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 "clang/Basic/SourceManager.h"
17 #include "llvm/ADT/SmallPtrSet.h"
18 #include "llvm/Support/Debug.h"
19 
20 #define DEBUG_TYPE "format-token-annotator"
21 
22 namespace clang {
23 namespace format {
24 
25 namespace {
26 
27 /// Returns \c true if the token can be used as an identifier in
28 /// an Objective-C \c @selector, \c false otherwise.
29 ///
30 /// Because getFormattingLangOpts() always lexes source code as
31 /// Objective-C++, C++ keywords like \c new and \c delete are
32 /// lexed as tok::kw_*, not tok::identifier, even for Objective-C.
33 ///
34 /// For Objective-C and Objective-C++, both identifiers and keywords
35 /// are valid inside @selector(...) (or a macro which
36 /// invokes @selector(...)). So, we allow treat any identifier or
37 /// keyword as a potential Objective-C selector component.
38 static bool canBeObjCSelectorComponent(const FormatToken &Tok) {
39   return Tok.Tok.getIdentifierInfo() != nullptr;
40 }
41 
42 /// A parser that gathers additional information about tokens.
43 ///
44 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
45 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
46 /// into template parameter lists.
47 class AnnotatingParser {
48 public:
49   AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
50                    const AdditionalKeywords &Keywords)
51       : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
52         Keywords(Keywords) {
53     Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
54     resetTokenMetadata(CurrentToken);
55   }
56 
57 private:
58   bool parseAngle() {
59     if (!CurrentToken || !CurrentToken->Previous)
60       return false;
61     if (NonTemplateLess.count(CurrentToken->Previous))
62       return false;
63 
64     const FormatToken &Previous = *CurrentToken->Previous;  // The '<'.
65     if (Previous.Previous) {
66       if (Previous.Previous->Tok.isLiteral())
67         return false;
68       if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
69           (!Previous.Previous->MatchingParen ||
70            !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen)))
71         return false;
72     }
73 
74     FormatToken *Left = CurrentToken->Previous;
75     Left->ParentBracket = Contexts.back().ContextKind;
76     ScopedContextCreator ContextCreator(*this, tok::less, 12);
77 
78     // If this angle is in the context of an expression, we need to be more
79     // hesitant to detect it as opening template parameters.
80     bool InExprContext = Contexts.back().IsExpression;
81 
82     Contexts.back().IsExpression = false;
83     // If there's a template keyword before the opening angle bracket, this is a
84     // template parameter, not an argument.
85     Contexts.back().InTemplateArgument =
86         Left->Previous && Left->Previous->Tok.isNot(tok::kw_template);
87 
88     if (Style.Language == FormatStyle::LK_Java &&
89         CurrentToken->is(tok::question))
90       next();
91 
92     while (CurrentToken) {
93       if (CurrentToken->is(tok::greater)) {
94         Left->MatchingParen = CurrentToken;
95         CurrentToken->MatchingParen = Left;
96         // In TT_Proto, we must distignuish between:
97         //   map<key, value>
98         //   msg < item: data >
99         //   msg: < item: data >
100         // In TT_TextProto, map<key, value> does not occur.
101         if (Style.Language == FormatStyle::LK_TextProto ||
102             (Style.Language == FormatStyle::LK_Proto && Left->Previous &&
103              Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral)))
104           CurrentToken->Type = TT_DictLiteral;
105         else
106           CurrentToken->Type = TT_TemplateCloser;
107         next();
108         return true;
109       }
110       if (CurrentToken->is(tok::question) &&
111           Style.Language == FormatStyle::LK_Java) {
112         next();
113         continue;
114       }
115       if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
116           (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext &&
117            Style.Language != FormatStyle::LK_Proto &&
118            Style.Language != FormatStyle::LK_TextProto))
119         return false;
120       // If a && or || is found and interpreted as a binary operator, this set
121       // of angles is likely part of something like "a < b && c > d". If the
122       // angles are inside an expression, the ||/&& might also be a binary
123       // operator that was misinterpreted because we are parsing template
124       // parameters.
125       // FIXME: This is getting out of hand, write a decent parser.
126       if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
127           CurrentToken->Previous->is(TT_BinaryOperator) &&
128           Contexts[Contexts.size() - 2].IsExpression &&
129           !Line.startsWith(tok::kw_template))
130         return false;
131       updateParameterCount(Left, CurrentToken);
132       if (Style.Language == FormatStyle::LK_Proto) {
133         if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) {
134           if (CurrentToken->is(tok::colon) ||
135               (CurrentToken->isOneOf(tok::l_brace, tok::less) &&
136                Previous->isNot(tok::colon)))
137             Previous->Type = TT_SelectorName;
138         }
139       }
140       if (!consumeToken())
141         return false;
142     }
143     return false;
144   }
145 
146   bool parseParens(bool LookForDecls = false) {
147     if (!CurrentToken)
148       return false;
149     FormatToken *Left = CurrentToken->Previous;
150     Left->ParentBracket = Contexts.back().ContextKind;
151     ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
152 
153     // FIXME: This is a bit of a hack. Do better.
154     Contexts.back().ColonIsForRangeExpr =
155         Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
156 
157     bool StartsObjCMethodExpr = false;
158     if (FormatToken *MaybeSel = Left->Previous) {
159       // @selector( starts a selector.
160       if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
161           MaybeSel->Previous->is(tok::at)) {
162         StartsObjCMethodExpr = true;
163       }
164     }
165 
166     if (Left->is(TT_OverloadedOperatorLParen)) {
167       Contexts.back().IsExpression = false;
168     } else if (Style.Language == FormatStyle::LK_JavaScript &&
169                (Line.startsWith(Keywords.kw_type, tok::identifier) ||
170                 Line.startsWith(tok::kw_export, Keywords.kw_type,
171                                 tok::identifier))) {
172       // type X = (...);
173       // export type X = (...);
174       Contexts.back().IsExpression = false;
175     } else if (Left->Previous &&
176                (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype,
177                                         tok::kw_if, tok::kw_while, tok::l_paren,
178                                         tok::comma) ||
179                 Left->Previous->endsSequence(tok::kw_constexpr, tok::kw_if) ||
180                 Left->Previous->is(TT_BinaryOperator))) {
181       // static_assert, if and while usually contain expressions.
182       Contexts.back().IsExpression = true;
183     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
184                (Left->Previous->is(Keywords.kw_function) ||
185                 (Left->Previous->endsSequence(tok::identifier,
186                                               Keywords.kw_function)))) {
187       // function(...) or function f(...)
188       Contexts.back().IsExpression = false;
189     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
190                Left->Previous->is(TT_JsTypeColon)) {
191       // let x: (SomeType);
192       Contexts.back().IsExpression = false;
193     } else if (Left->Previous && Left->Previous->is(tok::r_square) &&
194                Left->Previous->MatchingParen &&
195                Left->Previous->MatchingParen->is(TT_LambdaLSquare)) {
196       // This is a parameter list of a lambda expression.
197       Contexts.back().IsExpression = false;
198     } else if (Line.InPPDirective &&
199                (!Left->Previous || !Left->Previous->is(tok::identifier))) {
200       Contexts.back().IsExpression = true;
201     } else if (Contexts[Contexts.size() - 2].CaretFound) {
202       // This is the parameter list of an ObjC block.
203       Contexts.back().IsExpression = false;
204     } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) {
205       Left->Type = TT_AttributeParen;
206     } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) {
207       // The first argument to a foreach macro is a declaration.
208       Contexts.back().IsForEachMacro = true;
209       Contexts.back().IsExpression = false;
210     } else if (Left->Previous && Left->Previous->MatchingParen &&
211                Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
212       Contexts.back().IsExpression = false;
213     } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
214       bool IsForOrCatch =
215           Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch);
216       Contexts.back().IsExpression = !IsForOrCatch;
217     }
218 
219     if (StartsObjCMethodExpr) {
220       Contexts.back().ColonIsObjCMethodExpr = true;
221       Left->Type = TT_ObjCMethodExpr;
222     }
223 
224     // MightBeFunctionType and ProbablyFunctionType are used for
225     // function pointer and reference types as well as Objective-C
226     // block types:
227     //
228     // void (*FunctionPointer)(void);
229     // void (&FunctionReference)(void);
230     // void (^ObjCBlock)(void);
231     bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
232     bool ProbablyFunctionType =
233         CurrentToken->isOneOf(tok::star, tok::amp, tok::caret);
234     bool HasMultipleLines = false;
235     bool HasMultipleParametersOnALine = false;
236     bool MightBeObjCForRangeLoop =
237         Left->Previous && Left->Previous->is(tok::kw_for);
238     FormatToken *PossibleObjCForInToken = nullptr;
239     while (CurrentToken) {
240       // LookForDecls is set when "if (" has been seen. Check for
241       // 'identifier' '*' 'identifier' followed by not '=' -- this
242       // '*' has to be a binary operator but determineStarAmpUsage() will
243       // categorize it as an unary operator, so set the right type here.
244       if (LookForDecls && CurrentToken->Next) {
245         FormatToken *Prev = CurrentToken->getPreviousNonComment();
246         if (Prev) {
247           FormatToken *PrevPrev = Prev->getPreviousNonComment();
248           FormatToken *Next = CurrentToken->Next;
249           if (PrevPrev && PrevPrev->is(tok::identifier) &&
250               Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
251               CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
252             Prev->Type = TT_BinaryOperator;
253             LookForDecls = false;
254           }
255         }
256       }
257 
258       if (CurrentToken->Previous->is(TT_PointerOrReference) &&
259           CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
260                                                     tok::coloncolon))
261         ProbablyFunctionType = true;
262       if (CurrentToken->is(tok::comma))
263         MightBeFunctionType = false;
264       if (CurrentToken->Previous->is(TT_BinaryOperator))
265         Contexts.back().IsExpression = true;
266       if (CurrentToken->is(tok::r_paren)) {
267         if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next &&
268             (CurrentToken->Next->is(tok::l_paren) ||
269              (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration)))
270           Left->Type = Left->Next->is(tok::caret) ? TT_ObjCBlockLParen
271                                                   : TT_FunctionTypeLParen;
272         Left->MatchingParen = CurrentToken;
273         CurrentToken->MatchingParen = Left;
274 
275         if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
276             Left->Previous && Left->Previous->is(tok::l_paren)) {
277           // Detect the case where macros are used to generate lambdas or
278           // function bodies, e.g.:
279           //   auto my_lambda = MARCO((Type *type, int i) { .. body .. });
280           for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) {
281             if (Tok->is(TT_BinaryOperator) &&
282                 Tok->isOneOf(tok::star, tok::amp, tok::ampamp))
283               Tok->Type = TT_PointerOrReference;
284           }
285         }
286 
287         if (StartsObjCMethodExpr) {
288           CurrentToken->Type = TT_ObjCMethodExpr;
289           if (Contexts.back().FirstObjCSelectorName) {
290             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
291                 Contexts.back().LongestObjCSelectorName;
292           }
293         }
294 
295         if (Left->is(TT_AttributeParen))
296           CurrentToken->Type = TT_AttributeParen;
297         if (Left->Previous && Left->Previous->is(TT_JavaAnnotation))
298           CurrentToken->Type = TT_JavaAnnotation;
299         if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation))
300           CurrentToken->Type = TT_LeadingJavaAnnotation;
301 
302         if (!HasMultipleLines)
303           Left->PackingKind = PPK_Inconclusive;
304         else if (HasMultipleParametersOnALine)
305           Left->PackingKind = PPK_BinPacked;
306         else
307           Left->PackingKind = PPK_OnePerLine;
308 
309         next();
310         return true;
311       }
312       if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
313         return false;
314 
315       if (CurrentToken->is(tok::l_brace))
316         Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen
317       if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
318           !CurrentToken->Next->HasUnescapedNewline &&
319           !CurrentToken->Next->isTrailingComment())
320         HasMultipleParametersOnALine = true;
321       if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
322            CurrentToken->Previous->isSimpleTypeSpecifier()) &&
323           !CurrentToken->is(tok::l_brace))
324         Contexts.back().IsExpression = false;
325       if (CurrentToken->isOneOf(tok::semi, tok::colon)) {
326         MightBeObjCForRangeLoop = false;
327         if (PossibleObjCForInToken) {
328           PossibleObjCForInToken->Type = TT_Unknown;
329           PossibleObjCForInToken = nullptr;
330         }
331       }
332       if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) {
333         PossibleObjCForInToken = CurrentToken;
334         PossibleObjCForInToken->Type = TT_ObjCForIn;
335       }
336       // When we discover a 'new', we set CanBeExpression to 'false' in order to
337       // parse the type correctly. Reset that after a comma.
338       if (CurrentToken->is(tok::comma))
339         Contexts.back().CanBeExpression = true;
340 
341       FormatToken *Tok = CurrentToken;
342       if (!consumeToken())
343         return false;
344       updateParameterCount(Left, Tok);
345       if (CurrentToken && CurrentToken->HasUnescapedNewline)
346         HasMultipleLines = true;
347     }
348     return false;
349   }
350 
351   bool isCpp11AttributeSpecifier(const FormatToken &Tok) {
352     if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square))
353       return false;
354     const FormatToken *AttrTok = Tok.Next->Next;
355     if (!AttrTok)
356       return false;
357     // C++17 '[[using ns: foo, bar(baz, blech)]]'
358     // We assume nobody will name an ObjC variable 'using'.
359     if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon))
360       return true;
361     if (AttrTok->isNot(tok::identifier))
362       return false;
363     while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) {
364       // ObjC message send. We assume nobody will use : in a C++11 attribute
365       // specifier parameter, although this is technically valid:
366       // [[foo(:)]]
367       if (AttrTok->is(tok::colon) ||
368           AttrTok->startsSequence(tok::identifier, tok::identifier) ||
369           AttrTok->startsSequence(tok::r_paren, tok::identifier))
370         return false;
371       if (AttrTok->is(tok::ellipsis))
372         return true;
373       AttrTok = AttrTok->Next;
374     }
375     return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square);
376   }
377 
378   bool parseSquare() {
379     if (!CurrentToken)
380       return false;
381 
382     // A '[' could be an index subscript (after an identifier or after
383     // ')' or ']'), it could be the start of an Objective-C method
384     // expression, it could the start of an Objective-C array literal,
385     // or it could be a C++ attribute specifier [[foo::bar]].
386     FormatToken *Left = CurrentToken->Previous;
387     Left->ParentBracket = Contexts.back().ContextKind;
388     FormatToken *Parent = Left->getPreviousNonComment();
389 
390     // Cases where '>' is followed by '['.
391     // In C++, this can happen either in array of templates (foo<int>[10])
392     // or when array is a nested template type (unique_ptr<type1<type2>[]>).
