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.
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.
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.
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   // Check for `[...]`.
61   return Left->Previous && Left->Previous->is(tok::r_square) &&
62          Left->Previous->MatchingParen &&
63          Left->Previous->MatchingParen->is(TT_LambdaLSquare);
64 }
65 
66 /// Returns \c true if the token is followed by a boolean condition, \c false
67 /// otherwise.
68 static bool isKeywordWithCondition(const FormatToken &Tok) {
69   return Tok.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, tok::kw_switch,
70                      tok::kw_constexpr, tok::kw_catch);
71 }
72 
73 /// A parser that gathers additional information about tokens.
74 ///
75 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
76 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
77 /// into template parameter lists.
78 class AnnotatingParser {
79 public:
80   AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
81                    const AdditionalKeywords &Keywords)
82       : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
83         Keywords(Keywords) {
84     Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
85     resetTokenMetadata();
86   }
87 
88 private:
89   bool parseAngle() {
90     if (!CurrentToken || !CurrentToken->Previous)
91       return false;
92     if (NonTemplateLess.count(CurrentToken->Previous))
93       return false;
94 
95     const FormatToken &Previous = *CurrentToken->Previous; // The '<'.
96     if (Previous.Previous) {
97       if (Previous.Previous->Tok.isLiteral())
98         return false;
99       if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
100           (!Previous.Previous->MatchingParen ||
101            !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen)))
102         return false;
103     }
104 
105     FormatToken *Left = CurrentToken->Previous;
106     Left->ParentBracket = Contexts.back().ContextKind;
107     ScopedContextCreator ContextCreator(*this, tok::less, 12);
108 
109     // If this angle is in the context of an expression, we need to be more
110     // hesitant to detect it as opening template parameters.
111     bool InExprContext = Contexts.back().IsExpression;
112 
113     Contexts.back().IsExpression = false;
114     // If there's a template keyword before the opening angle bracket, this is a
115     // template parameter, not an argument.
116     if (Left->Previous && Left->Previous->isNot(tok::kw_template))
117       Contexts.back().ContextType = Context::TemplateArgument;
118 
119     if (Style.Language == FormatStyle::LK_Java &&
120         CurrentToken->is(tok::question))
121       next();
122 
123     while (CurrentToken) {
124       if (CurrentToken->is(tok::greater)) {
125         // Try to do a better job at looking for ">>" within the condition of
126         // a statement. Conservatively insert spaces between consecutive ">"
127         // tokens to prevent splitting right bitshift operators and potentially
128         // altering program semantics. This check is overly conservative and
129         // will prevent spaces from being inserted in select nested template
130         // parameter cases, but should not alter program semantics.
131         if (CurrentToken->Next && CurrentToken->Next->is(tok::greater) &&
132             Left->ParentBracket != tok::less &&
133             (isKeywordWithCondition(*Line.First) ||
134              CurrentToken->getStartOfNonWhitespace() ==
135                  CurrentToken->Next->getStartOfNonWhitespace().getLocWithOffset(
136                      -1)))
137           return false;
138         Left->MatchingParen = CurrentToken;
139         CurrentToken->MatchingParen = Left;
140         // In TT_Proto, we must distignuish between:
141         //   map<key, value>
142         //   msg < item: data >
143         //   msg: < item: data >
144         // In TT_TextProto, map<key, value> does not occur.
145         if (Style.Language == FormatStyle::LK_TextProto ||
146             (Style.Language == FormatStyle::LK_Proto && Left->Previous &&
147              Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral)))
148           CurrentToken->setType(TT_DictLiteral);
149         else
150           CurrentToken->setType(TT_TemplateCloser);
151         next();
152         return true;
153       }
154       if (CurrentToken->is(tok::question) &&
155           Style.Language == FormatStyle::LK_Java) {
156         next();
157         continue;
158       }
159       if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
160           (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext &&
161            !Style.isCSharp() && Style.Language != FormatStyle::LK_Proto &&
162            Style.Language != FormatStyle::LK_TextProto))
163         return false;
164       // If a && or || is found and interpreted as a binary operator, this set
165       // of angles is likely part of something like "a < b && c > d". If the
166       // angles are inside an expression, the ||/&& might also be a binary
167       // operator that was misinterpreted because we are parsing template
168       // parameters.
169       // FIXME: This is getting out of hand, write a decent parser.
170       if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
171           CurrentToken->Previous->is(TT_BinaryOperator) &&
172           Contexts[Contexts.size() - 2].IsExpression &&
173           !Line.startsWith(tok::kw_template))
174         return false;
175       updateParameterCount(Left, CurrentToken);
176       if (Style.Language == FormatStyle::LK_Proto) {
177         if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) {
178           if (CurrentToken->is(tok::colon) ||
179               (CurrentToken->isOneOf(tok::l_brace, tok::less) &&
180                Previous->isNot(tok::colon)))
181             Previous->setType(TT_SelectorName);
182         }
183       }
184       if (!consumeToken())
185         return false;
186     }
187     return false;
188   }
189 
190   bool parseUntouchableParens() {
191     while (CurrentToken) {
192       CurrentToken->Finalized = true;
193       switch (CurrentToken->Tok.getKind()) {
194       case tok::l_paren:
195         next();
196         if (!parseUntouchableParens())
197           return false;
198         continue;
199       case tok::r_paren:
200         next();
201         return true;
202       default:
203         // no-op
204         break;
205       }
206       next();
207     }
208     return false;
209   }
210 
211   bool parseParens(bool LookForDecls = false) {
212     if (!CurrentToken)
213       return false;
214     assert(CurrentToken->Previous && "Unknown previous token");
215     FormatToken &OpeningParen = *CurrentToken->Previous;
216     assert(OpeningParen.is(tok::l_paren));
217     FormatToken *PrevNonComment = OpeningParen.getPreviousNonComment();
218     OpeningParen.ParentBracket = Contexts.back().ContextKind;
219     ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
220 
221     // FIXME: This is a bit of a hack. Do better.
222     Contexts.back().ColonIsForRangeExpr =
223         Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
224 
225     if (OpeningParen.Previous &&
226         OpeningParen.Previous->is(TT_UntouchableMacroFunc)) {
227       OpeningParen.Finalized = true;
228       return parseUntouchableParens();
229     }
230 
231     bool StartsObjCMethodExpr = false;
232     if (FormatToken *MaybeSel = OpeningParen.Previous) {
233       // @selector( starts a selector.
234       if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
235           MaybeSel->Previous->is(tok::at))
236         StartsObjCMethodExpr = true;
237     }
238 
239     if (OpeningParen.is(TT_OverloadedOperatorLParen)) {
240       // Find the previous kw_operator token.
241       FormatToken *Prev = &OpeningParen;
242       while (!Prev->is(tok::kw_operator)) {
243         Prev = Prev->Previous;
244         assert(Prev && "Expect a kw_operator prior to the OperatorLParen!");
245       }
246 
247       // If faced with "a.operator*(argument)" or "a->operator*(argument)",
248       // i.e. the operator is called as a member function,
249       // then the argument must be an expression.
250       bool OperatorCalledAsMemberFunction =
251           Prev->Previous && Prev->Previous->isOneOf(tok::period, tok::arrow);
252       Contexts.back().IsExpression = OperatorCalledAsMemberFunction;
253     } else if (Style.isJavaScript() &&
254                (Line.startsWith(Keywords.kw_type, tok::identifier) ||
255                 Line.startsWith(tok::kw_export, Keywords.kw_type,
256                                 tok::identifier))) {
257       // type X = (...);
258       // export type X = (...);
259       Contexts.back().IsExpression = false;
260     } else if (OpeningParen.Previous &&
261                (OpeningParen.Previous->isOneOf(tok::kw_static_assert,
262                                                tok::kw_while, tok::l_paren,
263                                                tok::comma, TT_BinaryOperator) ||
264                 OpeningParen.Previous->isIf())) {
265       // static_assert, if and while usually contain expressions.
266       Contexts.back().IsExpression = true;
267     } else if (Style.isJavaScript() && OpeningParen.Previous &&
268                (OpeningParen.Previous->is(Keywords.kw_function) ||
269                 (OpeningParen.Previous->endsSequence(tok::identifier,
270                                                      Keywords.kw_function)))) {
271       // function(...) or function f(...)
272       Contexts.back().IsExpression = false;
273     } else if (Style.isJavaScript() && OpeningParen.Previous &&
274                OpeningParen.Previous->is(TT_JsTypeColon)) {
275       // let x: (SomeType);
276       Contexts.back().IsExpression = false;
277     } else if (isLambdaParameterList(&OpeningParen)) {
278       // This is a parameter list of a lambda expression.
279       Contexts.back().IsExpression = false;
280     } else if (Line.InPPDirective &&
281                (!OpeningParen.Previous ||
282                 !OpeningParen.Previous->is(tok::identifier))) {
283       Contexts.back().IsExpression = true;
284     } else if (Contexts[Contexts.size() - 2].CaretFound) {
285       // This is the parameter list of an ObjC block.
286       Contexts.back().IsExpression = false;
287     } else if (OpeningParen.Previous &&
288                OpeningParen.Previous->is(TT_ForEachMacro)) {
289       // The first argument to a foreach macro is a declaration.
290       Contexts.back().ContextType = Context::ForEachMacro;
291       Contexts.back().IsExpression = false;
292     } else if (OpeningParen.Previous && OpeningParen.Previous->MatchingParen &&
293                OpeningParen.Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
294       Contexts.back().IsExpression = false;
295     } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
296       bool IsForOrCatch =
297           OpeningParen.Previous &&
298           OpeningParen.Previous->isOneOf(tok::kw_for, tok::kw_catch);
299       Contexts.back().IsExpression = !IsForOrCatch;
300     }
301 
302     // Infer the role of the l_paren based on the previous token if we haven't
303     // detected one one yet.
304     if (PrevNonComment && OpeningParen.is(TT_Unknown)) {
305       if (PrevNonComment->is(tok::kw___attribute)) {
306         OpeningParen.setType(TT_AttributeParen);
307       } else if (PrevNonComment->isOneOf(TT_TypenameMacro, tok::kw_decltype,
308                                          tok::kw_typeof, tok::kw__Atomic,
309                                          tok::kw___underlying_type)) {
310         OpeningParen.setType(TT_TypeDeclarationParen);
311         // decltype() and typeof() usually contain expressions.
312         if (PrevNonComment->isOneOf(tok::kw_decltype, tok::kw_typeof))
313           Contexts.back().IsExpression = true;
314       }
315     }
316 
317     if (StartsObjCMethodExpr) {
318       Contexts.back().ColonIsObjCMethodExpr = true;
319       OpeningParen.setType(TT_ObjCMethodExpr);
320     }
321 
322     // MightBeFunctionType and ProbablyFunctionType are used for
323     // function pointer and reference types as well as Objective-C
324     // block types:
325     //
326     // void (*FunctionPointer)(void);
327     // void (&FunctionReference)(void);
328     // void (&&FunctionReference)(void);
329     // void (^ObjCBlock)(void);
330     bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
331     bool ProbablyFunctionType =
332         CurrentToken->isOneOf(tok::star, tok::amp, tok::ampamp, tok::caret);
333     bool HasMultipleLines = false;
334     bool HasMultipleParametersOnALine = false;
335     bool MightBeObjCForRangeLoop =
336         OpeningParen.Previous && OpeningParen.Previous->is(tok::kw_for);
337     FormatToken *PossibleObjCForInToken = nullptr;
338     while (CurrentToken) {
339       // LookForDecls is set when "if (" has been seen. Check for
340       // 'identifier' '*' 'identifier' followed by not '=' -- this
341       // '*' has to be a binary operator but determineStarAmpUsage() will
342       // categorize it as an unary operator, so set the right type here.
343       if (LookForDecls && CurrentToken->Next) {
344         FormatToken *Prev = CurrentToken->getPreviousNonComment();
345         if (Prev) {
346           FormatToken *PrevPrev = Prev->getPreviousNonComment();
347           FormatToken *Next = CurrentToken->Next;
348           if (PrevPrev && PrevPrev->is(tok::identifier) &&
349               Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
350               CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
351             Prev->setType(TT_BinaryOperator);
352             LookForDecls = false;
353           }
354         }
355       }
356 
357       if (CurrentToken->Previous->is(TT_PointerOrReference) &&
358           CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
359                                                     tok::coloncolon))
360         ProbablyFunctionType = true;
361       if (CurrentToken->is(tok::comma))
362         MightBeFunctionType = false;
363       if (CurrentToken->Previous->is(TT_BinaryOperator))
364         Contexts.back().IsExpression = true;
365       if (CurrentToken->is(tok::r_paren)) {
366         if (OpeningParen.isNot(TT_CppCastLParen) && MightBeFunctionType &&
367             ProbablyFunctionType && CurrentToken->Next &&
368             (CurrentToken->Next->is(tok::l_paren) ||
369              (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration)))
370           OpeningParen.setType(OpeningParen.Next->is(tok::caret)
371                                    ? TT_ObjCBlockLParen
372                                    : TT_FunctionTypeLParen);
373         OpeningParen.MatchingParen = CurrentToken;
374         CurrentToken->MatchingParen = &OpeningParen;
375 
376         if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
377             OpeningParen.Previous && OpeningParen.Previous->is(tok::l_paren)) {
378           // Detect the case where macros are used to generate lambdas or
379           // function bodies, e.g.:
380           //   auto my_lambda = MACRO((Type *type, int i) { .. body .. });
381           for (FormatToken *Tok = &OpeningParen; Tok != CurrentToken;
382                Tok = Tok->Next)
383             if (Tok->is(TT_BinaryOperator) &&
384                 Tok->isOneOf(tok::star, tok::amp, tok::ampamp))
385               Tok->setType(TT_PointerOrReference);
386         }
387 
388         if (StartsObjCMethodExpr) {
389           CurrentToken->setType(TT_ObjCMethodExpr);
390           if (Contexts.back().FirstObjCSelectorName) {
391             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
392                 Contexts.back().LongestObjCSelectorName;
393           }
394         }
395 
396         if (OpeningParen.is(TT_AttributeParen))
397           CurrentToken->setType(TT_AttributeParen);
398         if (OpeningParen.is(TT_TypeDeclarationParen))
399           CurrentToken->setType(TT_TypeDeclarationParen);
400         if (OpeningParen.Previous &&
401             OpeningParen.Previous->is(TT_JavaAnnotation))
402           CurrentToken->setType(TT_JavaAnnotation);
403         if (OpeningParen.Previous &&
404             OpeningParen.Previous->is(TT_LeadingJavaAnnotation))
405           CurrentToken->setType(TT_LeadingJavaAnnotation);
406         if (OpeningParen.Previous &&
407             OpeningParen.Previous->is(TT_AttributeSquare))
408           CurrentToken->setType(TT_AttributeSquare);
409 
410         if (!HasMultipleLines)
411           OpeningParen.setPackingKind(PPK_Inconclusive);
412         else if (HasMultipleParametersOnALine)
413           OpeningParen.setPackingKind(PPK_BinPacked);
414         else
415           OpeningParen.setPackingKind(PPK_OnePerLine);
416 
417         next();
418         return true;
419       }
420       if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
421         return false;
422 
423       if (CurrentToken->is(tok::l_brace) && OpeningParen.is(TT_ObjCBlockLParen))
424         OpeningParen.setType(TT_Unknown);
425       if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
426           !CurrentToken->Next->HasUnescapedNewline &&
427           !CurrentToken->Next->isTrailingComment())
428         HasMultipleParametersOnALine = true;
429       bool ProbablyFunctionTypeLParen =
430           (CurrentToken->is(tok::l_paren) && CurrentToken->Next &&
431            CurrentToken->Next->isOneOf(tok::star, tok::amp, tok::caret));
432       if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
433            CurrentToken->Previous->isSimpleTypeSpecifier()) &&
434           !(CurrentToken->is(tok::l_brace) ||
435             (CurrentToken->is(tok::l_paren) && !ProbablyFunctionTypeLParen)))
436         Contexts.back().IsExpression = false;
437       if (CurrentToken->isOneOf(tok::semi, tok::colon)) {
438         MightBeObjCForRangeLoop = false;
439         if (PossibleObjCForInToken) {
440           PossibleObjCForInToken->setType(TT_Unknown);
441           PossibleObjCForInToken = nullptr;
442         }
443       }
444       if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) {
445         PossibleObjCForInToken = CurrentToken;
446         PossibleObjCForInToken->setType(TT_ObjCForIn);
447       }
448       // When we discover a 'new', we set CanBeExpression to 'false' in order to
449       // parse the type correctly. Reset that after a comma.
450       if (CurrentToken->is(tok::comma))
451         Contexts.back().CanBeExpression = true;
452 
453       FormatToken *Tok = CurrentToken;
454       if (!consumeToken())
455         return false;
456       updateParameterCount(&OpeningParen, Tok);
457       if (CurrentToken && CurrentToken->HasUnescapedNewline)
458         HasMultipleLines = true;
459     }
460     return false;
461   }
462 
463   bool isCSharpAttributeSpecifier(const FormatToken &Tok) {
464     if (!Style.isCSharp())
465       return false;
466 
467     // `identifier[i]` is not an attribute.
468     if (Tok.Previous && Tok.Previous->is(tok::identifier))
469       return false;
470 
471     // Chains of [] in `identifier[i][j][k]` are not attributes.
472     if (Tok.Previous && Tok.Previous->is(tok::r_square)) {
473       auto *MatchingParen = Tok.Previous->MatchingParen;
474       if (!MatchingParen || MatchingParen->is(TT_ArraySubscriptLSquare))
475         return false;
476     }
477 
478     const FormatToken *AttrTok = Tok.Next;
479     if (!AttrTok)
480       return false;
481 
482     // Just an empty declaration e.g. string [].
483     if (AttrTok->is(tok::r_square))
484       return false;
485 
486     // Move along the tokens inbetween the '[' and ']' e.g. [STAThread].
487     while (AttrTok && AttrTok->isNot(tok::r_square))
488       AttrTok = AttrTok->Next;
489 
490     if (!AttrTok)
491       return false;
492 
493     // Allow an attribute to be the only content of a file.
494     AttrTok = AttrTok->Next;
495     if (!AttrTok)
496       return true;
497 
498     // Limit this to being an access modifier that follows.
499     if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
500                          tok::comment, tok::kw_class, tok::kw_static,
501                          tok::l_square, Keywords.kw_internal))
502       return true;
503 
504     // incase its a [XXX] retval func(....
505     if (AttrTok->Next &&
506         AttrTok->Next->startsSequence(tok::identifier, tok::l_paren))
507       return true;
508 
509     return false;
510   }
511 
512   bool isCpp11AttributeSpecifier(const FormatToken &Tok) {
513     if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square))
514       return false;
515     // The first square bracket is part of an ObjC array literal
516     if (Tok.Previous && Tok.Previous->is(tok::at))
517       return false;
518     const FormatToken *AttrTok = Tok.Next->Next;
519     if (!AttrTok)
520       return false;
521     // C++17 '[[using ns: foo, bar(baz, blech)]]'
522     // We assume nobody will name an ObjC variable 'using'.
523     if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon))
524       return true;
525     if (AttrTok->isNot(tok::identifier))
526       return false;
527     while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) {
528       // ObjC message send. We assume nobody will use : in a C++11 attribute
529       // specifier parameter, although this is technically valid:
530       // [[foo(:)]].
531       if (AttrTok->is(tok::colon) ||
532           AttrTok->startsSequence(tok::identifier, tok::identifier) ||
533           AttrTok->startsSequence(tok::r_paren, tok::identifier))
534         return false;
535       if (AttrTok->is(tok::ellipsis))
536         return true;
537       AttrTok = AttrTok->Next;
538     }
539     return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square);
540   }
541 
542   bool parseSquare() {
543     if (!CurrentToken)
544       return false;
545 
546     // A '[' could be an index subscript (after an identifier or after
547     // ')' or ']'), it could be the start of an Objective-C method
548     // expression, it could the start of an Objective-C array literal,
549     // or it could be a C++ attribute specifier [[foo::bar]].
550     FormatToken *Left = CurrentToken->Previous;
551     Left->ParentBracket = Contexts.back().ContextKind;
552     FormatToken *Parent = Left->getPreviousNonComment();
553 
554     // Cases where '>' is followed by '['.
555     // In C++, this can happen either in array of templates (foo<int>[10])
556     // or when array is a nested template type (unique_ptr<type1<type2>[]>).
557     bool CppArrayTemplates =
558         Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) &&
559         (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
560          Contexts.back().ContextType == Context::TemplateArgument);
561 
562     bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) ||
563                                      Contexts.back().InCpp11AttributeSpecifier;
564 
565     // Treat C# Attributes [STAThread] much like C++ attributes [[...]].
566     bool IsCSharpAttributeSpecifier =
567         isCSharpAttributeSpecifier(*Left) ||
568         Contexts.back().InCSharpAttributeSpecifier;
569 
570     bool InsideInlineASM = Line.startsWith(tok::kw_asm);
571     bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style);
572     bool StartsObjCMethodExpr =
573         !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates &&
574         Style.isCpp() && !IsCpp11AttributeSpecifier &&
575         !IsCSharpAttributeSpecifier && Contexts.back().CanBeExpression &&
576         Left->isNot(TT_LambdaLSquare) &&
577         !CurrentToken->isOneOf(tok::l_brace, tok::r_square) &&
578         (!Parent ||
579          Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
580                          tok::kw_return, tok::kw_throw) ||
581          Parent->isUnaryOperator() ||
582          // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
583          Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
584          (getBinOpPrecedence(Parent->Tok.getKind(), true, true) >
585           prec::Unknown));
586     bool ColonFound = false;
587 
588     unsigned BindingIncrease = 1;
589     if (IsCppStructuredBinding) {
590       Left->setType(TT_StructuredBindingLSquare);
591     } else if (Left->is(TT_Unknown)) {
592       if (StartsObjCMethodExpr) {
593         Left->setType(TT_ObjCMethodExpr);
594       } else if (InsideInlineASM) {
595         Left->setType(TT_InlineASMSymbolicNameLSquare);
596       } else if (IsCpp11AttributeSpecifier) {
597         Left->setType(TT_AttributeSquare);
598       } else if (Style.isJavaScript() && Parent &&
599                  Contexts.back().ContextKind == tok::l_brace &&
600                  Parent->isOneOf(tok::l_brace, tok::comma)) {
601         Left->setType(TT_JsComputedPropertyName);
602       } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace &&
603                  Parent && Parent->isOneOf(tok::l_brace, tok::comma)) {
604         Left->setType(TT_DesignatedInitializerLSquare);
605       } else if (IsCSharpAttributeSpecifier) {
606         Left->setType(TT_AttributeSquare);
607       } else if (CurrentToken->is(tok::r_square) && Parent &&
608                  Parent->is(TT_TemplateCloser)) {
609         Left->setType(TT_ArraySubscriptLSquare);
610       } else if (Style.Language == FormatStyle::LK_Proto ||
611                  Style.Language == FormatStyle::LK_TextProto) {
612         // Square braces in LK_Proto can either be message field attributes:
613         //
614         // optional Aaa aaa = 1 [
615         //   (aaa) = aaa
616         // ];
617         //
618         // extensions 123 [
619         //   (aaa) = aaa
620         // ];
621         //
622         // or text proto extensions (in options):
623         //
624         // option (Aaa.options) = {
625         //   [type.type/type] {
626         //     key: value
627         //   }
628         // }
629         //
630         // or repeated fields (in options):
631         //
632         // option (Aaa.options) = {
633         //   keys: [ 1, 2, 3 ]
634         // }
635         //
636         // In the first and the third case we want to spread the contents inside
637         // the square braces; in the second we want to keep them inline.
