(*CurrentToken->Previous)) {
717 CurrentToken->Previous->setType(TT_SelectorName);
718 }
719 // determineStarAmpUsage() thinks that '*' '[' is allocating an
720 // array of pointers, but if '[' starts a selector then '*' is a
721 // binary operator.
722 if (Parent && Parent->is(TT_PointerOrReference))
723 Parent->overwriteFixedType(TT_BinaryOperator);
724 }
725 // An arrow after an ObjC method expression is not a lambda arrow.
726 if (CurrentToken->getType() == TT_ObjCMethodExpr &&
727 CurrentToken->Next && CurrentToken->Next->is(TT_LambdaArrow)) {
728 CurrentToken->Next->overwriteFixedType(TT_Unknown);
729 }
730 Left->MatchingParen = CurrentToken;
731 CurrentToken->MatchingParen = Left;
732 // FirstObjCSelectorName is set when a colon is found. This does
733 // not work, however, when the method has no parameters.
734 // Here, we set FirstObjCSelectorName when the end of the method call is
735 // reached, in case it was not set already.
736 if (!Contexts.back().FirstObjCSelectorName) {
737 FormatToken *Previous = CurrentToken->getPreviousNonComment();
738 if (Previous && Previous->is(TT_SelectorName)) {
739 Previous->ObjCSelectorNameParts = 1;
740 Contexts.back().FirstObjCSelectorName = Previous;
741 }
742 } else {
743 Left->ParameterCount =
744 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
745 }
746 if (Contexts.back().FirstObjCSelectorName) {
747 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
748 Contexts.back().LongestObjCSelectorName;
749 if (Left->BlockParameterCount > 1)
750 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
751 }
752 next();
753 return true;
754 }
755 if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
756 return false;
757 if (CurrentToken->is(tok::colon)) {
758 if (IsCpp11AttributeSpecifier &&
759 CurrentToken->endsSequence(tok::colon, tok::identifier,
760 tok::kw_using)) {
761 // Remember that this is a [[using ns: foo]] C++ attribute, so we
762 // don't add a space before the colon (unlike other colons).
763 CurrentToken->setType(TT_AttributeColon);
764 } else if (Left->isOneOf(TT_ArraySubscriptLSquare,
765 TT_DesignatedInitializerLSquare)) {
766 Left->setType(TT_ObjCMethodExpr);
767 StartsObjCMethodExpr = true;
768 Contexts.back().ColonIsObjCMethodExpr = true;
769 if (Parent && Parent->is(tok::r_paren)) {
770 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
771 Parent->setType(TT_CastRParen);
772 }
773 }
774 ColonFound = true;
775 }
776 if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
777 !ColonFound) {
778 Left->setType(TT_ArrayInitializerLSquare);
779 }
780 FormatToken *Tok = CurrentToken;
781 if (!consumeToken())
782 return false;
783 updateParameterCount(Left, Tok);
784 }
785 return false;
786 }
787
couldBeInStructArrayInitializer() const788 bool couldBeInStructArrayInitializer() const {
789 if (Contexts.size() < 2)
790 return false;
791 // We want to back up no more then 2 context levels i.e.
792 // . { { <-
793 const auto End = std::next(Contexts.rbegin(), 2);
794 auto Last = Contexts.rbegin();
795 unsigned Depth = 0;
796 for (; Last != End; ++Last)
797 if (Last->ContextKind == tok::l_brace)
798 ++Depth;
799 return Depth == 2 && Last->ContextKind != tok::l_brace;
800 }
801
parseBrace()802 bool parseBrace() {
803 if (!CurrentToken)
804 return true;
805
806 assert(CurrentToken->Previous);
807 FormatToken &OpeningBrace = *CurrentToken->Previous;
808 assert(OpeningBrace.is(tok::l_brace));
809 OpeningBrace.ParentBracket = Contexts.back().ContextKind;
810
811 if (Contexts.back().CaretFound)
812 OpeningBrace.overwriteFixedType(TT_ObjCBlockLBrace);
813 Contexts.back().CaretFound = false;
814
815 ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
816 Contexts.back().ColonIsDictLiteral = true;
817 if (OpeningBrace.is(BK_BracedInit))
818 Contexts.back().IsExpression = true;
819 if (Style.isJavaScript() && OpeningBrace.Previous &&
820 OpeningBrace.Previous->is(TT_JsTypeColon)) {
821 Contexts.back().IsExpression = false;
822 }
823
824 unsigned CommaCount = 0;
825 while (CurrentToken) {
826 if (CurrentToken->is(tok::r_brace)) {
827 assert(OpeningBrace.Optional == CurrentToken->Optional);
828 OpeningBrace.MatchingParen = CurrentToken;
829 CurrentToken->MatchingParen = &OpeningBrace;
830 if (Style.AlignArrayOfStructures != FormatStyle::AIAS_None) {
831 if (OpeningBrace.ParentBracket == tok::l_brace &&
832 couldBeInStructArrayInitializer() && CommaCount > 0) {
833 Contexts.back().ContextType = Context::StructArrayInitializer;
834 }
835 }
836 next();
837 return true;
838 }
839 if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
840 return false;
841 updateParameterCount(&OpeningBrace, CurrentToken);
842 if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) {
843 FormatToken *Previous = CurrentToken->getPreviousNonComment();
844 if (Previous->is(TT_JsTypeOptionalQuestion))
845 Previous = Previous->getPreviousNonComment();
846 if ((CurrentToken->is(tok::colon) &&
847 (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) ||
848 Style.Language == FormatStyle::LK_Proto ||
849 Style.Language == FormatStyle::LK_TextProto) {
850 OpeningBrace.setType(TT_DictLiteral);
851 if (Previous->Tok.getIdentifierInfo() ||
852 Previous->is(tok::string_literal)) {
853 Previous->setType(TT_SelectorName);
854 }
855 }
856 if (CurrentToken->is(tok::colon) && OpeningBrace.is(TT_Unknown))
857 OpeningBrace.setType(TT_DictLiteral);
858 else if (Style.isJavaScript())
859 OpeningBrace.overwriteFixedType(TT_DictLiteral);
860 }
861 if (CurrentToken->is(tok::comma)) {
862 if (Style.isJavaScript())
863 OpeningBrace.overwriteFixedType(TT_DictLiteral);
864 ++CommaCount;
865 }
866 if (!consumeToken())
867 return false;
868 }
869 return true;
870 }
871
updateParameterCount(FormatToken * Left,FormatToken * Current)872 void updateParameterCount(FormatToken *Left, FormatToken *Current) {
873 // For ObjC methods, the number of parameters is calculated differently as
874 // method declarations have a different structure (the parameters are not
875 // inside a bracket scope).
876 if (Current->is(tok::l_brace) && Current->is(BK_Block))
877 ++Left->BlockParameterCount;
878 if (Current->is(tok::comma)) {
879 ++Left->ParameterCount;
880 if (!Left->Role)
881 Left->Role.reset(new CommaSeparatedList(Style));
882 Left->Role->CommaFound(Current);
883 } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
884 Left->ParameterCount = 1;
885 }
886 }
887
parseConditional()888 bool parseConditional() {
889 while (CurrentToken) {
890 if (CurrentToken->is(tok::colon)) {
891 CurrentToken->setType(TT_ConditionalExpr);
892 next();
893 return true;
894 }
895 if (!consumeToken())
896 return false;
897 }
898 return false;
899 }
900
parseTemplateDeclaration()901 bool parseTemplateDeclaration() {
902 if (CurrentToken && CurrentToken->is(tok::less)) {
903 CurrentToken->setType(TT_TemplateOpener);
904 next();
905 if (!parseAngle())
906 return false;
907 if (CurrentToken)
908 CurrentToken->Previous->ClosesTemplateDeclaration = true;
909 return true;
910 }
911 return false;
912 }
913
consumeToken()914 bool consumeToken() {
915 FormatToken *Tok = CurrentToken;
916 next();
917 switch (Tok->Tok.getKind()) {
918 case tok::plus:
919 case tok::minus:
920 if (!Tok->Previous && Line.MustBeDeclaration)
921 Tok->setType(TT_ObjCMethodSpecifier);
922 break;
923 case tok::colon:
924 if (!Tok->Previous)
925 return false;
926 // Colons from ?: are handled in parseConditional().
927 if (Style.isJavaScript()) {
928 if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
929 (Contexts.size() == 1 && // switch/case labels
930 !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
931 Contexts.back().ContextKind == tok::l_paren || // function params
932 Contexts.back().ContextKind == tok::l_square || // array type
933 (!Contexts.back().IsExpression &&
934 Contexts.back().ContextKind == tok::l_brace) || // object type
935 (Contexts.size() == 1 &&
936 Line.MustBeDeclaration)) { // method/property declaration
937 Contexts.back().IsExpression = false;
938 Tok->setType(TT_JsTypeColon);
939 break;
940 }
941 } else if (Style.isCSharp()) {
942 if (Contexts.back().InCSharpAttributeSpecifier) {
943 Tok->setType(TT_AttributeColon);
944 break;
945 }
946 if (Contexts.back().ContextKind == tok::l_paren) {
947 Tok->setType(TT_CSharpNamedArgumentColon);
948 break;
949 }
950 }
951 if (Line.First->isOneOf(Keywords.kw_module, Keywords.kw_import) ||
952 Line.First->startsSequence(tok::kw_export, Keywords.kw_module) ||
953 Line.First->startsSequence(tok::kw_export, Keywords.kw_import)) {
954 Tok->setType(TT_ModulePartitionColon);
955 } else if (Contexts.back().ColonIsDictLiteral ||
956 Style.Language == FormatStyle::LK_Proto ||
957 Style.Language == FormatStyle::LK_TextProto) {
958 Tok->setType(TT_DictLiteral);
959 if (Style.Language == FormatStyle::LK_TextProto) {
960 if (FormatToken *Previous = Tok->getPreviousNonComment())
961 Previous->setType(TT_SelectorName);
962 }
963 } else if (Contexts.back().ColonIsObjCMethodExpr ||
964 Line.startsWith(TT_ObjCMethodSpecifier)) {
965 Tok->setType(TT_ObjCMethodExpr);
966 const FormatToken *BeforePrevious = Tok->Previous->Previous;
967 // Ensure we tag all identifiers in method declarations as
968 // TT_SelectorName.
969 bool UnknownIdentifierInMethodDeclaration =
970 Line.startsWith(TT_ObjCMethodSpecifier) &&
971 Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown);
972 if (!BeforePrevious ||
973 // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
974 !(BeforePrevious->is(TT_CastRParen) ||
975 (BeforePrevious->is(TT_ObjCMethodExpr) &&
976 BeforePrevious->is(tok::colon))) ||
977 BeforePrevious->is(tok::r_square) ||
978 Contexts.back().LongestObjCSelectorName == 0 ||
979 UnknownIdentifierInMethodDeclaration) {
980 Tok->Previous->setType(TT_SelectorName);
981 if (!Contexts.back().FirstObjCSelectorName) {
982 Contexts.back().FirstObjCSelectorName = Tok->Previous;
983 } else if (Tok->Previous->ColumnWidth >
984 Contexts.back().LongestObjCSelectorName) {
985 Contexts.back().LongestObjCSelectorName =
986 Tok->Previous->ColumnWidth;
987 }
988 Tok->Previous->ParameterIndex =
989 Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
990 ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
991 }
992 } else if (Contexts.back().ColonIsForRangeExpr) {
993 Tok->setType(TT_RangeBasedForLoopColon);
994 } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
995 Tok->setType(TT_BitFieldColon);
996 } else if (Contexts.size() == 1 &&
997 !Line.First->isOneOf(tok::kw_enum, tok::kw_case,
998 tok::kw_default)) {
999 FormatToken *Prev = Tok->getPreviousNonComment();
1000 if (!Prev)
1001 break;
1002 if (Prev->isOneOf(tok::r_paren, tok::kw_noexcept) ||
1003 Prev->ClosesRequiresClause) {
1004 Tok->setType(TT_CtorInitializerColon);
1005 } else if (Prev->is(tok::kw_try)) {
1006 // Member initializer list within function try block.
1007 FormatToken *PrevPrev = Prev->getPreviousNonComment();
1008 if (!PrevPrev)
1009 break;
1010 if (PrevPrev && PrevPrev->isOneOf(tok::r_paren, tok::kw_noexcept))
1011 Tok->setType(TT_CtorInitializerColon);
1012 } else {
1013 Tok->setType(TT_InheritanceColon);
1014 }
1015 } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next &&
1016 (Tok->Next->isOneOf(tok::r_paren, tok::comma) ||
1017 (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next &&
1018 Tok->Next->Next->is(tok::colon)))) {
1019 // This handles a special macro in ObjC code where selectors including
1020 // the colon are passed as macro arguments.
1021 Tok->setType(TT_ObjCMethodExpr);
1022 } else if (Contexts.back().ContextKind == tok::l_paren) {
1023 Tok->setType(TT_InlineASMColon);
1024 }
1025 break;
1026 case tok::pipe:
1027 case tok::amp:
1028 // | and & in declarations/type expressions represent union and
1029 // intersection types, respectively.
1030 if (Style.isJavaScript() && !Contexts.back().IsExpression)
1031 Tok->setType(TT_JsTypeOperator);
1032 break;
1033 case tok::kw_if:
1034 if (CurrentToken &&
1035 CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier)) {
1036 next();
1037 }
1038 LLVM_FALLTHROUGH;
1039 case tok::kw_while:
1040 if (CurrentToken && CurrentToken->is(tok::l_paren)) {
1041 next();
1042 if (!parseParens(/*LookForDecls=*/true))
1043 return false;
1044 }
1045 break;
1046 case tok::kw_for:
1047 if (Style.isJavaScript()) {
1048 // x.for and {for: ...}
1049 if ((Tok->Previous && Tok->Previous->is(tok::period)) ||
1050 (Tok->Next && Tok->Next->is(tok::colon))) {
1051 break;
1052 }
1053 // JS' for await ( ...
1054 if (CurrentToken && CurrentToken->is(Keywords.kw_await))
1055 next();
1056 }
1057 if (Style.isCpp() && CurrentToken && CurrentToken->is(tok::kw_co_await))
1058 next();
1059 Contexts.back().ColonIsForRangeExpr = true;
1060 if (!CurrentToken || CurrentToken->isNot(tok::l_paren))
1061 return false;
1062 next();
1063 if (!parseParens())
1064 return false;
1065 break;
1066 case tok::l_paren:
1067 // When faced with 'operator()()', the kw_operator handler incorrectly
1068 // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
1069 // the first two parens OverloadedOperators and the second l_paren an
1070 // OverloadedOperatorLParen.
1071 if (Tok->Previous && Tok->Previous->is(tok::r_paren) &&
1072 Tok->Previous->MatchingParen &&
1073 Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
1074 Tok->Previous->setType(TT_OverloadedOperator);
1075 Tok->Previous->MatchingParen->setType(TT_OverloadedOperator);
1076 Tok->setType(TT_OverloadedOperatorLParen);
1077 }
1078
1079 if (!parseParens())
1080 return false;
1081 if (Line.MustBeDeclaration && Contexts.size() == 1 &&
1082 !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
1083 !Tok->isOneOf(TT_TypeDeclarationParen, TT_RequiresExpressionLParen) &&
1084 (!Tok->Previous ||
1085 !Tok->Previous->isOneOf(tok::kw___attribute,
1086 TT_LeadingJavaAnnotation))) {
1087 Line.MightBeFunctionDecl = true;
1088 }
1089 break;
1090 case tok::l_square:
1091 if (!parseSquare())
1092 return false;
1093 break;
1094 case tok::l_brace:
1095 if (Style.Language == FormatStyle::LK_TextProto) {
1096 FormatToken *Previous = Tok->getPreviousNonComment();
1097 if (Previous && Previous->getType() != TT_DictLiteral)
1098 Previous->setType(TT_SelectorName);
1099 }
1100 if (!parseBrace())
1101 return false;
1102 break;
1103 case tok::less:
1104 if (parseAngle()) {
1105 Tok->setType(TT_TemplateOpener);
1106 // In TT_Proto, we must distignuish between:
1107 // map<key, value>
1108 // msg < item: data >
1109 // msg: < item: data >
1110 // In TT_TextProto, map<key, value> does not occur.
1111 if (Style.Language == FormatStyle::LK_TextProto ||
1112 (Style.Language == FormatStyle::LK_Proto && Tok->Previous &&
1113 Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
1114 Tok->setType(TT_DictLiteral);
1115 FormatToken *Previous = Tok->getPreviousNonComment();
1116 if (Previous && Previous->getType() != TT_DictLiteral)
1117 Previous->setType(TT_SelectorName);
1118 }
1119 } else {
1120 Tok->setType(TT_BinaryOperator);
1121 NonTemplateLess.insert(Tok);
1122 CurrentToken = Tok;
1123 next();
1124 }
1125 break;
1126 case tok::r_paren:
1127 case tok::r_square:
1128 return false;
1129 case tok::r_brace:
1130 // Lines can start with '}'.
1131 if (Tok->Previous)
1132 return false;
1133 break;
1134 case tok::greater:
1135 if (Style.Language != FormatStyle::LK_TextProto)
1136 Tok->setType(TT_BinaryOperator);
1137 if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser))
1138 Tok->SpacesRequiredBefore = 1;
1139 break;
1140 case tok::kw_operator:
1141 if (Style.Language == FormatStyle::LK_TextProto ||
1142 Style.Language == FormatStyle::LK_Proto) {
1143 break;
1144 }
1145 while (CurrentToken &&
1146 !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
1147 if (CurrentToken->isOneOf(tok::star, tok::amp))
1148 CurrentToken->setType(TT_PointerOrReference);
1149 consumeToken();
1150 if (CurrentToken && CurrentToken->is(tok::comma) &&
1151 CurrentToken->Previous->isNot(tok::kw_operator)) {
1152 break;
1153 }
1154 if (CurrentToken && CurrentToken->Previous->isOneOf(
1155 TT_BinaryOperator, TT_UnaryOperator, tok::comma,
1156 tok::star, tok::arrow, tok::amp, tok::ampamp)) {
1157 CurrentToken->Previous->setType(TT_OverloadedOperator);
1158 }
1159 }
1160 if (CurrentToken && CurrentToken->is(tok::l_paren))
1161 CurrentToken->setType(TT_OverloadedOperatorLParen);
1162 if (CurrentToken && CurrentToken->Previous->is(TT_BinaryOperator))
1163 CurrentToken->Previous->setType(TT_OverloadedOperator);
1164 break;
1165 case tok::question:
1166 if (Style.isJavaScript() && Tok->Next &&
1167 Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
1168 tok::r_brace)) {
1169 // Question marks before semicolons, colons, etc. indicate optional
1170 // types (fields, parameters), e.g.
1171 // function(x?: string, y?) {...}
1172 // class X { y?; }
1173 Tok->setType(TT_JsTypeOptionalQuestion);
1174 break;
1175 }
1176 // Declarations cannot be conditional expressions, this can only be part
1177 // of a type declaration.
1178 if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
1179 Style.isJavaScript()) {
1180 break;
1181 }
1182 if (Style.isCSharp()) {
1183 // `Type?)`, `Type?>`, `Type? name;` and `Type? name =` can only be
1184 // nullable types.
1185 // Line.MustBeDeclaration will be true for `Type? name;`.
1186 if ((!Contexts.back().IsExpression && Line.MustBeDeclaration) ||
1187 (Tok->Next && Tok->Next->isOneOf(tok::r_paren, tok::greater)) ||
1188 (Tok->Next && Tok->Next->is(tok::identifier) && Tok->Next->Next &&
1189 Tok->Next->Next->is(tok::equal))) {
1190 Tok->setType(TT_CSharpNullable);
1191 break;
1192 }
1193 }
1194 parseConditional();
1195 break;
1196 case tok::kw_template:
1197 parseTemplateDeclaration();
1198 break;
1199 case tok::comma:
1200 switch (Contexts.back().ContextType) {
1201 case Context::CtorInitializer:
1202 Tok->setType(TT_CtorInitializerComma);
1203 break;
1204 case Context::InheritanceList:
1205 Tok->setType(TT_InheritanceComma);
1206 break;
1207 default:
1208 if (Contexts.back().FirstStartOfName &&
1209 (Contexts.size() == 1 || startsWithInitStatement(Line))) {
1210 Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
1211 Line.IsMultiVariableDeclStmt = true;
1212 }
1213 break;
1214 }
1215 if (Contexts.back().ContextType == Context::ForEachMacro)
1216 Contexts.back().IsExpression = true;
1217 break;
1218 case tok::identifier:
1219 if (Tok->isOneOf(Keywords.kw___has_include,
1220 Keywords.kw___has_include_next)) {
1221 parseHasInclude();
1222 }
1223 if (Style.isCSharp() && Tok->is(Keywords.kw_where) && Tok->Next &&
1224 Tok->Next->isNot(tok::l_paren)) {
1225 Tok->setType(TT_CSharpGenericTypeConstraint);
1226 parseCSharpGenericTypeConstraint();
1227 }
1228 break;
1229 case tok::arrow:
1230 if (Tok->isNot(TT_LambdaArrow) && Tok->Previous &&
1231 Tok->Previous->is(tok::kw_noexcept)) {
1232 Tok->setType(TT_TrailingReturnArrow);
1233 }
1234 break;
1235 default:
1236 break;
1237 }
1238 return true;
1239 }
1240
parseCSharpGenericTypeConstraint()1241 void parseCSharpGenericTypeConstraint() {
1242 int OpenAngleBracketsCount = 0;
1243 while (CurrentToken) {
1244 if (CurrentToken->is(tok::less)) {
1245 // parseAngle is too greedy and will consume the whole line.
