1 //===--- TokenAnnotator.cpp - Format C++ code -----------------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 ///
9 /// \file
10 /// This file implements a token annotator, i.e. creates
11 /// \c AnnotatedTokens out of \c FormatTokens with required extra information.
12 ///
13 //===----------------------------------------------------------------------===//
14 
15 #include "TokenAnnotator.h"
16 #include "FormatToken.h"
17 #include "clang/Basic/SourceManager.h"
18 #include "llvm/ADT/SmallPtrSet.h"
19 #include "llvm/Support/Debug.h"
20 
21 #define DEBUG_TYPE "format-token-annotator"
22 
23 namespace clang {
24 namespace format {
25 
26 namespace {
27 
28 /// Returns \c true if the token can be used as an identifier in
29 /// an Objective-C \c @selector, \c false otherwise.
30 ///
31 /// Because getFormattingLangOpts() always lexes source code as
32 /// Objective-C++, C++ keywords like \c new and \c delete are
33 /// lexed as tok::kw_*, not tok::identifier, even for Objective-C.
34 ///
35 /// For Objective-C and Objective-C++, both identifiers and keywords
36 /// are valid inside @selector(...) (or a macro which
37 /// invokes @selector(...)). So, we allow treat any identifier or
38 /// keyword as a potential Objective-C selector component.
39 static bool canBeObjCSelectorComponent(const FormatToken &Tok) {
40   return Tok.Tok.getIdentifierInfo() != nullptr;
41 }
42 
43 /// A parser that gathers additional information about tokens.
44 ///
45 /// The \c TokenAnnotator tries to match parenthesis and square brakets and
46 /// store a parenthesis levels. It also tries to resolve matching "<" and ">"
47 /// into template parameter lists.
48 class AnnotatingParser {
49 public:
50   AnnotatingParser(const FormatStyle &Style, AnnotatedLine &Line,
51                    const AdditionalKeywords &Keywords)
52       : Style(Style), Line(Line), CurrentToken(Line.First), AutoFound(false),
53         Keywords(Keywords) {
54     Contexts.push_back(Context(tok::unknown, 1, /*IsExpression=*/false));
55     resetTokenMetadata(CurrentToken);
56   }
57 
58 private:
59   bool parseAngle() {
60     if (!CurrentToken || !CurrentToken->Previous)
61       return false;
62     if (NonTemplateLess.count(CurrentToken->Previous))
63       return false;
64 
65     const FormatToken &Previous = *CurrentToken->Previous; // The '<'.
66     if (Previous.Previous) {
67       if (Previous.Previous->Tok.isLiteral())
68         return false;
69       if (Previous.Previous->is(tok::r_paren) && Contexts.size() > 1 &&
70           (!Previous.Previous->MatchingParen ||
71            !Previous.Previous->MatchingParen->is(TT_OverloadedOperatorLParen)))
72         return false;
73     }
74 
75     FormatToken *Left = CurrentToken->Previous;
76     Left->ParentBracket = Contexts.back().ContextKind;
77     ScopedContextCreator ContextCreator(*this, tok::less, 12);
78 
79     // If this angle is in the context of an expression, we need to be more
80     // hesitant to detect it as opening template parameters.
81     bool InExprContext = Contexts.back().IsExpression;
82 
83     Contexts.back().IsExpression = false;
84     // If there's a template keyword before the opening angle bracket, this is a
85     // template parameter, not an argument.
86     Contexts.back().InTemplateArgument =
87         Left->Previous && Left->Previous->Tok.isNot(tok::kw_template);
88 
89     if (Style.Language == FormatStyle::LK_Java &&
90         CurrentToken->is(tok::question))
91       next();
92 
93     while (CurrentToken) {
94       if (CurrentToken->is(tok::greater)) {
95         Left->MatchingParen = CurrentToken;
96         CurrentToken->MatchingParen = Left;
97         // In TT_Proto, we must distignuish between:
98         //   map<key, value>
99         //   msg < item: data >
100         //   msg: < item: data >
101         // In TT_TextProto, map<key, value> does not occur.
102         if (Style.Language == FormatStyle::LK_TextProto ||
103             (Style.Language == FormatStyle::LK_Proto && Left->Previous &&
104              Left->Previous->isOneOf(TT_SelectorName, TT_DictLiteral)))
105           CurrentToken->Type = TT_DictLiteral;
106         else
107           CurrentToken->Type = TT_TemplateCloser;
108         next();
109         return true;
110       }
111       if (CurrentToken->is(tok::question) &&
112           Style.Language == FormatStyle::LK_Java) {
113         next();
114         continue;
115       }
116       if (CurrentToken->isOneOf(tok::r_paren, tok::r_square, tok::r_brace) ||
117           (CurrentToken->isOneOf(tok::colon, tok::question) && InExprContext &&
118            Style.Language != FormatStyle::LK_Proto &&
119            Style.Language != FormatStyle::LK_TextProto))
120         return false;
121       // If a && or || is found and interpreted as a binary operator, this set
122       // of angles is likely part of something like "a < b && c > d". If the
123       // angles are inside an expression, the ||/&& might also be a binary
124       // operator that was misinterpreted because we are parsing template
125       // parameters.
126       // FIXME: This is getting out of hand, write a decent parser.
127       if (CurrentToken->Previous->isOneOf(tok::pipepipe, tok::ampamp) &&
128           CurrentToken->Previous->is(TT_BinaryOperator) &&
129           Contexts[Contexts.size() - 2].IsExpression &&
130           !Line.startsWith(tok::kw_template))
131         return false;
132       updateParameterCount(Left, CurrentToken);
133       if (Style.Language == FormatStyle::LK_Proto) {
134         if (FormatToken *Previous = CurrentToken->getPreviousNonComment()) {
135           if (CurrentToken->is(tok::colon) ||
136               (CurrentToken->isOneOf(tok::l_brace, tok::less) &&
137                Previous->isNot(tok::colon)))
138             Previous->Type = TT_SelectorName;
139         }
140       }
141       if (!consumeToken())
142         return false;
143     }
144     return false;
145   }
146 
147   bool parseParens(bool LookForDecls = false) {
148     if (!CurrentToken)
149       return false;
150     FormatToken *Left = CurrentToken->Previous;
151     Left->ParentBracket = Contexts.back().ContextKind;
152     ScopedContextCreator ContextCreator(*this, tok::l_paren, 1);
153 
154     // FIXME: This is a bit of a hack. Do better.
155     Contexts.back().ColonIsForRangeExpr =
156         Contexts.size() == 2 && Contexts[0].ColonIsForRangeExpr;
157 
158     bool StartsObjCMethodExpr = false;
159     if (FormatToken *MaybeSel = Left->Previous) {
160       // @selector( starts a selector.
161       if (MaybeSel->isObjCAtKeyword(tok::objc_selector) && MaybeSel->Previous &&
162           MaybeSel->Previous->is(tok::at)) {
163         StartsObjCMethodExpr = true;
164       }
165     }
166 
167     if (Left->is(TT_OverloadedOperatorLParen)) {
168       Contexts.back().IsExpression = false;
169     } else if (Style.Language == FormatStyle::LK_JavaScript &&
170                (Line.startsWith(Keywords.kw_type, tok::identifier) ||
171                 Line.startsWith(tok::kw_export, Keywords.kw_type,
172                                 tok::identifier))) {
173       // type X = (...);
174       // export type X = (...);
175       Contexts.back().IsExpression = false;
176     } else if (Left->Previous &&
177                (Left->Previous->isOneOf(tok::kw_static_assert, tok::kw_decltype,
178                                         tok::kw_if, tok::kw_while, tok::l_paren,
179                                         tok::comma) ||
180                 Left->Previous->endsSequence(tok::kw_constexpr, tok::kw_if) ||
181                 Left->Previous->is(TT_BinaryOperator))) {
182       // static_assert, if and while usually contain expressions.
183       Contexts.back().IsExpression = true;
184     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
185                (Left->Previous->is(Keywords.kw_function) ||
186                 (Left->Previous->endsSequence(tok::identifier,
187                                               Keywords.kw_function)))) {
188       // function(...) or function f(...)
189       Contexts.back().IsExpression = false;
190     } else if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
191                Left->Previous->is(TT_JsTypeColon)) {
192       // let x: (SomeType);
193       Contexts.back().IsExpression = false;
194     } else if (Left->Previous && Left->Previous->is(tok::r_square) &&
195                Left->Previous->MatchingParen &&
196                Left->Previous->MatchingParen->is(TT_LambdaLSquare)) {
197       // This is a parameter list of a lambda expression.
198       Contexts.back().IsExpression = false;
199     } else if (Line.InPPDirective &&
200                (!Left->Previous || !Left->Previous->is(tok::identifier))) {
201       Contexts.back().IsExpression = true;
202     } else if (Contexts[Contexts.size() - 2].CaretFound) {
203       // This is the parameter list of an ObjC block.
204       Contexts.back().IsExpression = false;
205     } else if (Left->Previous && Left->Previous->is(tok::kw___attribute)) {
206       Left->Type = TT_AttributeParen;
207     } else if (Left->Previous && Left->Previous->is(TT_ForEachMacro)) {
208       // The first argument to a foreach macro is a declaration.
209       Contexts.back().IsForEachMacro = true;
210       Contexts.back().IsExpression = false;
211     } else if (Left->Previous && Left->Previous->MatchingParen &&
212                Left->Previous->MatchingParen->is(TT_ObjCBlockLParen)) {
213       Contexts.back().IsExpression = false;
214     } else if (!Line.MustBeDeclaration && !Line.InPPDirective) {
215       bool IsForOrCatch =
216           Left->Previous && Left->Previous->isOneOf(tok::kw_for, tok::kw_catch);
217       Contexts.back().IsExpression = !IsForOrCatch;
218     }
219 
220     if (StartsObjCMethodExpr) {
221       Contexts.back().ColonIsObjCMethodExpr = true;
222       Left->Type = TT_ObjCMethodExpr;
223     }
224 
225     // MightBeFunctionType and ProbablyFunctionType are used for
226     // function pointer and reference types as well as Objective-C
227     // block types:
228     //
229     // void (*FunctionPointer)(void);
230     // void (&FunctionReference)(void);
231     // void (^ObjCBlock)(void);
232     bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
233     bool ProbablyFunctionType =
234         CurrentToken->isOneOf(tok::star, tok::amp, tok::caret);
235     bool HasMultipleLines = false;
236     bool HasMultipleParametersOnALine = false;
237     bool MightBeObjCForRangeLoop =
238         Left->Previous && Left->Previous->is(tok::kw_for);
239     FormatToken *PossibleObjCForInToken = nullptr;
240     while (CurrentToken) {
241       // LookForDecls is set when "if (" has been seen. Check for
242       // 'identifier' '*' 'identifier' followed by not '=' -- this
243       // '*' has to be a binary operator but determineStarAmpUsage() will
244       // categorize it as an unary operator, so set the right type here.
245       if (LookForDecls && CurrentToken->Next) {
246         FormatToken *Prev = CurrentToken->getPreviousNonComment();
247         if (Prev) {
248           FormatToken *PrevPrev = Prev->getPreviousNonComment();
249           FormatToken *Next = CurrentToken->Next;
250           if (PrevPrev && PrevPrev->is(tok::identifier) &&
251               Prev->isOneOf(tok::star, tok::amp, tok::ampamp) &&
252               CurrentToken->is(tok::identifier) && Next->isNot(tok::equal)) {
253             Prev->Type = TT_BinaryOperator;
254             LookForDecls = false;
255           }
256         }
257       }
258 
259       if (CurrentToken->Previous->is(TT_PointerOrReference) &&
260           CurrentToken->Previous->Previous->isOneOf(tok::l_paren,
261                                                     tok::coloncolon))
262         ProbablyFunctionType = true;
263       if (CurrentToken->is(tok::comma))
264         MightBeFunctionType = false;
265       if (CurrentToken->Previous->is(TT_BinaryOperator))
266         Contexts.back().IsExpression = true;
267       if (CurrentToken->is(tok::r_paren)) {
268         if (MightBeFunctionType && ProbablyFunctionType && CurrentToken->Next &&
269             (CurrentToken->Next->is(tok::l_paren) ||
270              (CurrentToken->Next->is(tok::l_square) && Line.MustBeDeclaration)))
271           Left->Type = Left->Next->is(tok::caret) ? TT_ObjCBlockLParen
272                                                   : TT_FunctionTypeLParen;
273         Left->MatchingParen = CurrentToken;
274         CurrentToken->MatchingParen = Left;
275 
276         if (CurrentToken->Next && CurrentToken->Next->is(tok::l_brace) &&
277             Left->Previous && Left->Previous->is(tok::l_paren)) {
278           // Detect the case where macros are used to generate lambdas or
279           // function bodies, e.g.:
280           //   auto my_lambda = MARCO((Type *type, int i) { .. body .. });
281           for (FormatToken *Tok = Left; Tok != CurrentToken; Tok = Tok->Next) {
282             if (Tok->is(TT_BinaryOperator) &&
283                 Tok->isOneOf(tok::star, tok::amp, tok::ampamp))
284               Tok->Type = TT_PointerOrReference;
285           }
286         }
287 
288         if (StartsObjCMethodExpr) {
289           CurrentToken->Type = TT_ObjCMethodExpr;
290           if (Contexts.back().FirstObjCSelectorName) {
291             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
292                 Contexts.back().LongestObjCSelectorName;
293           }
294         }
295 
296         if (Left->is(TT_AttributeParen))
297           CurrentToken->Type = TT_AttributeParen;
298         if (Left->Previous && Left->Previous->is(TT_JavaAnnotation))
299           CurrentToken->Type = TT_JavaAnnotation;
300         if (Left->Previous && Left->Previous->is(TT_LeadingJavaAnnotation))
301           CurrentToken->Type = TT_LeadingJavaAnnotation;
302         if (Left->Previous && Left->Previous->is(TT_AttributeSquare))
303           CurrentToken->Type = TT_AttributeSquare;
304 
305         if (!HasMultipleLines)
306           Left->PackingKind = PPK_Inconclusive;
307         else if (HasMultipleParametersOnALine)
308           Left->PackingKind = PPK_BinPacked;
309         else
310           Left->PackingKind = PPK_OnePerLine;
311 
312         next();
313         return true;
314       }
315       if (CurrentToken->isOneOf(tok::r_square, tok::r_brace))
316         return false;
317 
318       if (CurrentToken->is(tok::l_brace))
319         Left->Type = TT_Unknown; // Not TT_ObjCBlockLParen
320       if (CurrentToken->is(tok::comma) && CurrentToken->Next &&
321           !CurrentToken->Next->HasUnescapedNewline &&
322           !CurrentToken->Next->isTrailingComment())
323         HasMultipleParametersOnALine = true;
324       if ((CurrentToken->Previous->isOneOf(tok::kw_const, tok::kw_auto) ||
325            CurrentToken->Previous->isSimpleTypeSpecifier()) &&
326           !CurrentToken->is(tok::l_brace))
327         Contexts.back().IsExpression = false;
328       if (CurrentToken->isOneOf(tok::semi, tok::colon)) {
329         MightBeObjCForRangeLoop = false;
330         if (PossibleObjCForInToken) {
331           PossibleObjCForInToken->Type = TT_Unknown;
332           PossibleObjCForInToken = nullptr;
333         }
334       }
335       if (MightBeObjCForRangeLoop && CurrentToken->is(Keywords.kw_in)) {
336         PossibleObjCForInToken = CurrentToken;
337         PossibleObjCForInToken->Type = TT_ObjCForIn;
338       }
339       // When we discover a 'new', we set CanBeExpression to 'false' in order to
340       // parse the type correctly. Reset that after a comma.