638         Left->setType(TT_ArrayInitializerLSquare);
639         if (!Left->endsSequence(tok::l_square, tok::numeric_constant,
640                                 tok::equal) &&
641             !Left->endsSequence(tok::l_square, tok::numeric_constant,
642                                 tok::identifier) &&
643             !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) {
644           Left->setType(TT_ProtoExtensionLSquare);
645           BindingIncrease = 10;
646         }
647       } else if (!CppArrayTemplates && Parent &&
648                  Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
649                                  tok::comma, tok::l_paren, tok::l_square,
650                                  tok::question, tok::colon, tok::kw_return,
651                                  // Should only be relevant to JavaScript:
652                                  tok::kw_default)) {
653         Left->setType(TT_ArrayInitializerLSquare);
654       } else {
655         BindingIncrease = 10;
656         Left->setType(TT_ArraySubscriptLSquare);
657       }
658     }
659 
660     ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
661     Contexts.back().IsExpression = true;
662     if (Style.isJavaScript() && Parent && Parent->is(TT_JsTypeColon))
663       Contexts.back().IsExpression = false;
664 
665     Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
666     Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier;
667     Contexts.back().InCSharpAttributeSpecifier = IsCSharpAttributeSpecifier;
668 
669     while (CurrentToken) {
670       if (CurrentToken->is(tok::r_square)) {
671         if (IsCpp11AttributeSpecifier)
672           CurrentToken->setType(TT_AttributeSquare);
673         if (IsCSharpAttributeSpecifier)
674           CurrentToken->setType(TT_AttributeSquare);
675         else if (((CurrentToken->Next &&
676                    CurrentToken->Next->is(tok::l_paren)) ||
677                   (CurrentToken->Previous &&
678                    CurrentToken->Previous->Previous == Left)) &&
679                  Left->is(TT_ObjCMethodExpr)) {
680           // An ObjC method call is rarely followed by an open parenthesis. It
681           // also can't be composed of just one token, unless it's a macro that
682           // will be expanded to more tokens.
683           // FIXME: Do we incorrectly label ":" with this?
684           StartsObjCMethodExpr = false;
685           Left->setType(TT_Unknown);
686         }
687         if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
688           CurrentToken->setType(TT_ObjCMethodExpr);
689           // If we haven't seen a colon yet, make sure the last identifier
690           // before the r_square is tagged as a selector name component.
691           if (!ColonFound && CurrentToken->Previous &&
692               CurrentToken->Previous->is(TT_Unknown) &&
693               canBeObjCSelectorComponent(*CurrentToken->Previous))
694             CurrentToken->Previous->setType(TT_SelectorName);
695           // determineStarAmpUsage() thinks that '*' '[' is allocating an
696           // array of pointers, but if '[' starts a selector then '*' is a
697           // binary operator.
698           if (Parent && Parent->is(TT_PointerOrReference))
699             Parent->overwriteFixedType(TT_BinaryOperator);
700         }
701         // An arrow after an ObjC method expression is not a lambda arrow.
702         if (CurrentToken->getType() == TT_ObjCMethodExpr &&
703             CurrentToken->Next && CurrentToken->Next->is(TT_LambdaArrow))
704           CurrentToken->Next->overwriteFixedType(TT_Unknown);
705         Left->MatchingParen = CurrentToken;
706         CurrentToken->MatchingParen = Left;
707         // FirstObjCSelectorName is set when a colon is found. This does
708         // not work, however, when the method has no parameters.
709         // Here, we set FirstObjCSelectorName when the end of the method call is
710         // reached, in case it was not set already.
711         if (!Contexts.back().FirstObjCSelectorName) {
712           FormatToken *Previous = CurrentToken->getPreviousNonComment();
713           if (Previous && Previous->is(TT_SelectorName)) {
714             Previous->ObjCSelectorNameParts = 1;
715             Contexts.back().FirstObjCSelectorName = Previous;
716           }
717         } else {
718           Left->ParameterCount =
719               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
720         }
721         if (Contexts.back().FirstObjCSelectorName) {
722           Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
723               Contexts.back().LongestObjCSelectorName;
724           if (Left->BlockParameterCount > 1)
725             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
726         }
727         next();
728         return true;
729       }
730       if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
731         return false;
732       if (CurrentToken->is(tok::colon)) {
733         if (IsCpp11AttributeSpecifier &&
734             CurrentToken->endsSequence(tok::colon, tok::identifier,
735                                        tok::kw_using)) {
736           // Remember that this is a [[using ns: foo]] C++ attribute, so we
737           // don't add a space before the colon (unlike other colons).
738           CurrentToken->setType(TT_AttributeColon);
739         } else if (Left->isOneOf(TT_ArraySubscriptLSquare,
740                                  TT_DesignatedInitializerLSquare)) {
741           Left->setType(TT_ObjCMethodExpr);
742           StartsObjCMethodExpr = true;
743           Contexts.back().ColonIsObjCMethodExpr = true;
744           if (Parent && Parent->is(tok::r_paren))
745             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
746             Parent->setType(TT_CastRParen);
747         }
748         ColonFound = true;
749       }
750       if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
751           !ColonFound)
752         Left->setType(TT_ArrayInitializerLSquare);
753       FormatToken *Tok = CurrentToken;
754       if (!consumeToken())
755         return false;
756       updateParameterCount(Left, Tok);
757     }
758     return false;
759   }
760 
761   bool couldBeInStructArrayInitializer() const {
762     if (Contexts.size() < 2)
763       return false;
764     // We want to back up no more then 2 context levels i.e.
765     // . { { <-
766     const auto End = std::next(Contexts.rbegin(), 2);
767     auto Last = Contexts.rbegin();
768     unsigned Depth = 0;
769     for (; Last != End; ++Last)
770       if (Last->ContextKind == tok::l_brace)
771         ++Depth;
772     return Depth == 2 && Last->ContextKind != tok::l_brace;
773   }
774 
775   bool parseBrace() {
776     if (!CurrentToken)
777       return true;
778 
779     assert(CurrentToken->Previous);
780     FormatToken &OpeningBrace = *CurrentToken->Previous;
781     assert(OpeningBrace.is(tok::l_brace));
782     OpeningBrace.ParentBracket = Contexts.back().ContextKind;
783 
784     if (Contexts.back().CaretFound)
785       OpeningBrace.overwriteFixedType(TT_ObjCBlockLBrace);
786     Contexts.back().CaretFound = false;
787 
788     ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
789     Contexts.back().ColonIsDictLiteral = true;
790     if (OpeningBrace.is(BK_BracedInit))
791       Contexts.back().IsExpression = true;
792     if (Style.isJavaScript() && OpeningBrace.Previous &&
793         OpeningBrace.Previous->is(TT_JsTypeColon))
794       Contexts.back().IsExpression = false;
795 
796     unsigned CommaCount = 0;
797     while (CurrentToken) {
798       if (CurrentToken->is(tok::r_brace)) {
799         assert(OpeningBrace.Optional == CurrentToken->Optional);
800         OpeningBrace.MatchingParen = CurrentToken;
801         CurrentToken->MatchingParen = &OpeningBrace;
802         if (Style.AlignArrayOfStructures != FormatStyle::AIAS_None) {
803           if (OpeningBrace.ParentBracket == tok::l_brace &&
804               couldBeInStructArrayInitializer() && CommaCount > 0)
805             Contexts.back().ContextType = Context::StructArrayInitializer;
806         }
807         next();
808         return true;
809       }
810       if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
811         return false;
812       updateParameterCount(&OpeningBrace, CurrentToken);
813       if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) {
814         FormatToken *Previous = CurrentToken->getPreviousNonComment();
815         if (Previous->is(TT_JsTypeOptionalQuestion))
816           Previous = Previous->getPreviousNonComment();
817         if ((CurrentToken->is(tok::colon) &&
818              (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) ||
819             Style.Language == FormatStyle::LK_Proto ||
820             Style.Language == FormatStyle::LK_TextProto) {
821           OpeningBrace.setType(TT_DictLiteral);
822           if (Previous->Tok.getIdentifierInfo() ||
823               Previous->is(tok::string_literal))
824             Previous->setType(TT_SelectorName);
825         }
826         if (CurrentToken->is(tok::colon) && OpeningBrace.is(TT_Unknown))
827           OpeningBrace.setType(TT_DictLiteral);
828         else if (Style.isJavaScript())
829           OpeningBrace.overwriteFixedType(TT_DictLiteral);
830       }
831       if (CurrentToken->is(tok::comma)) {
832         if (Style.isJavaScript())
833           OpeningBrace.overwriteFixedType(TT_DictLiteral);
834         ++CommaCount;
835       }
836       if (!consumeToken())
837         return false;
838     }
839     return true;
840   }
841 
842   void updateParameterCount(FormatToken *Left, FormatToken *Current) {
843     // For ObjC methods, the number of parameters is calculated differently as
844     // method declarations have a different structure (the parameters are not
845     // inside a bracket scope).
846     if (Current->is(tok::l_brace) && Current->is(BK_Block))
847       ++Left->BlockParameterCount;
848     if (Current->is(tok::comma)) {
849       ++Left->ParameterCount;
850       if (!Left->Role)
851         Left->Role.reset(new CommaSeparatedList(Style));
852       Left->Role->CommaFound(Current);
853     } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
854       Left->ParameterCount = 1;
855     }
856   }
857 
858   bool parseConditional() {
859     while (CurrentToken) {
860       if (CurrentToken->is(tok::colon)) {
861         CurrentToken->setType(TT_ConditionalExpr);
862         next();
863         return true;
864       }
865       if (!consumeToken())
866         return false;
867     }
868     return false;
869   }
870 
871   bool parseTemplateDeclaration() {
872     if (CurrentToken && CurrentToken->is(tok::less)) {
873       CurrentToken->setType(TT_TemplateOpener);
874       next();
875       if (!parseAngle())
876         return false;
877       if (CurrentToken)
878         CurrentToken->Previous->ClosesTemplateDeclaration = true;
879       return true;
880     }
881     return false;
882   }
883 
884   bool consumeToken() {
885     FormatToken *Tok = CurrentToken;
886     next();
887     switch (Tok->Tok.getKind()) {
888     case tok::plus:
889     case tok::minus:
890       if (!Tok->Previous && Line.MustBeDeclaration)
891         Tok->setType(TT_ObjCMethodSpecifier);
892       break;
893     case tok::colon:
894       if (!Tok->Previous)
895         return false;
896       // Colons from ?: are handled in parseConditional().
897       if (Style.isJavaScript()) {
898         if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
899             (Contexts.size() == 1 &&               // switch/case labels
900              !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
901             Contexts.back().ContextKind == tok::l_paren ||  // function params
902             Contexts.back().ContextKind == tok::l_square || // array type
903             (!Contexts.back().IsExpression &&
904              Contexts.back().ContextKind == tok::l_brace) || // object type
905             (Contexts.size() == 1 &&
906              Line.MustBeDeclaration)) { // method/property declaration
907           Contexts.back().IsExpression = false;
908           Tok->setType(TT_JsTypeColon);
909           break;
910         }
911       } else if (Style.isCSharp()) {
912         if (Contexts.back().InCSharpAttributeSpecifier) {
913           Tok->setType(TT_AttributeColon);
914           break;
915         }
916         if (Contexts.back().ContextKind == tok::l_paren) {
917           Tok->setType(TT_CSharpNamedArgumentColon);
918           break;
919         }
920       }
921       if (Line.First->isOneOf(Keywords.kw_module, Keywords.kw_import) ||
922           Line.First->startsSequence(tok::kw_export, Keywords.kw_module) ||
923           Line.First->startsSequence(tok::kw_export, Keywords.kw_import)) {
924         Tok->setType(TT_ModulePartitionColon);
925       } else if (Contexts.back().ColonIsDictLiteral ||
926                  Style.Language == FormatStyle::LK_Proto ||
927                  Style.Language == FormatStyle::LK_TextProto) {
928         Tok->setType(TT_DictLiteral);
929         if (Style.Language == FormatStyle::LK_TextProto) {
930           if (FormatToken *Previous = Tok->getPreviousNonComment())
931             Previous->setType(TT_SelectorName);
932         }
933       } else if (Contexts.back().ColonIsObjCMethodExpr ||
934                  Line.startsWith(TT_ObjCMethodSpecifier)) {
935         Tok->setType(TT_ObjCMethodExpr);
936         const FormatToken *BeforePrevious = Tok->Previous->Previous;
937         // Ensure we tag all identifiers in method declarations as
938         // TT_SelectorName.
939         bool UnknownIdentifierInMethodDeclaration =
940             Line.startsWith(TT_ObjCMethodSpecifier) &&
941             Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown);
942         if (!BeforePrevious ||
943             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
944             !(BeforePrevious->is(TT_CastRParen) ||
945               (BeforePrevious->is(TT_ObjCMethodExpr) &&
946                BeforePrevious->is(tok::colon))) ||
947             BeforePrevious->is(tok::r_square) ||
948             Contexts.back().LongestObjCSelectorName == 0 ||
949             UnknownIdentifierInMethodDeclaration) {
950           Tok->Previous->setType(TT_SelectorName);
951           if (!Contexts.back().FirstObjCSelectorName)
952             Contexts.back().FirstObjCSelectorName = Tok->Previous;
953           else if (Tok->Previous->ColumnWidth >
954                    Contexts.back().LongestObjCSelectorName)
955             Contexts.back().LongestObjCSelectorName =
956                 Tok->Previous->ColumnWidth;
957           Tok->Previous->ParameterIndex =
958               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
959           ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
960         }
961       } else if (Contexts.back().ColonIsForRangeExpr) {
962         Tok->setType(TT_RangeBasedForLoopColon);
963       } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
964         Tok->setType(TT_BitFieldColon);
965       } else if (Contexts.size() == 1 &&
966                  !Line.First->isOneOf(tok::kw_enum, tok::kw_case,
967                                       tok::kw_default)) {
968         FormatToken *Prev = Tok->getPreviousNonComment();
969         if (!Prev)
970           break;
971         if (Prev->isOneOf(tok::r_paren, tok::kw_noexcept))
972           Tok->setType(TT_CtorInitializerColon);
973         else if (Prev->is(tok::kw_try)) {
974           // Member initializer list within function try block.
975           FormatToken *PrevPrev = Prev->getPreviousNonComment();
976           if (!PrevPrev)
977             break;
978           if (PrevPrev && PrevPrev->isOneOf(tok::r_paren, tok::kw_noexcept))
979             Tok->setType(TT_CtorInitializerColon);
980         } else
981           Tok->setType(TT_InheritanceColon);
982       } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next &&
983                  (Tok->Next->isOneOf(tok::r_paren, tok::comma) ||
984                   (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next &&
985                    Tok->Next->Next->is(tok::colon)))) {
986         // This handles a special macro in ObjC code where selectors including
987         // the colon are passed as macro arguments.
988         Tok->setType(TT_ObjCMethodExpr);
989       } else if (Contexts.back().ContextKind == tok::l_paren) {
990         Tok->setType(TT_InlineASMColon);
991       }
992       break;
993     case tok::pipe:
994     case tok::amp:
995       // | and & in declarations/type expressions represent union and
996       // intersection types, respectively.
997       if (Style.isJavaScript() && !Contexts.back().IsExpression)
998         Tok->setType(TT_JsTypeOperator);
999       break;
1000     case tok::kw_if:
1001       if (CurrentToken &&
1002           CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier))
1003         next();
1004       LLVM_FALLTHROUGH;
1005     case tok::kw_while:
1006       if (CurrentToken && CurrentToken->is(tok::l_paren)) {
1007         next();
1008         if (!parseParens(/*LookForDecls=*/true))
1009           return false;
1010       }
1011       break;
1012     case tok::kw_for:
1013       if (Style.isJavaScript()) {
1014         // x.for and {for: ...}
1015         if ((Tok->Previous && Tok->Previous->is(tok::period)) ||
1016             (Tok->Next && Tok->Next->is(tok::colon)))
1017           break;
1018         // JS' for await ( ...
1019         if (CurrentToken && CurrentToken->is(Keywords.kw_await))
1020           next();
1021       }
1022       if (Style.isCpp() && CurrentToken && CurrentToken->is(tok::kw_co_await))
1023         next();
1024       Contexts.back().ColonIsForRangeExpr = true;
1025       if (!CurrentToken || CurrentToken->isNot(tok::l_paren))
1026         return false;
1027       next();
1028       if (!parseParens())
1029         return false;
1030       break;
1031     case tok::l_paren:
1032       // When faced with 'operator()()', the kw_operator handler incorrectly
1033       // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
1034       // the first two parens OverloadedOperators and the second l_paren an
1035       // OverloadedOperatorLParen.
1036       if (Tok->Previous && Tok->Previous->is(tok::r_paren) &&
1037           Tok->Previous->MatchingParen &&
1038           Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
1039         Tok->Previous->setType(TT_OverloadedOperator);
1040         Tok->Previous->MatchingParen->setType(TT_OverloadedOperator);
1041         Tok->setType(TT_OverloadedOperatorLParen);
1042       }
1043 
1044       if (!parseParens())
1045         return false;
1046       if (Line.MustBeDeclaration && Contexts.size() == 1 &&
1047           !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
1048           !Tok->isOneOf(TT_TypeDeclarationParen, TT_RequiresExpressionLParen) &&
1049           (!Tok->Previous || !Tok->Previous->isOneOf(tok::kw___attribute,
1050                                                      TT_LeadingJavaAnnotation)))
1051         Line.MightBeFunctionDecl = true;
1052       break;
1053     case tok::l_square:
1054       if (!parseSquare())
1055         return false;
1056       break;
1057     case tok::l_brace:
1058       if (Style.Language == FormatStyle::LK_TextProto) {
1059         FormatToken *Previous = Tok->getPreviousNonComment();
1060         if (Previous && Previous->getType() != TT_DictLiteral)
1061           Previous->setType(TT_SelectorName);
1062       }
1063       if (!parseBrace())
1064         return false;
1065       break;
1066     case tok::less:
1067       if (parseAngle()) {
1068         Tok->setType(TT_TemplateOpener);
1069         // In TT_Proto, we must distignuish between:
1070         //   map<key, value>
1071         //   msg < item: data >
1072         //   msg: < item: data >
1073         // In TT_TextProto, map<key, value> does not occur.
1074         if (Style.Language == FormatStyle::LK_TextProto ||
1075             (Style.Language == FormatStyle::LK_Proto && Tok->Previous &&
1076              Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
1077           Tok->setType(TT_DictLiteral);
1078           FormatToken *Previous = Tok->getPreviousNonComment();
1079           if (Previous && Previous->getType() != TT_DictLiteral)
1080             Previous->setType(TT_SelectorName);
1081         }
1082       } else {
1083         Tok->setType(TT_BinaryOperator);
1084         NonTemplateLess.insert(Tok);
1085         CurrentToken = Tok;
1086         next();
1087       }
1088       break;
1089     case tok::r_paren:
1090     case tok::r_square:
1091       return false;
1092     case tok::r_brace:
1093       // Lines can start with '}'.
1094       if (Tok->Previous)
1095         return false;
1096       break;
1097     case tok::greater:
1098       if (Style.Language != FormatStyle::LK_TextProto)
1099         Tok->setType(TT_BinaryOperator);
1100       if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser))
1101         Tok->SpacesRequiredBefore = 1;
1102       break;
1103     case tok::kw_operator:
1104       if (Style.Language == FormatStyle::LK_TextProto ||
1105           Style.Language == FormatStyle::LK_Proto)
1106         break;
1107       while (CurrentToken &&
1108              !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
1109         if (CurrentToken->isOneOf(tok::star, tok::amp))
1110           CurrentToken->setType(TT_PointerOrReference);
1111         consumeToken();
1112         if (CurrentToken && CurrentToken->is(tok::comma) &&
1113             CurrentToken->Previous->isNot(tok::kw_operator))
1114           break;
1115         if (CurrentToken && CurrentToken->Previous->isOneOf(
1116                                 TT_BinaryOperator, TT_UnaryOperator, tok::comma,
1117                                 tok::star, tok::arrow, tok::amp, tok::ampamp))
1118           CurrentToken->Previous->setType(TT_OverloadedOperator);
1119       }
1120       if (CurrentToken && CurrentToken->is(tok::l_paren))
1121         CurrentToken->setType(TT_OverloadedOperatorLParen);
1122       if (CurrentToken && CurrentToken->Previous->is(TT_BinaryOperator))
1123         CurrentToken->Previous->setType(TT_OverloadedOperator);
1124       break;
1125     case tok::question:
1126       if (Style.isJavaScript() && Tok->Next &&
1127           Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
1128                              tok::r_brace)) {
1129         // Question marks before semicolons, colons, etc. indicate optional
1130         // types (fields, parameters), e.g.
1131         //   function(x?: string, y?) {...}
1132         //   class X { y?; }
1133         Tok->setType(TT_JsTypeOptionalQuestion);
1134         break;
1135       }
1136       // Declarations cannot be conditional expressions, this can only be part
1137       // of a type declaration.
1138       if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
1139           Style.isJavaScript())
1140         break;
1141       if (Style.isCSharp()) {
1142         // `Type?)`, `Type?>`, `Type? name;` and `Type? name =` can only be
1143         // nullable types.
1144         // Line.MustBeDeclaration will be true for `Type? name;`.
1145         if ((!Contexts.back().IsExpression && Line.MustBeDeclaration) ||
1146             (Tok->Next && Tok->Next->isOneOf(tok::r_paren, tok::greater)) ||
1147             (Tok->Next && Tok->Next->is(tok::identifier) && Tok->Next->Next &&
1148              Tok->Next->Next->is(tok::equal))) {
1149           Tok->setType(TT_CSharpNullable);
1150           break;
1151         }
1152       }
1153       parseConditional();
1154       break;
1155     case tok::kw_template:
1156       parseTemplateDeclaration();
1157       break;
1158     case tok::comma:
1159       switch (Contexts.back().ContextType) {
1160       case Context::CtorInitializer:
1161         Tok->setType(TT_CtorInitializerComma);
1162         break;
1163       case Context::InheritanceList:
1164         Tok->setType(TT_InheritanceComma);
1165         break;
1166       default:
1167         if (Contexts.back().FirstStartOfName &&
1168             (Contexts.size() == 1 || startsWithInitStatement(Line))) {
1169           Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
1170           Line.IsMultiVariableDeclStmt = true;
1171         }
1172         break;
1173       }
1174       if (Contexts.back().ContextType == Context::ForEachMacro)
1175         Contexts.back().IsExpression = true;
1176       break;
1177     case tok::identifier:
1178       if (Tok->isOneOf(Keywords.kw___has_include,
1179                        Keywords.kw___has_include_next))
1180         parseHasInclude();
1181       if (Style.isCSharp() && Tok->is(Keywords.kw_where) && Tok->Next &&
1182           Tok->Next->isNot(tok::l_paren)) {
1183         Tok->setType(TT_CSharpGenericTypeConstraint);
1184         parseCSharpGenericTypeConstraint();
1185       }
1186       break;
1187     case tok::arrow:
1188       if (Tok->Previous && Tok->Previous->is(tok::kw_noexcept))
1189         Tok->setType(TT_TrailingReturnArrow);
1190       break;
1191     default:
1192       break;
1193     }
1194     return true;
1195   }
1196 
1197   void parseCSharpGenericTypeConstraint() {
1198     int OpenAngleBracketsCount = 0;
1199     while (CurrentToken) {
1200       if (CurrentToken->is(tok::less)) {
1201         // parseAngle is too greedy and will consume the whole line.