1246 CurrentToken->setType(TT_TemplateOpener);
1247 ++OpenAngleBracketsCount;
1248 next();
1249 } else if (CurrentToken->is(tok::greater)) {
1250 CurrentToken->setType(TT_TemplateCloser);
1251 --OpenAngleBracketsCount;
1252 next();
1253 } else if (CurrentToken->is(tok::comma) && OpenAngleBracketsCount == 0) {
1254 // We allow line breaks after GenericTypeConstraintComma's
1255 // so do not flag commas in Generics as GenericTypeConstraintComma's.
1256 CurrentToken->setType(TT_CSharpGenericTypeConstraintComma);
1257 next();
1258 } else if (CurrentToken->is(Keywords.kw_where)) {
1259 CurrentToken->setType(TT_CSharpGenericTypeConstraint);
1260 next();
1261 } else if (CurrentToken->is(tok::colon)) {
1262 CurrentToken->setType(TT_CSharpGenericTypeConstraintColon);
1263 next();
1264 } else {
1265 next();
1266 }
1267 }
1268 }
1269
parseIncludeDirective()1270 void parseIncludeDirective() {
1271 if (CurrentToken && CurrentToken->is(tok::less)) {
1272 next();
1273 while (CurrentToken) {
1274 // Mark tokens up to the trailing line comments as implicit string
1275 // literals.
1276 if (CurrentToken->isNot(tok::comment) &&
1277 !CurrentToken->TokenText.startswith("//")) {
1278 CurrentToken->setType(TT_ImplicitStringLiteral);
1279 }
1280 next();
1281 }
1282 }
1283 }
1284
parseWarningOrError()1285 void parseWarningOrError() {
1286 next();
1287 // We still want to format the whitespace left of the first token of the
1288 // warning or error.
1289 next();
1290 while (CurrentToken) {
1291 CurrentToken->setType(TT_ImplicitStringLiteral);
1292 next();
1293 }
1294 }
1295
parsePragma()1296 void parsePragma() {
1297 next(); // Consume "pragma".
1298 if (CurrentToken &&
1299 CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option,
1300 Keywords.kw_region)) {
1301 bool IsMark = CurrentToken->is(Keywords.kw_mark);
1302 next();
1303 next(); // Consume first token (so we fix leading whitespace).
1304 while (CurrentToken) {
1305 if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
1306 CurrentToken->setType(TT_ImplicitStringLiteral);
1307 next();
1308 }
1309 }
1310 }
1311
parseHasInclude()1312 void parseHasInclude() {
1313 if (!CurrentToken || !CurrentToken->is(tok::l_paren))
1314 return;
1315 next(); // '('
1316 parseIncludeDirective();
1317 next(); // ')'
1318 }
1319
parsePreprocessorDirective()1320 LineType parsePreprocessorDirective() {
1321 bool IsFirstToken = CurrentToken->IsFirst;
1322 LineType Type = LT_PreprocessorDirective;
1323 next();
1324 if (!CurrentToken)
1325 return Type;
1326
1327 if (Style.isJavaScript() && IsFirstToken) {
1328 // JavaScript files can contain shebang lines of the form:
1329 // #!/usr/bin/env node
1330 // Treat these like C++ #include directives.
1331 while (CurrentToken) {
1332 // Tokens cannot be comments here.
1333 CurrentToken->setType(TT_ImplicitStringLiteral);
1334 next();
1335 }
1336 return LT_ImportStatement;
1337 }
1338
1339 if (CurrentToken->is(tok::numeric_constant)) {
1340 CurrentToken->SpacesRequiredBefore = 1;
1341 return Type;
1342 }
1343 // Hashes in the middle of a line can lead to any strange token
1344 // sequence.
1345 if (!CurrentToken->Tok.getIdentifierInfo())
1346 return Type;
1347 // In Verilog macro expansions start with a backtick just like preprocessor
1348 // directives. Thus we stop if the word is not a preprocessor directive.
1349 if (Style.isVerilog() && !Keywords.isVerilogPPDirective(*CurrentToken))
1350 return LT_Invalid;
1351 switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
1352 case tok::pp_include:
1353 case tok::pp_include_next:
1354 case tok::pp_import:
1355 next();
1356 parseIncludeDirective();
1357 Type = LT_ImportStatement;
1358 break;
1359 case tok::pp_error:
1360 case tok::pp_warning:
1361 parseWarningOrError();
1362 break;
1363 case tok::pp_pragma:
1364 parsePragma();
1365 break;
1366 case tok::pp_if:
1367 case tok::pp_elif:
1368 Contexts.back().IsExpression = true;
1369 next();
1370 parseLine();
1371 break;
1372 default:
1373 break;
1374 }
1375 while (CurrentToken) {
1376 FormatToken *Tok = CurrentToken;
1377 next();
1378 if (Tok->is(tok::l_paren)) {
1379 parseParens();
1380 } else if (Tok->isOneOf(Keywords.kw___has_include,
1381 Keywords.kw___has_include_next)) {
1382 parseHasInclude();
1383 }
1384 }
1385 return Type;
1386 }
1387
1388 public:
parseLine()1389 LineType parseLine() {
1390 if (!CurrentToken)
1391 return LT_Invalid;
1392 NonTemplateLess.clear();
1393 if (CurrentToken->is(tok::hash)) {
1394 // We were not yet allowed to use C++17 optional when this was being
1395 // written. So we used LT_Invalid to mark that the line is not a
1396 // preprocessor directive.
1397 auto Type = parsePreprocessorDirective();
1398 if (Type != LT_Invalid)
1399 return Type;
1400 }
1401
1402 // Directly allow to 'import <string-literal>' to support protocol buffer
1403 // definitions (github.com/google/protobuf) or missing "#" (either way we
1404 // should not break the line).
1405 IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
1406 if ((Style.Language == FormatStyle::LK_Java &&
1407 CurrentToken->is(Keywords.kw_package)) ||
1408 (Info && Info->getPPKeywordID() == tok::pp_import &&
1409 CurrentToken->Next &&
1410 CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
1411 tok::kw_static))) {
1412 next();
1413 parseIncludeDirective();
1414 return LT_ImportStatement;
1415 }
1416
1417 // If this line starts and ends in '<' and '>', respectively, it is likely
1418 // part of "#define <a/b.h>".
1419 if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
1420 parseIncludeDirective();
1421 return LT_ImportStatement;
1422 }
1423
1424 // In .proto files, top-level options and package statements are very
1425 // similar to import statements and should not be line-wrapped.
1426 if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
1427 CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) {
1428 next();
1429 if (CurrentToken && CurrentToken->is(tok::identifier)) {
1430 while (CurrentToken)
1431 next();
1432 return LT_ImportStatement;
1433 }
1434 }
1435
1436 bool KeywordVirtualFound = false;
1437 bool ImportStatement = false;
1438
1439 // import {...} from '...';
1440 if (Style.isJavaScript() && CurrentToken->is(Keywords.kw_import))
1441 ImportStatement = true;
1442
1443 while (CurrentToken) {
1444 if (CurrentToken->is(tok::kw_virtual))
1445 KeywordVirtualFound = true;
1446 if (Style.isJavaScript()) {
1447 // export {...} from '...';
1448 // An export followed by "from 'some string';" is a re-export from
1449 // another module identified by a URI and is treated as a
1450 // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
1451 // Just "export {...};" or "export class ..." should not be treated as
1452 // an import in this sense.
1453 if (Line.First->is(tok::kw_export) &&
1454 CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
1455 CurrentToken->Next->isStringLiteral()) {
1456 ImportStatement = true;
1457 }
1458 if (isClosureImportStatement(*CurrentToken))
1459 ImportStatement = true;
1460 }
1461 if (!consumeToken())
1462 return LT_Invalid;
1463 }
1464 if (KeywordVirtualFound)
1465 return LT_VirtualFunctionDecl;
1466 if (ImportStatement)
1467 return LT_ImportStatement;
1468
1469 if (Line.startsWith(TT_ObjCMethodSpecifier)) {
1470 if (Contexts.back().FirstObjCSelectorName) {
1471 Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
1472 Contexts.back().LongestObjCSelectorName;
1473 }
1474 return LT_ObjCMethodDecl;
1475 }
1476
1477 for (const auto &ctx : Contexts)
1478 if (ctx.ContextType == Context::StructArrayInitializer)
1479 return LT_ArrayOfStructInitializer;
1480
1481 return LT_Other;
1482 }
1483
1484 private:
isClosureImportStatement(const FormatToken & Tok)1485 bool isClosureImportStatement(const FormatToken &Tok) {
1486 // FIXME: Closure-library specific stuff should not be hard-coded but be
1487 // configurable.
1488 return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
1489 Tok.Next->Next &&
1490 (Tok.Next->Next->TokenText == "module" ||
1491 Tok.Next->Next->TokenText == "provide" ||
1492 Tok.Next->Next->TokenText == "require" ||
1493 Tok.Next->Next->TokenText == "requireType" ||
1494 Tok.Next->Next->TokenText == "forwardDeclare") &&
1495 Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
1496 }
1497
resetTokenMetadata()1498 void resetTokenMetadata() {
1499 if (!CurrentToken)
1500 return;
1501
1502 // Reset token type in case we have already looked at it and then
1503 // recovered from an error (e.g. failure to find the matching >).
1504 if (!CurrentToken->isTypeFinalized() &&
1505 !CurrentToken->isOneOf(
1506 TT_LambdaLSquare, TT_LambdaLBrace, TT_AttributeMacro, TT_IfMacro,
1507 TT_ForEachMacro, TT_TypenameMacro, TT_FunctionLBrace,
1508 TT_ImplicitStringLiteral, TT_InlineASMBrace, TT_FatArrow,
1509 TT_LambdaArrow, TT_NamespaceMacro, TT_OverloadedOperator,
1510 TT_RegexLiteral, TT_TemplateString, TT_ObjCStringLiteral,
1511 TT_UntouchableMacroFunc, TT_StatementAttributeLikeMacro,
1512 TT_FunctionLikeOrFreestandingMacro, TT_ClassLBrace, TT_EnumLBrace,
1513 TT_RecordLBrace, TT_StructLBrace, TT_UnionLBrace, TT_RequiresClause,
1514 TT_RequiresClauseInARequiresExpression, TT_RequiresExpression,
1515 TT_RequiresExpressionLParen, TT_RequiresExpressionLBrace,
1516 TT_CompoundRequirementLBrace, TT_BracedListLBrace)) {
1517 CurrentToken->setType(TT_Unknown);
1518 }
1519 CurrentToken->Role.reset();
1520 CurrentToken->MatchingParen = nullptr;
1521 CurrentToken->FakeLParens.clear();
1522 CurrentToken->FakeRParens = 0;
1523 }
1524
next()1525 void next() {
1526 if (!CurrentToken)
1527 return;
1528
1529 CurrentToken->NestingLevel = Contexts.size() - 1;
1530 CurrentToken->BindingStrength = Contexts.back().BindingStrength;
1531 modifyContext(*CurrentToken);
1532 determineTokenType(*CurrentToken);
1533 CurrentToken = CurrentToken->Next;
1534
1535 resetTokenMetadata();
1536 }
1537
1538 /// A struct to hold information valid in a specific context, e.g.
1539 /// a pair of parenthesis.
1540 struct Context {
Contextclang::format::__anon095ba8b90111::AnnotatingParser::Context1541 Context(tok::TokenKind ContextKind, unsigned BindingStrength,
1542 bool IsExpression)
1543 : ContextKind(ContextKind), BindingStrength(BindingStrength),
1544 IsExpression(IsExpression) {}
1545
1546 tok::TokenKind ContextKind;
1547 unsigned BindingStrength;
1548 bool IsExpression;
1549 unsigned LongestObjCSelectorName = 0;
1550 bool ColonIsForRangeExpr = false;
1551 bool ColonIsDictLiteral = false;
1552 bool ColonIsObjCMethodExpr = false;
1553 FormatToken *FirstObjCSelectorName = nullptr;
1554 FormatToken *FirstStartOfName = nullptr;
1555 bool CanBeExpression = true;
1556 bool CaretFound = false;
1557 bool InCpp11AttributeSpecifier = false;
1558 bool InCSharpAttributeSpecifier = false;
1559 enum {
1560 Unknown,
1561 // Like the part after `:` in a constructor.
1562 // Context(...) : IsExpression(IsExpression)
1563 CtorInitializer,
1564 // Like in the parentheses in a foreach.
1565 ForEachMacro,
1566 // Like the inheritance list in a class declaration.
1567 // class Input : public IO
1568 InheritanceList,
1569 // Like in the braced list.
1570 // int x[] = {};
1571 StructArrayInitializer,
1572 // Like in `static_cast<int>`.
1573 TemplateArgument,
1574 } ContextType = Unknown;
1575 };
1576
1577 /// Puts a new \c Context onto the stack \c Contexts for the lifetime
1578 /// of each instance.
1579 struct ScopedContextCreator {
1580 AnnotatingParser &P;
1581
ScopedContextCreatorclang::format::__anon095ba8b90111::AnnotatingParser::ScopedContextCreator1582 ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
1583 unsigned Increase)
1584 : P(P) {
1585 P.Contexts.push_back(Context(ContextKind,
1586 P.Contexts.back().BindingStrength + Increase,
1587 P.Contexts.back().IsExpression));
1588 }
1589
~ScopedContextCreatorclang::format::__anon095ba8b90111::AnnotatingParser::ScopedContextCreator1590 ~ScopedContextCreator() {
1591 if (P.Style.AlignArrayOfStructures != FormatStyle::AIAS_None) {
1592 if (P.Contexts.back().ContextType == Context::StructArrayInitializer) {
1593 P.Contexts.pop_back();
1594 P.Contexts.back().ContextType = Context::StructArrayInitializer;
1595 return;
1596 }
1597 }
1598 P.Contexts.pop_back();
1599 }
1600 };
1601
modifyContext(const FormatToken & Current)1602 void modifyContext(const FormatToken &Current) {
1603 auto AssignmentStartsExpression = [&]() {
1604 if (Current.getPrecedence() != prec::Assignment)
1605 return false;
1606
1607 if (Line.First->isOneOf(tok::kw_using, tok::kw_return))
1608 return false;
1609 if (Line.First->is(tok::kw_template)) {
1610 assert(Current.Previous);
1611 if (Current.Previous->is(tok::kw_operator)) {
1612 // `template ... operator=` cannot be an expression.
1613 return false;
1614 }
1615
1616 // `template` keyword can start a variable template.
1617 const FormatToken *Tok = Line.First->getNextNonComment();
1618 assert(Tok); // Current token is on the same line.
1619 if (Tok->isNot(TT_TemplateOpener)) {
1620 // Explicit template instantiations do not have `<>`.
1621 return false;
1622 }
1623
1624 Tok = Tok->MatchingParen;
1625 if (!Tok)
1626 return false;
1627 Tok = Tok->getNextNonComment();
1628 if (!Tok)
1629 return false;
1630
1631 if (Tok->isOneOf(tok::kw_class, tok::kw_enum, tok::kw_concept,
1632 tok::kw_struct, tok::kw_using)) {
1633 return false;
1634 }
1635
1636 return true;
1637 }
1638
1639 // Type aliases use `type X = ...;` in TypeScript and can be exported
1640 // using `export type ...`.
1641 if (Style.isJavaScript() &&
1642 (Line.startsWith(Keywords.kw_type, tok::identifier) ||
1643 Line.startsWith(tok::kw_export, Keywords.kw_type,
1644 tok::identifier))) {
1645 return false;
1646 }
1647
1648 return !Current.Previous || Current.Previous->isNot(tok::kw_operator);
1649 };
1650
1651 if (AssignmentStartsExpression()) {
1652 Contexts.back().IsExpression = true;
1653 if (!Line.startsWith(TT_UnaryOperator)) {
1654 for (FormatToken *Previous = Current.Previous;
1655 Previous && Previous->Previous &&
1656 !Previous->Previous->isOneOf(tok::comma, tok::semi);
1657 Previous = Previous->Previous) {
1658 if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
1659 Previous = Previous->MatchingParen;
1660 if (!Previous)
1661 break;
1662 }
1663 if (Previous->opensScope())
1664 break;
1665 if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
1666 Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
1667 Previous->Previous && Previous->Previous->isNot(tok::equal)) {
1668 Previous->setType(TT_PointerOrReference);
1669 }
1670 }
1671 }
1672 } else if (Current.is(tok::lessless) &&
1673 (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1674 Contexts.back().IsExpression = true;
1675 } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1676 Contexts.back().IsExpression = true;
1677 } else if (Current.is(TT_TrailingReturnArrow)) {
1678 Contexts.back().IsExpression = false;
1679 } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1680 Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1681 } else if (Current.Previous &&
1682 Current.Previous->is(TT_CtorInitializerColon)) {
1683 Contexts.back().IsExpression = true;
1684 Contexts.back().ContextType = Context::CtorInitializer;
1685 } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) {
1686 Contexts.back().ContextType = Context::InheritanceList;
1687 } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1688 for (FormatToken *Previous = Current.Previous;
1689 Previous && Previous->isOneOf(tok::star, tok::amp);
1690 Previous = Previous->Previous) {
1691 Previous->setType(TT_PointerOrReference);
1692 }
1693 if (Line.MustBeDeclaration &&
1694 Contexts.front().ContextType != Context::CtorInitializer) {
1695 Contexts.back().IsExpression = false;
1696 }
1697 } else if (Current.is(tok::kw_new)) {
1698 Contexts.back().CanBeExpression = false;
1699 } else if (Current.is(tok::semi) ||
1700 (Current.is(tok::exclaim) && Current.Previous &&
1701 !Current.Previous->is(tok::kw_operator))) {
1702 // This should be the condition or increment in a for-loop.
1703 // But not operator !() (can't use TT_OverloadedOperator here as its not
1704 // been annotated yet).
1705 Contexts.back().IsExpression = true;
1706 }
1707 }
1708
untilMatchingParen(FormatToken * Current)1709 static FormatToken *untilMatchingParen(FormatToken *Current) {
1710 // Used when `MatchingParen` is not yet established.
1711 int ParenLevel = 0;
1712 while (Current) {
1713 if (Current->is(tok::l_paren))
1714 ++ParenLevel;
1715 if (Current->is(tok::r_paren))
1716 --ParenLevel;
1717 if (ParenLevel < 1)
1718 break;
1719 Current = Current->Next;
1720 }
1721 return Current;
1722 }
1723
isDeductionGuide(FormatToken & Current)1724 static bool isDeductionGuide(FormatToken &Current) {
1725 // Look for a deduction guide template<T> A(...) -> A<...>;
1726 if (Current.Previous && Current.Previous->is(tok::r_paren) &&
1727 Current.startsSequence(tok::arrow, tok::identifier, tok::less)) {
1728 // Find the TemplateCloser.
1729 FormatToken *TemplateCloser = Current.Next->Next;
1730 int NestingLevel = 0;
1731 while (TemplateCloser) {
1732 // Skip over an expressions in parens A<(3 < 2)>;
1733 if (TemplateCloser->is(tok::l_paren)) {
1734 // No Matching Paren yet so skip to matching paren
1735 TemplateCloser = untilMatchingParen(TemplateCloser);
1736 if (!TemplateCloser)
1737 break;
1738 }
1739 if (TemplateCloser->is(tok::less))
1740 ++NestingLevel;
1741 if (TemplateCloser->is(tok::greater))
1742 --NestingLevel;
1743 if (NestingLevel < 1)
1744 break;
1745 TemplateCloser = TemplateCloser->Next;
1746 }
1747 // Assuming we have found the end of the template ensure its followed
1748 // with a semi-colon.
1749 if (TemplateCloser && TemplateCloser->Next &&
1750 TemplateCloser->Next->is(tok::semi) &&
1751 Current.Previous->MatchingParen) {
1752 // Determine if the identifier `A` prior to the A<..>; is the same as
1753 // prior to the A(..)
1754 FormatToken *LeadingIdentifier =
1755 Current.Previous->MatchingParen->Previous;
1756
1757 // Differentiate a deduction guide by seeing the
1758 // > of the template prior to the leading identifier.
1759 if (LeadingIdentifier) {
1760 FormatToken *PriorLeadingIdentifier = LeadingIdentifier->Previous;
1761 // Skip back past explicit decoration
1762 if (PriorLeadingIdentifier &&
1763 PriorLeadingIdentifier->is(tok::kw_explicit)) {
1764 PriorLeadingIdentifier = PriorLeadingIdentifier->Previous;
1765 }
1766
1767 return PriorLeadingIdentifier &&
1768 (PriorLeadingIdentifier->is(TT_TemplateCloser) ||
1769 PriorLeadingIdentifier->ClosesRequiresClause) &&
1770 LeadingIdentifier->TokenText == Current.Next->TokenText;
1771 }
1772 }
1773 }
1774 return false;
1775 }
1776
determineTokenType(FormatToken & Current)1777 void determineTokenType(FormatToken &Current) {
1778 if (!Current.is(TT_Unknown)) {
1779 // The token type is already known.
