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 /// A parser that gathers additional information about tokens.
60 ///
61 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
62 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
63 /// into template parameter lists.
64 class AnnotatingParser {
65 public:
66   AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
67                    const AdditionalKeywords &Keywords)
68       : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
69         Keywords(Keywords) {
70     Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
71     resetTokenMetadata(CurrentToken);
72   }
73 
74 private:
75   bool parseAngle() {
76     if (!CurrentToken || !CurrentToken->Previous)
77       return false;
78     if (NonTemplateLess.count(CurrentToken->Previous))
79       return false;
80 
81     const FormatToken &Previous = *CurrentToken->Previous; // The '<'.
82     if (Previous.Previous) {
83       if (Previous.Previous->Tok.isLiteral())
84         return false;
85       if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
86           (!Previous.Previous->MatchingParen ||
87            !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen)))
88         return false;
89     }
90 
91     FormatToken *Left = CurrentToken->Previous;
92     Left->ParentBracket = Contexts.back().ContextKind;
93     ScopedContextCreator ContextCreator(*this, tok::less, 12);
94 
95     // If this angle is in the context of an expression, we need to be more
96     // hesitant to detect it as opening template parameters.
97     bool InExprContext = Contexts.back().IsExpression;
98 
99     Contexts.back().IsExpression = false;
100     // If there's a template keyword before the opening angle bracket, this is a
101     // template parameter, not an argument.
102     Contexts.back().InTemplateArgument =
103         Left->Previous && Left->Previous->Tok.isNot(tok::kw_template);
104 
105     if (Style.Language == FormatStyle::LK_Java &&
106         CurrentToken->is(tok::question))
107       next();
108 
109     while (CurrentToken) {
110       if (CurrentToken->is(tok::greater)) {
111         Left->MatchingParen = CurrentToken;
112         CurrentToken->MatchingParen = Left;
113         // In TT_Proto, we must distignuish between:
114         //   map<key, value>
115         //   msg < item: data >
116         //   msg: < item: data >
117         // In TT_TextProto, map<key, value> does not occur.
118         if (Style.Language == FormatStyle::LK_TextProto ||
119             (Style.Language == FormatStyle::LK_Proto && Left->Previous &&
120              Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral)))
121           CurrentToken->Type = TT_DictLiteral;
122         else
123           CurrentToken->Type = TT_TemplateCloser;
124         next();
125         return true;
126       }
127       if (CurrentToken->is(tok::question) &&
128           Style.Language == FormatStyle::LK_Java) {
129         next();
130         continue;
131       }
132       if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
133           (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext &&
134            Style.Language != FormatStyle::LK_Proto &&
135            Style.Language != FormatStyle::LK_TextProto))
136         return false;
137       // If a && or || is found and interpreted as a binary operator, this set
138       // of angles is likely part of something like "a < b && c > d". If the
139       // angles are inside an expression, the ||/&& might also be a binary
140       // operator that was misinterpreted because we are parsing template
141       // parameters.
142       // FIXME: This is getting out of hand, write a decent parser.
143       if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
144           CurrentToken->Previous->is(TT_BinaryOperator) &&
145           Contexts[Contexts.size() - 2].IsExpression &&
146           !Line.startsWith(tok::kw_template))
147         return false;
148       updateParameterCount(Left, CurrentToken);
149       if (Style.Language == FormatStyle::LK_Proto) {
150         if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) {
151           if (CurrentToken->is(tok::colon) ||
152               (CurrentToken->isOneOf(tok::l_brace, tok::less) &&
153                Previous->isNot(tok::colon)))
154             Previous->Type = TT_SelectorName;
155         }
156       }
157       if (!consumeToken())
158         return false;
159     }
160     return false;
161   }
162 
163   bool parseParens(bool LookForDecls = false) {
164     if (!CurrentToken)
165       return false;
166     FormatToken *Left = CurrentToken->Previous;
167     Left->ParentBracket = Contexts.back().ContextKind;
168     ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
169 
170     // FIXME: This is a bit of a hack. Do better.
171     Contexts.back().ColonIsForRangeExpr =
172         Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
173 
174     bool StartsObjCMethodExpr = false;
175     if (FormatToken *MaybeSel = Left->Previous) {
176       // @selector( starts a selector.
177       if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
178           MaybeSel->Previous->is(tok::at)) {
179         StartsObjCMethodExpr = true;
180       }
181     }
182 
183     if (Left->is(TT_OverloadedOperatorLParen)) {
184       Contexts.back().IsExpression = false;
185     } else if (Style.Language == FormatStyle::LK_JavaScript &&
186                (Line.startsWith(Keywords.kw_type, tok::identifier) ||
187                 Line.startsWith(tok::kw_export, Keywords.kw_type,
188                                 tok::identifier))) {
189       // type X = (...);
190       // export type X = (...);
191       Contexts.back().IsExpression = false;
192     } else if (Left->Previous &&
193                (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype,
194                                         tok::kw_while, tok::l_paren,
195                                         tok::comma) ||
196                 Left->Previous->isIf() ||
197                 Left->Previous->is(TT_BinaryOperator))) {
198       // static_assert, if and while usually contain expressions.
199       Contexts.back().IsExpression = true;
200     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
201                (Left->Previous->is(Keywords.kw_function) ||
202                 (Left->Previous->endsSequence(tok::identifier,
203                                               Keywords.kw_function)))) {
204       // function(...) or function f(...)
205       Contexts.back().IsExpression = false;
206     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
207                Left->Previous->is(TT_JsTypeColon)) {
208       // let x: (SomeType);
209       Contexts.back().IsExpression = false;
210     } else if (isLambdaParameterList(Left)) {
211       // This is a parameter list of a lambda expression.
212       Contexts.back().IsExpression = false;
213     } else if (Line.InPPDirective &&
214                (!Left->Previous || !Left->Previous->is(tok::identifier))) {
215       Contexts.back().IsExpression = true;
216     } else if (Contexts[Contexts.size() - 2].CaretFound) {
217       // This is the parameter list of an ObjC block.
218       Contexts.back().IsExpression = false;
219     } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) {
220       Left->Type = TT_AttributeParen;
221     } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) {
222       // The first argument to a foreach macro is a declaration.
223       Contexts.back().IsForEachMacro = true;
224       Contexts.back().IsExpression = false;
225     } else if (Left->Previous && Left->Previous->MatchingParen &&
226                Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
227       Contexts.back().IsExpression = false;
228     } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
229       bool IsForOrCatch =
230           Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch);
231       Contexts.back().IsExpression = !IsForOrCatch;
232     }
233 
234     if (StartsObjCMethodExpr) {
235       Contexts.back().ColonIsObjCMethodExpr = true;
236       Left->Type = TT_ObjCMethodExpr;
237     }
238 
239     // MightBeFunctionType and ProbablyFunctionType are used for
240     // function pointer and reference types as well as Objective-C
241     // block types:
242     //
243     // void (*FunctionPointer)(void);
244     // void (&FunctionReference)(void);
245     // void (^ObjCBlock)(void);
246     bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
247     bool ProbablyFunctionType =
248         CurrentToken->isOneOf(tok::star, tok::amp, tok::caret);
249     bool HasMultipleLines = false;
250     bool HasMultipleParametersOnALine = false;
251     bool MightBeObjCForRangeLoop =
252         Left->Previous && Left->Previous->is(tok::kw_for);
253     FormatToken *PossibleObjCForInToken = nullptr;
254     while (CurrentToken) {
255       // LookForDecls is set when "if (" has been seen. Check for
256       // 'identifier' '*' 'identifier' followed by not '=' -- this
257       // '*' has to be a binary operator but determineStarAmpUsage() will
258       // categorize it as an unary operator, so set the right type here.
259       if (LookForDecls && CurrentToken->Next) {
260         FormatToken *Prev = CurrentToken->getPreviousNonComment();
261         if (Prev) {
262           FormatToken *PrevPrev = Prev->getPreviousNonComment();
263           FormatToken *Next = CurrentToken->Next;
264           if (PrevPrev && PrevPrev->is(tok::identifier) &&
265               Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
266               CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
267             Prev->Type = TT_BinaryOperator;
268             LookForDecls = false;
269           }
270         }
271       }
272 
273       if (CurrentToken->Previous->is(TT_PointerOrReference) &&
274           CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
275                                                     tok::coloncolon))
276         ProbablyFunctionType = true;
277       if (CurrentToken->is(tok::comma))
278         MightBeFunctionType = false;
279       if (CurrentToken->Previous->is(TT_BinaryOperator))
280         Contexts.back().IsExpression = true;
281       if (CurrentToken->is(tok::r_paren)) {
282         if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next &&
283             (CurrentToken->Next->is(tok::l_paren) ||
284              (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration)))
285           Left->Type = Left->Next->is(tok::caret) ? TT_ObjCBlockLParen
286                                                   : TT_FunctionTypeLParen;
287         Left->MatchingParen = CurrentToken;
288         CurrentToken->MatchingParen = Left;
289 
290         if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
291             Left->Previous && Left->Previous->is(tok::l_paren)) {
292           // Detect the case where macros are used to generate lambdas or
293           // function bodies, e.g.:
294           //   auto my_lambda = MARCO((Type *type, int i) { .. body .. });
295           for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) {
296             if (Tok->is(TT_BinaryOperator) &&
297                 Tok->isOneOf(tok::star, tok::amp, tok::ampamp))
298               Tok->Type = TT_PointerOrReference;
299           }
300         }
301 
302         if (StartsObjCMethodExpr) {
303           CurrentToken->Type = TT_ObjCMethodExpr;
304           if (Contexts.back().FirstObjCSelectorName) {
305             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
306                 Contexts.back().LongestObjCSelectorName;
307           }
308         }
309 
310         if (Left->is(TT_AttributeParen))
311           CurrentToken->Type = TT_AttributeParen;
312         if (Left->Previous && Left->Previous->is(TT_JavaAnnotation))
313           CurrentToken->Type = TT_JavaAnnotation;
314         if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation))
315           CurrentToken->Type = TT_LeadingJavaAnnotation;
316         if (Left->Previous && Left->Previous->is(TT_AttributeSquare))
317           CurrentToken->Type = TT_AttributeSquare;
318 
319         if (!HasMultipleLines)
320           Left->PackingKind = PPK_Inconclusive;
321         else if (HasMultipleParametersOnALine)
322           Left->PackingKind = PPK_BinPacked;
323         else
324           Left->PackingKind = PPK_OnePerLine;
325 
326         next();
327         return true;
328       }
329       if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
330         return false;
331 
332       if (CurrentToken->is(tok::l_brace))
333         Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen
334       if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
335           !CurrentToken->Next->HasUnescapedNewline &&
336           !CurrentToken->Next->isTrailingComment())
337         HasMultipleParametersOnALine = true;
338       if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
339            CurrentToken->Previous->isSimpleTypeSpecifier()) &&
340           !CurrentToken->is(tok::l_brace))
341         Contexts.back().IsExpression = false;
342       if (CurrentToken->isOneOf(tok::semi, tok::colon)) {
343         MightBeObjCForRangeLoop = false;
344         if (PossibleObjCForInToken) {
345           PossibleObjCForInToken->Type = TT_Unknown;
346           PossibleObjCForInToken = nullptr;
347         }
348       }
349       if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) {
350         PossibleObjCForInToken = CurrentToken;
351         PossibleObjCForInToken->Type = TT_ObjCForIn;
352       }
353       // When we discover a 'new', we set CanBeExpression to 'false' in order to
354       // parse the type correctly. Reset that after a comma.
355       if (CurrentToken->is(tok::comma))
356         Contexts.back().CanBeExpression = true;
357 
358       FormatToken *Tok = CurrentToken;
359       if (!consumeToken())
360         return false;
361       updateParameterCount(Left, Tok);
362       if (CurrentToken && CurrentToken->HasUnescapedNewline)
363         HasMultipleLines = true;
364     }
365     return false;
366   }
367 
368   bool isCSharpAttributeSpecifier(const FormatToken &Tok) {
369     if (!Style.isCSharp())
370       return false;
371 
372     const FormatToken *AttrTok = Tok.Next;
373     if (!AttrTok)
374       return false;
375 
376     // Just an empty declaration e.g. string [].
377     if (AttrTok->is(tok::r_square))
378       return false;
379 
380     // Move along the tokens inbetween the '[' and ']' e.g. [STAThread].
381     while (AttrTok && AttrTok->isNot(tok::r_square)) {
382       AttrTok = AttrTok->Next;
383     }
384 
385     if (!AttrTok)
386       return false;
387 
388     // Move past the end of ']'.
389     AttrTok = AttrTok->Next;
390     if (!AttrTok)
391       return false;
392 
393     // Limit this to being an access modifier that follows.
394     if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
395                          tok::kw_class, tok::kw_static, tok::l_square,
396                          Keywords.kw_internal)) {
397       return true;
398     }
399 
400     // incase its a [XXX] retval func(....
401     if (AttrTok->Next &&
402         AttrTok->Next->startsSequence(tok::identifier, tok::l_paren))
403       return true;
404 
405     return false;
406   }
407 
408   bool isCpp11AttributeSpecifier(const FormatToken &Tok) {
409     if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square))
410       return false;
411     // The first square bracket is part of an ObjC array literal
412     if (Tok.Previous && Tok.Previous->is(tok::at)) {
413       return false;
414     }
415     const FormatToken *AttrTok = Tok.Next->Next;
416     if (!AttrTok)
417       return false;
418     // C++17 '[[using ns: foo, bar(baz, blech)]]'
419     // We assume nobody will name an ObjC variable 'using'.
420     if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon))
421       return true;
422     if (AttrTok->isNot(tok::identifier))
423       return false;
424     while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) {
425       // ObjC message send. We assume nobody will use : in a C++11 attribute
426       // specifier parameter, although this is technically valid:
427       // [[foo(:)]].
428       if (AttrTok->is(tok::colon) ||
429           AttrTok->startsSequence(tok::identifier, tok::identifier) ||
430           AttrTok->startsSequence(tok::r_paren, tok::identifier))
431         return false;
432       if (AttrTok->is(tok::ellipsis))
433         return true;
434       AttrTok = AttrTok->Next;
435     }
436     return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square);
437   }
438 
439   bool parseSquare() {
440     if (!CurrentToken)
441       return false;
442 
443     // A '[' could be an index subscript (after an identifier or after
444     // ')' or ']'), it could be the start of an Objective-C method
445     // expression, it could the start of an Objective-C array literal,
446     // or it could be a C++ attribute specifier [[foo::bar]].
447     FormatToken *Left = CurrentToken->Previous;
448     Left->ParentBracket = Contexts.back().ContextKind;
449     FormatToken *Parent = Left->getPreviousNonComment();
450 
451     // Cases where '>' is followed by '['.
452     // In C++, this can happen either in array of templates (foo<int>[10])
453     // or when array is a nested template type (unique_ptr<type1<type2>[]>).
454     bool CppArrayTemplates =
455         Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) &&
456         (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
457          Contexts.back().InTemplateArgument);
458 
459     bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) ||
460                                      Contexts.back().InCpp11AttributeSpecifier;
461 
462     // Treat C# Attributes [STAThread] much like C++ attributes [[...]].
463     bool IsCSharp11AttributeSpecifier =
464         isCSharpAttributeSpecifier(*Left) ||
465         Contexts.back().InCSharpAttributeSpecifier;
466 
467     bool InsideInlineASM = Line.startsWith(tok::kw_asm);
468     bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style);
469     bool StartsObjCMethodExpr =
470         !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates &&
471         Style.isCpp() && !IsCpp11AttributeSpecifier &&
472         Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) &&
473         !CurrentToken->isOneOf(tok::l_brace, tok::r_square) &&
474         (!Parent ||
475          Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
476                          tok::kw_return, tok::kw_throw) ||
477          Parent->isUnaryOperator() ||
478          // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
479          Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
480          (getBinOpPrecedence(Parent->Tok.getKind(), true, true) >
481           prec::Unknown));
482     bool ColonFound = false;
483 
484     unsigned BindingIncrease = 1;
485     if (IsCppStructuredBinding) {
486       Left->Type = TT_StructuredBindingLSquare;
487     } else if (Left->is(TT_Unknown)) {
488       if (StartsObjCMethodExpr) {
489         Left->Type = TT_ObjCMethodExpr;
490       } else if (IsCpp11AttributeSpecifier) {
491         Left->Type = TT_AttributeSquare;
492       } else if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
493                  Contexts.back().ContextKind == tok::l_brace &&
494                  Parent->isOneOf(tok::l_brace, tok::comma)) {
495         Left->Type = TT_JsComputedPropertyName;
496       } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace &&
497                  Parent && Parent->isOneOf(tok::l_brace, tok::comma)) {
498         Left->Type = TT_DesignatedInitializerLSquare;
499       } else if (IsCSharp11AttributeSpecifier) {
500         Left->Type = TT_AttributeSquare;
501       } else if (CurrentToken->is(tok::r_square) && Parent &&
502                  Parent->is(TT_TemplateCloser)) {
503         Left->Type = TT_ArraySubscriptLSquare;
504       } else if (Style.Language == FormatStyle::LK_Proto ||
505                  Style.Language == FormatStyle::LK_TextProto) {
506         // Square braces in LK_Proto can either be message field attributes:
507         //
508         // optional Aaa aaa = 1 [
509         //   (aaa) = aaa
510         // ];
511         //
512         // extensions 123 [
513         //   (aaa) = aaa
514         // ];
515         //
516         // or text proto extensions (in options):
517         //
518         // option (Aaa.options) = {
519         //   [type.type/type] {
520         //     key: value
521         //   }
522         // }
523         //
524         // or repeated fields (in options):
525         //
526         // option (Aaa.options) = {
527         //   keys: [ 1, 2, 3 ]
528         // }
529         //
530         // In the first and the third case we want to spread the contents inside
531         // the square braces; in the second we want to keep them inline.
