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 token can be used as an identifier in
30 /// an Objective-C \c \@selector, \c false otherwise.
31 ///
32 /// Because getFormattingLangOpts() always lexes source code as
33 /// Objective-C++, C++ keywords like \c new and \c delete are
34 /// lexed as tok::kw_*, not tok::identifier, even for Objective-C.
35 ///
36 /// For Objective-C and Objective-C++, both identifiers and keywords
37 /// are valid inside @selector(...) (or a macro which
38 /// invokes @selector(...)). So, we allow treat any identifier or
39 /// keyword as a potential Objective-C selector component.
40 static bool canBeObjCSelectorComponent(const FormatToken &Tok) {
41   return Tok.Tok.getIdentifierInfo() != nullptr;
42 }
43 
44 /// With `Left` being '(', check if we're at either `[...](` or
45 /// `[...]<...>(`, where the [ opens a lambda capture list.
46 static bool isLambdaParameterList(const FormatToken *Left) {
47   // Skip <...> if present.
48   if (Left->Previous && Left->Previous->is(tok::greater) &&
49       Left->Previous->MatchingParen &&
50       Left->Previous->MatchingParen->is(TT_TemplateOpener))
51     Left = Left->Previous->MatchingParen;
52 
53   // Check for `[...]`.
54   return Left->Previous && Left->Previous->is(tok::r_square) &&
55          Left->Previous->MatchingParen &&
56          Left->Previous->MatchingParen->is(TT_LambdaLSquare);
57 }
58 
59 /// Returns \c true if the token is followed by a boolean condition, \c false
60 /// otherwise.
61 static bool isKeywordWithCondition(const FormatToken &Tok) {
62   return Tok.isOneOf(tok::kw_if, tok::kw_for, tok::kw_while, tok::kw_switch,
63                      tok::kw_constexpr, tok::kw_catch);
64 }
65 
66 /// A parser that gathers additional information about tokens.
67 ///
68 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
69 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
70 /// into template parameter lists.
71 class AnnotatingParser {
72 public:
73   AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
74                    const AdditionalKeywords &Keywords)
75       : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
76         Keywords(Keywords) {
77     Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
78     resetTokenMetadata(CurrentToken);
79   }
80 
81 private:
82   bool parseAngle() {
83     if (!CurrentToken || !CurrentToken->Previous)
84       return false;
85     if (NonTemplateLess.count(CurrentToken->Previous))
86       return false;
87 
88     const FormatToken &Previous = *CurrentToken->Previous; // The '<'.
89     if (Previous.Previous) {
90       if (Previous.Previous->Tok.isLiteral())
91         return false;
92       if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
93           (!Previous.Previous->MatchingParen ||
94            !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen)))
95         return false;
96     }
97 
98     FormatToken *Left = CurrentToken->Previous;
99     Left->ParentBracket = Contexts.back().ContextKind;
100     ScopedContextCreator ContextCreator(*this, tok::less, 12);
101 
102     // If this angle is in the context of an expression, we need to be more
103     // hesitant to detect it as opening template parameters.
104     bool InExprContext = Contexts.back().IsExpression;
105 
106     Contexts.back().IsExpression = false;
107     // If there's a template keyword before the opening angle bracket, this is a
108     // template parameter, not an argument.
109     Contexts.back().InTemplateArgument =
110         Left->Previous && Left->Previous->Tok.isNot(tok::kw_template);
111 
112     if (Style.Language == FormatStyle::LK_Java &&
113         CurrentToken->is(tok::question))
114       next();
115 
116     while (CurrentToken) {
117       if (CurrentToken->is(tok::greater)) {
118         // Try to do a better job at looking for ">>" within the condition of
119         // a statement.
120         if (CurrentToken->Next && CurrentToken->Next->is(tok::greater) &&
121             Left->ParentBracket != tok::less &&
122             isKeywordWithCondition(*Line.First))
123           return false;
124         Left->MatchingParen = CurrentToken;
125         CurrentToken->MatchingParen = Left;
126         // In TT_Proto, we must distignuish between:
127         //   map<key, value>
128         //   msg < item: data >
129         //   msg: < item: data >
130         // In TT_TextProto, map<key, value> does not occur.
131         if (Style.Language == FormatStyle::LK_TextProto ||
132             (Style.Language == FormatStyle::LK_Proto && Left->Previous &&
133              Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral)))
134           CurrentToken->setType(TT_DictLiteral);
135         else
136           CurrentToken->setType(TT_TemplateCloser);
137         next();
138         return true;
139       }
140       if (CurrentToken->is(tok::question) &&
141           Style.Language == FormatStyle::LK_Java) {
142         next();
143         continue;
144       }
145       if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
146           (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext &&
147            !Style.isCSharp() && Style.Language != FormatStyle::LK_Proto &&
148            Style.Language != FormatStyle::LK_TextProto))
149         return false;
150       // If a && or || is found and interpreted as a binary operator, this set
151       // of angles is likely part of something like "a < b && c > d". If the
152       // angles are inside an expression, the ||/&& might also be a binary
153       // operator that was misinterpreted because we are parsing template
154       // parameters.
155       // FIXME: This is getting out of hand, write a decent parser.
156       if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
157           CurrentToken->Previous->is(TT_BinaryOperator) &&
158           Contexts[Contexts.size() - 2].IsExpression &&
159           !Line.startsWith(tok::kw_template))
160         return false;
161       updateParameterCount(Left, CurrentToken);
162       if (Style.Language == FormatStyle::LK_Proto) {
163         if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) {
164           if (CurrentToken->is(tok::colon) ||
165               (CurrentToken->isOneOf(tok::l_brace, tok::less) &&
166                Previous->isNot(tok::colon)))
167             Previous->setType(TT_SelectorName);
168         }
169       }
170       if (!consumeToken())
171         return false;
172     }
173     return false;
174   }
175 
176   bool parseUntouchableParens() {
177     while (CurrentToken) {
178       CurrentToken->Finalized = true;
179       switch (CurrentToken->Tok.getKind()) {
180       case tok::l_paren:
181         next();
182         if (!parseUntouchableParens())
183           return false;
184         continue;
185       case tok::r_paren:
186         next();
187         return true;
188       default:
189         // no-op
190         break;
191       }
192       next();
193     }
194     return false;
195   }
196 
197   bool parseParens(bool LookForDecls = false) {
198     if (!CurrentToken)
199       return false;
200     FormatToken *Left = CurrentToken->Previous;
201     assert(Left && "Unknown previous token");
202     FormatToken *PrevNonComment = Left->getPreviousNonComment();
203     Left->ParentBracket = Contexts.back().ContextKind;
204     ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
205 
206     // FIXME: This is a bit of a hack. Do better.
207     Contexts.back().ColonIsForRangeExpr =
208         Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
209 
210     if (Left->Previous && Left->Previous->is(TT_UntouchableMacroFunc)) {
211       Left->Finalized = true;
212       return parseUntouchableParens();
213     }
214 
215     bool StartsObjCMethodExpr = false;
216     if (FormatToken *MaybeSel = Left->Previous) {
217       // @selector( starts a selector.
218       if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
219           MaybeSel->Previous->is(tok::at)) {
220         StartsObjCMethodExpr = true;
221       }
222     }
223 
224     if (Left->is(TT_OverloadedOperatorLParen)) {
225       Contexts.back().IsExpression = false;
226     } else if (Style.Language == FormatStyle::LK_JavaScript &&
227                (Line.startsWith(Keywords.kw_type, tok::identifier) ||
228                 Line.startsWith(tok::kw_export, Keywords.kw_type,
229                                 tok::identifier))) {
230       // type X = (...);
231       // export type X = (...);
232       Contexts.back().IsExpression = false;
233     } else if (Left->Previous &&
234                (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_while,
235                                         tok::l_paren, tok::comma) ||
236                 Left->Previous->isIf() ||
237                 Left->Previous->is(TT_BinaryOperator))) {
238       // static_assert, if and while usually contain expressions.
239       Contexts.back().IsExpression = true;
240     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
241                (Left->Previous->is(Keywords.kw_function) ||
242                 (Left->Previous->endsSequence(tok::identifier,
243                                               Keywords.kw_function)))) {
244       // function(...) or function f(...)
245       Contexts.back().IsExpression = false;
246     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
247                Left->Previous->is(TT_JsTypeColon)) {
248       // let x: (SomeType);
249       Contexts.back().IsExpression = false;
250     } else if (isLambdaParameterList(Left)) {
251       // This is a parameter list of a lambda expression.
252       Contexts.back().IsExpression = false;
253     } else if (Line.InPPDirective &&
254                (!Left->Previous || !Left->Previous->is(tok::identifier))) {
255       Contexts.back().IsExpression = true;
256     } else if (Contexts[Contexts.size() - 2].CaretFound) {
257       // This is the parameter list of an ObjC block.
258       Contexts.back().IsExpression = false;
259     } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) {
260       // The first argument to a foreach macro is a declaration.
261       Contexts.back().IsForEachMacro = true;
262       Contexts.back().IsExpression = false;
263     } else if (Left->Previous && Left->Previous->MatchingParen &&
264                Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
265       Contexts.back().IsExpression = false;
266     } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
267       bool IsForOrCatch =
268           Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch);
269       Contexts.back().IsExpression = !IsForOrCatch;
270     }
271 
272     // Infer the role of the l_paren based on the previous token if we haven't
273     // detected one one yet.
274     if (PrevNonComment && Left->is(TT_Unknown)) {
275       if (PrevNonComment->is(tok::kw___attribute)) {
276         Left->setType(TT_AttributeParen);
277       } else if (PrevNonComment->isOneOf(TT_TypenameMacro, tok::kw_decltype,
278                                          tok::kw_typeof, tok::kw__Atomic,
279                                          tok::kw___underlying_type)) {
280         Left->setType(TT_TypeDeclarationParen);
281         // decltype() and typeof() usually contain expressions.
282         if (PrevNonComment->isOneOf(tok::kw_decltype, tok::kw_typeof))
283           Contexts.back().IsExpression = true;
284       }
285     }
286 
287     if (StartsObjCMethodExpr) {
288       Contexts.back().ColonIsObjCMethodExpr = true;
289       Left->setType(TT_ObjCMethodExpr);
290     }
291 
292     // MightBeFunctionType and ProbablyFunctionType are used for
293     // function pointer and reference types as well as Objective-C
294     // block types:
295     //
296     // void (*FunctionPointer)(void);
297     // void (&FunctionReference)(void);
298     // void (^ObjCBlock)(void);
299     bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
300     bool ProbablyFunctionType =
301         CurrentToken->isOneOf(tok::star, tok::amp, tok::caret);
302     bool HasMultipleLines = false;
303     bool HasMultipleParametersOnALine = false;
304     bool MightBeObjCForRangeLoop =
305         Left->Previous && Left->Previous->is(tok::kw_for);
306     FormatToken *PossibleObjCForInToken = nullptr;
307     while (CurrentToken) {
308       // LookForDecls is set when "if (" has been seen. Check for
309       // 'identifier' '*' 'identifier' followed by not '=' -- this
310       // '*' has to be a binary operator but determineStarAmpUsage() will
311       // categorize it as an unary operator, so set the right type here.
312       if (LookForDecls && CurrentToken->Next) {
313         FormatToken *Prev = CurrentToken->getPreviousNonComment();
314         if (Prev) {
315           FormatToken *PrevPrev = Prev->getPreviousNonComment();
316           FormatToken *Next = CurrentToken->Next;
317           if (PrevPrev && PrevPrev->is(tok::identifier) &&
318               Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
319               CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
320             Prev->setType(TT_BinaryOperator);
321             LookForDecls = false;
322           }
323         }
324       }
325 
326       if (CurrentToken->Previous->is(TT_PointerOrReference) &&
327           CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
328                                                     tok::coloncolon))
329         ProbablyFunctionType = true;
330       if (CurrentToken->is(tok::comma))
331         MightBeFunctionType = false;
332       if (CurrentToken->Previous->is(TT_BinaryOperator))
333         Contexts.back().IsExpression = true;
334       if (CurrentToken->is(tok::r_paren)) {
335         if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next &&
336             (CurrentToken->Next->is(tok::l_paren) ||
337              (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration)))
338           Left->setType(Left->Next->is(tok::caret) ? TT_ObjCBlockLParen
339                                                    : TT_FunctionTypeLParen);
340         Left->MatchingParen = CurrentToken;
341         CurrentToken->MatchingParen = Left;
342 
343         if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
344             Left->Previous && Left->Previous->is(tok::l_paren)) {
345           // Detect the case where macros are used to generate lambdas or
346           // function bodies, e.g.:
347           //   auto my_lambda = MARCO((Type *type, int i) { .. body .. });
348           for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) {
349             if (Tok->is(TT_BinaryOperator) &&
350                 Tok->isOneOf(tok::star, tok::amp, tok::ampamp))
351               Tok->setType(TT_PointerOrReference);
352           }
353         }
354 
355         if (StartsObjCMethodExpr) {
356           CurrentToken->setType(TT_ObjCMethodExpr);
357           if (Contexts.back().FirstObjCSelectorName) {
358             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
359                 Contexts.back().LongestObjCSelectorName;
360           }
361         }
362 
363         if (Left->is(TT_AttributeParen))
364           CurrentToken->setType(TT_AttributeParen);
365         if (Left->is(TT_TypeDeclarationParen))
366           CurrentToken->setType(TT_TypeDeclarationParen);
367         if (Left->Previous && Left->Previous->is(TT_JavaAnnotation))
368           CurrentToken->setType(TT_JavaAnnotation);
369         if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation))
370           CurrentToken->setType(TT_LeadingJavaAnnotation);
371         if (Left->Previous && Left->Previous->is(TT_AttributeSquare))
372           CurrentToken->setType(TT_AttributeSquare);
373 
374         if (!HasMultipleLines)
375           Left->setPackingKind(PPK_Inconclusive);
376         else if (HasMultipleParametersOnALine)
377           Left->setPackingKind(PPK_BinPacked);
378         else
379           Left->setPackingKind(PPK_OnePerLine);
380 
381         next();
382         return true;
383       }
384       if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
385         return false;
386 
387       if (CurrentToken->is(tok::l_brace))
388         Left->setType(TT_Unknown); // Not TT_ObjCBlockLParen
389       if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
390           !CurrentToken->Next->HasUnescapedNewline &&
391           !CurrentToken->Next->isTrailingComment())
392         HasMultipleParametersOnALine = true;
393       if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
394            CurrentToken->Previous->isSimpleTypeSpecifier()) &&
395           !CurrentToken->is(tok::l_brace))
396         Contexts.back().IsExpression = false;
397       if (CurrentToken->isOneOf(tok::semi, tok::colon)) {
398         MightBeObjCForRangeLoop = false;
399         if (PossibleObjCForInToken) {
400           PossibleObjCForInToken->setType(TT_Unknown);
401           PossibleObjCForInToken = nullptr;
402         }
403       }
404       if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) {
405         PossibleObjCForInToken = CurrentToken;
406         PossibleObjCForInToken->setType(TT_ObjCForIn);
407       }
408       // When we discover a 'new', we set CanBeExpression to 'false' in order to
409       // parse the type correctly. Reset that after a comma.
410       if (CurrentToken->is(tok::comma))
411         Contexts.back().CanBeExpression = true;
412 
413       FormatToken *Tok = CurrentToken;
414       if (!consumeToken())
415         return false;
416       updateParameterCount(Left, Tok);
417       if (CurrentToken && CurrentToken->HasUnescapedNewline)
418         HasMultipleLines = true;
419     }
420     return false;
421   }
422 
423   bool isCSharpAttributeSpecifier(const FormatToken &Tok) {
424     if (!Style.isCSharp())
425       return false;
426 
427     // `identifier[i]` is not an attribute.
428     if (Tok.Previous && Tok.Previous->is(tok::identifier))
429       return false;
430 
431     // Chains of [] in `identifier[i][j][k]` are not attributes.
432     if (Tok.Previous && Tok.Previous->is(tok::r_square)) {
433       auto *MatchingParen = Tok.Previous->MatchingParen;
434       if (!MatchingParen || MatchingParen->is(TT_ArraySubscriptLSquare))
435         return false;
436     }
437 
438     const FormatToken *AttrTok = Tok.Next;
439     if (!AttrTok)
440       return false;
441 
442     // Just an empty declaration e.g. string [].
443     if (AttrTok->is(tok::r_square))
444       return false;
445 
446     // Move along the tokens inbetween the '[' and ']' e.g. [STAThread].
447     while (AttrTok && AttrTok->isNot(tok::r_square)) {
448       AttrTok = AttrTok->Next;
449     }
450 
451     if (!AttrTok)
452       return false;
453 
454     // Allow an attribute to be the only content of a file.
455     AttrTok = AttrTok->Next;
456     if (!AttrTok)
457       return true;
458 
459     // Limit this to being an access modifier that follows.
460     if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
461                          tok::comment, tok::kw_class, tok::kw_static,
462                          tok::l_square, Keywords.kw_internal)) {
463       return true;
464     }
465 
466     // incase its a [XXX] retval func(....
467     if (AttrTok->Next &&
468         AttrTok->Next->startsSequence(tok::identifier, tok::l_paren))
469       return true;
470 
471     return false;
472   }
473 
474   bool isCpp11AttributeSpecifier(const FormatToken &Tok) {
475     if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square))
476       return false;
477     // The first square bracket is part of an ObjC array literal
478     if (Tok.Previous && Tok.Previous->is(tok::at)) {
479       return false;
480     }
481     const FormatToken *AttrTok = Tok.Next->Next;
482     if (!AttrTok)
483       return false;
484     // C++17 '[[using ns: foo, bar(baz, blech)]]'
485     // We assume nobody will name an ObjC variable 'using'.
486     if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon))
487       return true;
488     if (AttrTok->isNot(tok::identifier))
489       return false;
490     while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) {
491       // ObjC message send. We assume nobody will use : in a C++11 attribute
492       // specifier parameter, although this is technically valid:
493       // [[foo(:)]].
494       if (AttrTok->is(tok::colon) ||
495           AttrTok->startsSequence(tok::identifier, tok::identifier) ||
496           AttrTok->startsSequence(tok::r_paren, tok::identifier))
497         return false;
498       if (AttrTok->is(tok::ellipsis))
499         return true;
500       AttrTok = AttrTok->Next;
501     }
502     return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square);
503   }
504 
505   bool parseSquare() {
506     if (!CurrentToken)
507       return false;
508 
509     // A '[' could be an index subscript (after an identifier or after
510     // ')' or ']'), it could be the start of an Objective-C method
511     // expression, it could the start of an Objective-C array literal,
512     // or it could be a C++ attribute specifier [[foo::bar]].
513     FormatToken *Left = CurrentToken->Previous;
514     Left->ParentBracket = Contexts.back().ContextKind;
515     FormatToken *Parent = Left->getPreviousNonComment();
516 
517     // Cases where '>' is followed by '['.
518     // In C++, this can happen either in array of templates (foo<int>[10])
519     // or when array is a nested template type (unique_ptr<type1<type2>[]>).
520     bool CppArrayTemplates =
521         Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) &&
522         (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
523          Contexts.back().InTemplateArgument);
524 
525     bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) ||
526                                      Contexts.back().InCpp11AttributeSpecifier;
527 
528     // Treat C# Attributes [STAThread] much like C++ attributes [[...]].