1780 return;
1781 }
1782
1783 if ((Style.isJavaScript() || Style.isCSharp()) &&
1784 Current.is(tok::exclaim)) {
1785 if (Current.Previous) {
1786 bool IsIdentifier =
1787 Style.isJavaScript()
1788 ? Keywords.IsJavaScriptIdentifier(
1789 *Current.Previous, /* AcceptIdentifierName= */ true)
1790 : Current.Previous->is(tok::identifier);
1791 if (IsIdentifier ||
1792 Current.Previous->isOneOf(
1793 tok::kw_default, tok::kw_namespace, tok::r_paren, tok::r_square,
1794 tok::r_brace, tok::kw_false, tok::kw_true, Keywords.kw_type,
1795 Keywords.kw_get, Keywords.kw_init, Keywords.kw_set) ||
1796 Current.Previous->Tok.isLiteral()) {
1797 Current.setType(TT_NonNullAssertion);
1798 return;
1799 }
1800 }
1801 if (Current.Next &&
1802 Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1803 Current.setType(TT_NonNullAssertion);
1804 return;
1805 }
1806 }
1807
1808 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1809 // function declaration have been found. In this case, 'Current' is a
1810 // trailing token of this declaration and thus cannot be a name.
1811 if (Current.is(Keywords.kw_instanceof)) {
1812 Current.setType(TT_BinaryOperator);
1813 } else if (isStartOfName(Current) &&
1814 (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1815 Contexts.back().FirstStartOfName = &Current;
1816 Current.setType(TT_StartOfName);
1817 } else if (Current.is(tok::semi)) {
1818 // Reset FirstStartOfName after finding a semicolon so that a for loop
1819 // with multiple increment statements is not confused with a for loop
1820 // having multiple variable declarations.
1821 Contexts.back().FirstStartOfName = nullptr;
1822 } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1823 AutoFound = true;
1824 } else if (Current.is(tok::arrow) &&
1825 Style.Language == FormatStyle::LK_Java) {
1826 Current.setType(TT_LambdaArrow);
1827 } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1828 Current.NestingLevel == 0 &&
1829 !Current.Previous->isOneOf(tok::kw_operator, tok::identifier)) {
1830 // not auto operator->() -> xxx;
1831 Current.setType(TT_TrailingReturnArrow);
1832 } else if (Current.is(tok::arrow) && Current.Previous &&
1833 Current.Previous->is(tok::r_brace)) {
1834 // Concept implicit conversion constraint needs to be treated like
1835 // a trailing return type ... } -> <type>.
1836 Current.setType(TT_TrailingReturnArrow);
1837 } else if (isDeductionGuide(Current)) {
1838 // Deduction guides trailing arrow " A(...) -> A<T>;".
1839 Current.setType(TT_TrailingReturnArrow);
1840 } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1841 Current.setType(determineStarAmpUsage(
1842 Current,
1843 Contexts.back().CanBeExpression && Contexts.back().IsExpression,
1844 Contexts.back().ContextType == Context::TemplateArgument));
1845 } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1846 Current.setType(determinePlusMinusCaretUsage(Current));
1847 if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1848 Contexts.back().CaretFound = true;
1849 } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1850 Current.setType(determineIncrementUsage(Current));
1851 } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1852 Current.setType(TT_UnaryOperator);
1853 } else if (Current.is(tok::question)) {
1854 if (Style.isJavaScript() && Line.MustBeDeclaration &&
1855 !Contexts.back().IsExpression) {
1856 // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1857 // on the interface, not a ternary expression.
1858 Current.setType(TT_JsTypeOptionalQuestion);
1859 } else {
1860 Current.setType(TT_ConditionalExpr);
1861 }
1862 } else if (Current.isBinaryOperator() &&
1863 (!Current.Previous || Current.Previous->isNot(tok::l_square)) &&
1864 (!Current.is(tok::greater) &&
1865 Style.Language != FormatStyle::LK_TextProto)) {
1866 Current.setType(TT_BinaryOperator);
1867 } else if (Current.is(tok::comment)) {
1868 if (Current.TokenText.startswith("/*")) {
1869 if (Current.TokenText.endswith("*/")) {
1870 Current.setType(TT_BlockComment);
1871 } else {
1872 // The lexer has for some reason determined a comment here. But we
1873 // cannot really handle it, if it isn't properly terminated.
1874 Current.Tok.setKind(tok::unknown);
1875 }
1876 } else {
1877 Current.setType(TT_LineComment);
1878 }
1879 } else if (Current.is(tok::l_paren)) {
1880 if (lParenStartsCppCast(Current))
1881 Current.setType(TT_CppCastLParen);
1882 } else if (Current.is(tok::r_paren)) {
1883 if (rParenEndsCast(Current))
1884 Current.setType(TT_CastRParen);
1885 if (Current.MatchingParen && Current.Next &&
1886 !Current.Next->isBinaryOperator() &&
1887 !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1888 tok::comma, tok::period, tok::arrow,
1889 tok::coloncolon)) {
1890 if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1891 // Make sure this isn't the return type of an Obj-C block declaration
1892 if (AfterParen->isNot(tok::caret)) {
1893 if (FormatToken *BeforeParen = Current.MatchingParen->Previous) {
1894 if (BeforeParen->is(tok::identifier) &&
1895 !BeforeParen->is(TT_TypenameMacro) &&
1896 BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1897 (!BeforeParen->Previous ||
1898 BeforeParen->Previous->ClosesTemplateDeclaration)) {
1899 Current.setType(TT_FunctionAnnotationRParen);
1900 }
1901 }
1902 }
1903 }
1904 }
1905 } else if (Current.is(tok::at) && Current.Next && !Style.isJavaScript() &&
1906 Style.Language != FormatStyle::LK_Java) {
1907 // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1908 // marks declarations and properties that need special formatting.
1909 switch (Current.Next->Tok.getObjCKeywordID()) {
1910 case tok::objc_interface:
1911 case tok::objc_implementation:
1912 case tok::objc_protocol:
1913 Current.setType(TT_ObjCDecl);
1914 break;
1915 case tok::objc_property:
1916 Current.setType(TT_ObjCProperty);
1917 break;
1918 default:
1919 break;
1920 }
1921 } else if (Current.is(tok::period)) {
1922 FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1923 if (PreviousNoComment &&
1924 PreviousNoComment->isOneOf(tok::comma, tok::l_brace)) {
1925 Current.setType(TT_DesignatedInitializerPeriod);
1926 } else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1927 Current.Previous->isOneOf(TT_JavaAnnotation,
1928 TT_LeadingJavaAnnotation)) {
1929 Current.setType(Current.Previous->getType());
1930 }
1931 } else if (canBeObjCSelectorComponent(Current) &&
1932 // FIXME(bug 36976): ObjC return types shouldn't use
1933 // TT_CastRParen.
1934 Current.Previous && Current.Previous->is(TT_CastRParen) &&
1935 Current.Previous->MatchingParen &&
1936 Current.Previous->MatchingParen->Previous &&
1937 Current.Previous->MatchingParen->Previous->is(
1938 TT_ObjCMethodSpecifier)) {
1939 // This is the first part of an Objective-C selector name. (If there's no
1940 // colon after this, this is the only place which annotates the identifier
1941 // as a selector.)
1942 Current.setType(TT_SelectorName);
1943 } else if (Current.isOneOf(tok::identifier, tok::kw_const, tok::kw_noexcept,
1944 tok::kw_requires) &&
1945 Current.Previous &&
1946 !Current.Previous->isOneOf(tok::equal, tok::at) &&
1947 Line.MightBeFunctionDecl && Contexts.size() == 1) {
1948 // Line.MightBeFunctionDecl can only be true after the parentheses of a
1949 // function declaration have been found.
1950 Current.setType(TT_TrailingAnnotation);
1951 } else if ((Style.Language == FormatStyle::LK_Java ||
1952 Style.isJavaScript()) &&
1953 Current.Previous) {
1954 if (Current.Previous->is(tok::at) &&
1955 Current.isNot(Keywords.kw_interface)) {
1956 const FormatToken &AtToken = *Current.Previous;
1957 const FormatToken *Previous = AtToken.getPreviousNonComment();
1958 if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1959 Current.setType(TT_LeadingJavaAnnotation);
1960 else
1961 Current.setType(TT_JavaAnnotation);
1962 } else if (Current.Previous->is(tok::period) &&
1963 Current.Previous->isOneOf(TT_JavaAnnotation,
1964 TT_LeadingJavaAnnotation)) {
1965 Current.setType(Current.Previous->getType());
1966 }
1967 }
1968 }
1969
1970 /// Take a guess at whether \p Tok starts a name of a function or
1971 /// variable declaration.
1972 ///
1973 /// This is a heuristic based on whether \p Tok is an identifier following
1974 /// something that is likely a type.
isStartOfName(const FormatToken & Tok)1975 bool isStartOfName(const FormatToken &Tok) {
1976 if (Tok.isNot(tok::identifier) || !Tok.Previous)
1977 return false;
1978
1979 if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1980 Keywords.kw_as)) {
1981 return false;
1982 }
1983 if (Style.isJavaScript() && Tok.Previous->is(Keywords.kw_in))
1984 return false;
1985
1986 // Skip "const" as it does not have an influence on whether this is a name.
1987 FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1988
1989 // For javascript const can be like "let" or "var"
1990 if (!Style.isJavaScript())
1991 while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1992 PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1993
1994 if (!PreviousNotConst)
1995 return false;
1996
1997 if (PreviousNotConst->ClosesRequiresClause)
1998 return false;
1999
2000 bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
2001 PreviousNotConst->Previous &&
2002 PreviousNotConst->Previous->is(tok::hash);
2003
2004 if (PreviousNotConst->is(TT_TemplateCloser)) {
2005 return PreviousNotConst && PreviousNotConst->MatchingParen &&
2006 PreviousNotConst->MatchingParen->Previous &&
2007 PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
2008 PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
2009 }
2010
2011 if (PreviousNotConst->is(tok::r_paren) &&
2012 PreviousNotConst->is(TT_TypeDeclarationParen)) {
2013 return true;
2014 }
2015
2016 // If is a preprocess keyword like #define.
2017 if (IsPPKeyword)
2018 return false;
2019
2020 // int a or auto a.
2021 if (PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto))
2022 return true;
2023
2024 // *a or &a or &&a.
2025 if (PreviousNotConst->is(TT_PointerOrReference))
2026 return true;
2027
2028 // MyClass a;
2029 if (PreviousNotConst->isSimpleTypeSpecifier())
2030 return true;
2031
2032 // const a = in JavaScript.
2033 return Style.isJavaScript() && PreviousNotConst->is(tok::kw_const);
2034 }
2035
2036 /// Determine whether '(' is starting a C++ cast.
lParenStartsCppCast(const FormatToken & Tok)2037 bool lParenStartsCppCast(const FormatToken &Tok) {
2038 // C-style casts are only used in C++.
2039 if (!Style.isCpp())
2040 return false;
2041
2042 FormatToken *LeftOfParens = Tok.getPreviousNonComment();
2043 if (LeftOfParens && LeftOfParens->is(TT_TemplateCloser) &&
2044 LeftOfParens->MatchingParen) {
2045 auto *Prev = LeftOfParens->MatchingParen->getPreviousNonComment();
2046 if (Prev &&
2047 Prev->isOneOf(tok::kw_const_cast, tok::kw_dynamic_cast,
2048 tok::kw_reinterpret_cast, tok::kw_static_cast)) {
2049 // FIXME: Maybe we should handle identifiers ending with "_cast",
2050 // e.g. any_cast?
2051 return true;
2052 }
2053 }
2054 return false;
2055 }
2056
2057 /// Determine whether ')' is ending a cast.
rParenEndsCast(const FormatToken & Tok)2058 bool rParenEndsCast(const FormatToken &Tok) {
2059 // C-style casts are only used in C++, C# and Java.
2060 if (!Style.isCSharp() && !Style.isCpp() &&
2061 Style.Language != FormatStyle::LK_Java) {
2062 return false;
2063 }
2064
2065 // Empty parens aren't casts and there are no casts at the end of the line.
2066 if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
2067 return false;
2068
2069 FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
2070 if (LeftOfParens) {
2071 // If there is a closing parenthesis left of the current
2072 // parentheses, look past it as these might be chained casts.
2073 if (LeftOfParens->is(tok::r_paren) &&
2074 LeftOfParens->isNot(TT_CastRParen)) {
2075 if (!LeftOfParens->MatchingParen ||
2076 !LeftOfParens->MatchingParen->Previous) {
2077 return false;
2078 }
2079 LeftOfParens = LeftOfParens->MatchingParen->Previous;
2080 }
2081
2082 if (LeftOfParens->is(tok::r_square)) {
2083 // delete[] (void *)ptr;
2084 auto MayBeArrayDelete = [](FormatToken *Tok) -> FormatToken * {
2085 if (Tok->isNot(tok::r_square))
2086 return nullptr;
2087
2088 Tok = Tok->getPreviousNonComment();
2089 if (!Tok || Tok->isNot(tok::l_square))
2090 return nullptr;
2091
2092 Tok = Tok->getPreviousNonComment();
2093 if (!Tok || Tok->isNot(tok::kw_delete))
2094 return nullptr;
2095 return Tok;
2096 };
2097 if (FormatToken *MaybeDelete = MayBeArrayDelete(LeftOfParens))
2098 LeftOfParens = MaybeDelete;
2099 }
2100
2101 // The Condition directly below this one will see the operator arguments
2102 // as a (void *foo) cast.
2103 // void operator delete(void *foo) ATTRIB;
2104 if (LeftOfParens->Tok.getIdentifierInfo() && LeftOfParens->Previous &&
2105 LeftOfParens->Previous->is(tok::kw_operator)) {
2106 return false;
2107 }
2108
2109 // If there is an identifier (or with a few exceptions a keyword) right
2110 // before the parentheses, this is unlikely to be a cast.
2111 if (LeftOfParens->Tok.getIdentifierInfo() &&
2112 !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
2113 tok::kw_delete)) {
2114 return false;
2115 }
2116
2117 // Certain other tokens right before the parentheses are also signals that
2118 // this cannot be a cast.
2119 if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
2120 TT_TemplateCloser, tok::ellipsis)) {
2121 return false;
2122 }
2123 }
2124
2125 if (Tok.Next->is(tok::question))
2126 return false;
2127
2128 // `foreach((A a, B b) in someList)` should not be seen as a cast.
2129 if (Tok.Next->is(Keywords.kw_in) && Style.isCSharp())
2130 return false;
2131
2132 // Functions which end with decorations like volatile, noexcept are unlikely
2133 // to be casts.
2134 if (Tok.Next->isOneOf(tok::kw_noexcept, tok::kw_volatile, tok::kw_const,
2135 tok::kw_requires, tok::kw_throw, tok::arrow,
2136 Keywords.kw_override, Keywords.kw_final) ||
2137 isCpp11AttributeSpecifier(*Tok.Next)) {
2138 return false;
2139 }
2140
2141 // As Java has no function types, a "(" after the ")" likely means that this
2142 // is a cast.
2143 if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
2144 return true;
2145
2146 // If a (non-string) literal follows, this is likely a cast.
2147 if (Tok.Next->isNot(tok::string_literal) &&
2148 (Tok.Next->Tok.isLiteral() ||
2149 Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof))) {
2150 return true;
2151 }
2152
2153 // Heuristically try to determine whether the parentheses contain a type.
2154 auto IsQualifiedPointerOrReference = [](FormatToken *T) {
2155 // This is used to handle cases such as x = (foo *const)&y;
2156 assert(!T->isSimpleTypeSpecifier() && "Should have already been checked");
2157 // Strip trailing qualifiers such as const or volatile when checking
2158 // whether the parens could be a cast to a pointer/reference type.
2159 while (T) {
2160 if (T->is(TT_AttributeParen)) {
2161 // Handle `x = (foo *__attribute__((foo)))&v;`:
2162 if (T->MatchingParen && T->MatchingParen->Previous &&
2163 T->MatchingParen->Previous->is(tok::kw___attribute)) {
2164 T = T->MatchingParen->Previous->Previous;
2165 continue;
2166 }
2167 } else if (T->is(TT_AttributeSquare)) {
2168 // Handle `x = (foo *[[clang::foo]])&v;`:
2169 if (T->MatchingParen && T->MatchingParen->Previous) {
2170 T = T->MatchingParen->Previous;
2171 continue;
2172 }
2173 } else if (T->canBePointerOrReferenceQualifier()) {
2174 T = T->Previous;
2175 continue;
2176 }
2177 break;
2178 }
2179 return T && T->is(TT_PointerOrReference);
2180 };
2181 bool ParensAreType =
2182 !Tok.Previous ||
2183 Tok.Previous->isOneOf(TT_TemplateCloser, TT_TypeDeclarationParen) ||
2184 Tok.Previous->isSimpleTypeSpecifier() ||
2185 IsQualifiedPointerOrReference(Tok.Previous);
2186 bool ParensCouldEndDecl =
2187 Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
2188 if (ParensAreType && !ParensCouldEndDecl)
2189 return true;
2190
2191 // At this point, we heuristically assume that there are no casts at the
2192 // start of the line. We assume that we have found most cases where there
2193 // are by the logic above, e.g. "(void)x;".
2194 if (!LeftOfParens)
2195 return false;
2196
2197 // Certain token types inside the parentheses mean that this can't be a
2198 // cast.
2199 for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
2200 Token = Token->Next) {
2201 if (Token->is(TT_BinaryOperator))
2202 return false;
2203 }
2204
2205 // If the following token is an identifier or 'this', this is a cast. All
2206 // cases where this can be something else are handled above.
2207 if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
2208 return true;
2209
2210 // Look for a cast `( x ) (`.
2211 if (Tok.Next->is(tok::l_paren) && Tok.Previous && Tok.Previous->Previous) {
2212 if (Tok.Previous->is(tok::identifier) &&
2213 Tok.Previous->Previous->is(tok::l_paren)) {
2214 return true;
2215 }
2216 }
2217
2218 if (!Tok.Next->Next)
2219 return false;
2220
2221 // If the next token after the parenthesis is a unary operator, assume
2222 // that this is cast, unless there are unexpected tokens inside the
2223 // parenthesis.
2224 bool NextIsUnary =
2225 Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
2226 if (!NextIsUnary || Tok.Next->is(tok::plus) ||
2227 !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant)) {
2228 return false;
2229 }
2230 // Search for unexpected tokens.
2231 for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
2232 Prev = Prev->Previous) {
2233 if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
2234 return false;
2235 }
2236 return true;
2237 }
2238
2239 /// Returns true if the token is used as a unary operator.
determineUnaryOperatorByUsage(const FormatToken & Tok)2240 bool determineUnaryOperatorByUsage(const FormatToken &Tok) {
2241 const FormatToken *PrevToken = Tok.getPreviousNonComment();
2242 if (!PrevToken)
2243 return true;
2244
2245 // These keywords are deliberately not included here because they may
2246 // precede only one of unary star/amp and plus/minus but not both. They are
2247 // either included in determineStarAmpUsage or determinePlusMinusCaretUsage.
2248 //
2249 // @ - It may be followed by a unary `-` in Objective-C literals. We don't
2250 // know how they can be followed by a star or amp.
2251 if (PrevToken->isOneOf(
2252 TT_ConditionalExpr, tok::l_paren, tok::comma, tok::colon, tok::semi,
2253 tok::equal, tok::question, tok::l_square, tok::l_brace,
2254 tok::kw_case, tok::kw_co_await, tok::kw_co_return, tok::kw_co_yield,
2255 tok::kw_delete, tok::kw_return, tok::kw_throw)) {
2256 return true;
2257 }
2258
2259 // We put sizeof here instead of only in determineStarAmpUsage. In the cases
2260 // where the unary `+` operator is overloaded, it is reasonable to write
2261 // things like `sizeof +x`. Like commit 446d6ec996c6c3.
2262 if (PrevToken->is(tok::kw_sizeof))
2263 return true;
2264
2265 // A sequence of leading unary operators.
2266 if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator))
2267 return true;
2268
2269 // There can't be two consecutive binary operators.
2270 if (PrevToken->is(TT_BinaryOperator))
2271 return true;
2272
2273 return false;
2274 }
2275
2276 /// Return the type of the given token assuming it is * or &.
determineStarAmpUsage(const FormatToken & Tok,bool IsExpression,bool InTemplateArgument)2277 TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
2278 bool InTemplateArgument) {
2279 if (Style.isJavaScript())
2280 return TT_BinaryOperator;
2281
2282 // && in C# must be a binary operator.
2283 if (Style.isCSharp() && Tok.is(tok::ampamp))
2284 return TT_BinaryOperator;
2285
2286 const FormatToken *PrevToken = Tok.getPreviousNonComment();
2287 if (!PrevToken)
2288 return TT_UnaryOperator;
2289
2290 const FormatToken *NextToken = Tok.getNextNonComment();
2291
2292 if (InTemplateArgument && NextToken && NextToken->is(tok::kw_noexcept))
2293 return TT_BinaryOperator;
2294
2295 if (!NextToken ||
2296 NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_noexcept) ||
2297 NextToken->canBePointerOrReferenceQualifier() ||
2298 (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment())) {
2299 return TT_PointerOrReference;
2300 }
2301
2302 if (PrevToken->is(tok::coloncolon))
2303 return TT_PointerOrReference;
2304
2305 if (PrevToken->is(tok::r_paren) && PrevToken->is(TT_TypeDeclarationParen))
2306 return TT_PointerOrReference;
2307
2308 if (determineUnaryOperatorByUsage(Tok))
2309 return TT_UnaryOperator;
2310
2311 if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
2312 return TT_PointerOrReference;
2313 if (NextToken->is(tok::kw_operator) && !IsExpression)
2314 return TT_PointerOrReference;
2315 if (NextToken->isOneOf(tok::comma, tok::semi))
2316 return TT_PointerOrReference;
2317
2318 // After right braces, star tokens are likely to be pointers to struct,
2319 // union, or class.