393     bool CppArrayTemplates =
394         Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) &&
395         (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
396          Contexts.back().InTemplateArgument);
397 
398     bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) ||
399                                      Contexts.back().InCpp11AttributeSpecifier;
400 
401     bool InsideInlineASM = Line.startsWith(tok::kw_asm);
402     bool StartsObjCMethodExpr =
403         !InsideInlineASM && !CppArrayTemplates && Style.isCpp() &&
404         !IsCpp11AttributeSpecifier && Contexts.back().CanBeExpression &&
405         Left->isNot(TT_LambdaLSquare) &&
406         !CurrentToken->isOneOf(tok::l_brace, tok::r_square) &&
407         (!Parent ||
408          Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
409                          tok::kw_return, tok::kw_throw) ||
410          Parent->isUnaryOperator() ||
411          // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
412          Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
413          getBinOpPrecedence(Parent->Tok.getKind(), true, true) > prec::Unknown);
414     bool ColonFound = false;
415 
416     unsigned BindingIncrease = 1;
417     if (Left->isCppStructuredBinding(Style)) {
418       Left->Type = TT_StructuredBindingLSquare;
419     } else if (Left->is(TT_Unknown)) {
420       if (StartsObjCMethodExpr) {
421         Left->Type = TT_ObjCMethodExpr;
422       } else if (IsCpp11AttributeSpecifier) {
423         Left->Type = TT_AttributeSquare;
424       } else if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
425                  Contexts.back().ContextKind == tok::l_brace &&
426                  Parent->isOneOf(tok::l_brace, tok::comma)) {
427         Left->Type = TT_JsComputedPropertyName;
428       } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace &&
429                  Parent && Parent->isOneOf(tok::l_brace, tok::comma)) {
430         Left->Type = TT_DesignatedInitializerLSquare;
431       } else if (CurrentToken->is(tok::r_square) && Parent &&
432                  Parent->is(TT_TemplateCloser)) {
433         Left->Type = TT_ArraySubscriptLSquare;
434       } else if (Style.Language == FormatStyle::LK_Proto ||
435                  Style.Language == FormatStyle::LK_TextProto) {
436         // Square braces in LK_Proto can either be message field attributes:
437         //
438         // optional Aaa aaa = 1 [
439         //   (aaa) = aaa
440         // ];
441         //
442         // extensions 123 [
443         //   (aaa) = aaa
444         // ];
445         //
446         // or text proto extensions (in options):
447         //
448         // option (Aaa.options) = {
449         //   [type.type/type] {
450         //     key: value
451         //   }
452         // }
453         //
454         // or repeated fields (in options):
455         //
456         // option (Aaa.options) = {
457         //   keys: [ 1, 2, 3 ]
458         // }
459         //
460         // In the first and the third case we want to spread the contents inside
461         // the square braces; in the second we want to keep them inline.
462         Left->Type = TT_ArrayInitializerLSquare;
463         if (!Left->endsSequence(tok::l_square, tok::numeric_constant,
464                                 tok::equal) &&
465             !Left->endsSequence(tok::l_square, tok::numeric_constant,
466                                 tok::identifier) &&
467             !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) {
468           Left->Type = TT_ProtoExtensionLSquare;
469           BindingIncrease = 10;
470         }
471       } else if (!CppArrayTemplates && Parent &&
472                  Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
473                                  tok::comma, tok::l_paren, tok::l_square,
474                                  tok::question, tok::colon, tok::kw_return,
475                                  // Should only be relevant to JavaScript:
476                                  tok::kw_default)) {
477         Left->Type = TT_ArrayInitializerLSquare;
478       } else {
479         BindingIncrease = 10;
480         Left->Type = TT_ArraySubscriptLSquare;
481       }
482     }
483 
484     ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
485     Contexts.back().IsExpression = true;
486     if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
487         Parent->is(TT_JsTypeColon))
488       Contexts.back().IsExpression = false;
489 
490     Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
491     Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier;
492 
493     while (CurrentToken) {
494       if (CurrentToken->is(tok::r_square)) {
495         if (IsCpp11AttributeSpecifier)
496           CurrentToken->Type = TT_AttributeSquare;
497         else if (((CurrentToken->Next &&
498                    CurrentToken->Next->is(tok::l_paren)) ||
499                   (CurrentToken->Previous &&
500                    CurrentToken->Previous->Previous == Left)) &&
501                  Left->is(TT_ObjCMethodExpr)) {
502           // An ObjC method call is rarely followed by an open parenthesis. It
503           // also can't be composed of just one token, unless it's a macro that
504           // will be expanded to more tokens.
505           // FIXME: Do we incorrectly label ":" with this?
506           StartsObjCMethodExpr = false;
507           Left->Type = TT_Unknown;
508         }
509         if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
510           CurrentToken->Type = TT_ObjCMethodExpr;
511           // If we haven't seen a colon yet, make sure the last identifier
512           // before the r_square is tagged as a selector name component.
513           if (!ColonFound && CurrentToken->Previous &&
514               CurrentToken->Previous->is(TT_Unknown) &&
515               canBeObjCSelectorComponent(*CurrentToken->Previous))
516             CurrentToken->Previous->Type = TT_SelectorName;
517           // determineStarAmpUsage() thinks that '*' '[' is allocating an
518           // array of pointers, but if '[' starts a selector then '*' is a
519           // binary operator.
520           if (Parent && Parent->is(TT_PointerOrReference))
521             Parent->Type = TT_BinaryOperator;
522         }
523         Left->MatchingParen = CurrentToken;
524         CurrentToken->MatchingParen = Left;
525         // FirstObjCSelectorName is set when a colon is found. This does
526         // not work, however, when the method has no parameters.
527         // Here, we set FirstObjCSelectorName when the end of the method call is
528         // reached, in case it was not set already.
529         if (!Contexts.back().FirstObjCSelectorName) {
530             FormatToken* Previous = CurrentToken->getPreviousNonComment();
531             if (Previous && Previous->is(TT_SelectorName)) {
532               Previous->ObjCSelectorNameParts = 1;
533               Contexts.back().FirstObjCSelectorName = Previous;
534             }
535         } else {
536           Left->ParameterCount =
537               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
538         }
539         if (Contexts.back().FirstObjCSelectorName) {
540           Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
541               Contexts.back().LongestObjCSelectorName;
542           if (Left->BlockParameterCount > 1)
543             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
544         }
545         next();
546         return true;
547       }
548       if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
549         return false;
550       if (CurrentToken->is(tok::colon)) {
551         if (IsCpp11AttributeSpecifier &&
552             CurrentToken->endsSequence(tok::colon, tok::identifier,
553                                        tok::kw_using)) {
554           // Remember that this is a [[using ns: foo]] C++ attribute, so we
555           // don't add a space before the colon (unlike other colons).
556           CurrentToken->Type = TT_AttributeColon;
557         } else if (Left->isOneOf(TT_ArraySubscriptLSquare,
558                                  TT_DesignatedInitializerLSquare)) {
559           Left->Type = TT_ObjCMethodExpr;
560           StartsObjCMethodExpr = true;
561           Contexts.back().ColonIsObjCMethodExpr = true;
562           if (Parent && Parent->is(tok::r_paren))
563             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
564             Parent->Type = TT_CastRParen;
565         }
566         ColonFound = true;
567       }
568       if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
569           !ColonFound)
570         Left->Type = TT_ArrayInitializerLSquare;
571       FormatToken *Tok = CurrentToken;
572       if (!consumeToken())
573         return false;
574       updateParameterCount(Left, Tok);
575     }
576     return false;
577   }
578 
579   bool parseBrace() {
580     if (CurrentToken) {
581       FormatToken *Left = CurrentToken->Previous;
582       Left->ParentBracket = Contexts.back().ContextKind;
583 
584       if (Contexts.back().CaretFound)
585         Left->Type = TT_ObjCBlockLBrace;
586       Contexts.back().CaretFound = false;
587 
588       ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
589       Contexts.back().ColonIsDictLiteral = true;
590       if (Left->BlockKind == BK_BracedInit)
591         Contexts.back().IsExpression = true;
592       if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
593           Left->Previous->is(TT_JsTypeColon))
594         Contexts.back().IsExpression = false;
595 
596       while (CurrentToken) {
597         if (CurrentToken->is(tok::r_brace)) {
598           Left->MatchingParen = CurrentToken;
599           CurrentToken->MatchingParen = Left;
600           next();
601           return true;
602         }
603         if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
604           return false;
605         updateParameterCount(Left, CurrentToken);
606         if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) {
607           FormatToken *Previous = CurrentToken->getPreviousNonComment();
608           if (Previous->is(TT_JsTypeOptionalQuestion))
609             Previous = Previous->getPreviousNonComment();
610           if ((CurrentToken->is(tok::colon) &&
611                (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) ||
612               Style.Language == FormatStyle::LK_Proto ||
613               Style.Language == FormatStyle::LK_TextProto) {
614             Left->Type = TT_DictLiteral;
615             if (Previous->Tok.getIdentifierInfo() ||
616                 Previous->is(tok::string_literal))
617               Previous->Type = TT_SelectorName;
618           }
619           if (CurrentToken->is(tok::colon) ||
620               Style.Language == FormatStyle::LK_JavaScript)
621             Left->Type = TT_DictLiteral;
622         }
623         if (CurrentToken->is(tok::comma) &&
624             Style.Language == FormatStyle::LK_JavaScript)
625           Left->Type = TT_DictLiteral;
626         if (!consumeToken())
627           return false;
628       }
629     }
630     return true;
631   }
632 
633   void updateParameterCount(FormatToken *Left, FormatToken *Current) {
634     // For ObjC methods, the number of parameters is calculated differently as
635     // method declarations have a different structure (the parameters are not
636     // inside a bracket scope).
637     if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block)
638       ++Left->BlockParameterCount;
639     if (Current->is(tok::comma)) {
640       ++Left->ParameterCount;
641       if (!Left->Role)
642         Left->Role.reset(new CommaSeparatedList(Style));
643       Left->Role->CommaFound(Current);
644     } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
645       Left->ParameterCount = 1;
646     }
647   }
648 
649   bool parseConditional() {
650     while (CurrentToken) {
651       if (CurrentToken->is(tok::colon)) {
652         CurrentToken->Type = TT_ConditionalExpr;
653         next();
654         return true;
655       }
656       if (!consumeToken())
657         return false;
658     }
659     return false;
660   }
661 
662   bool parseTemplateDeclaration() {
663     if (CurrentToken && CurrentToken->is(tok::less)) {
664       CurrentToken->Type = TT_TemplateOpener;
665       next();
666       if (!parseAngle())
667         return false;
668       if (CurrentToken)
669         CurrentToken->Previous->ClosesTemplateDeclaration = true;
670       return true;
671     }
672     return false;
673   }
674 
675   bool consumeToken() {
676     FormatToken *Tok = CurrentToken;
677     next();
678     switch (Tok->Tok.getKind()) {
679     case tok::plus:
680     case tok::minus:
681       if (!Tok->Previous && Line.MustBeDeclaration)
682         Tok->Type = TT_ObjCMethodSpecifier;
683       break;
684     case tok::colon:
685       if (!Tok->Previous)
686         return false;
687       // Colons from ?: are handled in parseConditional().
688       if (Style.Language == FormatStyle::LK_JavaScript) {
689         if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
690             (Contexts.size() == 1 &&               // switch/case labels
691              !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
692             Contexts.back().ContextKind == tok::l_paren ||  // function params
693             Contexts.back().ContextKind == tok::l_square || // array type
694             (!Contexts.back().IsExpression &&
695              Contexts.back().ContextKind == tok::l_brace) || // object type
696             (Contexts.size() == 1 &&
697              Line.MustBeDeclaration)) { // method/property declaration
698           Contexts.back().IsExpression = false;
699           Tok->Type = TT_JsTypeColon;
700           break;
701         }
702       }
703       if (Contexts.back().ColonIsDictLiteral ||
704           Style.Language == FormatStyle::LK_Proto ||
705           Style.Language == FormatStyle::LK_TextProto) {
706         Tok->Type = TT_DictLiteral;
707         if (Style.Language == FormatStyle::LK_TextProto) {
708           if (FormatToken *Previous = Tok->getPreviousNonComment())
709             Previous->Type = TT_SelectorName;
710         }
711       } else if (Contexts.back().ColonIsObjCMethodExpr ||
712                  Line.startsWith(TT_ObjCMethodSpecifier)) {
713         Tok->Type = TT_ObjCMethodExpr;
714         const FormatToken *BeforePrevious = Tok->Previous->Previous;
715         // Ensure we tag all identifiers in method declarations as
716         // TT_SelectorName.
717         bool UnknownIdentifierInMethodDeclaration =
718             Line.startsWith(TT_ObjCMethodSpecifier) &&
719             Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown);
720         if (!BeforePrevious ||
721             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
722             !(BeforePrevious->is(TT_CastRParen) ||
723               (BeforePrevious->is(TT_ObjCMethodExpr) &&
724                BeforePrevious->is(tok::colon))) ||
725             BeforePrevious->is(tok::r_square) ||
726             Contexts.back().LongestObjCSelectorName == 0 ||
727             UnknownIdentifierInMethodDeclaration) {
728           Tok->Previous->Type = TT_SelectorName;
729           if (!Contexts.back().FirstObjCSelectorName)
730             Contexts.back().FirstObjCSelectorName = Tok->Previous;
731           else if (Tok->Previous->ColumnWidth >
732                    Contexts.back().LongestObjCSelectorName)
733             Contexts.back().LongestObjCSelectorName =
734                 Tok->Previous->ColumnWidth;
735           Tok->Previous->ParameterIndex =
736               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
737           ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
738         }
739       } else if (Contexts.back().ColonIsForRangeExpr) {
740         Tok->Type = TT_RangeBasedForLoopColon;
741       } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
742         Tok->Type = TT_BitFieldColon;
743       } else if (Contexts.size() == 1 &&
744                  !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) {
745         if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren,
746                                                   tok::kw_noexcept))
747           Tok->Type = TT_CtorInitializerColon;
748         else
749           Tok->Type = TT_InheritanceColon;
750       } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next &&
751                  (Tok->Next->isOneOf(tok::r_paren, tok::comma) ||
752                   (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next &&
753                    Tok->Next->Next->is(tok::colon)))) {
754         // This handles a special macro in ObjC code where selectors including
755         // the colon are passed as macro arguments.
756         Tok->Type = TT_ObjCMethodExpr;
757       } else if (Contexts.back().ContextKind == tok::l_paren) {
758         Tok->Type = TT_InlineASMColon;
759       }
760       break;
761     case tok::pipe:
762     case tok::amp:
763       // | and & in declarations/type expressions represent union and
764       // intersection types, respectively.
765       if (Style.Language == FormatStyle::LK_JavaScript &&
766           !Contexts.back().IsExpression)
767         Tok->Type = TT_JsTypeOperator;
768       break;
769     case tok::kw_if:
770     case tok::kw_while:
771       if (Tok->is(tok::kw_if) && CurrentToken &&
772           CurrentToken->is(tok::kw_constexpr))
773         next();
774       if (CurrentToken && CurrentToken->is(tok::l_paren)) {
775         next();
776         if (!parseParens(/*LookForDecls=*/true))
777           return false;
778       }
779       break;
780     case tok::kw_for:
781       if (Style.Language == FormatStyle::LK_JavaScript) {
782         // x.for and {for: ...}
783         if ((Tok->Previous && Tok->Previous->is(tok::period)) ||
784             (Tok->Next && Tok->Next->is(tok::colon)))
785           break;
786         // JS' for await ( ...
787         if (CurrentToken && CurrentToken->is(Keywords.kw_await))
788           next();
789       }
790       Contexts.back().ColonIsForRangeExpr = true;
791       next();
792       if (!parseParens())
793         return false;
794       break;
795     case tok::l_paren:
796       // When faced with 'operator()()', the kw_operator handler incorrectly
797       // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
798       // the first two parens OverloadedOperators and the second l_paren an
799       // OverloadedOperatorLParen.
800       if (Tok->Previous && Tok->Previous->is(tok::r_paren) &&
801           Tok->Previous->MatchingParen &&
802           Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
803         Tok->Previous->Type = TT_OverloadedOperator;
804         Tok->Previous->MatchingParen->Type = TT_OverloadedOperator;
805         Tok->Type = TT_OverloadedOperatorLParen;
806       }
807 
808       if (!parseParens())
809         return false;
810       if (Line.MustBeDeclaration && Contexts.size() == 1 &&
811           !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
812           (!Tok->Previous ||
813            !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute,
814                                    TT_LeadingJavaAnnotation)))
815         Line.MightBeFunctionDecl = true;
816       break;
817     case tok::l_square:
818       if (!parseSquare())
819         return false;
820       break;
821     case tok::l_brace:
822       if (Style.Language == FormatStyle::LK_TextProto) {
823         FormatToken *Previous = Tok->getPreviousNonComment();
824         if (Previous && Previous->Type != TT_DictLiteral)
825           Previous->Type = TT_SelectorName;
826       }
827       if (!parseBrace())
828         return false;
829       break;
830     case tok::less:
831       if (parseAngle()) {
832         Tok->Type = TT_TemplateOpener;
833         // In TT_Proto, we must distignuish between:
834         //   map<key, value>
835         //   msg < item: data >
836         //   msg: < item: data >
837         // In TT_TextProto, map<key, value> does not occur.