529     bool IsCSharpAttributeSpecifier =
530         isCSharpAttributeSpecifier(*Left) ||
531         Contexts.back().InCSharpAttributeSpecifier;
532 
533     bool InsideInlineASM = Line.startsWith(tok::kw_asm);
534     bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style);
535     bool StartsObjCMethodExpr =
536         !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates &&
537         Style.isCpp() && !IsCpp11AttributeSpecifier &&
538         !IsCSharpAttributeSpecifier && Contexts.back().CanBeExpression &&
539         Left->isNot(TT_LambdaLSquare) &&
540         !CurrentToken->isOneOf(tok::l_brace, tok::r_square) &&
541         (!Parent ||
542          Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
543                          tok::kw_return, tok::kw_throw) ||
544          Parent->isUnaryOperator() ||
545          // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
546          Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
547          (getBinOpPrecedence(Parent->Tok.getKind(), true, true) >
548           prec::Unknown));
549     bool ColonFound = false;
550 
551     unsigned BindingIncrease = 1;
552     if (IsCppStructuredBinding) {
553       Left->setType(TT_StructuredBindingLSquare);
554     } else if (Left->is(TT_Unknown)) {
555       if (StartsObjCMethodExpr) {
556         Left->setType(TT_ObjCMethodExpr);
557       } else if (InsideInlineASM) {
558         Left->setType(TT_InlineASMSymbolicNameLSquare);
559       } else if (IsCpp11AttributeSpecifier) {
560         Left->setType(TT_AttributeSquare);
561       } else if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
562                  Contexts.back().ContextKind == tok::l_brace &&
563                  Parent->isOneOf(tok::l_brace, tok::comma)) {
564         Left->setType(TT_JsComputedPropertyName);
565       } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace &&
566                  Parent && Parent->isOneOf(tok::l_brace, tok::comma)) {
567         Left->setType(TT_DesignatedInitializerLSquare);
568       } else if (IsCSharpAttributeSpecifier) {
569         Left->setType(TT_AttributeSquare);
570       } else if (CurrentToken->is(tok::r_square) && Parent &&
571                  Parent->is(TT_TemplateCloser)) {
572         Left->setType(TT_ArraySubscriptLSquare);
573       } else if (Style.Language == FormatStyle::LK_Proto ||
574                  Style.Language == FormatStyle::LK_TextProto) {
575         // Square braces in LK_Proto can either be message field attributes:
576         //
577         // optional Aaa aaa = 1 [
578         //   (aaa) = aaa
579         // ];
580         //
581         // extensions 123 [
582         //   (aaa) = aaa
583         // ];
584         //
585         // or text proto extensions (in options):
586         //
587         // option (Aaa.options) = {
588         //   [type.type/type] {
589         //     key: value
590         //   }
591         // }
592         //
593         // or repeated fields (in options):
594         //
595         // option (Aaa.options) = {
596         //   keys: [ 1, 2, 3 ]
597         // }
598         //
599         // In the first and the third case we want to spread the contents inside
600         // the square braces; in the second we want to keep them inline.
601         Left->setType(TT_ArrayInitializerLSquare);
602         if (!Left->endsSequence(tok::l_square, tok::numeric_constant,
603                                 tok::equal) &&
604             !Left->endsSequence(tok::l_square, tok::numeric_constant,
605                                 tok::identifier) &&
606             !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) {
607           Left->setType(TT_ProtoExtensionLSquare);
608           BindingIncrease = 10;
609         }
610       } else if (!CppArrayTemplates && Parent &&
611                  Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
612                                  tok::comma, tok::l_paren, tok::l_square,
613                                  tok::question, tok::colon, tok::kw_return,
614                                  // Should only be relevant to JavaScript:
615                                  tok::kw_default)) {
616         Left->setType(TT_ArrayInitializerLSquare);
617       } else {
618         BindingIncrease = 10;
619         Left->setType(TT_ArraySubscriptLSquare);
620       }
621     }
622 
623     ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
624     Contexts.back().IsExpression = true;
625     if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
626         Parent->is(TT_JsTypeColon))
627       Contexts.back().IsExpression = false;
628 
629     Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
630     Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier;
631     Contexts.back().InCSharpAttributeSpecifier = IsCSharpAttributeSpecifier;
632 
633     while (CurrentToken) {
634       if (CurrentToken->is(tok::r_square)) {
635         if (IsCpp11AttributeSpecifier)
636           CurrentToken->setType(TT_AttributeSquare);
637         if (IsCSharpAttributeSpecifier)
638           CurrentToken->setType(TT_AttributeSquare);
639         else if (((CurrentToken->Next &&
640                    CurrentToken->Next->is(tok::l_paren)) ||
641                   (CurrentToken->Previous &&
642                    CurrentToken->Previous->Previous == Left)) &&
643                  Left->is(TT_ObjCMethodExpr)) {
644           // An ObjC method call is rarely followed by an open parenthesis. It
645           // also can't be composed of just one token, unless it's a macro that
646           // will be expanded to more tokens.
647           // FIXME: Do we incorrectly label ":" with this?
648           StartsObjCMethodExpr = false;
649           Left->setType(TT_Unknown);
650         }
651         if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
652           CurrentToken->setType(TT_ObjCMethodExpr);
653           // If we haven't seen a colon yet, make sure the last identifier
654           // before the r_square is tagged as a selector name component.
655           if (!ColonFound && CurrentToken->Previous &&
656               CurrentToken->Previous->is(TT_Unknown) &&
657               canBeObjCSelectorComponent(*CurrentToken->Previous))
658             CurrentToken->Previous->setType(TT_SelectorName);
659           // determineStarAmpUsage() thinks that '*' '[' is allocating an
660           // array of pointers, but if '[' starts a selector then '*' is a
661           // binary operator.
662           if (Parent && Parent->is(TT_PointerOrReference))
663             Parent->setType(TT_BinaryOperator);
664         }
665         // An arrow after an ObjC method expression is not a lambda arrow.
666         if (CurrentToken->getType() == TT_ObjCMethodExpr &&
667             CurrentToken->Next && CurrentToken->Next->is(TT_LambdaArrow))
668           CurrentToken->Next->setType(TT_Unknown);
669         Left->MatchingParen = CurrentToken;
670         CurrentToken->MatchingParen = Left;
671         // FirstObjCSelectorName is set when a colon is found. This does
672         // not work, however, when the method has no parameters.
673         // Here, we set FirstObjCSelectorName when the end of the method call is
674         // reached, in case it was not set already.
675         if (!Contexts.back().FirstObjCSelectorName) {
676           FormatToken *Previous = CurrentToken->getPreviousNonComment();
677           if (Previous && Previous->is(TT_SelectorName)) {
678             Previous->ObjCSelectorNameParts = 1;
679             Contexts.back().FirstObjCSelectorName = Previous;
680           }
681         } else {
682           Left->ParameterCount =
683               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
684         }
685         if (Contexts.back().FirstObjCSelectorName) {
686           Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
687               Contexts.back().LongestObjCSelectorName;
688           if (Left->BlockParameterCount > 1)
689             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
690         }
691         next();
692         return true;
693       }
694       if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
695         return false;
696       if (CurrentToken->is(tok::colon)) {
697         if (IsCpp11AttributeSpecifier &&
698             CurrentToken->endsSequence(tok::colon, tok::identifier,
699                                        tok::kw_using)) {
700           // Remember that this is a [[using ns: foo]] C++ attribute, so we
701           // don't add a space before the colon (unlike other colons).
702           CurrentToken->setType(TT_AttributeColon);
703         } else if (Left->isOneOf(TT_ArraySubscriptLSquare,
704                                  TT_DesignatedInitializerLSquare)) {
705           Left->setType(TT_ObjCMethodExpr);
706           StartsObjCMethodExpr = true;
707           Contexts.back().ColonIsObjCMethodExpr = true;
708           if (Parent && Parent->is(tok::r_paren))
709             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
710             Parent->setType(TT_CastRParen);
711         }
712         ColonFound = true;
713       }
714       if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
715           !ColonFound)
716         Left->setType(TT_ArrayInitializerLSquare);
717       FormatToken *Tok = CurrentToken;
718       if (!consumeToken())
719         return false;
720       updateParameterCount(Left, Tok);
721     }
722     return false;
723   }
724 
725   bool parseBrace() {
726     if (CurrentToken) {
727       FormatToken *Left = CurrentToken->Previous;
728       Left->ParentBracket = Contexts.back().ContextKind;
729 
730       if (Contexts.back().CaretFound)
731         Left->setType(TT_ObjCBlockLBrace);
732       Contexts.back().CaretFound = false;
733 
734       ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
735       Contexts.back().ColonIsDictLiteral = true;
736       if (Left->is(BK_BracedInit))
737         Contexts.back().IsExpression = true;
738       if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
739           Left->Previous->is(TT_JsTypeColon))
740         Contexts.back().IsExpression = false;
741 
742       while (CurrentToken) {
743         if (CurrentToken->is(tok::r_brace)) {
744           Left->MatchingParen = CurrentToken;
745           CurrentToken->MatchingParen = Left;
746           next();
747           return true;
748         }
749         if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
750           return false;
751         updateParameterCount(Left, CurrentToken);
752         if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) {
753           FormatToken *Previous = CurrentToken->getPreviousNonComment();
754           if (Previous->is(TT_JsTypeOptionalQuestion))
755             Previous = Previous->getPreviousNonComment();
756           if ((CurrentToken->is(tok::colon) &&
757                (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) ||
758               Style.Language == FormatStyle::LK_Proto ||
759               Style.Language == FormatStyle::LK_TextProto) {
760             Left->setType(TT_DictLiteral);
761             if (Previous->Tok.getIdentifierInfo() ||
762                 Previous->is(tok::string_literal))
763               Previous->setType(TT_SelectorName);
764           }
765           if (CurrentToken->is(tok::colon) ||
766               Style.Language == FormatStyle::LK_JavaScript)
767             Left->setType(TT_DictLiteral);
768         }
769         if (CurrentToken->is(tok::comma) &&
770             Style.Language == FormatStyle::LK_JavaScript)
771           Left->setType(TT_DictLiteral);
772         if (!consumeToken())
773           return false;
774       }
775     }
776     return true;
777   }
778 
779   void updateParameterCount(FormatToken *Left, FormatToken *Current) {
780     // For ObjC methods, the number of parameters is calculated differently as
781     // method declarations have a different structure (the parameters are not
782     // inside a bracket scope).
783     if (Current->is(tok::l_brace) && Current->is(BK_Block))
784       ++Left->BlockParameterCount;
785     if (Current->is(tok::comma)) {
786       ++Left->ParameterCount;
787       if (!Left->Role)
788         Left->Role.reset(new CommaSeparatedList(Style));
789       Left->Role->CommaFound(Current);
790     } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
791       Left->ParameterCount = 1;
792     }
793   }
794 
795   bool parseConditional() {
796     while (CurrentToken) {
797       if (CurrentToken->is(tok::colon)) {
798         CurrentToken->setType(TT_ConditionalExpr);
799         next();
800         return true;
801       }
802       if (!consumeToken())
803         return false;
804     }
805     return false;
806   }
807 
808   bool parseTemplateDeclaration() {
809     if (CurrentToken && CurrentToken->is(tok::less)) {
810       CurrentToken->setType(TT_TemplateOpener);
811       next();
812       if (!parseAngle())
813         return false;
814       if (CurrentToken)
815         CurrentToken->Previous->ClosesTemplateDeclaration = true;
816       return true;
817     }
818     return false;
819   }
820 
821   bool consumeToken() {
822     FormatToken *Tok = CurrentToken;
823     next();
824     switch (Tok->Tok.getKind()) {
825     case tok::plus:
826     case tok::minus:
827       if (!Tok->Previous && Line.MustBeDeclaration)
828         Tok->setType(TT_ObjCMethodSpecifier);
829       break;
830     case tok::colon:
831       if (!Tok->Previous)
832         return false;
833       // Colons from ?: are handled in parseConditional().
834       if (Style.Language == FormatStyle::LK_JavaScript) {
835         if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
836             (Contexts.size() == 1 &&               // switch/case labels
837              !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
838             Contexts.back().ContextKind == tok::l_paren ||  // function params
839             Contexts.back().ContextKind == tok::l_square || // array type
840             (!Contexts.back().IsExpression &&
841              Contexts.back().ContextKind == tok::l_brace) || // object type
842             (Contexts.size() == 1 &&
843              Line.MustBeDeclaration)) { // method/property declaration
844           Contexts.back().IsExpression = false;
845           Tok->setType(TT_JsTypeColon);
846           break;
847         }
848       } else if (Style.isCSharp()) {
849         if (Contexts.back().InCSharpAttributeSpecifier) {
850           Tok->setType(TT_AttributeColon);
851           break;
852         }
853         if (Contexts.back().ContextKind == tok::l_paren) {
854           Tok->setType(TT_CSharpNamedArgumentColon);
855           break;
856         }
857       }
858       if (Contexts.back().ColonIsDictLiteral ||
859           Style.Language == FormatStyle::LK_Proto ||
860           Style.Language == FormatStyle::LK_TextProto) {
861         Tok->setType(TT_DictLiteral);
862         if (Style.Language == FormatStyle::LK_TextProto) {
863           if (FormatToken *Previous = Tok->getPreviousNonComment())
864             Previous->setType(TT_SelectorName);
865         }
866       } else if (Contexts.back().ColonIsObjCMethodExpr ||
867                  Line.startsWith(TT_ObjCMethodSpecifier)) {
868         Tok->setType(TT_ObjCMethodExpr);
869         const FormatToken *BeforePrevious = Tok->Previous->Previous;
870         // Ensure we tag all identifiers in method declarations as
871         // TT_SelectorName.
872         bool UnknownIdentifierInMethodDeclaration =
873             Line.startsWith(TT_ObjCMethodSpecifier) &&
874             Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown);
875         if (!BeforePrevious ||
876             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
877             !(BeforePrevious->is(TT_CastRParen) ||
878               (BeforePrevious->is(TT_ObjCMethodExpr) &&
879                BeforePrevious->is(tok::colon))) ||
880             BeforePrevious->is(tok::r_square) ||
881             Contexts.back().LongestObjCSelectorName == 0 ||
882             UnknownIdentifierInMethodDeclaration) {
883           Tok->Previous->setType(TT_SelectorName);
884           if (!Contexts.back().FirstObjCSelectorName)
885             Contexts.back().FirstObjCSelectorName = Tok->Previous;
886           else if (Tok->Previous->ColumnWidth >
887                    Contexts.back().LongestObjCSelectorName)
888             Contexts.back().LongestObjCSelectorName =
889                 Tok->Previous->ColumnWidth;
890           Tok->Previous->ParameterIndex =
891               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
892           ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
893         }
894       } else if (Contexts.back().ColonIsForRangeExpr) {
895         Tok->setType(TT_RangeBasedForLoopColon);
896       } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
897         Tok->setType(TT_BitFieldColon);
898       } else if (Contexts.size() == 1 &&
899                  !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) {
900         FormatToken *Prev = Tok->getPreviousNonComment();
901         if (Prev->isOneOf(tok::r_paren, tok::kw_noexcept))
902           Tok->setType(TT_CtorInitializerColon);
903         else if (Prev->is(tok::kw_try)) {
904           // Member initializer list within function try block.
905           FormatToken *PrevPrev = Prev->getPreviousNonComment();
906           if (PrevPrev && PrevPrev->isOneOf(tok::r_paren, tok::kw_noexcept))
907             Tok->setType(TT_CtorInitializerColon);
908         } else
909           Tok->setType(TT_InheritanceColon);
910       } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next &&
911                  (Tok->Next->isOneOf(tok::r_paren, tok::comma) ||
912                   (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next &&
913                    Tok->Next->Next->is(tok::colon)))) {
914         // This handles a special macro in ObjC code where selectors including
915         // the colon are passed as macro arguments.
916         Tok->setType(TT_ObjCMethodExpr);
917       } else if (Contexts.back().ContextKind == tok::l_paren) {
918         Tok->setType(TT_InlineASMColon);
919       }
920       break;
921     case tok::pipe:
922     case tok::amp:
923       // | and & in declarations/type expressions represent union and
924       // intersection types, respectively.
925       if (Style.Language == FormatStyle::LK_JavaScript &&
926           !Contexts.back().IsExpression)
927         Tok->setType(TT_JsTypeOperator);
928       break;
929     case tok::kw_if:
930     case tok::kw_while:
931       if (Tok->is(tok::kw_if) && CurrentToken &&
932           CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier))
933         next();
934       if (CurrentToken && CurrentToken->is(tok::l_paren)) {
935         next();
936         if (!parseParens(/*LookForDecls=*/true))
937           return false;
938       }
939       break;
940     case tok::kw_for:
941       if (Style.Language == FormatStyle::LK_JavaScript) {
942         // x.for and {for: ...}
943         if ((Tok->Previous && Tok->Previous->is(tok::period)) ||
944             (Tok->Next && Tok->Next->is(tok::colon)))
945           break;
946         // JS' for await ( ...
947         if (CurrentToken && CurrentToken->is(Keywords.kw_await))
948           next();
949       }
950       Contexts.back().ColonIsForRangeExpr = true;
951       next();
952       if (!parseParens())
953         return false;
954       break;
955     case tok::l_paren:
956       // When faced with 'operator()()', the kw_operator handler incorrectly
957       // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
958       // the first two parens OverloadedOperators and the second l_paren an
959       // OverloadedOperatorLParen.
960       if (Tok->Previous && Tok->Previous->is(tok::r_paren) &&
961           Tok->Previous->MatchingParen &&
962           Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
963         Tok->Previous->setType(TT_OverloadedOperator);
964         Tok->Previous->MatchingParen->setType(TT_OverloadedOperator);
965         Tok->setType(TT_OverloadedOperatorLParen);
966       }
967 
968       if (!parseParens())
969         return false;
970       if (Line.MustBeDeclaration && Contexts.size() == 1 &&
971           !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
972           !Tok->is(TT_TypeDeclarationParen) &&
973           (!Tok->Previous || !Tok->Previous->isOneOf(tok::kw___attribute,
974                                                      TT_LeadingJavaAnnotation)))
975         Line.MightBeFunctionDecl = true;
976       break;
977     case tok::l_square:
978       if (!parseSquare())
979         return false;
980       break;
981     case tok::l_brace:
982       if (Style.Language == FormatStyle::LK_TextProto) {
983         FormatToken *Previous = Tok->getPreviousNonComment();
984         if (Previous && Previous->getType() != TT_DictLiteral)
985           Previous->setType(TT_SelectorName);
986       }
987       if (!parseBrace())
988         return false;
989       break;
990     case tok::less:
991       if (parseAngle()) {
992         Tok->setType(TT_TemplateOpener);
993         // In TT_Proto, we must distignuish between:
994         //   map<key, value>
995         //   msg < item: data >
996         //   msg: < item: data >
997         // In TT_TextProto, map<key, value> does not occur.
998         if (Style.Language == FormatStyle::LK_TextProto ||
999             (Style.Language == FormatStyle::LK_Proto && Tok->Previous &&
1000              Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
1001           Tok->setType(TT_DictLiteral);
1002           FormatToken *Previous = Tok->getPreviousNonComment();
1003           if (Previous && Previous->getType() != TT_DictLiteral)
1004             Previous->setType(TT_SelectorName);
1005         }
1006       } else {
1007         Tok->setType(TT_BinaryOperator);
1008         NonTemplateLess.insert(Tok);
1009         CurrentToken = Tok;
1010         next();
1011       }
1012       break;
1013     case tok::r_paren:
1014     case tok::r_square:
1015       return false;
1016     case tok::r_brace:
1017       // Lines can start with '}'.
1018       if (Tok->Previous)
1019         return false;
1020       break;
1021     case tok::greater:
1022       if (Style.Language != FormatStyle::LK_TextProto)
1023         Tok->setType(TT_BinaryOperator);
1024       if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser))
1025         Tok->SpacesRequiredBefore = 1;
1026       break;
1027     case tok::kw_operator:
1028       if (Style.Language == FormatStyle::LK_TextProto ||
1029           Style.Language == FormatStyle::LK_Proto)
1030         break;
1031       while (CurrentToken &&
1032              !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
1033         if (CurrentToken->isOneOf(tok::star, tok::amp))
1034           CurrentToken->setType(TT_PointerOrReference);
1035         consumeToken();
1036         if (CurrentToken && CurrentToken->is(tok::comma) &&
1037             CurrentToken->Previous->isNot(tok::kw_operator))
1038           break;
1039         if (CurrentToken && CurrentToken->Previous->isOneOf(
1040                                 TT_BinaryOperator, TT_UnaryOperator, tok::comma,
1041                                 tok::star, tok::arrow, tok::amp, tok::ampamp))
1042           CurrentToken->Previous->setType(TT_OverloadedOperator);
1043       }
1044       if (CurrentToken && CurrentToken->is(tok::l_paren))
1045         CurrentToken->setType(TT_OverloadedOperatorLParen);
1046       if (CurrentToken && CurrentToken->Previous->is(TT_BinaryOperator))
1047         CurrentToken->Previous->setType(TT_OverloadedOperator);
1048       break;
1049     case tok::question:
1050       if (Tok->is(TT_CSharpNullConditionalLSquare)) {
1051         if (!parseSquare())
1052           return false;
1053         break;
1054       }
1055       if (Tok->isOneOf(TT_CSharpNullConditional, TT_CSharpNullCoalescing))
1056         break;
1057       if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next &&
1058           Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
1059                              tok::r_brace)) {
1060         // Question marks before semicolons, colons, etc. indicate optional
1061         // types (fields, parameters), e.g.