1202         CurrentToken->setType(TT_TemplateOpener);
1203         ++OpenAngleBracketsCount;
1204         next();
1205       } else if (CurrentToken->is(tok::greater)) {
1206         CurrentToken->setType(TT_TemplateCloser);
1207         --OpenAngleBracketsCount;
1208         next();
1209       } else if (CurrentToken->is(tok::comma) && OpenAngleBracketsCount == 0) {
1210         // We allow line breaks after GenericTypeConstraintComma's
1211         // so do not flag commas in Generics as GenericTypeConstraintComma's.
1212         CurrentToken->setType(TT_CSharpGenericTypeConstraintComma);
1213         next();
1214       } else if (CurrentToken->is(Keywords.kw_where)) {
1215         CurrentToken->setType(TT_CSharpGenericTypeConstraint);
1216         next();
1217       } else if (CurrentToken->is(tok::colon)) {
1218         CurrentToken->setType(TT_CSharpGenericTypeConstraintColon);
1219         next();
1220       } else {
1221         next();
1222       }
1223     }
1224   }
1225 
1226   void parseIncludeDirective() {
1227     if (CurrentToken && CurrentToken->is(tok::less)) {
1228       next();
1229       while (CurrentToken) {
1230         // Mark tokens up to the trailing line comments as implicit string
1231         // literals.
1232         if (CurrentToken->isNot(tok::comment) &&
1233             !CurrentToken->TokenText.startswith("//"))
1234           CurrentToken->setType(TT_ImplicitStringLiteral);
1235         next();
1236       }
1237     }
1238   }
1239 
1240   void parseWarningOrError() {
1241     next();
1242     // We still want to format the whitespace left of the first token of the
1243     // warning or error.
1244     next();
1245     while (CurrentToken) {
1246       CurrentToken->setType(TT_ImplicitStringLiteral);
1247       next();
1248     }
1249   }
1250 
1251   void parsePragma() {
1252     next(); // Consume "pragma".
1253     if (CurrentToken &&
1254         CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
1255       bool IsMark = CurrentToken->is(Keywords.kw_mark);
1256       next(); // Consume "mark".
1257       next(); // Consume first token (so we fix leading whitespace).
1258       while (CurrentToken) {
1259         if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
1260           CurrentToken->setType(TT_ImplicitStringLiteral);
1261         next();
1262       }
1263     }
1264   }
1265 
1266   void parseHasInclude() {
1267     if (!CurrentToken || !CurrentToken->is(tok::l_paren))
1268       return;
1269     next(); // '('
1270     parseIncludeDirective();
1271     next(); // ')'
1272   }
1273 
1274   LineType parsePreprocessorDirective() {
1275     bool IsFirstToken = CurrentToken->IsFirst;
1276     LineType Type = LT_PreprocessorDirective;
1277     next();
1278     if (!CurrentToken)
1279       return Type;
1280 
1281     if (Style.isJavaScript() && IsFirstToken) {
1282       // JavaScript files can contain shebang lines of the form:
1283       // #!/usr/bin/env node
1284       // Treat these like C++ #include directives.
1285       while (CurrentToken) {
1286         // Tokens cannot be comments here.
1287         CurrentToken->setType(TT_ImplicitStringLiteral);
1288         next();
1289       }
1290       return LT_ImportStatement;
1291     }
1292 
1293     if (CurrentToken->is(tok::numeric_constant)) {
1294       CurrentToken->SpacesRequiredBefore = 1;
1295       return Type;
1296     }
1297     // Hashes in the middle of a line can lead to any strange token
1298     // sequence.
1299     if (!CurrentToken->Tok.getIdentifierInfo())
1300       return Type;
1301     switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
1302     case tok::pp_include:
1303     case tok::pp_include_next:
1304     case tok::pp_import:
1305       next();
1306       parseIncludeDirective();
1307       Type = LT_ImportStatement;
1308       break;
1309     case tok::pp_error:
1310     case tok::pp_warning:
1311       parseWarningOrError();
1312       break;
1313     case tok::pp_pragma:
1314       parsePragma();
1315       break;
1316     case tok::pp_if:
1317     case tok::pp_elif:
1318       Contexts.back().IsExpression = true;
1319       next();
1320       parseLine();
1321       break;
1322     default:
1323       break;
1324     }
1325     while (CurrentToken) {
1326       FormatToken *Tok = CurrentToken;
1327       next();
1328       if (Tok->is(tok::l_paren))
1329         parseParens();
1330       else if (Tok->isOneOf(Keywords.kw___has_include,
1331                             Keywords.kw___has_include_next))
1332         parseHasInclude();
1333     }
1334     return Type;
1335   }
1336 
1337 public:
1338   LineType parseLine() {
1339     if (!CurrentToken)
1340       return LT_Invalid;
1341     NonTemplateLess.clear();
1342     if (CurrentToken->is(tok::hash))
1343       return parsePreprocessorDirective();
1344 
1345     // Directly allow to 'import <string-literal>' to support protocol buffer
1346     // definitions (github.com/google/protobuf) or missing "#" (either way we
1347     // should not break the line).
1348     IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
1349     if ((Style.Language == FormatStyle::LK_Java &&
1350          CurrentToken->is(Keywords.kw_package)) ||
1351         (Info && Info->getPPKeywordID() == tok::pp_import &&
1352          CurrentToken->Next &&
1353          CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
1354                                      tok::kw_static))) {
1355       next();
1356       parseIncludeDirective();
1357       return LT_ImportStatement;
1358     }
1359 
1360     // If this line starts and ends in '<' and '>', respectively, it is likely
1361     // part of "#define <a/b.h>".
1362     if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
1363       parseIncludeDirective();
1364       return LT_ImportStatement;
1365     }
1366 
1367     // In .proto files, top-level options and package statements are very
1368     // similar to import statements and should not be line-wrapped.
1369     if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
1370         CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) {
1371       next();
1372       if (CurrentToken && CurrentToken->is(tok::identifier)) {
1373         while (CurrentToken)
1374           next();
1375         return LT_ImportStatement;
1376       }
1377     }
1378 
1379     bool KeywordVirtualFound = false;
1380     bool ImportStatement = false;
1381 
1382     // import {...} from '...';
1383     if (Style.isJavaScript() && CurrentToken->is(Keywords.kw_import))
1384       ImportStatement = true;
1385 
1386     while (CurrentToken) {
1387       if (CurrentToken->is(tok::kw_virtual))
1388         KeywordVirtualFound = true;
1389       if (Style.isJavaScript()) {
1390         // export {...} from '...';
1391         // An export followed by "from 'some string';" is a re-export from
1392         // another module identified by a URI and is treated as a
1393         // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
1394         // Just "export {...};" or "export class ..." should not be treated as
1395         // an import in this sense.
1396         if (Line.First->is(tok::kw_export) &&
1397             CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
1398             CurrentToken->Next->isStringLiteral())
1399           ImportStatement = true;
1400         if (isClosureImportStatement(*CurrentToken))
1401           ImportStatement = true;
1402       }
1403       if (!consumeToken())
1404         return LT_Invalid;
1405     }
1406     if (KeywordVirtualFound)
1407       return LT_VirtualFunctionDecl;
1408     if (ImportStatement)
1409       return LT_ImportStatement;
1410 
1411     if (Line.startsWith(TT_ObjCMethodSpecifier)) {
1412       if (Contexts.back().FirstObjCSelectorName)
1413         Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
1414             Contexts.back().LongestObjCSelectorName;
1415       return LT_ObjCMethodDecl;
1416     }
1417 
1418     for (const auto &ctx : Contexts)
1419       if (ctx.ContextType == Context::StructArrayInitializer)
1420         return LT_ArrayOfStructInitializer;
1421 
1422     return LT_Other;
1423   }
1424 
1425 private:
1426   bool isClosureImportStatement(const FormatToken &Tok) {
1427     // FIXME: Closure-library specific stuff should not be hard-coded but be
1428     // configurable.
1429     return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
1430            Tok.Next->Next &&
1431            (Tok.Next->Next->TokenText == "module" ||
1432             Tok.Next->Next->TokenText == "provide" ||
1433             Tok.Next->Next->TokenText == "require" ||
1434             Tok.Next->Next->TokenText == "requireType" ||
1435             Tok.Next->Next->TokenText == "forwardDeclare") &&
1436            Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
1437   }
1438 
1439   void resetTokenMetadata() {
1440     if (!CurrentToken)
1441       return;
1442 
1443     // Reset token type in case we have already looked at it and then
1444     // recovered from an error (e.g. failure to find the matching >).
1445     if (!CurrentToken->isTypeFinalized() &&
1446         !CurrentToken->isOneOf(
1447             TT_LambdaLSquare, TT_LambdaLBrace, TT_AttributeMacro, TT_IfMacro,
1448             TT_ForEachMacro, TT_TypenameMacro, TT_FunctionLBrace,
1449             TT_ImplicitStringLiteral, TT_InlineASMBrace, TT_FatArrow,
1450             TT_LambdaArrow, TT_NamespaceMacro, TT_OverloadedOperator,
1451             TT_RegexLiteral, TT_TemplateString, TT_ObjCStringLiteral,
1452             TT_UntouchableMacroFunc, TT_StatementAttributeLikeMacro,
1453             TT_FunctionLikeOrFreestandingMacro, TT_ClassLBrace, TT_EnumLBrace,
1454             TT_RecordLBrace, TT_StructLBrace, TT_UnionLBrace, TT_RequiresClause,
1455             TT_RequiresClauseInARequiresExpression, TT_RequiresExpression,
1456             TT_RequiresExpressionLParen, TT_RequiresExpressionLBrace,
1457             TT_CompoundRequirementLBrace, TT_BracedListLBrace))
1458       CurrentToken->setType(TT_Unknown);
1459     CurrentToken->Role.reset();
1460     CurrentToken->MatchingParen = nullptr;
1461     CurrentToken->FakeLParens.clear();
1462     CurrentToken->FakeRParens = 0;
1463   }
1464 
1465   void next() {
1466     if (!CurrentToken)
1467       return;
1468 
1469     CurrentToken->NestingLevel = Contexts.size() - 1;
1470     CurrentToken->BindingStrength = Contexts.back().BindingStrength;
1471     modifyContext(*CurrentToken);
1472     determineTokenType(*CurrentToken);
1473     CurrentToken = CurrentToken->Next;
1474 
1475     resetTokenMetadata();
1476   }
1477 
1478   /// A struct to hold information valid in a specific context, e.g.
1479   /// a pair of parenthesis.
1480   struct Context {
1481     Context(tok::TokenKind ContextKind, unsigned BindingStrength,
1482             bool IsExpression)
1483         : ContextKind(ContextKind), BindingStrength(BindingStrength),
1484           IsExpression(IsExpression) {}
1485 
1486     tok::TokenKind ContextKind;
1487     unsigned BindingStrength;
1488     bool IsExpression;
1489     unsigned LongestObjCSelectorName = 0;
1490     bool ColonIsForRangeExpr = false;
1491     bool ColonIsDictLiteral = false;
1492     bool ColonIsObjCMethodExpr = false;
1493     FormatToken *FirstObjCSelectorName = nullptr;
1494     FormatToken *FirstStartOfName = nullptr;
1495     bool CanBeExpression = true;
1496     bool CaretFound = false;
1497     bool InCpp11AttributeSpecifier = false;
1498     bool InCSharpAttributeSpecifier = false;
1499     enum {
1500       Unknown,
1501       // Like the part after `:` in a constructor.
1502       //   Context(...) : IsExpression(IsExpression)
1503       CtorInitializer,
1504       // Like in the parentheses in a foreach.
1505       ForEachMacro,
1506       // Like the inheritance list in a class declaration.
1507       //   class Input : public IO
1508       InheritanceList,
1509       // Like in the braced list.
1510       //   int x[] = {};
1511       StructArrayInitializer,
1512       // Like in `static_cast<int>`.
1513       TemplateArgument,
1514     } ContextType = Unknown;
1515   };
1516 
1517   /// Puts a new \c Context onto the stack \c Contexts for the lifetime
1518   /// of each instance.
1519   struct ScopedContextCreator {
1520     AnnotatingParser &P;
1521 
1522     ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
1523                          unsigned Increase)
1524         : P(P) {
1525       P.Contexts.push_back(Context(ContextKind,
1526                                    P.Contexts.back().BindingStrength + Increase,
1527                                    P.Contexts.back().IsExpression));
1528     }
1529 
1530     ~ScopedContextCreator() {
1531       if (P.Style.AlignArrayOfStructures != FormatStyle::AIAS_None) {
1532         if (P.Contexts.back().ContextType == Context::StructArrayInitializer) {
1533           P.Contexts.pop_back();
1534           P.Contexts.back().ContextType = Context::StructArrayInitializer;
1535           return;
1536         }
1537       }
1538       P.Contexts.pop_back();
1539     }
1540   };
1541 
1542   void modifyContext(const FormatToken &Current) {
1543     auto AssignmentStartsExpression = [&]() {
1544       if (Current.getPrecedence() != prec::Assignment)
1545         return false;
1546 
1547       if (Line.First->isOneOf(tok::kw_using, tok::kw_return))
1548         return false;
1549       if (Line.First->is(tok::kw_template)) {
1550         assert(Current.Previous);
1551         if (Current.Previous->is(tok::kw_operator)) {
1552           // `template ... operator=` cannot be an expression.
1553           return false;
1554         }
1555 
1556         // `template` keyword can start a variable template.
1557         const FormatToken *Tok = Line.First->getNextNonComment();
1558         assert(Tok); // Current token is on the same line.
1559         if (Tok->isNot(TT_TemplateOpener)) {
1560           // Explicit template instantiations do not have `<>`.
1561           return false;
1562         }
1563 
1564         Tok = Tok->MatchingParen;
1565         if (!Tok)
1566           return false;
1567         Tok = Tok->getNextNonComment();
1568         if (!Tok)
1569           return false;
1570 
1571         if (Tok->isOneOf(tok::kw_class, tok::kw_enum, tok::kw_concept,
1572                          tok::kw_struct, tok::kw_using))
1573           return false;
1574 
1575         return true;
1576       }
1577 
1578       // Type aliases use `type X = ...;` in TypeScript and can be exported
1579       // using `export type ...`.
1580       if (Style.isJavaScript() &&
1581           (Line.startsWith(Keywords.kw_type, tok::identifier) ||
1582            Line.startsWith(tok::kw_export, Keywords.kw_type, tok::identifier)))
1583         return false;
1584 
1585       return !Current.Previous || Current.Previous->isNot(tok::kw_operator);
1586     };
1587 
1588     if (AssignmentStartsExpression()) {
1589       Contexts.back().IsExpression = true;
1590       if (!Line.startsWith(TT_UnaryOperator)) {
1591         for (FormatToken *Previous = Current.Previous;
1592              Previous && Previous->Previous &&
1593              !Previous->Previous->isOneOf(tok::comma, tok::semi);
1594              Previous = Previous->Previous) {
1595           if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
1596             Previous = Previous->MatchingParen;
1597             if (!Previous)
1598               break;
1599           }
1600           if (Previous->opensScope())
1601             break;
1602           if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
1603               Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
1604               Previous->Previous && Previous->Previous->isNot(tok::equal))
1605             Previous->setType(TT_PointerOrReference);
1606         }
1607       }
1608     } else if (Current.is(tok::lessless) &&
1609                (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1610       Contexts.back().IsExpression = true;
1611     } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1612       Contexts.back().IsExpression = true;
1613     } else if (Current.is(TT_TrailingReturnArrow)) {
1614       Contexts.back().IsExpression = false;
1615     } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1616       Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1617     } else if (Current.Previous &&
1618                Current.Previous->is(TT_CtorInitializerColon)) {
1619       Contexts.back().IsExpression = true;
1620       Contexts.back().ContextType = Context::CtorInitializer;
1621     } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) {
1622       Contexts.back().ContextType = Context::InheritanceList;
1623     } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1624       for (FormatToken *Previous = Current.Previous;
1625            Previous && Previous->isOneOf(tok::star, tok::amp);
1626            Previous = Previous->Previous)
1627         Previous->setType(TT_PointerOrReference);
1628       if (Line.MustBeDeclaration &&
1629           Contexts.front().ContextType != Context::CtorInitializer)
1630         Contexts.back().IsExpression = false;
1631     } else if (Current.is(tok::kw_new)) {
1632       Contexts.back().CanBeExpression = false;
1633     } else if (Current.is(tok::semi) ||
1634                (Current.is(tok::exclaim) && Current.Previous &&
1635                 !Current.Previous->is(tok::kw_operator))) {
1636       // This should be the condition or increment in a for-loop.
1637       // But not operator !() (can't use TT_OverloadedOperator here as its not
1638       // been annotated yet).
1639       Contexts.back().IsExpression = true;
1640     }
1641   }
1642 
1643   static FormatToken *untilMatchingParen(FormatToken *Current) {
1644     // Used when `MatchingParen` is not yet established.
1645     int ParenLevel = 0;
1646     while (Current) {
1647       if (Current->is(tok::l_paren))
1648         ++ParenLevel;
1649       if (Current->is(tok::r_paren))
1650         --ParenLevel;
1651       if (ParenLevel < 1)
1652         break;
1653       Current = Current->Next;
1654     }
1655     return Current;
1656   }
1657 
1658   static bool isDeductionGuide(FormatToken &Current) {
1659     // Look for a deduction guide template<T> A(...) -> A<...>;
1660     if (Current.Previous && Current.Previous->is(tok::r_paren) &&
1661         Current.startsSequence(tok::arrow, tok::identifier, tok::less)) {
1662       // Find the TemplateCloser.
1663       FormatToken *TemplateCloser = Current.Next->Next;
1664       int NestingLevel = 0;
1665       while (TemplateCloser) {
1666         // Skip over an expressions in parens  A<(3 < 2)>;
1667         if (TemplateCloser->is(tok::l_paren)) {
1668           // No Matching Paren yet so skip to matching paren
1669           TemplateCloser = untilMatchingParen(TemplateCloser);
1670           if (!TemplateCloser)
1671             break;
1672         }
1673         if (TemplateCloser->is(tok::less))
1674           ++NestingLevel;
1675         if (TemplateCloser->is(tok::greater))
1676           --NestingLevel;
1677         if (NestingLevel < 1)
1678           break;
1679         TemplateCloser = TemplateCloser->Next;
1680       }
1681       // Assuming we have found the end of the template ensure its followed
1682       // with a semi-colon.
1683       if (TemplateCloser && TemplateCloser->Next &&
1684           TemplateCloser->Next->is(tok::semi) &&
1685           Current.Previous->MatchingParen) {
1686         // Determine if the identifier `A` prior to the A<..>; is the same as
1687         // prior to the A(..)
1688         FormatToken *LeadingIdentifier =
1689             Current.Previous->MatchingParen->Previous;
1690 
1691         // Differentiate a deduction guide by seeing the
1692         // > of the template prior to the leading identifier.
1693         if (LeadingIdentifier) {
1694           FormatToken *PriorLeadingIdentifier = LeadingIdentifier->Previous;
1695           // Skip back past explicit decoration
1696           if (PriorLeadingIdentifier &&
1697               PriorLeadingIdentifier->is(tok::kw_explicit))
1698             PriorLeadingIdentifier = PriorLeadingIdentifier->Previous;
1699 
1700           return PriorLeadingIdentifier &&
1701                  (PriorLeadingIdentifier->is(TT_TemplateCloser) ||
1702                   PriorLeadingIdentifier->ClosesRequiresClause) &&
1703                  LeadingIdentifier->TokenText == Current.Next->TokenText;
1704         }
1705       }
1706     }
1707     return false;
1708   }
1709 
1710   void determineTokenType(FormatToken &Current) {
1711     if (!Current.is(TT_Unknown))
1712       // The token type is already known.
1713       return;
1714 
1715     if ((Style.isJavaScript() || Style.isCSharp()) &&
1716         Current.is(tok::exclaim)) {
1717       if (Current.Previous) {
1718         bool IsIdentifier =
1719             Style.isJavaScript()
1720                 ? Keywords.IsJavaScriptIdentifier(
1721                       *Current.Previous, /* AcceptIdentifierName= */ true)
1722                 : Current.Previous->is(tok::identifier);
1723         if (IsIdentifier ||
1724             Current.Previous->isOneOf(
1725                 tok::kw_default, tok::kw_namespace, tok::r_paren, tok::r_square,
1726                 tok::r_brace, tok::kw_false, tok::kw_true, Keywords.kw_type,
1727                 Keywords.kw_get, Keywords.kw_init, Keywords.kw_set) ||
1728             Current.Previous->Tok.isLiteral()) {
1729           Current.setType(TT_NonNullAssertion);
1730           return;
1731         }
1732       }
1733       if (Current.Next &&
1734           Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1735         Current.setType(TT_NonNullAssertion);
1736         return;
1737       }
1738     }
1739 
1740     // Line.MightBeFunctionDecl can only be true after the parentheses of a
1741     // function declaration have been found. In this case, 'Current' is a
1742     // trailing token of this declaration and thus cannot be a name.
1743     if (Current.is(Keywords.kw_instanceof)) {
1744       Current.setType(TT_BinaryOperator);
1745     } else if (isStartOfName(Current) &&
1746                (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1747       Contexts.back().FirstStartOfName = &Current;
1748       Current.setType(TT_StartOfName);
1749     } else if (Current.is(tok::semi)) {
1750       // Reset FirstStartOfName after finding a semicolon so that a for loop
1751       // with multiple increment statements is not confused with a for loop
1752       // having multiple variable declarations.
1753       Contexts.back().FirstStartOfName = nullptr;
1754     } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1755       AutoFound = true;
1756     } else if (Current.is(tok::arrow) &&
1757                Style.Language == FormatStyle::LK_Java) {
1758       Current.setType(TT_LambdaArrow);
1759     } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1760                Current.NestingLevel == 0 &&
1761                !Current.Previous->isOneOf(tok::kw_operator, tok::identifier)) {
1762       // not auto operator->() -> xxx;
1763       Current.setType(TT_TrailingReturnArrow);
1764     } else if (Current.is(tok::arrow) && Current.Previous &&
1765                Current.Previous->is(tok::r_brace)) {
1766       // Concept implicit conversion constraint needs to be treated like
1767       // a trailing return type  ... } -> <type>.
1768       Current.setType(TT_TrailingReturnArrow);
1769     } else if (isDeductionGuide(Current)) {
1770       // Deduction guides trailing arrow " A(...) -> A<T>;".
1771       Current.setType(TT_TrailingReturnArrow);
1772     } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1773       Current.setType(determineStarAmpUsage(
1774           Current,
1775           Contexts.back().CanBeExpression && Contexts.back().IsExpression,
1776           Contexts.back().ContextType == Context::TemplateArgument));
1777     } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1778       Current.setType(determinePlusMinusCaretUsage(Current));
1779       if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1780         Contexts.back().CaretFound = true;
1781     } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1782       Current.setType(determineIncrementUsage(Current));
1783     } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1784       Current.setType(TT_UnaryOperator);
1785     } else if (Current.is(tok::question)) {
1786       if (Style.isJavaScript() && Line.MustBeDeclaration &&
1787           !Contexts.back().IsExpression) {
1788         // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1789         // on the interface, not a ternary expression.