838         if (Style.Language == FormatStyle::LK_TextProto ||
839             (Style.Language == FormatStyle::LK_Proto && Tok->Previous &&
840              Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
841           Tok->Type = TT_DictLiteral;
842           FormatToken *Previous = Tok->getPreviousNonComment();
843           if (Previous && Previous->Type != TT_DictLiteral)
844             Previous->Type = TT_SelectorName;
845         }
846       } else {
847         Tok->Type = TT_BinaryOperator;
848         NonTemplateLess.insert(Tok);
849         CurrentToken = Tok;
850         next();
851       }
852       break;
853     case tok::r_paren:
854     case tok::r_square:
855       return false;
856     case tok::r_brace:
857       // Lines can start with '}'.
858       if (Tok->Previous)
859         return false;
860       break;
861     case tok::greater:
862       if (Style.Language != FormatStyle::LK_TextProto)
863         Tok->Type = TT_BinaryOperator;
864       break;
865     case tok::kw_operator:
866       if (Style.Language == FormatStyle::LK_TextProto ||
867           Style.Language == FormatStyle::LK_Proto)
868         break;
869       while (CurrentToken &&
870              !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
871         if (CurrentToken->isOneOf(tok::star, tok::amp))
872           CurrentToken->Type = TT_PointerOrReference;
873         consumeToken();
874         if (CurrentToken &&
875             CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator,
876                                             tok::comma))
877           CurrentToken->Previous->Type = TT_OverloadedOperator;
878       }
879       if (CurrentToken) {
880         CurrentToken->Type = TT_OverloadedOperatorLParen;
881         if (CurrentToken->Previous->is(TT_BinaryOperator))
882           CurrentToken->Previous->Type = TT_OverloadedOperator;
883       }
884       break;
885     case tok::question:
886       if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next &&
887           Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
888                              tok::r_brace)) {
889         // Question marks before semicolons, colons, etc. indicate optional
890         // types (fields, parameters), e.g.
891         //   function(x?: string, y?) {...}
892         //   class X { y?; }
893         Tok->Type = TT_JsTypeOptionalQuestion;
894         break;
895       }
896       // Declarations cannot be conditional expressions, this can only be part
897       // of a type declaration.
898       if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
899           Style.Language == FormatStyle::LK_JavaScript)
900         break;
901       parseConditional();
902       break;
903     case tok::kw_template:
904       parseTemplateDeclaration();
905       break;
906     case tok::comma:
907       if (Contexts.back().InCtorInitializer)
908         Tok->Type = TT_CtorInitializerComma;
909       else if (Contexts.back().InInheritanceList)
910         Tok->Type = TT_InheritanceComma;
911       else if (Contexts.back().FirstStartOfName &&
912                (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
913         Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
914         Line.IsMultiVariableDeclStmt = true;
915       }
916       if (Contexts.back().IsForEachMacro)
917         Contexts.back().IsExpression = true;
918       break;
919     case tok::identifier:
920       if (Tok->isOneOf(Keywords.kw___has_include,
921                        Keywords.kw___has_include_next)) {
922         parseHasInclude();
923       }
924       break;
925     default:
926       break;
927     }
928     return true;
929   }
930 
931   void parseIncludeDirective() {
932     if (CurrentToken && CurrentToken->is(tok::less)) {
933       next();
934       while (CurrentToken) {
935         // Mark tokens up to the trailing line comments as implicit string
936         // literals.
937         if (CurrentToken->isNot(tok::comment) &&
938             !CurrentToken->TokenText.startswith("//"))
939           CurrentToken->Type = TT_ImplicitStringLiteral;
940         next();
941       }
942     }
943   }
944 
945   void parseWarningOrError() {
946     next();
947     // We still want to format the whitespace left of the first token of the
948     // warning or error.
949     next();
950     while (CurrentToken) {
951       CurrentToken->Type = TT_ImplicitStringLiteral;
952       next();
953     }
954   }
955 
956   void parsePragma() {
957     next(); // Consume "pragma".
958     if (CurrentToken &&
959         CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
960       bool IsMark = CurrentToken->is(Keywords.kw_mark);
961       next(); // Consume "mark".
962       next(); // Consume first token (so we fix leading whitespace).
963       while (CurrentToken) {
964         if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
965           CurrentToken->Type = TT_ImplicitStringLiteral;
966         next();
967       }
968     }
969   }
970 
971   void parseHasInclude() {
972     if (!CurrentToken || !CurrentToken->is(tok::l_paren))
973       return;
974     next(); // '('
975     parseIncludeDirective();
976     next(); // ')'
977   }
978 
979   LineType parsePreprocessorDirective() {
980     bool IsFirstToken = CurrentToken->IsFirst;
981     LineType Type = LT_PreprocessorDirective;
982     next();
983     if (!CurrentToken)
984       return Type;
985 
986     if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
987       // JavaScript files can contain shebang lines of the form:
988       // #!/usr/bin/env node
989       // Treat these like C++ #include directives.
990       while (CurrentToken) {
991         // Tokens cannot be comments here.
992         CurrentToken->Type = TT_ImplicitStringLiteral;
993         next();
994       }
995       return LT_ImportStatement;
996     }
997 
998     if (CurrentToken->Tok.is(tok::numeric_constant)) {
999       CurrentToken->SpacesRequiredBefore = 1;
1000       return Type;
1001     }
1002     // Hashes in the middle of a line can lead to any strange token
1003     // sequence.
1004     if (!CurrentToken->Tok.getIdentifierInfo())
1005       return Type;
1006     switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
1007     case tok::pp_include:
1008     case tok::pp_include_next:
1009     case tok::pp_import:
1010       next();
1011       parseIncludeDirective();
1012       Type = LT_ImportStatement;
1013       break;
1014     case tok::pp_error:
1015     case tok::pp_warning:
1016       parseWarningOrError();
1017       break;
1018     case tok::pp_pragma:
1019       parsePragma();
1020       break;
1021     case tok::pp_if:
1022     case tok::pp_elif:
1023       Contexts.back().IsExpression = true;
1024       parseLine();
1025       break;
1026     default:
1027       break;
1028     }
1029     while (CurrentToken) {
1030       FormatToken *Tok = CurrentToken;
1031       next();
1032       if (Tok->is(tok::l_paren))
1033         parseParens();
1034       else if (Tok->isOneOf(Keywords.kw___has_include,
1035                             Keywords.kw___has_include_next))
1036         parseHasInclude();
1037     }
1038     return Type;
1039   }
1040 
1041 public:
1042   LineType parseLine() {
1043     NonTemplateLess.clear();
1044     if (CurrentToken->is(tok::hash))
1045       return parsePreprocessorDirective();
1046 
1047     // Directly allow to 'import <string-literal>' to support protocol buffer
1048     // definitions (github.com/google/protobuf) or missing "#" (either way we
1049     // should not break the line).
1050     IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
1051     if ((Style.Language == FormatStyle::LK_Java &&
1052          CurrentToken->is(Keywords.kw_package)) ||
1053         (Info && Info->getPPKeywordID() == tok::pp_import &&
1054          CurrentToken->Next &&
1055          CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
1056                                      tok::kw_static))) {
1057       next();
1058       parseIncludeDirective();
1059       return LT_ImportStatement;
1060     }
1061 
1062     // If this line starts and ends in '<' and '>', respectively, it is likely
1063     // part of "#define <a/b.h>".
1064     if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
1065       parseIncludeDirective();
1066       return LT_ImportStatement;
1067     }
1068 
1069     // In .proto files, top-level options are very similar to import statements
1070     // and should not be line-wrapped.
1071     if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
1072         CurrentToken->is(Keywords.kw_option)) {
1073       next();
1074       if (CurrentToken && CurrentToken->is(tok::identifier))
1075         return LT_ImportStatement;
1076     }
1077 
1078     bool KeywordVirtualFound = false;
1079     bool ImportStatement = false;
1080 
1081     // import {...} from '...';
1082     if (Style.Language == FormatStyle::LK_JavaScript &&
1083         CurrentToken->is(Keywords.kw_import))
1084       ImportStatement = true;
1085 
1086     while (CurrentToken) {
1087       if (CurrentToken->is(tok::kw_virtual))
1088         KeywordVirtualFound = true;
1089       if (Style.Language == FormatStyle::LK_JavaScript) {
1090         // export {...} from '...';
1091         // An export followed by "from 'some string';" is a re-export from
1092         // another module identified by a URI and is treated as a
1093         // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
1094         // Just "export {...};" or "export class ..." should not be treated as
1095         // an import in this sense.
1096         if (Line.First->is(tok::kw_export) &&
1097             CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
1098             CurrentToken->Next->isStringLiteral())
1099           ImportStatement = true;
1100         if (isClosureImportStatement(*CurrentToken))
1101           ImportStatement = true;
1102       }
1103       if (!consumeToken())
1104         return LT_Invalid;
1105     }
1106     if (KeywordVirtualFound)
1107       return LT_VirtualFunctionDecl;
1108     if (ImportStatement)
1109       return LT_ImportStatement;
1110 
1111     if (Line.startsWith(TT_ObjCMethodSpecifier)) {
1112       if (Contexts.back().FirstObjCSelectorName)
1113         Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
1114             Contexts.back().LongestObjCSelectorName;
1115       return LT_ObjCMethodDecl;
1116     }
1117 
1118     return LT_Other;
1119   }
1120 
1121 private:
1122   bool isClosureImportStatement(const FormatToken &Tok) {
1123     // FIXME: Closure-library specific stuff should not be hard-coded but be
1124     // configurable.
1125     return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
1126            Tok.Next->Next &&
1127            (Tok.Next->Next->TokenText == "module" ||
1128             Tok.Next->Next->TokenText == "provide" ||
1129             Tok.Next->Next->TokenText == "require" ||
1130             Tok.Next->Next->TokenText == "requireType" ||
1131             Tok.Next->Next->TokenText == "forwardDeclare") &&
1132            Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
1133   }
1134 
1135   void resetTokenMetadata(FormatToken *Token) {
1136     if (!Token)
1137       return;
1138 
1139     // Reset token type in case we have already looked at it and then
1140     // recovered from an error (e.g. failure to find the matching >).
1141     if (!CurrentToken->isOneOf(TT_LambdaLSquare, TT_ForEachMacro,
1142                                TT_FunctionLBrace, TT_ImplicitStringLiteral,
1143                                TT_InlineASMBrace, TT_JsFatArrow, TT_LambdaArrow,
1144                                TT_OverloadedOperator, TT_RegexLiteral,
1145                                TT_TemplateString, TT_ObjCStringLiteral))
1146       CurrentToken->Type = TT_Unknown;
1147     CurrentToken->Role.reset();
1148     CurrentToken->MatchingParen = nullptr;
1149     CurrentToken->FakeLParens.clear();
1150     CurrentToken->FakeRParens = 0;
1151   }
1152 
1153   void next() {
1154     if (CurrentToken) {
1155       CurrentToken->NestingLevel = Contexts.size() - 1;
1156       CurrentToken->BindingStrength = Contexts.back().BindingStrength;
1157       modifyContext(*CurrentToken);
1158       determineTokenType(*CurrentToken);
1159       CurrentToken = CurrentToken->Next;
1160     }
1161 
1162     resetTokenMetadata(CurrentToken);
1163   }
1164 
1165   /// A struct to hold information valid in a specific context, e.g.
1166   /// a pair of parenthesis.
1167   struct Context {
1168     Context(tok::TokenKind ContextKind, unsigned BindingStrength,
1169             bool IsExpression)
1170         : ContextKind(ContextKind), BindingStrength(BindingStrength),
1171           IsExpression(IsExpression) {}
1172 
1173     tok::TokenKind ContextKind;
1174     unsigned BindingStrength;
1175     bool IsExpression;
1176     unsigned LongestObjCSelectorName = 0;
1177     bool ColonIsForRangeExpr = false;
1178     bool ColonIsDictLiteral = false;
1179     bool ColonIsObjCMethodExpr = false;
1180     FormatToken *FirstObjCSelectorName = nullptr;
1181     FormatToken *FirstStartOfName = nullptr;
1182     bool CanBeExpression = true;
1183     bool InTemplateArgument = false;
1184     bool InCtorInitializer = false;
1185     bool InInheritanceList = false;
1186     bool CaretFound = false;
1187     bool IsForEachMacro = false;
1188     bool InCpp11AttributeSpecifier = false;
1189   };
1190 
1191   /// Puts a new \c Context onto the stack \c Contexts for the lifetime
1192   /// of each instance.
1193   struct ScopedContextCreator {
1194     AnnotatingParser &P;
1195 
1196     ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
1197                          unsigned Increase)
1198         : P(P) {
1199       P.Contexts.push_back(Context(ContextKind,
1200                                    P.Contexts.back().BindingStrength + Increase,
1201                                    P.Contexts.back().IsExpression));
1202     }
1203 
1204     ~ScopedContextCreator() { P.Contexts.pop_back(); }
1205   };
1206 
1207   void modifyContext(const FormatToken &Current) {
1208     if (Current.getPrecedence() == prec::Assignment &&
1209         !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
1210         // Type aliases use `type X = ...;` in TypeScript and can be exported
1211         // using `export type ...`.
1212         !(Style.Language == FormatStyle::LK_JavaScript &&
1213           (Line.startsWith(Keywords.kw_type, tok::identifier) ||
1214            Line.startsWith(tok::kw_export, Keywords.kw_type,
1215                            tok::identifier))) &&
1216         (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
1217       Contexts.back().IsExpression = true;
1218       if (!Line.startsWith(TT_UnaryOperator)) {
1219         for (FormatToken *Previous = Current.Previous;
1220              Previous && Previous->Previous &&
1221              !Previous->Previous->isOneOf(tok::comma, tok::semi);
1222              Previous = Previous->Previous) {
1223           if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
1224             Previous = Previous->MatchingParen;
1225             if (!Previous)
1226               break;
1227           }
1228           if (Previous->opensScope())
1229             break;
1230           if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
1231               Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
1232               Previous->Previous && Previous->Previous->isNot(tok::equal))
1233             Previous->Type = TT_PointerOrReference;
1234         }
1235       }
1236     } else if (Current.is(tok::lessless) &&
1237                (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1238       Contexts.back().IsExpression = true;
1239     } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1240       Contexts.back().IsExpression = true;
1241     } else if (Current.is(TT_TrailingReturnArrow)) {
1242       Contexts.back().IsExpression = false;
1243     } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1244       Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1245     } else if (Current.Previous &&
1246                Current.Previous->is(TT_CtorInitializerColon)) {
1247       Contexts.back().IsExpression = true;
1248       Contexts.back().InCtorInitializer = true;
1249     } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) {
1250       Contexts.back().InInheritanceList = true;
1251     } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1252       for (FormatToken *Previous = Current.Previous;
1253            Previous && Previous->isOneOf(tok::star, tok::amp);
1254            Previous = Previous->Previous)
1255         Previous->Type = TT_PointerOrReference;
1256       if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1257         Contexts.back().IsExpression = false;
1258     } else if (Current.is(tok::kw_new)) {
1259       Contexts.back().CanBeExpression = false;
1260     } else if (Current.isOneOf(tok::semi, tok::exclaim)) {
1261       // This should be the condition or increment in a for-loop.
1262       Contexts.back().IsExpression = true;
1263     }
1264   }
1265 
1266   void determineTokenType(FormatToken &Current) {
1267     if (!Current.is(TT_Unknown))
1268       // The token type is already known.