341       if (CurrentToken->is(tok::comma))
342         Contexts.back().CanBeExpression = true;
343 
344       FormatToken *Tok = CurrentToken;
345       if (!consumeToken())
346         return false;
347       updateParameterCount(Left, Tok);
348       if (CurrentToken && CurrentToken->HasUnescapedNewline)
349         HasMultipleLines = true;
350     }
351     return false;
352   }
353 
354   bool isCSharpAttributeSpecifier(const FormatToken &Tok) {
355     if (!Style.isCSharp())
356       return false;
357 
358     const FormatToken *AttrTok = Tok.Next;
359     if (!AttrTok)
360       return false;
361 
362     // Just an empty declaration e.g. string [].
363     if (AttrTok->is(tok::r_square))
364       return false;
365 
366     // Move along the tokens inbetween the '[' and ']' e.g. [STAThread].
367     while (AttrTok && AttrTok->isNot(tok::r_square)) {
368       AttrTok = AttrTok->Next;
369     }
370 
371     if (!AttrTok)
372       return false;
373 
374     // Move past the end of ']'.
375     AttrTok = AttrTok->Next;
376     if (!AttrTok)
377       return false;
378 
379     // Limit this to being an access modifier that follows.
380     if (AttrTok->isOneOf(tok::kw_public, tok::kw_private, tok::kw_protected,
381                          tok::kw_class, tok::kw_static, tok::l_square,
382                          Keywords.kw_internal)) {
383       return true;
384     }
385     return false;
386   }
387 
388   bool isCpp11AttributeSpecifier(const FormatToken &Tok) {
389     if (!Style.isCpp() || !Tok.startsSequence(tok::l_square, tok::l_square))
390       return false;
391     const FormatToken *AttrTok = Tok.Next->Next;
392     if (!AttrTok)
393       return false;
394     // C++17 '[[using ns: foo, bar(baz, blech)]]'
395     // We assume nobody will name an ObjC variable 'using'.
396     if (AttrTok->startsSequence(tok::kw_using, tok::identifier, tok::colon))
397       return true;
398     if (AttrTok->isNot(tok::identifier))
399       return false;
400     while (AttrTok && !AttrTok->startsSequence(tok::r_square, tok::r_square)) {
401       // ObjC message send. We assume nobody will use : in a C++11 attribute
402       // specifier parameter, although this is technically valid:
403       // [[foo(:)]]
404       if (AttrTok->is(tok::colon) ||
405           AttrTok->startsSequence(tok::identifier, tok::identifier) ||
406           AttrTok->startsSequence(tok::r_paren, tok::identifier))
407         return false;
408       if (AttrTok->is(tok::ellipsis))
409         return true;
410       AttrTok = AttrTok->Next;
411     }
412     return AttrTok && AttrTok->startsSequence(tok::r_square, tok::r_square);
413   }
414 
415   bool parseSquare() {
416     if (!CurrentToken)
417       return false;
418 
419     // A '[' could be an index subscript (after an identifier or after
420     // ')' or ']'), it could be the start of an Objective-C method
421     // expression, it could the start of an Objective-C array literal,
422     // or it could be a C++ attribute specifier [[foo::bar]].
423     FormatToken *Left = CurrentToken->Previous;
424     Left->ParentBracket = Contexts.back().ContextKind;
425     FormatToken *Parent = Left->getPreviousNonComment();
426 
427     // Cases where '>' is followed by '['.
428     // In C++, this can happen either in array of templates (foo<int>[10])
429     // or when array is a nested template type (unique_ptr<type1<type2>[]>).
430     bool CppArrayTemplates =
431         Style.isCpp() && Parent && Parent->is(TT_TemplateCloser) &&
432         (Contexts.back().CanBeExpression || Contexts.back().IsExpression ||
433          Contexts.back().InTemplateArgument);
434 
435     bool IsCpp11AttributeSpecifier = isCpp11AttributeSpecifier(*Left) ||
436                                      Contexts.back().InCpp11AttributeSpecifier;
437 
438     // Treat C# Attributes [STAThread] much like C++ attributes [[...]].
439     bool IsCSharp11AttributeSpecifier =
440         isCSharpAttributeSpecifier(*Left) ||
441         Contexts.back().InCSharpAttributeSpecifier;
442 
443     bool InsideInlineASM = Line.startsWith(tok::kw_asm);
444     bool IsCppStructuredBinding = Left->isCppStructuredBinding(Style);
445     bool StartsObjCMethodExpr =
446         !IsCppStructuredBinding && !InsideInlineASM && !CppArrayTemplates &&
447         Style.isCpp() && !IsCpp11AttributeSpecifier &&
448         Contexts.back().CanBeExpression && Left->isNot(TT_LambdaLSquare) &&
449         !CurrentToken->isOneOf(tok::l_brace, tok::r_square) &&
450         (!Parent ||
451          Parent->isOneOf(tok::colon, tok::l_square, tok::l_paren,
452                          tok::kw_return, tok::kw_throw) ||
453          Parent->isUnaryOperator() ||
454          // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
455          Parent->isOneOf(TT_ObjCForIn, TT_CastRParen) ||
456          (getBinOpPrecedence(Parent->Tok.getKind(), true, true) >
457           prec::Unknown));
458     bool ColonFound = false;
459 
460     unsigned BindingIncrease = 1;
461     if (IsCppStructuredBinding) {
462       Left->Type = TT_StructuredBindingLSquare;
463     } else if (Left->is(TT_Unknown)) {
464       if (StartsObjCMethodExpr) {
465         Left->Type = TT_ObjCMethodExpr;
466       } else if (IsCpp11AttributeSpecifier) {
467         Left->Type = TT_AttributeSquare;
468       } else if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
469                  Contexts.back().ContextKind == tok::l_brace &&
470                  Parent->isOneOf(tok::l_brace, tok::comma)) {
471         Left->Type = TT_JsComputedPropertyName;
472       } else if (Style.isCpp() && Contexts.back().ContextKind == tok::l_brace &&
473                  Parent && Parent->isOneOf(tok::l_brace, tok::comma)) {
474         Left->Type = TT_DesignatedInitializerLSquare;
475       } else if (CurrentToken->is(tok::r_square) && Parent &&
476                  Parent->is(TT_TemplateCloser)) {
477         Left->Type = TT_ArraySubscriptLSquare;
478       } else if (Style.Language == FormatStyle::LK_Proto ||
479                  Style.Language == FormatStyle::LK_TextProto) {
480         // Square braces in LK_Proto can either be message field attributes:
481         //
482         // optional Aaa aaa = 1 [
483         //   (aaa) = aaa
484         // ];
485         //
486         // extensions 123 [
487         //   (aaa) = aaa
488         // ];
489         //
490         // or text proto extensions (in options):
491         //
492         // option (Aaa.options) = {
493         //   [type.type/type] {
494         //     key: value
495         //   }
496         // }
497         //
498         // or repeated fields (in options):
499         //
500         // option (Aaa.options) = {
501         //   keys: [ 1, 2, 3 ]
502         // }
503         //
504         // In the first and the third case we want to spread the contents inside
505         // the square braces; in the second we want to keep them inline.
506         Left->Type = TT_ArrayInitializerLSquare;
507         if (!Left->endsSequence(tok::l_square, tok::numeric_constant,
508                                 tok::equal) &&
509             !Left->endsSequence(tok::l_square, tok::numeric_constant,
510                                 tok::identifier) &&
511             !Left->endsSequence(tok::l_square, tok::colon, TT_SelectorName)) {
512           Left->Type = TT_ProtoExtensionLSquare;
513           BindingIncrease = 10;
514         }
515       } else if (!CppArrayTemplates && Parent &&
516                  Parent->isOneOf(TT_BinaryOperator, TT_TemplateCloser, tok::at,
517                                  tok::comma, tok::l_paren, tok::l_square,
518                                  tok::question, tok::colon, tok::kw_return,
519                                  // Should only be relevant to JavaScript:
520                                  tok::kw_default)) {
521         Left->Type = TT_ArrayInitializerLSquare;
522       } else if (IsCSharp11AttributeSpecifier) {
523         Left->Type = TT_AttributeSquare;
524       } else {
525         BindingIncrease = 10;
526         Left->Type = TT_ArraySubscriptLSquare;
527       }
528     }
529 
530     ScopedContextCreator ContextCreator(*this, tok::l_square, BindingIncrease);
531     Contexts.back().IsExpression = true;
532     if (Style.Language == FormatStyle::LK_JavaScript && Parent &&
533         Parent->is(TT_JsTypeColon))
534       Contexts.back().IsExpression = false;
535 
536     Contexts.back().ColonIsObjCMethodExpr = StartsObjCMethodExpr;
537     Contexts.back().InCpp11AttributeSpecifier = IsCpp11AttributeSpecifier;
538     Contexts.back().InCSharpAttributeSpecifier = IsCSharp11AttributeSpecifier;
539 
540     while (CurrentToken) {
541       if (CurrentToken->is(tok::r_square)) {
542         if (IsCpp11AttributeSpecifier)
543           CurrentToken->Type = TT_AttributeSquare;
544         if (IsCSharp11AttributeSpecifier)
545           CurrentToken->Type = TT_AttributeSquare;
546         else if (((CurrentToken->Next &&
547                    CurrentToken->Next->is(tok::l_paren)) ||
548                   (CurrentToken->Previous &&
549                    CurrentToken->Previous->Previous == Left)) &&
550                  Left->is(TT_ObjCMethodExpr)) {
551           // An ObjC method call is rarely followed by an open parenthesis. It
552           // also can't be composed of just one token, unless it's a macro that
553           // will be expanded to more tokens.
554           // FIXME: Do we incorrectly label ":" with this?
555           StartsObjCMethodExpr = false;
556           Left->Type = TT_Unknown;
557         }
558         if (StartsObjCMethodExpr && CurrentToken->Previous != Left) {
559           CurrentToken->Type = TT_ObjCMethodExpr;
560           // If we haven't seen a colon yet, make sure the last identifier
561           // before the r_square is tagged as a selector name component.
562           if (!ColonFound && CurrentToken->Previous &&
563               CurrentToken->Previous->is(TT_Unknown) &&
564               canBeObjCSelectorComponent(*CurrentToken->Previous))
565             CurrentToken->Previous->Type = TT_SelectorName;
566           // determineStarAmpUsage() thinks that '*' '[' is allocating an
567           // array of pointers, but if '[' starts a selector then '*' is a
568           // binary operator.
569           if (Parent && Parent->is(TT_PointerOrReference))
570             Parent->Type = TT_BinaryOperator;
571         }
572         // An arrow after an ObjC method expression is not a lambda arrow.
573         if (CurrentToken->Type == TT_ObjCMethodExpr && CurrentToken->Next &&
574             CurrentToken->Next->is(TT_LambdaArrow))
575           CurrentToken->Next->Type = TT_Unknown;
576         Left->MatchingParen = CurrentToken;
577         CurrentToken->MatchingParen = Left;
578         // FirstObjCSelectorName is set when a colon is found. This does
579         // not work, however, when the method has no parameters.
580         // Here, we set FirstObjCSelectorName when the end of the method call is
581         // reached, in case it was not set already.
582         if (!Contexts.back().FirstObjCSelectorName) {
583           FormatToken *Previous = CurrentToken->getPreviousNonComment();
584           if (Previous && Previous->is(TT_SelectorName)) {
585             Previous->ObjCSelectorNameParts = 1;
586             Contexts.back().FirstObjCSelectorName = Previous;
587           }
588         } else {
589           Left->ParameterCount =
590               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
591         }
592         if (Contexts.back().FirstObjCSelectorName) {
593           Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
594               Contexts.back().LongestObjCSelectorName;
595           if (Left->BlockParameterCount > 1)
596             Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName = 0;
597         }
598         next();
599         return true;
600       }
601       if (CurrentToken->isOneOf(tok::r_paren, tok::r_brace))
602         return false;
603       if (CurrentToken->is(tok::colon)) {
604         if (IsCpp11AttributeSpecifier &&
605             CurrentToken->endsSequence(tok::colon, tok::identifier,
606                                        tok::kw_using)) {
607           // Remember that this is a [[using ns: foo]] C++ attribute, so we
608           // don't add a space before the colon (unlike other colons).