532         Left->Type = TT_ArrayInitializerLSquare;
533         if (!Left->endsSequence(tok::l_square, tok::numeric_constant,
534                                 tok::equal) &&
535             !Left->endsSequence(tok::l_square, tok::numeric_constant,
536                                 tok::identifier) &&
537             !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) {
538           Left->Type = TT_ProtoExtensionLSquare;
539           BindingIncrease = 10;
540         }
541       } else if (!CppArrayTemplates && Parent &&
542                  Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
543                                  tok::comma, tok::l_paren, tok::l_square,
544                                  tok::question, tok::colon, tok::kw_return,
545                                  // Should only be relevant to JavaScript:
546                                  tok::kw_default)) {
547         Left->Type = TT_ArrayInitializerLSquare;
548       } else {
549         BindingIncrease = 10;
550         Left->Type = TT_ArraySubscriptLSquare;
551       }
552     }
553 
554     ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
555     Contexts.back().IsExpression = true;
556     if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
557         Parent->is(TT_JsTypeColon))
558       Contexts.back().IsExpression = false;
559 
560     Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
561     Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier;
562     Contexts.back().InCSharpAttributeSpecifier = IsCSharp11AttributeSpecifier;
563 
564     while (CurrentToken) {
565       if (CurrentToken->is(tok::r_square)) {
566         if (IsCpp11AttributeSpecifier)
567           CurrentToken->Type = TT_AttributeSquare;
568         if (IsCSharp11AttributeSpecifier)
569           CurrentToken->Type = TT_AttributeSquare;
570         else if (((CurrentToken->Next &&
571                    CurrentToken->Next->is(tok::l_paren)) ||
572                   (CurrentToken->Previous &&
573                    CurrentToken->Previous->Previous == Left)) &&
574                  Left->is(TT_ObjCMethodExpr)) {
575           // An ObjC method call is rarely followed by an open parenthesis. It
576           // also can't be composed of just one token, unless it's a macro that
577           // will be expanded to more tokens.
578           // FIXME: Do we incorrectly label ":" with this?
579           StartsObjCMethodExpr = false;
580           Left->Type = TT_Unknown;
581         }
582         if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
583           CurrentToken->Type = TT_ObjCMethodExpr;
584           // If we haven't seen a colon yet, make sure the last identifier
585           // before the r_square is tagged as a selector name component.
586           if (!ColonFound && CurrentToken->Previous &&
587               CurrentToken->Previous->is(TT_Unknown) &&
588               canBeObjCSelectorComponent(*CurrentToken->Previous))
589             CurrentToken->Previous->Type = TT_SelectorName;
590           // determineStarAmpUsage() thinks that '*' '[' is allocating an
591           // array of pointers, but if '[' starts a selector then '*' is a
592           // binary operator.
593           if (Parent && Parent->is(TT_PointerOrReference))
594             Parent->Type = TT_BinaryOperator;
595         }
596         // An arrow after an ObjC method expression is not a lambda arrow.
597         if (CurrentToken->Type == TT_ObjCMethodExpr && CurrentToken->Next &&
598             CurrentToken->Next->is(TT_LambdaArrow))
599           CurrentToken->Next->Type = TT_Unknown;
600         Left->MatchingParen = CurrentToken;
601         CurrentToken->MatchingParen = Left;
602         // FirstObjCSelectorName is set when a colon is found. This does
603         // not work, however, when the method has no parameters.
604         // Here, we set FirstObjCSelectorName when the end of the method call is
605         // reached, in case it was not set already.
606         if (!Contexts.back().FirstObjCSelectorName) {
607           FormatToken *Previous = CurrentToken->getPreviousNonComment();
608           if (Previous && Previous->is(TT_SelectorName)) {
609             Previous->ObjCSelectorNameParts = 1;
610             Contexts.back().FirstObjCSelectorName = Previous;
611           }
612         } else {
613           Left->ParameterCount =
614               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
615         }
616         if (Contexts.back().FirstObjCSelectorName) {
617           Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
618               Contexts.back().LongestObjCSelectorName;
619           if (Left->BlockParameterCount > 1)
620             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
621         }
622         next();
623         return true;
624       }
625       if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
626         return false;
627       if (CurrentToken->is(tok::colon)) {
628         if (IsCpp11AttributeSpecifier &&
629             CurrentToken->endsSequence(tok::colon, tok::identifier,
630                                        tok::kw_using)) {
631           // Remember that this is a [[using ns: foo]] C++ attribute, so we
632           // don't add a space before the colon (unlike other colons).
633           CurrentToken->Type = TT_AttributeColon;
634         } else if (Left->isOneOf(TT_ArraySubscriptLSquare,
635                                  TT_DesignatedInitializerLSquare)) {
636           Left->Type = TT_ObjCMethodExpr;
637           StartsObjCMethodExpr = true;
638           Contexts.back().ColonIsObjCMethodExpr = true;
639           if (Parent && Parent->is(tok::r_paren))
640             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
641             Parent->Type = TT_CastRParen;
642         }
643         ColonFound = true;
644       }
645       if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
646           !ColonFound)
647         Left->Type = TT_ArrayInitializerLSquare;
648       FormatToken *Tok = CurrentToken;
649       if (!consumeToken())
650         return false;
651       updateParameterCount(Left, Tok);
652     }
653     return false;
654   }
655 
656   bool parseBrace() {
657     if (CurrentToken) {
658       FormatToken *Left = CurrentToken->Previous;
659       Left->ParentBracket = Contexts.back().ContextKind;
660 
661       if (Contexts.back().CaretFound)
662         Left->Type = TT_ObjCBlockLBrace;
663       Contexts.back().CaretFound = false;
664 
665       ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
666       Contexts.back().ColonIsDictLiteral = true;
667       if (Left->BlockKind == BK_BracedInit)
668         Contexts.back().IsExpression = true;
669       if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
670           Left->Previous->is(TT_JsTypeColon))
671         Contexts.back().IsExpression = false;
672 
673       while (CurrentToken) {
674         if (CurrentToken->is(tok::r_brace)) {
675           Left->MatchingParen = CurrentToken;
676           CurrentToken->MatchingParen = Left;
677           next();
678           return true;
679         }
680         if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
681           return false;
682         updateParameterCount(Left, CurrentToken);
683         if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) {
684           FormatToken *Previous = CurrentToken->getPreviousNonComment();
685           if (Previous->is(TT_JsTypeOptionalQuestion))
686             Previous = Previous->getPreviousNonComment();
687           if ((CurrentToken->is(tok::colon) &&
688                (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) ||
689               Style.Language == FormatStyle::LK_Proto ||
690               Style.Language == FormatStyle::LK_TextProto) {
691             Left->Type = TT_DictLiteral;
692             if (Previous->Tok.getIdentifierInfo() ||
693                 Previous->is(tok::string_literal))
694               Previous->Type = TT_SelectorName;
695           }
696           if (CurrentToken->is(tok::colon) ||
697               Style.Language == FormatStyle::LK_JavaScript)
698             Left->Type = TT_DictLiteral;
699         }
700         if (CurrentToken->is(tok::comma) &&
701             Style.Language == FormatStyle::LK_JavaScript)
702           Left->Type = TT_DictLiteral;
703         if (!consumeToken())
704           return false;
705       }
706     }
707     return true;
708   }
709 
710   void updateParameterCount(FormatToken *Left, FormatToken *Current) {
711     // For ObjC methods, the number of parameters is calculated differently as
712     // method declarations have a different structure (the parameters are not
713     // inside a bracket scope).
714     if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block)
715       ++Left->BlockParameterCount;
716     if (Current->is(tok::comma)) {
717       ++Left->ParameterCount;
718       if (!Left->Role)
719         Left->Role.reset(new CommaSeparatedList(Style));
720       Left->Role->CommaFound(Current);
721     } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
722       Left->ParameterCount = 1;
723     }
724   }
725 
726   bool parseConditional() {
727     while (CurrentToken) {
728       if (CurrentToken->is(tok::colon)) {
729         CurrentToken->Type = TT_ConditionalExpr;
730         next();
731         return true;
732       }
733       if (!consumeToken())
734         return false;
735     }
736     return false;
737   }
738 
739   bool parseTemplateDeclaration() {
740     if (CurrentToken && CurrentToken->is(tok::less)) {
741       CurrentToken->Type = TT_TemplateOpener;
742       next();
743       if (!parseAngle())
744         return false;
745       if (CurrentToken)
746         CurrentToken->Previous->ClosesTemplateDeclaration = true;
747       return true;
748     }
749     return false;
750   }
751 
752   bool consumeToken() {
753     FormatToken *Tok = CurrentToken;
754     next();
755     switch (Tok->Tok.getKind()) {
756     case tok::plus:
757     case tok::minus:
758       if (!Tok->Previous && Line.MustBeDeclaration)
759         Tok->Type = TT_ObjCMethodSpecifier;
760       break;
761     case tok::colon:
762       if (!Tok->Previous)
763         return false;
764       // Colons from ?: are handled in parseConditional().
765       if (Style.Language == FormatStyle::LK_JavaScript) {
766         if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
767             (Contexts.size() == 1 &&               // switch/case labels
768              !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
769             Contexts.back().ContextKind == tok::l_paren ||  // function params
770             Contexts.back().ContextKind == tok::l_square || // array type
771             (!Contexts.back().IsExpression &&
772              Contexts.back().ContextKind == tok::l_brace) || // object type
773             (Contexts.size() == 1 &&
774              Line.MustBeDeclaration)) { // method/property declaration
775           Contexts.back().IsExpression = false;
776           Tok->Type = TT_JsTypeColon;
777           break;
778         }
779       }
780       if (Contexts.back().ColonIsDictLiteral ||
781           Style.Language == FormatStyle::LK_Proto ||
782           Style.Language == FormatStyle::LK_TextProto) {
783         Tok->Type = TT_DictLiteral;
784         if (Style.Language == FormatStyle::LK_TextProto) {
785           if (FormatToken *Previous = Tok->getPreviousNonComment())
786             Previous->Type = TT_SelectorName;
787         }
788       } else if (Contexts.back().ColonIsObjCMethodExpr ||
789                  Line.startsWith(TT_ObjCMethodSpecifier)) {
790         Tok->Type = TT_ObjCMethodExpr;
791         const FormatToken *BeforePrevious = Tok->Previous->Previous;
792         // Ensure we tag all identifiers in method declarations as
793         // TT_SelectorName.
794         bool UnknownIdentifierInMethodDeclaration =
795             Line.startsWith(TT_ObjCMethodSpecifier) &&
796             Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown);
797         if (!BeforePrevious ||
798             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
799             !(BeforePrevious->is(TT_CastRParen) ||
800               (BeforePrevious->is(TT_ObjCMethodExpr) &&
801                BeforePrevious->is(tok::colon))) ||
802             BeforePrevious->is(tok::r_square) ||
803             Contexts.back().LongestObjCSelectorName == 0 ||
804             UnknownIdentifierInMethodDeclaration) {
805           Tok->Previous->Type = TT_SelectorName;
806           if (!Contexts.back().FirstObjCSelectorName)
807             Contexts.back().FirstObjCSelectorName = Tok->Previous;
808           else if (Tok->Previous->ColumnWidth >
809                    Contexts.back().LongestObjCSelectorName)
810             Contexts.back().LongestObjCSelectorName =
811                 Tok->Previous->ColumnWidth;
812           Tok->Previous->ParameterIndex =
813               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
814           ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
815         }
816       } else if (Contexts.back().ColonIsForRangeExpr) {
817         Tok->Type = TT_RangeBasedForLoopColon;
818       } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
819         Tok->Type = TT_BitFieldColon;
820       } else if (Contexts.size() == 1 &&
821                  !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) {
822         if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren,
823                                                   tok::kw_noexcept))
824           Tok->Type = TT_CtorInitializerColon;
825         else
826           Tok->Type = TT_InheritanceColon;
827       } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next &&
828                  (Tok->Next->isOneOf(tok::r_paren, tok::comma) ||
829                   (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next &&
830                    Tok->Next->Next->is(tok::colon)))) {
831         // This handles a special macro in ObjC code where selectors including
832         // the colon are passed as macro arguments.
833         Tok->Type = TT_ObjCMethodExpr;
834       } else if (Contexts.back().ContextKind == tok::l_paren) {
835         Tok->Type = TT_InlineASMColon;
836       }
837       break;
838     case tok::pipe:
839     case tok::amp:
840       // | and & in declarations/type expressions represent union and
841       // intersection types, respectively.
842       if (Style.Language == FormatStyle::LK_JavaScript &&
843           !Contexts.back().IsExpression)
844         Tok->Type = TT_JsTypeOperator;
845       break;
846     case tok::kw_if:
847     case tok::kw_while:
848       if (Tok->is(tok::kw_if) && CurrentToken &&
849           CurrentToken->isOneOf(tok::kw_constexpr, tok::identifier))
850         next();
851       if (CurrentToken && CurrentToken->is(tok::l_paren)) {
852         next();
853         if (!parseParens(/*LookForDecls=*/true))
854           return false;
855       }
856       break;
857     case tok::kw_for:
858       if (Style.Language == FormatStyle::LK_JavaScript) {
859         // x.for and {for: ...}
860         if ((Tok->Previous && Tok->Previous->is(tok::period)) ||
861             (Tok->Next && Tok->Next->is(tok::colon)))
862           break;
863         // JS' for await ( ...
864         if (CurrentToken && CurrentToken->is(Keywords.kw_await))
865           next();
866       }
867       Contexts.back().ColonIsForRangeExpr = true;
868       next();
869       if (!parseParens())
870         return false;
871       break;
872     case tok::l_paren:
873       // When faced with 'operator()()', the kw_operator handler incorrectly
874       // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
875       // the first two parens OverloadedOperators and the second l_paren an
876       // OverloadedOperatorLParen.
877       if (Tok->Previous && Tok->Previous->is(tok::r_paren) &&
878           Tok->Previous->MatchingParen &&
879           Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
880         Tok->Previous->Type = TT_OverloadedOperator;
881         Tok->Previous->MatchingParen->Type = TT_OverloadedOperator;
882         Tok->Type = TT_OverloadedOperatorLParen;
883       }
884 
885       if (!parseParens())
886         return false;
887       if (Line.MustBeDeclaration && Contexts.size() == 1 &&
888           !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
889           (!Tok->Previous ||
890            !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute,
891                                    TT_LeadingJavaAnnotation)))
892         Line.MightBeFunctionDecl = true;
893       break;
894     case tok::l_square:
895       if (!parseSquare())
896         return false;
897       break;
898     case tok::l_brace:
899       if (Style.Language == FormatStyle::LK_TextProto) {
900         FormatToken *Previous = Tok->getPreviousNonComment();
901         if (Previous && Previous->Type != TT_DictLiteral)
902           Previous->Type = TT_SelectorName;
903       }
904       if (!parseBrace())
905         return false;
906       break;
907     case tok::less:
908       if (parseAngle()) {
909         Tok->Type = TT_TemplateOpener;
910         // In TT_Proto, we must distignuish between:
911         //   map<key, value>
912         //   msg < item: data >
913         //   msg: < item: data >
914         // In TT_TextProto, map<key, value> does not occur.
915         if (Style.Language == FormatStyle::LK_TextProto ||
916             (Style.Language == FormatStyle::LK_Proto && Tok->Previous &&
917              Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
918           Tok->Type = TT_DictLiteral;
919           FormatToken *Previous = Tok->getPreviousNonComment();
920           if (Previous && Previous->Type != TT_DictLiteral)
921             Previous->Type = TT_SelectorName;
922         }
923       } else {
924         Tok->Type = TT_BinaryOperator;
925         NonTemplateLess.insert(Tok);
926         CurrentToken = Tok;
927         next();
928       }
929       break;
930     case tok::r_paren:
931     case tok::r_square:
932       return false;
933     case tok::r_brace:
934       // Lines can start with '}'.