1269       return;
1270 
1271     if (Style.Language == FormatStyle::LK_JavaScript) {
1272       if (Current.is(tok::exclaim)) {
1273         if (Current.Previous &&
1274             (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace,
1275                                        tok::r_paren, tok::r_square,
1276                                        tok::r_brace) ||
1277              Current.Previous->Tok.isLiteral())) {
1278           Current.Type = TT_JsNonNullAssertion;
1279           return;
1280         }
1281         if (Current.Next &&
1282             Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1283           Current.Type = TT_JsNonNullAssertion;
1284           return;
1285         }
1286       }
1287     }
1288 
1289     // Line.MightBeFunctionDecl can only be true after the parentheses of a
1290     // function declaration have been found. In this case, 'Current' is a
1291     // trailing token of this declaration and thus cannot be a name.
1292     if (Current.is(Keywords.kw_instanceof)) {
1293       Current.Type = TT_BinaryOperator;
1294     } else if (isStartOfName(Current) &&
1295                (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1296       Contexts.back().FirstStartOfName = &Current;
1297       Current.Type = TT_StartOfName;
1298     } else if (Current.is(tok::semi)) {
1299       // Reset FirstStartOfName after finding a semicolon so that a for loop
1300       // with multiple increment statements is not confused with a for loop
1301       // having multiple variable declarations.
1302       Contexts.back().FirstStartOfName = nullptr;
1303     } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1304       AutoFound = true;
1305     } else if (Current.is(tok::arrow) &&
1306                Style.Language == FormatStyle::LK_Java) {
1307       Current.Type = TT_LambdaArrow;
1308     } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1309                Current.NestingLevel == 0) {
1310       Current.Type = TT_TrailingReturnArrow;
1311     } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1312       Current.Type = determineStarAmpUsage(Current,
1313                                            Contexts.back().CanBeExpression &&
1314                                                Contexts.back().IsExpression,
1315                                            Contexts.back().InTemplateArgument);
1316     } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1317       Current.Type = determinePlusMinusCaretUsage(Current);
1318       if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1319         Contexts.back().CaretFound = true;
1320     } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1321       Current.Type = determineIncrementUsage(Current);
1322     } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1323       Current.Type = TT_UnaryOperator;
1324     } else if (Current.is(tok::question)) {
1325       if (Style.Language == FormatStyle::LK_JavaScript &&
1326           Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1327         // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1328         // on the interface, not a ternary expression.
1329         Current.Type = TT_JsTypeOptionalQuestion;
1330       } else {
1331         Current.Type = TT_ConditionalExpr;
1332       }
1333     } else if (Current.isBinaryOperator() &&
1334                (!Current.Previous || Current.Previous->isNot(tok::l_square)) &&
1335                (!Current.is(tok::greater) &&
1336                 Style.Language != FormatStyle::LK_TextProto)) {
1337       Current.Type = TT_BinaryOperator;
1338     } else if (Current.is(tok::comment)) {
1339       if (Current.TokenText.startswith("/*")) {
1340         if (Current.TokenText.endswith("*/"))
1341           Current.Type = TT_BlockComment;
1342         else
1343           // The lexer has for some reason determined a comment here. But we
1344           // cannot really handle it, if it isn't properly terminated.
1345           Current.Tok.setKind(tok::unknown);
1346       } else {
1347         Current.Type = TT_LineComment;
1348       }
1349     } else if (Current.is(tok::r_paren)) {
1350       if (rParenEndsCast(Current))
1351         Current.Type = TT_CastRParen;
1352       if (Current.MatchingParen && Current.Next &&
1353           !Current.Next->isBinaryOperator() &&
1354           !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1355                                  tok::comma, tok::period, tok::arrow,
1356                                  tok::coloncolon))
1357         if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1358           // Make sure this isn't the return type of an Obj-C block declaration
1359           if (AfterParen->Tok.isNot(tok::caret)) {
1360             if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1361               if (BeforeParen->is(tok::identifier) &&
1362                   BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1363                   (!BeforeParen->Previous ||
1364                    BeforeParen->Previous->ClosesTemplateDeclaration))
1365                 Current.Type = TT_FunctionAnnotationRParen;
1366           }
1367         }
1368     } else if (Current.is(tok::at) && Current.Next &&
1369                Style.Language != FormatStyle::LK_JavaScript &&
1370                Style.Language != FormatStyle::LK_Java) {
1371       // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1372       // marks declarations and properties that need special formatting.
1373       switch (Current.Next->Tok.getObjCKeywordID()) {
1374       case tok::objc_interface:
1375       case tok::objc_implementation:
1376       case tok::objc_protocol:
1377         Current.Type = TT_ObjCDecl;
1378         break;
1379       case tok::objc_property:
1380         Current.Type = TT_ObjCProperty;
1381         break;
1382       default:
1383         break;
1384       }
1385     } else if (Current.is(tok::period)) {
1386       FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1387       if (PreviousNoComment &&
1388           PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1389         Current.Type = TT_DesignatedInitializerPeriod;
1390       else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1391                Current.Previous->isOneOf(TT_JavaAnnotation,
1392                                          TT_LeadingJavaAnnotation)) {
1393         Current.Type = Current.Previous->Type;
1394       }
1395     } else if (canBeObjCSelectorComponent(Current) &&
1396                // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
1397                Current.Previous && Current.Previous->is(TT_CastRParen) &&
1398                Current.Previous->MatchingParen &&
1399                Current.Previous->MatchingParen->Previous &&
1400                Current.Previous->MatchingParen->Previous->is(
1401                    TT_ObjCMethodSpecifier)) {
1402       // This is the first part of an Objective-C selector name. (If there's no
1403       // colon after this, this is the only place which annotates the identifier
1404       // as a selector.)
1405       Current.Type = TT_SelectorName;
1406     } else if (Current.isOneOf(tok::identifier, tok::kw_const) &&
1407                Current.Previous &&
1408                !Current.Previous->isOneOf(tok::equal, tok::at) &&
1409                Line.MightBeFunctionDecl && Contexts.size() == 1) {
1410       // Line.MightBeFunctionDecl can only be true after the parentheses of a
1411       // function declaration have been found.
1412       Current.Type = TT_TrailingAnnotation;
1413     } else if ((Style.Language == FormatStyle::LK_Java ||
1414                 Style.Language == FormatStyle::LK_JavaScript) &&
1415                Current.Previous) {
1416       if (Current.Previous->is(tok::at) &&
1417           Current.isNot(Keywords.kw_interface)) {
1418         const FormatToken &AtToken = *Current.Previous;
1419         const FormatToken *Previous = AtToken.getPreviousNonComment();
1420         if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1421           Current.Type = TT_LeadingJavaAnnotation;
1422         else
1423           Current.Type = TT_JavaAnnotation;
1424       } else if (Current.Previous->is(tok::period) &&
1425                  Current.Previous->isOneOf(TT_JavaAnnotation,
1426                                            TT_LeadingJavaAnnotation)) {
1427         Current.Type = Current.Previous->Type;
1428       }
1429     }
1430   }
1431 
1432   /// Take a guess at whether \p Tok starts a name of a function or
1433   /// variable declaration.
1434   ///
1435   /// This is a heuristic based on whether \p Tok is an identifier following
1436   /// something that is likely a type.
1437   bool isStartOfName(const FormatToken &Tok) {
1438     if (Tok.isNot(tok::identifier) || !Tok.Previous)
1439       return false;
1440 
1441     if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1442                               Keywords.kw_as))
1443       return false;
1444     if (Style.Language == FormatStyle::LK_JavaScript &&
1445         Tok.Previous->is(Keywords.kw_in))
1446       return false;
1447 
1448     // Skip "const" as it does not have an influence on whether this is a name.
1449     FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1450     while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1451       PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1452 
1453     if (!PreviousNotConst)
1454       return false;
1455 
1456     bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1457                        PreviousNotConst->Previous &&
1458                        PreviousNotConst->Previous->is(tok::hash);
1459 
1460     if (PreviousNotConst->is(TT_TemplateCloser))
1461       return PreviousNotConst && PreviousNotConst->MatchingParen &&
1462              PreviousNotConst->MatchingParen->Previous &&
1463              PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1464              PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1465 
1466     if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen &&
1467         PreviousNotConst->MatchingParen->Previous &&
1468         PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype))
1469       return true;
1470 
1471     return (!IsPPKeyword &&
1472             PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1473            PreviousNotConst->is(TT_PointerOrReference) ||
1474            PreviousNotConst->isSimpleTypeSpecifier();
1475   }
1476 
1477   /// Determine whether ')' is ending a cast.
1478   bool rParenEndsCast(const FormatToken &Tok) {
1479     // C-style casts are only used in C++ and Java.
1480     if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java)
1481       return false;
1482 
1483     // Empty parens aren't casts and there are no casts at the end of the line.
1484     if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1485       return false;
1486 
1487     FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1488     if (LeftOfParens) {
1489       // If there is a closing parenthesis left of the current parentheses,
1490       // look past it as these might be chained casts.
1491       if (LeftOfParens->is(tok::r_paren)) {
1492         if (!LeftOfParens->MatchingParen ||
1493             !LeftOfParens->MatchingParen->Previous)
1494           return false;
1495         LeftOfParens = LeftOfParens->MatchingParen->Previous;
1496       }
1497 
1498       // If there is an identifier (or with a few exceptions a keyword) right
1499       // before the parentheses, this is unlikely to be a cast.
1500       if (LeftOfParens->Tok.getIdentifierInfo() &&
1501           !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1502                                  tok::kw_delete))
1503         return false;
1504 
1505       // Certain other tokens right before the parentheses are also signals that
1506       // this cannot be a cast.
1507       if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1508                                 TT_TemplateCloser, tok::ellipsis))
1509         return false;
1510     }
1511 
1512     if (Tok.Next->is(tok::question))
1513       return false;
1514 
1515     // As Java has no function types, a "(" after the ")" likely means that this
1516     // is a cast.
1517     if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1518       return true;
1519 
1520     // If a (non-string) literal follows, this is likely a cast.
1521     if (Tok.Next->isNot(tok::string_literal) &&
1522         (Tok.Next->Tok.isLiteral() ||
1523          Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1524       return true;
1525 
1526     // Heuristically try to determine whether the parentheses contain a type.
1527     bool ParensAreType =
1528         !Tok.Previous ||
1529         Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) ||
1530         Tok.Previous->isSimpleTypeSpecifier();
1531     bool ParensCouldEndDecl =
1532         Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1533     if (ParensAreType && !ParensCouldEndDecl)
1534       return true;
1535 
1536     // At this point, we heuristically assume that there are no casts at the
1537     // start of the line. We assume that we have found most cases where there
1538     // are by the logic above, e.g. "(void)x;".
1539     if (!LeftOfParens)
1540       return false;
1541 
1542     // Certain token types inside the parentheses mean that this can't be a
1543     // cast.
1544     for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1545          Token = Token->Next)
1546       if (Token->is(TT_BinaryOperator))
1547         return false;
1548 
1549     // If the following token is an identifier or 'this', this is a cast. All
1550     // cases where this can be something else are handled above.
1551     if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1552       return true;
1553 
1554     if (!Tok.Next->Next)
1555       return false;
1556 
1557     // If the next token after the parenthesis is a unary operator, assume
1558     // that this is cast, unless there are unexpected tokens inside the
1559     // parenthesis.
1560     bool NextIsUnary =
1561         Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1562     if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1563         !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1564       return false;
1565     // Search for unexpected tokens.
1566     for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1567          Prev = Prev->Previous) {
1568       if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1569         return false;
1570     }
1571     return true;
1572   }
1573 
1574   /// Return the type of the given token assuming it is * or &.
1575   TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1576                                   bool InTemplateArgument) {
1577     if (Style.Language == FormatStyle::LK_JavaScript)
1578       return TT_BinaryOperator;
1579 
1580     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1581     if (!PrevToken)
1582       return TT_UnaryOperator;
1583 
1584     const FormatToken *NextToken = Tok.getNextNonComment();
1585     if (!NextToken ||
1586         NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) ||
1587         (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1588       return TT_PointerOrReference;
1589 
1590     if (PrevToken->is(tok::coloncolon))
1591       return TT_PointerOrReference;
1592 
1593     if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1594                            tok::comma, tok::semi, tok::kw_return, tok::colon,
1595                            tok::equal, tok::kw_delete, tok::kw_sizeof,
1596                            tok::kw_throw) ||
1597         PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1598                            TT_UnaryOperator, TT_CastRParen))
1599       return TT_UnaryOperator;
1600 
1601     if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1602       return TT_PointerOrReference;
1603     if (NextToken->is(tok::kw_operator) && !IsExpression)
1604       return TT_PointerOrReference;
1605     if (NextToken->isOneOf(tok::comma, tok::semi))
1606       return TT_PointerOrReference;
1607 
1608     if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) {
1609       FormatToken *TokenBeforeMatchingParen =
1610           PrevToken->MatchingParen->getPreviousNonComment();
1611       if (TokenBeforeMatchingParen &&
1612           TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype))
1613         return TT_PointerOrReference;
1614     }
1615 
1616     if (PrevToken->Tok.isLiteral() ||
1617         PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1618                            tok::kw_false, tok::r_brace) ||
1619         NextToken->Tok.isLiteral() ||
1620         NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1621         NextToken->isUnaryOperator() ||
1622         // If we know we're in a template argument, there are no named
1623         // declarations. Thus, having an identifier on the right-hand side
1624         // indicates a binary operator.
1625         (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1626       return TT_BinaryOperator;
1627 
1628     // "&&(" is quite unlikely to be two successive unary "&".
1629     if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren))
1630       return TT_BinaryOperator;
1631 
1632     // This catches some cases where evaluation order is used as control flow:
1633     //   aaa && aaa->f();
1634     const FormatToken *NextNextToken = NextToken->getNextNonComment();
1635     if (NextNextToken && NextNextToken->is(tok::arrow))
1636       return TT_BinaryOperator;
1637 
1638     // It is very unlikely that we are going to find a pointer or reference type
1639     // definition on the RHS of an assignment.
1640     if (IsExpression && !Contexts.back().CaretFound)
1641       return TT_BinaryOperator;
1642 
1643     return TT_PointerOrReference;
1644   }
1645 
1646   TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
1647     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1648     if (!PrevToken)
1649       return TT_UnaryOperator;
1650 
1651     if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator))
1652       // This must be a sequence of leading unary operators.
1653       return TT_UnaryOperator;
1654 
1655     // Use heuristics to recognize unary operators.
1656     if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
1657                            tok::question, tok::colon, tok::kw_return,
1658                            tok::kw_case, tok::at, tok::l_brace))
1659       return TT_UnaryOperator;
1660 
1661     // There can't be two consecutive binary operators.
1662     if (PrevToken->is(TT_BinaryOperator))
1663       return TT_UnaryOperator;
1664 
1665     // Fall back to marking the token as binary operator.
1666     return TT_BinaryOperator;
1667   }
1668 
1669   /// Determine whether ++/-- are pre- or post-increments/-decrements.
1670   TokenType determineIncrementUsage(const FormatToken &Tok) {
1671     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1672     if (!PrevToken || PrevToken->is(TT_CastRParen))
1673       return TT_UnaryOperator;
1674     if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
1675       return TT_TrailingUnaryOperator;
1676 
1677     return TT_UnaryOperator;
1678   }
1679 
1680   SmallVector<Context, 8> Contexts;
1681 
1682   const FormatStyle &Style;
1683   AnnotatedLine &Line;
1684   FormatToken *CurrentToken;
1685   bool AutoFound;
1686   const AdditionalKeywords &Keywords;
1687 
1688   // Set of "<" tokens that do not open a template parameter list. If parseAngle
1689   // determines that a specific token can't be a template opener, it will make
1690   // same decision irrespective of the decisions for tokens leading up to it.
1691   // Store this information to prevent this from causing exponential runtime.
1692   llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
1693 };
1694 
1695 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
1696 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
1697 
1698 /// Parses binary expressions by inserting fake parenthesis based on
1699 /// operator precedence.