2320 // struct {} *ptr;
2321 // This by itself is not sufficient to distinguish from multiplication
2322 // following a brace-initialized expression, as in:
2323 // int i = int{42} * 2;
2324 // In the struct case, the part of the struct declaration until the `{` and
2325 // the `}` are put on separate unwrapped lines; in the brace-initialized
2326 // case, the matching `{` is on the same unwrapped line, so check for the
2327 // presence of the matching brace to distinguish between those.
2328 if (PrevToken->is(tok::r_brace) && Tok.is(tok::star) &&
2329 !PrevToken->MatchingParen)
2330 return TT_PointerOrReference;
2331
2332 if (PrevToken->Tok.isLiteral() ||
2333 PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
2334 tok::kw_false, tok::r_brace)) {
2335 return TT_BinaryOperator;
2336 }
2337
2338 const FormatToken *NextNonParen = NextToken;
2339 while (NextNonParen && NextNonParen->is(tok::l_paren))
2340 NextNonParen = NextNonParen->getNextNonComment();
2341 if (NextNonParen && (NextNonParen->Tok.isLiteral() ||
2342 NextNonParen->isOneOf(tok::kw_true, tok::kw_false) ||
2343 NextNonParen->isUnaryOperator())) {
2344 return TT_BinaryOperator;
2345 }
2346
2347 // If we know we're in a template argument, there are no named declarations.
2348 // Thus, having an identifier on the right-hand side indicates a binary
2349 // operator.
2350 if (InTemplateArgument && NextToken->Tok.isAnyIdentifier())
2351 return TT_BinaryOperator;
2352
2353 // "&&(" is quite unlikely to be two successive unary "&".
2354 if (Tok.is(tok::ampamp) && NextToken->is(tok::l_paren))
2355 return TT_BinaryOperator;
2356
2357 // This catches some cases where evaluation order is used as control flow:
2358 // aaa && aaa->f();
2359 if (NextToken->Tok.isAnyIdentifier()) {
2360 const FormatToken *NextNextToken = NextToken->getNextNonComment();
2361 if (NextNextToken && NextNextToken->is(tok::arrow))
2362 return TT_BinaryOperator;
2363 }
2364
2365 // It is very unlikely that we are going to find a pointer or reference type
2366 // definition on the RHS of an assignment.
2367 if (IsExpression && !Contexts.back().CaretFound)
2368 return TT_BinaryOperator;
2369
2370 return TT_PointerOrReference;
2371 }
2372
determinePlusMinusCaretUsage(const FormatToken & Tok)2373 TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
2374 if (determineUnaryOperatorByUsage(Tok))
2375 return TT_UnaryOperator;
2376
2377 const FormatToken *PrevToken = Tok.getPreviousNonComment();
2378 if (!PrevToken)
2379 return TT_UnaryOperator;
2380
2381 if (PrevToken->is(tok::at))
2382 return TT_UnaryOperator;
2383
2384 // Fall back to marking the token as binary operator.
2385 return TT_BinaryOperator;
2386 }
2387
2388 /// Determine whether ++/-- are pre- or post-increments/-decrements.
determineIncrementUsage(const FormatToken & Tok)2389 TokenType determineIncrementUsage(const FormatToken &Tok) {
2390 const FormatToken *PrevToken = Tok.getPreviousNonComment();
2391 if (!PrevToken || PrevToken->is(TT_CastRParen))
2392 return TT_UnaryOperator;
2393 if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
2394 return TT_TrailingUnaryOperator;
2395
2396 return TT_UnaryOperator;
2397 }
2398
2399 SmallVector<Context, 8> Contexts;
2400
2401 const FormatStyle &Style;
2402 AnnotatedLine &Line;
2403 FormatToken *CurrentToken;
2404 bool AutoFound;
2405 const AdditionalKeywords &Keywords;
2406
2407 // Set of "<" tokens that do not open a template parameter list. If parseAngle
2408 // determines that a specific token can't be a template opener, it will make
2409 // same decision irrespective of the decisions for tokens leading up to it.
2410 // Store this information to prevent this from causing exponential runtime.
2411 llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
2412 };
2413
2414 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
2415 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
2416
2417 /// Parses binary expressions by inserting fake parenthesis based on
2418 /// operator precedence.
2419 class ExpressionParser {
2420 public:
ExpressionParser(const FormatStyle & Style,const AdditionalKeywords & Keywords,AnnotatedLine & Line)2421 ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
2422 AnnotatedLine &Line)
2423 : Style(Style), Keywords(Keywords), Line(Line), Current(Line.First) {}
2424
2425 /// Parse expressions with the given operator precedence.
parse(int Precedence=0)2426 void parse(int Precedence = 0) {
2427 // Skip 'return' and ObjC selector colons as they are not part of a binary
2428 // expression.
2429 while (Current && (Current->is(tok::kw_return) ||
2430 (Current->is(tok::colon) &&
2431 Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)))) {
2432 next();
2433 }
2434
2435 if (!Current || Precedence > PrecedenceArrowAndPeriod)
2436 return;
2437
2438 // Conditional expressions need to be parsed separately for proper nesting.
2439 if (Precedence == prec::Conditional) {
2440 parseConditionalExpr();
2441 return;
2442 }
2443
2444 // Parse unary operators, which all have a higher precedence than binary
2445 // operators.
2446 if (Precedence == PrecedenceUnaryOperator) {
2447 parseUnaryOperator();
2448 return;
2449 }
2450
2451 FormatToken *Start = Current;
2452 FormatToken *LatestOperator = nullptr;
2453 unsigned OperatorIndex = 0;
2454
2455 while (Current) {
2456 // Consume operators with higher precedence.
2457 parse(Precedence + 1);
2458
2459 int CurrentPrecedence = getCurrentPrecedence();
2460
2461 if (Precedence == CurrentPrecedence && Current &&
2462 Current->is(TT_SelectorName)) {
2463 if (LatestOperator)
2464 addFakeParenthesis(Start, prec::Level(Precedence));
2465 Start = Current;
2466 }
2467
2468 // At the end of the line or when an operator with higher precedence is
2469 // found, insert fake parenthesis and return.
2470 if (!Current ||
2471 (Current->closesScope() &&
2472 (Current->MatchingParen || Current->is(TT_TemplateString))) ||
2473 (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
2474 (CurrentPrecedence == prec::Conditional &&
2475 Precedence == prec::Assignment && Current->is(tok::colon))) {
2476 break;
2477 }
2478
2479 // Consume scopes: (), [], <> and {}
2480 // In addition to that we handle require clauses as scope, so that the
2481 // constraints in that are correctly indented.
2482 if (Current->opensScope() ||
2483 Current->isOneOf(TT_RequiresClause,
2484 TT_RequiresClauseInARequiresExpression)) {
2485 // In fragment of a JavaScript template string can look like '}..${' and
2486 // thus close a scope and open a new one at the same time.
2487 while (Current && (!Current->closesScope() || Current->opensScope())) {
2488 next();
2489 parse();
2490 }
2491 next();
2492 } else {
2493 // Operator found.
2494 if (CurrentPrecedence == Precedence) {
2495 if (LatestOperator)
2496 LatestOperator->NextOperator = Current;
2497 LatestOperator = Current;
2498 Current->OperatorIndex = OperatorIndex;
2499 ++OperatorIndex;
2500 }
2501 next(/*SkipPastLeadingComments=*/Precedence > 0);
2502 }
2503 }
2504
2505 if (LatestOperator && (Current || Precedence > 0)) {
2506 // The requires clauses do not neccessarily end in a semicolon or a brace,
2507 // but just go over to struct/class or a function declaration, we need to
2508 // intervene so that the fake right paren is inserted correctly.
2509 auto End =
2510 (Start->Previous &&
2511 Start->Previous->isOneOf(TT_RequiresClause,
2512 TT_RequiresClauseInARequiresExpression))
2513 ? [this](){
2514 auto Ret = Current ? Current : Line.Last;
2515 while (!Ret->ClosesRequiresClause && Ret->Previous)
2516 Ret = Ret->Previous;
2517 return Ret;
2518 }()
2519 : nullptr;
2520
2521 if (Precedence == PrecedenceArrowAndPeriod) {
2522 // Call expressions don't have a binary operator precedence.
2523 addFakeParenthesis(Start, prec::Unknown, End);
2524 } else {
2525 addFakeParenthesis(Start, prec::Level(Precedence), End);
2526 }
2527 }
2528 }
2529
2530 private:
2531 /// Gets the precedence (+1) of the given token for binary operators
2532 /// and other tokens that we treat like binary operators.
getCurrentPrecedence()2533 int getCurrentPrecedence() {
2534 if (Current) {
2535 const FormatToken *NextNonComment = Current->getNextNonComment();
2536 if (Current->is(TT_ConditionalExpr))
2537 return prec::Conditional;
2538 if (NextNonComment && Current->is(TT_SelectorName) &&
2539 (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) ||
2540 ((Style.Language == FormatStyle::LK_Proto ||
2541 Style.Language == FormatStyle::LK_TextProto) &&
2542 NextNonComment->is(tok::less)))) {
2543 return prec::Assignment;
2544 }
2545 if (Current->is(TT_JsComputedPropertyName))
2546 return prec::Assignment;
2547 if (Current->is(TT_LambdaArrow))
2548 return prec::Comma;
2549 if (Current->is(TT_FatArrow))
2550 return prec::Assignment;
2551 if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
2552 (Current->is(tok::comment) && NextNonComment &&
2553 NextNonComment->is(TT_SelectorName))) {
2554 return 0;
2555 }
2556 if (Current->is(TT_RangeBasedForLoopColon))
2557 return prec::Comma;
2558 if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) &&
2559 Current->is(Keywords.kw_instanceof)) {
2560 return prec::Relational;
2561 }
2562 if (Style.isJavaScript() &&
2563 Current->isOneOf(Keywords.kw_in, Keywords.kw_as)) {
2564 return prec::Relational;
2565 }
2566 if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
2567 return Current->getPrecedence();
2568 if (Current->isOneOf(tok::period, tok::arrow))
2569 return PrecedenceArrowAndPeriod;
2570 if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) &&
2571 Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
2572 Keywords.kw_throws)) {
2573 return 0;
2574 }
2575 }
2576 return -1;
2577 }
2578
addFakeParenthesis(FormatToken * Start,prec::Level Precedence,FormatToken * End=nullptr)2579 void addFakeParenthesis(FormatToken *Start, prec::Level Precedence,
2580 FormatToken *End = nullptr) {
2581 Start->FakeLParens.push_back(Precedence);
2582 if (Precedence > prec::Unknown)
2583 Start->StartsBinaryExpression = true;
2584 if (!End && Current)
2585 End = Current->getPreviousNonComment();
2586 if (End) {
2587 ++End->FakeRParens;
2588 if (Precedence > prec::Unknown)
2589 End->EndsBinaryExpression = true;
2590 }
2591 }
2592
2593 /// Parse unary operator expressions and surround them with fake
2594 /// parentheses if appropriate.
parseUnaryOperator()2595 void parseUnaryOperator() {
2596 llvm::SmallVector<FormatToken *, 2> Tokens;
2597 while (Current && Current->is(TT_UnaryOperator)) {
2598 Tokens.push_back(Current);
2599 next();
2600 }
2601 parse(PrecedenceArrowAndPeriod);
2602 for (FormatToken *Token : llvm::reverse(Tokens)) {
2603 // The actual precedence doesn't matter.
2604 addFakeParenthesis(Token, prec::Unknown);
2605 }
2606 }
2607
parseConditionalExpr()2608 void parseConditionalExpr() {
2609 while (Current && Current->isTrailingComment())
2610 next();
2611 FormatToken *Start = Current;
2612 parse(prec::LogicalOr);
2613 if (!Current || !Current->is(tok::question))
2614 return;
2615 next();
2616 parse(prec::Assignment);
2617 if (!Current || Current->isNot(TT_ConditionalExpr))
2618 return;
2619 next();
2620 parse(prec::Assignment);
2621 addFakeParenthesis(Start, prec::Conditional);
2622 }
2623
next(bool SkipPastLeadingComments=true)2624 void next(bool SkipPastLeadingComments = true) {
2625 if (Current)
2626 Current = Current->Next;
2627 while (Current &&
2628 (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
2629 Current->isTrailingComment()) {
2630 Current = Current->Next;
2631 }
2632 }
2633
2634 const FormatStyle &Style;
2635 const AdditionalKeywords &Keywords;
2636 const AnnotatedLine &Line;
2637 FormatToken *Current;
2638 };
2639
2640 } // end anonymous namespace
2641
setCommentLineLevels(SmallVectorImpl<AnnotatedLine * > & Lines) const2642 void TokenAnnotator::setCommentLineLevels(
2643 SmallVectorImpl<AnnotatedLine *> &Lines) const {
2644 const AnnotatedLine *NextNonCommentLine = nullptr;
2645 for (AnnotatedLine *Line : llvm::reverse(Lines)) {
2646 assert(Line->First);
2647
2648 // If the comment is currently aligned with the line immediately following
2649 // it, that's probably intentional and we should keep it.
2650 if (NextNonCommentLine && Line->isComment() &&
2651 NextNonCommentLine->First->NewlinesBefore <= 1 &&
2652 NextNonCommentLine->First->OriginalColumn ==
2653 Line->First->OriginalColumn) {
2654 // Align comments for preprocessor lines with the # in column 0 if
2655 // preprocessor lines are not indented. Otherwise, align with the next
2656 // line.
2657 Line->Level =
2658 (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
2659 (NextNonCommentLine->Type == LT_PreprocessorDirective ||
2660 NextNonCommentLine->Type == LT_ImportStatement))
2661 ? 0
2662 : NextNonCommentLine->Level;
2663 } else {
2664 NextNonCommentLine = Line->First->isNot(tok::r_brace) ? Line : nullptr;
2665 }
2666
2667 setCommentLineLevels(Line->Children);
2668 }
2669 }
2670
maxNestingDepth(const AnnotatedLine & Line)2671 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
2672 unsigned Result = 0;
2673 for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
2674 Result = std::max(Result, Tok->NestingLevel);
2675 return Result;
2676 }
2677
annotate(AnnotatedLine & Line) const2678 void TokenAnnotator::annotate(AnnotatedLine &Line) const {
2679 for (auto &Child : Line.Children)
2680 annotate(*Child);
2681
2682 AnnotatingParser Parser(Style, Line, Keywords);
2683 Line.Type = Parser.parseLine();
2684
2685 // With very deep nesting, ExpressionParser uses lots of stack and the
2686 // formatting algorithm is very slow. We're not going to do a good job here
2687 // anyway - it's probably generated code being formatted by mistake.
2688 // Just skip the whole line.
2689 if (maxNestingDepth(Line) > 50)
2690 Line.Type = LT_Invalid;
2691
2692 if (Line.Type == LT_Invalid)
2693 return;
2694
2695 ExpressionParser ExprParser(Style, Keywords, Line);
2696 ExprParser.parse();
2697
2698 if (Line.startsWith(TT_ObjCMethodSpecifier))
2699 Line.Type = LT_ObjCMethodDecl;
2700 else if (Line.startsWith(TT_ObjCDecl))
2701 Line.Type = LT_ObjCDecl;
2702 else if (Line.startsWith(TT_ObjCProperty))
2703 Line.Type = LT_ObjCProperty;
2704
2705 Line.First->SpacesRequiredBefore = 1;
2706 Line.First->CanBreakBefore = Line.First->MustBreakBefore;
2707 }
2708
2709 // This function heuristically determines whether 'Current' starts the name of a
2710 // function declaration.
isFunctionDeclarationName(bool IsCpp,const FormatToken & Current,const AnnotatedLine & Line)2711 static bool isFunctionDeclarationName(bool IsCpp, const FormatToken &Current,
2712 const AnnotatedLine &Line) {
2713 auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * {
2714 for (; Next; Next = Next->Next) {
2715 if (Next->is(TT_OverloadedOperatorLParen))
2716 return Next;
2717 if (Next->is(TT_OverloadedOperator))
2718 continue;
2719 if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
2720 // For 'new[]' and 'delete[]'.
2721 if (Next->Next &&
2722 Next->Next->startsSequence(tok::l_square, tok::r_square)) {
2723 Next = Next->Next->Next;
2724 }
2725 continue;
2726 }
2727 if (Next->startsSequence(tok::l_square, tok::r_square)) {
2728 // For operator[]().
2729 Next = Next->Next;
2730 continue;
2731 }
2732 if ((Next->isSimpleTypeSpecifier() || Next->is(tok::identifier)) &&
2733 Next->Next && Next->Next->isOneOf(tok::star, tok::amp, tok::ampamp)) {
2734 // For operator void*(), operator char*(), operator Foo*().
2735 Next = Next->Next;
2736 continue;
2737 }
2738 if (Next->is(TT_TemplateOpener) && Next->MatchingParen) {
2739 Next = Next->MatchingParen;
2740 continue;
2741 }
2742
2743 break;
2744 }
2745 return nullptr;
2746 };
2747
2748 // Find parentheses of parameter list.
2749 const FormatToken *Next = Current.Next;
2750 if (Current.is(tok::kw_operator)) {
2751 if (Current.Previous && Current.Previous->is(tok::coloncolon))
2752 return false;
2753 Next = skipOperatorName(Next);
2754 } else {
2755 if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
2756 return false;
2757 for (; Next; Next = Next->Next) {
2758 if (Next->is(TT_TemplateOpener)) {
2759 Next = Next->MatchingParen;
2760 } else if (Next->is(tok::coloncolon)) {
2761 Next = Next->Next;
2762 if (!Next)
2763 return false;
2764 if (Next->is(tok::kw_operator)) {
2765 Next = skipOperatorName(Next->Next);
2766 break;
2767 }
2768 if (!Next->is(tok::identifier))
2769 return false;
2770 } else if (Next->is(tok::l_paren)) {
2771 break;
2772 } else {
2773 return false;
2774 }
2775 }
2776 }
2777
2778 // Check whether parameter list can belong to a function declaration.
2779 if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
2780 return false;
2781 // If the lines ends with "{", this is likely a function definition.
2782 if (Line.Last->is(tok::l_brace))
2783 return true;
2784 if (Next->Next == Next->MatchingParen)
2785 return true; // Empty parentheses.
2786 // If there is an &/&& after the r_paren, this is likely a function.