935       if (Tok->Previous)
936         return false;
937       break;
938     case tok::greater:
939       if (Style.Language != FormatStyle::LK_TextProto)
940         Tok->Type = TT_BinaryOperator;
941       if (Tok->Previous && Tok->Previous->is(TT_TemplateCloser))
942         Tok->SpacesRequiredBefore = 1;
943       break;
944     case tok::kw_operator:
945       if (Style.Language == FormatStyle::LK_TextProto ||
946           Style.Language == FormatStyle::LK_Proto)
947         break;
948       while (CurrentToken &&
949              !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
950         if (CurrentToken->isOneOf(tok::star, tok::amp))
951           CurrentToken->Type = TT_PointerOrReference;
952         consumeToken();
953         if (CurrentToken &&
954             CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator,
955                                             tok::comma))
956           CurrentToken->Previous->Type = TT_OverloadedOperator;
957       }
958       if (CurrentToken) {
959         CurrentToken->Type = TT_OverloadedOperatorLParen;
960         if (CurrentToken->Previous->is(TT_BinaryOperator))
961           CurrentToken->Previous->Type = TT_OverloadedOperator;
962       }
963       break;
964     case tok::question:
965       if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next &&
966           Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
967                              tok::r_brace)) {
968         // Question marks before semicolons, colons, etc. indicate optional
969         // types (fields, parameters), e.g.
970         //   function(x?: string, y?) {...}
971         //   class X { y?; }
972         Tok->Type = TT_JsTypeOptionalQuestion;
973         break;
974       }
975       // Declarations cannot be conditional expressions, this can only be part
976       // of a type declaration.
977       if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
978           Style.Language == FormatStyle::LK_JavaScript)
979         break;
980       parseConditional();
981       break;
982     case tok::kw_template:
983       parseTemplateDeclaration();
984       break;
985     case tok::comma:
986       if (Contexts.back().InCtorInitializer)
987         Tok->Type = TT_CtorInitializerComma;
988       else if (Contexts.back().InInheritanceList)
989         Tok->Type = TT_InheritanceComma;
990       else if (Contexts.back().FirstStartOfName &&
991                (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
992         Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
993         Line.IsMultiVariableDeclStmt = true;
994       }
995       if (Contexts.back().IsForEachMacro)
996         Contexts.back().IsExpression = true;
997       break;
998     case tok::identifier:
999       if (Tok->isOneOf(Keywords.kw___has_include,
1000                        Keywords.kw___has_include_next)) {
1001         parseHasInclude();
1002       }
1003       break;
1004     default:
1005       break;
1006     }
1007     return true;
1008   }
1009 
1010   void parseIncludeDirective() {
1011     if (CurrentToken && CurrentToken->is(tok::less)) {
1012       next();
1013       while (CurrentToken) {
1014         // Mark tokens up to the trailing line comments as implicit string
1015         // literals.
1016         if (CurrentToken->isNot(tok::comment) &&
1017             !CurrentToken->TokenText.startswith("//"))
1018           CurrentToken->Type = TT_ImplicitStringLiteral;
1019         next();
1020       }
1021     }
1022   }
1023 
1024   void parseWarningOrError() {
1025     next();
1026     // We still want to format the whitespace left of the first token of the
1027     // warning or error.
1028     next();
1029     while (CurrentToken) {
1030       CurrentToken->Type = TT_ImplicitStringLiteral;
1031       next();
1032     }
1033   }
1034 
1035   void parsePragma() {
1036     next(); // Consume "pragma".
1037     if (CurrentToken &&
1038         CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
1039       bool IsMark = CurrentToken->is(Keywords.kw_mark);
1040       next(); // Consume "mark".
1041       next(); // Consume first token (so we fix leading whitespace).
1042       while (CurrentToken) {
1043         if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
1044           CurrentToken->Type = TT_ImplicitStringLiteral;
1045         next();
1046       }
1047     }
1048   }
1049 
1050   void parseHasInclude() {
1051     if (!CurrentToken || !CurrentToken->is(tok::l_paren))
1052       return;
1053     next(); // '('
1054     parseIncludeDirective();
1055     next(); // ')'
1056   }
1057 
1058   LineType parsePreprocessorDirective() {
1059     bool IsFirstToken = CurrentToken->IsFirst;
1060     LineType Type = LT_PreprocessorDirective;
1061     next();
1062     if (!CurrentToken)
1063       return Type;
1064 
1065     if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
1066       // JavaScript files can contain shebang lines of the form:
1067       // #!/usr/bin/env node
1068       // Treat these like C++ #include directives.
1069       while (CurrentToken) {
1070         // Tokens cannot be comments here.
1071         CurrentToken->Type = TT_ImplicitStringLiteral;
1072         next();
1073       }
1074       return LT_ImportStatement;
1075     }
1076 
1077     if (CurrentToken->Tok.is(tok::numeric_constant)) {
1078       CurrentToken->SpacesRequiredBefore = 1;
1079       return Type;
1080     }
1081     // Hashes in the middle of a line can lead to any strange token
1082     // sequence.
1083     if (!CurrentToken->Tok.getIdentifierInfo())
1084       return Type;
1085     switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
1086     case tok::pp_include:
1087     case tok::pp_include_next:
1088     case tok::pp_import:
1089       next();
1090       parseIncludeDirective();
1091       Type = LT_ImportStatement;
1092       break;
1093     case tok::pp_error:
1094     case tok::pp_warning:
1095       parseWarningOrError();
1096       break;
1097     case tok::pp_pragma:
1098       parsePragma();
1099       break;
1100     case tok::pp_if:
1101     case tok::pp_elif:
1102       Contexts.back().IsExpression = true;
1103       next();
1104       parseLine();
1105       break;
1106     default:
1107       break;
1108     }
1109     while (CurrentToken) {
1110       FormatToken *Tok = CurrentToken;
1111       next();
1112       if (Tok->is(tok::l_paren))
1113         parseParens();
1114       else if (Tok->isOneOf(Keywords.kw___has_include,
1115                             Keywords.kw___has_include_next))
1116         parseHasInclude();
1117     }
1118     return Type;
1119   }
1120 
1121 public:
1122   LineType parseLine() {
1123     if (!CurrentToken)
1124       return LT_Invalid;
1125     NonTemplateLess.clear();
1126     if (CurrentToken->is(tok::hash))
1127       return parsePreprocessorDirective();
1128 
1129     // Directly allow to 'import <string-literal>' to support protocol buffer
1130     // definitions (github.com/google/protobuf) or missing "#" (either way we
1131     // should not break the line).
1132     IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
1133     if ((Style.Language == FormatStyle::LK_Java &&
1134          CurrentToken->is(Keywords.kw_package)) ||
1135         (Info && Info->getPPKeywordID() == tok::pp_import &&
1136          CurrentToken->Next &&
1137          CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
1138                                      tok::kw_static))) {
1139       next();
1140       parseIncludeDirective();
1141       return LT_ImportStatement;
1142     }
1143 
1144     // If this line starts and ends in '<' and '>', respectively, it is likely
1145     // part of "#define <a/b.h>".
1146     if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
1147       parseIncludeDirective();
1148       return LT_ImportStatement;
1149     }
1150 
1151     // In .proto files, top-level options and package statements are very
1152     // similar to import statements and should not be line-wrapped.
1153     if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
1154         CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) {
1155       next();
1156       if (CurrentToken && CurrentToken->is(tok::identifier)) {
1157         while (CurrentToken)
1158           next();
1159         return LT_ImportStatement;
1160       }
1161     }
1162 
1163     bool KeywordVirtualFound = false;
1164     bool ImportStatement = false;
1165 
1166     // import {...} from '...';
1167     if (Style.Language == FormatStyle::LK_JavaScript &&
1168         CurrentToken->is(Keywords.kw_import))
1169       ImportStatement = true;
1170 
1171     while (CurrentToken) {
1172       if (CurrentToken->is(tok::kw_virtual))
1173         KeywordVirtualFound = true;
1174       if (Style.Language == FormatStyle::LK_JavaScript) {
1175         // export {...} from '...';
1176         // An export followed by "from 'some string';" is a re-export from
1177         // another module identified by a URI and is treated as a
1178         // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
1179         // Just "export {...};" or "export class ..." should not be treated as
1180         // an import in this sense.
1181         if (Line.First->is(tok::kw_export) &&
1182             CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
1183             CurrentToken->Next->isStringLiteral())
1184           ImportStatement = true;
1185         if (isClosureImportStatement(*CurrentToken))
1186           ImportStatement = true;
1187       }
1188       if (!consumeToken())
1189         return LT_Invalid;
1190     }
1191     if (KeywordVirtualFound)
1192       return LT_VirtualFunctionDecl;
1193     if (ImportStatement)
1194       return LT_ImportStatement;
1195 
1196     if (Line.startsWith(TT_ObjCMethodSpecifier)) {
1197       if (Contexts.back().FirstObjCSelectorName)
1198         Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
1199             Contexts.back().LongestObjCSelectorName;
1200       return LT_ObjCMethodDecl;
1201     }
1202 
1203     return LT_Other;
1204   }
1205 
1206 private:
1207   bool isClosureImportStatement(const FormatToken &Tok) {
1208     // FIXME: Closure-library specific stuff should not be hard-coded but be
1209     // configurable.
1210     return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
1211            Tok.Next->Next &&
1212            (Tok.Next->Next->TokenText == "module" ||
1213             Tok.Next->Next->TokenText == "provide" ||
1214             Tok.Next->Next->TokenText == "require" ||
1215             Tok.Next->Next->TokenText == "requireType" ||
1216             Tok.Next->Next->TokenText == "forwardDeclare") &&
1217            Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
1218   }
1219 
1220   void resetTokenMetadata(FormatToken *Token) {
1221     if (!Token)
1222       return;
1223 
1224     // Reset token type in case we have already looked at it and then
1225     // recovered from an error (e.g. failure to find the matching >).
1226     if (!CurrentToken->isOneOf(
1227             TT_LambdaLSquare, TT_LambdaLBrace, TT_ForEachMacro,
1228             TT_TypenameMacro, TT_FunctionLBrace, TT_ImplicitStringLiteral,
1229             TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow, TT_NamespaceMacro,
1230             TT_OverloadedOperator, TT_RegexLiteral, TT_TemplateString,
1231             TT_ObjCStringLiteral))
1232       CurrentToken->Type = TT_Unknown;
1233     CurrentToken->Role.reset();
1234     CurrentToken->MatchingParen = nullptr;
1235     CurrentToken->FakeLParens.clear();
1236     CurrentToken->FakeRParens = 0;
1237   }
1238 
1239   void next() {
1240     if (CurrentToken) {
1241       CurrentToken->NestingLevel = Contexts.size() - 1;
1242       CurrentToken->BindingStrength = Contexts.back().BindingStrength;
1243       modifyContext(*CurrentToken);
1244       determineTokenType(*CurrentToken);
1245       CurrentToken = CurrentToken->Next;
1246     }
1247 
1248     resetTokenMetadata(CurrentToken);
1249   }
1250 
1251   /// A struct to hold information valid in a specific context, e.g.
1252   /// a pair of parenthesis.
1253   struct Context {
1254     Context(tok::TokenKind ContextKind, unsigned BindingStrength,
1255             bool IsExpression)
1256         : ContextKind(ContextKind), BindingStrength(BindingStrength),
1257           IsExpression(IsExpression) {}
1258 
1259     tok::TokenKind ContextKind;
1260     unsigned BindingStrength;
1261     bool IsExpression;
1262     unsigned LongestObjCSelectorName = 0;
1263     bool ColonIsForRangeExpr = false;
1264     bool ColonIsDictLiteral = false;
1265     bool ColonIsObjCMethodExpr = false;
1266     FormatToken *FirstObjCSelectorName = nullptr;
1267     FormatToken *FirstStartOfName = nullptr;
1268     bool CanBeExpression = true;
1269     bool InTemplateArgument = false;
1270     bool InCtorInitializer = false;
1271     bool InInheritanceList = false;
1272     bool CaretFound = false;
1273     bool IsForEachMacro = false;
1274     bool InCpp11AttributeSpecifier = false;
1275     bool InCSharpAttributeSpecifier = false;
1276   };
1277 
1278   /// Puts a new \c Context onto the stack \c Contexts for the lifetime
1279   /// of each instance.
1280   struct ScopedContextCreator {
1281     AnnotatingParser &P;
1282 
1283     ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
1284                          unsigned Increase)
1285         : P(P) {
1286       P.Contexts.push_back(Context(ContextKind,
1287                                    P.Contexts.back().BindingStrength + Increase,
1288                                    P.Contexts.back().IsExpression));
1289     }
1290 
1291     ~ScopedContextCreator() { P.Contexts.pop_back(); }
1292   };
1293 
1294   void modifyContext(const FormatToken &Current) {
1295     if (Current.getPrecedence() == prec::Assignment &&
1296         !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
1297         // Type aliases use `type X = ...;` in TypeScript and can be exported
1298         // using `export type ...`.
1299         !(Style.Language == FormatStyle::LK_JavaScript &&
1300           (Line.startsWith(Keywords.kw_type, tok::identifier) ||
1301            Line.startsWith(tok::kw_export, Keywords.kw_type,
1302                            tok::identifier))) &&
1303         (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
1304       Contexts.back().IsExpression = true;
1305       if (!Line.startsWith(TT_UnaryOperator)) {
1306         for (FormatToken *Previous = Current.Previous;
1307              Previous && Previous->Previous &&
1308              !Previous->Previous->isOneOf(tok::comma, tok::semi);
1309              Previous = Previous->Previous) {
1310           if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
1311             Previous = Previous->MatchingParen;
1312             if (!Previous)
1313               break;
1314           }
1315           if (Previous->opensScope())
1316             break;
1317           if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
1318               Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
1319               Previous->Previous && Previous->Previous->isNot(tok::equal))
1320             Previous->Type = TT_PointerOrReference;
1321         }
1322       }
1323     } else if (Current.is(tok::lessless) &&
1324                (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1325       Contexts.back().IsExpression = true;
1326     } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1327       Contexts.back().IsExpression = true;
1328     } else if (Current.is(TT_TrailingReturnArrow)) {
1329       Contexts.back().IsExpression = false;
1330     } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1331       Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1332     } else if (Current.Previous &&
1333                Current.Previous->is(TT_CtorInitializerColon)) {
1334       Contexts.back().IsExpression = true;
1335       Contexts.back().InCtorInitializer = true;
1336     } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) {
1337       Contexts.back().InInheritanceList = true;
1338     } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1339       for (FormatToken *Previous = Current.Previous;
1340            Previous && Previous->isOneOf(tok::star, tok::amp);
1341            Previous = Previous->Previous)
1342         Previous->Type = TT_PointerOrReference;
1343       if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1344         Contexts.back().IsExpression = false;
1345     } else if (Current.is(tok::kw_new)) {
1346       Contexts.back().CanBeExpression = false;
1347     } else if (Current.isOneOf(tok::semi, tok::exclaim)) {
1348       // This should be the condition or increment in a for-loop.
1349       Contexts.back().IsExpression = true;
1350     }
1351   }
1352 
1353   static FormatToken *untilMatchingParen(FormatToken *Current) {
1354     // Used when `MatchingParen` is not yet established.
1355     int ParenLevel = 0;
1356     while (Current) {
1357       if (Current->is(tok::l_paren))
1358         ParenLevel++;
1359       if (Current->is(tok::r_paren))
1360         ParenLevel--;
1361       if (ParenLevel < 1)
1362         break;
1363       Current = Current->Next;
1364     }
1365     return Current;
1366   }
1367 
1368   static bool isDeductionGuide(FormatToken &Current) {
1369     // Look for a deduction guide template<T> A(...) -> A<...>;
1370     if (Current.Previous && Current.Previous->is(tok::r_paren) &&
1371         Current.startsSequence(tok::arrow, tok::identifier, tok::less)) {
1372       // Find the TemplateCloser.
1373       FormatToken *TemplateCloser = Current.Next->Next;
1374       int NestingLevel = 0;
1375       while (TemplateCloser) {
1376         // Skip over an expressions in parens  A<(3 < 2)>;
1377         if (TemplateCloser->is(tok::l_paren)) {
1378           // No Matching Paren yet so skip to matching paren
1379           TemplateCloser = untilMatchingParen(TemplateCloser);
1380         }
1381         if (TemplateCloser->is(tok::less))
1382           NestingLevel++;
1383         if (TemplateCloser->is(tok::greater))
1384           NestingLevel--;
1385         if (NestingLevel < 1)
1386           break;
1387         TemplateCloser = TemplateCloser->Next;
1388       }
1389       // Assuming we have found the end of the template ensure its followed
1390       // with a semi-colon.
1391       if (TemplateCloser && TemplateCloser->Next &&
1392           TemplateCloser->Next->is(tok::semi) &&
1393           Current.Previous->MatchingParen) {
1394         // Determine if the identifier `A` prior to the A<..>; is the same as
1395         // prior to the A(..)
1396         FormatToken *LeadingIdentifier =
1397             Current.Previous->MatchingParen->Previous;
1398 
1399         // Differentiate a deduction guide by seeing the
1400         // > of the template prior to the leading identifier.