1700 class ExpressionParser {
1701 public:
1702   ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
1703                    AnnotatedLine &Line)
1704       : Style(Style), Keywords(Keywords), Current(Line.First) {}
1705 
1706   /// Parse expressions with the given operator precedence.
1707   void parse(int Precedence = 0) {
1708     // Skip 'return' and ObjC selector colons as they are not part of a binary
1709     // expression.
1710     while (Current && (Current->is(tok::kw_return) ||
1711                        (Current->is(tok::colon) &&
1712                         Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
1713       next();
1714 
1715     if (!Current || Precedence > PrecedenceArrowAndPeriod)
1716       return;
1717 
1718     // Conditional expressions need to be parsed separately for proper nesting.
1719     if (Precedence == prec::Conditional) {
1720       parseConditionalExpr();
1721       return;
1722     }
1723 
1724     // Parse unary operators, which all have a higher precedence than binary
1725     // operators.
1726     if (Precedence == PrecedenceUnaryOperator) {
1727       parseUnaryOperator();
1728       return;
1729     }
1730 
1731     FormatToken *Start = Current;
1732     FormatToken *LatestOperator = nullptr;
1733     unsigned OperatorIndex = 0;
1734 
1735     while (Current) {
1736       // Consume operators with higher precedence.
1737       parse(Precedence + 1);
1738 
1739       int CurrentPrecedence = getCurrentPrecedence();
1740 
1741       if (Current && Current->is(TT_SelectorName) &&
1742           Precedence == CurrentPrecedence) {
1743         if (LatestOperator)
1744           addFakeParenthesis(Start, prec::Level(Precedence));
1745         Start = Current;
1746       }
1747 
1748       // At the end of the line or when an operator with higher precedence is
1749       // found, insert fake parenthesis and return.
1750       if (!Current ||
1751           (Current->closesScope() &&
1752            (Current->MatchingParen || Current->is(TT_TemplateString))) ||
1753           (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
1754           (CurrentPrecedence == prec::Conditional &&
1755            Precedence == prec::Assignment && Current->is(tok::colon))) {
1756         break;
1757       }
1758 
1759       // Consume scopes: (), [], <> and {}
1760       if (Current->opensScope()) {
1761         // In fragment of a JavaScript template string can look like '}..${' and
1762         // thus close a scope and open a new one at the same time.
1763         while (Current && (!Current->closesScope() || Current->opensScope())) {
1764           next();
1765           parse();
1766         }
1767         next();
1768       } else {
1769         // Operator found.
1770         if (CurrentPrecedence == Precedence) {
1771           if (LatestOperator)
1772             LatestOperator->NextOperator = Current;
1773           LatestOperator = Current;
1774           Current->OperatorIndex = OperatorIndex;
1775           ++OperatorIndex;
1776         }
1777         next(/*SkipPastLeadingComments=*/Precedence > 0);
1778       }
1779     }
1780 
1781     if (LatestOperator && (Current || Precedence > 0)) {
1782       // LatestOperator->LastOperator = true;
1783       if (Precedence == PrecedenceArrowAndPeriod) {
1784         // Call expressions don't have a binary operator precedence.
1785         addFakeParenthesis(Start, prec::Unknown);
1786       } else {
1787         addFakeParenthesis(Start, prec::Level(Precedence));
1788       }
1789     }
1790   }
1791 
1792 private:
1793   /// Gets the precedence (+1) of the given token for binary operators
1794   /// and other tokens that we treat like binary operators.
1795   int getCurrentPrecedence() {
1796     if (Current) {
1797       const FormatToken *NextNonComment = Current->getNextNonComment();
1798       if (Current->is(TT_ConditionalExpr))
1799         return prec::Conditional;
1800       if (NextNonComment && Current->is(TT_SelectorName) &&
1801           (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) ||
1802            ((Style.Language == FormatStyle::LK_Proto ||
1803              Style.Language == FormatStyle::LK_TextProto) &&
1804             NextNonComment->is(tok::less))))
1805         return prec::Assignment;
1806       if (Current->is(TT_JsComputedPropertyName))
1807         return prec::Assignment;
1808       if (Current->is(TT_LambdaArrow))
1809         return prec::Comma;
1810       if (Current->is(TT_JsFatArrow))
1811         return prec::Assignment;
1812       if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
1813           (Current->is(tok::comment) && NextNonComment &&
1814            NextNonComment->is(TT_SelectorName)))
1815         return 0;
1816       if (Current->is(TT_RangeBasedForLoopColon))
1817         return prec::Comma;
1818       if ((Style.Language == FormatStyle::LK_Java ||
1819            Style.Language == FormatStyle::LK_JavaScript) &&
1820           Current->is(Keywords.kw_instanceof))
1821         return prec::Relational;
1822       if (Style.Language == FormatStyle::LK_JavaScript &&
1823           Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
1824         return prec::Relational;
1825       if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
1826         return Current->getPrecedence();
1827       if (Current->isOneOf(tok::period, tok::arrow))
1828         return PrecedenceArrowAndPeriod;
1829       if ((Style.Language == FormatStyle::LK_Java ||
1830            Style.Language == FormatStyle::LK_JavaScript) &&
1831           Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
1832                            Keywords.kw_throws))
1833         return 0;
1834     }
1835     return -1;
1836   }
1837 
1838   void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
1839     Start->FakeLParens.push_back(Precedence);
1840     if (Precedence > prec::Unknown)
1841       Start->StartsBinaryExpression = true;
1842     if (Current) {
1843       FormatToken *Previous = Current->Previous;
1844       while (Previous->is(tok::comment) && Previous->Previous)
1845         Previous = Previous->Previous;
1846       ++Previous->FakeRParens;
1847       if (Precedence > prec::Unknown)
1848         Previous->EndsBinaryExpression = true;
1849     }
1850   }
1851 
1852   /// Parse unary operator expressions and surround them with fake
1853   /// parentheses if appropriate.
1854   void parseUnaryOperator() {
1855     llvm::SmallVector<FormatToken *, 2> Tokens;
1856     while (Current && Current->is(TT_UnaryOperator)) {
1857       Tokens.push_back(Current);
1858       next();
1859     }
1860     parse(PrecedenceArrowAndPeriod);
1861     for (FormatToken *Token : llvm::reverse(Tokens))
1862       // The actual precedence doesn't matter.
1863       addFakeParenthesis(Token, prec::Unknown);
1864   }
1865 
1866   void parseConditionalExpr() {
1867     while (Current && Current->isTrailingComment()) {
1868       next();
1869     }
1870     FormatToken *Start = Current;
1871     parse(prec::LogicalOr);
1872     if (!Current || !Current->is(tok::question))
1873       return;
1874     next();
1875     parse(prec::Assignment);
1876     if (!Current || Current->isNot(TT_ConditionalExpr))
1877       return;
1878     next();
1879     parse(prec::Assignment);
1880     addFakeParenthesis(Start, prec::Conditional);
1881   }
1882 
1883   void next(bool SkipPastLeadingComments = true) {
1884     if (Current)
1885       Current = Current->Next;
1886     while (Current &&
1887            (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
1888            Current->isTrailingComment())
1889       Current = Current->Next;
1890   }
1891 
1892   const FormatStyle &Style;
1893   const AdditionalKeywords &Keywords;
1894   FormatToken *Current;
1895 };
1896 
1897 } // end anonymous namespace
1898 
1899 void TokenAnnotator::setCommentLineLevels(
1900     SmallVectorImpl<AnnotatedLine *> &Lines) {
1901   const AnnotatedLine *NextNonCommentLine = nullptr;
1902   for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
1903                                                           E = Lines.rend();
1904        I != E; ++I) {
1905     bool CommentLine = true;
1906     for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
1907       if (!Tok->is(tok::comment)) {
1908         CommentLine = false;
1909         break;
1910       }
1911     }
1912 
1913     // If the comment is currently aligned with the line immediately following
1914     // it, that's probably intentional and we should keep it.
1915     if (NextNonCommentLine && CommentLine &&
1916         NextNonCommentLine->First->NewlinesBefore <= 1 &&
1917         NextNonCommentLine->First->OriginalColumn ==
1918             (*I)->First->OriginalColumn) {
1919       // Align comments for preprocessor lines with the # in column 0.
1920       // Otherwise, align with the next line.
1921       (*I)->Level = (NextNonCommentLine->Type == LT_PreprocessorDirective ||
1922                      NextNonCommentLine->Type == LT_ImportStatement)
1923                         ? 0
1924                         : NextNonCommentLine->Level;
1925     } else {
1926       NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
1927     }
1928 
1929     setCommentLineLevels((*I)->Children);
1930   }
1931 }
1932 
1933 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
1934   unsigned Result = 0;
1935   for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
1936     Result = std::max(Result, Tok->NestingLevel);
1937   return Result;
1938 }
1939 
1940 void TokenAnnotator::annotate(AnnotatedLine &Line) {
1941   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
1942                                                   E = Line.Children.end();
1943        I != E; ++I) {
1944     annotate(**I);
1945   }
1946   AnnotatingParser Parser(Style, Line, Keywords);
1947   Line.Type = Parser.parseLine();
1948 
1949   // With very deep nesting, ExpressionParser uses lots of stack and the
1950   // formatting algorithm is very slow. We're not going to do a good job here
1951   // anyway - it's probably generated code being formatted by mistake.
1952   // Just skip the whole line.
1953   if (maxNestingDepth(Line) > 50)
1954     Line.Type = LT_Invalid;
1955 
1956   if (Line.Type == LT_Invalid)
1957     return;
1958 
1959   ExpressionParser ExprParser(Style, Keywords, Line);
1960   ExprParser.parse();
1961 
1962   if (Line.startsWith(TT_ObjCMethodSpecifier))
1963     Line.Type = LT_ObjCMethodDecl;
1964   else if (Line.startsWith(TT_ObjCDecl))
1965     Line.Type = LT_ObjCDecl;
1966   else if (Line.startsWith(TT_ObjCProperty))
1967     Line.Type = LT_ObjCProperty;
1968 
1969   Line.First->SpacesRequiredBefore = 1;
1970   Line.First->CanBreakBefore = Line.First->MustBreakBefore;
1971 }
1972 
1973 // This function heuristically determines whether 'Current' starts the name of a
1974 // function declaration.
1975 static bool isFunctionDeclarationName(const FormatToken &Current,
1976                                       const AnnotatedLine &Line) {
1977   auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * {
1978     for (; Next; Next = Next->Next) {
1979       if (Next->is(TT_OverloadedOperatorLParen))
1980         return Next;
1981       if (Next->is(TT_OverloadedOperator))
1982         continue;
1983       if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
1984         // For 'new[]' and 'delete[]'.
1985         if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next &&
1986             Next->Next->Next->is(tok::r_square))
1987           Next = Next->Next->Next;
1988         continue;
1989       }
1990 
1991       break;
1992     }
1993     return nullptr;
1994   };
1995 
1996   // Find parentheses of parameter list.
1997   const FormatToken *Next = Current.Next;
1998   if (Current.is(tok::kw_operator)) {
1999     if (Current.Previous && Current.Previous->is(tok::coloncolon))
2000       return false;
2001     Next = skipOperatorName(Next);
2002   } else {
2003     if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
2004       return false;
2005     for (; Next; Next = Next->Next) {
2006       if (Next->is(TT_TemplateOpener)) {
2007         Next = Next->MatchingParen;
2008       } else if (Next->is(tok::coloncolon)) {
2009         Next = Next->Next;
2010         if (!Next)
2011           return false;
2012         if (Next->is(tok::kw_operator)) {
2013           Next = skipOperatorName(Next->Next);
2014           break;
2015         }
2016         if (!Next->is(tok::identifier))
2017           return false;
2018       } else if (Next->is(tok::l_paren)) {
2019         break;
2020       } else {
2021         return false;
2022       }
2023     }
2024   }
2025 
2026   // Check whether parameter list can belong to a function declaration.
2027   if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
2028     return false;
2029   // If the lines ends with "{", this is likely an function definition.
2030   if (Line.Last->is(tok::l_brace))
2031     return true;
2032   if (Next->Next == Next->MatchingParen)
2033     return true; // Empty parentheses.
2034   // If there is an &/&& after the r_paren, this is likely a function.
2035   if (Next->MatchingParen->Next &&
2036       Next->MatchingParen->Next->is(TT_PointerOrReference))
2037     return true;
2038   for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
2039        Tok = Tok->Next) {
2040     if (Tok->is(tok::l_paren) && Tok->MatchingParen) {
2041       Tok = Tok->MatchingParen;
2042       continue;
2043     }
2044     if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
2045         Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
2046       return true;
2047     if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
2048         Tok->Tok.isLiteral())
2049       return false;
2050   }
2051   return false;
2052 }
2053 
2054 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
2055   assert(Line.MightBeFunctionDecl);
2056 
2057   if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
2058        Style.AlwaysBreakAfterReturnType ==
2059            FormatStyle::RTBS_TopLevelDefinitions) &&
2060       Line.Level > 0)
2061     return false;
2062 
2063   switch (Style.AlwaysBreakAfterReturnType) {
2064   case FormatStyle::RTBS_None:
2065     return false;
2066   case FormatStyle::RTBS_All:
2067   case FormatStyle::RTBS_TopLevel:
2068     return true;
2069   case FormatStyle::RTBS_AllDefinitions:
2070   case FormatStyle::RTBS_TopLevelDefinitions:
2071     return Line.mightBeFunctionDefinition();
2072   }
2073 
2074   return false;
2075 }
2076 
2077 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
2078   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2079                                                   E = Line.Children.end();
2080        I != E; ++I) {
2081     calculateFormattingInformation(**I);
2082   }
2083 
2084   Line.First->TotalLength =
2085       Line.First->IsMultiline ? Style.ColumnLimit
2086                               : Line.FirstStartColumn + Line.First->ColumnWidth;
2087   FormatToken *Current = Line.First->Next;
2088   bool InFunctionDecl = Line.MightBeFunctionDecl;
2089   while (Current) {
2090     if (isFunctionDeclarationName(*Current, Line))
2091       Current->Type = TT_FunctionDeclarationName;
2092     if (Current->is(TT_LineComment)) {
2093       if (Current->Previous->BlockKind == BK_BracedInit &&
2094           Current->Previous->opensScope())
2095         Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1;
2096       else
2097         Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
2098 
2099       // If we find a trailing comment, iterate backwards to determine whether
2100       // it seems to relate to a specific parameter. If so, break before that
2101       // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
2102       // to the previous line in:
2103       //   SomeFunction(a,
2104       //                b, // comment
2105       //                c);
2106       if (!Current->HasUnescapedNewline) {
2107         for (FormatToken *Parameter = Current->Previous; Parameter;
2108              Parameter = Parameter->Previous) {
2109           if (Parameter->isOneOf(tok::comment, tok::r_brace))
2110             break;
2111           if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
2112             if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
2113                 Parameter->HasUnescapedNewline)
2114               Parameter->MustBreakBefore = true;
2115             break;
2116           }
2117         }
2118       }
2119     } else if (Current->SpacesRequiredBefore == 0 &&
2120                spaceRequiredBefore(Line, *Current)) {
2121       Current->SpacesRequiredBefore = 1;
2122     }
2123 
2124     Current->MustBreakBefore =
2125         Current->MustBreakBefore || mustBreakBefore(Line, *Current);
2126 
2127     if (!Current->MustBreakBefore && InFunctionDecl &&
2128         Current->is(TT_FunctionDeclarationName))
2129       Current->MustBreakBefore = mustBreakForReturnType(Line);
2130 
2131     Current->CanBreakBefore =
2132         Current->MustBreakBefore || canBreakBefore(Line, *Current);
2133     unsigned ChildSize = 0;
2134     if (Current->Previous->Children.size() == 1) {
2135       FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
2136       ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
2137                                                   : LastOfChild.TotalLength + 1;
2138     }
2139     const FormatToken *Prev = Current->Previous;
2140     if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
2141         (Prev->Children.size() == 1 &&
2142          Prev->Children[0]->First->MustBreakBefore) ||
2143         Current->IsMultiline)
2144       Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
2145     else
2146       Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
2147                              ChildSize + Current->SpacesRequiredBefore;
2148 
2149     if (Current->is(TT_CtorInitializerColon))
2150       InFunctionDecl = false;
2151 
2152     // FIXME: Only calculate this if CanBreakBefore is true once static
2153     // initializers etc. are sorted out.