2787 if (Next->MatchingParen->Next &&
2788 Next->MatchingParen->Next->is(TT_PointerOrReference)) {
2789 return true;
2790 }
2791
2792 // Check for K&R C function definitions (and C++ function definitions with
2793 // unnamed parameters), e.g.:
2794 // int f(i)
2795 // {
2796 // return i + 1;
2797 // }
2798 // bool g(size_t = 0, bool b = false)
2799 // {
2800 // return !b;
2801 // }
2802 if (IsCpp && Next->Next && Next->Next->is(tok::identifier) &&
2803 !Line.endsWith(tok::semi)) {
2804 return true;
2805 }
2806
2807 for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
2808 Tok = Tok->Next) {
2809 if (Tok->is(TT_TypeDeclarationParen))
2810 return true;
2811 if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) {
2812 Tok = Tok->MatchingParen;
2813 continue;
2814 }
2815 if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
2816 Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis)) {
2817 return true;
2818 }
2819 if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
2820 Tok->Tok.isLiteral()) {
2821 return false;
2822 }
2823 }
2824 return false;
2825 }
2826
mustBreakForReturnType(const AnnotatedLine & Line) const2827 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
2828 assert(Line.MightBeFunctionDecl);
2829
2830 if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
2831 Style.AlwaysBreakAfterReturnType ==
2832 FormatStyle::RTBS_TopLevelDefinitions) &&
2833 Line.Level > 0) {
2834 return false;
2835 }
2836
2837 switch (Style.AlwaysBreakAfterReturnType) {
2838 case FormatStyle::RTBS_None:
2839 return false;
2840 case FormatStyle::RTBS_All:
2841 case FormatStyle::RTBS_TopLevel:
2842 return true;
2843 case FormatStyle::RTBS_AllDefinitions:
2844 case FormatStyle::RTBS_TopLevelDefinitions:
2845 return Line.mightBeFunctionDefinition();
2846 }
2847
2848 return false;
2849 }
2850
calculateFormattingInformation(AnnotatedLine & Line) const2851 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) const {
2852 for (AnnotatedLine *ChildLine : Line.Children)
2853 calculateFormattingInformation(*ChildLine);
2854
2855 Line.First->TotalLength =
2856 Line.First->IsMultiline ? Style.ColumnLimit
2857 : Line.FirstStartColumn + Line.First->ColumnWidth;
2858 FormatToken *Current = Line.First->Next;
2859 bool InFunctionDecl = Line.MightBeFunctionDecl;
2860 bool AlignArrayOfStructures =
2861 (Style.AlignArrayOfStructures != FormatStyle::AIAS_None &&
2862 Line.Type == LT_ArrayOfStructInitializer);
2863 if (AlignArrayOfStructures)
2864 calculateArrayInitializerColumnList(Line);
2865
2866 while (Current) {
2867 if (isFunctionDeclarationName(Style.isCpp(), *Current, Line))
2868 Current->setType(TT_FunctionDeclarationName);
2869 const FormatToken *Prev = Current->Previous;
2870 if (Current->is(TT_LineComment)) {
2871 if (Prev->is(BK_BracedInit) && Prev->opensScope()) {
2872 Current->SpacesRequiredBefore =
2873 (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1;
2874 } else {
2875 Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
2876 }
2877
2878 // If we find a trailing comment, iterate backwards to determine whether
2879 // it seems to relate to a specific parameter. If so, break before that
2880 // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
2881 // to the previous line in:
2882 // SomeFunction(a,
2883 // b, // comment
2884 // c);
2885 if (!Current->HasUnescapedNewline) {
2886 for (FormatToken *Parameter = Current->Previous; Parameter;
2887 Parameter = Parameter->Previous) {
2888 if (Parameter->isOneOf(tok::comment, tok::r_brace))
2889 break;
2890 if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
2891 if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
2892 Parameter->HasUnescapedNewline) {
2893 Parameter->MustBreakBefore = true;
2894 }
2895 break;
2896 }
2897 }
2898 }
2899 } else if (Current->SpacesRequiredBefore == 0 &&
2900 spaceRequiredBefore(Line, *Current)) {
2901 Current->SpacesRequiredBefore = 1;
2902 }
2903
2904 const auto &Children = Prev->Children;
2905 if (!Children.empty() && Children.back()->Last->is(TT_LineComment)) {
2906 Current->MustBreakBefore = true;
2907 } else {
2908 Current->MustBreakBefore =
2909 Current->MustBreakBefore || mustBreakBefore(Line, *Current);
2910 if (!Current->MustBreakBefore && InFunctionDecl &&
2911 Current->is(TT_FunctionDeclarationName)) {
2912 Current->MustBreakBefore = mustBreakForReturnType(Line);
2913 }
2914 }
2915
2916 Current->CanBreakBefore =
2917 Current->MustBreakBefore || canBreakBefore(Line, *Current);
2918 unsigned ChildSize = 0;
2919 if (Prev->Children.size() == 1) {
2920 FormatToken &LastOfChild = *Prev->Children[0]->Last;
2921 ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
2922 : LastOfChild.TotalLength + 1;
2923 }
2924 if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
2925 (Prev->Children.size() == 1 &&
2926 Prev->Children[0]->First->MustBreakBefore) ||
2927 Current->IsMultiline) {
2928 Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
2929 } else {
2930 Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
2931 ChildSize + Current->SpacesRequiredBefore;
2932 }
2933
2934 if (Current->is(TT_CtorInitializerColon))
2935 InFunctionDecl = false;
2936
2937 // FIXME: Only calculate this if CanBreakBefore is true once static
2938 // initializers etc. are sorted out.
2939 // FIXME: Move magic numbers to a better place.
2940
2941 // Reduce penalty for aligning ObjC method arguments using the colon
2942 // alignment as this is the canonical way (still prefer fitting everything
2943 // into one line if possible). Trying to fit a whole expression into one
2944 // line should not force other line breaks (e.g. when ObjC method
2945 // expression is a part of other expression).
2946 Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl);
2947 if (Style.Language == FormatStyle::LK_ObjC &&
2948 Current->is(TT_SelectorName) && Current->ParameterIndex > 0) {
2949 if (Current->ParameterIndex == 1)
2950 Current->SplitPenalty += 5 * Current->BindingStrength;
2951 } else {
2952 Current->SplitPenalty += 20 * Current->BindingStrength;
2953 }
2954
2955 Current = Current->Next;
2956 }
2957
2958 calculateUnbreakableTailLengths(Line);
2959 unsigned IndentLevel = Line.Level;
2960 for (Current = Line.First; Current != nullptr; Current = Current->Next) {
2961 if (Current->Role)
2962 Current->Role->precomputeFormattingInfos(Current);
2963 if (Current->MatchingParen &&
2964 Current->MatchingParen->opensBlockOrBlockTypeList(Style) &&
2965 IndentLevel > 0) {
2966 --IndentLevel;
2967 }
2968 Current->IndentLevel = IndentLevel;
2969 if (Current->opensBlockOrBlockTypeList(Style))
2970 ++IndentLevel;
2971 }
2972
2973 LLVM_DEBUG({ printDebugInfo(Line); });
2974 }
2975
calculateUnbreakableTailLengths(AnnotatedLine & Line) const2976 void TokenAnnotator::calculateUnbreakableTailLengths(
2977 AnnotatedLine &Line) const {
2978 unsigned UnbreakableTailLength = 0;
2979 FormatToken *Current = Line.Last;
2980 while (Current) {
2981 Current->UnbreakableTailLength = UnbreakableTailLength;
2982 if (Current->CanBreakBefore ||
2983 Current->isOneOf(tok::comment, tok::string_literal)) {
2984 UnbreakableTailLength = 0;
2985 } else {
2986 UnbreakableTailLength +=
2987 Current->ColumnWidth + Current->SpacesRequiredBefore;
2988 }
2989 Current = Current->Previous;
2990 }
2991 }
2992
calculateArrayInitializerColumnList(AnnotatedLine & Line) const2993 void TokenAnnotator::calculateArrayInitializerColumnList(
2994 AnnotatedLine &Line) const {
2995 if (Line.First == Line.Last)
2996 return;
2997 auto *CurrentToken = Line.First;
2998 CurrentToken->ArrayInitializerLineStart = true;
2999 unsigned Depth = 0;
3000 while (CurrentToken != nullptr && CurrentToken != Line.Last) {
3001 if (CurrentToken->is(tok::l_brace)) {
3002 CurrentToken->IsArrayInitializer = true;
3003 if (CurrentToken->Next != nullptr)
3004 CurrentToken->Next->MustBreakBefore = true;
3005 CurrentToken =
3006 calculateInitializerColumnList(Line, CurrentToken->Next, Depth + 1);
3007 } else {
3008 CurrentToken = CurrentToken->Next;
3009 }
3010 }
3011 }
3012
calculateInitializerColumnList(AnnotatedLine & Line,FormatToken * CurrentToken,unsigned Depth) const3013 FormatToken *TokenAnnotator::calculateInitializerColumnList(
3014 AnnotatedLine &Line, FormatToken *CurrentToken, unsigned Depth) const {
3015 while (CurrentToken != nullptr && CurrentToken != Line.Last) {
3016 if (CurrentToken->is(tok::l_brace))
3017 ++Depth;
3018 else if (CurrentToken->is(tok::r_brace))
3019 --Depth;
3020 if (Depth == 2 && CurrentToken->isOneOf(tok::l_brace, tok::comma)) {
3021 CurrentToken = CurrentToken->Next;
3022 if (CurrentToken == nullptr)
3023 break;
3024 CurrentToken->StartsColumn = true;
3025 CurrentToken = CurrentToken->Previous;
3026 }
3027 CurrentToken = CurrentToken->Next;
3028 }
3029 return CurrentToken;
3030 }
3031
splitPenalty(const AnnotatedLine & Line,const FormatToken & Tok,bool InFunctionDecl) const3032 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
3033 const FormatToken &Tok,
3034 bool InFunctionDecl) const {
3035 const FormatToken &Left = *Tok.Previous;
3036 const FormatToken &Right = Tok;
3037
3038 if (Left.is(tok::semi))
3039 return 0;
3040
3041 if (Style.Language == FormatStyle::LK_Java) {
3042 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
3043 return 1;
3044 if (Right.is(Keywords.kw_implements))
3045 return 2;
3046 if (Left.is(tok::comma) && Left.NestingLevel == 0)
3047 return 3;
3048 } else if (Style.isJavaScript()) {
3049 if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
3050 return 100;
3051 if (Left.is(TT_JsTypeColon))
3052 return 35;
3053 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
3054 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) {
3055 return 100;
3056 }
3057 // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()".
3058 if (Left.opensScope() && Right.closesScope())
3059 return 200;
3060 }
3061
3062 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
3063 return 1;
3064 if (Right.is(tok::l_square)) {
3065 if (Style.Language == FormatStyle::LK_Proto)
3066 return 1;
3067 if (Left.is(tok::r_square))
3068 return 200;
3069 // Slightly prefer formatting local lambda definitions like functions.
3070 if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
3071 return 35;
3072 if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
3073 TT_ArrayInitializerLSquare,
3074 TT_DesignatedInitializerLSquare, TT_AttributeSquare)) {
3075 return 500;
3076 }
3077 }
3078
3079 if (Left.is(tok::coloncolon) ||
3080 (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto)) {
3081 return 500;
3082 }
3083 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
3084 Right.is(tok::kw_operator)) {
3085 if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
3086 return 3;
3087 if (Left.is(TT_StartOfName))
3088 return 110;
3089 if (InFunctionDecl && Right.NestingLevel == 0)
3090 return Style.PenaltyReturnTypeOnItsOwnLine;
3091 return 200;
3092 }
3093 if (Right.is(TT_PointerOrReference))
3094 return 190;
3095 if (Right.is(TT_LambdaArrow))
3096 return 110;
3097 if (Left.is(tok::equal) && Right.is(tok::l_brace))
3098 return 160;
3099 if (Left.is(TT_CastRParen))
3100 return 100;
3101 if (Left.isOneOf(tok::kw_class, tok::kw_struct))
3102 return 5000;
3103 if (Left.is(tok::comment))
3104 return 1000;
3105
3106 if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon,
3107 TT_CtorInitializerColon)) {
3108 return 2;
3109 }
3110
3111 if (Right.isMemberAccess()) {
3112 // Breaking before the "./->" of a chained call/member access is reasonably
3113 // cheap, as formatting those with one call per line is generally
3114 // desirable. In particular, it should be cheaper to break before the call
3115 // than it is to break inside a call's parameters, which could lead to weird
3116 // "hanging" indents. The exception is the very last "./->" to support this
3117 // frequent pattern:
3118 //
3119 // aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
3120 // dddddddd);
3121 //
3122 // which might otherwise be blown up onto many lines. Here, clang-format
3123 // won't produce "hanging" indents anyway as there is no other trailing
3124 // call.
3125 //
3126 // Also apply higher penalty is not a call as that might lead to a wrapping
3127 // like:
3128 //
3129 // aaaaaaa
3130 // .aaaaaaaaa.bbbbbbbb(cccccccc);
3131 return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
3132 ? 150
3133 : 35;
3134 }
3135
3136 if (Right.is(TT_TrailingAnnotation) &&
3137 (!Right.Next || Right.Next->isNot(tok::l_paren))) {
3138 // Moving trailing annotations to the next line is fine for ObjC method
3139 // declarations.
3140 if (Line.startsWith(TT_ObjCMethodSpecifier))
3141 return 10;
3142 // Generally, breaking before a trailing annotation is bad unless it is
3143 // function-like. It seems to be especially preferable to keep standard
3144 // annotations (i.e. "const", "final" and "override") on the same line.
3145 // Use a slightly higher penalty after ")" so that annotations like
3146 // "const override" are kept together.
3147 bool is_short_annotation = Right.TokenText.size() < 10;
3148 return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
3149 }
3150
3151 // In for-loops, prefer breaking at ',' and ';'.
3152 if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
3153 return 4;
3154
3155 // In Objective-C method expressions, prefer breaking before "param:" over
3156 // breaking after it.
3157 if (Right.is(TT_SelectorName))
3158 return 0;
3159 if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
3160 return Line.MightBeFunctionDecl ? 50 : 500;
3161
3162 // In Objective-C type declarations, avoid breaking after the category's
3163 // open paren (we'll prefer breaking after the protocol list's opening
3164 // angle bracket, if present).
3165 if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous &&
3166 Left.Previous->isOneOf(tok::identifier, tok::greater)) {
3167 return 500;
3168 }
3169
3170 if (Left.is(tok::l_paren) && Style.PenaltyBreakOpenParenthesis != 0)
3171 return Style.PenaltyBreakOpenParenthesis;
3172 if (Left.is(tok::l_paren) && InFunctionDecl &&
3173 Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign) {
3174 return 100;
3175 }
3176 if (Left.is(tok::l_paren) && Left.Previous &&
3177 (Left.Previous->is(tok::kw_for) || Left.Previous->isIf())) {
3178 return 1000;
3179 }
3180 if (Left.is(tok::equal) && InFunctionDecl)
3181 return 110;
3182 if (Right.is(tok::r_brace))
3183 return 1;
3184 if (Left.is(TT_TemplateOpener))
3185 return 100;
3186 if (Left.opensScope()) {
3187 // If we aren't aligning after opening parens/braces we can always break
3188 // here unless the style does not want us to place all arguments on the
3189 // next line.
3190 if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign &&
3191 (Left.ParameterCount <= 1 || Style.AllowAllArgumentsOnNextLine)) {
3192 return 0;
3193 }
3194 if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle)
3195 return 19;
3196 return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
3197 : 19;
3198 }
3199 if (Left.is(TT_JavaAnnotation))
3200 return 50;
3201
3202 if (Left.is(TT_UnaryOperator))
3203 return 60;
3204 if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
3205 Left.Previous->isLabelString() &&
3206 (Left.NextOperator || Left.OperatorIndex != 0)) {
3207 return 50;
3208 }
3209 if (Right.is(tok::plus) && Left.isLabelString() &&
3210 (Right.NextOperator || Right.OperatorIndex != 0)) {
3211 return 25;
3212 }
3213 if (Left.is(tok::comma))
3214 return 1;
3215 if (Right.is(tok::lessless) && Left.isLabelString() &&
3216 (Right.NextOperator || Right.OperatorIndex != 1)) {
3217 return 25;
3218 }
3219 if (Right.is(tok::lessless)) {
3220 // Breaking at a << is really cheap.
3221 if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0) {
3222 // Slightly prefer to break before the first one in log-like statements.
3223 return 2;
3224 }
3225 return 1;
3226 }
3227 if (Left.ClosesTemplateDeclaration)
3228 return Style.PenaltyBreakTemplateDeclaration;
3229 if (Left.ClosesRequiresClause)
3230 return 0;
3231 if (Left.is(TT_ConditionalExpr))
3232 return prec::Conditional;
3233 prec::Level Level = Left.getPrecedence();
3234 if (Level == prec::Unknown)
3235 Level = Right.getPrecedence();
3236 if (Level == prec::Assignment)
3237 return Style.PenaltyBreakAssignment;
3238 if (Level != prec::Unknown)
3239 return Level;
3240
3241 return 3;
3242 }
3243
spaceRequiredBeforeParens(const FormatToken & Right) const3244 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const {
3245 if (Style.SpaceBeforeParens == FormatStyle::SBPO_Always)
3246 return true;
3247 if (Right.is(TT_OverloadedOperatorLParen) &&
3248 Style.SpaceBeforeParensOptions.AfterOverloadedOperator) {
3249 return true;
3250 }
3251 if (Style.SpaceBeforeParensOptions.BeforeNonEmptyParentheses &&
3252 Right.ParameterCount > 0) {
3253 return true;
3254 }
3255 return false;
3256 }
3257
spaceRequiredBetween(const AnnotatedLine & Line,const FormatToken & Left,const FormatToken & Right) const3258 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
3259 const FormatToken &Left,
3260 const FormatToken &Right) const {
3261 if (Left.is(tok::kw_return) &&
3262 !Right.isOneOf(tok::semi, tok::r_paren, tok::hashhash)) {
3263 return true;
3264 }
3265 if (Style.isJson() && Left.is(tok::string_literal) && Right.is(tok::colon))
3266 return false;
3267 if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java)
3268 return true;
3269 if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
3270 Left.Tok.getObjCKeywordID() == tok::objc_property) {
3271 return true;
3272 }
3273 if (Right.is(tok::hashhash))
3274 return Left.is(tok::hash);
3275 if (Left.isOneOf(tok::hashhash, tok::hash))
3276 return Right.is(tok::hash);
3277 if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) ||
3278 (Left.is(tok::l_brace) && Left.isNot(BK_Block) &&
3279 Right.is(tok::r_brace) && Right.isNot(BK_Block))) {
3280 return Style.SpaceInEmptyParentheses;
3281 }
3282 if (Style.SpacesInConditionalStatement) {
3283 const FormatToken *LeftParen = nullptr;
3284 if (Left.is(tok::l_paren))
3285 LeftParen = &Left;
3286 else if (Right.is(tok::r_paren) && Right.MatchingParen)
3287 LeftParen = Right.MatchingParen;
3288 if (LeftParen && LeftParen->Previous &&
3289 isKeywordWithCondition(*LeftParen->Previous)) {
3290 return true;
3291 }
3292 }
3293
3294 // auto{x} auto(x)
3295 if (Left.is(tok::kw_auto) && Right.isOneOf(tok::l_paren, tok::l_brace))
3296 return false;
3297
3298 // operator co_await(x)
3299 if (Right.is(tok::l_paren) && Left.is(tok::kw_co_await) && Left.Previous &&
3300 Left.Previous->is(tok::kw_operator)) {
3301 return false;
3302 }
3303 // co_await (x), co_yield (x), co_return (x)
3304 if (Left.isOneOf(tok::kw_co_await, tok::kw_co_yield, tok::kw_co_return) &&
3305 !Right.isOneOf(tok::semi, tok::r_paren)) {
3306 return true;
3307 }
3308
3309 if (Left.is(tok::l_paren) || Right.is(tok::r_paren)) {
3310 return (Right.is(TT_CastRParen) ||
3311 (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
3312 ? Style.SpacesInCStyleCastParentheses
3313 : Style.SpacesInParentheses;
3314 }
3315 if (Right.isOneOf(tok::semi, tok::comma))
3316 return false;
3317 if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) {
3318 bool IsLightweightGeneric = Right.MatchingParen &&
3319 Right.MatchingParen->Next &&
3320 Right.MatchingParen->Next->is(tok::colon);
3321 return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList;
3322 }
3323 if (Right.is(tok::less) && Left.is(tok::kw_template))
3324 return Style.SpaceAfterTemplateKeyword;
3325 if (Left.isOneOf(tok::exclaim, tok::tilde))
3326 return false;
3327 if (Left.is(tok::at) &&
3328 Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
3329 tok::numeric_constant, tok::l_paren, tok::l_brace,
3330 tok::kw_true, tok::kw_false)) {
3331 return false;
3332 }
3333 if (Left.is(tok::colon))
3334 return !Left.is(TT_ObjCMethodExpr);
3335 if (Left.is(tok::coloncolon))
3336 return false;
3337 if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) {
3338 if (Style.Language == FormatStyle::LK_TextProto ||
3339 (Style.Language == FormatStyle::LK_Proto &&
3340 (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) {
3341 // Format empty list as `<>`.
3342 if (Left.is(tok::less) && Right.is(tok::greater))
3343 return false;
3344 return !Style.Cpp11BracedListStyle;
3345 }
3346 return false;
3347 }
3348 if (Right.is(tok::ellipsis)) {
3349 return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
3350 Left.Previous->is(tok::kw_case));
3351 }
3352 if (Left.is(tok::l_square) && Right.is(tok::amp))
3353 return Style.SpacesInSquareBrackets;
3354 if (Right.is(TT_PointerOrReference)) {
3355 if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) {
3356 if (!Left.MatchingParen)
3357 return true;
3358 FormatToken *TokenBeforeMatchingParen =
3359 Left.MatchingParen->getPreviousNonComment();
3360 if (!TokenBeforeMatchingParen || !Left.is(TT_TypeDeclarationParen))
3361 return true;
3362 }
3363 // Add a space if the previous token is a pointer qualifier or the closing
3364 // parenthesis of __attribute__(()) expression and the style requires spaces
3365 // after pointer qualifiers.
3366 if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_After ||
3367 Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) &&
3368 (Left.is(TT_AttributeParen) ||
3369 Left.canBePointerOrReferenceQualifier())) {
3370 return true;
3371 }
3372 if (Left.Tok.isLiteral())
3373 return true;
3374 // for (auto a = 0, b = 0; const auto & c : {1, 2, 3})
3375 if (Left.isTypeOrIdentifier() && Right.Next && Right.Next->Next &&
3376 Right.Next->Next->is(TT_RangeBasedForLoopColon)) {
3377 return getTokenPointerOrReferenceAlignment(Right) !=
3378 FormatStyle::PAS_Left;
3379 }
3380 return !Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
3381 (getTokenPointerOrReferenceAlignment(Right) !=
3382 FormatStyle::PAS_Left ||
3383 (Line.IsMultiVariableDeclStmt &&
3384 (Left.NestingLevel == 0 ||
3385 (Left.NestingLevel == 1 && startsWithInitStatement(Line)))));
3386 }
3387 if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
3388 (!Left.is(TT_PointerOrReference) ||
3389 (getTokenPointerOrReferenceAlignment(Left) != FormatStyle::PAS_Right &&
3390 !Line.IsMultiVariableDeclStmt))) {
3391 return true;
3392 }
3393 if (Left.is(TT_PointerOrReference)) {
3394 // Add a space if the next token is a pointer qualifier and the style
3395 // requires spaces before pointer qualifiers.