1790         Current.setType(TT_JsTypeOptionalQuestion);
1791       } else {
1792         Current.setType(TT_ConditionalExpr);
1793       }
1794     } else if (Current.isBinaryOperator() &&
1795                (!Current.Previous || Current.Previous->isNot(tok::l_square)) &&
1796                (!Current.is(tok::greater) &&
1797                 Style.Language != FormatStyle::LK_TextProto)) {
1798       Current.setType(TT_BinaryOperator);
1799     } else if (Current.is(tok::comment)) {
1800       if (Current.TokenText.startswith("/*"))
1801         if (Current.TokenText.endswith("*/"))
1802           Current.setType(TT_BlockComment);
1803         else
1804           // The lexer has for some reason determined a comment here. But we
1805           // cannot really handle it, if it isn't properly terminated.
1806           Current.Tok.setKind(tok::unknown);
1807       else
1808         Current.setType(TT_LineComment);
1809     } else if (Current.is(tok::l_paren)) {
1810       if (lParenStartsCppCast(Current))
1811         Current.setType(TT_CppCastLParen);
1812     } else if (Current.is(tok::r_paren)) {
1813       if (rParenEndsCast(Current))
1814         Current.setType(TT_CastRParen);
1815       if (Current.MatchingParen && Current.Next &&
1816           !Current.Next->isBinaryOperator() &&
1817           !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1818                                  tok::comma, tok::period, tok::arrow,
1819                                  tok::coloncolon))
1820         if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1821           // Make sure this isn't the return type of an Obj-C block declaration
1822           if (AfterParen->isNot(tok::caret)) {
1823             if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1824               if (BeforeParen->is(tok::identifier) &&
1825                   !BeforeParen->is(TT_TypenameMacro) &&
1826                   BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1827                   (!BeforeParen->Previous ||
1828                    BeforeParen->Previous->ClosesTemplateDeclaration))
1829                 Current.setType(TT_FunctionAnnotationRParen);
1830           }
1831         }
1832     } else if (Current.is(tok::at) && Current.Next && !Style.isJavaScript() &&
1833                Style.Language != FormatStyle::LK_Java) {
1834       // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1835       // marks declarations and properties that need special formatting.
1836       switch (Current.Next->Tok.getObjCKeywordID()) {
1837       case tok::objc_interface:
1838       case tok::objc_implementation:
1839       case tok::objc_protocol:
1840         Current.setType(TT_ObjCDecl);
1841         break;
1842       case tok::objc_property:
1843         Current.setType(TT_ObjCProperty);
1844         break;
1845       default:
1846         break;
1847       }
1848     } else if (Current.is(tok::period)) {
1849       FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1850       if (PreviousNoComment &&
1851           PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1852         Current.setType(TT_DesignatedInitializerPeriod);
1853       else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1854                Current.Previous->isOneOf(TT_JavaAnnotation,
1855                                          TT_LeadingJavaAnnotation))
1856         Current.setType(Current.Previous->getType());
1857     } else if (canBeObjCSelectorComponent(Current) &&
1858                // FIXME(bug 36976): ObjC return types shouldn't use
1859                // TT_CastRParen.
1860                Current.Previous && Current.Previous->is(TT_CastRParen) &&
1861                Current.Previous->MatchingParen &&
1862                Current.Previous->MatchingParen->Previous &&
1863                Current.Previous->MatchingParen->Previous->is(
1864                    TT_ObjCMethodSpecifier)) {
1865       // This is the first part of an Objective-C selector name. (If there's no
1866       // colon after this, this is the only place which annotates the identifier
1867       // as a selector.)
1868       Current.setType(TT_SelectorName);
1869     } else if (Current.isOneOf(tok::identifier, tok::kw_const, tok::kw_noexcept,
1870                                tok::kw_requires) &&
1871                Current.Previous &&
1872                !Current.Previous->isOneOf(tok::equal, tok::at) &&
1873                Line.MightBeFunctionDecl && Contexts.size() == 1) {
1874       // Line.MightBeFunctionDecl can only be true after the parentheses of a
1875       // function declaration have been found.
1876       Current.setType(TT_TrailingAnnotation);
1877     } else if ((Style.Language == FormatStyle::LK_Java ||
1878                 Style.isJavaScript()) &&
1879                Current.Previous) {
1880       if (Current.Previous->is(tok::at) &&
1881           Current.isNot(Keywords.kw_interface)) {
1882         const FormatToken &AtToken = *Current.Previous;
1883         const FormatToken *Previous = AtToken.getPreviousNonComment();
1884         if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1885           Current.setType(TT_LeadingJavaAnnotation);
1886         else
1887           Current.setType(TT_JavaAnnotation);
1888       } else if (Current.Previous->is(tok::period) &&
1889                  Current.Previous->isOneOf(TT_JavaAnnotation,
1890                                            TT_LeadingJavaAnnotation)) {
1891         Current.setType(Current.Previous->getType());
1892       }
1893     }
1894   }
1895 
1896   /// Take a guess at whether \p Tok starts a name of a function or
1897   /// variable declaration.
1898   ///
1899   /// This is a heuristic based on whether \p Tok is an identifier following
1900   /// something that is likely a type.
1901   bool isStartOfName(const FormatToken &Tok) {
1902     if (Tok.isNot(tok::identifier) || !Tok.Previous)
1903       return false;
1904 
1905     if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1906                               Keywords.kw_as))
1907       return false;
1908     if (Style.isJavaScript() && Tok.Previous->is(Keywords.kw_in))
1909       return false;
1910 
1911     // Skip "const" as it does not have an influence on whether this is a name.
1912     FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1913 
1914     // For javascript const can be like "let" or "var"
1915     if (!Style.isJavaScript())
1916       while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1917         PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1918 
1919     if (!PreviousNotConst)
1920       return false;
1921 
1922     if (PreviousNotConst->ClosesRequiresClause)
1923       return false;
1924 
1925     bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1926                        PreviousNotConst->Previous &&
1927                        PreviousNotConst->Previous->is(tok::hash);
1928 
1929     if (PreviousNotConst->is(TT_TemplateCloser))
1930       return PreviousNotConst && PreviousNotConst->MatchingParen &&
1931              PreviousNotConst->MatchingParen->Previous &&
1932              PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1933              PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1934 
1935     if (PreviousNotConst->is(tok::r_paren) &&
1936         PreviousNotConst->is(TT_TypeDeclarationParen))
1937       return true;
1938 
1939     // If is a preprocess keyword like #define.
1940     if (IsPPKeyword)
1941       return false;
1942 
1943     // int a or auto a.
1944     if (PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto))
1945       return true;
1946 
1947     // *a or &a or &&a.
1948     if (PreviousNotConst->is(TT_PointerOrReference))
1949       return true;
1950 
1951     // MyClass a;
1952     if (PreviousNotConst->isSimpleTypeSpecifier())
1953       return true;
1954 
1955     // const a = in JavaScript.
1956     return Style.isJavaScript() && PreviousNotConst->is(tok::kw_const);
1957   }
1958 
1959   /// Determine whether '(' is starting a C++ cast.
1960   bool lParenStartsCppCast(const FormatToken &Tok) {
1961     // C-style casts are only used in C++.
1962     if (!Style.isCpp())
1963       return false;
1964 
1965     FormatToken *LeftOfParens = Tok.getPreviousNonComment();
1966     if (LeftOfParens && LeftOfParens->is(TT_TemplateCloser) &&
1967         LeftOfParens->MatchingParen) {
1968       auto *Prev = LeftOfParens->MatchingParen->getPreviousNonComment();
1969       if (Prev && Prev->isOneOf(tok::kw_const_cast, tok::kw_dynamic_cast,
1970                                 tok::kw_reinterpret_cast, tok::kw_static_cast))
1971         // FIXME: Maybe we should handle identifiers ending with "_cast",
1972         // e.g. any_cast?
1973         return true;
1974     }
1975     return false;
1976   }
1977 
1978   /// Determine whether ')' is ending a cast.
1979   bool rParenEndsCast(const FormatToken &Tok) {
1980     // C-style casts are only used in C++, C# and Java.
1981     if (!Style.isCSharp() && !Style.isCpp() &&
1982         Style.Language != FormatStyle::LK_Java)
1983       return false;
1984 
1985     // Empty parens aren't casts and there are no casts at the end of the line.
1986     if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1987       return false;
1988 
1989     FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1990     if (LeftOfParens) {
1991       // If there is a closing parenthesis left of the current
1992       // parentheses, look past it as these might be chained casts.
1993       if (LeftOfParens->is(tok::r_paren) &&
1994           LeftOfParens->isNot(TT_CastRParen)) {
1995         if (!LeftOfParens->MatchingParen ||
1996             !LeftOfParens->MatchingParen->Previous)
1997           return false;
1998         LeftOfParens = LeftOfParens->MatchingParen->Previous;
1999       }
2000 
2001       if (LeftOfParens->is(tok::r_square)) {
2002         //   delete[] (void *)ptr;
2003         auto MayBeArrayDelete = [](FormatToken *Tok) -> FormatToken * {
2004           if (Tok->isNot(tok::r_square))
2005             return nullptr;
2006 
2007           Tok = Tok->getPreviousNonComment();
2008           if (!Tok || Tok->isNot(tok::l_square))
2009             return nullptr;
2010 
2011           Tok = Tok->getPreviousNonComment();
2012           if (!Tok || Tok->isNot(tok::kw_delete))
2013             return nullptr;
2014           return Tok;
2015         };
2016         if (FormatToken *MaybeDelete = MayBeArrayDelete(LeftOfParens))
2017           LeftOfParens = MaybeDelete;
2018       }
2019 
2020       // The Condition directly below this one will see the operator arguments
2021       // as a (void *foo) cast.
2022       //   void operator delete(void *foo) ATTRIB;
2023       if (LeftOfParens->Tok.getIdentifierInfo() && LeftOfParens->Previous &&
2024           LeftOfParens->Previous->is(tok::kw_operator))
2025         return false;
2026 
2027       // If there is an identifier (or with a few exceptions a keyword) right
2028       // before the parentheses, this is unlikely to be a cast.
2029       if (LeftOfParens->Tok.getIdentifierInfo() &&
2030           !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
2031                                  tok::kw_delete))
2032         return false;
2033 
2034       // Certain other tokens right before the parentheses are also signals that
2035       // this cannot be a cast.
2036       if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
2037                                 TT_TemplateCloser, tok::ellipsis))
2038         return false;
2039     }
2040 
2041     if (Tok.Next->is(tok::question))
2042       return false;
2043 
2044     // `foreach((A a, B b) in someList)` should not be seen as a cast.
2045     if (Tok.Next->is(Keywords.kw_in) && Style.isCSharp())
2046       return false;
2047 
2048     // Functions which end with decorations like volatile, noexcept are unlikely
2049     // to be casts.
2050     if (Tok.Next->isOneOf(tok::kw_noexcept, tok::kw_volatile, tok::kw_const,
2051                           tok::kw_requires, tok::kw_throw, tok::arrow,
2052                           Keywords.kw_override, Keywords.kw_final) ||
2053         isCpp11AttributeSpecifier(*Tok.Next))
2054       return false;
2055 
2056     // As Java has no function types, a "(" after the ")" likely means that this
2057     // is a cast.
2058     if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
2059       return true;
2060 
2061     // If a (non-string) literal follows, this is likely a cast.
2062     if (Tok.Next->isNot(tok::string_literal) &&
2063         (Tok.Next->Tok.isLiteral() ||
2064          Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
2065       return true;
2066 
2067     // Heuristically try to determine whether the parentheses contain a type.
2068     auto IsQualifiedPointerOrReference = [](FormatToken *T) {
2069       // This is used to handle cases such as x = (foo *const)&y;
2070       assert(!T->isSimpleTypeSpecifier() && "Should have already been checked");
2071       // Strip trailing qualifiers such as const or volatile when checking
2072       // whether the parens could be a cast to a pointer/reference type.
2073       while (T) {
2074         if (T->is(TT_AttributeParen)) {
2075           // Handle `x = (foo *__attribute__((foo)))&v;`:
2076           if (T->MatchingParen && T->MatchingParen->Previous &&
2077               T->MatchingParen->Previous->is(tok::kw___attribute)) {
2078             T = T->MatchingParen->Previous->Previous;
2079             continue;
2080           }
2081         } else if (T->is(TT_AttributeSquare)) {
2082           // Handle `x = (foo *[[clang::foo]])&v;`:
2083           if (T->MatchingParen && T->MatchingParen->Previous) {
2084             T = T->MatchingParen->Previous;
2085             continue;
2086           }
2087         } else if (T->canBePointerOrReferenceQualifier()) {
2088           T = T->Previous;
2089           continue;
2090         }
2091         break;
2092       }
2093       return T && T->is(TT_PointerOrReference);
2094     };
2095     bool ParensAreType =
2096         !Tok.Previous ||
2097         Tok.Previous->isOneOf(TT_TemplateCloser, TT_TypeDeclarationParen) ||
2098         Tok.Previous->isSimpleTypeSpecifier() ||
2099         IsQualifiedPointerOrReference(Tok.Previous);
2100     bool ParensCouldEndDecl =
2101         Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
2102     if (ParensAreType && !ParensCouldEndDecl)
2103       return true;
2104 
2105     // At this point, we heuristically assume that there are no casts at the
2106     // start of the line. We assume that we have found most cases where there
2107     // are by the logic above, e.g. "(void)x;".
2108     if (!LeftOfParens)
2109       return false;
2110 
2111     // Certain token types inside the parentheses mean that this can't be a
2112     // cast.
2113     for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
2114          Token = Token->Next)
2115       if (Token->is(TT_BinaryOperator))
2116         return false;
2117 
2118     // If the following token is an identifier or 'this', this is a cast. All
2119     // cases where this can be something else are handled above.
2120     if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
2121       return true;
2122 
2123     // Look for a cast `( x ) (`.
2124     if (Tok.Next->is(tok::l_paren) && Tok.Previous && Tok.Previous->Previous) {
2125       if (Tok.Previous->is(tok::identifier) &&
2126           Tok.Previous->Previous->is(tok::l_paren))
2127         return true;
2128     }
2129 
2130     if (!Tok.Next->Next)
2131       return false;
2132 
2133     // If the next token after the parenthesis is a unary operator, assume
2134     // that this is cast, unless there are unexpected tokens inside the
2135     // parenthesis.
2136     bool NextIsUnary =
2137         Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
2138     if (!NextIsUnary || Tok.Next->is(tok::plus) ||
2139         !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
2140       return false;
2141     // Search for unexpected tokens.
2142     for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
2143          Prev = Prev->Previous)
2144       if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
2145         return false;
2146     return true;
2147   }
2148 
2149   /// Returns true if the token is used as a unary operator.
2150   bool determineUnaryOperatorByUsage(const FormatToken &Tok) {
2151     const FormatToken *PrevToken = Tok.getPreviousNonComment();
2152     if (!PrevToken)
2153       return true;
2154 
2155     // These keywords are deliberately not included here because they may
2156     // precede only one of unary star/amp and plus/minus but not both.  They are
2157     // either included in determineStarAmpUsage or determinePlusMinusCaretUsage.
2158     //
2159     // @ - It may be followed by a unary `-` in Objective-C literals. We don't
2160     //   know how they can be followed by a star or amp.
2161     if (PrevToken->isOneOf(
2162             TT_ConditionalExpr, tok::l_paren, tok::comma, tok::colon, tok::semi,
2163             tok::equal, tok::question, tok::l_square, tok::l_brace,
2164             tok::kw_case, tok::kw_co_await, tok::kw_co_return, tok::kw_co_yield,
2165             tok::kw_delete, tok::kw_return, tok::kw_throw))
2166       return true;
2167 
2168     // We put sizeof here instead of only in determineStarAmpUsage. In the cases
2169     // where the unary `+` operator is overloaded, it is reasonable to write
2170     // things like `sizeof +x`. Like commit 446d6ec996c6c3.
2171     if (PrevToken->is(tok::kw_sizeof))
2172       return true;
2173 
2174     // A sequence of leading unary operators.
2175     if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator))
2176       return true;
2177 
2178     // There can't be two consecutive binary operators.
2179     if (PrevToken->is(TT_BinaryOperator))
2180       return true;
2181 
2182     return false;
2183   }
2184 
2185   /// Return the type of the given token assuming it is * or &.
2186   TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
2187                                   bool InTemplateArgument) {
2188     if (Style.isJavaScript())
2189       return TT_BinaryOperator;
2190 
2191     // && in C# must be a binary operator.
2192     if (Style.isCSharp() && Tok.is(tok::ampamp))
2193       return TT_BinaryOperator;
2194 
2195     const FormatToken *PrevToken = Tok.getPreviousNonComment();
2196     if (!PrevToken)
2197       return TT_UnaryOperator;
2198 
2199     const FormatToken *NextToken = Tok.getNextNonComment();
2200 
2201     if (InTemplateArgument && NextToken && NextToken->is(tok::kw_noexcept))
2202       return TT_BinaryOperator;
2203 
2204     if (!NextToken ||
2205         NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_noexcept) ||
2206         NextToken->canBePointerOrReferenceQualifier() ||
2207         (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
2208       return TT_PointerOrReference;
2209 
2210     if (PrevToken->is(tok::coloncolon))
2211       return TT_PointerOrReference;
2212 
2213     if (PrevToken->is(tok::r_paren) && PrevToken->is(TT_TypeDeclarationParen))
2214       return TT_PointerOrReference;
2215 
2216     if (determineUnaryOperatorByUsage(Tok))
2217       return TT_UnaryOperator;
2218 
2219     if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
2220       return TT_PointerOrReference;
2221     if (NextToken->is(tok::kw_operator) && !IsExpression)
2222       return TT_PointerOrReference;
2223     if (NextToken->isOneOf(tok::comma, tok::semi))
2224       return TT_PointerOrReference;
2225 
2226     if (PrevToken->Tok.isLiteral() ||
2227         PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
2228                            tok::kw_false, tok::r_brace))
2229       return TT_BinaryOperator;
2230 
2231     const FormatToken *NextNonParen = NextToken;
2232     while (NextNonParen && NextNonParen->is(tok::l_paren))
2233       NextNonParen = NextNonParen->getNextNonComment();
2234     if (NextNonParen && (NextNonParen->Tok.isLiteral() ||
2235                          NextNonParen->isOneOf(tok::kw_true, tok::kw_false) ||
2236                          NextNonParen->isUnaryOperator()))
2237       return TT_BinaryOperator;
2238 
2239     // If we know we're in a template argument, there are no named declarations.
2240     // Thus, having an identifier on the right-hand side indicates a binary
2241     // operator.
2242     if (InTemplateArgument && NextToken->Tok.isAnyIdentifier())
2243       return TT_BinaryOperator;
2244 
2245     // "&&(" is quite unlikely to be two successive unary "&".
2246     if (Tok.is(tok::ampamp) && NextToken->is(tok::l_paren))
2247       return TT_BinaryOperator;
2248 
2249     // This catches some cases where evaluation order is used as control flow:
2250     //   aaa && aaa->f();
2251     if (NextToken->Tok.isAnyIdentifier()) {
2252       const FormatToken *NextNextToken = NextToken->getNextNonComment();
2253       if (NextNextToken && NextNextToken->is(tok::arrow))
2254         return TT_BinaryOperator;
2255     }
2256 
2257     // It is very unlikely that we are going to find a pointer or reference type
2258     // definition on the RHS of an assignment.
2259     if (IsExpression && !Contexts.back().CaretFound)
2260       return TT_BinaryOperator;
2261 
2262     return TT_PointerOrReference;
2263   }
2264 
2265   TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
2266     if (determineUnaryOperatorByUsage(Tok))
2267       return TT_UnaryOperator;
2268 
2269     const FormatToken *PrevToken = Tok.getPreviousNonComment();
2270     if (!PrevToken)
2271       return TT_UnaryOperator;
2272 
2273     if (PrevToken->is(tok::at))
2274       return TT_UnaryOperator;
2275 
2276     // Fall back to marking the token as binary operator.
2277     return TT_BinaryOperator;
2278   }
2279 
2280   /// Determine whether ++/-- are pre- or post-increments/-decrements.
2281   TokenType determineIncrementUsage(const FormatToken &Tok) {
2282     const FormatToken *PrevToken = Tok.getPreviousNonComment();
2283     if (!PrevToken || PrevToken->is(TT_CastRParen))
2284       return TT_UnaryOperator;
2285     if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
2286       return TT_TrailingUnaryOperator;
2287 
2288     return TT_UnaryOperator;
2289   }
2290 
2291   SmallVector<Context, 8> Contexts;
2292 
2293   const FormatStyle &Style;
2294   AnnotatedLine &Line;
2295   FormatToken *CurrentToken;
2296   bool AutoFound;
2297   const AdditionalKeywords &Keywords;
2298 
2299   // Set of "<" tokens that do not open a template parameter list. If parseAngle
2300   // determines that a specific token can't be a template opener, it will make
2301   // same decision irrespective of the decisions for tokens leading up to it.
2302   // Store this information to prevent this from causing exponential runtime.
2303   llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
2304 };
2305 
2306 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
2307 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
2308 
2309 /// Parses binary expressions by inserting fake parenthesis based on
2310 /// operator precedence.
2311 class ExpressionParser {
2312 public:
2313   ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
2314                    AnnotatedLine &Line)
2315       : Style(Style), Keywords(Keywords), Line(Line), Current(Line.First) {}
2316 
2317   /// Parse expressions with the given operator precedence.
2318   void parse(int Precedence = 0) {
2319     // Skip 'return' and ObjC selector colons as they are not part of a binary
2320     // expression.
2321     while (Current && (Current->is(tok::kw_return) ||
2322                        (Current->is(tok::colon) &&
2323                         Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
2324       next();
2325 
2326     if (!Current || Precedence > PrecedenceArrowAndPeriod)
2327       return;
2328 
2329     // Conditional expressions need to be parsed separately for proper nesting.
2330     if (Precedence == prec::Conditional) {
2331       parseConditionalExpr();
2332       return;
2333     }
2334 
2335     // Parse unary operators, which all have a higher precedence than binary
2336     // operators.
2337     if (Precedence == PrecedenceUnaryOperator) {
2338       parseUnaryOperator();
2339       return;
2340     }
2341 
2342     FormatToken *Start = Current;
2343     FormatToken *LatestOperator = nullptr;
2344     unsigned OperatorIndex = 0;
2345 
2346     while (Current) {
2347       // Consume operators with higher precedence.
2348       parse(Precedence + 1);
2349 
2350       int CurrentPrecedence = getCurrentPrecedence();
2351 
2352       if (Precedence == CurrentPrecedence && Current &&
2353           Current->is(TT_SelectorName)) {
2354         if (LatestOperator)
2355           addFakeParenthesis(Start, prec::Level(Precedence));
2356         Start = Current;
2357       }
2358 
2359       // At the end of the line or when an operator with higher precedence is
2360       // found, insert fake parenthesis and return.
2361       if (!Current ||
2362           (Current->closesScope() &&
2363            (Current->MatchingParen || Current->is(TT_TemplateString))) ||
2364           (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
2365           (CurrentPrecedence == prec::Conditional &&
2366            Precedence == prec::Assignment && Current->is(tok::colon)))
2367         break;
2368 
2369       // Consume scopes: (), [], <> and {}
2370       // In addition to that we handle require clauses as scope, so that the
2371       // constraints in that are correctly indented.
2372       if (Current->opensScope() ||
2373           Current->isOneOf(TT_RequiresClause,
2374                            TT_RequiresClauseInARequiresExpression)) {
2375         // In fragment of a JavaScript template string can look like '}..${' and
2376         // thus close a scope and open a new one at the same time.