1062         //   function(x?: string, y?) {...}
1063         //   class X { y?; }
1064         Tok->setType(TT_JsTypeOptionalQuestion);
1065         break;
1066       }
1067       // Declarations cannot be conditional expressions, this can only be part
1068       // of a type declaration.
1069       if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
1070           Style.Language == FormatStyle::LK_JavaScript)
1071         break;
1072       if (Style.isCSharp()) {
1073         // `Type?)`, `Type?>`, `Type? name;` and `Type? name =` can only be
1074         // nullable types.
1075         // Line.MustBeDeclaration will be true for `Type? name;`.
1076         if ((!Contexts.back().IsExpression && Line.MustBeDeclaration) ||
1077             (Tok->Next && Tok->Next->isOneOf(tok::r_paren, tok::greater)) ||
1078             (Tok->Next && Tok->Next->is(tok::identifier) && Tok->Next->Next &&
1079              Tok->Next->Next->is(tok::equal))) {
1080           Tok->setType(TT_CSharpNullable);
1081           break;
1082         }
1083       }
1084       parseConditional();
1085       break;
1086     case tok::kw_template:
1087       parseTemplateDeclaration();
1088       break;
1089     case tok::comma:
1090       if (Contexts.back().InCtorInitializer)
1091         Tok->setType(TT_CtorInitializerComma);
1092       else if (Contexts.back().InInheritanceList)
1093         Tok->setType(TT_InheritanceComma);
1094       else if (Contexts.back().FirstStartOfName &&
1095                (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
1096         Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
1097         Line.IsMultiVariableDeclStmt = true;
1098       }
1099       if (Contexts.back().IsForEachMacro)
1100         Contexts.back().IsExpression = true;
1101       break;
1102     case tok::identifier:
1103       if (Tok->isOneOf(Keywords.kw___has_include,
1104                        Keywords.kw___has_include_next)) {
1105         parseHasInclude();
1106       }
1107       if (Style.isCSharp() && Tok->is(Keywords.kw_where) && Tok->Next &&
1108           Tok->Next->isNot(tok::l_paren)) {
1109         Tok->setType(TT_CSharpGenericTypeConstraint);
1110         parseCSharpGenericTypeConstraint();
1111       }
1112       break;
1113     default:
1114       break;
1115     }
1116     return true;
1117   }
1118 
1119   void parseCSharpGenericTypeConstraint() {
1120     int OpenAngleBracketsCount = 0;
1121     while (CurrentToken) {
1122       if (CurrentToken->is(tok::less)) {
1123         // parseAngle is too greedy and will consume the whole line.
1124         CurrentToken->setType(TT_TemplateOpener);
1125         ++OpenAngleBracketsCount;
1126         next();
1127       } else if (CurrentToken->is(tok::greater)) {
1128         CurrentToken->setType(TT_TemplateCloser);
1129         --OpenAngleBracketsCount;
1130         next();
1131       } else if (CurrentToken->is(tok::comma) && OpenAngleBracketsCount == 0) {
1132         // We allow line breaks after GenericTypeConstraintComma's
1133         // so do not flag commas in Generics as GenericTypeConstraintComma's.
1134         CurrentToken->setType(TT_CSharpGenericTypeConstraintComma);
1135         next();
1136       } else if (CurrentToken->is(Keywords.kw_where)) {
1137         CurrentToken->setType(TT_CSharpGenericTypeConstraint);
1138         next();
1139       } else if (CurrentToken->is(tok::colon)) {
1140         CurrentToken->setType(TT_CSharpGenericTypeConstraintColon);
1141         next();
1142       } else {
1143         next();
1144       }
1145     }
1146   }
1147 
1148   void parseIncludeDirective() {
1149     if (CurrentToken && CurrentToken->is(tok::less)) {
1150       next();
1151       while (CurrentToken) {
1152         // Mark tokens up to the trailing line comments as implicit string
1153         // literals.
1154         if (CurrentToken->isNot(tok::comment) &&
1155             !CurrentToken->TokenText.startswith("//"))
1156           CurrentToken->setType(TT_ImplicitStringLiteral);
1157         next();
1158       }
1159     }
1160   }
1161 
1162   void parseWarningOrError() {
1163     next();
1164     // We still want to format the whitespace left of the first token of the
1165     // warning or error.
1166     next();
1167     while (CurrentToken) {
1168       CurrentToken->setType(TT_ImplicitStringLiteral);
1169       next();
1170     }
1171   }
1172 
1173   void parsePragma() {
1174     next(); // Consume "pragma".
1175     if (CurrentToken &&
1176         CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
1177       bool IsMark = CurrentToken->is(Keywords.kw_mark);
1178       next(); // Consume "mark".
1179       next(); // Consume first token (so we fix leading whitespace).
1180       while (CurrentToken) {
1181         if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
1182           CurrentToken->setType(TT_ImplicitStringLiteral);
1183         next();
1184       }
1185     }
1186   }
1187 
1188   void parseHasInclude() {
1189     if (!CurrentToken || !CurrentToken->is(tok::l_paren))
1190       return;
1191     next(); // '('
1192     parseIncludeDirective();
1193     next(); // ')'
1194   }
1195 
1196   LineType parsePreprocessorDirective() {
1197     bool IsFirstToken = CurrentToken->IsFirst;
1198     LineType Type = LT_PreprocessorDirective;
1199     next();
1200     if (!CurrentToken)
1201       return Type;
1202 
1203     if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
1204       // JavaScript files can contain shebang lines of the form:
1205       // #!/usr/bin/env node
1206       // Treat these like C++ #include directives.
1207       while (CurrentToken) {
1208         // Tokens cannot be comments here.
1209         CurrentToken->setType(TT_ImplicitStringLiteral);
1210         next();
1211       }
1212       return LT_ImportStatement;
1213     }
1214 
1215     if (CurrentToken->Tok.is(tok::numeric_constant)) {
1216       CurrentToken->SpacesRequiredBefore = 1;
1217       return Type;
1218     }
1219     // Hashes in the middle of a line can lead to any strange token
1220     // sequence.
1221     if (!CurrentToken->Tok.getIdentifierInfo())
1222       return Type;
1223     switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
1224     case tok::pp_include:
1225     case tok::pp_include_next:
1226     case tok::pp_import:
1227       next();
1228       parseIncludeDirective();
1229       Type = LT_ImportStatement;
1230       break;
1231     case tok::pp_error:
1232     case tok::pp_warning:
1233       parseWarningOrError();
1234       break;
1235     case tok::pp_pragma:
1236       parsePragma();
1237       break;
1238     case tok::pp_if:
1239     case tok::pp_elif:
1240       Contexts.back().IsExpression = true;
1241       next();
1242       parseLine();
1243       break;
1244     default:
1245       break;
1246     }
1247     while (CurrentToken) {
1248       FormatToken *Tok = CurrentToken;
1249       next();
1250       if (Tok->is(tok::l_paren))
1251         parseParens();
1252       else if (Tok->isOneOf(Keywords.kw___has_include,
1253                             Keywords.kw___has_include_next))
1254         parseHasInclude();
1255     }
1256     return Type;
1257   }
1258 
1259 public:
1260   LineType parseLine() {
1261     if (!CurrentToken)
1262       return LT_Invalid;
1263     NonTemplateLess.clear();
1264     if (CurrentToken->is(tok::hash))
1265       return parsePreprocessorDirective();
1266 
1267     // Directly allow to 'import <string-literal>' to support protocol buffer
1268     // definitions (github.com/google/protobuf) or missing "#" (either way we
1269     // should not break the line).
1270     IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
1271     if ((Style.Language == FormatStyle::LK_Java &&
1272          CurrentToken->is(Keywords.kw_package)) ||
1273         (Info && Info->getPPKeywordID() == tok::pp_import &&
1274          CurrentToken->Next &&
1275          CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
1276                                      tok::kw_static))) {
1277       next();
1278       parseIncludeDirective();
1279       return LT_ImportStatement;
1280     }
1281 
1282     // If this line starts and ends in '<' and '>', respectively, it is likely
1283     // part of "#define <a/b.h>".
1284     if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
1285       parseIncludeDirective();
1286       return LT_ImportStatement;
1287     }
1288 
1289     // In .proto files, top-level options and package statements are very
1290     // similar to import statements and should not be line-wrapped.
1291     if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
1292         CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) {
1293       next();
1294       if (CurrentToken && CurrentToken->is(tok::identifier)) {
1295         while (CurrentToken)
1296           next();
1297         return LT_ImportStatement;
1298       }
1299     }
1300 
1301     bool KeywordVirtualFound = false;
1302     bool ImportStatement = false;
1303 
1304     // import {...} from '...';
1305     if (Style.Language == FormatStyle::LK_JavaScript &&
1306         CurrentToken->is(Keywords.kw_import))
1307       ImportStatement = true;
1308 
1309     while (CurrentToken) {
1310       if (CurrentToken->is(tok::kw_virtual))
1311         KeywordVirtualFound = true;
1312       if (Style.Language == FormatStyle::LK_JavaScript) {
1313         // export {...} from '...';
1314         // An export followed by "from 'some string';" is a re-export from
1315         // another module identified by a URI and is treated as a
1316         // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
1317         // Just "export {...};" or "export class ..." should not be treated as
1318         // an import in this sense.
1319         if (Line.First->is(tok::kw_export) &&
1320             CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
1321             CurrentToken->Next->isStringLiteral())
1322           ImportStatement = true;
1323         if (isClosureImportStatement(*CurrentToken))
1324           ImportStatement = true;
1325       }
1326       if (!consumeToken())
1327         return LT_Invalid;
1328     }
1329     if (KeywordVirtualFound)
1330       return LT_VirtualFunctionDecl;
1331     if (ImportStatement)
1332       return LT_ImportStatement;
1333 
1334     if (Line.startsWith(TT_ObjCMethodSpecifier)) {
1335       if (Contexts.back().FirstObjCSelectorName)
1336         Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
1337             Contexts.back().LongestObjCSelectorName;
1338       return LT_ObjCMethodDecl;
1339     }
1340 
1341     return LT_Other;
1342   }
1343 
1344 private:
1345   bool isClosureImportStatement(const FormatToken &Tok) {
1346     // FIXME: Closure-library specific stuff should not be hard-coded but be
1347     // configurable.
1348     return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
1349            Tok.Next->Next &&
1350            (Tok.Next->Next->TokenText == "module" ||
1351             Tok.Next->Next->TokenText == "provide" ||
1352             Tok.Next->Next->TokenText == "require" ||
1353             Tok.Next->Next->TokenText == "requireType" ||
1354             Tok.Next->Next->TokenText == "forwardDeclare") &&
1355            Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
1356   }
1357 
1358   void resetTokenMetadata(FormatToken *Token) {
1359     if (!Token)
1360       return;
1361 
1362     // Reset token type in case we have already looked at it and then
1363     // recovered from an error (e.g. failure to find the matching >).
1364     if (!CurrentToken->isOneOf(
1365             TT_LambdaLSquare, TT_LambdaLBrace, TT_AttributeMacro,
1366             TT_ForEachMacro, TT_TypenameMacro, TT_FunctionLBrace,
1367             TT_ImplicitStringLiteral, TT_InlineASMBrace, TT_JsFatArrow,
1368             TT_LambdaArrow, TT_NamespaceMacro, TT_OverloadedOperator,
1369             TT_RegexLiteral, TT_TemplateString, TT_ObjCStringLiteral,
1370             TT_UntouchableMacroFunc, TT_ConstraintJunctions))
1371       CurrentToken->setType(TT_Unknown);
1372     CurrentToken->Role.reset();
1373     CurrentToken->MatchingParen = nullptr;
1374     CurrentToken->FakeLParens.clear();
1375     CurrentToken->FakeRParens = 0;
1376   }
1377 
1378   void next() {
1379     if (CurrentToken) {
1380       CurrentToken->NestingLevel = Contexts.size() - 1;
1381       CurrentToken->BindingStrength = Contexts.back().BindingStrength;
1382       modifyContext(*CurrentToken);
1383       determineTokenType(*CurrentToken);
1384       CurrentToken = CurrentToken->Next;
1385     }
1386 
1387     resetTokenMetadata(CurrentToken);
1388   }
1389 
1390   /// A struct to hold information valid in a specific context, e.g.
1391   /// a pair of parenthesis.
1392   struct Context {
1393     Context(tok::TokenKind ContextKind, unsigned BindingStrength,
1394             bool IsExpression)
1395         : ContextKind(ContextKind), BindingStrength(BindingStrength),
1396           IsExpression(IsExpression) {}
1397 
1398     tok::TokenKind ContextKind;
1399     unsigned BindingStrength;
1400     bool IsExpression;
1401     unsigned LongestObjCSelectorName = 0;
1402     bool ColonIsForRangeExpr = false;
1403     bool ColonIsDictLiteral = false;
1404     bool ColonIsObjCMethodExpr = false;
1405     FormatToken *FirstObjCSelectorName = nullptr;
1406     FormatToken *FirstStartOfName = nullptr;
1407     bool CanBeExpression = true;
1408     bool InTemplateArgument = false;
1409     bool InCtorInitializer = false;
1410     bool InInheritanceList = false;
1411     bool CaretFound = false;
1412     bool IsForEachMacro = false;
1413     bool InCpp11AttributeSpecifier = false;
1414     bool InCSharpAttributeSpecifier = false;
1415   };
1416 
1417   /// Puts a new \c Context onto the stack \c Contexts for the lifetime
1418   /// of each instance.
1419   struct ScopedContextCreator {
1420     AnnotatingParser &P;
1421 
1422     ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
1423                          unsigned Increase)
1424         : P(P) {
1425       P.Contexts.push_back(Context(ContextKind,
1426                                    P.Contexts.back().BindingStrength + Increase,
1427                                    P.Contexts.back().IsExpression));
1428     }
1429 
1430     ~ScopedContextCreator() { P.Contexts.pop_back(); }
1431   };
1432 
1433   void modifyContext(const FormatToken &Current) {
1434     if (Current.getPrecedence() == prec::Assignment &&
1435         !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
1436         // Type aliases use `type X = ...;` in TypeScript and can be exported
1437         // using `export type ...`.
1438         !(Style.Language == FormatStyle::LK_JavaScript &&
1439           (Line.startsWith(Keywords.kw_type, tok::identifier) ||
1440            Line.startsWith(tok::kw_export, Keywords.kw_type,
1441                            tok::identifier))) &&
1442         (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
1443       Contexts.back().IsExpression = true;
1444       if (!Line.startsWith(TT_UnaryOperator)) {
1445         for (FormatToken *Previous = Current.Previous;
1446              Previous && Previous->Previous &&
1447              !Previous->Previous->isOneOf(tok::comma, tok::semi);
1448              Previous = Previous->Previous) {
1449           if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
1450             Previous = Previous->MatchingParen;
1451             if (!Previous)
1452               break;
1453           }
1454           if (Previous->opensScope())
1455             break;
1456           if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
1457               Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
1458               Previous->Previous && Previous->Previous->isNot(tok::equal))
1459             Previous->setType(TT_PointerOrReference);
1460         }
1461       }
1462     } else if (Current.is(tok::lessless) &&
1463                (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1464       Contexts.back().IsExpression = true;
1465     } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1466       Contexts.back().IsExpression = true;
1467     } else if (Current.is(TT_TrailingReturnArrow)) {
1468       Contexts.back().IsExpression = false;
1469     } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1470       Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1471     } else if (Current.Previous &&
1472                Current.Previous->is(TT_CtorInitializerColon)) {
1473       Contexts.back().IsExpression = true;
1474       Contexts.back().InCtorInitializer = true;
1475     } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) {
1476       Contexts.back().InInheritanceList = true;
1477     } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1478       for (FormatToken *Previous = Current.Previous;
1479            Previous && Previous->isOneOf(tok::star, tok::amp);
1480            Previous = Previous->Previous)
1481         Previous->setType(TT_PointerOrReference);
1482       if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1483         Contexts.back().IsExpression = false;
1484     } else if (Current.is(tok::kw_new)) {
1485       Contexts.back().CanBeExpression = false;
1486     } else if (Current.is(tok::semi) ||
1487                (Current.is(tok::exclaim) && Current.Previous &&
1488                 !Current.Previous->is(tok::kw_operator))) {
1489       // This should be the condition or increment in a for-loop.
1490       // But not operator !() (can't use TT_OverloadedOperator here as its not
1491       // been annotated yet).
1492       Contexts.back().IsExpression = true;
1493     }
1494   }
1495 
1496   static FormatToken *untilMatchingParen(FormatToken *Current) {
1497     // Used when `MatchingParen` is not yet established.
1498     int ParenLevel = 0;
1499     while (Current) {
1500       if (Current->is(tok::l_paren))
1501         ParenLevel++;
1502       if (Current->is(tok::r_paren))
1503         ParenLevel--;
1504       if (ParenLevel < 1)
1505         break;
1506       Current = Current->Next;
1507     }
1508     return Current;
1509   }
1510 
1511   static bool isDeductionGuide(FormatToken &Current) {
1512     // Look for a deduction guide template<T> A(...) -> A<...>;
1513     if (Current.Previous && Current.Previous->is(tok::r_paren) &&
1514         Current.startsSequence(tok::arrow, tok::identifier, tok::less)) {
1515       // Find the TemplateCloser.
1516       FormatToken *TemplateCloser = Current.Next->Next;
1517       int NestingLevel = 0;
1518       while (TemplateCloser) {
1519         // Skip over an expressions in parens  A<(3 < 2)>;
1520         if (TemplateCloser->is(tok::l_paren)) {
1521           // No Matching Paren yet so skip to matching paren
1522           TemplateCloser = untilMatchingParen(TemplateCloser);
1523         }
1524         if (TemplateCloser->is(tok::less))
1525           NestingLevel++;
1526         if (TemplateCloser->is(tok::greater))
1527           NestingLevel--;
1528         if (NestingLevel < 1)
1529           break;
1530         TemplateCloser = TemplateCloser->Next;
1531       }
1532       // Assuming we have found the end of the template ensure its followed
1533       // with a semi-colon.
1534       if (TemplateCloser && TemplateCloser->Next &&
1535           TemplateCloser->Next->is(tok::semi) &&
1536           Current.Previous->MatchingParen) {
1537         // Determine if the identifier `A` prior to the A<..>; is the same as
1538         // prior to the A(..)
1539         FormatToken *LeadingIdentifier =
1540             Current.Previous->MatchingParen->Previous;
1541 
1542         // Differentiate a deduction guide by seeing the
1543         // > of the template prior to the leading identifier.
1544         if (LeadingIdentifier) {
1545           FormatToken *PriorLeadingIdentifier = LeadingIdentifier->Previous;
1546           // Skip back past explicit decoration
1547           if (PriorLeadingIdentifier &&
1548               PriorLeadingIdentifier->is(tok::kw_explicit))
1549             PriorLeadingIdentifier = PriorLeadingIdentifier->Previous;
1550 
1551           return (PriorLeadingIdentifier &&
1552                   PriorLeadingIdentifier->is(TT_TemplateCloser) &&
1553                   LeadingIdentifier->TokenText == Current.Next->TokenText);
1554         }
1555       }
1556     }
1557     return false;
1558   }
1559 
1560   void determineTokenType(FormatToken &Current) {
1561     if (!Current.is(TT_Unknown))
1562       // The token type is already known.