3396 if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Before ||
3397 Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) &&
3398 Right.canBePointerOrReferenceQualifier()) {
3399 return true;
3400 }
3401 // & 1
3402 if (Right.Tok.isLiteral())
3403 return true;
3404 // & /* comment
3405 if (Right.is(TT_BlockComment))
3406 return true;
3407 // foo() -> const Bar * override/final
3408 if (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
3409 !Right.is(TT_StartOfName)) {
3410 return true;
3411 }
3412 // & {
3413 if (Right.is(tok::l_brace) && Right.is(BK_Block))
3414 return true;
3415 // for (auto a = 0, b = 0; const auto& c : {1, 2, 3})
3416 if (Left.Previous && Left.Previous->isTypeOrIdentifier() && Right.Next &&
3417 Right.Next->is(TT_RangeBasedForLoopColon)) {
3418 return getTokenPointerOrReferenceAlignment(Left) !=
3419 FormatStyle::PAS_Right;
3420 }
3421 if (Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
3422 tok::l_paren)) {
3423 return false;
3424 }
3425 if (getTokenPointerOrReferenceAlignment(Left) == FormatStyle::PAS_Right)
3426 return false;
3427 // FIXME: Setting IsMultiVariableDeclStmt for the whole line is error-prone,
3428 // because it does not take into account nested scopes like lambdas.
3429 // In multi-variable declaration statements, attach */& to the variable
3430 // independently of the style. However, avoid doing it if we are in a nested
3431 // scope, e.g. lambda. We still need to special-case statements with
3432 // initializers.
3433 if (Line.IsMultiVariableDeclStmt &&
3434 (Left.NestingLevel == Line.First->NestingLevel ||
3435 ((Left.NestingLevel == Line.First->NestingLevel + 1) &&
3436 startsWithInitStatement(Line)))) {
3437 return false;
3438 }
3439 return Left.Previous && !Left.Previous->isOneOf(
3440 tok::l_paren, tok::coloncolon, tok::l_square);
3441 }
3442 // Ensure right pointer alignment with ellipsis e.g. int *...P
3443 if (Left.is(tok::ellipsis) && Left.Previous &&
3444 Left.Previous->isOneOf(tok::star, tok::amp, tok::ampamp)) {
3445 return Style.PointerAlignment != FormatStyle::PAS_Right;
3446 }
3447
3448 if (Right.is(tok::star) && Left.is(tok::l_paren))
3449 return false;
3450 if (Left.is(tok::star) && Right.isOneOf(tok::star, tok::amp, tok::ampamp))
3451 return false;
3452 if (Right.isOneOf(tok::star, tok::amp, tok::ampamp)) {
3453 const FormatToken *Previous = &Left;
3454 while (Previous && !Previous->is(tok::kw_operator)) {
3455 if (Previous->is(tok::identifier) || Previous->isSimpleTypeSpecifier()) {
3456 Previous = Previous->getPreviousNonComment();
3457 continue;
3458 }
3459 if (Previous->is(TT_TemplateCloser) && Previous->MatchingParen) {
3460 Previous = Previous->MatchingParen->getPreviousNonComment();
3461 continue;
3462 }
3463 if (Previous->is(tok::coloncolon)) {
3464 Previous = Previous->getPreviousNonComment();
3465 continue;
3466 }
3467 break;
3468 }
3469 // Space between the type and the * in:
3470 // operator void*()
3471 // operator char*()
3472 // operator void const*()
3473 // operator void volatile*()
3474 // operator /*comment*/ const char*()
3475 // operator volatile /*comment*/ char*()
3476 // operator Foo*()
3477 // operator C<T>*()
3478 // operator std::Foo*()
3479 // operator C<T>::D<U>*()
3480 // dependent on PointerAlignment style.
3481 if (Previous) {
3482 if (Previous->endsSequence(tok::kw_operator))
3483 return Style.PointerAlignment != FormatStyle::PAS_Left;
3484 if (Previous->is(tok::kw_const) || Previous->is(tok::kw_volatile)) {
3485 return (Style.PointerAlignment != FormatStyle::PAS_Left) ||
3486 (Style.SpaceAroundPointerQualifiers ==
3487 FormatStyle::SAPQ_After) ||
3488 (Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both);
3489 }
3490 }
3491 }
3492 const auto SpaceRequiredForArrayInitializerLSquare =
3493 [](const FormatToken &LSquareTok, const FormatStyle &Style) {
3494 return Style.SpacesInContainerLiterals ||
3495 ((Style.Language == FormatStyle::LK_Proto ||
3496 Style.Language == FormatStyle::LK_TextProto) &&
3497 !Style.Cpp11BracedListStyle &&
3498 LSquareTok.endsSequence(tok::l_square, tok::colon,
3499 TT_SelectorName));
3500 };
3501 if (Left.is(tok::l_square)) {
3502 return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) &&
3503 SpaceRequiredForArrayInitializerLSquare(Left, Style)) ||
3504 (Left.isOneOf(TT_ArraySubscriptLSquare, TT_StructuredBindingLSquare,
3505 TT_LambdaLSquare) &&
3506 Style.SpacesInSquareBrackets && Right.isNot(tok::r_square));
3507 }
3508 if (Right.is(tok::r_square)) {
3509 return Right.MatchingParen &&
3510 ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) &&
3511 SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen,
3512 Style)) ||
3513 (Style.SpacesInSquareBrackets &&
3514 Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare,
3515 TT_StructuredBindingLSquare,
3516 TT_LambdaLSquare)) ||
3517 Right.MatchingParen->is(TT_AttributeParen));
3518 }
3519 if (Right.is(tok::l_square) &&
3520 !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
3521 TT_DesignatedInitializerLSquare,
3522 TT_StructuredBindingLSquare, TT_AttributeSquare) &&
3523 !Left.isOneOf(tok::numeric_constant, TT_DictLiteral) &&
3524 !(!Left.is(tok::r_square) && Style.SpaceBeforeSquareBrackets &&
3525 Right.is(TT_ArraySubscriptLSquare))) {
3526 return false;
3527 }
3528 if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
3529 return !Left.Children.empty(); // No spaces in "{}".
3530 if ((Left.is(tok::l_brace) && Left.isNot(BK_Block)) ||
3531 (Right.is(tok::r_brace) && Right.MatchingParen &&
3532 Right.MatchingParen->isNot(BK_Block))) {
3533 return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true;
3534 }
3535 if (Left.is(TT_BlockComment)) {
3536 // No whitespace in x(/*foo=*/1), except for JavaScript.
3537 return Style.isJavaScript() || !Left.TokenText.endswith("=*/");
3538 }
3539
3540 // Space between template and attribute.
3541 // e.g. template <typename T> [[nodiscard]] ...
3542 if (Left.is(TT_TemplateCloser) && Right.is(TT_AttributeSquare))
3543 return true;
3544 // Space before parentheses common for all languages
3545 if (Right.is(tok::l_paren)) {
3546 if (Left.is(TT_TemplateCloser) && Right.isNot(TT_FunctionTypeLParen))
3547 return spaceRequiredBeforeParens(Right);
3548 if (Left.isOneOf(TT_RequiresClause,
3549 TT_RequiresClauseInARequiresExpression)) {
3550 return Style.SpaceBeforeParensOptions.AfterRequiresInClause ||
3551 spaceRequiredBeforeParens(Right);
3552 }
3553 if (Left.is(TT_RequiresExpression)) {
3554 return Style.SpaceBeforeParensOptions.AfterRequiresInExpression ||
3555 spaceRequiredBeforeParens(Right);
3556 }
3557 if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) ||
3558 (Left.is(tok::r_square) && Left.is(TT_AttributeSquare))) {
3559 return true;
3560 }
3561 if (Left.is(TT_ForEachMacro)) {
3562 return Style.SpaceBeforeParensOptions.AfterForeachMacros ||
3563 spaceRequiredBeforeParens(Right);
3564 }
3565 if (Left.is(TT_IfMacro)) {
3566 return Style.SpaceBeforeParensOptions.AfterIfMacros ||
3567 spaceRequiredBeforeParens(Right);
3568 }
3569 if (Line.Type == LT_ObjCDecl)
3570 return true;
3571 if (Left.is(tok::semi))
3572 return true;
3573 if (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while, tok::kw_switch,
3574 tok::kw_case, TT_ForEachMacro, TT_ObjCForIn) ||
3575 Left.isIf(Line.Type != LT_PreprocessorDirective)) {
3576 return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3577 spaceRequiredBeforeParens(Right);
3578 }
3579
3580 // TODO add Operator overloading specific Options to
3581 // SpaceBeforeParensOptions
3582 if (Right.is(TT_OverloadedOperatorLParen))
3583 return spaceRequiredBeforeParens(Right);
3584 // Function declaration or definition
3585 if (Line.MightBeFunctionDecl && (Left.is(TT_FunctionDeclarationName))) {
3586 if (Line.mightBeFunctionDefinition()) {
3587 return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName ||
3588 spaceRequiredBeforeParens(Right);
3589 } else {
3590 return Style.SpaceBeforeParensOptions.AfterFunctionDeclarationName ||
3591 spaceRequiredBeforeParens(Right);
3592 }
3593 }
3594 // Lambda
3595 if (Line.Type != LT_PreprocessorDirective && Left.is(tok::r_square) &&
3596 Left.MatchingParen && Left.MatchingParen->is(TT_LambdaLSquare)) {
3597 return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName ||
3598 spaceRequiredBeforeParens(Right);
3599 }
3600 if (!Left.Previous || Left.Previous->isNot(tok::period)) {
3601 if (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch)) {
3602 return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3603 spaceRequiredBeforeParens(Right);
3604 }
3605 if (Left.isOneOf(tok::kw_new, tok::kw_delete)) {
3606 return ((!Line.MightBeFunctionDecl || !Left.Previous) &&
3607 Style.SpaceBeforeParens != FormatStyle::SBPO_Never) ||
3608 spaceRequiredBeforeParens(Right);
3609 }
3610
3611 if (Left.is(tok::r_square) && Left.MatchingParen &&
3612 Left.MatchingParen->Previous &&
3613 Left.MatchingParen->Previous->is(tok::kw_delete)) {
3614 return (Style.SpaceBeforeParens != FormatStyle::SBPO_Never) ||
3615 spaceRequiredBeforeParens(Right);
3616 }
3617 }
3618 // Handle builtins like identifiers.
3619 if (Line.Type != LT_PreprocessorDirective &&
3620 (Left.Tok.getIdentifierInfo() || Left.is(tok::r_paren))) {
3621 return spaceRequiredBeforeParens(Right);
3622 }
3623 return false;
3624 }
3625 if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
3626 return false;
3627 if (Right.is(TT_UnaryOperator)) {
3628 return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
3629 (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
3630 }
3631 if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
3632 tok::r_paren) ||
3633 Left.isSimpleTypeSpecifier()) &&
3634 Right.is(tok::l_brace) && Right.getNextNonComment() &&
3635 Right.isNot(BK_Block)) {
3636 return false;
3637 }
3638 if (Left.is(tok::period) || Right.is(tok::period))
3639 return false;
3640 // u#str, U#str, L#str, u8#str
3641 // uR#str, UR#str, LR#str, u8R#str
3642 if (Right.is(tok::hash) && Left.is(tok::identifier) &&
3643 (Left.TokenText == "L" || Left.TokenText == "u" ||
3644 Left.TokenText == "U" || Left.TokenText == "u8" ||
3645 Left.TokenText == "LR" || Left.TokenText == "uR" ||
3646 Left.TokenText == "UR" || Left.TokenText == "u8R")) {
3647 return false;
3648 }
3649 if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
3650 Left.MatchingParen->Previous &&
3651 (Left.MatchingParen->Previous->is(tok::period) ||
3652 Left.MatchingParen->Previous->is(tok::coloncolon))) {
3653 // Java call to generic function with explicit type:
3654 // A.<B<C<...>>>DoSomething();
3655 // A::<B<C<...>>>DoSomething(); // With a Java 8 method reference.
3656 return false;
3657 }
3658 if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
3659 return false;
3660 if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at)) {
3661 // Objective-C dictionary literal -> no space after opening brace.
3662 return false;
3663 }
3664 if (Right.is(tok::r_brace) && Right.MatchingParen &&
3665 Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at)) {
3666 // Objective-C dictionary literal -> no space before closing brace.
3667 return false;
3668 }
3669 if (Right.getType() == TT_TrailingAnnotation &&
3670 Right.isOneOf(tok::amp, tok::ampamp) &&
3671 Left.isOneOf(tok::kw_const, tok::kw_volatile) &&
3672 (!Right.Next || Right.Next->is(tok::semi))) {
3673 // Match const and volatile ref-qualifiers without any additional
3674 // qualifiers such as
3675 // void Fn() const &;
3676 return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left;
3677 }
3678
3679 return true;
3680 }
3681
spaceRequiredBefore(const AnnotatedLine & Line,const FormatToken & Right) const3682 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
3683 const FormatToken &Right) const {
3684 const FormatToken &Left = *Right.Previous;
3685
3686 // If the token is finalized don't touch it (as it could be in a
3687 // clang-format-off section).
3688 if (Left.Finalized)
3689 return Right.hasWhitespaceBefore();
3690
3691 // Never ever merge two words.
3692 if (Keywords.isWordLike(Right) && Keywords.isWordLike(Left))
3693 return true;
3694
3695 // Leave a space between * and /* to avoid C4138 `comment end` found outside
3696 // of comment.
3697 if (Left.is(tok::star) && Right.is(tok::comment))
3698 return true;
3699
3700 if (Style.isCpp()) {
3701 // Space between import <iostream>.
3702 // or import .....;
3703 if (Left.is(Keywords.kw_import) && Right.isOneOf(tok::less, tok::ellipsis))
3704 return true;
3705 // Space between `module :` and `import :`.
3706 if (Left.isOneOf(Keywords.kw_module, Keywords.kw_import) &&
3707 Right.is(TT_ModulePartitionColon)) {
3708 return true;
3709 }
3710 // No space between import foo:bar but keep a space between import :bar;
3711 if (Left.is(tok::identifier) && Right.is(TT_ModulePartitionColon))
3712 return false;
3713 // No space between :bar;
3714 if (Left.is(TT_ModulePartitionColon) &&
3715 Right.isOneOf(tok::identifier, tok::kw_private)) {
3716 return false;
3717 }
3718 if (Left.is(tok::ellipsis) && Right.is(tok::identifier) &&
3719 Line.First->is(Keywords.kw_import)) {
3720 return false;
3721 }
3722 // Space in __attribute__((attr)) ::type.
3723 if (Left.is(TT_AttributeParen) && Right.is(tok::coloncolon))
3724 return true;
3725
3726 if (Left.is(tok::kw_operator))
3727 return Right.is(tok::coloncolon);
3728 if (Right.is(tok::l_brace) && Right.is(BK_BracedInit) &&
3729 !Left.opensScope() && Style.SpaceBeforeCpp11BracedList) {
3730 return true;
3731 }
3732 if (Left.is(tok::less) && Left.is(TT_OverloadedOperator) &&
3733 Right.is(TT_TemplateOpener)) {
3734 return true;
3735 }
3736 } else if (Style.Language == FormatStyle::LK_Proto ||
3737 Style.Language == FormatStyle::LK_TextProto) {
3738 if (Right.is(tok::period) &&
3739 Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
3740 Keywords.kw_repeated, Keywords.kw_extend)) {
3741 return true;
3742 }
3743 if (Right.is(tok::l_paren) &&
3744 Left.isOneOf(Keywords.kw_returns, Keywords.kw_option)) {
3745 return true;
3746 }
3747 if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName))
3748 return true;
3749 // Slashes occur in text protocol extension syntax: [type/type] { ... }.
3750 if (Left.is(tok::slash) || Right.is(tok::slash))
3751 return false;
3752 if (Left.MatchingParen &&
3753 Left.MatchingParen->is(TT_ProtoExtensionLSquare) &&
3754 Right.isOneOf(tok::l_brace, tok::less)) {
3755 return !Style.Cpp11BracedListStyle;
3756 }
3757 // A percent is probably part of a formatting specification, such as %lld.
3758 if (Left.is(tok::percent))
3759 return false;
3760 // Preserve the existence of a space before a percent for cases like 0x%04x
3761 // and "%d %d"
3762 if (Left.is(tok::numeric_constant) && Right.is(tok::percent))
3763 return Right.hasWhitespaceBefore();
3764 } else if (Style.isJson()) {
3765 if (Right.is(tok::colon))
3766 return false;
3767 } else if (Style.isCSharp()) {
3768 // Require spaces around '{' and before '}' unless they appear in
3769 // interpolated strings. Interpolated strings are merged into a single token
3770 // so cannot have spaces inserted by this function.
3771
3772 // No space between 'this' and '['
3773 if (Left.is(tok::kw_this) && Right.is(tok::l_square))
3774 return false;
3775
3776 // No space between 'new' and '('
3777 if (Left.is(tok::kw_new) && Right.is(tok::l_paren))
3778 return false;
3779
3780 // Space before { (including space within '{ {').
3781 if (Right.is(tok::l_brace))
3782 return true;
3783
3784 // Spaces inside braces.
3785 if (Left.is(tok::l_brace) && Right.isNot(tok::r_brace))
3786 return true;
3787
3788 if (Left.isNot(tok::l_brace) && Right.is(tok::r_brace))
3789 return true;
3790
3791 // Spaces around '=>'.
3792 if (Left.is(TT_FatArrow) || Right.is(TT_FatArrow))
3793 return true;
3794
3795 // No spaces around attribute target colons
3796 if (Left.is(TT_AttributeColon) || Right.is(TT_AttributeColon))
3797 return false;
3798
3799 // space between type and variable e.g. Dictionary<string,string> foo;
3800 if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName))
3801 return true;
3802
3803 // spaces inside square brackets.
3804 if (Left.is(tok::l_square) || Right.is(tok::r_square))
3805 return Style.SpacesInSquareBrackets;
3806
3807 // No space before ? in nullable types.
3808 if (Right.is(TT_CSharpNullable))
3809 return false;
3810
3811 // No space before null forgiving '!'.
3812 if (Right.is(TT_NonNullAssertion))
3813 return false;
3814
3815 // No space between consecutive commas '[,,]'.
3816 if (Left.is(tok::comma) && Right.is(tok::comma))
3817 return false;
3818
3819 // space after var in `var (key, value)`
3820 if (Left.is(Keywords.kw_var) && Right.is(tok::l_paren))
3821 return true;
3822
3823 // space between keywords and paren e.g. "using ("
3824 if (Right.is(tok::l_paren)) {
3825 if (Left.isOneOf(tok::kw_using, Keywords.kw_async, Keywords.kw_when,
3826 Keywords.kw_lock)) {
3827 return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3828 spaceRequiredBeforeParens(Right);
3829 }
3830 }
3831
3832 // space between method modifier and opening parenthesis of a tuple return
3833 // type
3834 if (Left.isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
3835 tok::kw_virtual, tok::kw_extern, tok::kw_static,
3836 Keywords.kw_internal, Keywords.kw_abstract,
3837 Keywords.kw_sealed, Keywords.kw_override,
3838 Keywords.kw_async, Keywords.kw_unsafe) &&
3839 Right.is(tok::l_paren)) {
3840 return true;
3841 }
3842 } else if (Style.isJavaScript()) {
3843 if (Left.is(TT_FatArrow))
3844 return true;
3845 // for await ( ...
3846 if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous &&
3847 Left.Previous->is(tok::kw_for)) {
3848 return true;
3849 }
3850 if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
3851 Right.MatchingParen) {
3852 const FormatToken *Next = Right.MatchingParen->getNextNonComment();
3853 // An async arrow function, for example: `x = async () => foo();`,
3854 // as opposed to calling a function called async: `x = async();`
3855 if (Next && Next->is(TT_FatArrow))
3856 return true;
3857 }
3858 if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
3859 (Right.is(TT_TemplateString) && Right.TokenText.startswith("}"))) {
3860 return false;
3861 }
3862 // In tagged template literals ("html`bar baz`"), there is no space between
3863 // the tag identifier and the template string.
3864 if (Keywords.IsJavaScriptIdentifier(Left,
3865 /* AcceptIdentifierName= */ false) &&
3866 Right.is(TT_TemplateString)) {
3867 return false;
3868 }
3869 if (Right.is(tok::star) &&
3870 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield)) {
3871 return false;
3872 }
3873 if (Right.isOneOf(tok::l_brace, tok::l_square) &&
3874 Left.isOneOf(Keywords.kw_function, Keywords.kw_yield,
3875 Keywords.kw_extends, Keywords.kw_implements)) {
3876 return true;
3877 }
3878 if (Right.is(tok::l_paren)) {
3879 // JS methods can use some keywords as names (e.g. `delete()`).
3880 if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo())
3881 return false;
3882 // Valid JS method names can include keywords, e.g. `foo.delete()` or
3883 // `bar.instanceof()`. Recognize call positions by preceding period.
3884 if (Left.Previous && Left.Previous->is(tok::period) &&
3885 Left.Tok.getIdentifierInfo()) {
3886 return false;
3887 }
3888 // Additional unary JavaScript operators that need a space after.
3889 if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof,
3890 tok::kw_void)) {
3891 return true;
3892 }
3893 }
3894 // `foo as const;` casts into a const type.
3895 if (Left.endsSequence(tok::kw_const, Keywords.kw_as))
3896 return false;
3897 if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
3898 tok::kw_const) ||
3899 // "of" is only a keyword if it appears after another identifier
3900 // (e.g. as "const x of y" in a for loop), or after a destructuring
3901 // operation (const [x, y] of z, const {a, b} of c).