609           CurrentToken->Type = TT_AttributeColon;
610         } else if (Left->isOneOf(TT_ArraySubscriptLSquare,
611                                  TT_DesignatedInitializerLSquare)) {
612           Left->Type = TT_ObjCMethodExpr;
613           StartsObjCMethodExpr = true;
614           Contexts.back().ColonIsObjCMethodExpr = true;
615           if (Parent && Parent->is(tok::r_paren))
616             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
617             Parent->Type = TT_CastRParen;
618         }
619         ColonFound = true;
620       }
621       if (CurrentToken->is(tok::comma) && Left->is(TT_ObjCMethodExpr) &&
622           !ColonFound)
623         Left->Type = TT_ArrayInitializerLSquare;
624       FormatToken *Tok = CurrentToken;
625       if (!consumeToken())
626         return false;
627       updateParameterCount(Left, Tok);
628     }
629     return false;
630   }
631 
632   bool parseBrace() {
633     if (CurrentToken) {
634       FormatToken *Left = CurrentToken->Previous;
635       Left->ParentBracket = Contexts.back().ContextKind;
636 
637       if (Contexts.back().CaretFound)
638         Left->Type = TT_ObjCBlockLBrace;
639       Contexts.back().CaretFound = false;
640 
641       ScopedContextCreator ContextCreator(*this, tok::l_brace, 1);
642       Contexts.back().ColonIsDictLiteral = true;
643       if (Left->BlockKind == BK_BracedInit)
644         Contexts.back().IsExpression = true;
645       if (Style.Language == FormatStyle::LK_JavaScript && Left->Previous &&
646           Left->Previous->is(TT_JsTypeColon))
647         Contexts.back().IsExpression = false;
648 
649       while (CurrentToken) {
650         if (CurrentToken->is(tok::r_brace)) {
651           Left->MatchingParen = CurrentToken;
652           CurrentToken->MatchingParen = Left;
653           next();
654           return true;
655         }
656         if (CurrentToken->isOneOf(tok::r_paren, tok::r_square))
657           return false;
658         updateParameterCount(Left, CurrentToken);
659         if (CurrentToken->isOneOf(tok::colon, tok::l_brace, tok::less)) {
660           FormatToken *Previous = CurrentToken->getPreviousNonComment();
661           if (Previous->is(TT_JsTypeOptionalQuestion))
662             Previous = Previous->getPreviousNonComment();
663           if ((CurrentToken->is(tok::colon) &&
664                (!Contexts.back().ColonIsDictLiteral || !Style.isCpp())) ||
665               Style.Language == FormatStyle::LK_Proto ||
666               Style.Language == FormatStyle::LK_TextProto) {
667             Left->Type = TT_DictLiteral;
668             if (Previous->Tok.getIdentifierInfo() ||
669                 Previous->is(tok::string_literal))
670               Previous->Type = TT_SelectorName;
671           }
672           if (CurrentToken->is(tok::colon) ||
673               Style.Language == FormatStyle::LK_JavaScript)
674             Left->Type = TT_DictLiteral;
675         }
676         if (CurrentToken->is(tok::comma) &&
677             Style.Language == FormatStyle::LK_JavaScript)
678           Left->Type = TT_DictLiteral;
679         if (!consumeToken())
680           return false;
681       }
682     }
683     return true;
684   }
685 
686   void updateParameterCount(FormatToken *Left, FormatToken *Current) {
687     // For ObjC methods, the number of parameters is calculated differently as
688     // method declarations have a different structure (the parameters are not
689     // inside a bracket scope).
690     if (Current->is(tok::l_brace) && Current->BlockKind == BK_Block)
691       ++Left->BlockParameterCount;
692     if (Current->is(tok::comma)) {
693       ++Left->ParameterCount;
694       if (!Left->Role)
695         Left->Role.reset(new CommaSeparatedList(Style));
696       Left->Role->CommaFound(Current);
697     } else if (Left->ParameterCount == 0 && Current->isNot(tok::comment)) {
698       Left->ParameterCount = 1;
699     }
700   }
701 
702   bool parseConditional() {
703     while (CurrentToken) {
704       if (CurrentToken->is(tok::colon)) {
705         CurrentToken->Type = TT_ConditionalExpr;
706         next();
707         return true;
708       }
709       if (!consumeToken())
710         return false;
711     }
712     return false;
713   }
714 
715   bool parseTemplateDeclaration() {
716     if (CurrentToken && CurrentToken->is(tok::less)) {
717       CurrentToken->Type = TT_TemplateOpener;
718       next();
719       if (!parseAngle())
720         return false;
721       if (CurrentToken)
722         CurrentToken->Previous->ClosesTemplateDeclaration = true;
723       return true;
724     }
725     return false;
726   }
727 
728   bool consumeToken() {
729     FormatToken *Tok = CurrentToken;
730     next();
731     switch (Tok->Tok.getKind()) {
732     case tok::plus:
733     case tok::minus:
734       if (!Tok->Previous && Line.MustBeDeclaration)
735         Tok->Type = TT_ObjCMethodSpecifier;
736       break;
737     case tok::colon:
738       if (!Tok->Previous)
739         return false;
740       // Colons from ?: are handled in parseConditional().
741       if (Style.Language == FormatStyle::LK_JavaScript) {
742         if (Contexts.back().ColonIsForRangeExpr || // colon in for loop
743             (Contexts.size() == 1 &&               // switch/case labels
744              !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) ||
745             Contexts.back().ContextKind == tok::l_paren ||  // function params
746             Contexts.back().ContextKind == tok::l_square || // array type
747             (!Contexts.back().IsExpression &&
748              Contexts.back().ContextKind == tok::l_brace) || // object type
749             (Contexts.size() == 1 &&
750              Line.MustBeDeclaration)) { // method/property declaration
751           Contexts.back().IsExpression = false;
752           Tok->Type = TT_JsTypeColon;
753           break;
754         }
755       }
756       if (Contexts.back().ColonIsDictLiteral ||
757           Style.Language == FormatStyle::LK_Proto ||
758           Style.Language == FormatStyle::LK_TextProto) {
759         Tok->Type = TT_DictLiteral;
760         if (Style.Language == FormatStyle::LK_TextProto) {
761           if (FormatToken *Previous = Tok->getPreviousNonComment())
762             Previous->Type = TT_SelectorName;
763         }
764       } else if (Contexts.back().ColonIsObjCMethodExpr ||
765                  Line.startsWith(TT_ObjCMethodSpecifier)) {
766         Tok->Type = TT_ObjCMethodExpr;
767         const FormatToken *BeforePrevious = Tok->Previous->Previous;
768         // Ensure we tag all identifiers in method declarations as
769         // TT_SelectorName.
770         bool UnknownIdentifierInMethodDeclaration =
771             Line.startsWith(TT_ObjCMethodSpecifier) &&
772             Tok->Previous->is(tok::identifier) && Tok->Previous->is(TT_Unknown);
773         if (!BeforePrevious ||
774             // FIXME(bug 36976): ObjC return types shouldn't use TT_CastRParen.
775             !(BeforePrevious->is(TT_CastRParen) ||
776               (BeforePrevious->is(TT_ObjCMethodExpr) &&
777                BeforePrevious->is(tok::colon))) ||
778             BeforePrevious->is(tok::r_square) ||
779             Contexts.back().LongestObjCSelectorName == 0 ||
780             UnknownIdentifierInMethodDeclaration) {
781           Tok->Previous->Type = TT_SelectorName;
782           if (!Contexts.back().FirstObjCSelectorName)
783             Contexts.back().FirstObjCSelectorName = Tok->Previous;
784           else if (Tok->Previous->ColumnWidth >
785                    Contexts.back().LongestObjCSelectorName)
786             Contexts.back().LongestObjCSelectorName =
787                 Tok->Previous->ColumnWidth;
788           Tok->Previous->ParameterIndex =
789               Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
790           ++Contexts.back().FirstObjCSelectorName->ObjCSelectorNameParts;
791         }
792       } else if (Contexts.back().ColonIsForRangeExpr) {
793         Tok->Type = TT_RangeBasedForLoopColon;
794       } else if (CurrentToken && CurrentToken->is(tok::numeric_constant)) {
795         Tok->Type = TT_BitFieldColon;
796       } else if (Contexts.size() == 1 &&
797                  !Line.First->isOneOf(tok::kw_enum, tok::kw_case)) {
798         if (Tok->getPreviousNonComment()->isOneOf(tok::r_paren,
799                                                   tok::kw_noexcept))
800           Tok->Type = TT_CtorInitializerColon;
801         else
802           Tok->Type = TT_InheritanceColon;
803       } else if (canBeObjCSelectorComponent(*Tok->Previous) && Tok->Next &&
804                  (Tok->Next->isOneOf(tok::r_paren, tok::comma) ||
805                   (canBeObjCSelectorComponent(*Tok->Next) && Tok->Next->Next &&
806                    Tok->Next->Next->is(tok::colon)))) {
807         // This handles a special macro in ObjC code where selectors including
808         // the colon are passed as macro arguments.
809         Tok->Type = TT_ObjCMethodExpr;
810       } else if (Contexts.back().ContextKind == tok::l_paren) {
811         Tok->Type = TT_InlineASMColon;
812       }
813       break;
814     case tok::pipe:
815     case tok::amp:
816       // | and & in declarations/type expressions represent union and
817       // intersection types, respectively.
818       if (Style.Language == FormatStyle::LK_JavaScript &&
819           !Contexts.back().IsExpression)
820         Tok->Type = TT_JsTypeOperator;
821       break;
822     case tok::kw_if:
823     case tok::kw_while:
824       if (Tok->is(tok::kw_if) && CurrentToken &&
825           CurrentToken->is(tok::kw_constexpr))
826         next();
827       if (CurrentToken && CurrentToken->is(tok::l_paren)) {
828         next();
829         if (!parseParens(/*LookForDecls=*/true))
830           return false;
831       }
832       break;
833     case tok::kw_for:
834       if (Style.Language == FormatStyle::LK_JavaScript) {
835         // x.for and {for: ...}
836         if ((Tok->Previous && Tok->Previous->is(tok::period)) ||
837             (Tok->Next && Tok->Next->is(tok::colon)))
838           break;
839         // JS' for await ( ...
840         if (CurrentToken && CurrentToken->is(Keywords.kw_await))
841           next();
842       }
843       Contexts.back().ColonIsForRangeExpr = true;
844       next();
845       if (!parseParens())
846         return false;
847       break;
848     case tok::l_paren:
849       // When faced with 'operator()()', the kw_operator handler incorrectly
850       // marks the first l_paren as a OverloadedOperatorLParen. Here, we make
851       // the first two parens OverloadedOperators and the second l_paren an
852       // OverloadedOperatorLParen.
853       if (Tok->Previous && Tok->Previous->is(tok::r_paren) &&
854           Tok->Previous->MatchingParen &&
855           Tok->Previous->MatchingParen->is(TT_OverloadedOperatorLParen)) {
856         Tok->Previous->Type = TT_OverloadedOperator;
857         Tok->Previous->MatchingParen->Type = TT_OverloadedOperator;
858         Tok->Type = TT_OverloadedOperatorLParen;
859       }
860 
861       if (!parseParens())
862         return false;
863       if (Line.MustBeDeclaration && Contexts.size() == 1 &&
864           !Contexts.back().IsExpression && !Line.startsWith(TT_ObjCProperty) &&
865           (!Tok->Previous ||
866            !Tok->Previous->isOneOf(tok::kw_decltype, tok::kw___attribute,
867                                    TT_LeadingJavaAnnotation)))
868         Line.MightBeFunctionDecl = true;
869       break;
870     case tok::l_square:
871       if (!parseSquare())
872         return false;
873       break;
874     case tok::l_brace:
875       if (Style.Language == FormatStyle::LK_TextProto) {
876         FormatToken *Previous = Tok->getPreviousNonComment();
877         if (Previous && Previous->Type != TT_DictLiteral)
878           Previous->Type = TT_SelectorName;
879       }
880       if (!parseBrace())
881         return false;
882       break;
883     case tok::less:
884       if (parseAngle()) {
885         Tok->Type = TT_TemplateOpener;
886         // In TT_Proto, we must distignuish between:
887         //   map<key, value>
888         //   msg < item: data >
889         //   msg: < item: data >
890         // In TT_TextProto, map<key, value> does not occur.
891         if (Style.Language == FormatStyle::LK_TextProto ||
892             (Style.Language == FormatStyle::LK_Proto && Tok->Previous &&
893              Tok->Previous->isOneOf(TT_SelectorName, TT_DictLiteral))) {
894           Tok->Type = TT_DictLiteral;
895           FormatToken *Previous = Tok->getPreviousNonComment();
896           if (Previous && Previous->Type != TT_DictLiteral)
897             Previous->Type = TT_SelectorName;
898         }
899       } else {
900         Tok->Type = TT_BinaryOperator;
901         NonTemplateLess.insert(Tok);
902         CurrentToken = Tok;
903         next();
904       }
905       break;
906     case tok::r_paren:
907     case tok::r_square:
908       return false;
909     case tok::r_brace:
910       // Lines can start with '}'.
911       if (Tok->Previous)
912         return false;
913       break;
914     case tok::greater:
915       if (Style.Language != FormatStyle::LK_TextProto)
916         Tok->Type = TT_BinaryOperator;
917       break;
918     case tok::kw_operator:
919       if (Style.Language == FormatStyle::LK_TextProto ||
920           Style.Language == FormatStyle::LK_Proto)
921         break;
922       while (CurrentToken &&
923              !CurrentToken->isOneOf(tok::l_paren, tok::semi, tok::r_paren)) {
924         if (CurrentToken->isOneOf(tok::star, tok::amp))
925           CurrentToken->Type = TT_PointerOrReference;
926         consumeToken();
927         if (CurrentToken &&
928             CurrentToken->Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator,
929                                             tok::comma))
930           CurrentToken->Previous->Type = TT_OverloadedOperator;
931       }
932       if (CurrentToken) {
933         CurrentToken->Type = TT_OverloadedOperatorLParen;
934         if (CurrentToken->Previous->is(TT_BinaryOperator))
935           CurrentToken->Previous->Type = TT_OverloadedOperator;
936       }
937       break;
938     case tok::question:
939       if (Style.Language == FormatStyle::LK_JavaScript && Tok->Next &&
940           Tok->Next->isOneOf(tok::semi, tok::comma, tok::colon, tok::r_paren,
941                              tok::r_brace)) {
942         // Question marks before semicolons, colons, etc. indicate optional
943         // types (fields, parameters), e.g.
944         //   function(x?: string, y?) {...}
945         //   class X { y?; }
946         Tok->Type = TT_JsTypeOptionalQuestion;
947         break;
948       }
949       // Declarations cannot be conditional expressions, this can only be part
950       // of a type declaration.
951       if (Line.MustBeDeclaration && !Contexts.back().IsExpression &&
952           Style.Language == FormatStyle::LK_JavaScript)
953         break;
954       parseConditional();
955       break;
956     case tok::kw_template:
957       parseTemplateDeclaration();
958       break;
959     case tok::comma:
960       if (Contexts.back().InCtorInitializer)
961         Tok->Type = TT_CtorInitializerComma;
962       else if (Contexts.back().InInheritanceList)
963         Tok->Type = TT_InheritanceComma;
964       else if (Contexts.back().FirstStartOfName &&
965                (Contexts.size() == 1 || Line.startsWith(tok::kw_for))) {
966         Contexts.back().FirstStartOfName->PartOfMultiVariableDeclStmt = true;
967         Line.IsMultiVariableDeclStmt = true;
968       }
969       if (Contexts.back().IsForEachMacro)
970         Contexts.back().IsExpression = true;
971       break;
972     case tok::identifier:
973       if (Tok->isOneOf(Keywords.kw___has_include,
974                        Keywords.kw___has_include_next)) {
975         parseHasInclude();
976       }
977       break;
978     default:
979       break;
980     }
981     return true;
982   }
983 
984   void parseIncludeDirective() {
985     if (CurrentToken && CurrentToken->is(tok::less)) {
986       next();
987       while (CurrentToken) {
988         // Mark tokens up to the trailing line comments as implicit string
989         // literals.