1563       return;
1564 
1565     if (Style.isCSharp() && CurrentToken->is(tok::question)) {
1566       if (CurrentToken->TokenText == "??") {
1567         Current.setType(TT_CSharpNullCoalescing);
1568         return;
1569       }
1570       if (CurrentToken->TokenText == "?.") {
1571         Current.setType(TT_CSharpNullConditional);
1572         return;
1573       }
1574       if (CurrentToken->TokenText == "?[") {
1575         Current.setType(TT_CSharpNullConditionalLSquare);
1576         return;
1577       }
1578     }
1579 
1580     if (Style.Language == FormatStyle::LK_JavaScript) {
1581       if (Current.is(tok::exclaim)) {
1582         if (Current.Previous &&
1583             (Keywords.IsJavaScriptIdentifier(
1584                  *Current.Previous, /* AcceptIdentifierName= */ true) ||
1585              Current.Previous->isOneOf(
1586                  tok::kw_namespace, tok::r_paren, tok::r_square, tok::r_brace,
1587                  Keywords.kw_type, Keywords.kw_get, Keywords.kw_set) ||
1588              Current.Previous->Tok.isLiteral())) {
1589           Current.setType(TT_JsNonNullAssertion);
1590           return;
1591         }
1592         if (Current.Next &&
1593             Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1594           Current.setType(TT_JsNonNullAssertion);
1595           return;
1596         }
1597       }
1598     }
1599 
1600     // Line.MightBeFunctionDecl can only be true after the parentheses of a
1601     // function declaration have been found. In this case, 'Current' is a
1602     // trailing token of this declaration and thus cannot be a name.
1603     if (Current.is(Keywords.kw_instanceof)) {
1604       Current.setType(TT_BinaryOperator);
1605     } else if (isStartOfName(Current) &&
1606                (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1607       Contexts.back().FirstStartOfName = &Current;
1608       Current.setType(TT_StartOfName);
1609     } else if (Current.is(tok::semi)) {
1610       // Reset FirstStartOfName after finding a semicolon so that a for loop
1611       // with multiple increment statements is not confused with a for loop
1612       // having multiple variable declarations.
1613       Contexts.back().FirstStartOfName = nullptr;
1614     } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1615       AutoFound = true;
1616     } else if (Current.is(tok::arrow) &&
1617                Style.Language == FormatStyle::LK_Java) {
1618       Current.setType(TT_LambdaArrow);
1619     } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1620                Current.NestingLevel == 0 &&
1621                !Current.Previous->is(tok::kw_operator)) {
1622       // not auto operator->() -> xxx;
1623       Current.setType(TT_TrailingReturnArrow);
1624     } else if (Current.is(tok::arrow) && Current.Previous &&
1625                Current.Previous->is(tok::r_brace)) {
1626       // Concept implicit conversion contraint needs to be treated like
1627       // a trailing return type  ... } -> <type>.
1628       Current.setType(TT_TrailingReturnArrow);
1629     } else if (isDeductionGuide(Current)) {
1630       // Deduction guides trailing arrow " A(...) -> A<T>;".
1631       Current.setType(TT_TrailingReturnArrow);
1632     } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1633       Current.setType(determineStarAmpUsage(
1634           Current,
1635           Contexts.back().CanBeExpression && Contexts.back().IsExpression,
1636           Contexts.back().InTemplateArgument));
1637     } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1638       Current.setType(determinePlusMinusCaretUsage(Current));
1639       if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1640         Contexts.back().CaretFound = true;
1641     } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1642       Current.setType(determineIncrementUsage(Current));
1643     } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1644       Current.setType(TT_UnaryOperator);
1645     } else if (Current.is(tok::question)) {
1646       if (Style.Language == FormatStyle::LK_JavaScript &&
1647           Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1648         // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1649         // on the interface, not a ternary expression.
1650         Current.setType(TT_JsTypeOptionalQuestion);
1651       } else {
1652         Current.setType(TT_ConditionalExpr);
1653       }
1654     } else if (Current.isBinaryOperator() &&
1655                (!Current.Previous || Current.Previous->isNot(tok::l_square)) &&
1656                (!Current.is(tok::greater) &&
1657                 Style.Language != FormatStyle::LK_TextProto)) {
1658       Current.setType(TT_BinaryOperator);
1659     } else if (Current.is(tok::comment)) {
1660       if (Current.TokenText.startswith("/*")) {
1661         if (Current.TokenText.endswith("*/"))
1662           Current.setType(TT_BlockComment);
1663         else
1664           // The lexer has for some reason determined a comment here. But we
1665           // cannot really handle it, if it isn't properly terminated.
1666           Current.Tok.setKind(tok::unknown);
1667       } else {
1668         Current.setType(TT_LineComment);
1669       }
1670     } else if (Current.is(tok::r_paren)) {
1671       if (rParenEndsCast(Current))
1672         Current.setType(TT_CastRParen);
1673       if (Current.MatchingParen && Current.Next &&
1674           !Current.Next->isBinaryOperator() &&
1675           !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1676                                  tok::comma, tok::period, tok::arrow,
1677                                  tok::coloncolon))
1678         if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1679           // Make sure this isn't the return type of an Obj-C block declaration
1680           if (AfterParen->Tok.isNot(tok::caret)) {
1681             if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1682               if (BeforeParen->is(tok::identifier) &&
1683                   !BeforeParen->is(TT_TypenameMacro) &&
1684                   BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1685                   (!BeforeParen->Previous ||
1686                    BeforeParen->Previous->ClosesTemplateDeclaration))
1687                 Current.setType(TT_FunctionAnnotationRParen);
1688           }
1689         }
1690     } else if (Current.is(tok::at) && Current.Next &&
1691                Style.Language != FormatStyle::LK_JavaScript &&
1692                Style.Language != FormatStyle::LK_Java) {
1693       // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1694       // marks declarations and properties that need special formatting.
1695       switch (Current.Next->Tok.getObjCKeywordID()) {
1696       case tok::objc_interface:
1697       case tok::objc_implementation:
1698       case tok::objc_protocol:
1699         Current.setType(TT_ObjCDecl);
1700         break;
1701       case tok::objc_property:
1702         Current.setType(TT_ObjCProperty);
1703         break;
1704       default:
1705         break;
1706       }
1707     } else if (Current.is(tok::period)) {
1708       FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1709       if (PreviousNoComment &&
1710           PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1711         Current.setType(TT_DesignatedInitializerPeriod);
1712       else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1713                Current.Previous->isOneOf(TT_JavaAnnotation,
1714                                          TT_LeadingJavaAnnotation)) {
1715         Current.setType(Current.Previous->getType());
1716       }
1717     } else if (canBeObjCSelectorComponent(Current) &&
1718                // FIXME(bug 36976): ObjC return types shouldn't use
1719                // TT_CastRParen.
1720                Current.Previous && Current.Previous->is(TT_CastRParen) &&
1721                Current.Previous->MatchingParen &&
1722                Current.Previous->MatchingParen->Previous &&
1723                Current.Previous->MatchingParen->Previous->is(
1724                    TT_ObjCMethodSpecifier)) {
1725       // This is the first part of an Objective-C selector name. (If there's no
1726       // colon after this, this is the only place which annotates the identifier
1727       // as a selector.)
1728       Current.setType(TT_SelectorName);
1729     } else if (Current.isOneOf(tok::identifier, tok::kw_const, tok::kw_noexcept,
1730                                tok::kw_requires) &&
1731                Current.Previous &&
1732                !Current.Previous->isOneOf(tok::equal, tok::at) &&
1733                Line.MightBeFunctionDecl && Contexts.size() == 1) {
1734       // Line.MightBeFunctionDecl can only be true after the parentheses of a
1735       // function declaration have been found.
1736       Current.setType(TT_TrailingAnnotation);
1737     } else if ((Style.Language == FormatStyle::LK_Java ||
1738                 Style.Language == FormatStyle::LK_JavaScript) &&
1739                Current.Previous) {
1740       if (Current.Previous->is(tok::at) &&
1741           Current.isNot(Keywords.kw_interface)) {
1742         const FormatToken &AtToken = *Current.Previous;
1743         const FormatToken *Previous = AtToken.getPreviousNonComment();
1744         if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1745           Current.setType(TT_LeadingJavaAnnotation);
1746         else
1747           Current.setType(TT_JavaAnnotation);
1748       } else if (Current.Previous->is(tok::period) &&
1749                  Current.Previous->isOneOf(TT_JavaAnnotation,
1750                                            TT_LeadingJavaAnnotation)) {
1751         Current.setType(Current.Previous->getType());
1752       }
1753     }
1754   }
1755 
1756   /// Take a guess at whether \p Tok starts a name of a function or
1757   /// variable declaration.
1758   ///
1759   /// This is a heuristic based on whether \p Tok is an identifier following
1760   /// something that is likely a type.
1761   bool isStartOfName(const FormatToken &Tok) {
1762     if (Tok.isNot(tok::identifier) || !Tok.Previous)
1763       return false;
1764 
1765     if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1766                               Keywords.kw_as))
1767       return false;
1768     if (Style.Language == FormatStyle::LK_JavaScript &&
1769         Tok.Previous->is(Keywords.kw_in))
1770       return false;
1771 
1772     // Skip "const" as it does not have an influence on whether this is a name.
1773     FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1774     while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1775       PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1776 
1777     if (!PreviousNotConst)
1778       return false;
1779 
1780     bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1781                        PreviousNotConst->Previous &&
1782                        PreviousNotConst->Previous->is(tok::hash);
1783 
1784     if (PreviousNotConst->is(TT_TemplateCloser))
1785       return PreviousNotConst && PreviousNotConst->MatchingParen &&
1786              PreviousNotConst->MatchingParen->Previous &&
1787              PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1788              PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1789 
1790     if (PreviousNotConst->is(tok::r_paren) &&
1791         PreviousNotConst->is(TT_TypeDeclarationParen))
1792       return true;
1793 
1794     return (!IsPPKeyword &&
1795             PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1796            PreviousNotConst->is(TT_PointerOrReference) ||
1797            PreviousNotConst->isSimpleTypeSpecifier();
1798   }
1799 
1800   /// Determine whether ')' is ending a cast.
1801   bool rParenEndsCast(const FormatToken &Tok) {
1802     // C-style casts are only used in C++, C# and Java.
1803     if (!Style.isCSharp() && !Style.isCpp() &&
1804         Style.Language != FormatStyle::LK_Java)
1805       return false;
1806 
1807     // Empty parens aren't casts and there are no casts at the end of the line.
1808     if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1809       return false;
1810 
1811     FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1812     if (LeftOfParens) {
1813       // If there is a closing parenthesis left of the current parentheses,
1814       // look past it as these might be chained casts.
1815       if (LeftOfParens->is(tok::r_paren)) {
1816         if (!LeftOfParens->MatchingParen ||
1817             !LeftOfParens->MatchingParen->Previous)
1818           return false;
1819         LeftOfParens = LeftOfParens->MatchingParen->Previous;
1820       }
1821 
1822       // If there is an identifier (or with a few exceptions a keyword) right
1823       // before the parentheses, this is unlikely to be a cast.
1824       if (LeftOfParens->Tok.getIdentifierInfo() &&
1825           !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1826                                  tok::kw_delete))
1827         return false;
1828 
1829       // Certain other tokens right before the parentheses are also signals that
1830       // this cannot be a cast.
1831       if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1832                                 TT_TemplateCloser, tok::ellipsis))
1833         return false;
1834     }
1835 
1836     if (Tok.Next->is(tok::question))
1837       return false;
1838 
1839     // `foreach((A a, B b) in someList)` should not be seen as a cast.
1840     if (Tok.Next->is(Keywords.kw_in) && Style.isCSharp())
1841       return false;
1842 
1843     // Functions which end with decorations like volatile, noexcept are unlikely
1844     // to be casts.
1845     if (Tok.Next->isOneOf(tok::kw_noexcept, tok::kw_volatile, tok::kw_const,
1846                           tok::kw_requires, tok::kw_throw, tok::arrow,
1847                           Keywords.kw_override, Keywords.kw_final) ||
1848         isCpp11AttributeSpecifier(*Tok.Next))
1849       return false;
1850 
1851     // As Java has no function types, a "(" after the ")" likely means that this
1852     // is a cast.
1853     if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1854       return true;
1855 
1856     // If a (non-string) literal follows, this is likely a cast.
1857     if (Tok.Next->isNot(tok::string_literal) &&
1858         (Tok.Next->Tok.isLiteral() ||
1859          Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1860       return true;
1861 
1862     // Heuristically try to determine whether the parentheses contain a type.
1863     auto IsQualifiedPointerOrReference = [](FormatToken *T) {
1864       // This is used to handle cases such as x = (foo *const)&y;
1865       assert(!T->isSimpleTypeSpecifier() && "Should have already been checked");
1866       // Strip trailing qualifiers such as const or volatile when checking
1867       // whether the parens could be a cast to a pointer/reference type.
1868       while (T) {
1869         if (T->is(TT_AttributeParen)) {
1870           // Handle `x = (foo *__attribute__((foo)))&v;`:
1871           if (T->MatchingParen && T->MatchingParen->Previous &&
1872               T->MatchingParen->Previous->is(tok::kw___attribute)) {
1873             T = T->MatchingParen->Previous->Previous;
1874             continue;
1875           }
1876         } else if (T->is(TT_AttributeSquare)) {
1877           // Handle `x = (foo *[[clang::foo]])&v;`:
1878           if (T->MatchingParen && T->MatchingParen->Previous) {
1879             T = T->MatchingParen->Previous;
1880             continue;
1881           }
1882         } else if (T->canBePointerOrReferenceQualifier()) {
1883           T = T->Previous;
1884           continue;
1885         }
1886         break;
1887       }
1888       return T && T->is(TT_PointerOrReference);
1889     };
1890     bool ParensAreType =
1891         !Tok.Previous ||
1892         Tok.Previous->isOneOf(TT_TemplateCloser, TT_TypeDeclarationParen) ||
1893         Tok.Previous->isSimpleTypeSpecifier() ||
1894         IsQualifiedPointerOrReference(Tok.Previous);
1895     bool ParensCouldEndDecl =
1896         Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1897     if (ParensAreType && !ParensCouldEndDecl)
1898       return true;
1899 
1900     // At this point, we heuristically assume that there are no casts at the
1901     // start of the line. We assume that we have found most cases where there
1902     // are by the logic above, e.g. "(void)x;".
1903     if (!LeftOfParens)
1904       return false;
1905 
1906     // Certain token types inside the parentheses mean that this can't be a
1907     // cast.
1908     for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1909          Token = Token->Next)
1910       if (Token->is(TT_BinaryOperator))
1911         return false;
1912 
1913     // If the following token is an identifier or 'this', this is a cast. All
1914     // cases where this can be something else are handled above.
1915     if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1916       return true;
1917 
1918     if (!Tok.Next->Next)
1919       return false;
1920 
1921     // If the next token after the parenthesis is a unary operator, assume
1922     // that this is cast, unless there are unexpected tokens inside the
1923     // parenthesis.
1924     bool NextIsUnary =
1925         Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1926     if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1927         !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1928       return false;
1929     // Search for unexpected tokens.
1930     for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1931          Prev = Prev->Previous) {
1932       if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1933         return false;
1934     }
1935     return true;
1936   }
1937 
1938   /// Return the type of the given token assuming it is * or &.
1939   TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1940                                   bool InTemplateArgument) {
1941     if (Style.Language == FormatStyle::LK_JavaScript)
1942       return TT_BinaryOperator;
1943 
1944     // && in C# must be a binary operator.
1945     if (Style.isCSharp() && Tok.is(tok::ampamp))
1946       return TT_BinaryOperator;
1947 
1948     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1949     if (!PrevToken)
1950       return TT_UnaryOperator;
1951 
1952     const FormatToken *NextToken = Tok.getNextNonComment();
1953     if (!NextToken ||
1954         NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_noexcept) ||
1955         NextToken->canBePointerOrReferenceQualifier() ||
1956         (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1957       return TT_PointerOrReference;
1958 
1959     if (PrevToken->is(tok::coloncolon))
1960       return TT_PointerOrReference;
1961 
1962     if (PrevToken->is(tok::r_paren) && PrevToken->is(TT_TypeDeclarationParen))
1963       return TT_PointerOrReference;
1964 
1965     if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1966                            tok::comma, tok::semi, tok::kw_return, tok::colon,
1967                            tok::kw_co_return, tok::kw_co_await, tok::kw_co_yield,
1968                            tok::equal, tok::kw_delete, tok::kw_sizeof,
1969                            tok::kw_throw) ||
1970         PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1971                            TT_UnaryOperator, TT_CastRParen))
1972       return TT_UnaryOperator;
1973 
1974     if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1975       return TT_PointerOrReference;
1976     if (NextToken->is(tok::kw_operator) && !IsExpression)
1977       return TT_PointerOrReference;
1978     if (NextToken->isOneOf(tok::comma, tok::semi))
1979       return TT_PointerOrReference;
1980 
1981     if (PrevToken->Tok.isLiteral() ||
1982         PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1983                            tok::kw_false, tok::r_brace) ||
1984         NextToken->Tok.isLiteral() ||
1985         NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1986         NextToken->isUnaryOperator() ||
1987         // If we know we're in a template argument, there are no named
1988         // declarations. Thus, having an identifier on the right-hand side
1989         // indicates a binary operator.
1990         (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1991       return TT_BinaryOperator;
1992 
1993     // "&&(" is quite unlikely to be two successive unary "&".
1994     if (Tok.is(tok::ampamp) && NextToken->is(tok::l_paren))
1995       return TT_BinaryOperator;
1996 
1997     // This catches some cases where evaluation order is used as control flow:
1998     //   aaa && aaa->f();
1999     if (NextToken->Tok.isAnyIdentifier()) {
2000       const FormatToken *NextNextToken = NextToken->getNextNonComment();
2001       if (NextNextToken && NextNextToken->is(tok::arrow))
2002         return TT_BinaryOperator;
2003     }
2004 
2005     // It is very unlikely that we are going to find a pointer or reference type
2006     // definition on the RHS of an assignment.
2007     if (IsExpression && !Contexts.back().CaretFound)
2008       return TT_BinaryOperator;
2009 
2010     return TT_PointerOrReference;
2011   }
2012 
2013   TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
2014     const FormatToken *PrevToken = Tok.getPreviousNonComment();
2015     if (!PrevToken)
2016       return TT_UnaryOperator;
2017 
2018     if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator))
2019       // This must be a sequence of leading unary operators.
2020       return TT_UnaryOperator;
2021 
2022     // Use heuristics to recognize unary operators.
2023     if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
2024                            tok::question, tok::colon, tok::kw_return,
2025                            tok::kw_case, tok::at, tok::l_brace, tok::kw_throw,
2026                            tok::kw_co_return, tok::kw_co_yield))
2027       return TT_UnaryOperator;
2028 
2029     // There can't be two consecutive binary operators.
2030     if (PrevToken->is(TT_BinaryOperator))
2031       return TT_UnaryOperator;
2032 
2033     // Fall back to marking the token as binary operator.
2034     return TT_BinaryOperator;
2035   }
2036 
2037   /// Determine whether ++/-- are pre- or post-increments/-decrements.
2038   TokenType determineIncrementUsage(const FormatToken &Tok) {
2039     const FormatToken *PrevToken = Tok.getPreviousNonComment();
2040     if (!PrevToken || PrevToken->is(TT_CastRParen))
2041       return TT_UnaryOperator;
2042     if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
2043       return TT_TrailingUnaryOperator;
2044 
2045     return TT_UnaryOperator;
2046   }
2047 
2048   SmallVector<Context, 8> Contexts;
2049 
2050   const FormatStyle &Style;
2051   AnnotatedLine &Line;
2052   FormatToken *CurrentToken;
2053   bool AutoFound;
2054   const AdditionalKeywords &Keywords;
2055 
2056   // Set of "<" tokens that do not open a template parameter list. If parseAngle
2057   // determines that a specific token can't be a template opener, it will make
2058   // same decision irrespective of the decisions for tokens leading up to it.
2059   // Store this information to prevent this from causing exponential runtime.