2377         while (Current && (!Current->closesScope() || Current->opensScope())) {
2378           next();
2379           parse();
2380         }
2381         next();
2382       } else {
2383         // Operator found.
2384         if (CurrentPrecedence == Precedence) {
2385           if (LatestOperator)
2386             LatestOperator->NextOperator = Current;
2387           LatestOperator = Current;
2388           Current->OperatorIndex = OperatorIndex;
2389           ++OperatorIndex;
2390         }
2391         next(/*SkipPastLeadingComments=*/Precedence > 0);
2392       }
2393     }
2394 
2395     if (LatestOperator && (Current || Precedence > 0)) {
2396       // The requires clauses do not neccessarily end in a semicolon or a brace,
2397       // but just go over to struct/class or a function declaration, we need to
2398       // intervene so that the fake right paren is inserted correctly.
2399       auto End =
2400           (Start->Previous &&
2401            Start->Previous->isOneOf(TT_RequiresClause,
2402                                     TT_RequiresClauseInARequiresExpression))
2403               ? [this](){
2404                   auto Ret = Current ? Current : Line.Last;
2405                   while (!Ret->ClosesRequiresClause && Ret->Previous)
2406                     Ret = Ret->Previous;
2407                   return Ret;
2408                 }()
2409               : nullptr;
2410 
2411       if (Precedence == PrecedenceArrowAndPeriod) {
2412         // Call expressions don't have a binary operator precedence.
2413         addFakeParenthesis(Start, prec::Unknown, End);
2414       } else {
2415         addFakeParenthesis(Start, prec::Level(Precedence), End);
2416       }
2417     }
2418   }
2419 
2420 private:
2421   /// Gets the precedence (+1) of the given token for binary operators
2422   /// and other tokens that we treat like binary operators.
2423   int getCurrentPrecedence() {
2424     if (Current) {
2425       const FormatToken *NextNonComment = Current->getNextNonComment();
2426       if (Current->is(TT_ConditionalExpr))
2427         return prec::Conditional;
2428       if (NextNonComment && Current->is(TT_SelectorName) &&
2429           (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) ||
2430            ((Style.Language == FormatStyle::LK_Proto ||
2431              Style.Language == FormatStyle::LK_TextProto) &&
2432             NextNonComment->is(tok::less))))
2433         return prec::Assignment;
2434       if (Current->is(TT_JsComputedPropertyName))
2435         return prec::Assignment;
2436       if (Current->is(TT_LambdaArrow))
2437         return prec::Comma;
2438       if (Current->is(TT_FatArrow))
2439         return prec::Assignment;
2440       if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
2441           (Current->is(tok::comment) && NextNonComment &&
2442            NextNonComment->is(TT_SelectorName)))
2443         return 0;
2444       if (Current->is(TT_RangeBasedForLoopColon))
2445         return prec::Comma;
2446       if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) &&
2447           Current->is(Keywords.kw_instanceof))
2448         return prec::Relational;
2449       if (Style.isJavaScript() &&
2450           Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
2451         return prec::Relational;
2452       if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
2453         return Current->getPrecedence();
2454       if (Current->isOneOf(tok::period, tok::arrow))
2455         return PrecedenceArrowAndPeriod;
2456       if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) &&
2457           Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
2458                            Keywords.kw_throws))
2459         return 0;
2460     }
2461     return -1;
2462   }
2463 
2464   void addFakeParenthesis(FormatToken *Start, prec::Level Precedence,
2465                           FormatToken *End = nullptr) {
2466     Start->FakeLParens.push_back(Precedence);
2467     if (Precedence > prec::Unknown)
2468       Start->StartsBinaryExpression = true;
2469     if (!End && Current)
2470       End = Current->getPreviousNonComment();
2471     if (End) {
2472       ++End->FakeRParens;
2473       if (Precedence > prec::Unknown)
2474         End->EndsBinaryExpression = true;
2475     }
2476   }
2477 
2478   /// Parse unary operator expressions and surround them with fake
2479   /// parentheses if appropriate.
2480   void parseUnaryOperator() {
2481     llvm::SmallVector<FormatToken *, 2> Tokens;
2482     while (Current && Current->is(TT_UnaryOperator)) {
2483       Tokens.push_back(Current);
2484       next();
2485     }
2486     parse(PrecedenceArrowAndPeriod);
2487     for (FormatToken *Token : llvm::reverse(Tokens))
2488       // The actual precedence doesn't matter.
2489       addFakeParenthesis(Token, prec::Unknown);
2490   }
2491 
2492   void parseConditionalExpr() {
2493     while (Current && Current->isTrailingComment())
2494       next();
2495     FormatToken *Start = Current;
2496     parse(prec::LogicalOr);
2497     if (!Current || !Current->is(tok::question))
2498       return;
2499     next();
2500     parse(prec::Assignment);
2501     if (!Current || Current->isNot(TT_ConditionalExpr))
2502       return;
2503     next();
2504     parse(prec::Assignment);
2505     addFakeParenthesis(Start, prec::Conditional);
2506   }
2507 
2508   void next(bool SkipPastLeadingComments = true) {
2509     if (Current)
2510       Current = Current->Next;
2511     while (Current &&
2512            (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
2513            Current->isTrailingComment())
2514       Current = Current->Next;
2515   }
2516 
2517   const FormatStyle &Style;
2518   const AdditionalKeywords &Keywords;
2519   const AnnotatedLine &Line;
2520   FormatToken *Current;
2521 };
2522 
2523 } // end anonymous namespace
2524 
2525 void TokenAnnotator::setCommentLineLevels(
2526     SmallVectorImpl<AnnotatedLine *> &Lines) const {
2527   const AnnotatedLine *NextNonCommentLine = nullptr;
2528   for (AnnotatedLine *Line : llvm::reverse(Lines)) {
2529     assert(Line->First);
2530 
2531     // If the comment is currently aligned with the line immediately following
2532     // it, that's probably intentional and we should keep it.
2533     if (NextNonCommentLine && Line->isComment() &&
2534         NextNonCommentLine->First->NewlinesBefore <= 1 &&
2535         NextNonCommentLine->First->OriginalColumn ==
2536             Line->First->OriginalColumn) {
2537       // Align comments for preprocessor lines with the # in column 0 if
2538       // preprocessor lines are not indented. Otherwise, align with the next
2539       // line.
2540       Line->Level =
2541           (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
2542            (NextNonCommentLine->Type == LT_PreprocessorDirective ||
2543             NextNonCommentLine->Type == LT_ImportStatement))
2544               ? 0
2545               : NextNonCommentLine->Level;
2546     } else {
2547       NextNonCommentLine = Line->First->isNot(tok::r_brace) ? Line : nullptr;
2548     }
2549 
2550     setCommentLineLevels(Line->Children);
2551   }
2552 }
2553 
2554 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
2555   unsigned Result = 0;
2556   for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
2557     Result = std::max(Result, Tok->NestingLevel);
2558   return Result;
2559 }
2560 
2561 void TokenAnnotator::annotate(AnnotatedLine &Line) const {
2562   for (auto &Child : Line.Children)
2563     annotate(*Child);
2564 
2565   AnnotatingParser Parser(Style, Line, Keywords);
2566   Line.Type = Parser.parseLine();
2567 
2568   // With very deep nesting, ExpressionParser uses lots of stack and the
2569   // formatting algorithm is very slow. We're not going to do a good job here
2570   // anyway - it's probably generated code being formatted by mistake.
2571   // Just skip the whole line.
2572   if (maxNestingDepth(Line) > 50)
2573     Line.Type = LT_Invalid;
2574 
2575   if (Line.Type == LT_Invalid)
2576     return;
2577 
2578   ExpressionParser ExprParser(Style, Keywords, Line);
2579   ExprParser.parse();
2580 
2581   if (Line.startsWith(TT_ObjCMethodSpecifier))
2582     Line.Type = LT_ObjCMethodDecl;
2583   else if (Line.startsWith(TT_ObjCDecl))
2584     Line.Type = LT_ObjCDecl;
2585   else if (Line.startsWith(TT_ObjCProperty))
2586     Line.Type = LT_ObjCProperty;
2587 
2588   Line.First->SpacesRequiredBefore = 1;
2589   Line.First->CanBreakBefore = Line.First->MustBreakBefore;
2590 }
2591 
2592 // This function heuristically determines whether 'Current' starts the name of a
2593 // function declaration.
2594 static bool isFunctionDeclarationName(bool IsCpp, const FormatToken &Current,
2595                                       const AnnotatedLine &Line) {
2596   auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * {
2597     for (; Next; Next = Next->Next) {
2598       if (Next->is(TT_OverloadedOperatorLParen))
2599         return Next;
2600       if (Next->is(TT_OverloadedOperator))
2601         continue;
2602       if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
2603         // For 'new[]' and 'delete[]'.
2604         if (Next->Next &&
2605             Next->Next->startsSequence(tok::l_square, tok::r_square))
2606           Next = Next->Next->Next;
2607         continue;
2608       }
2609       if (Next->startsSequence(tok::l_square, tok::r_square)) {
2610         // For operator[]().
2611         Next = Next->Next;
2612         continue;
2613       }
2614       if ((Next->isSimpleTypeSpecifier() || Next->is(tok::identifier)) &&
2615           Next->Next && Next->Next->isOneOf(tok::star, tok::amp, tok::ampamp)) {
2616         // For operator void*(), operator char*(), operator Foo*().
2617         Next = Next->Next;
2618         continue;
2619       }
2620       if (Next->is(TT_TemplateOpener) && Next->MatchingParen) {
2621         Next = Next->MatchingParen;
2622         continue;
2623       }
2624 
2625       break;
2626     }
2627     return nullptr;
2628   };
2629 
2630   // Find parentheses of parameter list.
2631   const FormatToken *Next = Current.Next;
2632   if (Current.is(tok::kw_operator)) {
2633     if (Current.Previous && Current.Previous->is(tok::coloncolon))
2634       return false;
2635     Next = skipOperatorName(Next);
2636   } else {
2637     if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
2638       return false;
2639     for (; Next; Next = Next->Next) {
2640       if (Next->is(TT_TemplateOpener)) {
2641         Next = Next->MatchingParen;
2642       } else if (Next->is(tok::coloncolon)) {
2643         Next = Next->Next;
2644         if (!Next)
2645           return false;
2646         if (Next->is(tok::kw_operator)) {
2647           Next = skipOperatorName(Next->Next);
2648           break;
2649         }
2650         if (!Next->is(tok::identifier))
2651           return false;
2652       } else if (Next->is(tok::l_paren)) {
2653         break;
2654       } else {
2655         return false;
2656       }
2657     }
2658   }
2659 
2660   // Check whether parameter list can belong to a function declaration.
2661   if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
2662     return false;
2663   // If the lines ends with "{", this is likely a function definition.
2664   if (Line.Last->is(tok::l_brace))
2665     return true;
2666   if (Next->Next == Next->MatchingParen)
2667     return true; // Empty parentheses.
2668   // If there is an &/&& after the r_paren, this is likely a function.
2669   if (Next->MatchingParen->Next &&
2670       Next->MatchingParen->Next->is(TT_PointerOrReference))
2671     return true;
2672 
2673   // Check for K&R C function definitions (and C++ function definitions with
2674   // unnamed parameters), e.g.:
2675   //   int f(i)
2676   //   {
2677   //     return i + 1;
2678   //   }
2679   //   bool g(size_t = 0, bool b = false)
2680   //   {
2681   //     return !b;
2682   //   }
2683   if (IsCpp && Next->Next && Next->Next->is(tok::identifier) &&
2684       !Line.endsWith(tok::semi))
2685     return true;
2686 
2687   for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
2688        Tok = Tok->Next) {
2689     if (Tok->is(TT_TypeDeclarationParen))
2690       return true;
2691     if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) {
2692       Tok = Tok->MatchingParen;
2693       continue;
2694     }
2695     if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
2696         Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
2697       return true;
2698     if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
2699         Tok->Tok.isLiteral())
2700       return false;
2701   }
2702   return false;
2703 }
2704 
2705 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
2706   assert(Line.MightBeFunctionDecl);
2707 
2708   if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
2709        Style.AlwaysBreakAfterReturnType ==
2710            FormatStyle::RTBS_TopLevelDefinitions) &&
2711       Line.Level > 0)
2712     return false;
2713 
2714   switch (Style.AlwaysBreakAfterReturnType) {
2715   case FormatStyle::RTBS_None:
2716     return false;
2717   case FormatStyle::RTBS_All:
2718   case FormatStyle::RTBS_TopLevel:
2719     return true;
2720   case FormatStyle::RTBS_AllDefinitions:
2721   case FormatStyle::RTBS_TopLevelDefinitions:
2722     return Line.mightBeFunctionDefinition();
2723   }
2724 
2725   return false;
2726 }
2727 
2728 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) const {
2729   for (AnnotatedLine *ChildLine : Line.Children)
2730     calculateFormattingInformation(*ChildLine);
2731 
2732   Line.First->TotalLength =
2733       Line.First->IsMultiline ? Style.ColumnLimit
2734                               : Line.FirstStartColumn + Line.First->ColumnWidth;
2735   FormatToken *Current = Line.First->Next;
2736   bool InFunctionDecl = Line.MightBeFunctionDecl;
2737   bool AlignArrayOfStructures =
2738       (Style.AlignArrayOfStructures != FormatStyle::AIAS_None &&
2739        Line.Type == LT_ArrayOfStructInitializer);
2740   if (AlignArrayOfStructures)
2741     calculateArrayInitializerColumnList(Line);
2742 
2743   while (Current) {
2744     if (isFunctionDeclarationName(Style.isCpp(), *Current, Line))
2745       Current->setType(TT_FunctionDeclarationName);
2746     const FormatToken *Prev = Current->Previous;
2747     if (Current->is(TT_LineComment)) {
2748       if (Prev->is(BK_BracedInit) && Prev->opensScope())
2749         Current->SpacesRequiredBefore =
2750             (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1;
2751       else
2752         Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
2753 
2754       // If we find a trailing comment, iterate backwards to determine whether
2755       // it seems to relate to a specific parameter. If so, break before that
2756       // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
2757       // to the previous line in:
2758       //   SomeFunction(a,
2759       //                b, // comment
2760       //                c);
2761       if (!Current->HasUnescapedNewline) {
2762         for (FormatToken *Parameter = Current->Previous; Parameter;
2763              Parameter = Parameter->Previous) {
2764           if (Parameter->isOneOf(tok::comment, tok::r_brace))
2765             break;
2766           if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
2767             if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
2768                 Parameter->HasUnescapedNewline)
2769               Parameter->MustBreakBefore = true;
2770             break;
2771           }
2772         }
2773       }
2774     } else if (Current->SpacesRequiredBefore == 0 &&
2775                spaceRequiredBefore(Line, *Current)) {
2776       Current->SpacesRequiredBefore = 1;
2777     }
2778 
2779     const auto &Children = Prev->Children;
2780     if (!Children.empty() && Children.back()->Last->is(TT_LineComment)) {
2781       Current->MustBreakBefore = true;
2782     } else {
2783       Current->MustBreakBefore =
2784           Current->MustBreakBefore || mustBreakBefore(Line, *Current);
2785       if (!Current->MustBreakBefore && InFunctionDecl &&
2786           Current->is(TT_FunctionDeclarationName))
2787         Current->MustBreakBefore = mustBreakForReturnType(Line);
2788     }
2789 
2790     Current->CanBreakBefore =
2791         Current->MustBreakBefore || canBreakBefore(Line, *Current);
2792     unsigned ChildSize = 0;
2793     if (Prev->Children.size() == 1) {
2794       FormatToken &LastOfChild = *Prev->Children[0]->Last;
2795       ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
2796                                                   : LastOfChild.TotalLength + 1;
2797     }
2798     if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
2799         (Prev->Children.size() == 1 &&
2800          Prev->Children[0]->First->MustBreakBefore) ||
2801         Current->IsMultiline)
2802       Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
2803     else
2804       Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
2805                              ChildSize + Current->SpacesRequiredBefore;
2806 
2807     if (Current->is(TT_CtorInitializerColon))
2808       InFunctionDecl = false;
2809 
2810     // FIXME: Only calculate this if CanBreakBefore is true once static
2811     // initializers etc. are sorted out.
2812     // FIXME: Move magic numbers to a better place.
2813 
2814     // Reduce penalty for aligning ObjC method arguments using the colon
2815     // alignment as this is the canonical way (still prefer fitting everything
2816     // into one line if possible). Trying to fit a whole expression into one
2817     // line should not force other line breaks (e.g. when ObjC method
2818     // expression is a part of other expression).
2819     Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl);
2820     if (Style.Language == FormatStyle::LK_ObjC &&
2821         Current->is(TT_SelectorName) && Current->ParameterIndex > 0) {
2822       if (Current->ParameterIndex == 1)
2823         Current->SplitPenalty += 5 * Current->BindingStrength;
2824     } else {
2825       Current->SplitPenalty += 20 * Current->BindingStrength;
2826     }
2827 
2828     Current = Current->Next;
2829   }
2830 
2831   calculateUnbreakableTailLengths(Line);
2832   unsigned IndentLevel = Line.Level;
2833   for (Current = Line.First; Current != nullptr; Current = Current->Next) {
2834     if (Current->Role)
2835       Current->Role->precomputeFormattingInfos(Current);
2836     if (Current->MatchingParen &&
2837         Current->MatchingParen->opensBlockOrBlockTypeList(Style) &&
2838         IndentLevel > 0)
2839       --IndentLevel;
2840     Current->IndentLevel = IndentLevel;
2841     if (Current->opensBlockOrBlockTypeList(Style))
2842       ++IndentLevel;
2843   }
2844 
2845   LLVM_DEBUG({ printDebugInfo(Line); });
2846 }
2847 
2848 void TokenAnnotator::calculateUnbreakableTailLengths(
2849     AnnotatedLine &Line) const {
2850   unsigned UnbreakableTailLength = 0;
2851   FormatToken *Current = Line.Last;
2852   while (Current) {
2853     Current->UnbreakableTailLength = UnbreakableTailLength;
2854     if (Current->CanBreakBefore ||
2855         Current->isOneOf(tok::comment, tok::string_literal)) {
2856       UnbreakableTailLength = 0;
2857     } else {
2858       UnbreakableTailLength +=
2859           Current->ColumnWidth + Current->SpacesRequiredBefore;
2860     }
2861     Current = Current->Previous;
2862   }
2863 }
2864 
2865 void TokenAnnotator::calculateArrayInitializerColumnList(
2866     AnnotatedLine &Line) const {
2867   if (Line.First == Line.Last)
2868     return;
2869   auto *CurrentToken = Line.First;
2870   CurrentToken->ArrayInitializerLineStart = true;
2871   unsigned Depth = 0;
2872   while (CurrentToken != nullptr && CurrentToken != Line.Last) {
2873     if (CurrentToken->is(tok::l_brace)) {
2874       CurrentToken->IsArrayInitializer = true;
2875       if (CurrentToken->Next != nullptr)
2876         CurrentToken->Next->MustBreakBefore = true;
2877       CurrentToken =
2878           calculateInitializerColumnList(Line, CurrentToken->Next, Depth + 1);
2879     } else {
2880       CurrentToken = CurrentToken->Next;
2881     }
2882   }
2883 }
2884 
2885 FormatToken *TokenAnnotator::calculateInitializerColumnList(
2886     AnnotatedLine &Line, FormatToken *CurrentToken, unsigned Depth) const {
2887   while (CurrentToken != nullptr && CurrentToken != Line.Last) {
2888     if (CurrentToken->is(tok::l_brace))
2889       ++Depth;
2890     else if (CurrentToken->is(tok::r_brace))
2891       --Depth;
2892     if (Depth == 2 && CurrentToken->isOneOf(tok::l_brace, tok::comma)) {
2893       CurrentToken = CurrentToken->Next;
2894       if (CurrentToken == nullptr)
2895         break;
2896       CurrentToken->StartsColumn = true;
2897       CurrentToken = CurrentToken->Previous;
2898     }
2899     CurrentToken = CurrentToken->Next;
2900   }
2901   return CurrentToken;
2902 }
2903 
2904 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
2905                                       const FormatToken &Tok,
2906                                       bool InFunctionDecl) const {
2907   const FormatToken &Left = *Tok.Previous;
2908   const FormatToken &Right = Tok;
2909 
2910   if (Left.is(tok::semi))
2911     return 0;
2912 
2913   if (Style.Language == FormatStyle::LK_Java) {
2914     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
2915       return 1;
2916     if (Right.is(Keywords.kw_implements))
2917       return 2;
2918     if (Left.is(tok::comma) && Left.NestingLevel == 0)
2919       return 3;
2920   } else if (Style.isJavaScript()) {
2921     if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
2922       return 100;
2923     if (Left.is(TT_JsTypeColon))
2924       return 35;
2925     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2926         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2927       return 100;
2928     // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()".
2929     if (Left.opensScope() && Right.closesScope())
2930       return 200;
2931   }
2932 
2933   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2934     return 1;
2935   if (Right.is(tok::l_square)) {
2936     if (Style.Language == FormatStyle::LK_Proto)
2937       return 1;
2938     if (Left.is(tok::r_square))
2939       return 200;
2940     // Slightly prefer formatting local lambda definitions like functions.
2941     if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
2942       return 35;
2943     if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2944                        TT_ArrayInitializerLSquare,
2945                        TT_DesignatedInitializerLSquare, TT_AttributeSquare))
2946       return 500;
2947   }
2948 
2949   if (Left.is(tok::coloncolon) ||
2950       (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
2951     return 500;
2952   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2953       Right.is(tok::kw_operator)) {
2954     if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
2955       return 3;
2956     if (Left.is(TT_StartOfName))
2957       return 110;
2958     if (InFunctionDecl && Right.NestingLevel == 0)
2959       return Style.PenaltyReturnTypeOnItsOwnLine;
2960     return 200;
2961   }
2962   if (Right.is(TT_PointerOrReference))
2963     return 190;
2964   if (Right.is(TT_LambdaArrow))
2965     return 110;
2966   if (Left.is(tok::equal) && Right.is(tok::l_brace))
2967     return 160;
2968   if (Left.is(TT_CastRParen))
2969     return 100;
2970   if (Left.isOneOf(tok::kw_class, tok::kw_struct))
2971     return 5000;
2972   if (Left.is(tok::comment))
2973     return 1000;
2974 
2975   if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon,
2976                    TT_CtorInitializerColon))
2977     return 2;
2978 
2979   if (Right.isMemberAccess()) {
2980     // Breaking before the "./->" of a chained call/member access is reasonably
2981     // cheap, as formatting those with one call per line is generally
2982     // desirable. In particular, it should be cheaper to break before the call
2983     // than it is to break inside a call's parameters, which could lead to weird
2984     // "hanging" indents. The exception is the very last "./->" to support this
2985     // frequent pattern:
2986     //
2987     //   aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
2988     //       dddddddd);
2989     //
2990     // which might otherwise be blown up onto many lines. Here, clang-format
2991     // won't produce "hanging" indents anyway as there is no other trailing
2992     // call.
2993     //
2994     // Also apply higher penalty is not a call as that might lead to a wrapping
2995     // like:
2996     //
2997     //   aaaaaaa
2998     //       .aaaaaaaaa.bbbbbbbb(cccccccc);
2999     return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
3000                ? 150
3001                : 35;
3002   }
3003 
3004   if (Right.is(TT_TrailingAnnotation) &&
3005       (!Right.Next || Right.Next->isNot(tok::l_paren))) {
3006     // Moving trailing annotations to the next line is fine for ObjC method
3007     // declarations.
