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