2060   llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
2061 };
2062 
2063 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
2064 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
2065 
2066 /// Parses binary expressions by inserting fake parenthesis based on
2067 /// operator precedence.
2068 class ExpressionParser {
2069 public:
2070   ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
2071                    AnnotatedLine &Line)
2072       : Style(Style), Keywords(Keywords), Current(Line.First) {}
2073 
2074   /// Parse expressions with the given operator precedence.
2075   void parse(int Precedence = 0) {
2076     // Skip 'return' and ObjC selector colons as they are not part of a binary
2077     // expression.
2078     while (Current && (Current->is(tok::kw_return) ||
2079                        (Current->is(tok::colon) &&
2080                         Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
2081       next();
2082 
2083     if (!Current || Precedence > PrecedenceArrowAndPeriod)
2084       return;
2085 
2086     // Conditional expressions need to be parsed separately for proper nesting.
2087     if (Precedence == prec::Conditional) {
2088       parseConditionalExpr();
2089       return;
2090     }
2091 
2092     // Parse unary operators, which all have a higher precedence than binary
2093     // operators.
2094     if (Precedence == PrecedenceUnaryOperator) {
2095       parseUnaryOperator();
2096       return;
2097     }
2098 
2099     FormatToken *Start = Current;
2100     FormatToken *LatestOperator = nullptr;
2101     unsigned OperatorIndex = 0;
2102 
2103     while (Current) {
2104       // Consume operators with higher precedence.
2105       parse(Precedence + 1);
2106 
2107       int CurrentPrecedence = getCurrentPrecedence();
2108 
2109       if (Current && Current->is(TT_SelectorName) &&
2110           Precedence == CurrentPrecedence) {
2111         if (LatestOperator)
2112           addFakeParenthesis(Start, prec::Level(Precedence));
2113         Start = Current;
2114       }
2115 
2116       // At the end of the line or when an operator with higher precedence is
2117       // found, insert fake parenthesis and return.
2118       if (!Current ||
2119           (Current->closesScope() &&
2120            (Current->MatchingParen || Current->is(TT_TemplateString))) ||
2121           (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
2122           (CurrentPrecedence == prec::Conditional &&
2123            Precedence == prec::Assignment && Current->is(tok::colon))) {
2124         break;
2125       }
2126 
2127       // Consume scopes: (), [], <> and {}
2128       if (Current->opensScope()) {
2129         // In fragment of a JavaScript template string can look like '}..${' and
2130         // thus close a scope and open a new one at the same time.
2131         while (Current && (!Current->closesScope() || Current->opensScope())) {
2132           next();
2133           parse();
2134         }
2135         next();
2136       } else {
2137         // Operator found.
2138         if (CurrentPrecedence == Precedence) {
2139           if (LatestOperator)
2140             LatestOperator->NextOperator = Current;
2141           LatestOperator = Current;
2142           Current->OperatorIndex = OperatorIndex;
2143           ++OperatorIndex;
2144         }
2145         next(/*SkipPastLeadingComments=*/Precedence > 0);
2146       }
2147     }
2148 
2149     if (LatestOperator && (Current || Precedence > 0)) {
2150       // LatestOperator->LastOperator = true;
2151       if (Precedence == PrecedenceArrowAndPeriod) {
2152         // Call expressions don't have a binary operator precedence.
2153         addFakeParenthesis(Start, prec::Unknown);
2154       } else {
2155         addFakeParenthesis(Start, prec::Level(Precedence));
2156       }
2157     }
2158   }
2159 
2160 private:
2161   /// Gets the precedence (+1) of the given token for binary operators
2162   /// and other tokens that we treat like binary operators.
2163   int getCurrentPrecedence() {
2164     if (Current) {
2165       const FormatToken *NextNonComment = Current->getNextNonComment();
2166       if (Current->is(TT_ConditionalExpr))
2167         return prec::Conditional;
2168       if (NextNonComment && Current->is(TT_SelectorName) &&
2169           (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) ||
2170            ((Style.Language == FormatStyle::LK_Proto ||
2171              Style.Language == FormatStyle::LK_TextProto) &&
2172             NextNonComment->is(tok::less))))
2173         return prec::Assignment;
2174       if (Current->is(TT_JsComputedPropertyName))
2175         return prec::Assignment;
2176       if (Current->is(TT_LambdaArrow))
2177         return prec::Comma;
2178       if (Current->is(TT_JsFatArrow))
2179         return prec::Assignment;
2180       if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
2181           (Current->is(tok::comment) && NextNonComment &&
2182            NextNonComment->is(TT_SelectorName)))
2183         return 0;
2184       if (Current->is(TT_RangeBasedForLoopColon))
2185         return prec::Comma;
2186       if ((Style.Language == FormatStyle::LK_Java ||
2187            Style.Language == FormatStyle::LK_JavaScript) &&
2188           Current->is(Keywords.kw_instanceof))
2189         return prec::Relational;
2190       if (Style.Language == FormatStyle::LK_JavaScript &&
2191           Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
2192         return prec::Relational;
2193       if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
2194         return Current->getPrecedence();
2195       if (Current->isOneOf(tok::period, tok::arrow))
2196         return PrecedenceArrowAndPeriod;
2197       if ((Style.Language == FormatStyle::LK_Java ||
2198            Style.Language == FormatStyle::LK_JavaScript) &&
2199           Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
2200                            Keywords.kw_throws))
2201         return 0;
2202     }
2203     return -1;
2204   }
2205 
2206   void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
2207     Start->FakeLParens.push_back(Precedence);
2208     if (Precedence > prec::Unknown)
2209       Start->StartsBinaryExpression = true;
2210     if (Current) {
2211       FormatToken *Previous = Current->Previous;
2212       while (Previous->is(tok::comment) && Previous->Previous)
2213         Previous = Previous->Previous;
2214       ++Previous->FakeRParens;
2215       if (Precedence > prec::Unknown)
2216         Previous->EndsBinaryExpression = true;
2217     }
2218   }
2219 
2220   /// Parse unary operator expressions and surround them with fake
2221   /// parentheses if appropriate.
2222   void parseUnaryOperator() {
2223     llvm::SmallVector<FormatToken *, 2> Tokens;
2224     while (Current && Current->is(TT_UnaryOperator)) {
2225       Tokens.push_back(Current);
2226       next();
2227     }
2228     parse(PrecedenceArrowAndPeriod);
2229     for (FormatToken *Token : llvm::reverse(Tokens))
2230       // The actual precedence doesn't matter.
2231       addFakeParenthesis(Token, prec::Unknown);
2232   }
2233 
2234   void parseConditionalExpr() {
2235     while (Current && Current->isTrailingComment()) {
2236       next();
2237     }
2238     FormatToken *Start = Current;
2239     parse(prec::LogicalOr);
2240     if (!Current || !Current->is(tok::question))
2241       return;
2242     next();
2243     parse(prec::Assignment);
2244     if (!Current || Current->isNot(TT_ConditionalExpr))
2245       return;
2246     next();
2247     parse(prec::Assignment);
2248     addFakeParenthesis(Start, prec::Conditional);
2249   }
2250 
2251   void next(bool SkipPastLeadingComments = true) {
2252     if (Current)
2253       Current = Current->Next;
2254     while (Current &&
2255            (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
2256            Current->isTrailingComment())
2257       Current = Current->Next;
2258   }
2259 
2260   const FormatStyle &Style;
2261   const AdditionalKeywords &Keywords;
2262   FormatToken *Current;
2263 };
2264 
2265 } // end anonymous namespace
2266 
2267 void TokenAnnotator::setCommentLineLevels(
2268     SmallVectorImpl<AnnotatedLine *> &Lines) {
2269   const AnnotatedLine *NextNonCommentLine = nullptr;
2270   for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
2271                                                           E = Lines.rend();
2272        I != E; ++I) {
2273     bool CommentLine = true;
2274     for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
2275       if (!Tok->is(tok::comment)) {
2276         CommentLine = false;
2277         break;
2278       }
2279     }
2280 
2281     // If the comment is currently aligned with the line immediately following
2282     // it, that's probably intentional and we should keep it.
2283     if (NextNonCommentLine && CommentLine &&
2284         NextNonCommentLine->First->NewlinesBefore <= 1 &&
2285         NextNonCommentLine->First->OriginalColumn ==
2286             (*I)->First->OriginalColumn) {
2287       // Align comments for preprocessor lines with the # in column 0 if
2288       // preprocessor lines are not indented. Otherwise, align with the next
2289       // line.
2290       (*I)->Level =
2291           (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
2292            (NextNonCommentLine->Type == LT_PreprocessorDirective ||
2293             NextNonCommentLine->Type == LT_ImportStatement))
2294               ? 0
2295               : NextNonCommentLine->Level;
2296     } else {
2297       NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
2298     }
2299 
2300     setCommentLineLevels((*I)->Children);
2301   }
2302 }
2303 
2304 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
2305   unsigned Result = 0;
2306   for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
2307     Result = std::max(Result, Tok->NestingLevel);
2308   return Result;
2309 }
2310 
2311 void TokenAnnotator::annotate(AnnotatedLine &Line) {
2312   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2313                                                   E = Line.Children.end();
2314        I != E; ++I) {
2315     annotate(**I);
2316   }
2317   AnnotatingParser Parser(Style, Line, Keywords);
2318   Line.Type = Parser.parseLine();
2319 
2320   // With very deep nesting, ExpressionParser uses lots of stack and the
2321   // formatting algorithm is very slow. We're not going to do a good job here
2322   // anyway - it's probably generated code being formatted by mistake.
2323   // Just skip the whole line.
2324   if (maxNestingDepth(Line) > 50)
2325     Line.Type = LT_Invalid;
2326 
2327   if (Line.Type == LT_Invalid)
2328     return;
2329 
2330   ExpressionParser ExprParser(Style, Keywords, Line);
2331   ExprParser.parse();
2332 
2333   if (Line.startsWith(TT_ObjCMethodSpecifier))
2334     Line.Type = LT_ObjCMethodDecl;
2335   else if (Line.startsWith(TT_ObjCDecl))
2336     Line.Type = LT_ObjCDecl;
2337   else if (Line.startsWith(TT_ObjCProperty))
2338     Line.Type = LT_ObjCProperty;
2339 
2340   Line.First->SpacesRequiredBefore = 1;
2341   Line.First->CanBreakBefore = Line.First->MustBreakBefore;
2342 }
2343 
2344 // This function heuristically determines whether 'Current' starts the name of a
2345 // function declaration.
2346 static bool isFunctionDeclarationName(const FormatToken &Current,
2347                                       const AnnotatedLine &Line) {
2348   auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * {
2349     for (; Next; Next = Next->Next) {
2350       if (Next->is(TT_OverloadedOperatorLParen))
2351         return Next;
2352       if (Next->is(TT_OverloadedOperator))
2353         continue;
2354       if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
2355         // For 'new[]' and 'delete[]'.
2356         if (Next->Next &&
2357             Next->Next->startsSequence(tok::l_square, tok::r_square))
2358           Next = Next->Next->Next;
2359         continue;
2360       }
2361       if (Next->startsSequence(tok::l_square, tok::r_square)) {
2362         // For operator[]().
2363         Next = Next->Next;
2364         continue;
2365       }
2366       if ((Next->isSimpleTypeSpecifier() || Next->is(tok::identifier)) &&
2367           Next->Next && Next->Next->isOneOf(tok::star, tok::amp, tok::ampamp)) {
2368         // For operator void*(), operator char*(), operator Foo*().
2369         Next = Next->Next;
2370         continue;
2371       }
2372       if (Next->is(TT_TemplateOpener) && Next->MatchingParen) {
2373         Next = Next->MatchingParen;
2374         continue;
2375       }
2376 
2377       break;
2378     }
2379     return nullptr;
2380   };
2381 
2382   // Find parentheses of parameter list.
2383   const FormatToken *Next = Current.Next;
2384   if (Current.is(tok::kw_operator)) {
2385     if (Current.Previous && Current.Previous->is(tok::coloncolon))
2386       return false;
2387     Next = skipOperatorName(Next);
2388   } else {
2389     if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
2390       return false;
2391     for (; Next; Next = Next->Next) {
2392       if (Next->is(TT_TemplateOpener)) {
2393         Next = Next->MatchingParen;
2394       } else if (Next->is(tok::coloncolon)) {
2395         Next = Next->Next;
2396         if (!Next)
2397           return false;
2398         if (Next->is(tok::kw_operator)) {
2399           Next = skipOperatorName(Next->Next);
2400           break;
2401         }
2402         if (!Next->is(tok::identifier))
2403           return false;
2404       } else if (Next->is(tok::l_paren)) {
2405         break;
2406       } else {
2407         return false;
2408       }
2409     }
2410   }
2411 
2412   // Check whether parameter list can belong to a function declaration.
2413   if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
2414     return false;
2415   // If the lines ends with "{", this is likely an function definition.
2416   if (Line.Last->is(tok::l_brace))
2417     return true;
2418   if (Next->Next == Next->MatchingParen)
2419     return true; // Empty parentheses.
2420   // If there is an &/&& after the r_paren, this is likely a function.
2421   if (Next->MatchingParen->Next &&
2422       Next->MatchingParen->Next->is(TT_PointerOrReference))
2423     return true;
2424   for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
2425        Tok = Tok->Next) {
2426     if (Tok->is(TT_TypeDeclarationParen))
2427       return true;
2428     if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) {
2429       Tok = Tok->MatchingParen;
2430       continue;
2431     }
2432     if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
2433         Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
2434       return true;
2435     if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
2436         Tok->Tok.isLiteral())
2437       return false;
2438   }
2439   return false;
2440 }
2441 
2442 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
2443   assert(Line.MightBeFunctionDecl);
2444 
2445   if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
2446        Style.AlwaysBreakAfterReturnType ==
2447            FormatStyle::RTBS_TopLevelDefinitions) &&
2448       Line.Level > 0)
2449     return false;
2450 
2451   switch (Style.AlwaysBreakAfterReturnType) {
2452   case FormatStyle::RTBS_None:
2453     return false;
2454   case FormatStyle::RTBS_All:
2455   case FormatStyle::RTBS_TopLevel:
2456     return true;
2457   case FormatStyle::RTBS_AllDefinitions:
2458   case FormatStyle::RTBS_TopLevelDefinitions:
2459     return Line.mightBeFunctionDefinition();
2460   }
2461 
2462   return false;
2463 }
2464 
2465 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
2466   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2467                                                   E = Line.Children.end();
2468        I != E; ++I) {
2469     calculateFormattingInformation(**I);
2470   }
2471 
2472   Line.First->TotalLength =
2473       Line.First->IsMultiline ? Style.ColumnLimit
2474                               : Line.FirstStartColumn + Line.First->ColumnWidth;
2475   FormatToken *Current = Line.First->Next;
2476   bool InFunctionDecl = Line.MightBeFunctionDecl;
2477   while (Current) {
2478     if (isFunctionDeclarationName(*Current, Line))
2479       Current->setType(TT_FunctionDeclarationName);
2480     if (Current->is(TT_LineComment)) {
2481       if (Current->Previous->is(BK_BracedInit) &&
2482           Current->Previous->opensScope())
2483         Current->SpacesRequiredBefore =
2484             (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1;
2485       else
2486         Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
2487 
2488       // If we find a trailing comment, iterate backwards to determine whether
2489       // it seems to relate to a specific parameter. If so, break before that
2490       // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
2491       // to the previous line in:
2492       //   SomeFunction(a,
2493       //                b, // comment
2494       //                c);
2495       if (!Current->HasUnescapedNewline) {
2496         for (FormatToken *Parameter = Current->Previous; Parameter;
2497              Parameter = Parameter->Previous) {
2498           if (Parameter->isOneOf(tok::comment, tok::r_brace))
2499             break;
2500           if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
2501             if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
2502                 Parameter->HasUnescapedNewline)
2503               Parameter->MustBreakBefore = true;
2504             break;
2505           }
2506         }
2507       }
2508     } else if (Current->SpacesRequiredBefore == 0 &&
2509                spaceRequiredBefore(Line, *Current)) {
2510       Current->SpacesRequiredBefore = 1;
2511     }
2512 
2513     Current->MustBreakBefore =
2514         Current->MustBreakBefore || mustBreakBefore(Line, *Current);
2515 
2516     if (!Current->MustBreakBefore && InFunctionDecl &&
2517         Current->is(TT_FunctionDeclarationName))
2518       Current->MustBreakBefore = mustBreakForReturnType(Line);
2519 
2520     Current->CanBreakBefore =
2521         Current->MustBreakBefore || canBreakBefore(Line, *Current);
2522     unsigned ChildSize = 0;
2523     if (Current->Previous->Children.size() == 1) {
2524       FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
2525       ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
2526                                                   : LastOfChild.TotalLength + 1;
2527     }
2528     const FormatToken *Prev = Current->Previous;
2529     if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
2530         (Prev->Children.size() == 1 &&
2531          Prev->Children[0]->First->MustBreakBefore) ||
2532         Current->IsMultiline)
2533       Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
2534     else
2535       Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
2536                              ChildSize + Current->SpacesRequiredBefore;
2537 
2538     if (Current->is(TT_CtorInitializerColon))
2539       InFunctionDecl = false;
2540 
2541     // FIXME: Only calculate this if CanBreakBefore is true once static
2542     // initializers etc. are sorted out.
2543     // FIXME: Move magic numbers to a better place.
2544 
2545     // Reduce penalty for aligning ObjC method arguments using the colon
2546     // alignment as this is the canonical way (still prefer fitting everything
2547     // into one line if possible). Trying to fit a whole expression into one
2548     // line should not force other line breaks (e.g. when ObjC method
2549     // expression is a part of other expression).
2550     Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl);
2551     if (Style.Language == FormatStyle::LK_ObjC &&
2552         Current->is(TT_SelectorName) && Current->ParameterIndex > 0) {
2553       if (Current->ParameterIndex == 1)
2554         Current->SplitPenalty += 5 * Current->BindingStrength;
2555     } else {
2556       Current->SplitPenalty += 20 * Current->BindingStrength;
2557     }
2558 
2559     Current = Current->Next;
2560   }
2561 
2562   calculateUnbreakableTailLengths(Line);
2563   unsigned IndentLevel = Line.Level;
2564   for (Current = Line.First; Current != nullptr; Current = Current->Next) {
2565     if (Current->Role)
2566       Current->Role->precomputeFormattingInfos(Current);
2567     if (Current->MatchingParen &&
2568         Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
2569       assert(IndentLevel > 0);
2570       --IndentLevel;
2571     }
2572     Current->IndentLevel = IndentLevel;
2573     if (Current->opensBlockOrBlockTypeList(Style))
2574       ++IndentLevel;
2575   }
2576 
2577   LLVM_DEBUG({ printDebugInfo(Line); });
2578 }
2579 
2580 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
2581   unsigned UnbreakableTailLength = 0;
2582   FormatToken *Current = Line.Last;
2583   while (Current) {
2584     Current->UnbreakableTailLength = UnbreakableTailLength;
2585     if (Current->CanBreakBefore ||
2586         Current->isOneOf(tok::comment, tok::string_literal)) {
2587       UnbreakableTailLength = 0;
2588     } else {
2589       UnbreakableTailLength +=
2590           Current->ColumnWidth + Current->SpacesRequiredBefore;
2591     }
2592     Current = Current->Previous;
2593   }
2594 }
2595 
2596 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
2597                                       const FormatToken &Tok,
2598                                       bool InFunctionDecl) {
2599   const FormatToken &Left = *Tok.Previous;
2600   const FormatToken &Right = Tok;
2601 
2602   if (Left.is(tok::semi))
2603     return 0;
2604 
2605   if (Style.Language == FormatStyle::LK_Java) {
2606     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
2607       return 1;
2608     if (Right.is(Keywords.kw_implements))
2609       return 2;
2610     if (Left.is(tok::comma) && Left.NestingLevel == 0)
2611       return 3;
2612   } else if (Style.Language == FormatStyle::LK_JavaScript) {
2613     if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
2614       return 100;
2615     if (Left.is(TT_JsTypeColon))
2616       return 35;
2617     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2618         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2619       return 100;
2620     // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()".