2154     // FIXME: Move magic numbers to a better place.
2155 
2156     // Reduce penalty for aligning ObjC method arguments using the colon
2157     // alignment as this is the canonical way (still prefer fitting everything
2158     // into one line if possible). Trying to fit a whole expression into one
2159     // line should not force other line breaks (e.g. when ObjC method
2160     // expression is a part of other expression).
2161     Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl);
2162     if (Style.Language == FormatStyle::LK_ObjC &&
2163         Current->is(TT_SelectorName) && Current->ParameterIndex > 0) {
2164       if (Current->ParameterIndex == 1)
2165         Current->SplitPenalty += 5 * Current->BindingStrength;
2166     } else {
2167       Current->SplitPenalty += 20 * Current->BindingStrength;
2168     }
2169 
2170     Current = Current->Next;
2171   }
2172 
2173   calculateUnbreakableTailLengths(Line);
2174   unsigned IndentLevel = Line.Level;
2175   for (Current = Line.First; Current != nullptr; Current = Current->Next) {
2176     if (Current->Role)
2177       Current->Role->precomputeFormattingInfos(Current);
2178     if (Current->MatchingParen &&
2179         Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
2180       assert(IndentLevel > 0);
2181       --IndentLevel;
2182     }
2183     Current->IndentLevel = IndentLevel;
2184     if (Current->opensBlockOrBlockTypeList(Style))
2185       ++IndentLevel;
2186   }
2187 
2188   LLVM_DEBUG({ printDebugInfo(Line); });
2189 }
2190 
2191 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
2192   unsigned UnbreakableTailLength = 0;
2193   FormatToken *Current = Line.Last;
2194   while (Current) {
2195     Current->UnbreakableTailLength = UnbreakableTailLength;
2196     if (Current->CanBreakBefore ||
2197         Current->isOneOf(tok::comment, tok::string_literal)) {
2198       UnbreakableTailLength = 0;
2199     } else {
2200       UnbreakableTailLength +=
2201           Current->ColumnWidth + Current->SpacesRequiredBefore;
2202     }
2203     Current = Current->Previous;
2204   }
2205 }
2206 
2207 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
2208                                       const FormatToken &Tok,
2209                                       bool InFunctionDecl) {
2210   const FormatToken &Left = *Tok.Previous;
2211   const FormatToken &Right = Tok;
2212 
2213   if (Left.is(tok::semi))
2214     return 0;
2215 
2216   if (Style.Language == FormatStyle::LK_Java) {
2217     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
2218       return 1;
2219     if (Right.is(Keywords.kw_implements))
2220       return 2;
2221     if (Left.is(tok::comma) && Left.NestingLevel == 0)
2222       return 3;
2223   } else if (Style.Language == FormatStyle::LK_JavaScript) {
2224     if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
2225       return 100;
2226     if (Left.is(TT_JsTypeColon))
2227       return 35;
2228     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2229         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2230       return 100;
2231     // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()".
2232     if (Left.opensScope() && Right.closesScope())
2233       return 200;
2234   }
2235 
2236   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2237     return 1;
2238   if (Right.is(tok::l_square)) {
2239     if (Style.Language == FormatStyle::LK_Proto)
2240       return 1;
2241     if (Left.is(tok::r_square))
2242       return 200;
2243     // Slightly prefer formatting local lambda definitions like functions.
2244     if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
2245       return 35;
2246     if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2247                        TT_ArrayInitializerLSquare,
2248                        TT_DesignatedInitializerLSquare, TT_AttributeSquare))
2249       return 500;
2250   }
2251 
2252   if (Left.is(tok::coloncolon) ||
2253       (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
2254     return 500;
2255   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2256       Right.is(tok::kw_operator)) {
2257     if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
2258       return 3;
2259     if (Left.is(TT_StartOfName))
2260       return 110;
2261     if (InFunctionDecl && Right.NestingLevel == 0)
2262       return Style.PenaltyReturnTypeOnItsOwnLine;
2263     return 200;
2264   }
2265   if (Right.is(TT_PointerOrReference))
2266     return 190;
2267   if (Right.is(TT_LambdaArrow))
2268     return 110;
2269   if (Left.is(tok::equal) && Right.is(tok::l_brace))
2270     return 160;
2271   if (Left.is(TT_CastRParen))
2272     return 100;
2273   if (Left.isOneOf(tok::kw_class, tok::kw_struct))
2274     return 5000;
2275   if (Left.is(tok::comment))
2276     return 1000;
2277 
2278   if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon,
2279                    TT_CtorInitializerColon))
2280     return 2;
2281 
2282   if (Right.isMemberAccess()) {
2283     // Breaking before the "./->" of a chained call/member access is reasonably
2284     // cheap, as formatting those with one call per line is generally
2285     // desirable. In particular, it should be cheaper to break before the call
2286     // than it is to break inside a call's parameters, which could lead to weird
2287     // "hanging" indents. The exception is the very last "./->" to support this
2288     // frequent pattern:
2289     //
2290     //   aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
2291     //       dddddddd);
2292     //
2293     // which might otherwise be blown up onto many lines. Here, clang-format
2294     // won't produce "hanging" indents anyway as there is no other trailing
2295     // call.
2296     //
2297     // Also apply higher penalty is not a call as that might lead to a wrapping
2298     // like:
2299     //
2300     //   aaaaaaa
2301     //       .aaaaaaaaa.bbbbbbbb(cccccccc);
2302     return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
2303                ? 150
2304                : 35;
2305   }
2306 
2307   if (Right.is(TT_TrailingAnnotation) &&
2308       (!Right.Next || Right.Next->isNot(tok::l_paren))) {
2309     // Moving trailing annotations to the next line is fine for ObjC method
2310     // declarations.
2311     if (Line.startsWith(TT_ObjCMethodSpecifier))
2312       return 10;
2313     // Generally, breaking before a trailing annotation is bad unless it is
2314     // function-like. It seems to be especially preferable to keep standard
2315     // annotations (i.e. "const", "final" and "override") on the same line.
2316     // Use a slightly higher penalty after ")" so that annotations like
2317     // "const override" are kept together.
2318     bool is_short_annotation = Right.TokenText.size() < 10;
2319     return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2320   }
2321 
2322   // In for-loops, prefer breaking at ',' and ';'.
2323   if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2324     return 4;
2325 
2326   // In Objective-C method expressions, prefer breaking before "param:" over
2327   // breaking after it.
2328   if (Right.is(TT_SelectorName))
2329     return 0;
2330   if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2331     return Line.MightBeFunctionDecl ? 50 : 500;
2332 
2333   // In Objective-C type declarations, avoid breaking after the category's
2334   // open paren (we'll prefer breaking after the protocol list's opening
2335   // angle bracket, if present).
2336   if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous &&
2337       Left.Previous->isOneOf(tok::identifier, tok::greater))
2338     return 500;
2339 
2340   if (Left.is(tok::l_paren) && InFunctionDecl &&
2341       Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2342     return 100;
2343   if (Left.is(tok::l_paren) && Left.Previous &&
2344       (Left.Previous->isOneOf(tok::kw_if, tok::kw_for) ||
2345        Left.Previous->endsSequence(tok::kw_constexpr, tok::kw_if)))
2346     return 1000;
2347   if (Left.is(tok::equal) && InFunctionDecl)
2348     return 110;
2349   if (Right.is(tok::r_brace))
2350     return 1;
2351   if (Left.is(TT_TemplateOpener))
2352     return 100;
2353   if (Left.opensScope()) {
2354     if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
2355       return 0;
2356     if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle)
2357       return 19;
2358     return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2359                                    : 19;
2360   }
2361   if (Left.is(TT_JavaAnnotation))
2362     return 50;
2363 
2364   if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2365       Left.Previous->isLabelString() &&
2366       (Left.NextOperator || Left.OperatorIndex != 0))
2367     return 50;
2368   if (Right.is(tok::plus) && Left.isLabelString() &&
2369       (Right.NextOperator || Right.OperatorIndex != 0))
2370     return 25;
2371   if (Left.is(tok::comma))
2372     return 1;
2373   if (Right.is(tok::lessless) && Left.isLabelString() &&
2374       (Right.NextOperator || Right.OperatorIndex != 1))
2375     return 25;
2376   if (Right.is(tok::lessless)) {
2377     // Breaking at a << is really cheap.
2378     if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2379       // Slightly prefer to break before the first one in log-like statements.
2380       return 2;
2381     return 1;
2382   }
2383   if (Left.ClosesTemplateDeclaration)
2384     return Style.PenaltyBreakTemplateDeclaration;
2385   if (Left.is(TT_ConditionalExpr))
2386     return prec::Conditional;
2387   prec::Level Level = Left.getPrecedence();
2388   if (Level == prec::Unknown)
2389     Level = Right.getPrecedence();
2390   if (Level == prec::Assignment)
2391     return Style.PenaltyBreakAssignment;
2392   if (Level != prec::Unknown)
2393     return Level;
2394 
2395   return 3;
2396 }
2397 
2398 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2399                                           const FormatToken &Left,
2400                                           const FormatToken &Right) {
2401   if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2402     return true;
2403   if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java)
2404     return true;
2405   if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2406       Left.Tok.getObjCKeywordID() == tok::objc_property)
2407     return true;
2408   if (Right.is(tok::hashhash))
2409     return Left.is(tok::hash);
2410   if (Left.isOneOf(tok::hashhash, tok::hash))
2411     return Right.is(tok::hash);
2412   if (Left.is(tok::l_paren) && Right.is(tok::r_paren))
2413     return Style.SpaceInEmptyParentheses;
2414   if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2415     return (Right.is(TT_CastRParen) ||
2416             (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2417                ? Style.SpacesInCStyleCastParentheses
2418                : Style.SpacesInParentheses;
2419   if (Right.isOneOf(tok::semi, tok::comma))
2420     return false;
2421   if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) {
2422     bool IsLightweightGeneric =
2423         Right.MatchingParen && Right.MatchingParen->Next &&
2424         Right.MatchingParen->Next->is(tok::colon);
2425     return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList;
2426   }
2427   if (Right.is(tok::less) && Left.is(tok::kw_template))
2428     return Style.SpaceAfterTemplateKeyword;
2429   if (Left.isOneOf(tok::exclaim, tok::tilde))
2430     return false;
2431   if (Left.is(tok::at) &&
2432       Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2433                     tok::numeric_constant, tok::l_paren, tok::l_brace,
2434                     tok::kw_true, tok::kw_false))
2435     return false;
2436   if (Left.is(tok::colon))
2437     return !Left.is(TT_ObjCMethodExpr);
2438   if (Left.is(tok::coloncolon))
2439     return false;
2440   if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) {
2441     if (Style.Language == FormatStyle::LK_TextProto ||
2442         (Style.Language == FormatStyle::LK_Proto &&
2443          (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) {
2444       // Format empty list as `<>`.
2445       if (Left.is(tok::less) && Right.is(tok::greater))
2446         return false;
2447       return !Style.Cpp11BracedListStyle;
2448     }
2449     return false;
2450   }
2451   if (Right.is(tok::ellipsis))
2452     return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2453                                     Left.Previous->is(tok::kw_case));
2454   if (Left.is(tok::l_square) && Right.is(tok::amp))
2455     return false;
2456   if (Right.is(TT_PointerOrReference)) {
2457     if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) {
2458       if (!Left.MatchingParen)
2459         return true;
2460       FormatToken *TokenBeforeMatchingParen =
2461           Left.MatchingParen->getPreviousNonComment();
2462       if (!TokenBeforeMatchingParen ||
2463           !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype))
2464         return true;
2465     }
2466     return (Left.Tok.isLiteral() ||
2467             (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2468              (Style.PointerAlignment != FormatStyle::PAS_Left ||
2469               (Line.IsMultiVariableDeclStmt &&
2470                (Left.NestingLevel == 0 ||
2471                 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2472   }
2473   if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2474       (!Left.is(TT_PointerOrReference) ||
2475        (Style.PointerAlignment != FormatStyle::PAS_Right &&
2476         !Line.IsMultiVariableDeclStmt)))
2477     return true;
2478   if (Left.is(TT_PointerOrReference))
2479     return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2480            (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2481             !Right.is(TT_StartOfName)) ||
2482            (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) ||
2483            (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2484                            tok::l_paren) &&
2485             (Style.PointerAlignment != FormatStyle::PAS_Right &&
2486              !Line.IsMultiVariableDeclStmt) &&
2487             Left.Previous &&
2488             !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon));
2489   if (Right.is(tok::star) && Left.is(tok::l_paren))
2490     return false;
2491   const auto SpaceRequiredForArrayInitializerLSquare =
2492       [](const FormatToken &LSquareTok, const FormatStyle &Style) {
2493         return Style.SpacesInContainerLiterals ||
2494                ((Style.Language == FormatStyle::LK_Proto ||
2495                  Style.Language == FormatStyle::LK_TextProto) &&
2496                 !Style.Cpp11BracedListStyle &&
2497                 LSquareTok.endsSequence(tok::l_square, tok::colon,
2498                                         TT_SelectorName));
2499       };
2500   if (Left.is(tok::l_square))
2501     return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) &&
2502             SpaceRequiredForArrayInitializerLSquare(Left, Style)) ||
2503            (Left.isOneOf(TT_ArraySubscriptLSquare,
2504                          TT_StructuredBindingLSquare) &&
2505             Style.SpacesInSquareBrackets && Right.isNot(tok::r_square));
2506   if (Right.is(tok::r_square))
2507     return Right.MatchingParen &&
2508            ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) &&
2509              SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen,
2510                                                      Style)) ||
2511             (Style.SpacesInSquareBrackets &&
2512              Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare,
2513                                           TT_StructuredBindingLSquare)) ||
2514             Right.MatchingParen->is(TT_AttributeParen));
2515   if (Right.is(tok::l_square) &&
2516       !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2517                      TT_DesignatedInitializerLSquare,
2518                      TT_StructuredBindingLSquare, TT_AttributeSquare) &&
2519       !Left.isOneOf(tok::numeric_constant, TT_DictLiteral))
2520     return false;
2521   if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2522     return !Left.Children.empty(); // No spaces in "{}".
2523   if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) ||
2524       (Right.is(tok::r_brace) && Right.MatchingParen &&
2525        Right.MatchingParen->BlockKind != BK_Block))
2526     return !Style.Cpp11BracedListStyle;
2527   if (Left.is(TT_BlockComment))
2528     // No whitespace in x(/*foo=*/1), except for JavaScript.
2529     return Style.Language == FormatStyle::LK_JavaScript ||
2530            !Left.TokenText.endswith("=*/");
2531   if (Right.is(tok::l_paren)) {
2532     if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) ||
2533         (Left.is(tok::r_square) && Left.is(TT_AttributeSquare)))
2534       return true;
2535     return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
2536            (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
2537             (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while,
2538                           tok::kw_switch, tok::kw_case, TT_ForEachMacro,
2539                           TT_ObjCForIn) ||
2540              Left.endsSequence(tok::kw_constexpr, tok::kw_if) ||
2541              (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
2542                            tok::kw_new, tok::kw_delete) &&
2543               (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
2544            (Style.SpaceBeforeParens == FormatStyle::SBPO_Always &&
2545             (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
2546              Left.is(tok::r_paren)) &&
2547             Line.Type != LT_PreprocessorDirective);
2548   }
2549   if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
2550     return false;
2551   if (Right.is(TT_UnaryOperator))
2552     return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
2553            (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
2554   if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
2555                     tok::r_paren) ||
2556        Left.isSimpleTypeSpecifier()) &&
2557       Right.is(tok::l_brace) && Right.getNextNonComment() &&
2558       Right.BlockKind != BK_Block)
2559     return false;
2560   if (Left.is(tok::period) || Right.is(tok::period))
2561     return false;
2562   if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
2563     return false;
2564   if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
2565       Left.MatchingParen->Previous &&
2566       (Left.MatchingParen->Previous->is(tok::period) ||
2567        Left.MatchingParen->Previous->is(tok::coloncolon)))
2568     // Java call to generic function with explicit type:
2569     // A.<B<C<...>>>DoSomething();
2570     // A::<B<C<...>>>DoSomething();  // With a Java 8 method reference.