1401         if (LeadingIdentifier) {
1402           FormatToken *PriorLeadingIdentifier = LeadingIdentifier->Previous;
1403           // Skip back past explicit decoration
1404           if (PriorLeadingIdentifier &&
1405               PriorLeadingIdentifier->is(tok::kw_explicit))
1406             PriorLeadingIdentifier = PriorLeadingIdentifier->Previous;
1407 
1408           return (PriorLeadingIdentifier &&
1409                   PriorLeadingIdentifier->is(TT_TemplateCloser) &&
1410                   LeadingIdentifier->TokenText == Current.Next->TokenText);
1411         }
1412       }
1413     }
1414     return false;
1415   }
1416 
1417   void determineTokenType(FormatToken &Current) {
1418     if (!Current.is(TT_Unknown))
1419       // The token type is already known.
1420       return;
1421 
1422     if (Style.Language == FormatStyle::LK_JavaScript) {
1423       if (Current.is(tok::exclaim)) {
1424         if (Current.Previous &&
1425             (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace,
1426                                        tok::r_paren, tok::r_square,
1427                                        tok::r_brace) ||
1428              Current.Previous->Tok.isLiteral())) {
1429           Current.Type = TT_JsNonNullAssertion;
1430           return;
1431         }
1432         if (Current.Next &&
1433             Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1434           Current.Type = TT_JsNonNullAssertion;
1435           return;
1436         }
1437       }
1438     }
1439 
1440     // Line.MightBeFunctionDecl can only be true after the parentheses of a
1441     // function declaration have been found. In this case, 'Current' is a
1442     // trailing token of this declaration and thus cannot be a name.
1443     if (Current.is(Keywords.kw_instanceof)) {
1444       Current.Type = TT_BinaryOperator;
1445     } else if (isStartOfName(Current) &&
1446                (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1447       Contexts.back().FirstStartOfName = &Current;
1448       Current.Type = TT_StartOfName;
1449     } else if (Current.is(tok::semi)) {
1450       // Reset FirstStartOfName after finding a semicolon so that a for loop
1451       // with multiple increment statements is not confused with a for loop
1452       // having multiple variable declarations.
1453       Contexts.back().FirstStartOfName = nullptr;
1454     } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1455       AutoFound = true;
1456     } else if (Current.is(tok::arrow) &&
1457                Style.Language == FormatStyle::LK_Java) {
1458       Current.Type = TT_LambdaArrow;
1459     } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1460                Current.NestingLevel == 0 &&
1461                !Current.Previous->is(tok::kw_operator)) {
1462       // not auto operator->() -> xxx;
1463       Current.Type = TT_TrailingReturnArrow;
1464 
1465     } else if (isDeductionGuide(Current)) {
1466       // Deduction guides trailing arrow " A(...) -> A<T>;".
1467       Current.Type = TT_TrailingReturnArrow;
1468     } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1469       Current.Type = determineStarAmpUsage(Current,
1470                                            Contexts.back().CanBeExpression &&
1471                                                Contexts.back().IsExpression,
1472                                            Contexts.back().InTemplateArgument);
1473     } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1474       Current.Type = determinePlusMinusCaretUsage(Current);
1475       if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1476         Contexts.back().CaretFound = true;
1477     } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1478       Current.Type = determineIncrementUsage(Current);
1479     } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1480       Current.Type = TT_UnaryOperator;
1481     } else if (Current.is(tok::question)) {
1482       if (Style.Language == FormatStyle::LK_JavaScript &&
1483           Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1484         // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1485         // on the interface, not a ternary expression.
1486         Current.Type = TT_JsTypeOptionalQuestion;
1487       } else {
1488         Current.Type = TT_ConditionalExpr;
1489       }
1490     } else if (Current.isBinaryOperator() &&
1491                (!Current.Previous || Current.Previous->isNot(tok::l_square)) &&
1492                (!Current.is(tok::greater) &&
1493                 Style.Language != FormatStyle::LK_TextProto)) {
1494       Current.Type = TT_BinaryOperator;
1495     } else if (Current.is(tok::comment)) {
1496       if (Current.TokenText.startswith("/*")) {
1497         if (Current.TokenText.endswith("*/"))
1498           Current.Type = TT_BlockComment;
1499         else
1500           // The lexer has for some reason determined a comment here. But we
1501           // cannot really handle it, if it isn't properly terminated.
1502           Current.Tok.setKind(tok::unknown);
1503       } else {
1504         Current.Type = TT_LineComment;
1505       }
1506     } else if (Current.is(tok::r_paren)) {
1507       if (rParenEndsCast(Current))
1508         Current.Type = TT_CastRParen;
1509       if (Current.MatchingParen && Current.Next &&
1510           !Current.Next->isBinaryOperator() &&
1511           !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1512                                  tok::comma, tok::period, tok::arrow,
1513                                  tok::coloncolon))
1514         if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1515           // Make sure this isn't the return type of an Obj-C block declaration
1516           if (AfterParen->Tok.isNot(tok::caret)) {
1517             if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1518               if (BeforeParen->is(tok::identifier) &&
1519                   !BeforeParen->is(TT_TypenameMacro) &&
1520                   BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1521                   (!BeforeParen->Previous ||
1522                    BeforeParen->Previous->ClosesTemplateDeclaration))
1523                 Current.Type = TT_FunctionAnnotationRParen;
1524           }
1525         }
1526     } else if (Current.is(tok::at) && Current.Next &&
1527                Style.Language != FormatStyle::LK_JavaScript &&
1528                Style.Language != FormatStyle::LK_Java) {
1529       // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1530       // marks declarations and properties that need special formatting.
1531       switch (Current.Next->Tok.getObjCKeywordID()) {
1532       case tok::objc_interface:
1533       case tok::objc_implementation:
1534       case tok::objc_protocol:
1535         Current.Type = TT_ObjCDecl;
1536         break;
1537       case tok::objc_property:
1538         Current.Type = TT_ObjCProperty;
1539         break;
1540       default:
1541         break;
1542       }
1543     } else if (Current.is(tok::period)) {
1544       FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1545       if (PreviousNoComment &&
1546           PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1547         Current.Type = TT_DesignatedInitializerPeriod;
1548       else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1549                Current.Previous->isOneOf(TT_JavaAnnotation,
1550                                          TT_LeadingJavaAnnotation)) {
1551         Current.Type = Current.Previous->Type;
1552       }
1553     } else if (canBeObjCSelectorComponent(Current) &&
1554                // FIXME(bug 36976): ObjC return types shouldn't use
1555                // TT_CastRParen.
1556                Current.Previous && Current.Previous->is(TT_CastRParen) &&
1557                Current.Previous->MatchingParen &&
1558                Current.Previous->MatchingParen->Previous &&
1559                Current.Previous->MatchingParen->Previous->is(
1560                    TT_ObjCMethodSpecifier)) {
1561       // This is the first part of an Objective-C selector name. (If there's no
1562       // colon after this, this is the only place which annotates the identifier
1563       // as a selector.)
1564       Current.Type = TT_SelectorName;
1565     } else if (Current.isOneOf(tok::identifier, tok::kw_const,
1566                                tok::kw_noexcept) &&
1567                Current.Previous &&
1568                !Current.Previous->isOneOf(tok::equal, tok::at) &&
1569                Line.MightBeFunctionDecl && Contexts.size() == 1) {
1570       // Line.MightBeFunctionDecl can only be true after the parentheses of a
1571       // function declaration have been found.
1572       Current.Type = TT_TrailingAnnotation;
1573     } else if ((Style.Language == FormatStyle::LK_Java ||
1574                 Style.Language == FormatStyle::LK_JavaScript) &&
1575                Current.Previous) {
1576       if (Current.Previous->is(tok::at) &&
1577           Current.isNot(Keywords.kw_interface)) {
1578         const FormatToken &AtToken = *Current.Previous;
1579         const FormatToken *Previous = AtToken.getPreviousNonComment();
1580         if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1581           Current.Type = TT_LeadingJavaAnnotation;
1582         else
1583           Current.Type = TT_JavaAnnotation;
1584       } else if (Current.Previous->is(tok::period) &&
1585                  Current.Previous->isOneOf(TT_JavaAnnotation,
1586                                            TT_LeadingJavaAnnotation)) {
1587         Current.Type = Current.Previous->Type;
1588       }
1589     }
1590   }
1591 
1592   /// Take a guess at whether \p Tok starts a name of a function or
1593   /// variable declaration.
1594   ///
1595   /// This is a heuristic based on whether \p Tok is an identifier following
1596   /// something that is likely a type.
1597   bool isStartOfName(const FormatToken &Tok) {
1598     if (Tok.isNot(tok::identifier) || !Tok.Previous)
1599       return false;
1600 
1601     if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1602                               Keywords.kw_as))
1603       return false;
1604     if (Style.Language == FormatStyle::LK_JavaScript &&
1605         Tok.Previous->is(Keywords.kw_in))
1606       return false;
1607 
1608     // Skip "const" as it does not have an influence on whether this is a name.
1609     FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1610     while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1611       PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1612 
1613     if (!PreviousNotConst)
1614       return false;
1615 
1616     bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1617                        PreviousNotConst->Previous &&
1618                        PreviousNotConst->Previous->is(tok::hash);
1619 
1620     if (PreviousNotConst->is(TT_TemplateCloser))
1621       return PreviousNotConst && PreviousNotConst->MatchingParen &&
1622              PreviousNotConst->MatchingParen->Previous &&
1623              PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1624              PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1625 
1626     if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen &&
1627         PreviousNotConst->MatchingParen->Previous &&
1628         PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype))
1629       return true;
1630 
1631     return (!IsPPKeyword &&
1632             PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1633            PreviousNotConst->is(TT_PointerOrReference) ||
1634            PreviousNotConst->isSimpleTypeSpecifier();
1635   }
1636 
1637   /// Determine whether ')' is ending a cast.
1638   bool rParenEndsCast(const FormatToken &Tok) {
1639     // C-style casts are only used in C++ and Java.
1640     if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java)
1641       return false;
1642 
1643     // Empty parens aren't casts and there are no casts at the end of the line.
1644     if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1645       return false;
1646 
1647     FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1648     if (LeftOfParens) {
1649       // If there is a closing parenthesis left of the current parentheses,
1650       // look past it as these might be chained casts.
1651       if (LeftOfParens->is(tok::r_paren)) {
1652         if (!LeftOfParens->MatchingParen ||
1653             !LeftOfParens->MatchingParen->Previous)
1654           return false;
1655         LeftOfParens = LeftOfParens->MatchingParen->Previous;
1656       }
1657 
1658       // If there is an identifier (or with a few exceptions a keyword) right
1659       // before the parentheses, this is unlikely to be a cast.
1660       if (LeftOfParens->Tok.getIdentifierInfo() &&
1661           !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1662                                  tok::kw_delete))
1663         return false;
1664 
1665       // Certain other tokens right before the parentheses are also signals that
1666       // this cannot be a cast.
1667       if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1668                                 TT_TemplateCloser, tok::ellipsis))
1669         return false;
1670     }
1671 
1672     if (Tok.Next->is(tok::question))
1673       return false;
1674 
1675     // Functions which end with decorations like volatile, noexcept are unlikely
1676     // to be casts.
1677     if (Tok.Next->isOneOf(tok::kw_noexcept, tok::kw_volatile, tok::kw_const,
1678                           tok::kw_throw, tok::arrow, Keywords.kw_override,
1679                           Keywords.kw_final) ||
1680         isCpp11AttributeSpecifier(*Tok.Next))
1681       return false;
1682 
1683     // As Java has no function types, a "(" after the ")" likely means that this
1684     // is a cast.
1685     if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1686       return true;
1687 
1688     // If a (non-string) literal follows, this is likely a cast.
1689     if (Tok.Next->isNot(tok::string_literal) &&
1690         (Tok.Next->Tok.isLiteral() ||
1691          Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1692       return true;
1693 
1694     // Heuristically try to determine whether the parentheses contain a type.
1695     bool ParensAreType =
1696         !Tok.Previous ||
1697         Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) ||
1698         Tok.Previous->isSimpleTypeSpecifier();
1699     bool ParensCouldEndDecl =
1700         Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1701     if (ParensAreType && !ParensCouldEndDecl)
1702       return true;
1703 
1704     // At this point, we heuristically assume that there are no casts at the
1705     // start of the line. We assume that we have found most cases where there
1706     // are by the logic above, e.g. "(void)x;".
1707     if (!LeftOfParens)
1708       return false;
1709 
1710     // Certain token types inside the parentheses mean that this can't be a
1711     // cast.
1712     for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1713          Token = Token->Next)
1714       if (Token->is(TT_BinaryOperator))
1715         return false;
1716 
1717     // If the following token is an identifier or 'this', this is a cast. All
1718     // cases where this can be something else are handled above.
1719     if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1720       return true;
1721 
1722     if (!Tok.Next->Next)
1723       return false;
1724 
1725     // If the next token after the parenthesis is a unary operator, assume
1726     // that this is cast, unless there are unexpected tokens inside the
1727     // parenthesis.
1728     bool NextIsUnary =
1729         Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1730     if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1731         !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1732       return false;
1733     // Search for unexpected tokens.
1734     for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1735          Prev = Prev->Previous) {
1736       if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1737         return false;
1738     }
1739     return true;
1740   }
1741 
1742   /// Return the type of the given token assuming it is * or &.
1743   TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1744                                   bool InTemplateArgument) {
1745     if (Style.Language == FormatStyle::LK_JavaScript)
1746       return TT_BinaryOperator;
1747 
1748     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1749     if (!PrevToken)
1750       return TT_UnaryOperator;
1751 
1752     const FormatToken *NextToken = Tok.getNextNonComment();
1753     if (!NextToken ||
1754         NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const,
1755                            tok::kw_noexcept) ||
1756         (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1757       return TT_PointerOrReference;
1758 
1759     if (PrevToken->is(tok::coloncolon))
1760       return TT_PointerOrReference;
1761 
1762     if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1763                            tok::comma, tok::semi, tok::kw_return, tok::colon,
1764                            tok::equal, tok::kw_delete, tok::kw_sizeof,
1765                            tok::kw_throw) ||
1766         PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1767                            TT_UnaryOperator, TT_CastRParen))
1768       return TT_UnaryOperator;
1769 
1770     if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1771       return TT_PointerOrReference;
1772     if (NextToken->is(tok::kw_operator) && !IsExpression)
1773       return TT_PointerOrReference;
1774     if (NextToken->isOneOf(tok::comma, tok::semi))
1775       return TT_PointerOrReference;
1776 
1777     if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) {
1778       FormatToken *TokenBeforeMatchingParen =
1779           PrevToken->MatchingParen->getPreviousNonComment();
1780       if (TokenBeforeMatchingParen &&
1781           TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype,
1782                                             TT_TypenameMacro))
1783         return TT_PointerOrReference;
1784     }
1785 
1786     if (PrevToken->Tok.isLiteral() ||
1787         PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1788                            tok::kw_false, tok::r_brace) ||
1789         NextToken->Tok.isLiteral() ||
1790         NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1791         NextToken->isUnaryOperator() ||
1792         // If we know we're in a template argument, there are no named
1793         // declarations. Thus, having an identifier on the right-hand side
1794         // indicates a binary operator.
1795         (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1796       return TT_BinaryOperator;
1797 
1798     // "&&(" is quite unlikely to be two successive unary "&".
1799     if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren))
1800       return TT_BinaryOperator;
1801 
1802     // This catches some cases where evaluation order is used as control flow:
1803     //   aaa && aaa->f();
1804     const FormatToken *NextNextToken = NextToken->getNextNonComment();
1805     if (NextNextToken && NextNextToken->is(tok::arrow))
1806       return TT_BinaryOperator;
1807 
1808     // It is very unlikely that we are going to find a pointer or reference type
1809     // definition on the RHS of an assignment.
1810     if (IsExpression && !Contexts.back().CaretFound)
1811       return TT_BinaryOperator;
1812 
1813     return TT_PointerOrReference;
1814   }
1815 
1816   TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
1817     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1818     if (!PrevToken)
1819       return TT_UnaryOperator;
1820 
1821     if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator))
1822       // This must be a sequence of leading unary operators.
1823       return TT_UnaryOperator;
1824 
1825     // Use heuristics to recognize unary operators.
1826     if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
1827                            tok::question, tok::colon, tok::kw_return,
1828                            tok::kw_case, tok::at, tok::l_brace, tok::kw_throw))
1829       return TT_UnaryOperator;
1830 
1831     // There can't be two consecutive binary operators.
1832     if (PrevToken->is(TT_BinaryOperator))
1833       return TT_UnaryOperator;
1834 
1835     // Fall back to marking the token as binary operator.
1836     return TT_BinaryOperator;
1837   }
1838 
1839   /// Determine whether ++/-- are pre- or post-increments/-decrements.