3902 (Left.is(Keywords.kw_of) && Left.Previous &&
3903 (Left.Previous->is(tok::identifier) ||
3904 Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) &&
3905 (!Left.Previous || !Left.Previous->is(tok::period))) {
3906 return true;
3907 }
3908 if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
3909 Left.Previous->is(tok::period) && Right.is(tok::l_paren)) {
3910 return false;
3911 }
3912 if (Left.is(Keywords.kw_as) &&
3913 Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren)) {
3914 return true;
3915 }
3916 if (Left.is(tok::kw_default) && Left.Previous &&
3917 Left.Previous->is(tok::kw_export)) {
3918 return true;
3919 }
3920 if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
3921 return true;
3922 if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
3923 return false;
3924 if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
3925 return false;
3926 if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
3927 Line.First->isOneOf(Keywords.kw_import, tok::kw_export)) {
3928 return false;
3929 }
3930 if (Left.is(tok::ellipsis))
3931 return false;
3932 if (Left.is(TT_TemplateCloser) &&
3933 !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
3934 Keywords.kw_implements, Keywords.kw_extends)) {
3935 // Type assertions ('<type>expr') are not followed by whitespace. Other
3936 // locations that should have whitespace following are identified by the
3937 // above set of follower tokens.
3938 return false;
3939 }
3940 if (Right.is(TT_NonNullAssertion))
3941 return false;
3942 if (Left.is(TT_NonNullAssertion) &&
3943 Right.isOneOf(Keywords.kw_as, Keywords.kw_in)) {
3944 return true; // "x! as string", "x! in y"
3945 }
3946 } else if (Style.Language == FormatStyle::LK_Java) {
3947 if (Left.is(tok::r_square) && Right.is(tok::l_brace))
3948 return true;
3949 if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren)) {
3950 return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3951 spaceRequiredBeforeParens(Right);
3952 }
3953 if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
3954 tok::kw_protected) ||
3955 Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
3956 Keywords.kw_native)) &&
3957 Right.is(TT_TemplateOpener)) {
3958 return true;
3959 }
3960 } else if (Style.isVerilog()) {
3961 // Don't add space within a delay like `#0`.
3962 if (!Left.is(TT_BinaryOperator) &&
3963 Left.isOneOf(Keywords.kw_verilogHash, Keywords.kw_verilogHashHash)) {
3964 return false;
3965 }
3966 // Add space after a delay.
3967 if (!Right.is(tok::semi) &&
3968 (Left.endsSequence(tok::numeric_constant, Keywords.kw_verilogHash) ||
3969 Left.endsSequence(tok::numeric_constant,
3970 Keywords.kw_verilogHashHash) ||
3971 (Left.is(tok::r_paren) && Left.MatchingParen &&
3972 Left.MatchingParen->endsSequence(tok::l_paren, tok::at)))) {
3973 return true;
3974 }
3975 }
3976 if (Left.is(TT_ImplicitStringLiteral))
3977 return Right.hasWhitespaceBefore();
3978 if (Line.Type == LT_ObjCMethodDecl) {
3979 if (Left.is(TT_ObjCMethodSpecifier))
3980 return true;
3981 if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right)) {
3982 // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a
3983 // keyword in Objective-C, and '+ (instancetype)new;' is a standard class
3984 // method declaration.
3985 return false;
3986 }
3987 }
3988 if (Line.Type == LT_ObjCProperty &&
3989 (Right.is(tok::equal) || Left.is(tok::equal))) {
3990 return false;
3991 }
3992
3993 if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
3994 Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow)) {
3995 return true;
3996 }
3997 if (Left.is(tok::comma) && !Right.is(TT_OverloadedOperatorLParen))
3998 return true;
3999 if (Right.is(tok::comma))
4000 return false;
4001 if (Right.is(TT_ObjCBlockLParen))
4002 return true;
4003 if (Right.is(TT_CtorInitializerColon))
4004 return Style.SpaceBeforeCtorInitializerColon;
4005 if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon)
4006 return false;
4007 if (Right.is(TT_RangeBasedForLoopColon) &&
4008 !Style.SpaceBeforeRangeBasedForLoopColon) {
4009 return false;
4010 }
4011 if (Left.is(TT_BitFieldColon)) {
4012 return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both ||
4013 Style.BitFieldColonSpacing == FormatStyle::BFCS_After;
4014 }
4015 if (Right.is(tok::colon)) {
4016 if (Line.First->isOneOf(tok::kw_default, tok::kw_case))
4017 return Style.SpaceBeforeCaseColon;
4018 const FormatToken *Next = Right.getNextNonComment();
4019 if (!Next || Next->is(tok::semi))
4020 return false;
4021 if (Right.is(TT_ObjCMethodExpr))
4022 return false;
4023 if (Left.is(tok::question))
4024 return false;
4025 if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
4026 return false;
4027 if (Right.is(TT_DictLiteral))
4028 return Style.SpacesInContainerLiterals;
4029 if (Right.is(TT_AttributeColon))
4030 return false;
4031 if (Right.is(TT_CSharpNamedArgumentColon))
4032 return false;
4033 if (Right.is(TT_BitFieldColon)) {
4034 return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both ||
4035 Style.BitFieldColonSpacing == FormatStyle::BFCS_Before;
4036 }
4037 return true;
4038 }
4039 // Do not merge "- -" into "--".
4040 if ((Left.isOneOf(tok::minus, tok::minusminus) &&
4041 Right.isOneOf(tok::minus, tok::minusminus)) ||
4042 (Left.isOneOf(tok::plus, tok::plusplus) &&
4043 Right.isOneOf(tok::plus, tok::plusplus))) {
4044 return true;
4045 }
4046 if (Left.is(TT_UnaryOperator)) {
4047 if (!Right.is(tok::l_paren)) {
4048 // The alternative operators for ~ and ! are "compl" and "not".
4049 // If they are used instead, we do not want to combine them with
4050 // the token to the right, unless that is a left paren.
4051 if (Left.is(tok::exclaim) && Left.TokenText == "not")
4052 return true;
4053 if (Left.is(tok::tilde) && Left.TokenText == "compl")
4054 return true;
4055 // Lambda captures allow for a lone &, so "&]" needs to be properly
4056 // handled.
4057 if (Left.is(tok::amp) && Right.is(tok::r_square))
4058 return Style.SpacesInSquareBrackets;
4059 }
4060 return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) ||
4061 Right.is(TT_BinaryOperator);
4062 }
4063
4064 // If the next token is a binary operator or a selector name, we have
4065 // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
4066 if (Left.is(TT_CastRParen)) {
4067 return Style.SpaceAfterCStyleCast ||
4068 Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
4069 }
4070
4071 auto ShouldAddSpacesInAngles = [this, &Right]() {
4072 if (this->Style.SpacesInAngles == FormatStyle::SIAS_Always)
4073 return true;
4074 if (this->Style.SpacesInAngles == FormatStyle::SIAS_Leave)
4075 return Right.hasWhitespaceBefore();
4076 return false;
4077 };
4078
4079 if (Left.is(tok::greater) && Right.is(tok::greater)) {
4080 if (Style.Language == FormatStyle::LK_TextProto ||
4081 (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral))) {
4082 return !Style.Cpp11BracedListStyle;
4083 }
4084 return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
4085 ((Style.Standard < FormatStyle::LS_Cpp11) ||
4086 ShouldAddSpacesInAngles());
4087 }
4088 if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) ||
4089 Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
4090 (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod))) {
4091 return false;
4092 }
4093 if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) &&
4094 Right.getPrecedence() == prec::Assignment) {
4095 return false;
4096 }
4097 if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) &&
4098 (Left.is(tok::identifier) || Left.is(tok::kw_this))) {
4099 return false;
4100 }
4101 if (Right.is(tok::coloncolon) && Left.is(tok::identifier)) {
4102 // Generally don't remove existing spaces between an identifier and "::".
4103 // The identifier might actually be a macro name such as ALWAYS_INLINE. If
4104 // this turns out to be too lenient, add analysis of the identifier itself.
4105 return Right.hasWhitespaceBefore();
4106 }
4107 if (Right.is(tok::coloncolon) &&
4108 !Left.isOneOf(tok::l_brace, tok::comment, tok::l_paren)) {
4109 // Put a space between < and :: in vector< ::std::string >
4110 return (Left.is(TT_TemplateOpener) &&
4111 ((Style.Standard < FormatStyle::LS_Cpp11) ||
4112 ShouldAddSpacesInAngles())) ||
4113 !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
4114 tok::kw___super, TT_TemplateOpener,
4115 TT_TemplateCloser)) ||
4116 (Left.is(tok::l_paren) && Style.SpacesInParentheses);
4117 }
4118 if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
4119 return ShouldAddSpacesInAngles();
4120 // Space before TT_StructuredBindingLSquare.
4121 if (Right.is(TT_StructuredBindingLSquare)) {
4122 return !Left.isOneOf(tok::amp, tok::ampamp) ||
4123 getTokenReferenceAlignment(Left) != FormatStyle::PAS_Right;
4124 }
4125 // Space before & or && following a TT_StructuredBindingLSquare.
4126 if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) &&
4127 Right.isOneOf(tok::amp, tok::ampamp)) {
4128 return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left;
4129 }
4130 if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
4131 (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
4132 !Right.is(tok::r_paren))) {
4133 return true;
4134 }
4135 if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
4136 Left.MatchingParen &&
4137 Left.MatchingParen->is(TT_OverloadedOperatorLParen)) {
4138 return false;
4139 }
4140 if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
4141 Line.startsWith(tok::hash)) {
4142 return true;
4143 }
4144 if (Right.is(TT_TrailingUnaryOperator))
4145 return false;
4146 if (Left.is(TT_RegexLiteral))
4147 return false;
4148 return spaceRequiredBetween(Line, Left, Right);
4149 }
4150
4151 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
isAllmanBrace(const FormatToken & Tok)4152 static bool isAllmanBrace(const FormatToken &Tok) {
4153 return Tok.is(tok::l_brace) && Tok.is(BK_Block) &&
4154 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral);
4155 }
4156
4157 // Returns 'true' if 'Tok' is a function argument.
IsFunctionArgument(const FormatToken & Tok)4158 static bool IsFunctionArgument(const FormatToken &Tok) {
4159 return Tok.MatchingParen && Tok.MatchingParen->Next &&
4160 Tok.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren);
4161 }
4162
4163 static bool
isItAnEmptyLambdaAllowed(const FormatToken & Tok,FormatStyle::ShortLambdaStyle ShortLambdaOption)4164 isItAnEmptyLambdaAllowed(const FormatToken &Tok,
4165 FormatStyle::ShortLambdaStyle ShortLambdaOption) {
4166 return Tok.Children.empty() && ShortLambdaOption != FormatStyle::SLS_None;
4167 }
4168
isAllmanLambdaBrace(const FormatToken & Tok)4169 static bool isAllmanLambdaBrace(const FormatToken &Tok) {
4170 return Tok.is(tok::l_brace) && Tok.is(BK_Block) &&
4171 !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral);
4172 }
4173
4174 // Returns the first token on the line that is not a comment.
getFirstNonComment(const AnnotatedLine & Line)4175 static const FormatToken *getFirstNonComment(const AnnotatedLine &Line) {
4176 const FormatToken *Next = Line.First;
4177 if (!Next)
4178 return Next;
4179 if (Next->is(tok::comment))
4180 Next = Next->getNextNonComment();
4181 return Next;
4182 }
4183
mustBreakBefore(const AnnotatedLine & Line,const FormatToken & Right) const4184 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
4185 const FormatToken &Right) const {
4186 const FormatToken &Left = *Right.Previous;
4187 if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
4188 return true;
4189
4190 if (Style.isCSharp()) {
4191 if (Left.is(TT_FatArrow) && Right.is(tok::l_brace) &&
4192 Style.BraceWrapping.AfterFunction) {
4193 return true;
4194 }
4195 if (Right.is(TT_CSharpNamedArgumentColon) ||
4196 Left.is(TT_CSharpNamedArgumentColon)) {
4197 return false;
4198 }
4199 if (Right.is(TT_CSharpGenericTypeConstraint))
4200 return true;
4201 if (Right.Next && Right.Next->is(TT_FatArrow) &&
4202 (Right.is(tok::numeric_constant) ||
4203 (Right.is(tok::identifier) && Right.TokenText == "_"))) {
4204 return true;
4205 }
4206
4207 // Break after C# [...] and before public/protected/private/internal.
4208 if (Left.is(TT_AttributeSquare) && Left.is(tok::r_square) &&
4209 (Right.isAccessSpecifier(/*ColonRequired=*/false) ||
4210 Right.is(Keywords.kw_internal))) {
4211 return true;
4212 }
4213 // Break between ] and [ but only when there are really 2 attributes.
4214 if (Left.is(TT_AttributeSquare) && Right.is(TT_AttributeSquare) &&
4215 Left.is(tok::r_square) && Right.is(tok::l_square)) {
4216 return true;
4217 }
4218
4219 } else if (Style.isJavaScript()) {
4220 // FIXME: This might apply to other languages and token kinds.
4221 if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
4222 Left.Previous->is(tok::string_literal)) {
4223 return true;
4224 }
4225 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
4226 Left.Previous && Left.Previous->is(tok::equal) &&
4227 Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
4228 tok::kw_const) &&
4229 // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
4230 // above.
4231 !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let)) {
4232 // Object literals on the top level of a file are treated as "enum-style".
4233 // Each key/value pair is put on a separate line, instead of bin-packing.
4234 return true;
4235 }
4236 if (Left.is(tok::l_brace) && Line.Level == 0 &&
4237 (Line.startsWith(tok::kw_enum) ||
4238 Line.startsWith(tok::kw_const, tok::kw_enum) ||
4239 Line.startsWith(tok::kw_export, tok::kw_enum) ||
4240 Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum))) {
4241 // JavaScript top-level enum key/value pairs are put on separate lines
4242 // instead of bin-packing.
4243 return true;
4244 }
4245 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && Left.Previous &&
4246 Left.Previous->is(TT_FatArrow)) {
4247 // JS arrow function (=> {...}).
4248 switch (Style.AllowShortLambdasOnASingleLine) {
4249 case FormatStyle::SLS_All:
4250 return false;
4251 case FormatStyle::SLS_None:
4252 return true;
4253 case FormatStyle::SLS_Empty:
4254 return !Left.Children.empty();
4255 case FormatStyle::SLS_Inline:
4256 // allow one-lining inline (e.g. in function call args) and empty arrow
4257 // functions.
4258 return (Left.NestingLevel == 0 && Line.Level == 0) &&
4259 !Left.Children.empty();
4260 }
4261 llvm_unreachable("Unknown FormatStyle::ShortLambdaStyle enum");
4262 }
4263
4264 if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
4265 !Left.Children.empty()) {
4266 // Support AllowShortFunctionsOnASingleLine for JavaScript.
4267 return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
4268 Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
4269 (Left.NestingLevel == 0 && Line.Level == 0 &&
4270 Style.AllowShortFunctionsOnASingleLine &
4271 FormatStyle::SFS_InlineOnly);
4272 }
4273 } else if (Style.Language == FormatStyle::LK_Java) {
4274 if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
4275 Right.Next->is(tok::string_literal)) {
4276 return true;
4277 }
4278 } else if (Style.Language == FormatStyle::LK_Cpp ||
4279 Style.Language == FormatStyle::LK_ObjC ||
4280 Style.Language == FormatStyle::LK_Proto ||
4281 Style.Language == FormatStyle::LK_TableGen ||
4282 Style.Language == FormatStyle::LK_TextProto) {
4283 if (Left.isStringLiteral() && Right.isStringLiteral())
4284 return true;
4285 }
4286
4287 // Basic JSON newline processing.
4288 if (Style.isJson()) {
4289 // Always break after a JSON record opener.
4290 // {
4291 // }
4292 if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace))
4293 return true;
4294 // Always break after a JSON array opener.
4295 // [
4296 // ]
4297 if (Left.is(TT_ArrayInitializerLSquare) && Left.is(tok::l_square) &&
4298 !Right.is(tok::r_square)) {
4299 return true;
4300 }
4301 // Always break after successive entries.
4302 // 1,
4303 // 2
4304 if (Left.is(tok::comma))
4305 return true;
4306 }
4307
4308 // If the last token before a '}', ']', or ')' is a comma or a trailing
4309 // comment, the intention is to insert a line break after it in order to make
4310 // shuffling around entries easier. Import statements, especially in
4311 // JavaScript, can be an exception to this rule.
4312 if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) {
4313 const FormatToken *BeforeClosingBrace = nullptr;
4314 if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
4315 (Style.isJavaScript() && Left.is(tok::l_paren))) &&
4316 Left.isNot(BK_Block) && Left.MatchingParen) {
4317 BeforeClosingBrace = Left.MatchingParen->Previous;
4318 } else if (Right.MatchingParen &&
4319 (Right.MatchingParen->isOneOf(tok::l_brace,
4320 TT_ArrayInitializerLSquare) ||
4321 (Style.isJavaScript() &&
4322 Right.MatchingParen->is(tok::l_paren)))) {
4323 BeforeClosingBrace = &Left;
4324 }
4325 if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
4326 BeforeClosingBrace->isTrailingComment())) {
4327 return true;
4328 }
4329 }
4330
4331 if (Right.is(tok::comment)) {
4332 return Left.isNot(BK_BracedInit) && Left.isNot(TT_CtorInitializerColon) &&
4333 (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
4334 }
4335 if (Left.isTrailingComment())
4336 return true;
4337 if (Left.IsUnterminatedLiteral)
4338 return true;
4339 if (Right.is(tok::lessless) && Right.Next && Left.is(tok::string_literal) &&
4340 Right.Next->is(tok::string_literal)) {
4341 return true;
4342 }
4343 if (Right.is(TT_RequiresClause)) {
4344 switch (Style.RequiresClausePosition) {
4345 case FormatStyle::RCPS_OwnLine:
4346 case FormatStyle::RCPS_WithFollowing:
4347 return true;
4348 default:
4349 break;
4350 }
4351 }
4352 // Can break after template<> declaration
4353 if (Left.ClosesTemplateDeclaration && Left.MatchingParen &&
4354 Left.MatchingParen->NestingLevel == 0) {
4355 // Put concepts on the next line e.g.
4356 // template<typename T>
4357 // concept ...
4358 if (Right.is(tok::kw_concept))
4359 return Style.BreakBeforeConceptDeclarations == FormatStyle::BBCDS_Always;
4360 return Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes;
4361 }
4362 if (Left.ClosesRequiresClause && Right.isNot(tok::semi)) {
4363 switch (Style.RequiresClausePosition) {
4364 case FormatStyle::RCPS_OwnLine:
4365 case FormatStyle::RCPS_WithPreceding:
4366 return true;
4367 default:
4368 break;
4369 }
4370 }
4371 if (Style.PackConstructorInitializers == FormatStyle::PCIS_Never) {
4372 if (Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeColon &&
4373 (Left.is(TT_CtorInitializerComma) ||
4374 Right.is(TT_CtorInitializerColon))) {
4375 return true;
4376 }
4377
4378 if (Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon &&
4379 Left.isOneOf(TT_CtorInitializerColon, TT_CtorInitializerComma)) {
4380 return true;
4381 }
4382 }
4383 if (Style.PackConstructorInitializers < FormatStyle::PCIS_CurrentLine &&
4384 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
4385 Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon)) {
4386 return true;
4387 }
4388 // Break only if we have multiple inheritance.
4389 if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma &&
4390 Right.is(TT_InheritanceComma)) {
4391 return true;
4392 }
4393 if (Style.BreakInheritanceList == FormatStyle::BILS_AfterComma &&
4394 Left.is(TT_InheritanceComma)) {
4395 return true;
4396 }
4397 if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\"")) {
4398 // Multiline raw string literals are special wrt. line breaks. The author
4399 // has made a deliberate choice and might have aligned the contents of the
4400 // string literal accordingly. Thus, we try keep existing line breaks.
4401 return Right.IsMultiline && Right.NewlinesBefore > 0;
4402 }
4403 if ((Left.is(tok::l_brace) || (Left.is(tok::less) && Left.Previous &&
4404 Left.Previous->is(tok::equal))) &&
4405 Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) {
4406 // Don't put enums or option definitions onto single lines in protocol
4407 // buffers.
4408 return true;
4409 }
4410 if (Right.is(TT_InlineASMBrace))
4411 return Right.HasUnescapedNewline;
4412
4413 if (isAllmanBrace(Left) || isAllmanBrace(Right)) {
4414 auto FirstNonComment = getFirstNonComment(Line);
4415 bool AccessSpecifier =
4416 FirstNonComment &&
4417 FirstNonComment->isOneOf(Keywords.kw_internal, tok::kw_public,
4418 tok::kw_private, tok::kw_protected);
4419
4420 if (Style.BraceWrapping.AfterEnum) {
4421 if (Line.startsWith(tok::kw_enum) ||
4422 Line.startsWith(tok::kw_typedef, tok::kw_enum)) {
4423 return true;
4424 }
4425 // Ensure BraceWrapping for `public enum A {`.
4426 if (AccessSpecifier && FirstNonComment->Next &&
4427 FirstNonComment->Next->is(tok::kw_enum)) {
4428 return true;
4429 }
4430 }
4431
4432 // Ensure BraceWrapping for `public interface A {`.