3008     if (Line.startsWith(TT_ObjCMethodSpecifier))
3009       return 10;
3010     // Generally, breaking before a trailing annotation is bad unless it is
3011     // function-like. It seems to be especially preferable to keep standard
3012     // annotations (i.e. "const", "final" and "override") on the same line.
3013     // Use a slightly higher penalty after ")" so that annotations like
3014     // "const override" are kept together.
3015     bool is_short_annotation = Right.TokenText.size() < 10;
3016     return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
3017   }
3018 
3019   // In for-loops, prefer breaking at ',' and ';'.
3020   if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
3021     return 4;
3022 
3023   // In Objective-C method expressions, prefer breaking before "param:" over
3024   // breaking after it.
3025   if (Right.is(TT_SelectorName))
3026     return 0;
3027   if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
3028     return Line.MightBeFunctionDecl ? 50 : 500;
3029 
3030   // In Objective-C type declarations, avoid breaking after the category's
3031   // open paren (we'll prefer breaking after the protocol list's opening
3032   // angle bracket, if present).
3033   if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous &&
3034       Left.Previous->isOneOf(tok::identifier, tok::greater))
3035     return 500;
3036 
3037   if (Left.is(tok::l_paren) && Style.PenaltyBreakOpenParenthesis != 0)
3038     return Style.PenaltyBreakOpenParenthesis;
3039   if (Left.is(tok::l_paren) && InFunctionDecl &&
3040       Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
3041     return 100;
3042   if (Left.is(tok::l_paren) && Left.Previous &&
3043       (Left.Previous->is(tok::kw_for) || Left.Previous->isIf()))
3044     return 1000;
3045   if (Left.is(tok::equal) && InFunctionDecl)
3046     return 110;
3047   if (Right.is(tok::r_brace))
3048     return 1;
3049   if (Left.is(TT_TemplateOpener))
3050     return 100;
3051   if (Left.opensScope()) {
3052     // If we aren't aligning after opening parens/braces we can always break
3053     // here unless the style does not want us to place all arguments on the
3054     // next line.
3055     if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign &&
3056         (Left.ParameterCount <= 1 || Style.AllowAllArgumentsOnNextLine))
3057       return 0;
3058     if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle)
3059       return 19;
3060     return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
3061                                    : 19;
3062   }
3063   if (Left.is(TT_JavaAnnotation))
3064     return 50;
3065 
3066   if (Left.is(TT_UnaryOperator))
3067     return 60;
3068   if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
3069       Left.Previous->isLabelString() &&
3070       (Left.NextOperator || Left.OperatorIndex != 0))
3071     return 50;
3072   if (Right.is(tok::plus) && Left.isLabelString() &&
3073       (Right.NextOperator || Right.OperatorIndex != 0))
3074     return 25;
3075   if (Left.is(tok::comma))
3076     return 1;
3077   if (Right.is(tok::lessless) && Left.isLabelString() &&
3078       (Right.NextOperator || Right.OperatorIndex != 1))
3079     return 25;
3080   if (Right.is(tok::lessless)) {
3081     // Breaking at a << is really cheap.
3082     if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
3083       // Slightly prefer to break before the first one in log-like statements.
3084       return 2;
3085     return 1;
3086   }
3087   if (Left.ClosesTemplateDeclaration)
3088     return Style.PenaltyBreakTemplateDeclaration;
3089   if (Left.ClosesRequiresClause)
3090     return 0;
3091   if (Left.is(TT_ConditionalExpr))
3092     return prec::Conditional;
3093   prec::Level Level = Left.getPrecedence();
3094   if (Level == prec::Unknown)
3095     Level = Right.getPrecedence();
3096   if (Level == prec::Assignment)
3097     return Style.PenaltyBreakAssignment;
3098   if (Level != prec::Unknown)
3099     return Level;
3100 
3101   return 3;
3102 }
3103 
3104 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const {
3105   if (Style.SpaceBeforeParens == FormatStyle::SBPO_Always)
3106     return true;
3107   if (Right.is(TT_OverloadedOperatorLParen) &&
3108       Style.SpaceBeforeParensOptions.AfterOverloadedOperator)
3109     return true;
3110   if (Style.SpaceBeforeParensOptions.BeforeNonEmptyParentheses &&
3111       Right.ParameterCount > 0)
3112     return true;
3113   return false;
3114 }
3115 
3116 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
3117                                           const FormatToken &Left,
3118                                           const FormatToken &Right) const {
3119   if (Left.is(tok::kw_return) &&
3120       !Right.isOneOf(tok::semi, tok::r_paren, tok::hashhash))
3121     return true;
3122   if (Style.isJson() && Left.is(tok::string_literal) && Right.is(tok::colon))
3123     return false;
3124   if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java)
3125     return true;
3126   if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
3127       Left.Tok.getObjCKeywordID() == tok::objc_property)
3128     return true;
3129   if (Right.is(tok::hashhash))
3130     return Left.is(tok::hash);
3131   if (Left.isOneOf(tok::hashhash, tok::hash))
3132     return Right.is(tok::hash);
3133   if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) ||
3134       (Left.is(tok::l_brace) && Left.isNot(BK_Block) &&
3135        Right.is(tok::r_brace) && Right.isNot(BK_Block)))
3136     return Style.SpaceInEmptyParentheses;
3137   if (Style.SpacesInConditionalStatement) {
3138     const FormatToken *LeftParen = nullptr;
3139     if (Left.is(tok::l_paren))
3140       LeftParen = &Left;
3141     else if (Right.is(tok::r_paren) && Right.MatchingParen)
3142       LeftParen = Right.MatchingParen;
3143     if (LeftParen && LeftParen->Previous &&
3144         isKeywordWithCondition(*LeftParen->Previous))
3145       return true;
3146   }
3147 
3148   // auto{x} auto(x)
3149   if (Left.is(tok::kw_auto) && Right.isOneOf(tok::l_paren, tok::l_brace))
3150     return false;
3151 
3152   // operator co_await(x)
3153   if (Right.is(tok::l_paren) && Left.is(tok::kw_co_await) && Left.Previous &&
3154       Left.Previous->is(tok::kw_operator))
3155     return false;
3156   // co_await (x), co_yield (x), co_return (x)
3157   if (Left.isOneOf(tok::kw_co_await, tok::kw_co_yield, tok::kw_co_return) &&
3158       !Right.isOneOf(tok::semi, tok::r_paren))
3159     return true;
3160 
3161   if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
3162     return (Right.is(TT_CastRParen) ||
3163             (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
3164                ? Style.SpacesInCStyleCastParentheses
3165                : Style.SpacesInParentheses;
3166   if (Right.isOneOf(tok::semi, tok::comma))
3167     return false;
3168   if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) {
3169     bool IsLightweightGeneric = Right.MatchingParen &&
3170                                 Right.MatchingParen->Next &&
3171                                 Right.MatchingParen->Next->is(tok::colon);
3172     return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList;
3173   }
3174   if (Right.is(tok::less) && Left.is(tok::kw_template))
3175     return Style.SpaceAfterTemplateKeyword;
3176   if (Left.isOneOf(tok::exclaim, tok::tilde))
3177     return false;
3178   if (Left.is(tok::at) &&
3179       Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
3180                     tok::numeric_constant, tok::l_paren, tok::l_brace,
3181                     tok::kw_true, tok::kw_false))
3182     return false;
3183   if (Left.is(tok::colon))
3184     return !Left.is(TT_ObjCMethodExpr);
3185   if (Left.is(tok::coloncolon))
3186     return false;
3187   if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) {
3188     if (Style.Language == FormatStyle::LK_TextProto ||
3189         (Style.Language == FormatStyle::LK_Proto &&
3190          (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) {
3191       // Format empty list as `<>`.
3192       if (Left.is(tok::less) && Right.is(tok::greater))
3193         return false;
3194       return !Style.Cpp11BracedListStyle;
3195     }
3196     return false;
3197   }
3198   if (Right.is(tok::ellipsis))
3199     return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
3200                                     Left.Previous->is(tok::kw_case));
3201   if (Left.is(tok::l_square) && Right.is(tok::amp))
3202     return Style.SpacesInSquareBrackets;
3203   if (Right.is(TT_PointerOrReference)) {
3204     if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) {
3205       if (!Left.MatchingParen)
3206         return true;
3207       FormatToken *TokenBeforeMatchingParen =
3208           Left.MatchingParen->getPreviousNonComment();
3209       if (!TokenBeforeMatchingParen || !Left.is(TT_TypeDeclarationParen))
3210         return true;
3211     }
3212     // Add a space if the previous token is a pointer qualifier or the closing
3213     // parenthesis of __attribute__(()) expression and the style requires spaces
3214     // after pointer qualifiers.
3215     if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_After ||
3216          Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) &&
3217         (Left.is(TT_AttributeParen) || Left.canBePointerOrReferenceQualifier()))
3218       return true;
3219     if (Left.Tok.isLiteral())
3220       return true;
3221     // for (auto a = 0, b = 0; const auto & c : {1, 2, 3})
3222     if (Left.isTypeOrIdentifier() && Right.Next && Right.Next->Next &&
3223         Right.Next->Next->is(TT_RangeBasedForLoopColon))
3224       return getTokenPointerOrReferenceAlignment(Right) !=
3225              FormatStyle::PAS_Left;
3226     return !Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
3227            (getTokenPointerOrReferenceAlignment(Right) !=
3228                 FormatStyle::PAS_Left ||
3229             (Line.IsMultiVariableDeclStmt &&
3230              (Left.NestingLevel == 0 ||
3231               (Left.NestingLevel == 1 && startsWithInitStatement(Line)))));
3232   }
3233   if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
3234       (!Left.is(TT_PointerOrReference) ||
3235        (getTokenPointerOrReferenceAlignment(Left) != FormatStyle::PAS_Right &&
3236         !Line.IsMultiVariableDeclStmt)))
3237     return true;
3238   if (Left.is(TT_PointerOrReference)) {
3239     // Add a space if the next token is a pointer qualifier and the style
3240     // requires spaces before pointer qualifiers.
3241     if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Before ||
3242          Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) &&
3243         Right.canBePointerOrReferenceQualifier())
3244       return true;
3245     // & 1
3246     if (Right.Tok.isLiteral())
3247       return true;
3248     // & /* comment
3249     if (Right.is(TT_BlockComment))
3250       return true;
3251     // foo() -> const Bar * override/final
3252     if (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
3253         !Right.is(TT_StartOfName))
3254       return true;
3255     // & {
3256     if (Right.is(tok::l_brace) && Right.is(BK_Block))
3257       return true;
3258     // for (auto a = 0, b = 0; const auto& c : {1, 2, 3})
3259     if (Left.Previous && Left.Previous->isTypeOrIdentifier() && Right.Next &&
3260         Right.Next->is(TT_RangeBasedForLoopColon))
3261       return getTokenPointerOrReferenceAlignment(Left) !=
3262              FormatStyle::PAS_Right;
3263     if (Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
3264                       tok::l_paren))
3265       return false;
3266     if (getTokenPointerOrReferenceAlignment(Left) == FormatStyle::PAS_Right)
3267       return false;
3268     // FIXME: Setting IsMultiVariableDeclStmt for the whole line is error-prone,
3269     // because it does not take into account nested scopes like lambdas.
3270     // In multi-variable declaration statements, attach */& to the variable
3271     // independently of the style. However, avoid doing it if we are in a nested
3272     // scope, e.g. lambda. We still need to special-case statements with
3273     // initializers.
3274     if (Line.IsMultiVariableDeclStmt &&
3275         (Left.NestingLevel == Line.First->NestingLevel ||
3276          ((Left.NestingLevel == Line.First->NestingLevel + 1) &&
3277           startsWithInitStatement(Line))))
3278       return false;
3279     return Left.Previous && !Left.Previous->isOneOf(
3280                                 tok::l_paren, tok::coloncolon, tok::l_square);
3281   }
3282   // Ensure right pointer alignment with ellipsis e.g. int *...P
3283   if (Left.is(tok::ellipsis) && Left.Previous &&
3284       Left.Previous->isOneOf(tok::star, tok::amp, tok::ampamp))
3285     return Style.PointerAlignment != FormatStyle::PAS_Right;
3286 
3287   if (Right.is(tok::star) && Left.is(tok::l_paren))
3288     return false;
3289   if (Left.is(tok::star) && Right.isOneOf(tok::star, tok::amp, tok::ampamp))
3290     return false;
3291   if (Right.isOneOf(tok::star, tok::amp, tok::ampamp)) {
3292     const FormatToken *Previous = &Left;
3293     while (Previous && !Previous->is(tok::kw_operator)) {
3294       if (Previous->is(tok::identifier) || Previous->isSimpleTypeSpecifier()) {
3295         Previous = Previous->getPreviousNonComment();
3296         continue;
3297       }
3298       if (Previous->is(TT_TemplateCloser) && Previous->MatchingParen) {
3299         Previous = Previous->MatchingParen->getPreviousNonComment();
3300         continue;
3301       }
3302       if (Previous->is(tok::coloncolon)) {
3303         Previous = Previous->getPreviousNonComment();
3304         continue;
3305       }
3306       break;
3307     }
3308     // Space between the type and the * in:
3309     //   operator void*()
3310     //   operator char*()
3311     //   operator void const*()
3312     //   operator void volatile*()
3313     //   operator /*comment*/ const char*()
3314     //   operator volatile /*comment*/ char*()
3315     //   operator Foo*()
3316     //   operator C<T>*()
3317     //   operator std::Foo*()
3318     //   operator C<T>::D<U>*()
3319     // dependent on PointerAlignment style.
3320     if (Previous) {
3321       if (Previous->endsSequence(tok::kw_operator))
3322         return Style.PointerAlignment != FormatStyle::PAS_Left;
3323       if (Previous->is(tok::kw_const) || Previous->is(tok::kw_volatile))
3324         return (Style.PointerAlignment != FormatStyle::PAS_Left) ||
3325                (Style.SpaceAroundPointerQualifiers ==
3326                 FormatStyle::SAPQ_After) ||
3327                (Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both);
3328     }
3329   }
3330   const auto SpaceRequiredForArrayInitializerLSquare =
3331       [](const FormatToken &LSquareTok, const FormatStyle &Style) {
3332         return Style.SpacesInContainerLiterals ||
3333                ((Style.Language == FormatStyle::LK_Proto ||
3334                  Style.Language == FormatStyle::LK_TextProto) &&
3335                 !Style.Cpp11BracedListStyle &&
3336                 LSquareTok.endsSequence(tok::l_square, tok::colon,
3337                                         TT_SelectorName));
3338       };
3339   if (Left.is(tok::l_square))
3340     return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) &&
3341             SpaceRequiredForArrayInitializerLSquare(Left, Style)) ||
3342            (Left.isOneOf(TT_ArraySubscriptLSquare, TT_StructuredBindingLSquare,
3343                          TT_LambdaLSquare) &&
3344             Style.SpacesInSquareBrackets && Right.isNot(tok::r_square));
3345   if (Right.is(tok::r_square))
3346     return Right.MatchingParen &&
3347            ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) &&
3348              SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen,
3349                                                      Style)) ||
3350             (Style.SpacesInSquareBrackets &&
3351              Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare,
3352                                           TT_StructuredBindingLSquare,
3353                                           TT_LambdaLSquare)) ||
3354             Right.MatchingParen->is(TT_AttributeParen));
3355   if (Right.is(tok::l_square) &&
3356       !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
3357                      TT_DesignatedInitializerLSquare,
3358                      TT_StructuredBindingLSquare, TT_AttributeSquare) &&
3359       !Left.isOneOf(tok::numeric_constant, TT_DictLiteral) &&
3360       !(!Left.is(tok::r_square) && Style.SpaceBeforeSquareBrackets &&
3361         Right.is(TT_ArraySubscriptLSquare)))
3362     return false;
3363   if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
3364     return !Left.Children.empty(); // No spaces in "{}".
3365   if ((Left.is(tok::l_brace) && Left.isNot(BK_Block)) ||
3366       (Right.is(tok::r_brace) && Right.MatchingParen &&
3367        Right.MatchingParen->isNot(BK_Block)))
3368     return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true;
3369   if (Left.is(TT_BlockComment))
3370     // No whitespace in x(/*foo=*/1), except for JavaScript.
3371     return Style.isJavaScript() || !Left.TokenText.endswith("=*/");
3372 
3373   // Space between template and attribute.
3374   // e.g. template <typename T> [[nodiscard]] ...
3375   if (Left.is(TT_TemplateCloser) && Right.is(TT_AttributeSquare))
3376     return true;
3377   // Space before parentheses common for all languages
3378   if (Right.is(tok::l_paren)) {
3379     if (Left.is(TT_TemplateCloser) && Right.isNot(TT_FunctionTypeLParen))
3380       return spaceRequiredBeforeParens(Right);
3381     if (Left.isOneOf(TT_RequiresClause, TT_RequiresClauseInARequiresExpression))
3382       return Style.SpaceBeforeParensOptions.AfterRequiresInClause ||
3383              spaceRequiredBeforeParens(Right);
3384     if (Left.is(TT_RequiresExpression))
3385       return Style.SpaceBeforeParensOptions.AfterRequiresInExpression ||
3386              spaceRequiredBeforeParens(Right);
3387     if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) ||
3388         (Left.is(tok::r_square) && Left.is(TT_AttributeSquare)))
3389       return true;
3390     if (Left.is(TT_ForEachMacro))
3391       return Style.SpaceBeforeParensOptions.AfterForeachMacros ||
3392              spaceRequiredBeforeParens(Right);
3393     if (Left.is(TT_IfMacro))
3394       return Style.SpaceBeforeParensOptions.AfterIfMacros ||
3395              spaceRequiredBeforeParens(Right);
3396     if (Line.Type == LT_ObjCDecl)
3397       return true;
3398     if (Left.is(tok::semi))
3399       return true;
3400     if (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while, tok::kw_switch,
3401                      tok::kw_case, TT_ForEachMacro, TT_ObjCForIn) ||
3402         Left.isIf(Line.Type != LT_PreprocessorDirective))
3403       return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3404              spaceRequiredBeforeParens(Right);
3405 
3406     // TODO add Operator overloading specific Options to
3407     // SpaceBeforeParensOptions
3408     if (Right.is(TT_OverloadedOperatorLParen))
3409       return spaceRequiredBeforeParens(Right);
3410     // Function declaration or definition
3411     if (Line.MightBeFunctionDecl && (Left.is(TT_FunctionDeclarationName))) {
3412       if (Line.mightBeFunctionDefinition())
3413         return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName ||
3414                spaceRequiredBeforeParens(Right);
3415       else
3416         return Style.SpaceBeforeParensOptions.AfterFunctionDeclarationName ||
3417                spaceRequiredBeforeParens(Right);
3418     }
3419     // Lambda
3420     if (Line.Type != LT_PreprocessorDirective && Left.is(tok::r_square) &&
3421         Left.MatchingParen && Left.MatchingParen->is(TT_LambdaLSquare))
3422       return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName ||
3423              spaceRequiredBeforeParens(Right);
3424     if (!Left.Previous || Left.Previous->isNot(tok::period)) {
3425       if (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch))
3426         return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3427                spaceRequiredBeforeParens(Right);
3428       if (Left.isOneOf(tok::kw_new, tok::kw_delete))
3429         return ((!Line.MightBeFunctionDecl || !Left.Previous) &&
3430                 Style.SpaceBeforeParens != FormatStyle::SBPO_Never) ||
3431                spaceRequiredBeforeParens(Right);
3432 
3433       if (Left.is(tok::r_square) && Left.MatchingParen &&
3434           Left.MatchingParen->Previous &&
3435           Left.MatchingParen->Previous->is(tok::kw_delete))
3436         return (Style.SpaceBeforeParens != FormatStyle::SBPO_Never) ||
3437                spaceRequiredBeforeParens(Right);
3438     }
3439     // Handle builtins like identifiers.
3440     if (Line.Type != LT_PreprocessorDirective &&
3441         (Left.Tok.getIdentifierInfo() || Left.is(tok::r_paren)))
3442       return spaceRequiredBeforeParens(Right);
3443     return false;
3444   }
3445   if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
3446     return false;
3447   if (Right.is(TT_UnaryOperator))
3448     return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
3449            (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
3450   if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
3451                     tok::r_paren) ||
3452        Left.isSimpleTypeSpecifier()) &&
3453       Right.is(tok::l_brace) && Right.getNextNonComment() &&
3454       Right.isNot(BK_Block))
3455     return false;
3456   if (Left.is(tok::period) || Right.is(tok::period))
3457     return false;
3458   // u#str, U#str, L#str, u8#str
3459   // uR#str, UR#str, LR#str, u8R#str
3460   if (Right.is(tok::hash) && Left.is(tok::identifier) &&
3461       (Left.TokenText == "L" || Left.TokenText == "u" ||
3462        Left.TokenText == "U" || Left.TokenText == "u8" ||
3463        Left.TokenText == "LR" || Left.TokenText == "uR" ||
3464        Left.TokenText == "UR" || Left.TokenText == "u8R"))
3465     return false;
3466   if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
3467       Left.MatchingParen->Previous &&
3468       (Left.MatchingParen->Previous->is(tok::period) ||
3469        Left.MatchingParen->Previous->is(tok::coloncolon)))
3470     // Java call to generic function with explicit type:
3471     // A.<B<C<...>>>DoSomething();
3472     // A::<B<C<...>>>DoSomething();  // With a Java 8 method reference.
3473     return false;
3474   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
3475     return false;
3476   if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at))
3477     // Objective-C dictionary literal -> no space after opening brace.
3478     return false;
3479   if (Right.is(tok::r_brace) && Right.MatchingParen &&
3480       Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at))
3481     // Objective-C dictionary literal -> no space before closing brace.
3482     return false;
3483   if (Right.getType() == TT_TrailingAnnotation &&
3484       Right.isOneOf(tok::amp, tok::ampamp) &&
3485       Left.isOneOf(tok::kw_const, tok::kw_volatile) &&
3486       (!Right.Next || Right.Next->is(tok::semi)))
3487     // Match const and volatile ref-qualifiers without any additional
3488     // qualifiers such as
3489     // void Fn() const &;
3490     return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left;
3491 
3492   return true;
3493 }
3494 
3495 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
3496                                          const FormatToken &Right) const {
3497   const FormatToken &Left = *Right.Previous;
3498 
3499   // If the token is finalized don't touch it (as it could be in a
3500   // clang-format-off section).
3501   if (Left.Finalized)
3502     return Right.hasWhitespaceBefore();
3503 
3504   if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
3505     return true; // Never ever merge two identifiers.
3506 
3507   // Leave a space between * and /* to avoid C4138 `comment end` found outside
3508   // of comment.
3509   if (Left.is(tok::star) && Right.is(tok::comment))
3510     return true;
3511 
3512   if (Style.isCpp()) {
3513     // Space between import <iostream>.
3514     // or import .....;
3515     if (Left.is(Keywords.kw_import) && Right.isOneOf(tok::less, tok::ellipsis))
3516       return true;
3517     // Space between `module :` and `import :`.