2621     if (Left.opensScope() && Right.closesScope())
2622       return 200;
2623   }
2624 
2625   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2626     return 1;
2627   if (Right.is(tok::l_square)) {
2628     if (Style.Language == FormatStyle::LK_Proto)
2629       return 1;
2630     if (Left.is(tok::r_square))
2631       return 200;
2632     // Slightly prefer formatting local lambda definitions like functions.
2633     if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
2634       return 35;
2635     if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2636                        TT_ArrayInitializerLSquare,
2637                        TT_DesignatedInitializerLSquare, TT_AttributeSquare))
2638       return 500;
2639   }
2640 
2641   if (Left.is(tok::coloncolon) ||
2642       (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
2643     return 500;
2644   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2645       Right.is(tok::kw_operator)) {
2646     if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
2647       return 3;
2648     if (Left.is(TT_StartOfName))
2649       return 110;
2650     if (InFunctionDecl && Right.NestingLevel == 0)
2651       return Style.PenaltyReturnTypeOnItsOwnLine;
2652     return 200;
2653   }
2654   if (Right.is(TT_PointerOrReference))
2655     return 190;
2656   if (Right.is(TT_LambdaArrow))
2657     return 110;
2658   if (Left.is(tok::equal) && Right.is(tok::l_brace))
2659     return 160;
2660   if (Left.is(TT_CastRParen))
2661     return 100;
2662   if (Left.isOneOf(tok::kw_class, tok::kw_struct))
2663     return 5000;
2664   if (Left.is(tok::comment))
2665     return 1000;
2666 
2667   if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon,
2668                    TT_CtorInitializerColon))
2669     return 2;
2670 
2671   if (Right.isMemberAccess()) {
2672     // Breaking before the "./->" of a chained call/member access is reasonably
2673     // cheap, as formatting those with one call per line is generally
2674     // desirable. In particular, it should be cheaper to break before the call
2675     // than it is to break inside a call's parameters, which could lead to weird
2676     // "hanging" indents. The exception is the very last "./->" to support this
2677     // frequent pattern:
2678     //
2679     //   aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
2680     //       dddddddd);
2681     //
2682     // which might otherwise be blown up onto many lines. Here, clang-format
2683     // won't produce "hanging" indents anyway as there is no other trailing
2684     // call.
2685     //
2686     // Also apply higher penalty is not a call as that might lead to a wrapping
2687     // like:
2688     //
2689     //   aaaaaaa
2690     //       .aaaaaaaaa.bbbbbbbb(cccccccc);
2691     return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
2692                ? 150
2693                : 35;
2694   }
2695 
2696   if (Right.is(TT_TrailingAnnotation) &&
2697       (!Right.Next || Right.Next->isNot(tok::l_paren))) {
2698     // Moving trailing annotations to the next line is fine for ObjC method
2699     // declarations.
2700     if (Line.startsWith(TT_ObjCMethodSpecifier))
2701       return 10;
2702     // Generally, breaking before a trailing annotation is bad unless it is
2703     // function-like. It seems to be especially preferable to keep standard
2704     // annotations (i.e. "const", "final" and "override") on the same line.
2705     // Use a slightly higher penalty after ")" so that annotations like
2706     // "const override" are kept together.
2707     bool is_short_annotation = Right.TokenText.size() < 10;
2708     return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2709   }
2710 
2711   // In for-loops, prefer breaking at ',' and ';'.
2712   if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2713     return 4;
2714 
2715   // In Objective-C method expressions, prefer breaking before "param:" over
2716   // breaking after it.
2717   if (Right.is(TT_SelectorName))
2718     return 0;
2719   if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2720     return Line.MightBeFunctionDecl ? 50 : 500;
2721 
2722   // In Objective-C type declarations, avoid breaking after the category's
2723   // open paren (we'll prefer breaking after the protocol list's opening
2724   // angle bracket, if present).
2725   if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous &&
2726       Left.Previous->isOneOf(tok::identifier, tok::greater))
2727     return 500;
2728 
2729   if (Left.is(tok::l_paren) && InFunctionDecl &&
2730       Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2731     return 100;
2732   if (Left.is(tok::l_paren) && Left.Previous &&
2733       (Left.Previous->is(tok::kw_for) || Left.Previous->isIf()))
2734     return 1000;
2735   if (Left.is(tok::equal) && InFunctionDecl)
2736     return 110;
2737   if (Right.is(tok::r_brace))
2738     return 1;
2739   if (Left.is(TT_TemplateOpener))
2740     return 100;
2741   if (Left.opensScope()) {
2742     // If we aren't aligning after opening parens/braces we can always break
2743     // here unless the style does not want us to place all arguments on the
2744     // next line.
2745     if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign &&
2746         (Left.ParameterCount <= 1 || Style.AllowAllArgumentsOnNextLine))
2747       return 0;
2748     if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle)
2749       return 19;
2750     return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2751                                    : 19;
2752   }
2753   if (Left.is(TT_JavaAnnotation))
2754     return 50;
2755 
2756   if (Left.is(TT_UnaryOperator))
2757     return 60;
2758   if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2759       Left.Previous->isLabelString() &&
2760       (Left.NextOperator || Left.OperatorIndex != 0))
2761     return 50;
2762   if (Right.is(tok::plus) && Left.isLabelString() &&
2763       (Right.NextOperator || Right.OperatorIndex != 0))
2764     return 25;
2765   if (Left.is(tok::comma))
2766     return 1;
2767   if (Right.is(tok::lessless) && Left.isLabelString() &&
2768       (Right.NextOperator || Right.OperatorIndex != 1))
2769     return 25;
2770   if (Right.is(tok::lessless)) {
2771     // Breaking at a << is really cheap.
2772     if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2773       // Slightly prefer to break before the first one in log-like statements.
2774       return 2;
2775     return 1;
2776   }
2777   if (Left.ClosesTemplateDeclaration)
2778     return Style.PenaltyBreakTemplateDeclaration;
2779   if (Left.is(TT_ConditionalExpr))
2780     return prec::Conditional;
2781   prec::Level Level = Left.getPrecedence();
2782   if (Level == prec::Unknown)
2783     Level = Right.getPrecedence();
2784   if (Level == prec::Assignment)
2785     return Style.PenaltyBreakAssignment;
2786   if (Level != prec::Unknown)
2787     return Level;
2788 
2789   return 3;
2790 }
2791 
2792 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const {
2793   return Style.SpaceBeforeParens == FormatStyle::SBPO_Always ||
2794          (Style.SpaceBeforeParens == FormatStyle::SBPO_NonEmptyParentheses &&
2795           Right.ParameterCount > 0);
2796 }
2797 
2798 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2799                                           const FormatToken &Left,
2800                                           const FormatToken &Right) {
2801   if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2802     return true;
2803   if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java)
2804     return true;
2805   if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2806       Left.Tok.getObjCKeywordID() == tok::objc_property)
2807     return true;
2808   if (Right.is(tok::hashhash))
2809     return Left.is(tok::hash);
2810   if (Left.isOneOf(tok::hashhash, tok::hash))
2811     return Right.is(tok::hash);
2812   if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) ||
2813       (Left.is(tok::l_brace) && Left.isNot(BK_Block) &&
2814        Right.is(tok::r_brace) && Right.isNot(BK_Block)))
2815     return Style.SpaceInEmptyParentheses;
2816   if (Style.SpacesInConditionalStatement) {
2817     if (Left.is(tok::l_paren) && Left.Previous &&
2818         isKeywordWithCondition(*Left.Previous))
2819       return true;
2820     if (Right.is(tok::r_paren) && Right.MatchingParen &&
2821         Right.MatchingParen->Previous &&
2822         isKeywordWithCondition(*Right.MatchingParen->Previous))
2823       return true;
2824   }
2825 
2826   // requires ( or requires(
2827   if (Right.is(tok::l_paren) && Left.is(tok::kw_requires))
2828     return spaceRequiredBeforeParens(Right);
2829   // requires clause Concept1<T> && Concept2<T>
2830   if (Left.is(TT_ConstraintJunctions) && Right.is(tok::identifier))
2831     return true;
2832 
2833   if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2834     return (Right.is(TT_CastRParen) ||
2835             (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2836                ? Style.SpacesInCStyleCastParentheses
2837                : Style.SpacesInParentheses;
2838   if (Right.isOneOf(tok::semi, tok::comma))
2839     return false;
2840   if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) {
2841     bool IsLightweightGeneric = Right.MatchingParen &&
2842                                 Right.MatchingParen->Next &&
2843                                 Right.MatchingParen->Next->is(tok::colon);
2844     return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList;
2845   }
2846   if (Right.is(tok::less) && Left.is(tok::kw_template))
2847     return Style.SpaceAfterTemplateKeyword;
2848   if (Left.isOneOf(tok::exclaim, tok::tilde))
2849     return false;
2850   if (Left.is(tok::at) &&
2851       Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2852                     tok::numeric_constant, tok::l_paren, tok::l_brace,
2853                     tok::kw_true, tok::kw_false))
2854     return false;
2855   if (Left.is(tok::colon))
2856     return !Left.is(TT_ObjCMethodExpr);
2857   if (Left.is(tok::coloncolon))
2858     return false;
2859   if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) {
2860     if (Style.Language == FormatStyle::LK_TextProto ||
2861         (Style.Language == FormatStyle::LK_Proto &&
2862          (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) {
2863       // Format empty list as `<>`.
2864       if (Left.is(tok::less) && Right.is(tok::greater))
2865         return false;
2866       return !Style.Cpp11BracedListStyle;
2867     }
2868     return false;
2869   }
2870   if (Right.is(tok::ellipsis))
2871     return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2872                                     Left.Previous->is(tok::kw_case));
2873   if (Left.is(tok::l_square) && Right.is(tok::amp))
2874     return Style.SpacesInSquareBrackets;
2875   if (Right.is(TT_PointerOrReference)) {
2876     if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) {
2877       if (!Left.MatchingParen)
2878         return true;
2879       FormatToken *TokenBeforeMatchingParen =
2880           Left.MatchingParen->getPreviousNonComment();
2881       if (!TokenBeforeMatchingParen || !Left.is(TT_TypeDeclarationParen))
2882         return true;
2883     }
2884     // Add a space if the previous token is a pointer qualifer or the closing
2885     // parenthesis of __attribute__(()) expression and the style requires spaces
2886     // after pointer qualifiers.
2887     if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_After ||
2888          Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) &&
2889         (Left.is(TT_AttributeParen) || Left.canBePointerOrReferenceQualifier()))
2890       return true;
2891     return (Left.Tok.isLiteral() ||
2892             (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2893              (Style.PointerAlignment != FormatStyle::PAS_Left ||
2894               (Line.IsMultiVariableDeclStmt &&
2895                (Left.NestingLevel == 0 ||
2896                 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2897   }
2898   if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2899       (!Left.is(TT_PointerOrReference) ||
2900        (Style.PointerAlignment != FormatStyle::PAS_Right &&
2901         !Line.IsMultiVariableDeclStmt)))
2902     return true;
2903   if (Left.is(TT_PointerOrReference)) {
2904     // Add a space if the next token is a pointer qualifer and the style
2905     // requires spaces before pointer qualifiers.
2906     if ((Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Before ||
2907          Style.SpaceAroundPointerQualifiers == FormatStyle::SAPQ_Both) &&
2908         Right.canBePointerOrReferenceQualifier())
2909       return true;
2910     return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2911            (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2912             !Right.is(TT_StartOfName)) ||
2913            (Right.is(tok::l_brace) && Right.is(BK_Block)) ||
2914            (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2915                            tok::l_paren) &&
2916             (Style.PointerAlignment != FormatStyle::PAS_Right &&
2917              !Line.IsMultiVariableDeclStmt) &&
2918             Left.Previous &&
2919             !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon,
2920                                     tok::l_square));
2921   }
2922   // Ensure right pointer alignement with ellipsis e.g. int *...P
2923   if (Left.is(tok::ellipsis) && Left.Previous &&
2924       Left.Previous->isOneOf(tok::star, tok::amp, tok::ampamp))
2925     return Style.PointerAlignment != FormatStyle::PAS_Right;
2926 
2927   if (Right.is(tok::star) && Left.is(tok::l_paren))
2928     return false;
2929   if (Left.is(tok::star) && Right.isOneOf(tok::star, tok::amp, tok::ampamp))
2930     return false;
2931   if (Right.isOneOf(tok::star, tok::amp, tok::ampamp)) {
2932     const FormatToken *Previous = &Left;
2933     while (Previous && !Previous->is(tok::kw_operator)) {
2934       if (Previous->is(tok::identifier) || Previous->isSimpleTypeSpecifier()) {
2935         Previous = Previous->getPreviousNonComment();
2936         continue;
2937       }
2938       if (Previous->is(TT_TemplateCloser) && Previous->MatchingParen) {
2939         Previous = Previous->MatchingParen->getPreviousNonComment();
2940         continue;
2941       }
2942       if (Previous->is(tok::coloncolon)) {
2943         Previous = Previous->getPreviousNonComment();
2944         continue;
2945       }
2946       break;
2947     }
2948     // Space between the type and the * in:
2949     //   operator void*()
2950     //   operator char*()
2951     //   operator /*comment*/ const char*()
2952     //   operator volatile /*comment*/ char*()
2953     //   operator Foo*()
2954     //   operator C<T>*()
2955     //   operator std::Foo*()
2956     //   operator C<T>::D<U>*()
2957     // dependent on PointerAlignment style.
2958     if (Previous &&
2959         (Previous->endsSequence(tok::kw_operator) ||
2960          Previous->endsSequence(tok::kw_const, tok::kw_operator) ||
2961          Previous->endsSequence(tok::kw_volatile, tok::kw_operator)))
2962       return (Style.PointerAlignment != FormatStyle::PAS_Left);
2963   }
2964   const auto SpaceRequiredForArrayInitializerLSquare =
2965       [](const FormatToken &LSquareTok, const FormatStyle &Style) {
2966         return Style.SpacesInContainerLiterals ||
2967                ((Style.Language == FormatStyle::LK_Proto ||
2968                  Style.Language == FormatStyle::LK_TextProto) &&
2969                 !Style.Cpp11BracedListStyle &&
2970                 LSquareTok.endsSequence(tok::l_square, tok::colon,
2971                                         TT_SelectorName));
2972       };
2973   if (Left.is(tok::l_square))
2974     return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) &&
2975             SpaceRequiredForArrayInitializerLSquare(Left, Style)) ||
2976            (Left.isOneOf(TT_ArraySubscriptLSquare, TT_StructuredBindingLSquare,
2977                          TT_LambdaLSquare) &&
2978             Style.SpacesInSquareBrackets && Right.isNot(tok::r_square));
2979   if (Right.is(tok::r_square))
2980     return Right.MatchingParen &&
2981            ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) &&
2982              SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen,
2983                                                      Style)) ||
2984             (Style.SpacesInSquareBrackets &&
2985              Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare,
2986                                           TT_StructuredBindingLSquare,
2987                                           TT_LambdaLSquare)) ||
2988             Right.MatchingParen->is(TT_AttributeParen));
2989   if (Right.is(tok::l_square) &&
2990       !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2991                      TT_DesignatedInitializerLSquare,
2992                      TT_StructuredBindingLSquare, TT_AttributeSquare) &&
2993       !Left.isOneOf(tok::numeric_constant, TT_DictLiteral) &&
2994       !(!Left.is(tok::r_square) && Style.SpaceBeforeSquareBrackets &&
2995         Right.is(TT_ArraySubscriptLSquare)))
2996     return false;
2997   if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2998     return !Left.Children.empty(); // No spaces in "{}".
2999   if ((Left.is(tok::l_brace) && Left.isNot(BK_Block)) ||
3000       (Right.is(tok::r_brace) && Right.MatchingParen &&
3001        Right.MatchingParen->isNot(BK_Block)))
3002     return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true;
3003   if (Left.is(TT_BlockComment))
3004     // No whitespace in x(/*foo=*/1), except for JavaScript.
3005     return Style.Language == FormatStyle::LK_JavaScript ||
3006            !Left.TokenText.endswith("=*/");
3007 
3008   // Space between template and attribute.
3009   // e.g. template <typename T> [[nodiscard]] ...
3010   if (Left.is(TT_TemplateCloser) && Right.is(TT_AttributeSquare))
3011     return true;
3012   if (Right.is(tok::l_paren)) {
3013     if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) ||
3014         (Left.is(tok::r_square) && Left.is(TT_AttributeSquare)))
3015       return true;
3016     if (Style.SpaceBeforeParens ==
3017             FormatStyle::SBPO_ControlStatementsExceptForEachMacros &&
3018         Left.is(TT_ForEachMacro))
3019       return false;
3020     return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
3021            (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
3022             (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while,
3023                           tok::kw_switch, tok::kw_case, TT_ForEachMacro,
3024                           TT_ObjCForIn) ||
3025              Left.isIf(Line.Type != LT_PreprocessorDirective) ||
3026              (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
3027                            tok::kw_new, tok::kw_delete) &&
3028               (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
3029            (spaceRequiredBeforeParens(Right) &&
3030             (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
3031              Left.is(tok::r_paren) || Left.isSimpleTypeSpecifier() ||
3032              (Left.is(tok::r_square) && Left.MatchingParen &&
3033               Left.MatchingParen->is(TT_LambdaLSquare))) &&
3034             Line.Type != LT_PreprocessorDirective);
3035   }
3036   if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
3037     return false;
3038   if (Right.is(TT_UnaryOperator))
3039     return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
3040            (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
3041   if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
3042                     tok::r_paren) ||
3043        Left.isSimpleTypeSpecifier()) &&
3044       Right.is(tok::l_brace) && Right.getNextNonComment() &&
3045       Right.isNot(BK_Block))
3046     return false;
3047   if (Left.is(tok::period) || Right.is(tok::period))
3048     return false;
3049   if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
3050     return false;
3051   if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
3052       Left.MatchingParen->Previous &&
3053       (Left.MatchingParen->Previous->is(tok::period) ||
3054        Left.MatchingParen->Previous->is(tok::coloncolon)))
3055     // Java call to generic function with explicit type:
3056     // A.<B<C<...>>>DoSomething();
3057     // A::<B<C<...>>>DoSomething();  // With a Java 8 method reference.
3058     return false;
3059   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
3060     return false;
3061   if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at))
3062     // Objective-C dictionary literal -> no space after opening brace.
3063     return false;
3064   if (Right.is(tok::r_brace) && Right.MatchingParen &&
3065       Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at))
3066     // Objective-C dictionary literal -> no space before closing brace.
3067     return false;
3068   if (Right.getType() == TT_TrailingAnnotation &&
3069       Right.isOneOf(tok::amp, tok::ampamp) &&
3070       Left.isOneOf(tok::kw_const, tok::kw_volatile) &&
3071       (!Right.Next || Right.Next->is(tok::semi)))
3072     // Match const and volatile ref-qualifiers without any additional
3073     // qualifiers such as
3074     // void Fn() const &;
3075     return Style.PointerAlignment != FormatStyle::PAS_Left;
3076   return true;
3077 }
3078 
3079 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
3080                                          const FormatToken &Right) {
3081   const FormatToken &Left = *Right.Previous;
3082   if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
3083     return true; // Never ever merge two identifiers.
3084   if (Style.isCpp()) {
3085     if (Left.is(tok::kw_operator))
3086       return Right.is(tok::coloncolon);
3087     if (Right.is(tok::l_brace) && Right.is(BK_BracedInit) &&
3088         !Left.opensScope() && Style.SpaceBeforeCpp11BracedList)
3089       return true;
3090   } else if (Style.Language == FormatStyle::LK_Proto ||
3091              Style.Language == FormatStyle::LK_TextProto) {
3092     if (Right.is(tok::period) &&
3093         Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
3094                      Keywords.kw_repeated, Keywords.kw_extend))
3095       return true;
3096     if (Right.is(tok::l_paren) &&
3097         Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
3098       return true;
3099     if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName))
3100       return true;
3101     // Slashes occur in text protocol extension syntax: [type/type] { ... }.