990         if (CurrentToken->isNot(tok::comment) &&
991             !CurrentToken->TokenText.startswith("//"))
992           CurrentToken->Type = TT_ImplicitStringLiteral;
993         next();
994       }
995     }
996   }
997 
998   void parseWarningOrError() {
999     next();
1000     // We still want to format the whitespace left of the first token of the
1001     // warning or error.
1002     next();
1003     while (CurrentToken) {
1004       CurrentToken->Type = TT_ImplicitStringLiteral;
1005       next();
1006     }
1007   }
1008 
1009   void parsePragma() {
1010     next(); // Consume "pragma".
1011     if (CurrentToken &&
1012         CurrentToken->isOneOf(Keywords.kw_mark, Keywords.kw_option)) {
1013       bool IsMark = CurrentToken->is(Keywords.kw_mark);
1014       next(); // Consume "mark".
1015       next(); // Consume first token (so we fix leading whitespace).
1016       while (CurrentToken) {
1017         if (IsMark || CurrentToken->Previous->is(TT_BinaryOperator))
1018           CurrentToken->Type = TT_ImplicitStringLiteral;
1019         next();
1020       }
1021     }
1022   }
1023 
1024   void parseHasInclude() {
1025     if (!CurrentToken || !CurrentToken->is(tok::l_paren))
1026       return;
1027     next(); // '('
1028     parseIncludeDirective();
1029     next(); // ')'
1030   }
1031 
1032   LineType parsePreprocessorDirective() {
1033     bool IsFirstToken = CurrentToken->IsFirst;
1034     LineType Type = LT_PreprocessorDirective;
1035     next();
1036     if (!CurrentToken)
1037       return Type;
1038 
1039     if (Style.Language == FormatStyle::LK_JavaScript && IsFirstToken) {
1040       // JavaScript files can contain shebang lines of the form:
1041       // #!/usr/bin/env node
1042       // Treat these like C++ #include directives.
1043       while (CurrentToken) {
1044         // Tokens cannot be comments here.
1045         CurrentToken->Type = TT_ImplicitStringLiteral;
1046         next();
1047       }
1048       return LT_ImportStatement;
1049     }
1050 
1051     if (CurrentToken->Tok.is(tok::numeric_constant)) {
1052       CurrentToken->SpacesRequiredBefore = 1;
1053       return Type;
1054     }
1055     // Hashes in the middle of a line can lead to any strange token
1056     // sequence.
1057     if (!CurrentToken->Tok.getIdentifierInfo())
1058       return Type;
1059     switch (CurrentToken->Tok.getIdentifierInfo()->getPPKeywordID()) {
1060     case tok::pp_include:
1061     case tok::pp_include_next:
1062     case tok::pp_import:
1063       next();
1064       parseIncludeDirective();
1065       Type = LT_ImportStatement;
1066       break;
1067     case tok::pp_error:
1068     case tok::pp_warning:
1069       parseWarningOrError();
1070       break;
1071     case tok::pp_pragma:
1072       parsePragma();
1073       break;
1074     case tok::pp_if:
1075     case tok::pp_elif:
1076       Contexts.back().IsExpression = true;
1077       parseLine();
1078       break;
1079     default:
1080       break;
1081     }
1082     while (CurrentToken) {
1083       FormatToken *Tok = CurrentToken;
1084       next();
1085       if (Tok->is(tok::l_paren))
1086         parseParens();
1087       else if (Tok->isOneOf(Keywords.kw___has_include,
1088                             Keywords.kw___has_include_next))
1089         parseHasInclude();
1090     }
1091     return Type;
1092   }
1093 
1094 public:
1095   LineType parseLine() {
1096     NonTemplateLess.clear();
1097     if (CurrentToken->is(tok::hash))
1098       return parsePreprocessorDirective();
1099 
1100     // Directly allow to 'import <string-literal>' to support protocol buffer
1101     // definitions (github.com/google/protobuf) or missing "#" (either way we
1102     // should not break the line).
1103     IdentifierInfo *Info = CurrentToken->Tok.getIdentifierInfo();
1104     if ((Style.Language == FormatStyle::LK_Java &&
1105          CurrentToken->is(Keywords.kw_package)) ||
1106         (Info && Info->getPPKeywordID() == tok::pp_import &&
1107          CurrentToken->Next &&
1108          CurrentToken->Next->isOneOf(tok::string_literal, tok::identifier,
1109                                      tok::kw_static))) {
1110       next();
1111       parseIncludeDirective();
1112       return LT_ImportStatement;
1113     }
1114 
1115     // If this line starts and ends in '<' and '>', respectively, it is likely
1116     // part of "#define <a/b.h>".
1117     if (CurrentToken->is(tok::less) && Line.Last->is(tok::greater)) {
1118       parseIncludeDirective();
1119       return LT_ImportStatement;
1120     }
1121 
1122     // In .proto files, top-level options and package statements are very
1123     // similar to import statements and should not be line-wrapped.
1124     if (Style.Language == FormatStyle::LK_Proto && Line.Level == 0 &&
1125         CurrentToken->isOneOf(Keywords.kw_option, Keywords.kw_package)) {
1126       next();
1127       if (CurrentToken && CurrentToken->is(tok::identifier)) {
1128         while (CurrentToken)
1129           next();
1130         return LT_ImportStatement;
1131       }
1132     }
1133 
1134     bool KeywordVirtualFound = false;
1135     bool ImportStatement = false;
1136 
1137     // import {...} from '...';
1138     if (Style.Language == FormatStyle::LK_JavaScript &&
1139         CurrentToken->is(Keywords.kw_import))
1140       ImportStatement = true;
1141 
1142     while (CurrentToken) {
1143       if (CurrentToken->is(tok::kw_virtual))
1144         KeywordVirtualFound = true;
1145       if (Style.Language == FormatStyle::LK_JavaScript) {
1146         // export {...} from '...';
1147         // An export followed by "from 'some string';" is a re-export from
1148         // another module identified by a URI and is treated as a
1149         // LT_ImportStatement (i.e. prevent wraps on it for long URIs).
1150         // Just "export {...};" or "export class ..." should not be treated as
1151         // an import in this sense.
1152         if (Line.First->is(tok::kw_export) &&
1153             CurrentToken->is(Keywords.kw_from) && CurrentToken->Next &&
1154             CurrentToken->Next->isStringLiteral())
1155           ImportStatement = true;
1156         if (isClosureImportStatement(*CurrentToken))
1157           ImportStatement = true;
1158       }
1159       if (!consumeToken())
1160         return LT_Invalid;
1161     }
1162     if (KeywordVirtualFound)
1163       return LT_VirtualFunctionDecl;
1164     if (ImportStatement)
1165       return LT_ImportStatement;
1166 
1167     if (Line.startsWith(TT_ObjCMethodSpecifier)) {
1168       if (Contexts.back().FirstObjCSelectorName)
1169         Contexts.back().FirstObjCSelectorName->LongestObjCSelectorName =
1170             Contexts.back().LongestObjCSelectorName;
1171       return LT_ObjCMethodDecl;
1172     }
1173 
1174     return LT_Other;
1175   }
1176 
1177 private:
1178   bool isClosureImportStatement(const FormatToken &Tok) {
1179     // FIXME: Closure-library specific stuff should not be hard-coded but be
1180     // configurable.
1181     return Tok.TokenText == "goog" && Tok.Next && Tok.Next->is(tok::period) &&
1182            Tok.Next->Next &&
1183            (Tok.Next->Next->TokenText == "module" ||
1184             Tok.Next->Next->TokenText == "provide" ||
1185             Tok.Next->Next->TokenText == "require" ||
1186             Tok.Next->Next->TokenText == "requireType" ||
1187             Tok.Next->Next->TokenText == "forwardDeclare") &&
1188            Tok.Next->Next->Next && Tok.Next->Next->Next->is(tok::l_paren);
1189   }
1190 
1191   void resetTokenMetadata(FormatToken *Token) {
1192     if (!Token)
1193       return;
1194 
1195     // Reset token type in case we have already looked at it and then
1196     // recovered from an error (e.g. failure to find the matching >).
1197     if (!CurrentToken->isOneOf(
1198             TT_LambdaLSquare, TT_LambdaLBrace, TT_ForEachMacro,
1199             TT_FunctionLBrace, TT_ImplicitStringLiteral, TT_InlineASMBrace,
1200             TT_JsFatArrow, TT_LambdaArrow, TT_OverloadedOperator,
1201             TT_RegexLiteral, TT_TemplateString, TT_ObjCStringLiteral))
1202       CurrentToken->Type = TT_Unknown;
1203     CurrentToken->Role.reset();
1204     CurrentToken->MatchingParen = nullptr;
1205     CurrentToken->FakeLParens.clear();
1206     CurrentToken->FakeRParens = 0;
1207   }
1208 
1209   void next() {
1210     if (CurrentToken) {
1211       CurrentToken->NestingLevel = Contexts.size() - 1;
1212       CurrentToken->BindingStrength = Contexts.back().BindingStrength;
1213       modifyContext(*CurrentToken);
1214       determineTokenType(*CurrentToken);
1215       CurrentToken = CurrentToken->Next;
1216     }
1217 
1218     resetTokenMetadata(CurrentToken);
1219   }
1220 
1221   /// A struct to hold information valid in a specific context, e.g.
1222   /// a pair of parenthesis.
1223   struct Context {
1224     Context(tok::TokenKind ContextKind, unsigned BindingStrength,
1225             bool IsExpression)
1226         : ContextKind(ContextKind), BindingStrength(BindingStrength),
1227           IsExpression(IsExpression) {}
1228 
1229     tok::TokenKind ContextKind;
1230     unsigned BindingStrength;
1231     bool IsExpression;
1232     unsigned LongestObjCSelectorName = 0;
1233     bool ColonIsForRangeExpr = false;
1234     bool ColonIsDictLiteral = false;
1235     bool ColonIsObjCMethodExpr = false;
1236     FormatToken *FirstObjCSelectorName = nullptr;
1237     FormatToken *FirstStartOfName = nullptr;
1238     bool CanBeExpression = true;
1239     bool InTemplateArgument = false;
1240     bool InCtorInitializer = false;
1241     bool InInheritanceList = false;
1242     bool CaretFound = false;
1243     bool IsForEachMacro = false;
1244     bool InCpp11AttributeSpecifier = false;
1245     bool InCSharpAttributeSpecifier = false;
1246   };
1247 
1248   /// Puts a new \c Context onto the stack \c Contexts for the lifetime
1249   /// of each instance.
1250   struct ScopedContextCreator {
1251     AnnotatingParser &P;
1252 
1253     ScopedContextCreator(AnnotatingParser &P, tok::TokenKind ContextKind,
1254                          unsigned Increase)
1255         : P(P) {
1256       P.Contexts.push_back(Context(ContextKind,
1257                                    P.Contexts.back().BindingStrength + Increase,
1258                                    P.Contexts.back().IsExpression));
1259     }
1260 
1261     ~ScopedContextCreator() { P.Contexts.pop_back(); }
1262   };
1263 
1264   void modifyContext(const FormatToken &Current) {
1265     if (Current.getPrecedence() == prec::Assignment &&
1266         !Line.First->isOneOf(tok::kw_template, tok::kw_using, tok::kw_return) &&
1267         // Type aliases use `type X = ...;` in TypeScript and can be exported
1268         // using `export type ...`.
1269         !(Style.Language == FormatStyle::LK_JavaScript &&
1270           (Line.startsWith(Keywords.kw_type, tok::identifier) ||
1271            Line.startsWith(tok::kw_export, Keywords.kw_type,
1272                            tok::identifier))) &&
1273         (!Current.Previous || Current.Previous->isNot(tok::kw_operator))) {
1274       Contexts.back().IsExpression = true;
1275       if (!Line.startsWith(TT_UnaryOperator)) {
1276         for (FormatToken *Previous = Current.Previous;
1277              Previous && Previous->Previous &&
1278              !Previous->Previous->isOneOf(tok::comma, tok::semi);
1279              Previous = Previous->Previous) {
1280           if (Previous->isOneOf(tok::r_square, tok::r_paren)) {
1281             Previous = Previous->MatchingParen;
1282             if (!Previous)
1283               break;
1284           }
1285           if (Previous->opensScope())
1286             break;
1287           if (Previous->isOneOf(TT_BinaryOperator, TT_UnaryOperator) &&
1288               Previous->isOneOf(tok::star, tok::amp, tok::ampamp) &&
1289               Previous->Previous && Previous->Previous->isNot(tok::equal))
1290             Previous->Type = TT_PointerOrReference;
1291         }
1292       }
1293     } else if (Current.is(tok::lessless) &&
1294                (!Current.Previous || !Current.Previous->is(tok::kw_operator))) {
1295       Contexts.back().IsExpression = true;
1296     } else if (Current.isOneOf(tok::kw_return, tok::kw_throw)) {
1297       Contexts.back().IsExpression = true;
1298     } else if (Current.is(TT_TrailingReturnArrow)) {
1299       Contexts.back().IsExpression = false;
1300     } else if (Current.is(TT_LambdaArrow) || Current.is(Keywords.kw_assert)) {
1301       Contexts.back().IsExpression = Style.Language == FormatStyle::LK_Java;
1302     } else if (Current.Previous &&
1303                Current.Previous->is(TT_CtorInitializerColon)) {
1304       Contexts.back().IsExpression = true;
1305       Contexts.back().InCtorInitializer = true;
1306     } else if (Current.Previous && Current.Previous->is(TT_InheritanceColon)) {
1307       Contexts.back().InInheritanceList = true;
1308     } else if (Current.isOneOf(tok::r_paren, tok::greater, tok::comma)) {
1309       for (FormatToken *Previous = Current.Previous;
1310            Previous && Previous->isOneOf(tok::star, tok::amp);
1311            Previous = Previous->Previous)
1312         Previous->Type = TT_PointerOrReference;
1313       if (Line.MustBeDeclaration && !Contexts.front().InCtorInitializer)
1314         Contexts.back().IsExpression = false;
1315     } else if (Current.is(tok::kw_new)) {
1316       Contexts.back().CanBeExpression = false;
1317     } else if (Current.isOneOf(tok::semi, tok::exclaim)) {
1318       // This should be the condition or increment in a for-loop.
1319       Contexts.back().IsExpression = true;
1320     }
1321   }
1322 
1323   void determineTokenType(FormatToken &Current) {
1324     if (!Current.is(TT_Unknown))
1325       // The token type is already known.