3518     if (Left.isOneOf(Keywords.kw_module, Keywords.kw_import) &&
3519         Right.is(TT_ModulePartitionColon))
3520       return true;
3521     // No space between import foo:bar but keep a space between import :bar;
3522     if (Left.is(tok::identifier) && Right.is(TT_ModulePartitionColon))
3523       return false;
3524     // No space between :bar;
3525     if (Left.is(TT_ModulePartitionColon) &&
3526         Right.isOneOf(tok::identifier, tok::kw_private))
3527       return false;
3528     if (Left.is(tok::ellipsis) && Right.is(tok::identifier) &&
3529         Line.First->is(Keywords.kw_import))
3530       return false;
3531     // Space in __attribute__((attr)) ::type.
3532     if (Left.is(TT_AttributeParen) && Right.is(tok::coloncolon))
3533       return true;
3534 
3535     if (Left.is(tok::kw_operator))
3536       return Right.is(tok::coloncolon);
3537     if (Right.is(tok::l_brace) && Right.is(BK_BracedInit) &&
3538         !Left.opensScope() && Style.SpaceBeforeCpp11BracedList)
3539       return true;
3540     if (Left.is(tok::less) && Left.is(TT_OverloadedOperator) &&
3541         Right.is(TT_TemplateOpener))
3542       return true;
3543   } else if (Style.Language == FormatStyle::LK_Proto ||
3544              Style.Language == FormatStyle::LK_TextProto) {
3545     if (Right.is(tok::period) &&
3546         Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
3547                      Keywords.kw_repeated, Keywords.kw_extend))
3548       return true;
3549     if (Right.is(tok::l_paren) &&
3550         Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
3551       return true;
3552     if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName))
3553       return true;
3554     // Slashes occur in text protocol extension syntax: [type/type] { ... }.
3555     if (Left.is(tok::slash) || Right.is(tok::slash))
3556       return false;
3557     if (Left.MatchingParen &&
3558         Left.MatchingParen->is(TT_ProtoExtensionLSquare) &&
3559         Right.isOneOf(tok::l_brace, tok::less))
3560       return !Style.Cpp11BracedListStyle;
3561     // A percent is probably part of a formatting specification, such as %lld.
3562     if (Left.is(tok::percent))
3563       return false;
3564     // Preserve the existence of a space before a percent for cases like 0x%04x
3565     // and "%d %d"
3566     if (Left.is(tok::numeric_constant) && Right.is(tok::percent))
3567       return Right.hasWhitespaceBefore();
3568   } else if (Style.isJson()) {
3569     if (Right.is(tok::colon))
3570       return false;
3571   } else if (Style.isCSharp()) {
3572     // Require spaces around '{' and  before '}' unless they appear in
3573     // interpolated strings. Interpolated strings are merged into a single token
3574     // so cannot have spaces inserted by this function.
3575 
3576     // No space between 'this' and '['
3577     if (Left.is(tok::kw_this) && Right.is(tok::l_square))
3578       return false;
3579 
3580     // No space between 'new' and '('
3581     if (Left.is(tok::kw_new) && Right.is(tok::l_paren))
3582       return false;
3583 
3584     // Space before { (including space within '{ {').
3585     if (Right.is(tok::l_brace))
3586       return true;
3587 
3588     // Spaces inside braces.
3589     if (Left.is(tok::l_brace) && Right.isNot(tok::r_brace))
3590       return true;
3591 
3592     if (Left.isNot(tok::l_brace) && Right.is(tok::r_brace))
3593       return true;
3594 
3595     // Spaces around '=>'.
3596     if (Left.is(TT_FatArrow) || Right.is(TT_FatArrow))
3597       return true;
3598 
3599     // No spaces around attribute target colons
3600     if (Left.is(TT_AttributeColon) || Right.is(TT_AttributeColon))
3601       return false;
3602 
3603     // space between type and variable e.g. Dictionary<string,string> foo;
3604     if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName))
3605       return true;
3606 
3607     // spaces inside square brackets.
3608     if (Left.is(tok::l_square) || Right.is(tok::r_square))
3609       return Style.SpacesInSquareBrackets;
3610 
3611     // No space before ? in nullable types.
3612     if (Right.is(TT_CSharpNullable))
3613       return false;
3614 
3615     // No space before null forgiving '!'.
3616     if (Right.is(TT_NonNullAssertion))
3617       return false;
3618 
3619     // No space between consecutive commas '[,,]'.
3620     if (Left.is(tok::comma) && Right.is(tok::comma))
3621       return false;
3622 
3623     // space after var in `var (key, value)`
3624     if (Left.is(Keywords.kw_var) && Right.is(tok::l_paren))
3625       return true;
3626 
3627     // space between keywords and paren e.g. "using ("
3628     if (Right.is(tok::l_paren))
3629       if (Left.isOneOf(tok::kw_using, Keywords.kw_async, Keywords.kw_when,
3630                        Keywords.kw_lock))
3631         return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3632                spaceRequiredBeforeParens(Right);
3633 
3634     // space between method modifier and opening parenthesis of a tuple return
3635     // type
3636     if (Left.isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
3637                      tok::kw_virtual, tok::kw_extern, tok::kw_static,
3638                      Keywords.kw_internal, Keywords.kw_abstract,
3639                      Keywords.kw_sealed, Keywords.kw_override,
3640                      Keywords.kw_async, Keywords.kw_unsafe) &&
3641         Right.is(tok::l_paren))
3642       return true;
3643   } else if (Style.isJavaScript()) {
3644     if (Left.is(TT_FatArrow))
3645       return true;
3646     // for await ( ...
3647     if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous &&
3648         Left.Previous->is(tok::kw_for))
3649       return true;
3650     if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
3651         Right.MatchingParen) {
3652       const FormatToken *Next = Right.MatchingParen->getNextNonComment();
3653       // An async arrow function, for example: `x = async () => foo();`,
3654       // as opposed to calling a function called async: `x = async();`
3655       if (Next && Next->is(TT_FatArrow))
3656         return true;
3657     }
3658     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
3659         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
3660       return false;
3661     // In tagged template literals ("html`bar baz`"), there is no space between
3662     // the tag identifier and the template string.
3663     if (Keywords.IsJavaScriptIdentifier(Left,
3664                                         /* AcceptIdentifierName= */ false) &&
3665         Right.is(TT_TemplateString))
3666       return false;
3667     if (Right.is(tok::star) &&
3668         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
3669       return false;
3670     if (Right.isOneOf(tok::l_brace, tok::l_square) &&
3671         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield,
3672                      Keywords.kw_extends, Keywords.kw_implements))
3673       return true;
3674     if (Right.is(tok::l_paren)) {
3675       // JS methods can use some keywords as names (e.g. `delete()`).
3676       if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo())
3677         return false;
3678       // Valid JS method names can include keywords, e.g. `foo.delete()` or
3679       // `bar.instanceof()`. Recognize call positions by preceding period.
3680       if (Left.Previous && Left.Previous->is(tok::period) &&
3681           Left.Tok.getIdentifierInfo())
3682         return false;
3683       // Additional unary JavaScript operators that need a space after.
3684       if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof,
3685                        tok::kw_void))
3686         return true;
3687     }
3688     // `foo as const;` casts into a const type.
3689     if (Left.endsSequence(tok::kw_const, Keywords.kw_as))
3690       return false;
3691     if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
3692                       tok::kw_const) ||
3693          // "of" is only a keyword if it appears after another identifier
3694          // (e.g. as "const x of y" in a for loop), or after a destructuring
3695          // operation (const [x, y] of z, const {a, b} of c).
3696          (Left.is(Keywords.kw_of) && Left.Previous &&
3697           (Left.Previous->is(tok::identifier) ||
3698            Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) &&
3699         (!Left.Previous || !Left.Previous->is(tok::period)))
3700       return true;
3701     if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
3702         Left.Previous->is(tok::period) && Right.is(tok::l_paren))
3703       return false;
3704     if (Left.is(Keywords.kw_as) &&
3705         Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
3706       return true;
3707     if (Left.is(tok::kw_default) && Left.Previous &&
3708         Left.Previous->is(tok::kw_export))
3709       return true;
3710     if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
3711       return true;
3712     if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
3713       return false;
3714     if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
3715       return false;
3716     if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
3717         Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
3718       return false;
3719     if (Left.is(tok::ellipsis))
3720       return false;
3721     if (Left.is(TT_TemplateCloser) &&
3722         !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
3723                        Keywords.kw_implements, Keywords.kw_extends))
3724       // Type assertions ('<type>expr') are not followed by whitespace. Other
3725       // locations that should have whitespace following are identified by the
3726       // above set of follower tokens.
3727       return false;
3728     if (Right.is(TT_NonNullAssertion))
3729       return false;
3730     if (Left.is(TT_NonNullAssertion) &&
3731         Right.isOneOf(Keywords.kw_as, Keywords.kw_in))
3732       return true; // "x! as string", "x! in y"
3733   } else if (Style.Language == FormatStyle::LK_Java) {
3734     if (Left.is(tok::r_square) && Right.is(tok::l_brace))
3735       return true;
3736     if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
3737       return Style.SpaceBeforeParensOptions.AfterControlStatements ||
3738              spaceRequiredBeforeParens(Right);
3739     if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
3740                       tok::kw_protected) ||
3741          Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
3742                       Keywords.kw_native)) &&
3743         Right.is(TT_TemplateOpener))
3744       return true;
3745   }
3746   if (Left.is(TT_ImplicitStringLiteral))
3747     return Right.hasWhitespaceBefore();
3748   if (Line.Type == LT_ObjCMethodDecl) {
3749     if (Left.is(TT_ObjCMethodSpecifier))
3750       return true;
3751     if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right))
3752       // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a
3753       // keyword in Objective-C, and '+ (instancetype)new;' is a standard class
3754       // method declaration.
3755       return false;
3756   }
3757   if (Line.Type == LT_ObjCProperty &&
3758       (Right.is(tok::equal) || Left.is(tok::equal)))
3759     return false;
3760 
3761   if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
3762       Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
3763     return true;
3764   if (Left.is(tok::comma) && !Right.is(TT_OverloadedOperatorLParen))
3765     return true;
3766   if (Right.is(tok::comma))
3767     return false;
3768   if (Right.is(TT_ObjCBlockLParen))
3769     return true;
3770   if (Right.is(TT_CtorInitializerColon))
3771     return Style.SpaceBeforeCtorInitializerColon;
3772   if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon)
3773     return false;
3774   if (Right.is(TT_RangeBasedForLoopColon) &&
3775       !Style.SpaceBeforeRangeBasedForLoopColon)
3776     return false;
3777   if (Left.is(TT_BitFieldColon))
3778     return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both ||
3779            Style.BitFieldColonSpacing == FormatStyle::BFCS_After;
3780   if (Right.is(tok::colon)) {
3781     if (Line.First->isOneOf(tok::kw_default, tok::kw_case))
3782       return Style.SpaceBeforeCaseColon;
3783     const FormatToken *Next = Right.getNextNonComment();
3784     if (!Next || Next->is(tok::semi))
3785       return false;
3786     if (Right.is(TT_ObjCMethodExpr))
3787       return false;
3788     if (Left.is(tok::question))
3789       return false;
3790     if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
3791       return false;
3792     if (Right.is(TT_DictLiteral))
3793       return Style.SpacesInContainerLiterals;
3794     if (Right.is(TT_AttributeColon))
3795       return false;
3796     if (Right.is(TT_CSharpNamedArgumentColon))
3797       return false;
3798     if (Right.is(TT_BitFieldColon))
3799       return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both ||
3800              Style.BitFieldColonSpacing == FormatStyle::BFCS_Before;
3801     return true;
3802   }
3803   // Do not merge "- -" into "--".
3804   if ((Left.isOneOf(tok::minus, tok::minusminus) &&
3805        Right.isOneOf(tok::minus, tok::minusminus)) ||
3806       (Left.isOneOf(tok::plus, tok::plusplus) &&
3807        Right.isOneOf(tok::plus, tok::plusplus)))
3808     return true;
3809   if (Left.is(TT_UnaryOperator)) {
3810     if (!Right.is(tok::l_paren)) {
3811       // The alternative operators for ~ and ! are "compl" and "not".
3812       // If they are used instead, we do not want to combine them with
3813       // the token to the right, unless that is a left paren.
3814       if (Left.is(tok::exclaim) && Left.TokenText == "not")
3815         return true;
3816       if (Left.is(tok::tilde) && Left.TokenText == "compl")
3817         return true;
3818       // Lambda captures allow for a lone &, so "&]" needs to be properly
3819       // handled.
3820       if (Left.is(tok::amp) && Right.is(tok::r_square))
3821         return Style.SpacesInSquareBrackets;
3822     }
3823     return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) ||
3824            Right.is(TT_BinaryOperator);
3825   }
3826 
3827   // If the next token is a binary operator or a selector name, we have
3828   // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
3829   if (Left.is(TT_CastRParen))
3830     return Style.SpaceAfterCStyleCast ||
3831            Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
3832 
3833   auto ShouldAddSpacesInAngles = [this, &Right]() {
3834     if (this->Style.SpacesInAngles == FormatStyle::SIAS_Always)
3835       return true;
3836     if (this->Style.SpacesInAngles == FormatStyle::SIAS_Leave)
3837       return Right.hasWhitespaceBefore();
3838     return false;
3839   };
3840 
3841   if (Left.is(tok::greater) && Right.is(tok::greater)) {
3842     if (Style.Language == FormatStyle::LK_TextProto ||
3843         (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral)))
3844       return !Style.Cpp11BracedListStyle;
3845     return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
3846            ((Style.Standard < FormatStyle::LS_Cpp11) ||
3847             ShouldAddSpacesInAngles());
3848   }
3849   if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) ||
3850       Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
3851       (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod)))
3852     return false;
3853   if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) &&
3854       Right.getPrecedence() == prec::Assignment)
3855     return false;
3856   if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) &&
3857       (Left.is(tok::identifier) || Left.is(tok::kw_this)))
3858     return false;
3859   if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
3860     // Generally don't remove existing spaces between an identifier and "::".
3861     // The identifier might actually be a macro name such as ALWAYS_INLINE. If
3862     // this turns out to be too lenient, add analysis of the identifier itself.
3863     return Right.hasWhitespaceBefore();
3864   if (Right.is(tok::coloncolon) &&
3865       !Left.isOneOf(tok::l_brace, tok::comment, tok::l_paren))
3866     // Put a space between < and :: in vector< ::std::string >
3867     return (Left.is(TT_TemplateOpener) &&
3868             ((Style.Standard < FormatStyle::LS_Cpp11) ||
3869              ShouldAddSpacesInAngles())) ||
3870            !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
3871                           tok::kw___super, TT_TemplateOpener,
3872                           TT_TemplateCloser)) ||
3873            (Left.is(tok::l_paren) && Style.SpacesInParentheses);
3874   if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
3875     return ShouldAddSpacesInAngles();
3876   // Space before TT_StructuredBindingLSquare.
3877   if (Right.is(TT_StructuredBindingLSquare))
3878     return !Left.isOneOf(tok::amp, tok::ampamp) ||
3879            getTokenReferenceAlignment(Left) != FormatStyle::PAS_Right;
3880   // Space before & or && following a TT_StructuredBindingLSquare.
3881   if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) &&
3882       Right.isOneOf(tok::amp, tok::ampamp))
3883     return getTokenReferenceAlignment(Right) != FormatStyle::PAS_Left;
3884   if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
3885       (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
3886        !Right.is(tok::r_paren)))
3887     return true;
3888   if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
3889       Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
3890     return false;
3891   if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
3892       Line.startsWith(tok::hash))
3893     return true;
3894   if (Right.is(TT_TrailingUnaryOperator))
3895     return false;
3896   if (Left.is(TT_RegexLiteral))
3897     return false;
3898   return spaceRequiredBetween(Line, Left, Right);
3899 }
3900 
3901 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
3902 static bool isAllmanBrace(const FormatToken &Tok) {
3903   return Tok.is(tok::l_brace) && Tok.is(BK_Block) &&
3904          !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral);
3905 }
3906 
3907 // Returns 'true' if 'Tok' is a function argument.
3908 static bool IsFunctionArgument(const FormatToken &Tok) {
3909   return Tok.MatchingParen && Tok.MatchingParen->Next &&
3910          Tok.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren);
3911 }
3912 
3913 static bool
3914 isItAnEmptyLambdaAllowed(const FormatToken &Tok,
3915                          FormatStyle::ShortLambdaStyle ShortLambdaOption) {
3916   return Tok.Children.empty() && ShortLambdaOption != FormatStyle::SLS_None;
3917 }
3918 
3919 static bool isAllmanLambdaBrace(const FormatToken &Tok) {
3920   return Tok.is(tok::l_brace) && Tok.is(BK_Block) &&
3921          !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral);
3922 }
3923 
3924 // Returns the first token on the line that is not a comment.
3925 static const FormatToken *getFirstNonComment(const AnnotatedLine &Line) {
3926   const FormatToken *Next = Line.First;
3927   if (!Next)
3928     return Next;
3929   if (Next->is(tok::comment))
3930     Next = Next->getNextNonComment();
3931   return Next;
3932 }
3933 
3934 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
3935                                      const FormatToken &Right) const {
3936   const FormatToken &Left = *Right.Previous;
3937   if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
3938     return true;
3939 
3940   if (Style.isCSharp()) {
3941     if (Left.is(TT_FatArrow) && Right.is(tok::l_brace) &&
3942         Style.BraceWrapping.AfterFunction)
3943       return true;
3944     if (Right.is(TT_CSharpNamedArgumentColon) ||
3945         Left.is(TT_CSharpNamedArgumentColon))
3946       return false;
3947     if (Right.is(TT_CSharpGenericTypeConstraint))
3948       return true;
3949     if (Right.Next && Right.Next->is(TT_FatArrow) &&
3950         (Right.is(tok::numeric_constant) ||
3951          (Right.is(tok::identifier) && Right.TokenText == "_")))
3952       return true;
3953 
3954     // Break after C# [...] and before public/protected/private/internal.
3955     if (Left.is(TT_AttributeSquare) && Left.is(tok::r_square) &&
3956         (Right.isAccessSpecifier(/*ColonRequired=*/false) ||
3957          Right.is(Keywords.kw_internal)))
3958       return true;
3959     // Break between ] and [ but only when there are really 2 attributes.
3960     if (Left.is(TT_AttributeSquare) && Right.is(TT_AttributeSquare) &&
3961         Left.is(tok::r_square) && Right.is(tok::l_square))
3962       return true;
3963 
3964   } else if (Style.isJavaScript()) {
3965     // FIXME: This might apply to other languages and token kinds.
3966     if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
3967         Left.Previous->is(tok::string_literal))
3968       return true;
3969     if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
3970         Left.Previous && Left.Previous->is(tok::equal) &&
3971         Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
3972                             tok::kw_const) &&
3973         // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
3974         // above.
3975         !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
3976       // Object literals on the top level of a file are treated as "enum-style".
3977       // Each key/value pair is put on a separate line, instead of bin-packing.
3978       return true;
3979     if (Left.is(tok::l_brace) && Line.Level == 0 &&
3980         (Line.startsWith(tok::kw_enum) ||
3981          Line.startsWith(tok::kw_const, tok::kw_enum) ||
3982          Line.startsWith(tok::kw_export, tok::kw_enum) ||
3983          Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum)))
3984       // JavaScript top-level enum key/value pairs are put on separate lines
3985       // instead of bin-packing.
3986       return true;
3987     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && Left.Previous &&
3988         Left.Previous->is(TT_FatArrow)) {
3989       // JS arrow function (=> {...}).
3990       switch (Style.AllowShortLambdasOnASingleLine) {
3991       case FormatStyle::SLS_All:
3992         return false;
3993       case FormatStyle::SLS_None:
3994         return true;
3995       case FormatStyle::SLS_Empty:
3996         return !Left.Children.empty();
3997       case FormatStyle::SLS_Inline:
3998         // allow one-lining inline (e.g. in function call args) and empty arrow
3999         // functions.
4000         return (Left.NestingLevel == 0 && Line.Level == 0) &&
4001                !Left.Children.empty();
4002       }
4003       llvm_unreachable("Unknown FormatStyle::ShortLambdaStyle enum");
4004     }
4005 
4006     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
4007         !Left.Children.empty())
4008       // Support AllowShortFunctionsOnASingleLine for JavaScript.
4009       return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
4010              Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
4011              (Left.NestingLevel == 0 && Line.Level == 0 &&
4012               Style.AllowShortFunctionsOnASingleLine &
4013                   FormatStyle::SFS_InlineOnly);
4014   } else if (Style.Language == FormatStyle::LK_Java) {
4015     if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
4016         Right.Next->is(tok::string_literal))
4017       return true;
4018   } else if (Style.Language == FormatStyle::LK_Cpp ||
4019              Style.Language == FormatStyle::LK_ObjC ||
4020              Style.Language == FormatStyle::LK_Proto ||
4021              Style.Language == FormatStyle::LK_TableGen ||
4022              Style.Language == FormatStyle::LK_TextProto) {
4023     if (Left.isStringLiteral() && Right.isStringLiteral())
4024       return true;
4025   }
4026 
4027   // Basic JSON newline processing.
4028   if (Style.isJson()) {
4029     // Always break after a JSON record opener.
4030     // {
4031     // }
4032     if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace))
4033       return true;
4034     // Always break after a JSON array opener.
4035     // [
4036     // ]
4037     if (Left.is(TT_ArrayInitializerLSquare) && Left.is(tok::l_square) &&
4038         !Right.is(tok::r_square))
4039       return true;
4040     // Always break after successive entries.
4041     // 1,
4042     // 2
4043     if (Left.is(tok::comma))
4044       return true;
4045   }
4046 
4047   // If the last token before a '}', ']', or ')' is a comma or a trailing
4048   // comment, the intention is to insert a line break after it in order to make
4049   // shuffling around entries easier. Import statements, especially in
4050   // JavaScript, can be an exception to this rule.
4051   if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) {
4052     const FormatToken *BeforeClosingBrace = nullptr;
4053     if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
4054          (Style.isJavaScript() && Left.is(tok::l_paren))) &&
4055         Left.isNot(BK_Block) && Left.MatchingParen)
4056       BeforeClosingBrace = Left.MatchingParen->Previous;
4057     else if (Right.MatchingParen &&
4058              (Right.MatchingParen->isOneOf(tok::l_brace,
4059                                            TT_ArrayInitializerLSquare) ||
4060               (Style.isJavaScript() && Right.MatchingParen->is(tok::l_paren))))
4061       BeforeClosingBrace = &Left;
4062     if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
4063                                BeforeClosingBrace->isTrailingComment()))
4064       return true;
4065   }
4066 
4067   if (Right.is(tok::comment))
4068     return Left.isNot(BK_BracedInit) && Left.isNot(TT_CtorInitializerColon) &&
4069            (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
4070   if (Left.isTrailingComment())
4071     return true;
4072   if (Left.IsUnterminatedLiteral)
4073     return true;
4074   if (Right.is(tok::lessless) && Right.Next && Left.is(tok::string_literal) &&
4075       Right.Next->is(tok::string_literal))
4076     return true;
4077   if (Right.is(TT_RequiresClause)) {
4078     switch (Style.RequiresClausePosition) {
4079     case FormatStyle::RCPS_OwnLine:
4080     case FormatStyle::RCPS_WithFollowing:
4081       return true;
4082     default:
4083       break;
4084     }
4085   }
4086   // Can break after template<> declaration
4087   if (Left.ClosesTemplateDeclaration && Left.MatchingParen &&
4088       Left.MatchingParen->NestingLevel == 0) {
4089     // Put concepts on the next line e.g.
4090     // template<typename T>
4091     // concept ...