3102     if (Left.is(tok::slash) || Right.is(tok::slash))
3103       return false;
3104     if (Left.MatchingParen &&
3105         Left.MatchingParen->is(TT_ProtoExtensionLSquare) &&
3106         Right.isOneOf(tok::l_brace, tok::less))
3107       return !Style.Cpp11BracedListStyle;
3108     // A percent is probably part of a formatting specification, such as %lld.
3109     if (Left.is(tok::percent))
3110       return false;
3111     // Preserve the existence of a space before a percent for cases like 0x%04x
3112     // and "%d %d"
3113     if (Left.is(tok::numeric_constant) && Right.is(tok::percent))
3114       return Right.WhitespaceRange.getEnd() != Right.WhitespaceRange.getBegin();
3115   } else if (Style.isCSharp()) {
3116     // Require spaces around '{' and  before '}' unless they appear in
3117     // interpolated strings. Interpolated strings are merged into a single token
3118     // so cannot have spaces inserted by this function.
3119 
3120     // No space between 'this' and '['
3121     if (Left.is(tok::kw_this) && Right.is(tok::l_square))
3122       return false;
3123 
3124     // No space between 'new' and '('
3125     if (Left.is(tok::kw_new) && Right.is(tok::l_paren))
3126       return false;
3127 
3128     // Space before { (including space within '{ {').
3129     if (Right.is(tok::l_brace))
3130       return true;
3131 
3132     // Spaces inside braces.
3133     if (Left.is(tok::l_brace) && Right.isNot(tok::r_brace))
3134       return true;
3135 
3136     if (Left.isNot(tok::l_brace) && Right.is(tok::r_brace))
3137       return true;
3138 
3139     // Spaces around '=>'.
3140     if (Left.is(TT_JsFatArrow) || Right.is(TT_JsFatArrow))
3141       return true;
3142 
3143     // No spaces around attribute target colons
3144     if (Left.is(TT_AttributeColon) || Right.is(TT_AttributeColon))
3145       return false;
3146 
3147     // space between type and variable e.g. Dictionary<string,string> foo;
3148     if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName))
3149       return true;
3150 
3151     // spaces inside square brackets.
3152     if (Left.is(tok::l_square) || Right.is(tok::r_square))
3153       return Style.SpacesInSquareBrackets;
3154 
3155     // No space before ? in nullable types.
3156     if (Right.is(TT_CSharpNullable))
3157       return false;
3158 
3159     // Require space after ? in nullable types except in generics and casts.
3160     if (Left.is(TT_CSharpNullable))
3161       return !Right.isOneOf(TT_TemplateCloser, tok::r_paren);
3162 
3163     // No space before or after '?.'.
3164     if (Left.is(TT_CSharpNullConditional) || Right.is(TT_CSharpNullConditional))
3165       return false;
3166 
3167     // Space before and after '??'.
3168     if (Left.is(TT_CSharpNullCoalescing) || Right.is(TT_CSharpNullCoalescing))
3169       return true;
3170 
3171     // No space before '?['.
3172     if (Right.is(TT_CSharpNullConditionalLSquare))
3173       return false;
3174 
3175     // No space between consecutive commas '[,,]'.
3176     if (Left.is(tok::comma) && Right.is(tok::comma))
3177       return false;
3178 
3179     // Possible space inside `?[ 0 ]`.
3180     if (Left.is(TT_CSharpNullConditionalLSquare))
3181       return Style.SpacesInSquareBrackets;
3182 
3183     // space after var in `var (key, value)`
3184     if (Left.is(Keywords.kw_var) && Right.is(tok::l_paren))
3185       return true;
3186 
3187     // space between keywords and paren e.g. "using ("
3188     if (Right.is(tok::l_paren))
3189       if (Left.isOneOf(tok::kw_using, Keywords.kw_async, Keywords.kw_when,
3190                        Keywords.kw_lock))
3191         return Style.SpaceBeforeParens == FormatStyle::SBPO_ControlStatements ||
3192                spaceRequiredBeforeParens(Right);
3193 
3194     // space between method modifier and opening parenthesis of a tuple return
3195     // type
3196     if (Left.isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
3197                      tok::kw_virtual, tok::kw_extern, tok::kw_static,
3198                      Keywords.kw_internal, Keywords.kw_abstract,
3199                      Keywords.kw_sealed, Keywords.kw_override,
3200                      Keywords.kw_async, Keywords.kw_unsafe) &&
3201         Right.is(tok::l_paren))
3202       return true;
3203   } else if (Style.Language == FormatStyle::LK_JavaScript) {
3204     if (Left.is(TT_JsFatArrow))
3205       return true;
3206     // for await ( ...
3207     if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous &&
3208         Left.Previous->is(tok::kw_for))
3209       return true;
3210     if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
3211         Right.MatchingParen) {
3212       const FormatToken *Next = Right.MatchingParen->getNextNonComment();
3213       // An async arrow function, for example: `x = async () => foo();`,
3214       // as opposed to calling a function called async: `x = async();`
3215       if (Next && Next->is(TT_JsFatArrow))
3216         return true;
3217     }
3218     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
3219         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
3220       return false;
3221     // In tagged template literals ("html`bar baz`"), there is no space between
3222     // the tag identifier and the template string.
3223     if (Keywords.IsJavaScriptIdentifier(Left,
3224                                         /* AcceptIdentifierName= */ false) &&
3225         Right.is(TT_TemplateString))
3226       return false;
3227     if (Right.is(tok::star) &&
3228         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
3229       return false;
3230     if (Right.isOneOf(tok::l_brace, tok::l_square) &&
3231         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield,
3232                      Keywords.kw_extends, Keywords.kw_implements))
3233       return true;
3234     if (Right.is(tok::l_paren)) {
3235       // JS methods can use some keywords as names (e.g. `delete()`).
3236       if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo())
3237         return false;
3238       // Valid JS method names can include keywords, e.g. `foo.delete()` or
3239       // `bar.instanceof()`. Recognize call positions by preceding period.
3240       if (Left.Previous && Left.Previous->is(tok::period) &&
3241           Left.Tok.getIdentifierInfo())
3242         return false;
3243       // Additional unary JavaScript operators that need a space after.
3244       if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof,
3245                        tok::kw_void))
3246         return true;
3247     }
3248     // `foo as const;` casts into a const type.
3249     if (Left.endsSequence(tok::kw_const, Keywords.kw_as)) {
3250       return false;
3251     }
3252     if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
3253                       tok::kw_const) ||
3254          // "of" is only a keyword if it appears after another identifier
3255          // (e.g. as "const x of y" in a for loop), or after a destructuring
3256          // operation (const [x, y] of z, const {a, b} of c).
3257          (Left.is(Keywords.kw_of) && Left.Previous &&
3258           (Left.Previous->Tok.is(tok::identifier) ||
3259            Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) &&
3260         (!Left.Previous || !Left.Previous->is(tok::period)))
3261       return true;
3262     if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
3263         Left.Previous->is(tok::period) && Right.is(tok::l_paren))
3264       return false;
3265     if (Left.is(Keywords.kw_as) &&
3266         Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
3267       return true;
3268     if (Left.is(tok::kw_default) && Left.Previous &&
3269         Left.Previous->is(tok::kw_export))
3270       return true;
3271     if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
3272       return true;
3273     if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
3274       return false;
3275     if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
3276       return false;
3277     if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
3278         Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
3279       return false;
3280     if (Left.is(tok::ellipsis))
3281       return false;
3282     if (Left.is(TT_TemplateCloser) &&
3283         !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
3284                        Keywords.kw_implements, Keywords.kw_extends))
3285       // Type assertions ('<type>expr') are not followed by whitespace. Other
3286       // locations that should have whitespace following are identified by the
3287       // above set of follower tokens.
3288       return false;
3289     if (Right.is(TT_JsNonNullAssertion))
3290       return false;
3291     if (Left.is(TT_JsNonNullAssertion) &&
3292         Right.isOneOf(Keywords.kw_as, Keywords.kw_in))
3293       return true; // "x! as string", "x! in y"
3294   } else if (Style.Language == FormatStyle::LK_Java) {
3295     if (Left.is(tok::r_square) && Right.is(tok::l_brace))
3296       return true;
3297     if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
3298       return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
3299     if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
3300                       tok::kw_protected) ||
3301          Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
3302                       Keywords.kw_native)) &&
3303         Right.is(TT_TemplateOpener))
3304       return true;
3305   }
3306   if (Left.is(TT_ImplicitStringLiteral))
3307     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
3308   if (Line.Type == LT_ObjCMethodDecl) {
3309     if (Left.is(TT_ObjCMethodSpecifier))
3310       return true;
3311     if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right))
3312       // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a
3313       // keyword in Objective-C, and '+ (instancetype)new;' is a standard class
3314       // method declaration.
3315       return false;
3316   }
3317   if (Line.Type == LT_ObjCProperty &&
3318       (Right.is(tok::equal) || Left.is(tok::equal)))
3319     return false;
3320 
3321   if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
3322       Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
3323     return true;
3324   if (Right.is(TT_OverloadedOperatorLParen))
3325     return spaceRequiredBeforeParens(Right);
3326   if (Left.is(tok::comma))
3327     return true;
3328   if (Right.is(tok::comma))
3329     return false;
3330   if (Right.is(TT_ObjCBlockLParen))
3331     return true;
3332   if (Right.is(TT_CtorInitializerColon))
3333     return Style.SpaceBeforeCtorInitializerColon;
3334   if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon)
3335     return false;
3336   if (Right.is(TT_RangeBasedForLoopColon) &&
3337       !Style.SpaceBeforeRangeBasedForLoopColon)
3338     return false;
3339   if (Left.is(TT_BitFieldColon))
3340     return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both ||
3341            Style.BitFieldColonSpacing == FormatStyle::BFCS_After;
3342   if (Right.is(tok::colon)) {
3343     if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
3344         !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
3345       return false;
3346     if (Right.is(TT_ObjCMethodExpr))
3347       return false;
3348     if (Left.is(tok::question))
3349       return false;
3350     if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
3351       return false;
3352     if (Right.is(TT_DictLiteral))
3353       return Style.SpacesInContainerLiterals;
3354     if (Right.is(TT_AttributeColon))
3355       return false;
3356     if (Right.is(TT_CSharpNamedArgumentColon))
3357       return false;
3358     if (Right.is(TT_BitFieldColon))
3359       return Style.BitFieldColonSpacing == FormatStyle::BFCS_Both ||
3360              Style.BitFieldColonSpacing == FormatStyle::BFCS_Before;
3361     return true;
3362   }
3363   if (Left.is(TT_UnaryOperator)) {
3364     if (!Right.is(tok::l_paren)) {
3365       // The alternative operators for ~ and ! are "compl" and "not".
3366       // If they are used instead, we do not want to combine them with
3367       // the token to the right, unless that is a left paren.
3368       if (Left.is(tok::exclaim) && Left.TokenText == "not")
3369         return true;
3370       if (Left.is(tok::tilde) && Left.TokenText == "compl")
3371         return true;
3372       // Lambda captures allow for a lone &, so "&]" needs to be properly
3373       // handled.
3374       if (Left.is(tok::amp) && Right.is(tok::r_square))
3375         return Style.SpacesInSquareBrackets;
3376     }
3377     return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) ||
3378            Right.is(TT_BinaryOperator);
3379   }
3380 
3381   // If the next token is a binary operator or a selector name, we have
3382   // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
3383   if (Left.is(TT_CastRParen))
3384     return Style.SpaceAfterCStyleCast ||
3385            Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
3386 
3387   if (Left.is(tok::greater) && Right.is(tok::greater)) {
3388     if (Style.Language == FormatStyle::LK_TextProto ||
3389         (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral)))
3390       return !Style.Cpp11BracedListStyle;
3391     return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
3392            (Style.Standard < FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
3393   }
3394   if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) ||
3395       Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
3396       (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod)))
3397     return false;
3398   if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) &&
3399       Right.getPrecedence() == prec::Assignment)
3400     return false;
3401   if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) &&
3402       (Left.is(tok::identifier) || Left.is(tok::kw_this)))
3403     return false;
3404   if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
3405     // Generally don't remove existing spaces between an identifier and "::".
3406     // The identifier might actually be a macro name such as ALWAYS_INLINE. If
3407     // this turns out to be too lenient, add analysis of the identifier itself.
3408     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
3409   if (Right.is(tok::coloncolon) &&
3410       !Left.isOneOf(tok::l_brace, tok::comment, tok::l_paren))
3411     // Put a space between < and :: in vector< ::std::string >
3412     return (Left.is(TT_TemplateOpener) &&
3413             (Style.Standard < FormatStyle::LS_Cpp11 || Style.SpacesInAngles)) ||
3414            !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
3415                           tok::kw___super, TT_TemplateOpener,
3416                           TT_TemplateCloser)) ||
3417            (Left.is(tok::l_paren) && Style.SpacesInParentheses);
3418   if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
3419     return Style.SpacesInAngles;
3420   // Space before TT_StructuredBindingLSquare.
3421   if (Right.is(TT_StructuredBindingLSquare))
3422     return !Left.isOneOf(tok::amp, tok::ampamp) ||
3423            Style.PointerAlignment != FormatStyle::PAS_Right;
3424   // Space before & or && following a TT_StructuredBindingLSquare.
3425   if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) &&
3426       Right.isOneOf(tok::amp, tok::ampamp))
3427     return Style.PointerAlignment != FormatStyle::PAS_Left;
3428   if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
3429       (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
3430        !Right.is(tok::r_paren)))
3431     return true;
3432   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
3433       Right.isNot(TT_FunctionTypeLParen))
3434     return spaceRequiredBeforeParens(Right);
3435   if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
3436       Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
3437     return false;
3438   if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
3439       Line.startsWith(tok::hash))
3440     return true;
3441   if (Right.is(TT_TrailingUnaryOperator))
3442     return false;
3443   if (Left.is(TT_RegexLiteral))
3444     return false;
3445   return spaceRequiredBetween(Line, Left, Right);
3446 }
3447 
3448 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
3449 static bool isAllmanBrace(const FormatToken &Tok) {
3450   return Tok.is(tok::l_brace) && Tok.is(BK_Block) &&
3451          !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral);
3452 }
3453 
3454 // Returns 'true' if 'Tok' is an function argument.
3455 static bool IsFunctionArgument(const FormatToken &Tok) {
3456   return Tok.MatchingParen && Tok.MatchingParen->Next &&
3457          Tok.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren);
3458 }
3459 
3460 static bool
3461 isItAnEmptyLambdaAllowed(const FormatToken &Tok,
3462                          FormatStyle::ShortLambdaStyle ShortLambdaOption) {
3463   return Tok.Children.empty() && ShortLambdaOption != FormatStyle::SLS_None;
3464 }
3465 
3466 static bool
3467 isItAInlineLambdaAllowed(const FormatToken &Tok,
3468                          FormatStyle::ShortLambdaStyle ShortLambdaOption) {
3469   return (ShortLambdaOption == FormatStyle::SLS_Inline &&
3470           IsFunctionArgument(Tok)) ||
3471          (ShortLambdaOption == FormatStyle::SLS_All);
3472 }
3473 
3474 static bool isOneChildWithoutMustBreakBefore(const FormatToken &Tok) {
3475   if (Tok.Children.size() != 1)
3476     return false;
3477   FormatToken *curElt = Tok.Children[0]->First;
3478   while (curElt) {
3479     if (curElt->MustBreakBefore)
3480       return false;
3481     curElt = curElt->Next;
3482   }
3483   return true;
3484 }
3485 static bool isAllmanLambdaBrace(const FormatToken &Tok) {
3486   return (Tok.is(tok::l_brace) && Tok.is(BK_Block) &&
3487           !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral));
3488 }
3489 
3490 static bool isAllmanBraceIncludedBreakableLambda(
3491     const FormatToken &Tok, FormatStyle::ShortLambdaStyle ShortLambdaOption) {
3492   if (!isAllmanLambdaBrace(Tok))
3493     return false;
3494 
3495   if (isItAnEmptyLambdaAllowed(Tok, ShortLambdaOption))
3496     return false;
3497 
3498   return !isItAInlineLambdaAllowed(Tok, ShortLambdaOption) ||
3499          !isOneChildWithoutMustBreakBefore(Tok);
3500 }
3501 
3502 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
3503                                      const FormatToken &Right) {
3504   const FormatToken &Left = *Right.Previous;
3505   if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
3506     return true;
3507 
3508   if (Style.isCSharp()) {
3509     if (Right.is(TT_CSharpNamedArgumentColon) ||
3510         Left.is(TT_CSharpNamedArgumentColon))
3511       return false;
3512     if (Right.is(TT_CSharpGenericTypeConstraint))
3513       return true;
3514   } else if (Style.Language == FormatStyle::LK_JavaScript) {
3515     // FIXME: This might apply to other languages and token kinds.
3516     if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
3517         Left.Previous->is(tok::string_literal))
3518       return true;
3519     if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
3520         Left.Previous && Left.Previous->is(tok::equal) &&
3521         Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
3522                             tok::kw_const) &&
3523         // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
3524         // above.
3525         !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
3526       // Object literals on the top level of a file are treated as "enum-style".
3527       // Each key/value pair is put on a separate line, instead of bin-packing.
3528       return true;
3529     if (Left.is(tok::l_brace) && Line.Level == 0 &&
3530         (Line.startsWith(tok::kw_enum) ||
3531          Line.startsWith(tok::kw_const, tok::kw_enum) ||
3532          Line.startsWith(tok::kw_export, tok::kw_enum) ||
3533          Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum)))
3534       // JavaScript top-level enum key/value pairs are put on separate lines
3535       // instead of bin-packing.
3536       return true;
3537     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) && Left.Previous &&
3538         Left.Previous->is(TT_JsFatArrow)) {
3539       // JS arrow function (=> {...}).
3540       switch (Style.AllowShortLambdasOnASingleLine) {
3541       case FormatStyle::SLS_All:
3542         return false;
3543       case FormatStyle::SLS_None:
3544         return true;
3545       case FormatStyle::SLS_Empty:
3546         return !Left.Children.empty();
3547       case FormatStyle::SLS_Inline:
3548         // allow one-lining inline (e.g. in function call args) and empty arrow
3549         // functions.
3550         return (Left.NestingLevel == 0 && Line.Level == 0) &&
3551                !Left.Children.empty();
3552       }
3553       llvm_unreachable("Unknown FormatStyle::ShortLambdaStyle enum");
3554     }
3555 
3556     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
3557         !Left.Children.empty())
3558       // Support AllowShortFunctionsOnASingleLine for JavaScript.
3559       return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
3560              Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
3561              (Left.NestingLevel == 0 && Line.Level == 0 &&
3562               Style.AllowShortFunctionsOnASingleLine &
3563                   FormatStyle::SFS_InlineOnly);
3564   } else if (Style.Language == FormatStyle::LK_Java) {
3565     if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
3566         Right.Next->is(tok::string_literal))
3567       return true;
3568   } else if (Style.Language == FormatStyle::LK_Cpp ||
3569              Style.Language == FormatStyle::LK_ObjC ||
3570              Style.Language == FormatStyle::LK_Proto ||
3571              Style.Language == FormatStyle::LK_TableGen ||
3572              Style.Language == FormatStyle::LK_TextProto) {
3573     if (Left.isStringLiteral() && Right.isStringLiteral())
3574       return true;
3575   }
3576 
3577   // If the last token before a '}', ']', or ')' is a comma or a trailing
3578   // comment, the intention is to insert a line break after it in order to make
3579   // shuffling around entries easier. Import statements, especially in
3580   // JavaScript, can be an exception to this rule.