4433 if (Style.BraceWrapping.AfterClass &&
4434 ((AccessSpecifier && FirstNonComment->Next &&
4435 FirstNonComment->Next->is(Keywords.kw_interface)) ||
4436 Line.startsWith(Keywords.kw_interface))) {
4437 return true;
4438 }
4439
4440 return (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
4441 (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
4442 }
4443
4444 if (Left.is(TT_ObjCBlockLBrace) &&
4445 Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never) {
4446 return true;
4447 }
4448
4449 // Ensure wrapping after __attribute__((XX)) and @interface etc.
4450 if (Left.is(TT_AttributeParen) && Right.is(TT_ObjCDecl))
4451 return true;
4452
4453 if (Left.is(TT_LambdaLBrace)) {
4454 if (IsFunctionArgument(Left) &&
4455 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline) {
4456 return false;
4457 }
4458
4459 if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None ||
4460 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline ||
4461 (!Left.Children.empty() &&
4462 Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty)) {
4463 return true;
4464 }
4465 }
4466
4467 if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace) &&
4468 Left.isOneOf(tok::star, tok::amp, tok::ampamp, TT_TemplateCloser)) {
4469 return true;
4470 }
4471
4472 // Put multiple Java annotation on a new line.
4473 if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) &&
4474 Left.is(TT_LeadingJavaAnnotation) &&
4475 Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
4476 (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations)) {
4477 return true;
4478 }
4479
4480 if (Right.is(TT_ProtoExtensionLSquare))
4481 return true;
4482
4483 // In text proto instances if a submessage contains at least 2 entries and at
4484 // least one of them is a submessage, like A { ... B { ... } ... },
4485 // put all of the entries of A on separate lines by forcing the selector of
4486 // the submessage B to be put on a newline.
4487 //
4488 // Example: these can stay on one line:
4489 // a { scalar_1: 1 scalar_2: 2 }
4490 // a { b { key: value } }
4491 //
4492 // and these entries need to be on a new line even if putting them all in one
4493 // line is under the column limit:
4494 // a {
4495 // scalar: 1
4496 // b { key: value }
4497 // }
4498 //
4499 // We enforce this by breaking before a submessage field that has previous
4500 // siblings, *and* breaking before a field that follows a submessage field.
4501 //
4502 // Be careful to exclude the case [proto.ext] { ... } since the `]` is
4503 // the TT_SelectorName there, but we don't want to break inside the brackets.
4504 //
4505 // Another edge case is @submessage { key: value }, which is a common
4506 // substitution placeholder. In this case we want to keep `@` and `submessage`
4507 // together.
4508 //
4509 // We ensure elsewhere that extensions are always on their own line.
4510 if ((Style.Language == FormatStyle::LK_Proto ||
4511 Style.Language == FormatStyle::LK_TextProto) &&
4512 Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) {
4513 // Keep `@submessage` together in:
4514 // @submessage { key: value }
4515 if (Left.is(tok::at))
4516 return false;
4517 // Look for the scope opener after selector in cases like:
4518 // selector { ...
4519 // selector: { ...
4520 // selector: @base { ...
4521 FormatToken *LBrace = Right.Next;
4522 if (LBrace && LBrace->is(tok::colon)) {
4523 LBrace = LBrace->Next;
4524 if (LBrace && LBrace->is(tok::at)) {
4525 LBrace = LBrace->Next;
4526 if (LBrace)
4527 LBrace = LBrace->Next;
4528 }
4529 }
4530 if (LBrace &&
4531 // The scope opener is one of {, [, <:
4532 // selector { ... }
4533 // selector [ ... ]
4534 // selector < ... >
4535 //
4536 // In case of selector { ... }, the l_brace is TT_DictLiteral.
4537 // In case of an empty selector {}, the l_brace is not TT_DictLiteral,
4538 // so we check for immediately following r_brace.
4539 ((LBrace->is(tok::l_brace) &&
4540 (LBrace->is(TT_DictLiteral) ||
4541 (LBrace->Next && LBrace->Next->is(tok::r_brace)))) ||
4542 LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) {
4543 // If Left.ParameterCount is 0, then this submessage entry is not the
4544 // first in its parent submessage, and we want to break before this entry.
4545 // If Left.ParameterCount is greater than 0, then its parent submessage
4546 // might contain 1 or more entries and we want to break before this entry
4547 // if it contains at least 2 entries. We deal with this case later by
4548 // detecting and breaking before the next entry in the parent submessage.
4549 if (Left.ParameterCount == 0)
4550 return true;
4551 // However, if this submessage is the first entry in its parent
4552 // submessage, Left.ParameterCount might be 1 in some cases.
4553 // We deal with this case later by detecting an entry
4554 // following a closing paren of this submessage.
4555 }
4556
4557 // If this is an entry immediately following a submessage, it will be
4558 // preceded by a closing paren of that submessage, like in:
4559 // left---. .---right
4560 // v v
4561 // sub: { ... } key: value
4562 // If there was a comment between `}` an `key` above, then `key` would be
4563 // put on a new line anyways.
4564 if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square))
4565 return true;
4566 }
4567
4568 // Deal with lambda arguments in C++ - we want consistent line breaks whether
4569 // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced
4570 // as aggressive line breaks are placed when the lambda is not the last arg.
4571 if ((Style.Language == FormatStyle::LK_Cpp ||
4572 Style.Language == FormatStyle::LK_ObjC) &&
4573 Left.is(tok::l_paren) && Left.BlockParameterCount > 0 &&
4574 !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) {
4575 // Multiple lambdas in the same function call force line breaks.
4576 if (Left.BlockParameterCount > 1)
4577 return true;
4578
4579 // A lambda followed by another arg forces a line break.
4580 if (!Left.Role)
4581 return false;
4582 auto Comma = Left.Role->lastComma();
4583 if (!Comma)
4584 return false;
4585 auto Next = Comma->getNextNonComment();
4586 if (!Next)
4587 return false;
4588 if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret))
4589 return true;
4590 }
4591
4592 return false;
4593 }
4594
canBreakBefore(const AnnotatedLine & Line,const FormatToken & Right) const4595 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
4596 const FormatToken &Right) const {
4597 const FormatToken &Left = *Right.Previous;
4598 // Language-specific stuff.
4599 if (Style.isCSharp()) {
4600 if (Left.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon) ||
4601 Right.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon)) {
4602 return false;
4603 }
4604 // Only break after commas for generic type constraints.
4605 if (Line.First->is(TT_CSharpGenericTypeConstraint))
4606 return Left.is(TT_CSharpGenericTypeConstraintComma);
4607 // Keep nullable operators attached to their identifiers.
4608 if (Right.is(TT_CSharpNullable))
4609 return false;
4610 } else if (Style.Language == FormatStyle::LK_Java) {
4611 if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
4612 Keywords.kw_implements)) {
4613 return false;
4614 }
4615 if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
4616 Keywords.kw_implements)) {
4617 return true;
4618 }
4619 } else if (Style.isJavaScript()) {
4620 const FormatToken *NonComment = Right.getPreviousNonComment();
4621 if (NonComment &&
4622 NonComment->isOneOf(
4623 tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break,
4624 tok::kw_throw, Keywords.kw_interface, Keywords.kw_type,
4625 tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected,
4626 Keywords.kw_readonly, Keywords.kw_override, Keywords.kw_abstract,
4627 Keywords.kw_get, Keywords.kw_set, Keywords.kw_async,
4628 Keywords.kw_await)) {
4629 return false; // Otherwise automatic semicolon insertion would trigger.
4630 }
4631 if (Right.NestingLevel == 0 &&
4632 (Left.Tok.getIdentifierInfo() ||
4633 Left.isOneOf(tok::r_square, tok::r_paren)) &&
4634 Right.isOneOf(tok::l_square, tok::l_paren)) {
4635 return false; // Otherwise automatic semicolon insertion would trigger.
4636 }
4637 if (NonComment && NonComment->is(tok::identifier) &&
4638 NonComment->TokenText == "asserts") {
4639 return false;
4640 }
4641 if (Left.is(TT_FatArrow) && Right.is(tok::l_brace))
4642 return false;
4643 if (Left.is(TT_JsTypeColon))
4644 return true;
4645 // Don't wrap between ":" and "!" of a strict prop init ("field!: type;").
4646 if (Left.is(tok::exclaim) && Right.is(tok::colon))
4647 return false;
4648 // Look for is type annotations like:
4649 // function f(): a is B { ... }
4650 // Do not break before is in these cases.
4651 if (Right.is(Keywords.kw_is)) {
4652 const FormatToken *Next = Right.getNextNonComment();
4653 // If `is` is followed by a colon, it's likely that it's a dict key, so
4654 // ignore it for this check.
4655 // For example this is common in Polymer:
4656 // Polymer({
4657 // is: 'name',
4658 // ...
4659 // });
4660 if (!Next || !Next->is(tok::colon))
4661 return false;
4662 }
4663 if (Left.is(Keywords.kw_in))
4664 return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
4665 if (Right.is(Keywords.kw_in))
4666 return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
4667 if (Right.is(Keywords.kw_as))
4668 return false; // must not break before as in 'x as type' casts
4669 if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) {
4670 // extends and infer can appear as keywords in conditional types:
4671 // https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types
4672 // do not break before them, as the expressions are subject to ASI.
4673 return false;
4674 }
4675 if (Left.is(Keywords.kw_as))
4676 return true;
4677 if (Left.is(TT_NonNullAssertion))
4678 return true;
4679 if (Left.is(Keywords.kw_declare) &&
4680 Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
4681 Keywords.kw_function, tok::kw_class, tok::kw_enum,
4682 Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
4683 Keywords.kw_let, tok::kw_const)) {
4684 // See grammar for 'declare' statements at:
4685 // https://github.com/Microsoft/TypeScript/blob/main/doc/spec-ARCHIVED.md#A.10
4686 return false;
4687 }
4688 if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
4689 Right.isOneOf(tok::identifier, tok::string_literal)) {
4690 return false; // must not break in "module foo { ...}"
4691 }
4692 if (Right.is(TT_TemplateString) && Right.closesScope())
4693 return false;
4694 // Don't split tagged template literal so there is a break between the tag
4695 // identifier and template string.
4696 if (Left.is(tok::identifier) && Right.is(TT_TemplateString))
4697 return false;
4698 if (Left.is(TT_TemplateString) && Left.opensScope())
4699 return true;
4700 }
4701
4702 if (Left.is(tok::at))
4703 return false;
4704 if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
4705 return false;
4706 if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
4707 return !Right.is(tok::l_paren);
4708 if (Right.is(TT_PointerOrReference)) {
4709 return Line.IsMultiVariableDeclStmt ||
4710 (getTokenPointerOrReferenceAlignment(Right) ==
4711 FormatStyle::PAS_Right &&
4712 (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
4713 }
4714 if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
4715 Right.is(tok::kw_operator)) {
4716 return true;
4717 }
4718 if (Left.is(TT_PointerOrReference))
4719 return false;
4720 if (Right.isTrailingComment()) {
4721 // We rely on MustBreakBefore being set correctly here as we should not
4722 // change the "binding" behavior of a comment.
4723 // The first comment in a braced lists is always interpreted as belonging to
4724 // the first list element. Otherwise, it should be placed outside of the
4725 // list.
4726 return Left.is(BK_BracedInit) ||
4727 (Left.is(TT_CtorInitializerColon) && Right.NewlinesBefore > 0 &&
4728 Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon);
4729 }
4730 if (Left.is(tok::question) && Right.is(tok::colon))
4731 return false;
4732 if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
4733 return Style.BreakBeforeTernaryOperators;
4734 if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
4735 return !Style.BreakBeforeTernaryOperators;
4736 if (Left.is(TT_InheritanceColon))
4737 return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon;
4738 if (Right.is(TT_InheritanceColon))
4739 return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon;
4740 if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
4741 Left.isNot(TT_SelectorName)) {
4742 return true;
4743 }
4744
4745 if (Right.is(tok::colon) &&
4746 !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon)) {
4747 return false;
4748 }
4749 if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
4750 if (Style.Language == FormatStyle::LK_Proto ||
4751 Style.Language == FormatStyle::LK_TextProto) {
4752 if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral())
4753 return false;
4754 // Prevent cases like:
4755 //
4756 // submessage:
4757 // { key: valueeeeeeeeeeee }
4758 //
4759 // when the snippet does not fit into one line.
4760 // Prefer:
4761 //
4762 // submessage: {
4763 // key: valueeeeeeeeeeee
4764 // }
4765 //
4766 // instead, even if it is longer by one line.
4767 //
4768 // Note that this allows allows the "{" to go over the column limit
4769 // when the column limit is just between ":" and "{", but that does
4770 // not happen too often and alternative formattings in this case are
4771 // not much better.
4772 //
4773 // The code covers the cases:
4774 //
4775 // submessage: { ... }
4776 // submessage: < ... >
4777 // repeated: [ ... ]
4778 if (((Right.is(tok::l_brace) || Right.is(tok::less)) &&
4779 Right.is(TT_DictLiteral)) ||
4780 Right.is(TT_ArrayInitializerLSquare)) {
4781 return false;
4782 }
4783 }
4784 return true;
4785 }
4786 if (Right.is(tok::r_square) && Right.MatchingParen &&
4787 Right.MatchingParen->is(TT_ProtoExtensionLSquare)) {
4788 return false;
4789 }
4790 if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
4791 Right.Next->is(TT_ObjCMethodExpr))) {
4792 return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
4793 }
4794 if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
4795 return true;
4796 if (Right.is(tok::kw_concept))
4797 return Style.BreakBeforeConceptDeclarations != FormatStyle::BBCDS_Never;
4798 if (Right.is(TT_RequiresClause))
4799 return true;
4800 if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
4801 return true;
4802 if (Left.ClosesRequiresClause)
4803 return true;
4804 if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
4805 TT_OverloadedOperator)) {
4806 return false;
4807 }
4808 if (Left.is(TT_RangeBasedForLoopColon))
4809 return true;
4810 if (Right.is(TT_RangeBasedForLoopColon))
4811 return false;
4812 if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
4813 return true;
4814 if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
4815 (Left.is(tok::less) && Right.is(tok::less))) {
4816 return false;
4817 }
4818 if (Right.is(TT_BinaryOperator) &&
4819 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
4820 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
4821 Right.getPrecedence() != prec::Assignment)) {
4822 return true;
4823 }
4824 if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
4825 Left.is(tok::kw_operator)) {
4826 return false;
4827 }
4828 if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
4829 Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0) {
4830 return false;
4831 }
4832 if (Left.is(tok::equal) && Right.is(tok::l_brace) &&
4833 !Style.Cpp11BracedListStyle) {
4834 return false;
4835 }
4836 if (Left.is(tok::l_paren) &&
4837 Left.isOneOf(TT_AttributeParen, TT_TypeDeclarationParen)) {
4838 return false;
4839 }
4840 if (Left.is(tok::l_paren) && Left.Previous &&
4841 (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen))) {
4842 return false;
4843 }
4844 if (Right.is(TT_ImplicitStringLiteral))
4845 return false;
4846
4847 if (Right.is(TT_TemplateCloser))
4848 return false;
4849 if (Right.is(tok::r_square) && Right.MatchingParen &&
4850 Right.MatchingParen->is(TT_LambdaLSquare)) {
4851 return false;
4852 }
4853
4854 // We only break before r_brace if there was a corresponding break before
4855 // the l_brace, which is tracked by BreakBeforeClosingBrace.
4856 if (Right.is(tok::r_brace))
4857 return Right.MatchingParen && Right.MatchingParen->is(BK_Block);
4858
4859 // We only break before r_paren if we're in a block indented context.
4860 if (Right.is(tok::r_paren)) {
4861 if (Style.AlignAfterOpenBracket != FormatStyle::BAS_BlockIndent ||
4862 !Right.MatchingParen) {
4863 return false;
4864 }
4865 const FormatToken *Previous = Right.MatchingParen->Previous;
4866 return !(Previous && (Previous->is(tok::kw_for) || Previous->isIf()));
4867 }
4868
4869 // Allow breaking after a trailing annotation, e.g. after a method
4870 // declaration.
4871 if (Left.is(TT_TrailingAnnotation)) {
4872 return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
4873 tok::less, tok::coloncolon);
4874 }
4875
4876 if (Right.is(tok::kw___attribute) ||
4877 (Right.is(tok::l_square) && Right.is(TT_AttributeSquare))) {
4878 return !Left.is(TT_AttributeSquare);
4879 }
4880
4881 if (Left.is(tok::identifier) && Right.is(tok::string_literal))
4882 return true;
4883
4884 if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
4885 return true;
4886
4887 if (Left.is(TT_CtorInitializerColon)) {
4888 return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon &&
4889 (!Right.isTrailingComment() || Right.NewlinesBefore > 0);
4890 }
4891 if (Right.is(TT_CtorInitializerColon))
4892 return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon;
4893 if (Left.is(TT_CtorInitializerComma) &&
4894 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) {
4895 return false;
4896 }
4897 if (Right.is(TT_CtorInitializerComma) &&
4898 Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma) {
4899 return true;
4900 }
4901 if (Left.is(TT_InheritanceComma) &&
4902 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) {
4903 return false;
4904 }
4905 if (Right.is(TT_InheritanceComma) &&
4906 Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma) {
4907 return true;
4908 }
4909 if (Left.is(TT_ArrayInitializerLSquare))
4910 return true;
4911 if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
4912 return true;
4913 if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
4914 !Left.isOneOf(tok::arrowstar, tok::lessless) &&
4915 Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
4916 (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
4917 Left.getPrecedence() == prec::Assignment)) {
4918 return true;
4919 }
4920 if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) ||
4921 (Left.is(tok::r_square) && Right.is(TT_AttributeSquare))) {
4922 return false;
4923 }
4924
4925 auto ShortLambdaOption = Style.AllowShortLambdasOnASingleLine;
4926 if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace)) {
4927 if (isAllmanLambdaBrace(Left))
4928 return !isItAnEmptyLambdaAllowed(Left, ShortLambdaOption);
4929 if (isAllmanLambdaBrace(Right))
4930 return !isItAnEmptyLambdaAllowed(Right, ShortLambdaOption);
4931 }
4932
4933 return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
4934 tok::kw_class, tok::kw_struct, tok::comment) ||
4935 Right.isMemberAccess() ||
4936 Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
4937 tok::colon, tok::l_square, tok::at) ||
4938 (Left.is(tok::r_paren) &&
4939 Right.isOneOf(tok::identifier, tok::kw_const)) ||
4940 (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
4941 (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
4942 }
4943
printDebugInfo(const AnnotatedLine & Line) const4944 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) const {
4945 llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n";
4946 const FormatToken *Tok = Line.First;
4947 while (Tok) {
4948 llvm::errs() << " M=" << Tok->MustBreakBefore
4949 << " C=" << Tok->CanBreakBefore
4950 << " T=" << getTokenTypeName(Tok->getType())
4951 << " S=" << Tok->SpacesRequiredBefore
4952 << " F=" << Tok->Finalized << " B=" << Tok->BlockParameterCount
4953 << " BK=" << Tok->getBlockKind() << " P=" << Tok->SplitPenalty
4954 << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength
4955 << " PPK=" << Tok->getPackingKind() << " FakeLParens=";
4956 for (prec::Level LParen : Tok->FakeLParens)
4957 llvm::errs() << LParen << "/";
4958 llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
4959 llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo();
4960 llvm::errs() << " Text='" << Tok->TokenText << "'\n";
4961 if (!Tok->Next)
4962 assert(Tok == Line.Last);
4963 Tok = Tok->Next;
4964 }
4965 llvm::errs() << "----\n";
4966 }
4967
4968 FormatStyle::PointerAlignmentStyle
getTokenReferenceAlignment(const FormatToken & Reference) const4969 TokenAnnotator::getTokenReferenceAlignment(const FormatToken &Reference) const {
4970 assert(Reference.isOneOf(tok::amp, tok::ampamp));
4971 switch (Style.ReferenceAlignment) {
4972 case FormatStyle::RAS_Pointer:
4973 return Style.PointerAlignment;
4974 case FormatStyle::RAS_Left:
4975 return FormatStyle::PAS_Left;
4976 case FormatStyle::RAS_Right:
4977 return FormatStyle::PAS_Right;
4978 case FormatStyle::RAS_Middle:
4979 return FormatStyle::PAS_Middle;
4980 }
4981 assert(0); //"Unhandled value of ReferenceAlignment"
4982 return Style.PointerAlignment;
4983 }
4984
4985 FormatStyle::PointerAlignmentStyle
getTokenPointerOrReferenceAlignment(const FormatToken & PointerOrReference) const4986 TokenAnnotator::getTokenPointerOrReferenceAlignment(
4987 const FormatToken &PointerOrReference) const {
4988 if (PointerOrReference.isOneOf(tok::amp, tok::ampamp)) {
4989 switch (Style.ReferenceAlignment) {
4990 case FormatStyle::RAS_Pointer:
4991 return Style.PointerAlignment;
4992 case FormatStyle::RAS_Left:
4993 return FormatStyle::PAS_Left;
4994 case FormatStyle::RAS_Right:
4995 return FormatStyle::PAS_Right;
4996 case FormatStyle::RAS_Middle:
4997 return FormatStyle::PAS_Middle;
4998 }
4999 }
5000 assert(PointerOrReference.is(tok::star));
5001 return Style.PointerAlignment;
5002 }
5003
5004 } // namespace format
5005 } // namespace clang
5006