1840   TokenType determineIncrementUsage(const FormatToken &Tok) {
1841     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1842     if (!PrevToken || PrevToken->is(TT_CastRParen))
1843       return TT_UnaryOperator;
1844     if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
1845       return TT_TrailingUnaryOperator;
1846 
1847     return TT_UnaryOperator;
1848   }
1849 
1850   SmallVector<Context, 8> Contexts;
1851 
1852   const FormatStyle &Style;
1853   AnnotatedLine &Line;
1854   FormatToken *CurrentToken;
1855   bool AutoFound;
1856   const AdditionalKeywords &Keywords;
1857 
1858   // Set of "<" tokens that do not open a template parameter list. If parseAngle
1859   // determines that a specific token can't be a template opener, it will make
1860   // same decision irrespective of the decisions for tokens leading up to it.
1861   // Store this information to prevent this from causing exponential runtime.
1862   llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
1863 };
1864 
1865 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
1866 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
1867 
1868 /// Parses binary expressions by inserting fake parenthesis based on
1869 /// operator precedence.
1870 class ExpressionParser {
1871 public:
1872   ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
1873                    AnnotatedLine &Line)
1874       : Style(Style), Keywords(Keywords), Current(Line.First) {}
1875 
1876   /// Parse expressions with the given operator precedence.
1877   void parse(int Precedence = 0) {
1878     // Skip 'return' and ObjC selector colons as they are not part of a binary
1879     // expression.
1880     while (Current && (Current->is(tok::kw_return) ||
1881                        (Current->is(tok::colon) &&
1882                         Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
1883       next();
1884 
1885     if (!Current || Precedence > PrecedenceArrowAndPeriod)
1886       return;
1887 
1888     // Conditional expressions need to be parsed separately for proper nesting.
1889     if (Precedence == prec::Conditional) {
1890       parseConditionalExpr();
1891       return;
1892     }
1893 
1894     // Parse unary operators, which all have a higher precedence than binary
1895     // operators.
1896     if (Precedence == PrecedenceUnaryOperator) {
1897       parseUnaryOperator();
1898       return;
1899     }
1900 
1901     FormatToken *Start = Current;
1902     FormatToken *LatestOperator = nullptr;
1903     unsigned OperatorIndex = 0;
1904 
1905     while (Current) {
1906       // Consume operators with higher precedence.
1907       parse(Precedence + 1);
1908 
1909       int CurrentPrecedence = getCurrentPrecedence();
1910 
1911       if (Current && Current->is(TT_SelectorName) &&
1912           Precedence == CurrentPrecedence) {
1913         if (LatestOperator)
1914           addFakeParenthesis(Start, prec::Level(Precedence));
1915         Start = Current;
1916       }
1917 
1918       // At the end of the line or when an operator with higher precedence is
1919       // found, insert fake parenthesis and return.
1920       if (!Current ||
1921           (Current->closesScope() &&
1922            (Current->MatchingParen || Current->is(TT_TemplateString))) ||
1923           (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
1924           (CurrentPrecedence == prec::Conditional &&
1925            Precedence == prec::Assignment && Current->is(tok::colon))) {
1926         break;
1927       }
1928 
1929       // Consume scopes: (), [], <> and {}
1930       if (Current->opensScope()) {
1931         // In fragment of a JavaScript template string can look like '}..${' and
1932         // thus close a scope and open a new one at the same time.
1933         while (Current && (!Current->closesScope() || Current->opensScope())) {
1934           next();
1935           parse();
1936         }
1937         next();
1938       } else {
1939         // Operator found.
1940         if (CurrentPrecedence == Precedence) {
1941           if (LatestOperator)
1942             LatestOperator->NextOperator = Current;
1943           LatestOperator = Current;
1944           Current->OperatorIndex = OperatorIndex;
1945           ++OperatorIndex;
1946         }
1947         next(/*SkipPastLeadingComments=*/Precedence > 0);
1948       }
1949     }
1950 
1951     if (LatestOperator && (Current || Precedence > 0)) {
1952       // LatestOperator->LastOperator = true;
1953       if (Precedence == PrecedenceArrowAndPeriod) {
1954         // Call expressions don't have a binary operator precedence.
1955         addFakeParenthesis(Start, prec::Unknown);
1956       } else {
1957         addFakeParenthesis(Start, prec::Level(Precedence));
1958       }
1959     }
1960   }
1961 
1962 private:
1963   /// Gets the precedence (+1) of the given token for binary operators
1964   /// and other tokens that we treat like binary operators.
1965   int getCurrentPrecedence() {
1966     if (Current) {
1967       const FormatToken *NextNonComment = Current->getNextNonComment();
1968       if (Current->is(TT_ConditionalExpr))
1969         return prec::Conditional;
1970       if (NextNonComment && Current->is(TT_SelectorName) &&
1971           (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) ||
1972            ((Style.Language == FormatStyle::LK_Proto ||
1973              Style.Language == FormatStyle::LK_TextProto) &&
1974             NextNonComment->is(tok::less))))
1975         return prec::Assignment;
1976       if (Current->is(TT_JsComputedPropertyName))
1977         return prec::Assignment;
1978       if (Current->is(TT_LambdaArrow))
1979         return prec::Comma;
1980       if (Current->is(TT_JsFatArrow))
1981         return prec::Assignment;
1982       if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
1983           (Current->is(tok::comment) && NextNonComment &&
1984            NextNonComment->is(TT_SelectorName)))
1985         return 0;
1986       if (Current->is(TT_RangeBasedForLoopColon))
1987         return prec::Comma;
1988       if ((Style.Language == FormatStyle::LK_Java ||
1989            Style.Language == FormatStyle::LK_JavaScript) &&
1990           Current->is(Keywords.kw_instanceof))
1991         return prec::Relational;
1992       if (Style.Language == FormatStyle::LK_JavaScript &&
1993           Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
1994         return prec::Relational;
1995       if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
1996         return Current->getPrecedence();
1997       if (Current->isOneOf(tok::period, tok::arrow))
1998         return PrecedenceArrowAndPeriod;
1999       if ((Style.Language == FormatStyle::LK_Java ||
2000            Style.Language == FormatStyle::LK_JavaScript) &&
2001           Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
2002                            Keywords.kw_throws))
2003         return 0;
2004     }
2005     return -1;
2006   }
2007 
2008   void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
2009     Start->FakeLParens.push_back(Precedence);
2010     if (Precedence > prec::Unknown)
2011       Start->StartsBinaryExpression = true;
2012     if (Current) {
2013       FormatToken *Previous = Current->Previous;
2014       while (Previous->is(tok::comment) && Previous->Previous)
2015         Previous = Previous->Previous;
2016       ++Previous->FakeRParens;
2017       if (Precedence > prec::Unknown)
2018         Previous->EndsBinaryExpression = true;
2019     }
2020   }
2021 
2022   /// Parse unary operator expressions and surround them with fake
2023   /// parentheses if appropriate.
2024   void parseUnaryOperator() {
2025     llvm::SmallVector<FormatToken *, 2> Tokens;
2026     while (Current && Current->is(TT_UnaryOperator)) {
2027       Tokens.push_back(Current);
2028       next();
2029     }
2030     parse(PrecedenceArrowAndPeriod);
2031     for (FormatToken *Token : llvm::reverse(Tokens))
2032       // The actual precedence doesn't matter.
2033       addFakeParenthesis(Token, prec::Unknown);
2034   }
2035 
2036   void parseConditionalExpr() {
2037     while (Current && Current->isTrailingComment()) {
2038       next();
2039     }
2040     FormatToken *Start = Current;
2041     parse(prec::LogicalOr);
2042     if (!Current || !Current->is(tok::question))
2043       return;
2044     next();
2045     parse(prec::Assignment);
2046     if (!Current || Current->isNot(TT_ConditionalExpr))
2047       return;
2048     next();
2049     parse(prec::Assignment);
2050     addFakeParenthesis(Start, prec::Conditional);
2051   }
2052 
2053   void next(bool SkipPastLeadingComments = true) {
2054     if (Current)
2055       Current = Current->Next;
2056     while (Current &&
2057            (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
2058            Current->isTrailingComment())
2059       Current = Current->Next;
2060   }
2061 
2062   const FormatStyle &Style;
2063   const AdditionalKeywords &Keywords;
2064   FormatToken *Current;
2065 };
2066 
2067 } // end anonymous namespace
2068 
2069 void TokenAnnotator::setCommentLineLevels(
2070     SmallVectorImpl<AnnotatedLine *> &Lines) {
2071   const AnnotatedLine *NextNonCommentLine = nullptr;
2072   for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
2073                                                           E = Lines.rend();
2074        I != E; ++I) {
2075     bool CommentLine = true;
2076     for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
2077       if (!Tok->is(tok::comment)) {
2078         CommentLine = false;
2079         break;
2080       }
2081     }
2082 
2083     // If the comment is currently aligned with the line immediately following
2084     // it, that's probably intentional and we should keep it.
2085     if (NextNonCommentLine && CommentLine &&
2086         NextNonCommentLine->First->NewlinesBefore <= 1 &&
2087         NextNonCommentLine->First->OriginalColumn ==
2088             (*I)->First->OriginalColumn) {
2089       // Align comments for preprocessor lines with the # in column 0 if
2090       // preprocessor lines are not indented. Otherwise, align with the next
2091       // line.
2092       (*I)->Level =
2093           (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
2094            (NextNonCommentLine->Type == LT_PreprocessorDirective ||
2095             NextNonCommentLine->Type == LT_ImportStatement))
2096               ? 0
2097               : NextNonCommentLine->Level;
2098     } else {
2099       NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
2100     }
2101 
2102     setCommentLineLevels((*I)->Children);
2103   }
2104 }
2105 
2106 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
2107   unsigned Result = 0;
2108   for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
2109     Result = std::max(Result, Tok->NestingLevel);
2110   return Result;
2111 }
2112 
2113 void TokenAnnotator::annotate(AnnotatedLine &Line) {
2114   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2115                                                   E = Line.Children.end();
2116        I != E; ++I) {
2117     annotate(**I);
2118   }
2119   AnnotatingParser Parser(Style, Line, Keywords);
2120   Line.Type = Parser.parseLine();
2121 
2122   // With very deep nesting, ExpressionParser uses lots of stack and the
2123   // formatting algorithm is very slow. We're not going to do a good job here
2124   // anyway - it's probably generated code being formatted by mistake.
2125   // Just skip the whole line.
2126   if (maxNestingDepth(Line) > 50)
2127     Line.Type = LT_Invalid;
2128 
2129   if (Line.Type == LT_Invalid)
2130     return;
2131 
2132   ExpressionParser ExprParser(Style, Keywords, Line);
2133   ExprParser.parse();
2134 
2135   if (Line.startsWith(TT_ObjCMethodSpecifier))
2136     Line.Type = LT_ObjCMethodDecl;
2137   else if (Line.startsWith(TT_ObjCDecl))
2138     Line.Type = LT_ObjCDecl;
2139   else if (Line.startsWith(TT_ObjCProperty))
2140     Line.Type = LT_ObjCProperty;
2141 
2142   Line.First->SpacesRequiredBefore = 1;
2143   Line.First->CanBreakBefore = Line.First->MustBreakBefore;
2144 }
2145 
2146 // This function heuristically determines whether 'Current' starts the name of a
2147 // function declaration.
2148 static bool isFunctionDeclarationName(const FormatToken &Current,
2149                                       const AnnotatedLine &Line) {
2150   auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * {
2151     for (; Next; Next = Next->Next) {
2152       if (Next->is(TT_OverloadedOperatorLParen))
2153         return Next;
2154       if (Next->is(TT_OverloadedOperator))
2155         continue;
2156       if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
2157         // For 'new[]' and 'delete[]'.
2158         if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next &&
2159             Next->Next->Next->is(tok::r_square))
2160           Next = Next->Next->Next;
2161         continue;
2162       }
2163 
2164       break;
2165     }
2166     return nullptr;
2167   };
2168 
2169   // Find parentheses of parameter list.
2170   const FormatToken *Next = Current.Next;
2171   if (Current.is(tok::kw_operator)) {
2172     if (Current.Previous && Current.Previous->is(tok::coloncolon))
2173       return false;
2174     Next = skipOperatorName(Next);
2175   } else {
2176     if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
2177       return false;
2178     for (; Next; Next = Next->Next) {
2179       if (Next->is(TT_TemplateOpener)) {
2180         Next = Next->MatchingParen;
2181       } else if (Next->is(tok::coloncolon)) {
2182         Next = Next->Next;
2183         if (!Next)
2184           return false;
2185         if (Next->is(tok::kw_operator)) {
2186           Next = skipOperatorName(Next->Next);
2187           break;
2188         }
2189         if (!Next->is(tok::identifier))
2190           return false;
2191       } else if (Next->is(tok::l_paren)) {
2192         break;
2193       } else {
2194         return false;
2195       }
2196     }
2197   }
2198 
2199   // Check whether parameter list can belong to a function declaration.
2200   if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
2201     return false;
2202   // If the lines ends with "{", this is likely an function definition.
2203   if (Line.Last->is(tok::l_brace))
2204     return true;
2205   if (Next->Next == Next->MatchingParen)
2206     return true; // Empty parentheses.
2207   // If there is an &/&& after the r_paren, this is likely a function.
2208   if (Next->MatchingParen->Next &&
2209       Next->MatchingParen->Next->is(TT_PointerOrReference))
2210     return true;
2211   for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
2212        Tok = Tok->Next) {
2213     if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) {
2214       Tok = Tok->MatchingParen;
2215       continue;
2216     }
2217     if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
2218         Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
2219       return true;
2220     if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
2221         Tok->Tok.isLiteral())
2222       return false;
2223   }
2224   return false;
2225 }
2226 
2227 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
2228   assert(Line.MightBeFunctionDecl);
2229 
2230   if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
2231        Style.AlwaysBreakAfterReturnType ==
2232            FormatStyle::RTBS_TopLevelDefinitions) &&
2233       Line.Level > 0)
2234     return false;
2235 
2236   switch (Style.AlwaysBreakAfterReturnType) {
2237   case FormatStyle::RTBS_None:
2238     return false;
2239   case FormatStyle::RTBS_All:
2240   case FormatStyle::RTBS_TopLevel:
2241     return true;
2242   case FormatStyle::RTBS_AllDefinitions:
2243   case FormatStyle::RTBS_TopLevelDefinitions:
2244     return Line.mightBeFunctionDefinition();
2245   }
2246 
2247   return false;
2248 }
2249 
2250 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
2251   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2252                                                   E = Line.Children.end();
2253        I != E; ++I) {
2254     calculateFormattingInformation(**I);
2255   }
2256 
2257   Line.First->TotalLength =
2258       Line.First->IsMultiline ? Style.ColumnLimit
2259                               : Line.FirstStartColumn + Line.First->ColumnWidth;
2260   FormatToken *Current = Line.First->Next;
2261   bool InFunctionDecl = Line.MightBeFunctionDecl;
2262   while (Current) {
2263     if (isFunctionDeclarationName(*Current, Line))
2264       Current->Type = TT_FunctionDeclarationName;
2265     if (Current->is(TT_LineComment)) {
2266       if (Current->Previous->BlockKind == BK_BracedInit &&
2267           Current->Previous->opensScope())
2268         Current->SpacesRequiredBefore =
2269             (Style.Cpp11BracedListStyle && !Style.SpacesInParentheses) ? 0 : 1;
2270       else
2271         Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
2272 
2273       // If we find a trailing comment, iterate backwards to determine whether
2274       // it seems to relate to a specific parameter. If so, break before that
2275       // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
2276       // to the previous line in:
2277       //   SomeFunction(a,
2278       //                b, // comment
2279       //                c);
2280       if (!Current->HasUnescapedNewline) {
2281         for (FormatToken *Parameter = Current->Previous; Parameter;
2282              Parameter = Parameter->Previous) {
2283           if (Parameter->isOneOf(tok::comment, tok::r_brace))
2284             break;
2285           if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
2286             if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
2287                 Parameter->HasUnescapedNewline)
2288               Parameter->MustBreakBefore = true;
2289             break;
2290           }
2291         }
2292       }
2293     } else if (Current->SpacesRequiredBefore == 0 &&
2294                spaceRequiredBefore(Line, *Current)) {
2295       Current->SpacesRequiredBefore = 1;
2296     }
2297 
2298     Current->MustBreakBefore =
2299         Current->MustBreakBefore || mustBreakBefore(Line, *Current);
2300 
2301     if (!Current->MustBreakBefore && InFunctionDecl &&
2302         Current->is(TT_FunctionDeclarationName))
2303       Current->MustBreakBefore = mustBreakForReturnType(Line);
2304 
2305     Current->CanBreakBefore =
2306         Current->MustBreakBefore || canBreakBefore(Line, *Current);
2307     unsigned ChildSize = 0;
2308     if (Current->Previous->Children.size() == 1) {
2309       FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
2310       ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
2311                                                   : LastOfChild.TotalLength + 1;
2312     }
2313     const FormatToken *Prev = Current->Previous;
2314     if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
2315         (Prev->Children.size() == 1 &&
2316          Prev->Children[0]->First->MustBreakBefore) ||
2317         Current->IsMultiline)
2318       Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
2319     else
2320       Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
2321                              ChildSize + Current->SpacesRequiredBefore;
2322 
2323     if (Current->is(TT_CtorInitializerColon))
2324       InFunctionDecl = false;
2325 
2326     // FIXME: Only calculate this if CanBreakBefore is true once static
2327     // initializers etc. are sorted out.