2571     return false;
2572   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
2573     return false;
2574   if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at))
2575     // Objective-C dictionary literal -> no space after opening brace.
2576     return false;
2577   if (Right.is(tok::r_brace) && Right.MatchingParen &&
2578       Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at))
2579     // Objective-C dictionary literal -> no space before closing brace.
2580     return false;
2581   return true;
2582 }
2583 
2584 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
2585                                          const FormatToken &Right) {
2586   const FormatToken &Left = *Right.Previous;
2587   if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
2588     return true; // Never ever merge two identifiers.
2589   if (Style.isCpp()) {
2590     if (Left.is(tok::kw_operator))
2591       return Right.is(tok::coloncolon);
2592     if (Right.is(tok::l_brace) && Right.BlockKind == BK_BracedInit &&
2593         !Left.opensScope() && Style.SpaceBeforeCpp11BracedList)
2594       return true;
2595   } else if (Style.Language == FormatStyle::LK_Proto ||
2596              Style.Language == FormatStyle::LK_TextProto) {
2597     if (Right.is(tok::period) &&
2598         Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
2599                      Keywords.kw_repeated, Keywords.kw_extend))
2600       return true;
2601     if (Right.is(tok::l_paren) &&
2602         Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
2603       return true;
2604     if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName))
2605       return true;
2606     // Slashes occur in text protocol extension syntax: [type/type] { ... }.
2607     if (Left.is(tok::slash) || Right.is(tok::slash))
2608       return false;
2609     if (Left.MatchingParen && Left.MatchingParen->is(TT_ProtoExtensionLSquare) &&
2610         Right.isOneOf(tok::l_brace, tok::less))
2611       return !Style.Cpp11BracedListStyle;
2612     // A percent is probably part of a formatting specification, such as %lld.
2613     if (Left.is(tok::percent))
2614       return false;
2615     // Preserve the existence of a space before a percent for cases like 0x%04x
2616     // and "%d %d"
2617     if (Left.is(tok::numeric_constant) && Right.is(tok::percent))
2618       return Right.WhitespaceRange.getEnd() != Right.WhitespaceRange.getBegin();
2619   } else if (Style.Language == FormatStyle::LK_JavaScript) {
2620     if (Left.is(TT_JsFatArrow))
2621       return true;
2622     // for await ( ...
2623     if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous &&
2624         Left.Previous->is(tok::kw_for))
2625       return true;
2626     if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
2627         Right.MatchingParen) {
2628       const FormatToken *Next = Right.MatchingParen->getNextNonComment();
2629       // An async arrow function, for example: `x = async () => foo();`,
2630       // as opposed to calling a function called async: `x = async();`
2631       if (Next && Next->is(TT_JsFatArrow))
2632         return true;
2633     }
2634     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2635         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2636       return false;
2637     // In tagged template literals ("html`bar baz`"), there is no space between
2638     // the tag identifier and the template string. getIdentifierInfo makes sure
2639     // that the identifier is not a pseudo keyword like `yield`, either.
2640     if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) &&
2641         Right.is(TT_TemplateString))
2642       return false;
2643     if (Right.is(tok::star) &&
2644         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2645       return false;
2646     if (Right.isOneOf(tok::l_brace, tok::l_square) &&
2647         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield,
2648                      Keywords.kw_extends, Keywords.kw_implements))
2649       return true;
2650     if (Right.is(tok::l_paren)) {
2651       // JS methods can use some keywords as names (e.g. `delete()`).
2652       if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo())
2653         return false;
2654       // Valid JS method names can include keywords, e.g. `foo.delete()` or
2655       // `bar.instanceof()`. Recognize call positions by preceding period.
2656       if (Left.Previous && Left.Previous->is(tok::period) &&
2657           Left.Tok.getIdentifierInfo())
2658         return false;
2659       // Additional unary JavaScript operators that need a space after.
2660       if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof,
2661                        tok::kw_void))
2662         return true;
2663     }
2664     if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
2665                       tok::kw_const) ||
2666          // "of" is only a keyword if it appears after another identifier
2667          // (e.g. as "const x of y" in a for loop), or after a destructuring
2668          // operation (const [x, y] of z, const {a, b} of c).
2669          (Left.is(Keywords.kw_of) && Left.Previous &&
2670           (Left.Previous->Tok.is(tok::identifier) ||
2671            Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) &&
2672         (!Left.Previous || !Left.Previous->is(tok::period)))
2673       return true;
2674     if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
2675         Left.Previous->is(tok::period) && Right.is(tok::l_paren))
2676       return false;
2677     if (Left.is(Keywords.kw_as) &&
2678         Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
2679       return true;
2680     if (Left.is(tok::kw_default) && Left.Previous &&
2681         Left.Previous->is(tok::kw_export))
2682       return true;
2683     if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
2684       return true;
2685     if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
2686       return false;
2687     if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
2688       return false;
2689     if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
2690         Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
2691       return false;
2692     if (Left.is(tok::ellipsis))
2693       return false;
2694     if (Left.is(TT_TemplateCloser) &&
2695         !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
2696                        Keywords.kw_implements, Keywords.kw_extends))
2697       // Type assertions ('<type>expr') are not followed by whitespace. Other
2698       // locations that should have whitespace following are identified by the
2699       // above set of follower tokens.
2700       return false;
2701     if (Right.is(TT_JsNonNullAssertion))
2702       return false;
2703     if (Left.is(TT_JsNonNullAssertion) &&
2704         Right.isOneOf(Keywords.kw_as, Keywords.kw_in))
2705       return true; // "x! as string", "x! in y"
2706   } else if (Style.Language == FormatStyle::LK_Java) {
2707     if (Left.is(tok::r_square) && Right.is(tok::l_brace))
2708       return true;
2709     if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
2710       return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
2711     if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
2712                       tok::kw_protected) ||
2713          Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
2714                       Keywords.kw_native)) &&
2715         Right.is(TT_TemplateOpener))
2716       return true;
2717   }
2718   if (Left.is(TT_ImplicitStringLiteral))
2719     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2720   if (Line.Type == LT_ObjCMethodDecl) {
2721     if (Left.is(TT_ObjCMethodSpecifier))
2722       return true;
2723     if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right))
2724       // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a
2725       // keyword in Objective-C, and '+ (instancetype)new;' is a standard class
2726       // method declaration.
2727       return false;
2728   }
2729   if (Line.Type == LT_ObjCProperty &&
2730       (Right.is(tok::equal) || Left.is(tok::equal)))
2731     return false;
2732 
2733   if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
2734       Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
2735     return true;
2736   if (Right.is(TT_OverloadedOperatorLParen))
2737     return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2738   if (Left.is(tok::comma))
2739     return true;
2740   if (Right.is(tok::comma))
2741     return false;
2742   if (Right.is(TT_ObjCBlockLParen))
2743     return true;
2744   if (Right.is(TT_CtorInitializerColon))
2745     return Style.SpaceBeforeCtorInitializerColon;
2746   if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon)
2747     return false;
2748   if (Right.is(TT_RangeBasedForLoopColon) &&
2749       !Style.SpaceBeforeRangeBasedForLoopColon)
2750     return false;
2751   if (Right.is(tok::colon)) {
2752     if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
2753         !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
2754       return false;
2755     if (Right.is(TT_ObjCMethodExpr))
2756       return false;
2757     if (Left.is(tok::question))
2758       return false;
2759     if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
2760       return false;
2761     if (Right.is(TT_DictLiteral))
2762       return Style.SpacesInContainerLiterals;
2763     if (Right.is(TT_AttributeColon))
2764       return false;
2765     return true;
2766   }
2767   if (Left.is(TT_UnaryOperator))
2768     return Right.is(TT_BinaryOperator);
2769 
2770   // If the next token is a binary operator or a selector name, we have
2771   // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
2772   if (Left.is(TT_CastRParen))
2773     return Style.SpaceAfterCStyleCast ||
2774            Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
2775 
2776   if (Left.is(tok::greater) && Right.is(tok::greater)) {
2777     if (Style.Language == FormatStyle::LK_TextProto ||
2778         (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral)))
2779       return !Style.Cpp11BracedListStyle;
2780     return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
2781            (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
2782   }
2783   if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) ||
2784       Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
2785       (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod)))
2786     return false;
2787   if (!Style.SpaceBeforeAssignmentOperators &&
2788       Right.getPrecedence() == prec::Assignment)
2789     return false;
2790   if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) &&
2791       (Left.is(tok::identifier) || Left.is(tok::kw_this)))
2792     return false;
2793   if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
2794     // Generally don't remove existing spaces between an identifier and "::".
2795     // The identifier might actually be a macro name such as ALWAYS_INLINE. If
2796     // this turns out to be too lenient, add analysis of the identifier itself.
2797     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2798   if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment))
2799     return (Left.is(TT_TemplateOpener) &&
2800             Style.Standard == FormatStyle::LS_Cpp03) ||
2801            !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
2802                           tok::kw___super, TT_TemplateCloser,
2803                           TT_TemplateOpener)) ||
2804            (Left.is(tok ::l_paren) && Style.SpacesInParentheses);
2805   if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
2806     return Style.SpacesInAngles;
2807   // Space before TT_StructuredBindingLSquare.
2808   if (Right.is(TT_StructuredBindingLSquare))
2809     return !Left.isOneOf(tok::amp, tok::ampamp) ||
2810            Style.PointerAlignment != FormatStyle::PAS_Right;
2811   // Space before & or && following a TT_StructuredBindingLSquare.
2812   if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) &&
2813       Right.isOneOf(tok::amp, tok::ampamp))
2814     return Style.PointerAlignment != FormatStyle::PAS_Left;
2815   if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
2816       (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
2817        !Right.is(tok::r_paren)))
2818     return true;
2819   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
2820       Right.isNot(TT_FunctionTypeLParen))
2821     return Style.SpaceBeforeParens == FormatStyle::SBPO_Always;
2822   if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
2823       Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
2824     return false;
2825   if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
2826       Line.startsWith(tok::hash))
2827     return true;
2828   if (Right.is(TT_TrailingUnaryOperator))
2829     return false;
2830   if (Left.is(TT_RegexLiteral))
2831     return false;
2832   return spaceRequiredBetween(Line, Left, Right);
2833 }
2834 
2835 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
2836 static bool isAllmanBrace(const FormatToken &Tok) {
2837   return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block &&
2838          !Tok.isOneOf(TT_ObjCBlockLBrace, TT_DictLiteral);
2839 }
2840 
2841 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
2842                                      const FormatToken &Right) {
2843   const FormatToken &Left = *Right.Previous;
2844   if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
2845     return true;
2846 
2847   if (Style.Language == FormatStyle::LK_JavaScript) {
2848     // FIXME: This might apply to other languages and token kinds.
2849     if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
2850         Left.Previous->is(tok::string_literal))
2851       return true;
2852     if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
2853         Left.Previous && Left.Previous->is(tok::equal) &&
2854         Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
2855                             tok::kw_const) &&
2856         // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
2857         // above.
2858         !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
2859       // Object literals on the top level of a file are treated as "enum-style".
2860       // Each key/value pair is put on a separate line, instead of bin-packing.
2861       return true;
2862     if (Left.is(tok::l_brace) && Line.Level == 0 &&
2863         (Line.startsWith(tok::kw_enum) ||
2864          Line.startsWith(tok::kw_const, tok::kw_enum) ||
2865          Line.startsWith(tok::kw_export, tok::kw_enum) ||
2866          Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum)))
2867       // JavaScript top-level enum key/value pairs are put on separate lines
2868       // instead of bin-packing.
2869       return true;
2870     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
2871         !Left.Children.empty())
2872       // Support AllowShortFunctionsOnASingleLine for JavaScript.
2873       return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
2874              Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
2875              (Left.NestingLevel == 0 && Line.Level == 0 &&
2876               Style.AllowShortFunctionsOnASingleLine &
2877                   FormatStyle::SFS_InlineOnly);
2878   } else if (Style.Language == FormatStyle::LK_Java) {
2879     if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
2880         Right.Next->is(tok::string_literal))
2881       return true;
2882   } else if (Style.Language == FormatStyle::LK_Cpp ||
2883              Style.Language == FormatStyle::LK_ObjC ||
2884              Style.Language == FormatStyle::LK_Proto ||
2885              Style.Language == FormatStyle::LK_TableGen ||
2886              Style.Language == FormatStyle::LK_TextProto) {
2887     if (Left.isStringLiteral() && Right.isStringLiteral())
2888       return true;
2889   }
2890 
2891   // If the last token before a '}', ']', or ')' is a comma or a trailing
2892   // comment, the intention is to insert a line break after it in order to make
2893   // shuffling around entries easier. Import statements, especially in
2894   // JavaScript, can be an exception to this rule.
2895   if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) {
2896     const FormatToken *BeforeClosingBrace = nullptr;
2897     if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
2898          (Style.Language == FormatStyle::LK_JavaScript &&
2899           Left.is(tok::l_paren))) &&
2900         Left.BlockKind != BK_Block && Left.MatchingParen)
2901       BeforeClosingBrace = Left.MatchingParen->Previous;
2902     else if (Right.MatchingParen &&
2903              (Right.MatchingParen->isOneOf(tok::l_brace,
2904                                            TT_ArrayInitializerLSquare) ||
2905               (Style.Language == FormatStyle::LK_JavaScript &&
2906                Right.MatchingParen->is(tok::l_paren))))
2907       BeforeClosingBrace = &Left;
2908     if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
2909                                BeforeClosingBrace->isTrailingComment()))
2910       return true;
2911   }
2912 
2913   if (Right.is(tok::comment))
2914     return Left.BlockKind != BK_BracedInit &&
2915            Left.isNot(TT_CtorInitializerColon) &&
2916            (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
2917   if (Left.isTrailingComment())
2918     return true;
2919   if (Right.Previous->IsUnterminatedLiteral)
2920     return true;
2921   if (Right.is(tok::lessless) && Right.Next &&
2922       Right.Previous->is(tok::string_literal) &&
2923       Right.Next->is(tok::string_literal))
2924     return true;
2925   if (Right.Previous->ClosesTemplateDeclaration &&
2926       Right.Previous->MatchingParen &&
2927       Right.Previous->MatchingParen->NestingLevel == 0 &&
2928       Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes)
2929     return true;
2930   if (Right.is(TT_CtorInitializerComma) &&
2931       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
2932       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
2933     return true;
2934   if (Right.is(TT_CtorInitializerColon) &&
2935       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
2936       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
2937     return true;
2938   // Break only if we have multiple inheritance.
2939   if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma &&
2940       Right.is(TT_InheritanceComma))
2941     return true;
2942   if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
2943     // Multiline raw string literals are special wrt. line breaks. The author
2944     // has made a deliberate choice and might have aligned the contents of the
2945     // string literal accordingly. Thus, we try keep existing line breaks.
2946     return Right.IsMultiline && Right.NewlinesBefore > 0;
2947   if ((Right.Previous->is(tok::l_brace) ||
2948        (Right.Previous->is(tok::less) && Right.Previous->Previous &&
2949         Right.Previous->Previous->is(tok::equal))) &&
2950       Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) {
2951     // Don't put enums or option definitions onto single lines in protocol
2952     // buffers.