3581   if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) {
3582     const FormatToken *BeforeClosingBrace = nullptr;
3583     if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
3584          (Style.Language == FormatStyle::LK_JavaScript &&
3585           Left.is(tok::l_paren))) &&
3586         Left.isNot(BK_Block) && Left.MatchingParen)
3587       BeforeClosingBrace = Left.MatchingParen->Previous;
3588     else if (Right.MatchingParen &&
3589              (Right.MatchingParen->isOneOf(tok::l_brace,
3590                                            TT_ArrayInitializerLSquare) ||
3591               (Style.Language == FormatStyle::LK_JavaScript &&
3592                Right.MatchingParen->is(tok::l_paren))))
3593       BeforeClosingBrace = &Left;
3594     if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
3595                                BeforeClosingBrace->isTrailingComment()))
3596       return true;
3597   }
3598 
3599   if (Right.is(tok::comment))
3600     return Left.isNot(BK_BracedInit) && Left.isNot(TT_CtorInitializerColon) &&
3601            (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
3602   if (Left.isTrailingComment())
3603     return true;
3604   if (Right.Previous->IsUnterminatedLiteral)
3605     return true;
3606   if (Right.is(tok::lessless) && Right.Next &&
3607       Right.Previous->is(tok::string_literal) &&
3608       Right.Next->is(tok::string_literal))
3609     return true;
3610   // Can break after template<> declaration
3611   if (Right.Previous->ClosesTemplateDeclaration &&
3612       Right.Previous->MatchingParen &&
3613       Right.Previous->MatchingParen->NestingLevel == 0) {
3614     // Put concepts on the next line e.g.
3615     // template<typename T>
3616     // concept ...
3617     if (Right.is(tok::kw_concept))
3618       return Style.BreakBeforeConceptDeclarations;
3619     return (Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes);
3620   }
3621   if (Right.is(TT_CtorInitializerComma) &&
3622       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
3623       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
3624     return true;
3625   if (Right.is(TT_CtorInitializerColon) &&
3626       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
3627       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
3628     return true;
3629   // Break only if we have multiple inheritance.
3630   if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma &&
3631       Right.is(TT_InheritanceComma))
3632     return true;
3633   if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
3634     // Multiline raw string literals are special wrt. line breaks. The author
3635     // has made a deliberate choice and might have aligned the contents of the
3636     // string literal accordingly. Thus, we try keep existing line breaks.
3637     return Right.IsMultiline && Right.NewlinesBefore > 0;
3638   if ((Right.Previous->is(tok::l_brace) ||
3639        (Right.Previous->is(tok::less) && Right.Previous->Previous &&
3640         Right.Previous->Previous->is(tok::equal))) &&
3641       Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) {
3642     // Don't put enums or option definitions onto single lines in protocol
3643     // buffers.
3644     return true;
3645   }
3646   if (Right.is(TT_InlineASMBrace))
3647     return Right.HasUnescapedNewline;
3648 
3649   auto ShortLambdaOption = Style.AllowShortLambdasOnASingleLine;
3650   if (Style.BraceWrapping.BeforeLambdaBody &&
3651       (isAllmanBraceIncludedBreakableLambda(Left, ShortLambdaOption) ||
3652        isAllmanBraceIncludedBreakableLambda(Right, ShortLambdaOption))) {
3653     return true;
3654   }
3655 
3656   if (isAllmanBrace(Left) || isAllmanBrace(Right))
3657     return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
3658            (Line.startsWith(tok::kw_typedef, tok::kw_enum) &&
3659             Style.BraceWrapping.AfterEnum) ||
3660            (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
3661            (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
3662   if (Left.is(TT_ObjCBlockLBrace) &&
3663       Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never)
3664     return true;
3665 
3666   if (Left.is(TT_LambdaLBrace)) {
3667     if (IsFunctionArgument(Left) &&
3668         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline)
3669       return false;
3670 
3671     if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None ||
3672         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline ||
3673         (!Left.Children.empty() &&
3674          Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty))
3675       return true;
3676   }
3677 
3678   // Put multiple Java annotation on a new line.
3679   if ((Style.Language == FormatStyle::LK_Java ||
3680        Style.Language == FormatStyle::LK_JavaScript) &&
3681       Left.is(TT_LeadingJavaAnnotation) &&
3682       Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
3683       (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
3684     return true;
3685 
3686   if (Right.is(TT_ProtoExtensionLSquare))
3687     return true;
3688 
3689   // In text proto instances if a submessage contains at least 2 entries and at
3690   // least one of them is a submessage, like A { ... B { ... } ... },
3691   // put all of the entries of A on separate lines by forcing the selector of
3692   // the submessage B to be put on a newline.
3693   //
3694   // Example: these can stay on one line:
3695   // a { scalar_1: 1 scalar_2: 2 }
3696   // a { b { key: value } }
3697   //
3698   // and these entries need to be on a new line even if putting them all in one
3699   // line is under the column limit:
3700   // a {
3701   //   scalar: 1
3702   //   b { key: value }
3703   // }
3704   //
3705   // We enforce this by breaking before a submessage field that has previous
3706   // siblings, *and* breaking before a field that follows a submessage field.
3707   //
3708   // Be careful to exclude the case  [proto.ext] { ... } since the `]` is
3709   // the TT_SelectorName there, but we don't want to break inside the brackets.
3710   //
3711   // Another edge case is @submessage { key: value }, which is a common
3712   // substitution placeholder. In this case we want to keep `@` and `submessage`
3713   // together.
3714   //
3715   // We ensure elsewhere that extensions are always on their own line.
3716   if ((Style.Language == FormatStyle::LK_Proto ||
3717        Style.Language == FormatStyle::LK_TextProto) &&
3718       Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) {
3719     // Keep `@submessage` together in:
3720     // @submessage { key: value }
3721     if (Right.Previous && Right.Previous->is(tok::at))
3722       return false;
3723     // Look for the scope opener after selector in cases like:
3724     // selector { ...
3725     // selector: { ...
3726     // selector: @base { ...
3727     FormatToken *LBrace = Right.Next;
3728     if (LBrace && LBrace->is(tok::colon)) {
3729       LBrace = LBrace->Next;
3730       if (LBrace && LBrace->is(tok::at)) {
3731         LBrace = LBrace->Next;
3732         if (LBrace)
3733           LBrace = LBrace->Next;
3734       }
3735     }
3736     if (LBrace &&
3737         // The scope opener is one of {, [, <:
3738         // selector { ... }
3739         // selector [ ... ]
3740         // selector < ... >
3741         //
3742         // In case of selector { ... }, the l_brace is TT_DictLiteral.
3743         // In case of an empty selector {}, the l_brace is not TT_DictLiteral,
3744         // so we check for immediately following r_brace.
3745         ((LBrace->is(tok::l_brace) &&
3746           (LBrace->is(TT_DictLiteral) ||
3747            (LBrace->Next && LBrace->Next->is(tok::r_brace)))) ||
3748          LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) {
3749       // If Left.ParameterCount is 0, then this submessage entry is not the
3750       // first in its parent submessage, and we want to break before this entry.
3751       // If Left.ParameterCount is greater than 0, then its parent submessage
3752       // might contain 1 or more entries and we want to break before this entry
3753       // if it contains at least 2 entries. We deal with this case later by
3754       // detecting and breaking before the next entry in the parent submessage.
3755       if (Left.ParameterCount == 0)
3756         return true;
3757       // However, if this submessage is the first entry in its parent
3758       // submessage, Left.ParameterCount might be 1 in some cases.
3759       // We deal with this case later by detecting an entry
3760       // following a closing paren of this submessage.
3761     }
3762 
3763     // If this is an entry immediately following a submessage, it will be
3764     // preceded by a closing paren of that submessage, like in:
3765     //     left---.  .---right
3766     //            v  v
3767     // sub: { ... } key: value
3768     // If there was a comment between `}` an `key` above, then `key` would be
3769     // put on a new line anyways.
3770     if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square))
3771       return true;
3772   }
3773 
3774   // Deal with lambda arguments in C++ - we want consistent line breaks whether
3775   // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced
3776   // as aggressive line breaks are placed when the lambda is not the last arg.
3777   if ((Style.Language == FormatStyle::LK_Cpp ||
3778        Style.Language == FormatStyle::LK_ObjC) &&
3779       Left.is(tok::l_paren) && Left.BlockParameterCount > 0 &&
3780       !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) {
3781     // Multiple lambdas in the same function call force line breaks.
3782     if (Left.BlockParameterCount > 1)
3783       return true;
3784 
3785     // A lambda followed by another arg forces a line break.
3786     if (!Left.Role)
3787       return false;
3788     auto Comma = Left.Role->lastComma();
3789     if (!Comma)
3790       return false;
3791     auto Next = Comma->getNextNonComment();
3792     if (!Next)
3793       return false;
3794     if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret))
3795       return true;
3796   }
3797 
3798   return false;
3799 }
3800 
3801 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
3802                                     const FormatToken &Right) {
3803   const FormatToken &Left = *Right.Previous;
3804   // Language-specific stuff.
3805   if (Style.isCSharp()) {
3806     if (Left.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon) ||
3807         Right.isOneOf(TT_CSharpNamedArgumentColon, TT_AttributeColon))
3808       return false;
3809     // Only break after commas for generic type constraints.
3810     if (Line.First->is(TT_CSharpGenericTypeConstraint))
3811       return Left.is(TT_CSharpGenericTypeConstraintComma);
3812   } else if (Style.Language == FormatStyle::LK_Java) {
3813     if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3814                      Keywords.kw_implements))
3815       return false;
3816     if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3817                       Keywords.kw_implements))
3818       return true;
3819   } else if (Style.Language == FormatStyle::LK_JavaScript) {
3820     const FormatToken *NonComment = Right.getPreviousNonComment();
3821     if (NonComment &&
3822         NonComment->isOneOf(
3823             tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break,
3824             tok::kw_throw, Keywords.kw_interface, Keywords.kw_type,
3825             tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected,
3826             Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get,
3827             Keywords.kw_set, Keywords.kw_async, Keywords.kw_await))
3828       return false; // Otherwise automatic semicolon insertion would trigger.
3829     if (Right.NestingLevel == 0 &&
3830         (Left.Tok.getIdentifierInfo() ||
3831          Left.isOneOf(tok::r_square, tok::r_paren)) &&
3832         Right.isOneOf(tok::l_square, tok::l_paren))
3833       return false; // Otherwise automatic semicolon insertion would trigger.
3834     if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
3835       return false;
3836     if (Left.is(TT_JsTypeColon))
3837       return true;
3838     // Don't wrap between ":" and "!" of a strict prop init ("field!: type;").
3839     if (Left.is(tok::exclaim) && Right.is(tok::colon))
3840       return false;
3841     // Look for is type annotations like:
3842     // function f(): a is B { ... }
3843     // Do not break before is in these cases.
3844     if (Right.is(Keywords.kw_is)) {
3845       const FormatToken *Next = Right.getNextNonComment();
3846       // If `is` is followed by a colon, it's likely that it's a dict key, so
3847       // ignore it for this check.
3848       // For example this is common in Polymer:
3849       // Polymer({
3850       //   is: 'name',
3851       //   ...
3852       // });
3853       if (!Next || !Next->is(tok::colon))
3854         return false;
3855     }
3856     if (Left.is(Keywords.kw_in))
3857       return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
3858     if (Right.is(Keywords.kw_in))
3859       return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
3860     if (Right.is(Keywords.kw_as))
3861       return false; // must not break before as in 'x as type' casts
3862     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) {
3863       // extends and infer can appear as keywords in conditional types:
3864       //   https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types
3865       // do not break before them, as the expressions are subject to ASI.
3866       return false;
3867     }
3868     if (Left.is(Keywords.kw_as))
3869       return true;
3870     if (Left.is(TT_JsNonNullAssertion))
3871       return true;
3872     if (Left.is(Keywords.kw_declare) &&
3873         Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
3874                       Keywords.kw_function, tok::kw_class, tok::kw_enum,
3875                       Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
3876                       Keywords.kw_let, tok::kw_const))
3877       // See grammar for 'declare' statements at:
3878       // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
3879       return false;
3880     if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
3881         Right.isOneOf(tok::identifier, tok::string_literal))
3882       return false; // must not break in "module foo { ...}"
3883     if (Right.is(TT_TemplateString) && Right.closesScope())
3884       return false;
3885     // Don't split tagged template literal so there is a break between the tag
3886     // identifier and template string.
3887     if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) {
3888       return false;
3889     }
3890     if (Left.is(TT_TemplateString) && Left.opensScope())
3891       return true;
3892   }
3893 
3894   if (Left.is(tok::at))
3895     return false;
3896   if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
3897     return false;
3898   if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
3899     return !Right.is(tok::l_paren);
3900   if (Right.is(TT_PointerOrReference))
3901     return Line.IsMultiVariableDeclStmt ||
3902            (Style.PointerAlignment == FormatStyle::PAS_Right &&
3903             (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
3904   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
3905       Right.is(tok::kw_operator))
3906     return true;
3907   if (Left.is(TT_PointerOrReference))
3908     return false;
3909   if (Right.isTrailingComment())
3910     // We rely on MustBreakBefore being set correctly here as we should not
3911     // change the "binding" behavior of a comment.
3912     // The first comment in a braced lists is always interpreted as belonging to
3913     // the first list element. Otherwise, it should be placed outside of the
3914     // list.
3915     return Left.is(BK_BracedInit) ||
3916            (Left.is(TT_CtorInitializerColon) &&
3917             Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon);
3918   if (Left.is(tok::question) && Right.is(tok::colon))
3919     return false;
3920   if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
3921     return Style.BreakBeforeTernaryOperators;
3922   if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
3923     return !Style.BreakBeforeTernaryOperators;
3924   if (Left.is(TT_InheritanceColon))
3925     return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon;
3926   if (Right.is(TT_InheritanceColon))
3927     return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon;
3928   if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
3929       Left.isNot(TT_SelectorName))
3930     return true;
3931 
3932   if (Right.is(tok::colon) &&
3933       !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
3934     return false;
3935   if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
3936     if (Style.Language == FormatStyle::LK_Proto ||
3937         Style.Language == FormatStyle::LK_TextProto) {
3938       if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral())
3939         return false;
3940       // Prevent cases like:
3941       //
3942       // submessage:
3943       //     { key: valueeeeeeeeeeee }
3944       //
3945       // when the snippet does not fit into one line.
3946       // Prefer:
3947       //
3948       // submessage: {
3949       //   key: valueeeeeeeeeeee
3950       // }
3951       //
3952       // instead, even if it is longer by one line.
3953       //
3954       // Note that this allows allows the "{" to go over the column limit
3955       // when the column limit is just between ":" and "{", but that does
3956       // not happen too often and alternative formattings in this case are
3957       // not much better.
3958       //
3959       // The code covers the cases:
3960       //
3961       // submessage: { ... }
3962       // submessage: < ... >
3963       // repeated: [ ... ]
3964       if (((Right.is(tok::l_brace) || Right.is(tok::less)) &&
3965            Right.is(TT_DictLiteral)) ||
3966           Right.is(TT_ArrayInitializerLSquare))
3967         return false;
3968     }
3969     return true;
3970   }
3971   if (Right.is(tok::r_square) && Right.MatchingParen &&
3972       Right.MatchingParen->is(TT_ProtoExtensionLSquare))
3973     return false;
3974   if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
3975                                     Right.Next->is(TT_ObjCMethodExpr)))
3976     return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
3977   if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
3978     return true;
3979   if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
3980     return true;
3981   if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
3982                     TT_OverloadedOperator))
3983     return false;
3984   if (Left.is(TT_RangeBasedForLoopColon))
3985     return true;
3986   if (Right.is(TT_RangeBasedForLoopColon))
3987     return false;
3988   if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
3989     return true;
3990   if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
3991       Left.is(tok::kw_operator))
3992     return false;
3993   if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
3994       Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
3995     return false;
3996   if (Left.is(tok::equal) && Right.is(tok::l_brace) &&
3997       !Style.Cpp11BracedListStyle)
3998     return false;
3999   if (Left.is(tok::l_paren) &&
4000       Left.isOneOf(TT_AttributeParen, TT_TypeDeclarationParen))
4001     return false;
4002   if (Left.is(tok::l_paren) && Left.Previous &&
4003       (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
4004     return false;
4005   if (Right.is(TT_ImplicitStringLiteral))
4006     return false;
4007 
4008   if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
4009     return false;
4010   if (Right.is(tok::r_square) && Right.MatchingParen &&
4011       Right.MatchingParen->is(TT_LambdaLSquare))
4012     return false;
4013 
4014   // We only break before r_brace if there was a corresponding break before
4015   // the l_brace, which is tracked by BreakBeforeClosingBrace.
4016   if (Right.is(tok::r_brace))
4017     return Right.MatchingParen && Right.MatchingParen->is(BK_Block);
4018 
4019   // Allow breaking after a trailing annotation, e.g. after a method
4020   // declaration.
4021   if (Left.is(TT_TrailingAnnotation))
4022     return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
4023                           tok::less, tok::coloncolon);
4024 
4025   if (Right.is(tok::kw___attribute) ||
4026       (Right.is(tok::l_square) && Right.is(TT_AttributeSquare)))
4027     return !Left.is(TT_AttributeSquare);
4028 
4029   if (Left.is(tok::identifier) && Right.is(tok::string_literal))
4030     return true;
4031 
4032   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
4033     return true;
4034 
4035   if (Left.is(TT_CtorInitializerColon))
4036     return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon;
4037   if (Right.is(TT_CtorInitializerColon))
4038     return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon;
4039   if (Left.is(TT_CtorInitializerComma) &&
4040       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
4041     return false;
4042   if (Right.is(TT_CtorInitializerComma) &&
4043       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
4044     return true;
4045   if (Left.is(TT_InheritanceComma) &&
4046       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
4047     return false;
4048   if (Right.is(TT_InheritanceComma) &&
4049       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
4050     return true;
4051   if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
4052       (Left.is(tok::less) && Right.is(tok::less)))
4053     return false;
4054   if (Right.is(TT_BinaryOperator) &&
4055       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
4056       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
4057        Right.getPrecedence() != prec::Assignment))
4058     return true;
4059   if (Left.is(TT_ArrayInitializerLSquare))
4060     return true;
4061   if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
4062     return true;
4063   if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
4064       !Left.isOneOf(tok::arrowstar, tok::lessless) &&
4065       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
4066       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
4067        Left.getPrecedence() == prec::Assignment))
4068     return true;
4069   if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) ||
4070       (Left.is(tok::r_square) && Right.is(TT_AttributeSquare)))
4071     return false;
4072 
4073   auto ShortLambdaOption = Style.AllowShortLambdasOnASingleLine;
4074   if (Style.BraceWrapping.BeforeLambdaBody) {
4075     if (isAllmanLambdaBrace(Left))
4076       return !isItAnEmptyLambdaAllowed(Left, ShortLambdaOption);
4077     if (isAllmanLambdaBrace(Right))
4078       return !isItAnEmptyLambdaAllowed(Right, ShortLambdaOption);
4079   }
4080 
4081   return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
4082                       tok::kw_class, tok::kw_struct, tok::comment) ||
4083          Right.isMemberAccess() ||
4084          Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
4085                        tok::colon, tok::l_square, tok::at) ||
4086          (Style.BraceWrapping.BeforeLambdaBody && Right.is(TT_LambdaLBrace)) ||
4087          (Left.is(tok::r_paren) &&
4088           Right.isOneOf(tok::identifier, tok::kw_const)) ||
4089          (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
4090          (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
4091 }
4092 
4093 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
4094   llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n";
4095   const FormatToken *Tok = Line.First;
4096   while (Tok) {
4097     llvm::errs() << " M=" << Tok->MustBreakBefore
4098                  << " C=" << Tok->CanBreakBefore
4099                  << " T=" << getTokenTypeName(Tok->getType())
4100                  << " S=" << Tok->SpacesRequiredBefore
4101                  << " F=" << Tok->Finalized << " B=" << Tok->BlockParameterCount
4102                  << " BK=" << Tok->getBlockKind() << " P=" << Tok->SplitPenalty
4103                  << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength
4104                  << " PPK=" << Tok->getPackingKind() << " FakeLParens=";
4105     for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
4106       llvm::errs() << Tok->FakeLParens[i] << "/";
4107     llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
4108     llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo();
4109     llvm::errs() << " Text='" << Tok->TokenText << "'\n";
4110     if (!Tok->Next)
4111       assert(Tok == Line.Last);
4112     Tok = Tok->Next;
4113   }
4114   llvm::errs() << "----\n";
4115 }
4116 
4117 } // namespace format
4118 } // namespace clang
4119