2328     // FIXME: Move magic numbers to a better place.
2329 
2330     // Reduce penalty for aligning ObjC method arguments using the colon
2331     // alignment as this is the canonical way (still prefer fitting everything
2332     // into one line if possible). Trying to fit a whole expression into one
2333     // line should not force other line breaks (e.g. when ObjC method
2334     // expression is a part of other expression).
2335     Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl);
2336     if (Style.Language == FormatStyle::LK_ObjC &&
2337         Current->is(TT_SelectorName) && Current->ParameterIndex > 0) {
2338       if (Current->ParameterIndex == 1)
2339         Current->SplitPenalty += 5 * Current->BindingStrength;
2340     } else {
2341       Current->SplitPenalty += 20 * Current->BindingStrength;
2342     }
2343 
2344     Current = Current->Next;
2345   }
2346 
2347   calculateUnbreakableTailLengths(Line);
2348   unsigned IndentLevel = Line.Level;
2349   for (Current = Line.First; Current != nullptr; Current = Current->Next) {
2350     if (Current->Role)
2351       Current->Role->precomputeFormattingInfos(Current);
2352     if (Current->MatchingParen &&
2353         Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
2354       assert(IndentLevel > 0);
2355       --IndentLevel;
2356     }
2357     Current->IndentLevel = IndentLevel;
2358     if (Current->opensBlockOrBlockTypeList(Style))
2359       ++IndentLevel;
2360   }
2361 
2362   LLVM_DEBUG({ printDebugInfo(Line); });
2363 }
2364 
2365 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
2366   unsigned UnbreakableTailLength = 0;
2367   FormatToken *Current = Line.Last;
2368   while (Current) {
2369     Current->UnbreakableTailLength = UnbreakableTailLength;
2370     if (Current->CanBreakBefore ||
2371         Current->isOneOf(tok::comment, tok::string_literal)) {
2372       UnbreakableTailLength = 0;
2373     } else {
2374       UnbreakableTailLength +=
2375           Current->ColumnWidth + Current->SpacesRequiredBefore;
2376     }
2377     Current = Current->Previous;
2378   }
2379 }
2380 
2381 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
2382                                       const FormatToken &Tok,
2383                                       bool InFunctionDecl) {
2384   const FormatToken &Left = *Tok.Previous;
2385   const FormatToken &Right = Tok;
2386 
2387   if (Left.is(tok::semi))
2388     return 0;
2389 
2390   if (Style.Language == FormatStyle::LK_Java) {
2391     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
2392       return 1;
2393     if (Right.is(Keywords.kw_implements))
2394       return 2;
2395     if (Left.is(tok::comma) && Left.NestingLevel == 0)
2396       return 3;
2397   } else if (Style.Language == FormatStyle::LK_JavaScript) {
2398     if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
2399       return 100;
2400     if (Left.is(TT_JsTypeColon))
2401       return 35;
2402     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2403         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2404       return 100;
2405     // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()".
2406     if (Left.opensScope() && Right.closesScope())
2407       return 200;
2408   }
2409 
2410   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2411     return 1;
2412   if (Right.is(tok::l_square)) {
2413     if (Style.Language == FormatStyle::LK_Proto)
2414       return 1;
2415     if (Left.is(tok::r_square))
2416       return 200;
2417     // Slightly prefer formatting local lambda definitions like functions.
2418     if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
2419       return 35;
2420     if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2421                        TT_ArrayInitializerLSquare,
2422                        TT_DesignatedInitializerLSquare, TT_AttributeSquare))
2423       return 500;
2424   }
2425 
2426   if (Left.is(tok::coloncolon) ||
2427       (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
2428     return 500;
2429   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2430       Right.is(tok::kw_operator)) {
2431     if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
2432       return 3;
2433     if (Left.is(TT_StartOfName))
2434       return 110;
2435     if (InFunctionDecl && Right.NestingLevel == 0)
2436       return Style.PenaltyReturnTypeOnItsOwnLine;
2437     return 200;
2438   }
2439   if (Right.is(TT_PointerOrReference))
2440     return 190;
2441   if (Right.is(TT_LambdaArrow))
2442     return 110;
2443   if (Left.is(tok::equal) && Right.is(tok::l_brace))
2444     return 160;
2445   if (Left.is(TT_CastRParen))
2446     return 100;
2447   if (Left.isOneOf(tok::kw_class, tok::kw_struct))
2448     return 5000;
2449   if (Left.is(tok::comment))
2450     return 1000;
2451 
2452   if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon,
2453                    TT_CtorInitializerColon))
2454     return 2;
2455 
2456   if (Right.isMemberAccess()) {
2457     // Breaking before the "./->" of a chained call/member access is reasonably
2458     // cheap, as formatting those with one call per line is generally
2459     // desirable. In particular, it should be cheaper to break before the call
2460     // than it is to break inside a call's parameters, which could lead to weird
2461     // "hanging" indents. The exception is the very last "./->" to support this
2462     // frequent pattern:
2463     //
2464     //   aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
2465     //       dddddddd);
2466     //
2467     // which might otherwise be blown up onto many lines. Here, clang-format
2468     // won't produce "hanging" indents anyway as there is no other trailing
2469     // call.
2470     //
2471     // Also apply higher penalty is not a call as that might lead to a wrapping
2472     // like:
2473     //
2474     //   aaaaaaa
2475     //       .aaaaaaaaa.bbbbbbbb(cccccccc);
2476     return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
2477                ? 150
2478                : 35;
2479   }
2480 
2481   if (Right.is(TT_TrailingAnnotation) &&
2482       (!Right.Next || Right.Next->isNot(tok::l_paren))) {
2483     // Moving trailing annotations to the next line is fine for ObjC method
2484     // declarations.
2485     if (Line.startsWith(TT_ObjCMethodSpecifier))
2486       return 10;
2487     // Generally, breaking before a trailing annotation is bad unless it is
2488     // function-like. It seems to be especially preferable to keep standard
2489     // annotations (i.e. "const", "final" and "override") on the same line.
2490     // Use a slightly higher penalty after ")" so that annotations like
2491     // "const override" are kept together.
2492     bool is_short_annotation = Right.TokenText.size() < 10;
2493     return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2494   }
2495 
2496   // In for-loops, prefer breaking at ',' and ';'.
2497   if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2498     return 4;
2499 
2500   // In Objective-C method expressions, prefer breaking before "param:" over
2501   // breaking after it.
2502   if (Right.is(TT_SelectorName))
2503     return 0;
2504   if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2505     return Line.MightBeFunctionDecl ? 50 : 500;
2506 
2507   // In Objective-C type declarations, avoid breaking after the category's
2508   // open paren (we'll prefer breaking after the protocol list's opening
2509   // angle bracket, if present).
2510   if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous &&
2511       Left.Previous->isOneOf(tok::identifier, tok::greater))
2512     return 500;
2513 
2514   if (Left.is(tok::l_paren) && InFunctionDecl &&
2515       Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2516     return 100;
2517   if (Left.is(tok::l_paren) && Left.Previous &&
2518       (Left.Previous->is(tok::kw_for) || Left.Previous->isIf()))
2519     return 1000;
2520   if (Left.is(tok::equal) && InFunctionDecl)
2521     return 110;
2522   if (Right.is(tok::r_brace))
2523     return 1;
2524   if (Left.is(TT_TemplateOpener))
2525     return 100;
2526   if (Left.opensScope()) {
2527     if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
2528       return 0;
2529     if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle)
2530       return 19;
2531     return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2532                                    : 19;
2533   }
2534   if (Left.is(TT_JavaAnnotation))
2535     return 50;
2536 
2537   if (Left.is(TT_UnaryOperator))
2538     return 60;
2539   if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2540       Left.Previous->isLabelString() &&
2541       (Left.NextOperator || Left.OperatorIndex != 0))
2542     return 50;
2543   if (Right.is(tok::plus) && Left.isLabelString() &&
2544       (Right.NextOperator || Right.OperatorIndex != 0))
2545     return 25;
2546   if (Left.is(tok::comma))
2547     return 1;
2548   if (Right.is(tok::lessless) && Left.isLabelString() &&
2549       (Right.NextOperator || Right.OperatorIndex != 1))
2550     return 25;
2551   if (Right.is(tok::lessless)) {
2552     // Breaking at a << is really cheap.
2553     if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2554       // Slightly prefer to break before the first one in log-like statements.
2555       return 2;
2556     return 1;
2557   }
2558   if (Left.ClosesTemplateDeclaration)
2559     return Style.PenaltyBreakTemplateDeclaration;
2560   if (Left.is(TT_ConditionalExpr))
2561     return prec::Conditional;
2562   prec::Level Level = Left.getPrecedence();
2563   if (Level == prec::Unknown)
2564     Level = Right.getPrecedence();
2565   if (Level == prec::Assignment)
2566     return Style.PenaltyBreakAssignment;
2567   if (Level != prec::Unknown)
2568     return Level;
2569 
2570   return 3;
2571 }
2572 
2573 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const {
2574   return Style.SpaceBeforeParens == FormatStyle::SBPO_Always ||
2575          (Style.SpaceBeforeParens == FormatStyle::SBPO_NonEmptyParentheses &&
2576           Right.ParameterCount > 0);
2577 }
2578 
2579 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2580                                           const FormatToken &Left,
2581                                           const FormatToken &Right) {
2582   if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2583     return true;
2584   if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java)
2585     return true;
2586   if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2587       Left.Tok.getObjCKeywordID() == tok::objc_property)
2588     return true;
2589   if (Right.is(tok::hashhash))
2590     return Left.is(tok::hash);
2591   if (Left.isOneOf(tok::hashhash, tok::hash))
2592     return Right.is(tok::hash);
2593   if ((Left.is(tok::l_paren) && Right.is(tok::r_paren)) ||
2594       (Left.is(tok::l_brace) && Left.BlockKind != BK_Block &&
2595        Right.is(tok::r_brace) && Right.BlockKind != BK_Block))
2596     return Style.SpaceInEmptyParentheses;
2597   if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2598     return (Right.is(TT_CastRParen) ||
2599             (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2600                ? Style.SpacesInCStyleCastParentheses
2601                : Style.SpacesInParentheses;
2602   if (Right.isOneOf(tok::semi, tok::comma))
2603     return false;
2604   if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) {
2605     bool IsLightweightGeneric = Right.MatchingParen &&
2606                                 Right.MatchingParen->Next &&
2607                                 Right.MatchingParen->Next->is(tok::colon);
2608     return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList;
2609   }
2610   if (Right.is(tok::less) && Left.is(tok::kw_template))
2611     return Style.SpaceAfterTemplateKeyword;
2612   if (Left.isOneOf(tok::exclaim, tok::tilde))
2613     return false;
2614   if (Left.is(tok::at) &&
2615       Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2616                     tok::numeric_constant, tok::l_paren, tok::l_brace,
2617                     tok::kw_true, tok::kw_false))
2618     return false;
2619   if (Left.is(tok::colon))
2620     return !Left.is(TT_ObjCMethodExpr);
2621   if (Left.is(tok::coloncolon))
2622     return false;
2623   if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) {
2624     if (Style.Language == FormatStyle::LK_TextProto ||
2625         (Style.Language == FormatStyle::LK_Proto &&
2626          (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) {
2627       // Format empty list as `<>`.
2628       if (Left.is(tok::less) && Right.is(tok::greater))
2629         return false;
2630       return !Style.Cpp11BracedListStyle;
2631     }
2632     return false;
2633   }
2634   if (Right.is(tok::ellipsis))
2635     return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2636                                     Left.Previous->is(tok::kw_case));
2637   if (Left.is(tok::l_square) && Right.is(tok::amp))
2638     return Style.SpacesInSquareBrackets;
2639   if (Right.is(TT_PointerOrReference)) {
2640     if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) {
2641       if (!Left.MatchingParen)
2642         return true;
2643       FormatToken *TokenBeforeMatchingParen =
2644           Left.MatchingParen->getPreviousNonComment();
2645       if (!TokenBeforeMatchingParen ||
2646           !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype,
2647                                              TT_TypenameMacro))
2648         return true;
2649     }
2650     return (Left.Tok.isLiteral() ||
2651             (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2652              (Style.PointerAlignment != FormatStyle::PAS_Left ||
2653               (Line.IsMultiVariableDeclStmt &&
2654                (Left.NestingLevel == 0 ||
2655                 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2656   }
2657   if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2658       (!Left.is(TT_PointerOrReference) ||
2659        (Style.PointerAlignment != FormatStyle::PAS_Right &&
2660         !Line.IsMultiVariableDeclStmt)))
2661     return true;
2662   if (Left.is(TT_PointerOrReference))
2663     return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2664            (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2665             !Right.is(TT_StartOfName)) ||
2666            (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) ||
2667            (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2668                            tok::l_paren) &&
2669             (Style.PointerAlignment != FormatStyle::PAS_Right &&
2670              !Line.IsMultiVariableDeclStmt) &&
2671             Left.Previous &&
2672             !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon,
2673                                     tok::l_square));
2674   if (Right.is(tok::star) && Left.is(tok::l_paren))
2675     return false;
2676   const auto SpaceRequiredForArrayInitializerLSquare =
2677       [](const FormatToken &LSquareTok, const FormatStyle &Style) {
2678         return Style.SpacesInContainerLiterals ||
2679                ((Style.Language == FormatStyle::LK_Proto ||
2680                  Style.Language == FormatStyle::LK_TextProto) &&
2681                 !Style.Cpp11BracedListStyle &&
2682                 LSquareTok.endsSequence(tok::l_square, tok::colon,
2683                                         TT_SelectorName));
2684       };
2685   if (Left.is(tok::l_square))
2686     return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) &&
2687             SpaceRequiredForArrayInitializerLSquare(Left, Style)) ||
2688            (Left.isOneOf(TT_ArraySubscriptLSquare, TT_StructuredBindingLSquare,
2689                          TT_LambdaLSquare) &&
2690             Style.SpacesInSquareBrackets && Right.isNot(tok::r_square));
2691   if (Right.is(tok::r_square))
2692     return Right.MatchingParen &&
2693            ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) &&
2694              SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen,
2695                                                      Style)) ||
2696             (Style.SpacesInSquareBrackets &&
2697              Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare,
2698                                           TT_StructuredBindingLSquare,
2699                                           TT_LambdaLSquare)) ||
2700             Right.MatchingParen->is(TT_AttributeParen));
2701   if (Right.is(tok::l_square) &&
2702       !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2703                      TT_DesignatedInitializerLSquare,
2704                      TT_StructuredBindingLSquare, TT_AttributeSquare) &&
2705       !Left.isOneOf(tok::numeric_constant, TT_DictLiteral))
2706     return false;
2707   if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2708     return !Left.Children.empty(); // No spaces in "{}".
2709   if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) ||
2710       (Right.is(tok::r_brace) && Right.MatchingParen &&
2711        Right.MatchingParen->BlockKind != BK_Block))
2712     return Style.Cpp11BracedListStyle ? Style.SpacesInParentheses : true;
2713   if (Left.is(TT_BlockComment))
2714     // No whitespace in x(/*foo=*/1), except for JavaScript.
2715     return Style.Language == FormatStyle::LK_JavaScript ||
2716            !Left.TokenText.endswith("=*/");
2717   if (Right.is(tok::l_paren)) {
2718     if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) ||
2719         (Left.is(tok::r_square) && Left.is(TT_AttributeSquare)))
2720       return true;
2721     return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
2722            (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
2723             (Left.isOneOf(tok::pp_elif, tok::kw_for, tok::kw_while,
2724                           tok::kw_switch, tok::kw_case, TT_ForEachMacro,
2725                           TT_ObjCForIn) ||
2726              Left.isIf(Line.Type != LT_PreprocessorDirective) ||
2727              (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
2728                            tok::kw_new, tok::kw_delete) &&
2729               (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
2730            (spaceRequiredBeforeParens(Right) &&
2731             (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
2732              Left.is(tok::r_paren) || Left.isSimpleTypeSpecifier() ||
2733              (Left.is(tok::r_square) && Left.MatchingParen &&
2734               Left.MatchingParen->is(TT_LambdaLSquare))) &&
2735             Line.Type != LT_PreprocessorDirective);
2736   }
2737   if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
2738     return false;
2739   if (Right.is(TT_UnaryOperator))
2740     return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
2741            (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
2742   if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
2743                     tok::r_paren) ||
2744        Left.isSimpleTypeSpecifier()) &&
2745       Right.is(tok::l_brace) && Right.getNextNonComment() &&
2746       Right.BlockKind != BK_Block)
2747     return false;
2748   if (Left.is(tok::period) || Right.is(tok::period))
2749     return false;
2750   if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
2751     return false;
2752   if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
2753       Left.MatchingParen->Previous &&
2754       (Left.MatchingParen->Previous->is(tok::period) ||
2755        Left.MatchingParen->Previous->is(tok::coloncolon)))
2756     // Java call to generic function with explicit type:
2757     // A.<B<C<...>>>DoSomething();
2758     // A::<B<C<...>>>DoSomething();  // With a Java 8 method reference.