2953     return true;
2954   }
2955   if (Right.is(TT_InlineASMBrace))
2956     return Right.HasUnescapedNewline;
2957   if (isAllmanBrace(Left) || isAllmanBrace(Right))
2958     return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
2959            (Line.startsWith(tok::kw_typedef, tok::kw_enum) &&
2960             Style.BraceWrapping.AfterEnum) ||
2961            (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
2962            (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
2963   if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine)
2964     return true;
2965 
2966   if ((Style.Language == FormatStyle::LK_Java ||
2967        Style.Language == FormatStyle::LK_JavaScript) &&
2968       Left.is(TT_LeadingJavaAnnotation) &&
2969       Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
2970       (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
2971     return true;
2972 
2973   if (Right.is(TT_ProtoExtensionLSquare))
2974     return true;
2975 
2976   // In text proto instances if a submessage contains at least 2 entries and at
2977   // least one of them is a submessage, like A { ... B { ... } ... },
2978   // put all of the entries of A on separate lines by forcing the selector of
2979   // the submessage B to be put on a newline.
2980   //
2981   // Example: these can stay on one line:
2982   // a { scalar_1: 1 scalar_2: 2 }
2983   // a { b { key: value } }
2984   //
2985   // and these entries need to be on a new line even if putting them all in one
2986   // line is under the column limit:
2987   // a {
2988   //   scalar: 1
2989   //   b { key: value }
2990   // }
2991   //
2992   // We enforce this by breaking before a submessage field that has previous
2993   // siblings, *and* breaking before a field that follows a submessage field.
2994   //
2995   // Be careful to exclude the case  [proto.ext] { ... } since the `]` is
2996   // the TT_SelectorName there, but we don't want to break inside the brackets.
2997   //
2998   // Another edge case is @submessage { key: value }, which is a common
2999   // substitution placeholder. In this case we want to keep `@` and `submessage`
3000   // together.
3001   //
3002   // We ensure elsewhere that extensions are always on their own line.
3003   if ((Style.Language == FormatStyle::LK_Proto ||
3004        Style.Language == FormatStyle::LK_TextProto) &&
3005       Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) {
3006     // Keep `@submessage` together in:
3007     // @submessage { key: value }
3008     if (Right.Previous && Right.Previous->is(tok::at))
3009       return false;
3010     // Look for the scope opener after selector in cases like:
3011     // selector { ...
3012     // selector: { ...
3013     // selector: @base { ...
3014     FormatToken *LBrace = Right.Next;
3015     if (LBrace && LBrace->is(tok::colon)) {
3016       LBrace = LBrace->Next;
3017       if (LBrace && LBrace->is(tok::at)) {
3018         LBrace = LBrace->Next;
3019         if (LBrace)
3020           LBrace = LBrace->Next;
3021       }
3022     }
3023     if (LBrace &&
3024         // The scope opener is one of {, [, <:
3025         // selector { ... }
3026         // selector [ ... ]
3027         // selector < ... >
3028         //
3029         // In case of selector { ... }, the l_brace is TT_DictLiteral.
3030         // In case of an empty selector {}, the l_brace is not TT_DictLiteral,
3031         // so we check for immediately following r_brace.
3032         ((LBrace->is(tok::l_brace) &&
3033           (LBrace->is(TT_DictLiteral) ||
3034            (LBrace->Next && LBrace->Next->is(tok::r_brace)))) ||
3035          LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) {
3036       // If Left.ParameterCount is 0, then this submessage entry is not the
3037       // first in its parent submessage, and we want to break before this entry.
3038       // If Left.ParameterCount is greater than 0, then its parent submessage
3039       // might contain 1 or more entries and we want to break before this entry
3040       // if it contains at least 2 entries. We deal with this case later by
3041       // detecting and breaking before the next entry in the parent submessage.
3042       if (Left.ParameterCount == 0)
3043         return true;
3044       // However, if this submessage is the first entry in its parent
3045       // submessage, Left.ParameterCount might be 1 in some cases.
3046       // We deal with this case later by detecting an entry
3047       // following a closing paren of this submessage.
3048     }
3049 
3050     // If this is an entry immediately following a submessage, it will be
3051     // preceded by a closing paren of that submessage, like in:
3052     //     left---.  .---right
3053     //            v  v
3054     // sub: { ... } key: value
3055     // If there was a comment between `}` an `key` above, then `key` would be
3056     // put on a new line anyways.
3057     if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square))
3058       return true;
3059   }
3060 
3061   // Deal with lambda arguments in C++ - we want consistent line breaks whether
3062   // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced
3063   // as aggressive line breaks are placed when the lambda is not the last arg.
3064   if ((Style.Language == FormatStyle::LK_Cpp ||
3065        Style.Language == FormatStyle::LK_ObjC) &&
3066       Left.is(tok::l_paren) && Left.BlockParameterCount > 0 &&
3067       !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) {
3068     // Multiple lambdas in the same function call force line breaks.
3069     if (Left.BlockParameterCount > 1)
3070       return true;
3071 
3072     // A lambda followed by another arg forces a line break.
3073     if (!Left.Role)
3074       return false;
3075     auto Comma = Left.Role->lastComma();
3076     if (!Comma)
3077       return false;
3078     auto Next = Comma->getNextNonComment();
3079     if (!Next)
3080       return false;
3081     if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret))
3082       return true;
3083   }
3084 
3085   return false;
3086 }
3087 
3088 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
3089                                     const FormatToken &Right) {
3090   const FormatToken &Left = *Right.Previous;
3091 
3092   // Language-specific stuff.
3093   if (Style.Language == FormatStyle::LK_Java) {
3094     if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3095                      Keywords.kw_implements))
3096       return false;
3097     if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3098                       Keywords.kw_implements))
3099       return true;
3100   } else if (Style.Language == FormatStyle::LK_JavaScript) {
3101     const FormatToken *NonComment = Right.getPreviousNonComment();
3102     if (NonComment &&
3103         NonComment->isOneOf(
3104             tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break,
3105             tok::kw_throw, Keywords.kw_interface, Keywords.kw_type,
3106             tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected,
3107             Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get,
3108             Keywords.kw_set, Keywords.kw_async, Keywords.kw_await))
3109       return false; // Otherwise automatic semicolon insertion would trigger.
3110     if (Right.NestingLevel == 0 &&
3111         (Left.Tok.getIdentifierInfo() ||
3112          Left.isOneOf(tok::r_square, tok::r_paren)) &&
3113         Right.isOneOf(tok::l_square, tok::l_paren))
3114       return false; // Otherwise automatic semicolon insertion would trigger.
3115     if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
3116       return false;
3117     if (Left.is(TT_JsTypeColon))
3118       return true;
3119     // Don't wrap between ":" and "!" of a strict prop init ("field!: type;").
3120     if (Left.is(tok::exclaim) && Right.is(tok::colon))
3121       return false;
3122     // Look for is type annotations like:
3123     // function f(): a is B { ... }
3124     // Do not break before is in these cases.
3125     if (Right.is(Keywords.kw_is)) {
3126       const FormatToken* Next = Right.getNextNonComment();
3127       // If `is` is followed by a colon, it's likely that it's a dict key, so
3128       // ignore it for this check.
3129       // For example this is common in Polymer:
3130       // Polymer({
3131       //   is: 'name',
3132       //   ...
3133       // });
3134       if (!Next || !Next->is(tok::colon))
3135         return false;
3136     }
3137     if (Left.is(Keywords.kw_in))
3138       return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
3139     if (Right.is(Keywords.kw_in))
3140       return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
3141     if (Right.is(Keywords.kw_as))
3142       return false; // must not break before as in 'x as type' casts
3143     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) {
3144       // extends and infer can appear as keywords in conditional types:
3145       //   https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types
3146       // do not break before them, as the expressions are subject to ASI.
3147       return false;
3148     }
3149     if (Left.is(Keywords.kw_as))
3150       return true;
3151     if (Left.is(TT_JsNonNullAssertion))
3152       return true;
3153     if (Left.is(Keywords.kw_declare) &&
3154         Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
3155                       Keywords.kw_function, tok::kw_class, tok::kw_enum,
3156                       Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
3157                       Keywords.kw_let, tok::kw_const))
3158       // See grammar for 'declare' statements at:
3159       // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
3160       return false;
3161     if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
3162         Right.isOneOf(tok::identifier, tok::string_literal))
3163       return false; // must not break in "module foo { ...}"
3164     if (Right.is(TT_TemplateString) && Right.closesScope())
3165       return false;
3166     if (Left.is(TT_TemplateString) && Left.opensScope())
3167       return true;
3168   }
3169 
3170   if (Left.is(tok::at))
3171     return false;
3172   if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
3173     return false;
3174   if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
3175     return !Right.is(tok::l_paren);
3176   if (Right.is(TT_PointerOrReference))
3177     return Line.IsMultiVariableDeclStmt ||
3178            (Style.PointerAlignment == FormatStyle::PAS_Right &&
3179             (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
3180   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
3181       Right.is(tok::kw_operator))
3182     return true;
3183   if (Left.is(TT_PointerOrReference))
3184     return false;
3185   if (Right.isTrailingComment())
3186     // We rely on MustBreakBefore being set correctly here as we should not
3187     // change the "binding" behavior of a comment.
3188     // The first comment in a braced lists is always interpreted as belonging to
3189     // the first list element. Otherwise, it should be placed outside of the
3190     // list.
3191     return Left.BlockKind == BK_BracedInit ||
3192            (Left.is(TT_CtorInitializerColon) &&
3193             Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon);
3194   if (Left.is(tok::question) && Right.is(tok::colon))
3195     return false;
3196   if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
3197     return Style.BreakBeforeTernaryOperators;
3198   if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
3199     return !Style.BreakBeforeTernaryOperators;
3200   if (Left.is(TT_InheritanceColon))
3201     return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon;
3202   if (Right.is(TT_InheritanceColon))
3203     return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon;
3204   if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
3205       Left.isNot(TT_SelectorName))
3206     return true;
3207 
3208   if (Right.is(tok::colon) &&
3209       !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
3210     return false;
3211   if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
3212     if (Style.Language == FormatStyle::LK_Proto ||
3213         Style.Language == FormatStyle::LK_TextProto) {
3214       if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral())
3215         return false;
3216       // Prevent cases like:
3217       //
3218       // submessage:
3219       //     { key: valueeeeeeeeeeee }
3220       //
3221       // when the snippet does not fit into one line.
3222       // Prefer:
3223       //
3224       // submessage: {
3225       //   key: valueeeeeeeeeeee
3226       // }
3227       //
3228       // instead, even if it is longer by one line.
3229       //
3230       // Note that this allows allows the "{" to go over the column limit
3231       // when the column limit is just between ":" and "{", but that does
3232       // not happen too often and alternative formattings in this case are
3233       // not much better.
3234       //
3235       // The code covers the cases:
3236       //
3237       // submessage: { ... }
3238       // submessage: < ... >
3239       // repeated: [ ... ]
3240       if (((Right.is(tok::l_brace) || Right.is(tok::less)) &&
3241            Right.is(TT_DictLiteral)) ||
3242           Right.is(TT_ArrayInitializerLSquare))
3243         return false;
3244     }
3245     return true;
3246   }
3247   if (Right.is(tok::r_square) && Right.MatchingParen &&
3248       Right.MatchingParen->is(TT_ProtoExtensionLSquare))
3249     return false;
3250   if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
3251                                     Right.Next->is(TT_ObjCMethodExpr)))
3252     return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
3253   if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
3254     return true;
3255   if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
3256     return true;
3257   if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
3258                     TT_OverloadedOperator))
3259     return false;
3260   if (Left.is(TT_RangeBasedForLoopColon))
3261     return true;
3262   if (Right.is(TT_RangeBasedForLoopColon))
3263     return false;
3264   if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
3265     return true;
3266   if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
3267       Left.is(tok::kw_operator))
3268     return false;
3269   if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
3270       Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
3271     return false;
3272   if (Left.is(tok::equal) && Right.is(tok::l_brace) &&
3273       !Style.Cpp11BracedListStyle)
3274     return false;
3275   if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen))
3276     return false;
3277   if (Left.is(tok::l_paren) && Left.Previous &&
3278       (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
3279     return false;
3280   if (Right.is(TT_ImplicitStringLiteral))
3281     return false;
3282 
3283   if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
3284     return false;
3285   if (Right.is(tok::r_square) && Right.MatchingParen &&
3286       Right.MatchingParen->is(TT_LambdaLSquare))
3287     return false;
3288 
3289   // We only break before r_brace if there was a corresponding break before
3290   // the l_brace, which is tracked by BreakBeforeClosingBrace.
3291   if (Right.is(tok::r_brace))
3292     return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block;
3293 
3294   // Allow breaking after a trailing annotation, e.g. after a method
3295   // declaration.
3296   if (Left.is(TT_TrailingAnnotation))
3297     return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
3298                           tok::less, tok::coloncolon);
3299 
3300   if (Right.is(tok::kw___attribute) ||
3301       (Right.is(tok::l_square) && Right.is(TT_AttributeSquare)))
3302     return true;
3303 
3304   if (Left.is(tok::identifier) && Right.is(tok::string_literal))
3305     return true;
3306 
3307   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
3308     return true;
3309 
3310   if (Left.is(TT_CtorInitializerColon))
3311     return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon;
3312   if (Right.is(TT_CtorInitializerColon))
3313     return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon;
3314   if (Left.is(TT_CtorInitializerComma) &&
3315       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
3316     return false;
3317   if (Right.is(TT_CtorInitializerComma) &&
3318       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
3319     return true;
3320   if (Left.is(TT_InheritanceComma) &&
3321       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
3322     return false;
3323   if (Right.is(TT_InheritanceComma) &&
3324       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
3325     return true;
3326   if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
3327       (Left.is(tok::less) && Right.is(tok::less)))
3328     return false;
3329   if (Right.is(TT_BinaryOperator) &&
3330       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
3331       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
3332        Right.getPrecedence() != prec::Assignment))
3333     return true;
3334   if (Left.is(TT_ArrayInitializerLSquare))
3335     return true;
3336   if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
3337     return true;
3338   if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
3339       !Left.isOneOf(tok::arrowstar, tok::lessless) &&
3340       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
3341       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
3342        Left.getPrecedence() == prec::Assignment))
3343     return true;
3344   if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) ||
3345       (Left.is(tok::r_square) && Right.is(TT_AttributeSquare)))
3346     return false;
3347   return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
3348                       tok::kw_class, tok::kw_struct, tok::comment) ||
3349          Right.isMemberAccess() ||
3350          Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
3351                        tok::colon, tok::l_square, tok::at) ||
3352          (Left.is(tok::r_paren) &&
3353           Right.isOneOf(tok::identifier, tok::kw_const)) ||
3354          (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
3355          (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
3356 }
3357 
3358 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
3359   llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n";
3360   const FormatToken *Tok = Line.First;
3361   while (Tok) {
3362     llvm::errs() << " M=" << Tok->MustBreakBefore
3363                  << " C=" << Tok->CanBreakBefore
3364                  << " T=" << getTokenTypeName(Tok->Type)
3365                  << " S=" << Tok->SpacesRequiredBefore
3366                  << " B=" << Tok->BlockParameterCount
3367                  << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty
3368                  << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength
3369                  << " PPK=" << Tok->PackingKind << " FakeLParens=";
3370     for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
3371       llvm::errs() << Tok->FakeLParens[i] << "/";
3372     llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
3373     llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo();
3374     llvm::errs() << " Text='" << Tok->TokenText << "'\n";
3375     if (!Tok->Next)
3376       assert(Tok == Line.Last);
3377     Tok = Tok->Next;
3378   }
3379   llvm::errs() << "----\n";
3380 }
3381 
3382 } // namespace format
3383 } // namespace clang
3384