1326       return;
1327 
1328     if (Style.Language == FormatStyle::LK_JavaScript) {
1329       if (Current.is(tok::exclaim)) {
1330         if (Current.Previous &&
1331             (Current.Previous->isOneOf(tok::identifier, tok::kw_namespace,
1332                                        tok::r_paren, tok::r_square,
1333                                        tok::r_brace) ||
1334              Current.Previous->Tok.isLiteral())) {
1335           Current.Type = TT_JsNonNullAssertion;
1336           return;
1337         }
1338         if (Current.Next &&
1339             Current.Next->isOneOf(TT_BinaryOperator, Keywords.kw_as)) {
1340           Current.Type = TT_JsNonNullAssertion;
1341           return;
1342         }
1343       }
1344     }
1345 
1346     // Line.MightBeFunctionDecl can only be true after the parentheses of a
1347     // function declaration have been found. In this case, 'Current' is a
1348     // trailing token of this declaration and thus cannot be a name.
1349     if (Current.is(Keywords.kw_instanceof)) {
1350       Current.Type = TT_BinaryOperator;
1351     } else if (isStartOfName(Current) &&
1352                (!Line.MightBeFunctionDecl || Current.NestingLevel != 0)) {
1353       Contexts.back().FirstStartOfName = &Current;
1354       Current.Type = TT_StartOfName;
1355     } else if (Current.is(tok::semi)) {
1356       // Reset FirstStartOfName after finding a semicolon so that a for loop
1357       // with multiple increment statements is not confused with a for loop
1358       // having multiple variable declarations.
1359       Contexts.back().FirstStartOfName = nullptr;
1360     } else if (Current.isOneOf(tok::kw_auto, tok::kw___auto_type)) {
1361       AutoFound = true;
1362     } else if (Current.is(tok::arrow) &&
1363                Style.Language == FormatStyle::LK_Java) {
1364       Current.Type = TT_LambdaArrow;
1365     } else if (Current.is(tok::arrow) && AutoFound && Line.MustBeDeclaration &&
1366                Current.NestingLevel == 0) {
1367       Current.Type = TT_TrailingReturnArrow;
1368     } else if (Current.isOneOf(tok::star, tok::amp, tok::ampamp)) {
1369       Current.Type = determineStarAmpUsage(Current,
1370                                            Contexts.back().CanBeExpression &&
1371                                                Contexts.back().IsExpression,
1372                                            Contexts.back().InTemplateArgument);
1373     } else if (Current.isOneOf(tok::minus, tok::plus, tok::caret)) {
1374       Current.Type = determinePlusMinusCaretUsage(Current);
1375       if (Current.is(TT_UnaryOperator) && Current.is(tok::caret))
1376         Contexts.back().CaretFound = true;
1377     } else if (Current.isOneOf(tok::minusminus, tok::plusplus)) {
1378       Current.Type = determineIncrementUsage(Current);
1379     } else if (Current.isOneOf(tok::exclaim, tok::tilde)) {
1380       Current.Type = TT_UnaryOperator;
1381     } else if (Current.is(tok::question)) {
1382       if (Style.Language == FormatStyle::LK_JavaScript &&
1383           Line.MustBeDeclaration && !Contexts.back().IsExpression) {
1384         // In JavaScript, `interface X { foo?(): bar; }` is an optional method
1385         // on the interface, not a ternary expression.
1386         Current.Type = TT_JsTypeOptionalQuestion;
1387       } else {
1388         Current.Type = TT_ConditionalExpr;
1389       }
1390     } else if (Current.isBinaryOperator() &&
1391                (!Current.Previous || Current.Previous->isNot(tok::l_square)) &&
1392                (!Current.is(tok::greater) &&
1393                 Style.Language != FormatStyle::LK_TextProto)) {
1394       Current.Type = TT_BinaryOperator;
1395     } else if (Current.is(tok::comment)) {
1396       if (Current.TokenText.startswith("/*")) {
1397         if (Current.TokenText.endswith("*/"))
1398           Current.Type = TT_BlockComment;
1399         else
1400           // The lexer has for some reason determined a comment here. But we
1401           // cannot really handle it, if it isn't properly terminated.
1402           Current.Tok.setKind(tok::unknown);
1403       } else {
1404         Current.Type = TT_LineComment;
1405       }
1406     } else if (Current.is(tok::r_paren)) {
1407       if (rParenEndsCast(Current))
1408         Current.Type = TT_CastRParen;
1409       if (Current.MatchingParen && Current.Next &&
1410           !Current.Next->isBinaryOperator() &&
1411           !Current.Next->isOneOf(tok::semi, tok::colon, tok::l_brace,
1412                                  tok::comma, tok::period, tok::arrow,
1413                                  tok::coloncolon))
1414         if (FormatToken *AfterParen = Current.MatchingParen->Next) {
1415           // Make sure this isn't the return type of an Obj-C block declaration
1416           if (AfterParen->Tok.isNot(tok::caret)) {
1417             if (FormatToken *BeforeParen = Current.MatchingParen->Previous)
1418               if (BeforeParen->is(tok::identifier) &&
1419                   BeforeParen->TokenText == BeforeParen->TokenText.upper() &&
1420                   (!BeforeParen->Previous ||
1421                    BeforeParen->Previous->ClosesTemplateDeclaration))
1422                 Current.Type = TT_FunctionAnnotationRParen;
1423           }
1424         }
1425     } else if (Current.is(tok::at) && Current.Next &&
1426                Style.Language != FormatStyle::LK_JavaScript &&
1427                Style.Language != FormatStyle::LK_Java) {
1428       // In Java & JavaScript, "@..." is a decorator or annotation. In ObjC, it
1429       // marks declarations and properties that need special formatting.
1430       switch (Current.Next->Tok.getObjCKeywordID()) {
1431       case tok::objc_interface:
1432       case tok::objc_implementation:
1433       case tok::objc_protocol:
1434         Current.Type = TT_ObjCDecl;
1435         break;
1436       case tok::objc_property:
1437         Current.Type = TT_ObjCProperty;
1438         break;
1439       default:
1440         break;
1441       }
1442     } else if (Current.is(tok::period)) {
1443       FormatToken *PreviousNoComment = Current.getPreviousNonComment();
1444       if (PreviousNoComment &&
1445           PreviousNoComment->isOneOf(tok::comma, tok::l_brace))
1446         Current.Type = TT_DesignatedInitializerPeriod;
1447       else if (Style.Language == FormatStyle::LK_Java && Current.Previous &&
1448                Current.Previous->isOneOf(TT_JavaAnnotation,
1449                                          TT_LeadingJavaAnnotation)) {
1450         Current.Type = Current.Previous->Type;
1451       }
1452     } else if (canBeObjCSelectorComponent(Current) &&
1453                // FIXME(bug 36976): ObjC return types shouldn't use
1454                // TT_CastRParen.
1455                Current.Previous && Current.Previous->is(TT_CastRParen) &&
1456                Current.Previous->MatchingParen &&
1457                Current.Previous->MatchingParen->Previous &&
1458                Current.Previous->MatchingParen->Previous->is(
1459                    TT_ObjCMethodSpecifier)) {
1460       // This is the first part of an Objective-C selector name. (If there's no
1461       // colon after this, this is the only place which annotates the identifier
1462       // as a selector.)
1463       Current.Type = TT_SelectorName;
1464     } else if (Current.isOneOf(tok::identifier, tok::kw_const) &&
1465                Current.Previous &&
1466                !Current.Previous->isOneOf(tok::equal, tok::at) &&
1467                Line.MightBeFunctionDecl && Contexts.size() == 1) {
1468       // Line.MightBeFunctionDecl can only be true after the parentheses of a
1469       // function declaration have been found.
1470       Current.Type = TT_TrailingAnnotation;
1471     } else if ((Style.Language == FormatStyle::LK_Java ||
1472                 Style.Language == FormatStyle::LK_JavaScript) &&
1473                Current.Previous) {
1474       if (Current.Previous->is(tok::at) &&
1475           Current.isNot(Keywords.kw_interface)) {
1476         const FormatToken &AtToken = *Current.Previous;
1477         const FormatToken *Previous = AtToken.getPreviousNonComment();
1478         if (!Previous || Previous->is(TT_LeadingJavaAnnotation))
1479           Current.Type = TT_LeadingJavaAnnotation;
1480         else
1481           Current.Type = TT_JavaAnnotation;
1482       } else if (Current.Previous->is(tok::period) &&
1483                  Current.Previous->isOneOf(TT_JavaAnnotation,
1484                                            TT_LeadingJavaAnnotation)) {
1485         Current.Type = Current.Previous->Type;
1486       }
1487     }
1488   }
1489 
1490   /// Take a guess at whether \p Tok starts a name of a function or
1491   /// variable declaration.
1492   ///
1493   /// This is a heuristic based on whether \p Tok is an identifier following
1494   /// something that is likely a type.
1495   bool isStartOfName(const FormatToken &Tok) {
1496     if (Tok.isNot(tok::identifier) || !Tok.Previous)
1497       return false;
1498 
1499     if (Tok.Previous->isOneOf(TT_LeadingJavaAnnotation, Keywords.kw_instanceof,
1500                               Keywords.kw_as))
1501       return false;
1502     if (Style.Language == FormatStyle::LK_JavaScript &&
1503         Tok.Previous->is(Keywords.kw_in))
1504       return false;
1505 
1506     // Skip "const" as it does not have an influence on whether this is a name.
1507     FormatToken *PreviousNotConst = Tok.getPreviousNonComment();
1508     while (PreviousNotConst && PreviousNotConst->is(tok::kw_const))
1509       PreviousNotConst = PreviousNotConst->getPreviousNonComment();
1510 
1511     if (!PreviousNotConst)
1512       return false;
1513 
1514     bool IsPPKeyword = PreviousNotConst->is(tok::identifier) &&
1515                        PreviousNotConst->Previous &&
1516                        PreviousNotConst->Previous->is(tok::hash);
1517 
1518     if (PreviousNotConst->is(TT_TemplateCloser))
1519       return PreviousNotConst && PreviousNotConst->MatchingParen &&
1520              PreviousNotConst->MatchingParen->Previous &&
1521              PreviousNotConst->MatchingParen->Previous->isNot(tok::period) &&
1522              PreviousNotConst->MatchingParen->Previous->isNot(tok::kw_template);
1523 
1524     if (PreviousNotConst->is(tok::r_paren) && PreviousNotConst->MatchingParen &&
1525         PreviousNotConst->MatchingParen->Previous &&
1526         PreviousNotConst->MatchingParen->Previous->is(tok::kw_decltype))
1527       return true;
1528 
1529     return (!IsPPKeyword &&
1530             PreviousNotConst->isOneOf(tok::identifier, tok::kw_auto)) ||
1531            PreviousNotConst->is(TT_PointerOrReference) ||
1532            PreviousNotConst->isSimpleTypeSpecifier();
1533   }
1534 
1535   /// Determine whether ')' is ending a cast.
1536   bool rParenEndsCast(const FormatToken &Tok) {
1537     // C-style casts are only used in C++ and Java.
1538     if (!Style.isCpp() && Style.Language != FormatStyle::LK_Java)
1539       return false;
1540 
1541     // Empty parens aren't casts and there are no casts at the end of the line.
1542     if (Tok.Previous == Tok.MatchingParen || !Tok.Next || !Tok.MatchingParen)
1543       return false;
1544 
1545     FormatToken *LeftOfParens = Tok.MatchingParen->getPreviousNonComment();
1546     if (LeftOfParens) {
1547       // If there is a closing parenthesis left of the current parentheses,
1548       // look past it as these might be chained casts.
1549       if (LeftOfParens->is(tok::r_paren)) {
1550         if (!LeftOfParens->MatchingParen ||
1551             !LeftOfParens->MatchingParen->Previous)
1552           return false;
1553         LeftOfParens = LeftOfParens->MatchingParen->Previous;
1554       }
1555 
1556       // If there is an identifier (or with a few exceptions a keyword) right
1557       // before the parentheses, this is unlikely to be a cast.
1558       if (LeftOfParens->Tok.getIdentifierInfo() &&
1559           !LeftOfParens->isOneOf(Keywords.kw_in, tok::kw_return, tok::kw_case,
1560                                  tok::kw_delete))
1561         return false;
1562 
1563       // Certain other tokens right before the parentheses are also signals that
1564       // this cannot be a cast.
1565       if (LeftOfParens->isOneOf(tok::at, tok::r_square, TT_OverloadedOperator,
1566                                 TT_TemplateCloser, tok::ellipsis))
1567         return false;
1568     }
1569 
1570     if (Tok.Next->is(tok::question))
1571       return false;
1572 
1573     // As Java has no function types, a "(" after the ")" likely means that this
1574     // is a cast.
1575     if (Style.Language == FormatStyle::LK_Java && Tok.Next->is(tok::l_paren))
1576       return true;
1577 
1578     // If a (non-string) literal follows, this is likely a cast.
1579     if (Tok.Next->isNot(tok::string_literal) &&
1580         (Tok.Next->Tok.isLiteral() ||
1581          Tok.Next->isOneOf(tok::kw_sizeof, tok::kw_alignof)))
1582       return true;
1583 
1584     // Heuristically try to determine whether the parentheses contain a type.
1585     bool ParensAreType =
1586         !Tok.Previous ||
1587         Tok.Previous->isOneOf(TT_PointerOrReference, TT_TemplateCloser) ||
1588         Tok.Previous->isSimpleTypeSpecifier();
1589     bool ParensCouldEndDecl =
1590         Tok.Next->isOneOf(tok::equal, tok::semi, tok::l_brace, tok::greater);
1591     if (ParensAreType && !ParensCouldEndDecl)
1592       return true;
1593 
1594     // At this point, we heuristically assume that there are no casts at the
1595     // start of the line. We assume that we have found most cases where there
1596     // are by the logic above, e.g. "(void)x;".
1597     if (!LeftOfParens)
1598       return false;
1599 
1600     // Certain token types inside the parentheses mean that this can't be a
1601     // cast.
1602     for (const FormatToken *Token = Tok.MatchingParen->Next; Token != &Tok;
1603          Token = Token->Next)
1604       if (Token->is(TT_BinaryOperator))
1605         return false;
1606 
1607     // If the following token is an identifier or 'this', this is a cast. All
1608     // cases where this can be something else are handled above.
1609     if (Tok.Next->isOneOf(tok::identifier, tok::kw_this))
1610       return true;
1611 
1612     if (!Tok.Next->Next)
1613       return false;
1614 
1615     // If the next token after the parenthesis is a unary operator, assume
1616     // that this is cast, unless there are unexpected tokens inside the
1617     // parenthesis.
1618     bool NextIsUnary =
1619         Tok.Next->isUnaryOperator() || Tok.Next->isOneOf(tok::amp, tok::star);
1620     if (!NextIsUnary || Tok.Next->is(tok::plus) ||
1621         !Tok.Next->Next->isOneOf(tok::identifier, tok::numeric_constant))
1622       return false;
1623     // Search for unexpected tokens.