2759     return false;
2760   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
2761     return false;
2762   if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at))
2763     // Objective-C dictionary literal -> no space after opening brace.
2764     return false;
2765   if (Right.is(tok::r_brace) && Right.MatchingParen &&
2766       Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at))
2767     // Objective-C dictionary literal -> no space before closing brace.
2768     return false;
2769   if (Right.Type == TT_TrailingAnnotation &&
2770       Right.isOneOf(tok::amp, tok::ampamp) &&
2771       Left.isOneOf(tok::kw_const, tok::kw_volatile) &&
2772       (!Right.Next || Right.Next->is(tok::semi)))
2773     // Match const and volatile ref-qualifiers without any additional
2774     // qualifiers such as
2775     // void Fn() const &;
2776     return Style.PointerAlignment != FormatStyle::PAS_Left;
2777   return true;
2778 }
2779 
2780 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
2781                                          const FormatToken &Right) {
2782   const FormatToken &Left = *Right.Previous;
2783   if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
2784     return true; // Never ever merge two identifiers.
2785   if (Style.isCpp()) {
2786     if (Left.is(tok::kw_operator))
2787       return Right.is(tok::coloncolon);
2788     if (Right.is(tok::l_brace) && Right.BlockKind == BK_BracedInit &&
2789         !Left.opensScope() && Style.SpaceBeforeCpp11BracedList)
2790       return true;
2791   } else if (Style.Language == FormatStyle::LK_Proto ||
2792              Style.Language == FormatStyle::LK_TextProto) {
2793     if (Right.is(tok::period) &&
2794         Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
2795                      Keywords.kw_repeated, Keywords.kw_extend))
2796       return true;
2797     if (Right.is(tok::l_paren) &&
2798         Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
2799       return true;
2800     if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName))
2801       return true;
2802     // Slashes occur in text protocol extension syntax: [type/type] { ... }.
2803     if (Left.is(tok::slash) || Right.is(tok::slash))
2804       return false;
2805     if (Left.MatchingParen &&
2806         Left.MatchingParen->is(TT_ProtoExtensionLSquare) &&
2807         Right.isOneOf(tok::l_brace, tok::less))
2808       return !Style.Cpp11BracedListStyle;
2809     // A percent is probably part of a formatting specification, such as %lld.
2810     if (Left.is(tok::percent))
2811       return false;
2812     // Preserve the existence of a space before a percent for cases like 0x%04x
2813     // and "%d %d"
2814     if (Left.is(tok::numeric_constant) && Right.is(tok::percent))
2815       return Right.WhitespaceRange.getEnd() != Right.WhitespaceRange.getBegin();
2816   } else if (Style.isCSharp()) {
2817     // space between type and variable e.g. Dictionary<string,string> foo;
2818     if (Left.is(TT_TemplateCloser) && Right.is(TT_StartOfName))
2819       return true;
2820     // space between keywords and paren e.g. "using ("
2821     if (Right.is(tok::l_paren))
2822       if (Left.is(tok::kw_using))
2823         return spaceRequiredBeforeParens(Left);
2824   } else if (Style.Language == FormatStyle::LK_JavaScript) {
2825     if (Left.is(TT_JsFatArrow))
2826       return true;
2827     // for await ( ...
2828     if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous &&
2829         Left.Previous->is(tok::kw_for))
2830       return true;
2831     if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
2832         Right.MatchingParen) {
2833       const FormatToken *Next = Right.MatchingParen->getNextNonComment();
2834       // An async arrow function, for example: `x = async () => foo();`,
2835       // as opposed to calling a function called async: `x = async();`
2836       if (Next && Next->is(TT_JsFatArrow))
2837         return true;
2838     }
2839     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2840         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2841       return false;
2842     // In tagged template literals ("html`bar baz`"), there is no space between
2843     // the tag identifier and the template string. getIdentifierInfo makes sure
2844     // that the identifier is not a pseudo keyword like `yield`, either.
2845     if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) &&
2846         Right.is(TT_TemplateString))
2847       return false;
2848     if (Right.is(tok::star) &&
2849         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2850       return false;
2851     if (Right.isOneOf(tok::l_brace, tok::l_square) &&
2852         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield,
2853                      Keywords.kw_extends, Keywords.kw_implements))
2854       return true;
2855     if (Right.is(tok::l_paren)) {
2856       // JS methods can use some keywords as names (e.g. `delete()`).
2857       if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo())
2858         return false;
2859       // Valid JS method names can include keywords, e.g. `foo.delete()` or
2860       // `bar.instanceof()`. Recognize call positions by preceding period.
2861       if (Left.Previous && Left.Previous->is(tok::period) &&
2862           Left.Tok.getIdentifierInfo())
2863         return false;
2864       // Additional unary JavaScript operators that need a space after.
2865       if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof,
2866                        tok::kw_void))
2867         return true;
2868     }
2869     // `foo as const;` casts into a const type.
2870     if (Left.endsSequence(tok::kw_const, Keywords.kw_as)) {
2871       return false;
2872     }
2873     if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
2874                       tok::kw_const) ||
2875          // "of" is only a keyword if it appears after another identifier
2876          // (e.g. as "const x of y" in a for loop), or after a destructuring
2877          // operation (const [x, y] of z, const {a, b} of c).
2878          (Left.is(Keywords.kw_of) && Left.Previous &&
2879           (Left.Previous->Tok.is(tok::identifier) ||
2880            Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) &&
2881         (!Left.Previous || !Left.Previous->is(tok::period)))
2882       return true;
2883     if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
2884         Left.Previous->is(tok::period) && Right.is(tok::l_paren))
2885       return false;
2886     if (Left.is(Keywords.kw_as) &&
2887         Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
2888       return true;
2889     if (Left.is(tok::kw_default) && Left.Previous &&
2890         Left.Previous->is(tok::kw_export))
2891       return true;
2892     if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
2893       return true;
2894     if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
2895       return false;
2896     if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
2897       return false;
2898     if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
2899         Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
2900       return false;
2901     if (Left.is(tok::ellipsis))
2902       return false;
2903     if (Left.is(TT_TemplateCloser) &&
2904         !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
2905                        Keywords.kw_implements, Keywords.kw_extends))
2906       // Type assertions ('<type>expr') are not followed by whitespace. Other
2907       // locations that should have whitespace following are identified by the
2908       // above set of follower tokens.
2909       return false;
2910     if (Right.is(TT_JsNonNullAssertion))
2911       return false;
2912     if (Left.is(TT_JsNonNullAssertion) &&
2913         Right.isOneOf(Keywords.kw_as, Keywords.kw_in))
2914       return true; // "x! as string", "x! in y"
2915   } else if (Style.Language == FormatStyle::LK_Java) {
2916     if (Left.is(tok::r_square) && Right.is(tok::l_brace))
2917       return true;
2918     if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
2919       return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
2920     if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
2921                       tok::kw_protected) ||
2922          Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
2923                       Keywords.kw_native)) &&
2924         Right.is(TT_TemplateOpener))
2925       return true;
2926   }
2927   if (Left.is(TT_ImplicitStringLiteral))
2928     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2929   if (Line.Type == LT_ObjCMethodDecl) {
2930     if (Left.is(TT_ObjCMethodSpecifier))
2931       return true;
2932     if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right))
2933       // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a
2934       // keyword in Objective-C, and '+ (instancetype)new;' is a standard class
2935       // method declaration.
2936       return false;
2937   }
2938   if (Line.Type == LT_ObjCProperty &&
2939       (Right.is(tok::equal) || Left.is(tok::equal)))
2940     return false;
2941 
2942   if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
2943       Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
2944     return true;
2945   if (Right.is(TT_OverloadedOperatorLParen))
2946     return spaceRequiredBeforeParens(Right);
2947   if (Left.is(tok::comma))
2948     return true;
2949   if (Right.is(tok::comma))
2950     return false;
2951   if (Right.is(TT_ObjCBlockLParen))
2952     return true;
2953   if (Right.is(TT_CtorInitializerColon))
2954     return Style.SpaceBeforeCtorInitializerColon;
2955   if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon)
2956     return false;
2957   if (Right.is(TT_RangeBasedForLoopColon) &&
2958       !Style.SpaceBeforeRangeBasedForLoopColon)
2959     return false;
2960   if (Right.is(tok::colon)) {
2961     if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
2962         !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
2963       return false;
2964     if (Right.is(TT_ObjCMethodExpr))
2965       return false;
2966     if (Left.is(tok::question))
2967       return false;
2968     if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
2969       return false;
2970     if (Right.is(TT_DictLiteral))
2971       return Style.SpacesInContainerLiterals;
2972     if (Right.is(TT_AttributeColon))
2973       return false;
2974     return true;
2975   }
2976   if (Left.is(TT_UnaryOperator)) {
2977     // The alternative operators for ~ and ! are "compl" and "not".
2978     // If they are used instead, we do not want to combine them with
2979     // the token to the right, unless that is a left paren.
2980     if (!Right.is(tok::l_paren)) {
2981       if (Left.is(tok::exclaim) && Left.TokenText == "not")
2982         return true;
2983       if (Left.is(tok::tilde) && Left.TokenText == "compl")
2984         return true;
2985     }
2986     return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) ||
2987            Right.is(TT_BinaryOperator);
2988   }
2989 
2990   // If the next token is a binary operator or a selector name, we have
2991   // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
2992   if (Left.is(TT_CastRParen))
2993     return Style.SpaceAfterCStyleCast ||
2994            Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
2995 
2996   if (Left.is(tok::greater) && Right.is(tok::greater)) {
2997     if (Style.Language == FormatStyle::LK_TextProto ||
2998         (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral)))
2999       return !Style.Cpp11BracedListStyle;
3000     return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
3001            (Style.Standard < FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
3002   }
3003   if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) ||
3004       Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
3005       (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod)))
3006     return false;
3007   if (!Style.SpaceBeforeAssignmentOperators && Left.isNot(TT_TemplateCloser) &&
3008       Right.getPrecedence() == prec::Assignment)
3009     return false;
3010   if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) &&
3011       (Left.is(tok::identifier) || Left.is(tok::kw_this)))
3012     return false;
3013   if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
3014     // Generally don't remove existing spaces between an identifier and "::".
3015     // The identifier might actually be a macro name such as ALWAYS_INLINE. If
3016     // this turns out to be too lenient, add analysis of the identifier itself.
3017     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
3018   if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment))
3019     return (Left.is(TT_TemplateOpener) &&
3020             Style.Standard < FormatStyle::LS_Cpp11) ||
3021            !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
3022                           tok::kw___super, TT_TemplateCloser,
3023                           TT_TemplateOpener)) ||
3024            (Left.is(tok ::l_paren) && Style.SpacesInParentheses);
3025   if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
3026     return Style.SpacesInAngles;
3027   // Space before TT_StructuredBindingLSquare.
3028   if (Right.is(TT_StructuredBindingLSquare))
3029     return !Left.isOneOf(tok::amp, tok::ampamp) ||
3030            Style.PointerAlignment != FormatStyle::PAS_Right;
3031   // Space before & or && following a TT_StructuredBindingLSquare.
3032   if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) &&
3033       Right.isOneOf(tok::amp, tok::ampamp))
3034     return Style.PointerAlignment != FormatStyle::PAS_Left;
3035   if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
3036       (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
3037        !Right.is(tok::r_paren)))
3038     return true;
3039   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
3040       Right.isNot(TT_FunctionTypeLParen))
3041     return spaceRequiredBeforeParens(Right);
3042   if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
3043       Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
3044     return false;
3045   if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
3046       Line.startsWith(tok::hash))
3047     return true;
3048   if (Right.is(TT_TrailingUnaryOperator))
3049     return false;
3050   if (Left.is(TT_RegexLiteral))
3051     return false;
3052   return spaceRequiredBetween(Line, Left, Right);
3053 }
3054 
3055 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
3056 static bool isAllmanBrace(const FormatToken &Tok) {
3057   return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block &&
3058          !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral);
3059 }
3060 
3061 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
3062                                      const FormatToken &Right) {
3063   const FormatToken &Left = *Right.Previous;
3064   if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
3065     return true;
3066 
3067   if (Style.Language == FormatStyle::LK_JavaScript) {
3068     // FIXME: This might apply to other languages and token kinds.
3069     if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
3070         Left.Previous->is(tok::string_literal))
3071       return true;
3072     if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
3073         Left.Previous && Left.Previous->is(tok::equal) &&
3074         Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
3075                             tok::kw_const) &&
3076         // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
3077         // above.
3078         !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
3079       // Object literals on the top level of a file are treated as "enum-style".
3080       // Each key/value pair is put on a separate line, instead of bin-packing.
3081       return true;
3082     if (Left.is(tok::l_brace) && Line.Level == 0 &&
3083         (Line.startsWith(tok::kw_enum) ||
3084          Line.startsWith(tok::kw_const, tok::kw_enum) ||
3085          Line.startsWith(tok::kw_export, tok::kw_enum) ||
3086          Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum)))
3087       // JavaScript top-level enum key/value pairs are put on separate lines
3088       // instead of bin-packing.
3089       return true;
3090     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
3091         !Left.Children.empty())
3092       // Support AllowShortFunctionsOnASingleLine for JavaScript.
3093       return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
3094              Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
3095              (Left.NestingLevel == 0 && Line.Level == 0 &&
3096               Style.AllowShortFunctionsOnASingleLine &
3097                   FormatStyle::SFS_InlineOnly);
3098   } else if (Style.Language == FormatStyle::LK_Java) {
3099     if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
3100         Right.Next->is(tok::string_literal))
3101       return true;
3102   } else if (Style.Language == FormatStyle::LK_Cpp ||
3103              Style.Language == FormatStyle::LK_ObjC ||
3104              Style.Language == FormatStyle::LK_Proto ||
3105              Style.Language == FormatStyle::LK_TableGen ||
3106              Style.Language == FormatStyle::LK_TextProto) {
3107     if (Left.isStringLiteral() && Right.isStringLiteral())
3108       return true;
3109   }
3110 
3111   // If the last token before a '}', ']', or ')' is a comma or a trailing
3112   // comment, the intention is to insert a line break after it in order to make
3113   // shuffling around entries easier. Import statements, especially in
3114   // JavaScript, can be an exception to this rule.
3115   if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) {
3116     const FormatToken *BeforeClosingBrace = nullptr;
3117     if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
3118          (Style.Language == FormatStyle::LK_JavaScript &&
3119           Left.is(tok::l_paren))) &&
3120         Left.BlockKind != BK_Block && Left.MatchingParen)
3121       BeforeClosingBrace = Left.MatchingParen->Previous;
3122     else if (Right.MatchingParen &&
3123              (Right.MatchingParen->isOneOf(tok::l_brace,
3124                                            TT_ArrayInitializerLSquare) ||
3125               (Style.Language == FormatStyle::LK_JavaScript &&
3126                Right.MatchingParen->is(tok::l_paren))))
3127       BeforeClosingBrace = &Left;
3128     if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
3129                                BeforeClosingBrace->isTrailingComment()))
3130       return true;
3131   }
3132 
3133   if (Right.is(tok::comment))
3134     return Left.BlockKind != BK_BracedInit &&
3135            Left.isNot(TT_CtorInitializerColon) &&
3136            (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
3137   if (Left.isTrailingComment())
3138     return true;
3139   if (Right.Previous->IsUnterminatedLiteral)
3140     return true;
3141   if (Right.is(tok::lessless) && Right.Next &&
3142       Right.Previous->is(tok::string_literal) &&
3143       Right.Next->is(tok::string_literal))
3144     return true;
3145   if (Right.Previous->ClosesTemplateDeclaration &&
3146       Right.Previous->MatchingParen &&
3147       Right.Previous->MatchingParen->NestingLevel == 0 &&
3148       Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes)
3149     return true;
3150   if (Right.is(TT_CtorInitializerComma) &&
3151       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
3152       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
3153     return true;
3154   if (Right.is(TT_CtorInitializerColon) &&
3155       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
3156       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
3157     return true;
3158   // Break only if we have multiple inheritance.
3159   if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma &&
3160       Right.is(TT_InheritanceComma))
3161     return true;
3162   if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
3163     // Multiline raw string literals are special wrt. line breaks. The author
3164     // has made a deliberate choice and might have aligned the contents of the
3165     // string literal accordingly. Thus, we try keep existing line breaks.