4092     if (Right.is(tok::kw_concept))
4093       return Style.BreakBeforeConceptDeclarations == FormatStyle::BBCDS_Always;
4094     return Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes;
4095   }
4096   if (Left.ClosesRequiresClause && Right.isNot(tok::semi)) {
4097     switch (Style.RequiresClausePosition) {
4098     case FormatStyle::RCPS_OwnLine:
4099     case FormatStyle::RCPS_WithPreceding:
4100       return true;
4101     default:
4102       break;
4103     }
4104   }
4105   if (Style.PackConstructorInitializers == FormatStyle::PCIS_Never) {
4106     if (Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeColon &&
4107         (Left.is(TT_CtorInitializerComma) || Right.is(TT_CtorInitializerColon)))
4108       return true;
4109 
4110     if (Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon &&
4111         Left.isOneOf(TT_CtorInitializerColon, TT_CtorInitializerComma))
4112       return true;
4113   }
4114   if (Style.PackConstructorInitializers < FormatStyle::PCIS_CurrentLine &&
4115       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
4116       Right.isOneOf(TT_CtorInitializerComma, TT_CtorInitializerColon))
4117     return true;
4118   // Break only if we have multiple inheritance.
4119   if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma &&
4120       Right.is(TT_InheritanceComma))
4121     return true;
4122   if (Style.BreakInheritanceList == FormatStyle::BILS_AfterComma &&
4123       Left.is(TT_InheritanceComma))
4124     return true;
4125   if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
4126     // Multiline raw string literals are special wrt. line breaks. The author
4127     // has made a deliberate choice and might have aligned the contents of the
4128     // string literal accordingly. Thus, we try keep existing line breaks.
4129     return Right.IsMultiline && Right.NewlinesBefore > 0;
4130   if ((Left.is(tok::l_brace) || (Left.is(tok::less) && Left.Previous &&
4131                                  Left.Previous->is(tok::equal))) &&
4132       Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) {
4133     // Don't put enums or option definitions onto single lines in protocol
4134     // buffers.
4135     return true;
4136   }
4137   if (Right.is(TT_InlineASMBrace))
4138     return Right.HasUnescapedNewline;
4139 
4140   if (isAllmanBrace(Left) || isAllmanBrace(Right)) {
4141     auto FirstNonComment = getFirstNonComment(Line);
4142     bool AccessSpecifier =
4143         FirstNonComment &&
4144         FirstNonComment->isOneOf(Keywords.kw_internal, tok::kw_public,
4145                                  tok::kw_private, tok::kw_protected);
4146 
4147     if (Style.BraceWrapping.AfterEnum) {
4148       if (Line.startsWith(tok::kw_enum) ||
4149           Line.startsWith(tok::kw_typedef, tok::kw_enum))
4150         return true;
4151       // Ensure BraceWrapping for `public enum A {`.
4152       if (AccessSpecifier && FirstNonComment->Next &&
4153           FirstNonComment->Next->is(tok::kw_enum))
4154         return true;
4155     }
4156 
4157     // Ensure BraceWrapping for `public interface A {`.
4158     if (Style.BraceWrapping.AfterClass &&
4159         ((AccessSpecifier && FirstNonComment->Next &&
4160           FirstNonComment->Next->is(Keywords.kw_interface)) ||
4161          Line.startsWith(Keywords.kw_interface)))
4162       return true;
4163 
4164     return (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
4165            (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
4166   }
4167 
4168   if (Left.is(TT_ObjCBlockLBrace) &&
4169       Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never)
4170     return true;
4171 
4172   // Ensure wrapping after __attribute__((XX)) and @interface etc.
4173   if (Left.is(TT_AttributeParen) && Right.is(TT_ObjCDecl))
4174     return true;
4175 
4176   if (Left.is(TT_LambdaLBrace)) {
4177     if (IsFunctionArgument(Left) &&
4178         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline)
4179       return false;
4180 
4181     if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None ||
4182         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline ||
4183         (!Left.Children.empty() &&
4184          Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty))
4185       return true;
4186   }
4187 
4188   if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace) &&
4189       Left.isOneOf(tok::star, tok::amp, tok::ampamp, TT_TemplateCloser))
4190     return true;
4191 
4192   // Put multiple Java annotation on a new line.
4193   if ((Style.Language == FormatStyle::LK_Java || Style.isJavaScript()) &&
4194       Left.is(TT_LeadingJavaAnnotation) &&
4195       Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
4196       (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
4197     return true;
4198 
4199   if (Right.is(TT_ProtoExtensionLSquare))
4200     return true;
4201 
4202   // In text proto instances if a submessage contains at least 2 entries and at
4203   // least one of them is a submessage, like A { ... B { ... } ... },
4204   // put all of the entries of A on separate lines by forcing the selector of
4205   // the submessage B to be put on a newline.
4206   //
4207   // Example: these can stay on one line:
4208   // a { scalar_1: 1 scalar_2: 2 }
4209   // a { b { key: value } }
4210   //
4211   // and these entries need to be on a new line even if putting them all in one
4212   // line is under the column limit:
4213   // a {
4214   //   scalar: 1
4215   //   b { key: value }
4216   // }
4217   //
4218   // We enforce this by breaking before a submessage field that has previous
4219   // siblings, *and* breaking before a field that follows a submessage field.
4220   //
4221   // Be careful to exclude the case  [proto.ext] { ... } since the `]` is
4222   // the TT_SelectorName there, but we don't want to break inside the brackets.
4223   //
4224   // Another edge case is @submessage { key: value }, which is a common
4225   // substitution placeholder. In this case we want to keep `@` and `submessage`
4226   // together.
4227   //
4228   // We ensure elsewhere that extensions are always on their own line.
4229   if ((Style.Language == FormatStyle::LK_Proto ||
4230        Style.Language == FormatStyle::LK_TextProto) &&
4231       Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) {
4232     // Keep `@submessage` together in:
4233     // @submessage { key: value }
4234     if (Left.is(tok::at))
4235       return false;
4236     // Look for the scope opener after selector in cases like:
4237     // selector { ...
4238     // selector: { ...
4239     // selector: @base { ...
4240     FormatToken *LBrace = Right.Next;
4241     if (LBrace && LBrace->is(tok::colon)) {
4242       LBrace = LBrace->Next;
4243       if (LBrace && LBrace->is(tok::at)) {
4244         LBrace = LBrace->Next;
4245         if (LBrace)
4246           LBrace = LBrace->Next;
4247       }
4248     }
4249     if (LBrace &&
4250         // The scope opener is one of {, [, <:
4251         // selector { ... }
4252         // selector [ ... ]
4253         // selector < ... >
4254         //
4255         // In case of selector { ... }, the l_brace is TT_DictLiteral.
4256         // In case of an empty selector {}, the l_brace is not TT_DictLiteral,
4257         // so we check for immediately following r_brace.
4258         ((LBrace->is(tok::l_brace) &&
4259           (LBrace->is(TT_DictLiteral) ||
4260            (LBrace->Next && LBrace->Next->is(tok::r_brace)))) ||
4261          LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) {
4262       // If Left.ParameterCount is 0, then this submessage entry is not the
4263       // first in its parent submessage, and we want to break before this entry.
4264       // If Left.ParameterCount is greater than 0, then its parent submessage
4265       // might contain 1 or more entries and we want to break before this entry
4266       // if it contains at least 2 entries. We deal with this case later by
4267       // detecting and breaking before the next entry in the parent submessage.
4268       if (Left.ParameterCount == 0)
4269         return true;
4270       // However, if this submessage is the first entry in its parent
4271       // submessage, Left.ParameterCount might be 1 in some cases.
4272       // We deal with this case later by detecting an entry
4273       // following a closing paren of this submessage.
4274     }
4275 
4276     // If this is an entry immediately following a submessage, it will be
4277     // preceded by a closing paren of that submessage, like in:
4278     //     left---.  .---right
4279     //            v  v
4280     // sub: { ... } key: value
4281     // If there was a comment between `}` an `key` above, then `key` would be
4282     // put on a new line anyways.
4283     if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square))
4284       return true;
4285   }
4286 
4287   // Deal with lambda arguments in C++ - we want consistent line breaks whether
4288   // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced
4289   // as aggressive line breaks are placed when the lambda is not the last arg.
4290   if ((Style.Language == FormatStyle::LK_Cpp ||
4291        Style.Language == FormatStyle::LK_ObjC) &&
4292       Left.is(tok::l_paren) && Left.BlockParameterCount > 0 &&
4293       !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) {
4294     // Multiple lambdas in the same function call force line breaks.
4295     if (Left.BlockParameterCount > 1)
4296       return true;
4297 
4298     // A lambda followed by another arg forces a line break.
4299     if (!Left.Role)
4300       return false;
4301     auto Comma = Left.Role->lastComma();
4302     if (!Comma)
4303       return false;
4304     auto Next = Comma->getNextNonComment();
4305     if (!Next)
4306       return false;
4307     if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret))
4308       return true;
4309   }
4310 
4311   return false;
4312 }
4313 
4314 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
4315                                     const FormatToken &Right) const {
4316   const FormatToken &Left = *Right.Previous;
4317   // Language-specific stuff.
4318   if (Style.isCSharp()) {
4319     if (Left.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon) ||
4320         Right.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon))
4321       return false;
4322     // Only break after commas for generic type constraints.
4323     if (Line.First->is(TT_CSharpGenericTypeConstraint))
4324       return Left.is(TT_CSharpGenericTypeConstraintComma);
4325     // Keep nullable operators attached to their identifiers.
4326     if (Right.is(TT_CSharpNullable))
4327       return false;
4328   } else if (Style.Language == FormatStyle::LK_Java) {
4329     if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
4330                      Keywords.kw_implements))
4331       return false;
4332     if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
4333                       Keywords.kw_implements))
4334       return true;
4335   } else if (Style.isJavaScript()) {
4336     const FormatToken *NonComment = Right.getPreviousNonComment();
4337     if (NonComment &&
4338         NonComment->isOneOf(
4339             tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break,
4340             tok::kw_throw, Keywords.kw_interface, Keywords.kw_type,
4341             tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected,
4342             Keywords.kw_readonly, Keywords.kw_override, Keywords.kw_abstract,
4343             Keywords.kw_get, Keywords.kw_set, Keywords.kw_async,
4344             Keywords.kw_await))
4345       return false; // Otherwise automatic semicolon insertion would trigger.
4346     if (Right.NestingLevel == 0 &&
4347         (Left.Tok.getIdentifierInfo() ||
4348          Left.isOneOf(tok::r_square, tok::r_paren)) &&
4349         Right.isOneOf(tok::l_square, tok::l_paren))
4350       return false; // Otherwise automatic semicolon insertion would trigger.
4351     if (NonComment && NonComment->is(tok::identifier) &&
4352         NonComment->TokenText == "asserts")
4353       return false;
4354     if (Left.is(TT_FatArrow) && Right.is(tok::l_brace))
4355       return false;
4356     if (Left.is(TT_JsTypeColon))
4357       return true;
4358     // Don't wrap between ":" and "!" of a strict prop init ("field!: type;").
4359     if (Left.is(tok::exclaim) && Right.is(tok::colon))
4360       return false;
4361     // Look for is type annotations like:
4362     // function f(): a is B { ... }
4363     // Do not break before is in these cases.
4364     if (Right.is(Keywords.kw_is)) {
4365       const FormatToken *Next = Right.getNextNonComment();
4366       // If `is` is followed by a colon, it's likely that it's a dict key, so
4367       // ignore it for this check.
4368       // For example this is common in Polymer:
4369       // Polymer({
4370       //   is: 'name',
4371       //   ...
4372       // });
4373       if (!Next || !Next->is(tok::colon))
4374         return false;
4375     }
4376     if (Left.is(Keywords.kw_in))
4377       return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
4378     if (Right.is(Keywords.kw_in))
4379       return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
4380     if (Right.is(Keywords.kw_as))
4381       return false; // must not break before as in 'x as type' casts
4382     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) {
4383       // extends and infer can appear as keywords in conditional types:
4384       //   https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types
4385       // do not break before them, as the expressions are subject to ASI.
4386       return false;
4387     }
4388     if (Left.is(Keywords.kw_as))
4389       return true;
4390     if (Left.is(TT_NonNullAssertion))
4391       return true;
4392     if (Left.is(Keywords.kw_declare) &&
4393         Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
4394                       Keywords.kw_function, tok::kw_class, tok::kw_enum,
4395                       Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
4396                       Keywords.kw_let, tok::kw_const))
4397       // See grammar for 'declare' statements at:
4398       // https://github.com/Microsoft/TypeScript/blob/main/doc/spec-ARCHIVED.md#A.10
4399       return false;
4400     if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
4401         Right.isOneOf(tok::identifier, tok::string_literal))
4402       return false; // must not break in "module foo { ...}"
4403     if (Right.is(TT_TemplateString) && Right.closesScope())
4404       return false;
4405     // Don't split tagged template literal so there is a break between the tag
4406     // identifier and template string.
4407     if (Left.is(tok::identifier) && Right.is(TT_TemplateString))
4408       return false;
4409     if (Left.is(TT_TemplateString) && Left.opensScope())
4410       return true;
4411   }
4412 
4413   if (Left.is(tok::at))
4414     return false;
4415   if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
4416     return false;
4417   if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
4418     return !Right.is(tok::l_paren);
4419   if (Right.is(TT_PointerOrReference))
4420     return Line.IsMultiVariableDeclStmt ||
4421            (getTokenPointerOrReferenceAlignment(Right) ==
4422                 FormatStyle::PAS_Right &&
4423             (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
4424   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
4425       Right.is(tok::kw_operator))
4426     return true;
4427   if (Left.is(TT_PointerOrReference))
4428     return false;
4429   if (Right.isTrailingComment())
4430     // We rely on MustBreakBefore being set correctly here as we should not
4431     // change the "binding" behavior of a comment.
4432     // The first comment in a braced lists is always interpreted as belonging to
4433     // the first list element. Otherwise, it should be placed outside of the
4434     // list.
4435     return Left.is(BK_BracedInit) ||
4436            (Left.is(TT_CtorInitializerColon) &&
4437             Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon);
4438   if (Left.is(tok::question) && Right.is(tok::colon))
4439     return false;
4440   if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
4441     return Style.BreakBeforeTernaryOperators;
4442   if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
4443     return !Style.BreakBeforeTernaryOperators;
4444   if (Left.is(TT_InheritanceColon))
4445     return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon;
4446   if (Right.is(TT_InheritanceColon))
4447     return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon;
4448   if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
4449       Left.isNot(TT_SelectorName))
4450     return true;
4451 
4452   if (Right.is(tok::colon) &&
4453       !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
4454     return false;
4455   if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
4456     if (Style.Language == FormatStyle::LK_Proto ||
4457         Style.Language == FormatStyle::LK_TextProto) {
4458       if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral())
4459         return false;
4460       // Prevent cases like:
4461       //
4462       // submessage:
4463       //     { key: valueeeeeeeeeeee }
4464       //
4465       // when the snippet does not fit into one line.
4466       // Prefer:
4467       //
4468       // submessage: {
4469       //   key: valueeeeeeeeeeee
4470       // }
4471       //
4472       // instead, even if it is longer by one line.
4473       //
4474       // Note that this allows allows the "{" to go over the column limit
4475       // when the column limit is just between ":" and "{", but that does
4476       // not happen too often and alternative formattings in this case are
4477       // not much better.
4478       //
4479       // The code covers the cases:
4480       //
4481       // submessage: { ... }
4482       // submessage: < ... >
4483       // repeated: [ ... ]
4484       if (((Right.is(tok::l_brace) || Right.is(tok::less)) &&
4485            Right.is(TT_DictLiteral)) ||
4486           Right.is(TT_ArrayInitializerLSquare))
4487         return false;
4488     }
4489     return true;
4490   }
4491   if (Right.is(tok::r_square) && Right.MatchingParen &&
4492       Right.MatchingParen->is(TT_ProtoExtensionLSquare))
4493     return false;
4494   if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
4495                                     Right.Next->is(TT_ObjCMethodExpr)))
4496     return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
4497   if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
4498     return true;
4499   if (Right.is(tok::kw_concept))
4500     return Style.BreakBeforeConceptDeclarations != FormatStyle::BBCDS_Never;
4501   if (Right.is(TT_RequiresClause))
4502     return true;
4503   if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
4504     return true;
4505   if (Left.ClosesRequiresClause)
4506     return true;
4507   if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
4508                     TT_OverloadedOperator))
4509     return false;
4510   if (Left.is(TT_RangeBasedForLoopColon))
4511     return true;
4512   if (Right.is(TT_RangeBasedForLoopColon))
4513     return false;
4514   if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
4515     return true;
4516   if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
4517       (Left.is(tok::less) && Right.is(tok::less)))
4518     return false;
4519   if (Right.is(TT_BinaryOperator) &&
4520       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
4521       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
4522        Right.getPrecedence() != prec::Assignment))
4523     return true;
4524   if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
4525       Left.is(tok::kw_operator))
4526     return false;
4527   if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
4528       Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
4529     return false;
4530   if (Left.is(tok::equal) && Right.is(tok::l_brace) &&
4531       !Style.Cpp11BracedListStyle)
4532     return false;
4533   if (Left.is(tok::l_paren) &&
4534       Left.isOneOf(TT_AttributeParen, TT_TypeDeclarationParen))
4535     return false;
4536   if (Left.is(tok::l_paren) && Left.Previous &&
4537       (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
4538     return false;
4539   if (Right.is(TT_ImplicitStringLiteral))
4540     return false;
4541 
4542   if (Right.is(TT_TemplateCloser))
4543     return false;
4544   if (Right.is(tok::r_square) && Right.MatchingParen &&
4545       Right.MatchingParen->is(TT_LambdaLSquare))
4546     return false;
4547 
4548   // We only break before r_brace if there was a corresponding break before
4549   // the l_brace, which is tracked by BreakBeforeClosingBrace.
4550   if (Right.is(tok::r_brace))
4551     return Right.MatchingParen && Right.MatchingParen->is(BK_Block);
4552 
4553   // We only break before r_paren if we're in a block indented context.
4554   if (Right.is(tok::r_paren)) {
4555     if (Style.AlignAfterOpenBracket != FormatStyle::BAS_BlockIndent ||
4556         !Right.MatchingParen)
4557       return false;
4558     const FormatToken *Previous = Right.MatchingParen->Previous;
4559     return !(Previous && (Previous->is(tok::kw_for) || Previous->isIf()));
4560   }
4561 
4562   // Allow breaking after a trailing annotation, e.g. after a method
4563   // declaration.
4564   if (Left.is(TT_TrailingAnnotation))
4565     return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
4566                           tok::less, tok::coloncolon);
4567 
4568   if (Right.is(tok::kw___attribute) ||
4569       (Right.is(tok::l_square) && Right.is(TT_AttributeSquare)))
4570     return !Left.is(TT_AttributeSquare);
4571 
4572   if (Left.is(tok::identifier) && Right.is(tok::string_literal))
4573     return true;
4574 
4575   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
4576     return true;
4577 
4578   if (Left.is(TT_CtorInitializerColon))
4579     return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon;
4580   if (Right.is(TT_CtorInitializerColon))
4581     return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon;
4582   if (Left.is(TT_CtorInitializerComma) &&
4583       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
4584     return false;
4585   if (Right.is(TT_CtorInitializerComma) &&
4586       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
4587     return true;
4588   if (Left.is(TT_InheritanceComma) &&
4589       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
4590     return false;
4591   if (Right.is(TT_InheritanceComma) &&
4592       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
4593     return true;
4594   if (Left.is(TT_ArrayInitializerLSquare))
4595     return true;
4596   if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
4597     return true;
4598   if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
4599       !Left.isOneOf(tok::arrowstar, tok::lessless) &&
4600       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
4601       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
4602        Left.getPrecedence() == prec::Assignment))
4603     return true;
4604   if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) ||
4605       (Left.is(tok::r_square) && Right.is(TT_AttributeSquare)))
4606     return false;
4607 
4608   auto ShortLambdaOption = Style.AllowShortLambdasOnASingleLine;
4609   if (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace)) {
4610     if (isAllmanLambdaBrace(Left))
4611       return !isItAnEmptyLambdaAllowed(Left, ShortLambdaOption);
4612     if (isAllmanLambdaBrace(Right))
4613       return !isItAnEmptyLambdaAllowed(Right, ShortLambdaOption);
4614   }
4615 
4616   return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
4617                       tok::kw_class, tok::kw_struct, tok::comment) ||
4618          Right.isMemberAccess() ||
4619          Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
4620                        tok::colon, tok::l_square, tok::at) ||
4621          (Left.is(tok::r_paren) &&
4622           Right.isOneOf(tok::identifier, tok::kw_const)) ||
4623          (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
4624          (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
4625 }
4626 
4627 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) const {
4628   llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n";
4629   const FormatToken *Tok = Line.First;
4630   while (Tok) {
4631     llvm::errs() << " M=" << Tok->MustBreakBefore
4632                  << " C=" << Tok->CanBreakBefore
4633                  << " T=" << getTokenTypeName(Tok->getType())
4634                  << " S=" << Tok->SpacesRequiredBefore
4635                  << " F=" << Tok->Finalized << " B=" << Tok->BlockParameterCount
4636                  << " BK=" << Tok->getBlockKind() << " P=" << Tok->SplitPenalty
4637                  << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength
4638                  << " PPK=" << Tok->getPackingKind() << " FakeLParens=";
4639     for (prec::Level LParen : Tok->FakeLParens)
4640       llvm::errs() << LParen << "/";
4641     llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
4642     llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo();
4643     llvm::errs() << " Text='" << Tok->TokenText << "'\n";
4644     if (!Tok->Next)
4645       assert(Tok == Line.Last);
4646     Tok = Tok->Next;
4647   }
4648   llvm::errs() << "----\n";
4649 }
4650 
4651 FormatStyle::PointerAlignmentStyle
4652 TokenAnnotator::getTokenReferenceAlignment(const FormatToken &Reference) const {
4653   assert(Reference.isOneOf(tok::amp, tok::ampamp));
4654   switch (Style.ReferenceAlignment) {
4655   case FormatStyle::RAS_Pointer:
4656     return Style.PointerAlignment;
4657   case FormatStyle::RAS_Left:
4658     return FormatStyle::PAS_Left;
4659   case FormatStyle::RAS_Right:
4660     return FormatStyle::PAS_Right;
4661   case FormatStyle::RAS_Middle:
4662     return FormatStyle::PAS_Middle;
4663   }
4664   assert(0); //"Unhandled value of ReferenceAlignment"
4665   return Style.PointerAlignment;
4666 }
4667 
4668 FormatStyle::PointerAlignmentStyle
4669 TokenAnnotator::getTokenPointerOrReferenceAlignment(
4670     const FormatToken &PointerOrReference) const {
4671   if (PointerOrReference.isOneOf(tok::amp, tok::ampamp)) {
4672     switch (Style.ReferenceAlignment) {
4673     case FormatStyle::RAS_Pointer:
4674       return Style.PointerAlignment;
4675     case FormatStyle::RAS_Left:
4676       return FormatStyle::PAS_Left;
4677     case FormatStyle::RAS_Right:
4678       return FormatStyle::PAS_Right;
4679     case FormatStyle::RAS_Middle:
4680       return FormatStyle::PAS_Middle;
4681     }
4682   }
4683   assert(PointerOrReference.is(tok::star));
4684   return Style.PointerAlignment;
4685 }
4686 
4687 } // namespace format
4688 } // namespace clang
4689