1624     for (FormatToken *Prev = Tok.Previous; Prev != Tok.MatchingParen;
1625          Prev = Prev->Previous) {
1626       if (!Prev->isOneOf(tok::kw_const, tok::identifier, tok::coloncolon))
1627         return false;
1628     }
1629     return true;
1630   }
1631 
1632   /// Return the type of the given token assuming it is * or &.
1633   TokenType determineStarAmpUsage(const FormatToken &Tok, bool IsExpression,
1634                                   bool InTemplateArgument) {
1635     if (Style.Language == FormatStyle::LK_JavaScript)
1636       return TT_BinaryOperator;
1637 
1638     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1639     if (!PrevToken)
1640       return TT_UnaryOperator;
1641 
1642     const FormatToken *NextToken = Tok.getNextNonComment();
1643     if (!NextToken ||
1644         NextToken->isOneOf(tok::arrow, tok::equal, tok::kw_const) ||
1645         (NextToken->is(tok::l_brace) && !NextToken->getNextNonComment()))
1646       return TT_PointerOrReference;
1647 
1648     if (PrevToken->is(tok::coloncolon))
1649       return TT_PointerOrReference;
1650 
1651     if (PrevToken->isOneOf(tok::l_paren, tok::l_square, tok::l_brace,
1652                            tok::comma, tok::semi, tok::kw_return, tok::colon,
1653                            tok::equal, tok::kw_delete, tok::kw_sizeof,
1654                            tok::kw_throw) ||
1655         PrevToken->isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
1656                            TT_UnaryOperator, TT_CastRParen))
1657       return TT_UnaryOperator;
1658 
1659     if (NextToken->is(tok::l_square) && NextToken->isNot(TT_LambdaLSquare))
1660       return TT_PointerOrReference;
1661     if (NextToken->is(tok::kw_operator) && !IsExpression)
1662       return TT_PointerOrReference;
1663     if (NextToken->isOneOf(tok::comma, tok::semi))
1664       return TT_PointerOrReference;
1665 
1666     if (PrevToken->is(tok::r_paren) && PrevToken->MatchingParen) {
1667       FormatToken *TokenBeforeMatchingParen =
1668           PrevToken->MatchingParen->getPreviousNonComment();
1669       if (TokenBeforeMatchingParen &&
1670           TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype))
1671         return TT_PointerOrReference;
1672     }
1673 
1674     if (PrevToken->Tok.isLiteral() ||
1675         PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::kw_true,
1676                            tok::kw_false, tok::r_brace) ||
1677         NextToken->Tok.isLiteral() ||
1678         NextToken->isOneOf(tok::kw_true, tok::kw_false) ||
1679         NextToken->isUnaryOperator() ||
1680         // If we know we're in a template argument, there are no named
1681         // declarations. Thus, having an identifier on the right-hand side
1682         // indicates a binary operator.
1683         (InTemplateArgument && NextToken->Tok.isAnyIdentifier()))
1684       return TT_BinaryOperator;
1685 
1686     // "&&(" is quite unlikely to be two successive unary "&".
1687     if (Tok.is(tok::ampamp) && NextToken && NextToken->is(tok::l_paren))
1688       return TT_BinaryOperator;
1689 
1690     // This catches some cases where evaluation order is used as control flow:
1691     //   aaa && aaa->f();
1692     const FormatToken *NextNextToken = NextToken->getNextNonComment();
1693     if (NextNextToken && NextNextToken->is(tok::arrow))
1694       return TT_BinaryOperator;
1695 
1696     // It is very unlikely that we are going to find a pointer or reference type
1697     // definition on the RHS of an assignment.
1698     if (IsExpression && !Contexts.back().CaretFound)
1699       return TT_BinaryOperator;
1700 
1701     return TT_PointerOrReference;
1702   }
1703 
1704   TokenType determinePlusMinusCaretUsage(const FormatToken &Tok) {
1705     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1706     if (!PrevToken)
1707       return TT_UnaryOperator;
1708 
1709     if (PrevToken->isOneOf(TT_CastRParen, TT_UnaryOperator))
1710       // This must be a sequence of leading unary operators.
1711       return TT_UnaryOperator;
1712 
1713     // Use heuristics to recognize unary operators.
1714     if (PrevToken->isOneOf(tok::equal, tok::l_paren, tok::comma, tok::l_square,
1715                            tok::question, tok::colon, tok::kw_return,
1716                            tok::kw_case, tok::at, tok::l_brace))
1717       return TT_UnaryOperator;
1718 
1719     // There can't be two consecutive binary operators.
1720     if (PrevToken->is(TT_BinaryOperator))
1721       return TT_UnaryOperator;
1722 
1723     // Fall back to marking the token as binary operator.
1724     return TT_BinaryOperator;
1725   }
1726 
1727   /// Determine whether ++/-- are pre- or post-increments/-decrements.
1728   TokenType determineIncrementUsage(const FormatToken &Tok) {
1729     const FormatToken *PrevToken = Tok.getPreviousNonComment();
1730     if (!PrevToken || PrevToken->is(TT_CastRParen))
1731       return TT_UnaryOperator;
1732     if (PrevToken->isOneOf(tok::r_paren, tok::r_square, tok::identifier))
1733       return TT_TrailingUnaryOperator;
1734 
1735     return TT_UnaryOperator;
1736   }
1737 
1738   SmallVector<Context, 8> Contexts;
1739 
1740   const FormatStyle &Style;
1741   AnnotatedLine &Line;
1742   FormatToken *CurrentToken;
1743   bool AutoFound;
1744   const AdditionalKeywords &Keywords;
1745 
1746   // Set of "<" tokens that do not open a template parameter list. If parseAngle
1747   // determines that a specific token can't be a template opener, it will make
1748   // same decision irrespective of the decisions for tokens leading up to it.
1749   // Store this information to prevent this from causing exponential runtime.
1750   llvm::SmallPtrSet<FormatToken *, 16> NonTemplateLess;
1751 };
1752 
1753 static const int PrecedenceUnaryOperator = prec::PointerToMember + 1;
1754 static const int PrecedenceArrowAndPeriod = prec::PointerToMember + 2;
1755 
1756 /// Parses binary expressions by inserting fake parenthesis based on
1757 /// operator precedence.
1758 class ExpressionParser {
1759 public:
1760   ExpressionParser(const FormatStyle &Style, const AdditionalKeywords &Keywords,
1761                    AnnotatedLine &Line)
1762       : Style(Style), Keywords(Keywords), Current(Line.First) {}
1763 
1764   /// Parse expressions with the given operator precedence.
1765   void parse(int Precedence = 0) {
1766     // Skip 'return' and ObjC selector colons as they are not part of a binary
1767     // expression.
1768     while (Current && (Current->is(tok::kw_return) ||
1769                        (Current->is(tok::colon) &&
1770                         Current->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))))
1771       next();
1772 
1773     if (!Current || Precedence > PrecedenceArrowAndPeriod)
1774       return;
1775 
1776     // Conditional expressions need to be parsed separately for proper nesting.
1777     if (Precedence == prec::Conditional) {
1778       parseConditionalExpr();
1779       return;
1780     }
1781 
1782     // Parse unary operators, which all have a higher precedence than binary
1783     // operators.
1784     if (Precedence == PrecedenceUnaryOperator) {
1785       parseUnaryOperator();
1786       return;
1787     }
1788 
1789     FormatToken *Start = Current;
1790     FormatToken *LatestOperator = nullptr;
1791     unsigned OperatorIndex = 0;
1792 
1793     while (Current) {
1794       // Consume operators with higher precedence.
1795       parse(Precedence + 1);
1796 
1797       int CurrentPrecedence = getCurrentPrecedence();
1798 
1799       if (Current && Current->is(TT_SelectorName) &&
1800           Precedence == CurrentPrecedence) {
1801         if (LatestOperator)
1802           addFakeParenthesis(Start, prec::Level(Precedence));
1803         Start = Current;
1804       }
1805 
1806       // At the end of the line or when an operator with higher precedence is
1807       // found, insert fake parenthesis and return.
1808       if (!Current ||
1809           (Current->closesScope() &&
1810            (Current->MatchingParen || Current->is(TT_TemplateString))) ||
1811           (CurrentPrecedence != -1 && CurrentPrecedence < Precedence) ||
1812           (CurrentPrecedence == prec::Conditional &&
1813            Precedence == prec::Assignment && Current->is(tok::colon))) {
1814         break;
1815       }
1816 
1817       // Consume scopes: (), [], <> and {}
1818       if (Current->opensScope()) {
1819         // In fragment of a JavaScript template string can look like '}..${' and
1820         // thus close a scope and open a new one at the same time.
1821         while (Current && (!Current->closesScope() || Current->opensScope())) {
1822           next();
1823           parse();
1824         }
1825         next();
1826       } else {
1827         // Operator found.
1828         if (CurrentPrecedence == Precedence) {
1829           if (LatestOperator)
1830             LatestOperator->NextOperator = Current;
1831           LatestOperator = Current;
1832           Current->OperatorIndex = OperatorIndex;
1833           ++OperatorIndex;
1834         }
1835         next(/*SkipPastLeadingComments=*/Precedence > 0);
1836       }
1837     }
1838 
1839     if (LatestOperator && (Current || Precedence > 0)) {
1840       // LatestOperator->LastOperator = true;
1841       if (Precedence == PrecedenceArrowAndPeriod) {
1842         // Call expressions don't have a binary operator precedence.
1843         addFakeParenthesis(Start, prec::Unknown);
1844       } else {
1845         addFakeParenthesis(Start, prec::Level(Precedence));
1846       }
1847     }
1848   }
1849 
1850 private:
1851   /// Gets the precedence (+1) of the given token for binary operators
1852   /// and other tokens that we treat like binary operators.
1853   int getCurrentPrecedence() {
1854     if (Current) {
1855       const FormatToken *NextNonComment = Current->getNextNonComment();
1856       if (Current->is(TT_ConditionalExpr))
1857         return prec::Conditional;
1858       if (NextNonComment && Current->is(TT_SelectorName) &&
1859           (NextNonComment->isOneOf(TT_DictLiteral, TT_JsTypeColon) ||
1860            ((Style.Language == FormatStyle::LK_Proto ||
1861              Style.Language == FormatStyle::LK_TextProto) &&
1862             NextNonComment->is(tok::less))))
1863         return prec::Assignment;
1864       if (Current->is(TT_JsComputedPropertyName))
1865         return prec::Assignment;
1866       if (Current->is(TT_LambdaArrow))
1867         return prec::Comma;
1868       if (Current->is(TT_JsFatArrow))
1869         return prec::Assignment;
1870       if (Current->isOneOf(tok::semi, TT_InlineASMColon, TT_SelectorName) ||
1871           (Current->is(tok::comment) && NextNonComment &&
1872            NextNonComment->is(TT_SelectorName)))
1873         return 0;
1874       if (Current->is(TT_RangeBasedForLoopColon))
1875         return prec::Comma;
1876       if ((Style.Language == FormatStyle::LK_Java ||
1877            Style.Language == FormatStyle::LK_JavaScript) &&
1878           Current->is(Keywords.kw_instanceof))
1879         return prec::Relational;
1880       if (Style.Language == FormatStyle::LK_JavaScript &&
1881           Current->isOneOf(Keywords.kw_in, Keywords.kw_as))
1882         return prec::Relational;
1883       if (Current->is(TT_BinaryOperator) || Current->is(tok::comma))
1884         return Current->getPrecedence();
1885       if (Current->isOneOf(tok::period, tok::arrow))
1886         return PrecedenceArrowAndPeriod;
1887       if ((Style.Language == FormatStyle::LK_Java ||
1888            Style.Language == FormatStyle::LK_JavaScript) &&
1889           Current->isOneOf(Keywords.kw_extends, Keywords.kw_implements,
1890                            Keywords.kw_throws))
1891         return 0;
1892     }
1893     return -1;
1894   }
1895 
1896   void addFakeParenthesis(FormatToken *Start, prec::Level Precedence) {
1897     Start->FakeLParens.push_back(Precedence);
1898     if (Precedence > prec::Unknown)
1899       Start->StartsBinaryExpression = true;
1900     if (Current) {
1901       FormatToken *Previous = Current->Previous;
1902       while (Previous->is(tok::comment) && Previous->Previous)
1903         Previous = Previous->Previous;
1904       ++Previous->FakeRParens;
1905       if (Precedence > prec::Unknown)
1906         Previous->EndsBinaryExpression = true;
1907     }
1908   }
1909 
1910   /// Parse unary operator expressions and surround them with fake
1911   /// parentheses if appropriate.
1912   void parseUnaryOperator() {
1913     llvm::SmallVector<FormatToken *, 2> Tokens;
1914     while (Current && Current->is(TT_UnaryOperator)) {
1915       Tokens.push_back(Current);
1916       next();
1917     }
1918     parse(PrecedenceArrowAndPeriod);
1919     for (FormatToken *Token : llvm::reverse(Tokens))
1920       // The actual precedence doesn't matter.
1921       addFakeParenthesis(Token, prec::Unknown);
1922   }
1923 
1924   void parseConditionalExpr() {
1925     while (Current && Current->isTrailingComment()) {
1926       next();
1927     }
1928     FormatToken *Start = Current;
1929     parse(prec::LogicalOr);
1930     if (!Current || !Current->is(tok::question))
1931       return;
1932     next();
1933     parse(prec::Assignment);
1934     if (!Current || Current->isNot(TT_ConditionalExpr))
1935       return;
1936     next();
1937     parse(prec::Assignment);
1938     addFakeParenthesis(Start, prec::Conditional);
1939   }
1940 
1941   void next(bool SkipPastLeadingComments = true) {
1942     if (Current)
1943       Current = Current->Next;
1944     while (Current &&
1945            (Current->NewlinesBefore == 0 || SkipPastLeadingComments) &&
1946            Current->isTrailingComment())
1947       Current = Current->Next;
1948   }
1949 
1950   const FormatStyle &Style;
1951   const AdditionalKeywords &Keywords;
1952   FormatToken *Current;
1953 };
1954 
1955 } // end anonymous namespace
1956 
1957 void TokenAnnotator::setCommentLineLevels(
1958     SmallVectorImpl<AnnotatedLine *> &Lines) {
1959   const AnnotatedLine *NextNonCommentLine = nullptr;
1960   for (SmallVectorImpl<AnnotatedLine *>::reverse_iterator I = Lines.rbegin(),
1961                                                           E = Lines.rend();
1962        I != E; ++I) {
1963     bool CommentLine = true;
1964     for (const FormatToken *Tok = (*I)->First; Tok; Tok = Tok->Next) {
1965       if (!Tok->is(tok::comment)) {
1966         CommentLine = false;
1967         break;
1968       }
1969     }
1970 
1971     // If the comment is currently aligned with the line immediately following
1972     // it, that's probably intentional and we should keep it.