3166     return Right.IsMultiline && Right.NewlinesBefore > 0;
3167   if ((Right.Previous->is(tok::l_brace) ||
3168        (Right.Previous->is(tok::less) && Right.Previous->Previous &&
3169         Right.Previous->Previous->is(tok::equal))) &&
3170       Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) {
3171     // Don't put enums or option definitions onto single lines in protocol
3172     // buffers.
3173     return true;
3174   }
3175   if (Right.is(TT_InlineASMBrace))
3176     return Right.HasUnescapedNewline;
3177   if (isAllmanBrace(Left) || isAllmanBrace(Right))
3178     return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
3179            (Line.startsWith(tok::kw_typedef, tok::kw_enum) &&
3180             Style.BraceWrapping.AfterEnum) ||
3181            (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
3182            (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
3183   if (Left.is(TT_ObjCBlockLBrace) &&
3184       Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never)
3185     return true;
3186 
3187   if (Left.is(TT_LambdaLBrace)) {
3188     if (Left.MatchingParen && Left.MatchingParen->Next &&
3189         Left.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren) &&
3190         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline)
3191       return false;
3192 
3193     if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None ||
3194         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline ||
3195         (!Left.Children.empty() &&
3196          Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty))
3197       return true;
3198   }
3199 
3200   // Put multiple C# attributes on a new line.
3201   if (Style.isCSharp() &&
3202       ((Left.is(TT_AttributeSquare) && Left.is(tok::r_square)) ||
3203        (Left.is(tok::r_square) && Right.is(TT_AttributeSquare) &&
3204         Right.is(tok::l_square))))
3205     return true;
3206 
3207   // Put multiple Java annotation on a new line.
3208   if ((Style.Language == FormatStyle::LK_Java ||
3209        Style.Language == FormatStyle::LK_JavaScript) &&
3210       Left.is(TT_LeadingJavaAnnotation) &&
3211       Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
3212       (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
3213     return true;
3214 
3215   if (Right.is(TT_ProtoExtensionLSquare))
3216     return true;
3217 
3218   // In text proto instances if a submessage contains at least 2 entries and at
3219   // least one of them is a submessage, like A { ... B { ... } ... },
3220   // put all of the entries of A on separate lines by forcing the selector of
3221   // the submessage B to be put on a newline.
3222   //
3223   // Example: these can stay on one line:
3224   // a { scalar_1: 1 scalar_2: 2 }
3225   // a { b { key: value } }
3226   //
3227   // and these entries need to be on a new line even if putting them all in one
3228   // line is under the column limit:
3229   // a {
3230   //   scalar: 1
3231   //   b { key: value }
3232   // }
3233   //
3234   // We enforce this by breaking before a submessage field that has previous
3235   // siblings, *and* breaking before a field that follows a submessage field.
3236   //
3237   // Be careful to exclude the case  [proto.ext] { ... } since the `]` is
3238   // the TT_SelectorName there, but we don't want to break inside the brackets.
3239   //
3240   // Another edge case is @submessage { key: value }, which is a common
3241   // substitution placeholder. In this case we want to keep `@` and `submessage`
3242   // together.
3243   //
3244   // We ensure elsewhere that extensions are always on their own line.
3245   if ((Style.Language == FormatStyle::LK_Proto ||
3246        Style.Language == FormatStyle::LK_TextProto) &&
3247       Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) {
3248     // Keep `@submessage` together in:
3249     // @submessage { key: value }
3250     if (Right.Previous && Right.Previous->is(tok::at))
3251       return false;
3252     // Look for the scope opener after selector in cases like:
3253     // selector { ...
3254     // selector: { ...
3255     // selector: @base { ...
3256     FormatToken *LBrace = Right.Next;
3257     if (LBrace && LBrace->is(tok::colon)) {
3258       LBrace = LBrace->Next;
3259       if (LBrace && LBrace->is(tok::at)) {
3260         LBrace = LBrace->Next;
3261         if (LBrace)
3262           LBrace = LBrace->Next;
3263       }
3264     }
3265     if (LBrace &&
3266         // The scope opener is one of {, [, <:
3267         // selector { ... }
3268         // selector [ ... ]
3269         // selector < ... >
3270         //
3271         // In case of selector { ... }, the l_brace is TT_DictLiteral.
3272         // In case of an empty selector {}, the l_brace is not TT_DictLiteral,
3273         // so we check for immediately following r_brace.
3274         ((LBrace->is(tok::l_brace) &&
3275           (LBrace->is(TT_DictLiteral) ||
3276            (LBrace->Next && LBrace->Next->is(tok::r_brace)))) ||
3277          LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) {
3278       // If Left.ParameterCount is 0, then this submessage entry is not the
3279       // first in its parent submessage, and we want to break before this entry.
3280       // If Left.ParameterCount is greater than 0, then its parent submessage
3281       // might contain 1 or more entries and we want to break before this entry
3282       // if it contains at least 2 entries. We deal with this case later by
3283       // detecting and breaking before the next entry in the parent submessage.
3284       if (Left.ParameterCount == 0)
3285         return true;
3286       // However, if this submessage is the first entry in its parent
3287       // submessage, Left.ParameterCount might be 1 in some cases.
3288       // We deal with this case later by detecting an entry
3289       // following a closing paren of this submessage.
3290     }
3291 
3292     // If this is an entry immediately following a submessage, it will be
3293     // preceded by a closing paren of that submessage, like in:
3294     //     left---.  .---right
3295     //            v  v
3296     // sub: { ... } key: value
3297     // If there was a comment between `}` an `key` above, then `key` would be
3298     // put on a new line anyways.
3299     if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square))
3300       return true;
3301   }
3302 
3303   // Deal with lambda arguments in C++ - we want consistent line breaks whether
3304   // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced
3305   // as aggressive line breaks are placed when the lambda is not the last arg.
3306   if ((Style.Language == FormatStyle::LK_Cpp ||
3307        Style.Language == FormatStyle::LK_ObjC) &&
3308       Left.is(tok::l_paren) && Left.BlockParameterCount > 0 &&
3309       !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) {
3310     // Multiple lambdas in the same function call force line breaks.
3311     if (Left.BlockParameterCount > 1)
3312       return true;
3313 
3314     // A lambda followed by another arg forces a line break.
3315     if (!Left.Role)
3316       return false;
3317     auto Comma = Left.Role->lastComma();
3318     if (!Comma)
3319       return false;
3320     auto Next = Comma->getNextNonComment();
3321     if (!Next)
3322       return false;
3323     if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret))
3324       return true;
3325   }
3326 
3327   return false;
3328 }
3329 
3330 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
3331                                     const FormatToken &Right) {
3332   const FormatToken &Left = *Right.Previous;
3333 
3334   // Language-specific stuff.
3335   if (Style.Language == FormatStyle::LK_Java) {
3336     if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3337                      Keywords.kw_implements))
3338       return false;
3339     if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3340                       Keywords.kw_implements))
3341       return true;
3342   } else if (Style.Language == FormatStyle::LK_JavaScript) {
3343     const FormatToken *NonComment = Right.getPreviousNonComment();
3344     if (NonComment &&
3345         NonComment->isOneOf(
3346             tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break,
3347             tok::kw_throw, Keywords.kw_interface, Keywords.kw_type,
3348             tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected,
3349             Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get,
3350             Keywords.kw_set, Keywords.kw_async, Keywords.kw_await))
3351       return false; // Otherwise automatic semicolon insertion would trigger.
3352     if (Right.NestingLevel == 0 &&
3353         (Left.Tok.getIdentifierInfo() ||
3354          Left.isOneOf(tok::r_square, tok::r_paren)) &&
3355         Right.isOneOf(tok::l_square, tok::l_paren))
3356       return false; // Otherwise automatic semicolon insertion would trigger.
3357     if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
3358       return false;
3359     if (Left.is(TT_JsTypeColon))
3360       return true;
3361     // Don't wrap between ":" and "!" of a strict prop init ("field!: type;").
3362     if (Left.is(tok::exclaim) && Right.is(tok::colon))
3363       return false;
3364     // Look for is type annotations like:
3365     // function f(): a is B { ... }
3366     // Do not break before is in these cases.
3367     if (Right.is(Keywords.kw_is)) {
3368       const FormatToken *Next = Right.getNextNonComment();
3369       // If `is` is followed by a colon, it's likely that it's a dict key, so
3370       // ignore it for this check.
3371       // For example this is common in Polymer:
3372       // Polymer({
3373       //   is: 'name',
3374       //   ...
3375       // });
3376       if (!Next || !Next->is(tok::colon))
3377         return false;
3378     }
3379     if (Left.is(Keywords.kw_in))
3380       return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
3381     if (Right.is(Keywords.kw_in))
3382       return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
3383     if (Right.is(Keywords.kw_as))
3384       return false; // must not break before as in 'x as type' casts
3385     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) {
3386       // extends and infer can appear as keywords in conditional types:
3387       //   https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types
3388       // do not break before them, as the expressions are subject to ASI.
3389       return false;
3390     }
3391     if (Left.is(Keywords.kw_as))
3392       return true;
3393     if (Left.is(TT_JsNonNullAssertion))
3394       return true;
3395     if (Left.is(Keywords.kw_declare) &&
3396         Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
3397                       Keywords.kw_function, tok::kw_class, tok::kw_enum,
3398                       Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
3399                       Keywords.kw_let, tok::kw_const))
3400       // See grammar for 'declare' statements at:
3401       // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
3402       return false;
3403     if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
3404         Right.isOneOf(tok::identifier, tok::string_literal))
3405       return false; // must not break in "module foo { ...}"
3406     if (Right.is(TT_TemplateString) && Right.closesScope())
3407       return false;
3408     // Don't split tagged template literal so there is a break between the tag
3409     // identifier and template string.
3410     if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) {
3411       return false;
3412     }
3413     if (Left.is(TT_TemplateString) && Left.opensScope())
3414       return true;
3415   }
3416 
3417   if (Left.is(tok::at))
3418     return false;
3419   if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
3420     return false;
3421   if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
3422     return !Right.is(tok::l_paren);
3423   if (Right.is(TT_PointerOrReference))
3424     return Line.IsMultiVariableDeclStmt ||
3425            (Style.PointerAlignment == FormatStyle::PAS_Right &&
3426             (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
3427   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
3428       Right.is(tok::kw_operator))
3429     return true;
3430   if (Left.is(TT_PointerOrReference))
3431     return false;
3432   if (Right.isTrailingComment())
3433     // We rely on MustBreakBefore being set correctly here as we should not
3434     // change the "binding" behavior of a comment.
3435     // The first comment in a braced lists is always interpreted as belonging to
3436     // the first list element. Otherwise, it should be placed outside of the
3437     // list.
3438     return Left.BlockKind == BK_BracedInit ||
3439            (Left.is(TT_CtorInitializerColon) &&
3440             Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon);
3441   if (Left.is(tok::question) && Right.is(tok::colon))
3442     return false;
3443   if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
3444     return Style.BreakBeforeTernaryOperators;
3445   if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
3446     return !Style.BreakBeforeTernaryOperators;
3447   if (Left.is(TT_InheritanceColon))
3448     return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon;
3449   if (Right.is(TT_InheritanceColon))
3450     return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon;
3451   if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
3452       Left.isNot(TT_SelectorName))
3453     return true;
3454 
3455   if (Right.is(tok::colon) &&
3456       !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
3457     return false;
3458   if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
3459     if (Style.Language == FormatStyle::LK_Proto ||
3460         Style.Language == FormatStyle::LK_TextProto) {
3461       if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral())
3462         return false;
3463       // Prevent cases like:
3464       //
3465       // submessage:
3466       //     { key: valueeeeeeeeeeee }
3467       //
3468       // when the snippet does not fit into one line.
3469       // Prefer:
3470       //
3471       // submessage: {
3472       //   key: valueeeeeeeeeeee
3473       // }
3474       //
3475       // instead, even if it is longer by one line.
3476       //
3477       // Note that this allows allows the "{" to go over the column limit
3478       // when the column limit is just between ":" and "{", but that does
3479       // not happen too often and alternative formattings in this case are
3480       // not much better.
3481       //
3482       // The code covers the cases:
3483       //
3484       // submessage: { ... }
3485       // submessage: < ... >
3486       // repeated: [ ... ]
3487       if (((Right.is(tok::l_brace) || Right.is(tok::less)) &&
3488            Right.is(TT_DictLiteral)) ||
3489           Right.is(TT_ArrayInitializerLSquare))
3490         return false;
3491     }
3492     return true;
3493   }
3494   if (Right.is(tok::r_square) && Right.MatchingParen &&
3495       Right.MatchingParen->is(TT_ProtoExtensionLSquare))
3496     return false;
3497   if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
3498                                     Right.Next->is(TT_ObjCMethodExpr)))
3499     return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
3500   if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
3501     return true;
3502   if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
3503     return true;
3504   if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
3505                     TT_OverloadedOperator))
3506     return false;
3507   if (Left.is(TT_RangeBasedForLoopColon))
3508     return true;
3509   if (Right.is(TT_RangeBasedForLoopColon))
3510     return false;
3511   if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
3512     return true;
3513   if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
3514       Left.is(tok::kw_operator))
3515     return false;
3516   if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
3517       Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
3518     return false;
3519   if (Left.is(tok::equal) && Right.is(tok::l_brace) &&
3520       !Style.Cpp11BracedListStyle)
3521     return false;
3522   if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen))
3523     return false;
3524   if (Left.is(tok::l_paren) && Left.Previous &&
3525       (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
3526     return false;
3527   if (Right.is(TT_ImplicitStringLiteral))
3528     return false;
3529 
3530   if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
3531     return false;
3532   if (Right.is(tok::r_square) && Right.MatchingParen &&
3533       Right.MatchingParen->is(TT_LambdaLSquare))
3534     return false;
3535 
3536   // We only break before r_brace if there was a corresponding break before
3537   // the l_brace, which is tracked by BreakBeforeClosingBrace.
3538   if (Right.is(tok::r_brace))
3539     return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block;
3540 
3541   // Allow breaking after a trailing annotation, e.g. after a method
3542   // declaration.
3543   if (Left.is(TT_TrailingAnnotation))
3544     return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
3545                           tok::less, tok::coloncolon);
3546 
3547   if (Right.is(tok::kw___attribute) ||
3548       (Right.is(tok::l_square) && Right.is(TT_AttributeSquare)))
3549     return true;
3550 
3551   if (Left.is(tok::identifier) && Right.is(tok::string_literal))
3552     return true;
3553 
3554   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
3555     return true;
3556 
3557   if (Left.is(TT_CtorInitializerColon))
3558     return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon;
3559   if (Right.is(TT_CtorInitializerColon))
3560     return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon;
3561   if (Left.is(TT_CtorInitializerComma) &&
3562       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
3563     return false;
3564   if (Right.is(TT_CtorInitializerComma) &&
3565       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
3566     return true;
3567   if (Left.is(TT_InheritanceComma) &&
3568       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
3569     return false;
3570   if (Right.is(TT_InheritanceComma) &&
3571       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
3572     return true;
3573   if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
3574       (Left.is(tok::less) && Right.is(tok::less)))
3575     return false;
3576   if (Right.is(TT_BinaryOperator) &&
3577       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
3578       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
3579        Right.getPrecedence() != prec::Assignment))
3580     return true;
3581   if (Left.is(TT_ArrayInitializerLSquare))
3582     return true;
3583   if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
3584     return true;
3585   if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
3586       !Left.isOneOf(tok::arrowstar, tok::lessless) &&
3587       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
3588       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
3589        Left.getPrecedence() == prec::Assignment))
3590     return true;
3591   if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) ||
3592       (Left.is(tok::r_square) && Right.is(TT_AttributeSquare)))
3593     return false;
3594   return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
3595                       tok::kw_class, tok::kw_struct, tok::comment) ||
3596          Right.isMemberAccess() ||
3597          Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
3598                        tok::colon, tok::l_square, tok::at) ||
3599          (Left.is(tok::r_paren) &&
3600           Right.isOneOf(tok::identifier, tok::kw_const)) ||
3601          (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
3602          (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
3603 }
3604 
3605 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
3606   llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n";
3607   const FormatToken *Tok = Line.First;
3608   while (Tok) {
3609     llvm::errs() << " M=" << Tok->MustBreakBefore
3610                  << " C=" << Tok->CanBreakBefore
3611                  << " T=" << getTokenTypeName(Tok->Type)
3612                  << " S=" << Tok->SpacesRequiredBefore
3613                  << " B=" << Tok->BlockParameterCount
3614                  << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty
3615                  << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength
3616                  << " PPK=" << Tok->PackingKind << " FakeLParens=";
3617     for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
3618       llvm::errs() << Tok->FakeLParens[i] << "/";
3619     llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
3620     llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo();
3621     llvm::errs() << " Text='" << Tok->TokenText << "'\n";
3622     if (!Tok->Next)
3623       assert(Tok == Line.Last);
3624     Tok = Tok->Next;
3625   }
3626   llvm::errs() << "----\n";
3627 }
3628 
3629 } // namespace format
3630 } // namespace clang
3631