1973     if (NextNonCommentLine && CommentLine &&
1974         NextNonCommentLine->First->NewlinesBefore <= 1 &&
1975         NextNonCommentLine->First->OriginalColumn ==
1976             (*I)->First->OriginalColumn) {
1977       // Align comments for preprocessor lines with the # in column 0 if
1978       // preprocessor lines are not indented. Otherwise, align with the next
1979       // line.
1980       (*I)->Level =
1981           (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
1982            (NextNonCommentLine->Type == LT_PreprocessorDirective ||
1983             NextNonCommentLine->Type == LT_ImportStatement))
1984               ? 0
1985               : NextNonCommentLine->Level;
1986     } else {
1987       NextNonCommentLine = (*I)->First->isNot(tok::r_brace) ? (*I) : nullptr;
1988     }
1989 
1990     setCommentLineLevels((*I)->Children);
1991   }
1992 }
1993 
1994 static unsigned maxNestingDepth(const AnnotatedLine &Line) {
1995   unsigned Result = 0;
1996   for (const auto *Tok = Line.First; Tok != nullptr; Tok = Tok->Next)
1997     Result = std::max(Result, Tok->NestingLevel);
1998   return Result;
1999 }
2000 
2001 void TokenAnnotator::annotate(AnnotatedLine &Line) {
2002   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2003                                                   E = Line.Children.end();
2004        I != E; ++I) {
2005     annotate(**I);
2006   }
2007   AnnotatingParser Parser(Style, Line, Keywords);
2008   Line.Type = Parser.parseLine();
2009 
2010   // With very deep nesting, ExpressionParser uses lots of stack and the
2011   // formatting algorithm is very slow. We're not going to do a good job here
2012   // anyway - it's probably generated code being formatted by mistake.
2013   // Just skip the whole line.
2014   if (maxNestingDepth(Line) > 50)
2015     Line.Type = LT_Invalid;
2016 
2017   if (Line.Type == LT_Invalid)
2018     return;
2019 
2020   ExpressionParser ExprParser(Style, Keywords, Line);
2021   ExprParser.parse();
2022 
2023   if (Line.startsWith(TT_ObjCMethodSpecifier))
2024     Line.Type = LT_ObjCMethodDecl;
2025   else if (Line.startsWith(TT_ObjCDecl))
2026     Line.Type = LT_ObjCDecl;
2027   else if (Line.startsWith(TT_ObjCProperty))
2028     Line.Type = LT_ObjCProperty;
2029 
2030   Line.First->SpacesRequiredBefore = 1;
2031   Line.First->CanBreakBefore = Line.First->MustBreakBefore;
2032 }
2033 
2034 // This function heuristically determines whether 'Current' starts the name of a
2035 // function declaration.
2036 static bool isFunctionDeclarationName(const FormatToken &Current,
2037                                       const AnnotatedLine &Line) {
2038   auto skipOperatorName = [](const FormatToken *Next) -> const FormatToken * {
2039     for (; Next; Next = Next->Next) {
2040       if (Next->is(TT_OverloadedOperatorLParen))
2041         return Next;
2042       if (Next->is(TT_OverloadedOperator))
2043         continue;
2044       if (Next->isOneOf(tok::kw_new, tok::kw_delete)) {
2045         // For 'new[]' and 'delete[]'.
2046         if (Next->Next && Next->Next->is(tok::l_square) && Next->Next->Next &&
2047             Next->Next->Next->is(tok::r_square))
2048           Next = Next->Next->Next;
2049         continue;
2050       }
2051 
2052       break;
2053     }
2054     return nullptr;
2055   };
2056 
2057   // Find parentheses of parameter list.
2058   const FormatToken *Next = Current.Next;
2059   if (Current.is(tok::kw_operator)) {
2060     if (Current.Previous && Current.Previous->is(tok::coloncolon))
2061       return false;
2062     Next = skipOperatorName(Next);
2063   } else {
2064     if (!Current.is(TT_StartOfName) || Current.NestingLevel != 0)
2065       return false;
2066     for (; Next; Next = Next->Next) {
2067       if (Next->is(TT_TemplateOpener)) {
2068         Next = Next->MatchingParen;
2069       } else if (Next->is(tok::coloncolon)) {
2070         Next = Next->Next;
2071         if (!Next)
2072           return false;
2073         if (Next->is(tok::kw_operator)) {
2074           Next = skipOperatorName(Next->Next);
2075           break;
2076         }
2077         if (!Next->is(tok::identifier))
2078           return false;
2079       } else if (Next->is(tok::l_paren)) {
2080         break;
2081       } else {
2082         return false;
2083       }
2084     }
2085   }
2086 
2087   // Check whether parameter list can belong to a function declaration.
2088   if (!Next || !Next->is(tok::l_paren) || !Next->MatchingParen)
2089     return false;
2090   // If the lines ends with "{", this is likely an function definition.
2091   if (Line.Last->is(tok::l_brace))
2092     return true;
2093   if (Next->Next == Next->MatchingParen)
2094     return true; // Empty parentheses.
2095   // If there is an &/&& after the r_paren, this is likely a function.
2096   if (Next->MatchingParen->Next &&
2097       Next->MatchingParen->Next->is(TT_PointerOrReference))
2098     return true;
2099   for (const FormatToken *Tok = Next->Next; Tok && Tok != Next->MatchingParen;
2100        Tok = Tok->Next) {
2101     if (Tok->isOneOf(tok::l_paren, TT_TemplateOpener) && Tok->MatchingParen) {
2102       Tok = Tok->MatchingParen;
2103       continue;
2104     }
2105     if (Tok->is(tok::kw_const) || Tok->isSimpleTypeSpecifier() ||
2106         Tok->isOneOf(TT_PointerOrReference, TT_StartOfName, tok::ellipsis))
2107       return true;
2108     if (Tok->isOneOf(tok::l_brace, tok::string_literal, TT_ObjCMethodExpr) ||
2109         Tok->Tok.isLiteral())
2110       return false;
2111   }
2112   return false;
2113 }
2114 
2115 bool TokenAnnotator::mustBreakForReturnType(const AnnotatedLine &Line) const {
2116   assert(Line.MightBeFunctionDecl);
2117 
2118   if ((Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_TopLevel ||
2119        Style.AlwaysBreakAfterReturnType ==
2120            FormatStyle::RTBS_TopLevelDefinitions) &&
2121       Line.Level > 0)
2122     return false;
2123 
2124   switch (Style.AlwaysBreakAfterReturnType) {
2125   case FormatStyle::RTBS_None:
2126     return false;
2127   case FormatStyle::RTBS_All:
2128   case FormatStyle::RTBS_TopLevel:
2129     return true;
2130   case FormatStyle::RTBS_AllDefinitions:
2131   case FormatStyle::RTBS_TopLevelDefinitions:
2132     return Line.mightBeFunctionDefinition();
2133   }
2134 
2135   return false;
2136 }
2137 
2138 void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
2139   for (SmallVectorImpl<AnnotatedLine *>::iterator I = Line.Children.begin(),
2140                                                   E = Line.Children.end();
2141        I != E; ++I) {
2142     calculateFormattingInformation(**I);
2143   }
2144 
2145   Line.First->TotalLength =
2146       Line.First->IsMultiline ? Style.ColumnLimit
2147                               : Line.FirstStartColumn + Line.First->ColumnWidth;
2148   FormatToken *Current = Line.First->Next;
2149   bool InFunctionDecl = Line.MightBeFunctionDecl;
2150   while (Current) {
2151     if (isFunctionDeclarationName(*Current, Line))
2152       Current->Type = TT_FunctionDeclarationName;
2153     if (Current->is(TT_LineComment)) {
2154       if (Current->Previous->BlockKind == BK_BracedInit &&
2155           Current->Previous->opensScope())
2156         Current->SpacesRequiredBefore = Style.Cpp11BracedListStyle ? 0 : 1;
2157       else
2158         Current->SpacesRequiredBefore = Style.SpacesBeforeTrailingComments;
2159 
2160       // If we find a trailing comment, iterate backwards to determine whether
2161       // it seems to relate to a specific parameter. If so, break before that
2162       // parameter to avoid changing the comment's meaning. E.g. don't move 'b'
2163       // to the previous line in:
2164       //   SomeFunction(a,
2165       //                b, // comment
2166       //                c);
2167       if (!Current->HasUnescapedNewline) {
2168         for (FormatToken *Parameter = Current->Previous; Parameter;
2169              Parameter = Parameter->Previous) {
2170           if (Parameter->isOneOf(tok::comment, tok::r_brace))
2171             break;
2172           if (Parameter->Previous && Parameter->Previous->is(tok::comma)) {
2173             if (!Parameter->Previous->is(TT_CtorInitializerComma) &&
2174                 Parameter->HasUnescapedNewline)
2175               Parameter->MustBreakBefore = true;
2176             break;
2177           }
2178         }
2179       }
2180     } else if (Current->SpacesRequiredBefore == 0 &&
2181                spaceRequiredBefore(Line, *Current)) {
2182       Current->SpacesRequiredBefore = 1;
2183     }
2184 
2185     Current->MustBreakBefore =
2186         Current->MustBreakBefore || mustBreakBefore(Line, *Current);
2187 
2188     if (!Current->MustBreakBefore && InFunctionDecl &&
2189         Current->is(TT_FunctionDeclarationName))
2190       Current->MustBreakBefore = mustBreakForReturnType(Line);
2191 
2192     Current->CanBreakBefore =
2193         Current->MustBreakBefore || canBreakBefore(Line, *Current);
2194     unsigned ChildSize = 0;
2195     if (Current->Previous->Children.size() == 1) {
2196       FormatToken &LastOfChild = *Current->Previous->Children[0]->Last;
2197       ChildSize = LastOfChild.isTrailingComment() ? Style.ColumnLimit
2198                                                   : LastOfChild.TotalLength + 1;
2199     }
2200     const FormatToken *Prev = Current->Previous;
2201     if (Current->MustBreakBefore || Prev->Children.size() > 1 ||
2202         (Prev->Children.size() == 1 &&
2203          Prev->Children[0]->First->MustBreakBefore) ||
2204         Current->IsMultiline)
2205       Current->TotalLength = Prev->TotalLength + Style.ColumnLimit;
2206     else
2207       Current->TotalLength = Prev->TotalLength + Current->ColumnWidth +
2208                              ChildSize + Current->SpacesRequiredBefore;
2209 
2210     if (Current->is(TT_CtorInitializerColon))
2211       InFunctionDecl = false;
2212 
2213     // FIXME: Only calculate this if CanBreakBefore is true once static
2214     // initializers etc. are sorted out.
2215     // FIXME: Move magic numbers to a better place.
2216 
2217     // Reduce penalty for aligning ObjC method arguments using the colon
2218     // alignment as this is the canonical way (still prefer fitting everything
2219     // into one line if possible). Trying to fit a whole expression into one
2220     // line should not force other line breaks (e.g. when ObjC method
2221     // expression is a part of other expression).
2222     Current->SplitPenalty = splitPenalty(Line, *Current, InFunctionDecl);
2223     if (Style.Language == FormatStyle::LK_ObjC &&
2224         Current->is(TT_SelectorName) && Current->ParameterIndex > 0) {
2225       if (Current->ParameterIndex == 1)
2226         Current->SplitPenalty += 5 * Current->BindingStrength;
2227     } else {
2228       Current->SplitPenalty += 20 * Current->BindingStrength;
2229     }
2230 
2231     Current = Current->Next;
2232   }
2233 
2234   calculateUnbreakableTailLengths(Line);
2235   unsigned IndentLevel = Line.Level;
2236   for (Current = Line.First; Current != nullptr; Current = Current->Next) {
2237     if (Current->Role)
2238       Current->Role->precomputeFormattingInfos(Current);
2239     if (Current->MatchingParen &&
2240         Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
2241       assert(IndentLevel > 0);
2242       --IndentLevel;
2243     }
2244     Current->IndentLevel = IndentLevel;
2245     if (Current->opensBlockOrBlockTypeList(Style))
2246       ++IndentLevel;
2247   }
2248 
2249   LLVM_DEBUG({ printDebugInfo(Line); });
2250 }
2251 
2252 void TokenAnnotator::calculateUnbreakableTailLengths(AnnotatedLine &Line) {
2253   unsigned UnbreakableTailLength = 0;
2254   FormatToken *Current = Line.Last;
2255   while (Current) {
2256     Current->UnbreakableTailLength = UnbreakableTailLength;
2257     if (Current->CanBreakBefore ||
2258         Current->isOneOf(tok::comment, tok::string_literal)) {
2259       UnbreakableTailLength = 0;
2260     } else {
2261       UnbreakableTailLength +=
2262           Current->ColumnWidth + Current->SpacesRequiredBefore;
2263     }
2264     Current = Current->Previous;
2265   }
2266 }
2267 
2268 unsigned TokenAnnotator::splitPenalty(const AnnotatedLine &Line,
2269                                       const FormatToken &Tok,
2270                                       bool InFunctionDecl) {
2271   const FormatToken &Left = *Tok.Previous;
2272   const FormatToken &Right = Tok;
2273 
2274   if (Left.is(tok::semi))
2275     return 0;
2276 
2277   if (Style.Language == FormatStyle::LK_Java) {
2278     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_throws))
2279       return 1;
2280     if (Right.is(Keywords.kw_implements))
2281       return 2;
2282     if (Left.is(tok::comma) && Left.NestingLevel == 0)
2283       return 3;
2284   } else if (Style.Language == FormatStyle::LK_JavaScript) {
2285     if (Right.is(Keywords.kw_function) && Left.isNot(tok::comma))
2286       return 100;
2287     if (Left.is(TT_JsTypeColon))
2288       return 35;
2289     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2290         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2291       return 100;
2292     // Prefer breaking call chains (".foo") over empty "{}", "[]" or "()".
2293     if (Left.opensScope() && Right.closesScope())
2294       return 200;
2295   }
2296 
2297   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
2298     return 1;
2299   if (Right.is(tok::l_square)) {
2300     if (Style.Language == FormatStyle::LK_Proto)
2301       return 1;
2302     if (Left.is(tok::r_square))
2303       return 200;
2304     // Slightly prefer formatting local lambda definitions like functions.
2305     if (Right.is(TT_LambdaLSquare) && Left.is(tok::equal))
2306       return 35;
2307     if (!Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2308                        TT_ArrayInitializerLSquare,
2309                        TT_DesignatedInitializerLSquare, TT_AttributeSquare))
2310       return 500;
2311   }
2312 
2313   if (Left.is(tok::coloncolon) ||
2314       (Right.is(tok::period) && Style.Language == FormatStyle::LK_Proto))
2315     return 500;
2316   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
2317       Right.is(tok::kw_operator)) {
2318     if (Line.startsWith(tok::kw_for) && Right.PartOfMultiVariableDeclStmt)
2319       return 3;
2320     if (Left.is(TT_StartOfName))
2321       return 110;
2322     if (InFunctionDecl && Right.NestingLevel == 0)
2323       return Style.PenaltyReturnTypeOnItsOwnLine;
2324     return 200;
2325   }
2326   if (Right.is(TT_PointerOrReference))
2327     return 190;
2328   if (Right.is(TT_LambdaArrow))
2329     return 110;
2330   if (Left.is(tok::equal) && Right.is(tok::l_brace))
2331     return 160;
2332   if (Left.is(TT_CastRParen))
2333     return 100;
2334   if (Left.isOneOf(tok::kw_class, tok::kw_struct))
2335     return 5000;
2336   if (Left.is(tok::comment))
2337     return 1000;
2338 
2339   if (Left.isOneOf(TT_RangeBasedForLoopColon, TT_InheritanceColon,
2340                    TT_CtorInitializerColon))
2341     return 2;
2342 
2343   if (Right.isMemberAccess()) {
2344     // Breaking before the "./->" of a chained call/member access is reasonably
2345     // cheap, as formatting those with one call per line is generally
2346     // desirable. In particular, it should be cheaper to break before the call
2347     // than it is to break inside a call's parameters, which could lead to weird
2348     // "hanging" indents. The exception is the very last "./->" to support this
2349     // frequent pattern:
2350     //
2351     //   aaaaaaaa.aaaaaaaa.bbbbbbb().ccccccccccccccccccccc(
2352     //       dddddddd);
2353     //
2354     // which might otherwise be blown up onto many lines. Here, clang-format
2355     // won't produce "hanging" indents anyway as there is no other trailing
2356     // call.
2357     //
2358     // Also apply higher penalty is not a call as that might lead to a wrapping
2359     // like:
2360     //
2361     //   aaaaaaa
2362     //       .aaaaaaaaa.bbbbbbbb(cccccccc);
2363     return !Right.NextOperator || !Right.NextOperator->Previous->closesScope()
2364                ? 150
2365                : 35;
2366   }
2367 
2368   if (Right.is(TT_TrailingAnnotation) &&
2369       (!Right.Next || Right.Next->isNot(tok::l_paren))) {
2370     // Moving trailing annotations to the next line is fine for ObjC method
2371     // declarations.
2372     if (Line.startsWith(TT_ObjCMethodSpecifier))
2373       return 10;
2374     // Generally, breaking before a trailing annotation is bad unless it is
2375     // function-like. It seems to be especially preferable to keep standard
2376     // annotations (i.e. "const", "final" and "override") on the same line.
2377     // Use a slightly higher penalty after ")" so that annotations like
2378     // "const override" are kept together.
2379     bool is_short_annotation = Right.TokenText.size() < 10;
2380     return (Left.is(tok::r_paren) ? 100 : 120) + (is_short_annotation ? 50 : 0);
2381   }
2382 
2383   // In for-loops, prefer breaking at ',' and ';'.
2384   if (Line.startsWith(tok::kw_for) && Left.is(tok::equal))
2385     return 4;
2386 
2387   // In Objective-C method expressions, prefer breaking before "param:" over
2388   // breaking after it.
2389   if (Right.is(TT_SelectorName))
2390     return 0;
2391   if (Left.is(tok::colon) && Left.is(TT_ObjCMethodExpr))
2392     return Line.MightBeFunctionDecl ? 50 : 500;
2393 
2394   // In Objective-C type declarations, avoid breaking after the category's
2395   // open paren (we'll prefer breaking after the protocol list's opening
2396   // angle bracket, if present).
2397   if (Line.Type == LT_ObjCDecl && Left.is(tok::l_paren) && Left.Previous &&
2398       Left.Previous->isOneOf(tok::identifier, tok::greater))
2399     return 500;
2400 
2401   if (Left.is(tok::l_paren) && InFunctionDecl &&
2402       Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
2403     return 100;
2404   if (Left.is(tok::l_paren) && Left.Previous &&
2405       (Left.Previous->isOneOf(tok::kw_if, tok::kw_for) ||
2406        Left.Previous->endsSequence(tok::kw_constexpr, tok::kw_if)))
2407     return 1000;
2408   if (Left.is(tok::equal) && InFunctionDecl)
2409     return 110;
2410   if (Right.is(tok::r_brace))
2411     return 1;
2412   if (Left.is(TT_TemplateOpener))
2413     return 100;
2414   if (Left.opensScope()) {
2415     if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
2416       return 0;
2417     if (Left.is(tok::l_brace) && !Style.Cpp11BracedListStyle)
2418       return 19;
2419     return Left.ParameterCount > 1 ? Style.PenaltyBreakBeforeFirstCallParameter
2420                                    : 19;
2421   }
2422   if (Left.is(TT_JavaAnnotation))
2423     return 50;
2424 
2425   if (Left.isOneOf(tok::plus, tok::comma) && Left.Previous &&
2426       Left.Previous->isLabelString() &&
2427       (Left.NextOperator || Left.OperatorIndex != 0))
2428     return 50;
2429   if (Right.is(tok::plus) && Left.isLabelString() &&
2430       (Right.NextOperator || Right.OperatorIndex != 0))
2431     return 25;
2432   if (Left.is(tok::comma))
2433     return 1;
2434   if (Right.is(tok::lessless) && Left.isLabelString() &&
2435       (Right.NextOperator || Right.OperatorIndex != 1))
2436     return 25;
2437   if (Right.is(tok::lessless)) {
2438     // Breaking at a << is really cheap.
2439     if (!Left.is(tok::r_paren) || Right.OperatorIndex > 0)
2440       // Slightly prefer to break before the first one in log-like statements.
2441       return 2;
2442     return 1;
2443   }
2444   if (Left.ClosesTemplateDeclaration)
2445     return Style.PenaltyBreakTemplateDeclaration;
2446   if (Left.is(TT_ConditionalExpr))
2447     return prec::Conditional;
2448   prec::Level Level = Left.getPrecedence();
2449   if (Level == prec::Unknown)
2450     Level = Right.getPrecedence();
2451   if (Level == prec::Assignment)
2452     return Style.PenaltyBreakAssignment;
2453   if (Level != prec::Unknown)
2454     return Level;
2455 
2456   return 3;
2457 }
2458 
2459 bool TokenAnnotator::spaceRequiredBeforeParens(const FormatToken &Right) const {
2460   return Style.SpaceBeforeParens == FormatStyle::SBPO_Always ||
2461          (Style.SpaceBeforeParens == FormatStyle::SBPO_NonEmptyParentheses &&
2462           Right.ParameterCount > 0);
2463 }
2464 
2465 bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
2466                                           const FormatToken &Left,
2467                                           const FormatToken &Right) {
2468   if (Left.is(tok::kw_return) && Right.isNot(tok::semi))
2469     return true;
2470   if (Left.is(Keywords.kw_assert) && Style.Language == FormatStyle::LK_Java)
2471     return true;
2472   if (Style.ObjCSpaceAfterProperty && Line.Type == LT_ObjCProperty &&
2473       Left.Tok.getObjCKeywordID() == tok::objc_property)
2474     return true;
2475   if (Right.is(tok::hashhash))
2476     return Left.is(tok::hash);
2477   if (Left.isOneOf(tok::hashhash, tok::hash))
2478     return Right.is(tok::hash);
2479   if (Left.is(tok::l_paren) && Right.is(tok::r_paren))
2480     return Style.SpaceInEmptyParentheses;
2481   if (Left.is(tok::l_paren) || Right.is(tok::r_paren))
2482     return (Right.is(TT_CastRParen) ||
2483             (Left.MatchingParen && Left.MatchingParen->is(TT_CastRParen)))
2484                ? Style.SpacesInCStyleCastParentheses
2485                : Style.SpacesInParentheses;
2486   if (Right.isOneOf(tok::semi, tok::comma))
2487     return false;
2488   if (Right.is(tok::less) && Line.Type == LT_ObjCDecl) {
2489     bool IsLightweightGeneric = Right.MatchingParen &&
2490                                 Right.MatchingParen->Next &&
2491                                 Right.MatchingParen->Next->is(tok::colon);
2492     return !IsLightweightGeneric && Style.ObjCSpaceBeforeProtocolList;
2493   }
2494   if (Right.is(tok::less) && Left.is(tok::kw_template))
2495     return Style.SpaceAfterTemplateKeyword;
2496   if (Left.isOneOf(tok::exclaim, tok::tilde))
2497     return false;
2498   if (Left.is(tok::at) &&
2499       Right.isOneOf(tok::identifier, tok::string_literal, tok::char_constant,
2500                     tok::numeric_constant, tok::l_paren, tok::l_brace,
2501                     tok::kw_true, tok::kw_false))
2502     return false;
2503   if (Left.is(tok::colon))
2504     return !Left.is(TT_ObjCMethodExpr);
2505   if (Left.is(tok::coloncolon))
2506     return false;
2507   if (Left.is(tok::less) || Right.isOneOf(tok::greater, tok::less)) {
2508     if (Style.Language == FormatStyle::LK_TextProto ||
2509         (Style.Language == FormatStyle::LK_Proto &&
2510          (Left.is(TT_DictLiteral) || Right.is(TT_DictLiteral)))) {
2511       // Format empty list as `<>`.
2512       if (Left.is(tok::less) && Right.is(tok::greater))
2513         return false;
2514       return !Style.Cpp11BracedListStyle;
2515     }
2516     return false;
2517   }
2518   if (Right.is(tok::ellipsis))
2519     return Left.Tok.isLiteral() || (Left.is(tok::identifier) && Left.Previous &&
2520                                     Left.Previous->is(tok::kw_case));
2521   if (Left.is(tok::l_square) && Right.is(tok::amp))
2522     return false;
2523   if (Right.is(TT_PointerOrReference)) {
2524     if (Left.is(tok::r_paren) && Line.MightBeFunctionDecl) {
2525       if (!Left.MatchingParen)
2526         return true;
2527       FormatToken *TokenBeforeMatchingParen =
2528           Left.MatchingParen->getPreviousNonComment();
2529       if (!TokenBeforeMatchingParen ||
2530           !TokenBeforeMatchingParen->isOneOf(tok::kw_typeof, tok::kw_decltype))
2531         return true;
2532     }
2533     return (Left.Tok.isLiteral() ||
2534             (!Left.isOneOf(TT_PointerOrReference, tok::l_paren) &&
2535              (Style.PointerAlignment != FormatStyle::PAS_Left ||
2536               (Line.IsMultiVariableDeclStmt &&
2537                (Left.NestingLevel == 0 ||
2538                 (Left.NestingLevel == 1 && Line.First->is(tok::kw_for)))))));
2539   }
2540   if (Right.is(TT_FunctionTypeLParen) && Left.isNot(tok::l_paren) &&
2541       (!Left.is(TT_PointerOrReference) ||
2542        (Style.PointerAlignment != FormatStyle::PAS_Right &&
2543         !Line.IsMultiVariableDeclStmt)))
2544     return true;
2545   if (Left.is(TT_PointerOrReference))
2546     return Right.Tok.isLiteral() || Right.is(TT_BlockComment) ||
2547            (Right.isOneOf(Keywords.kw_override, Keywords.kw_final) &&
2548             !Right.is(TT_StartOfName)) ||
2549            (Right.is(tok::l_brace) && Right.BlockKind == BK_Block) ||
2550            (!Right.isOneOf(TT_PointerOrReference, TT_ArraySubscriptLSquare,
2551                            tok::l_paren) &&
2552             (Style.PointerAlignment != FormatStyle::PAS_Right &&
2553              !Line.IsMultiVariableDeclStmt) &&
2554             Left.Previous &&
2555             !Left.Previous->isOneOf(tok::l_paren, tok::coloncolon));
2556   if (Right.is(tok::star) && Left.is(tok::l_paren))
2557     return false;
2558   const auto SpaceRequiredForArrayInitializerLSquare =
2559       [](const FormatToken &LSquareTok, const FormatStyle &Style) {
2560         return Style.SpacesInContainerLiterals ||
2561                ((Style.Language == FormatStyle::LK_Proto ||
2562                  Style.Language == FormatStyle::LK_TextProto) &&
2563                 !Style.Cpp11BracedListStyle &&
2564                 LSquareTok.endsSequence(tok::l_square, tok::colon,
2565                                         TT_SelectorName));
2566       };
2567   if (Left.is(tok::l_square))
2568     return (Left.is(TT_ArrayInitializerLSquare) && Right.isNot(tok::r_square) &&
2569             SpaceRequiredForArrayInitializerLSquare(Left, Style)) ||
2570            (Left.isOneOf(TT_ArraySubscriptLSquare,
2571                          TT_StructuredBindingLSquare) &&
2572             Style.SpacesInSquareBrackets && Right.isNot(tok::r_square));
2573   if (Right.is(tok::r_square))
2574     return Right.MatchingParen &&
2575            ((Right.MatchingParen->is(TT_ArrayInitializerLSquare) &&
2576              SpaceRequiredForArrayInitializerLSquare(*Right.MatchingParen,
2577                                                      Style)) ||
2578             (Style.SpacesInSquareBrackets &&
2579              Right.MatchingParen->isOneOf(TT_ArraySubscriptLSquare,
2580                                           TT_StructuredBindingLSquare)) ||
2581             Right.MatchingParen->is(TT_AttributeParen));
2582   if (Right.is(tok::l_square) &&
2583       !Right.isOneOf(TT_ObjCMethodExpr, TT_LambdaLSquare,
2584                      TT_DesignatedInitializerLSquare,
2585                      TT_StructuredBindingLSquare, TT_AttributeSquare) &&
2586       !Left.isOneOf(tok::numeric_constant, TT_DictLiteral))
2587     return false;
2588   if (Left.is(tok::l_brace) && Right.is(tok::r_brace))
2589     return !Left.Children.empty(); // No spaces in "{}".
2590   if ((Left.is(tok::l_brace) && Left.BlockKind != BK_Block) ||
2591       (Right.is(tok::r_brace) && Right.MatchingParen &&
2592        Right.MatchingParen->BlockKind != BK_Block))
2593     return !Style.Cpp11BracedListStyle;
2594   if (Left.is(TT_BlockComment))
2595     // No whitespace in x(/*foo=*/1), except for JavaScript.
2596     return Style.Language == FormatStyle::LK_JavaScript ||
2597            !Left.TokenText.endswith("=*/");
2598   if (Right.is(tok::l_paren)) {
2599     if ((Left.is(tok::r_paren) && Left.is(TT_AttributeParen)) ||
2600         (Left.is(tok::r_square) && Left.is(TT_AttributeSquare)))
2601       return true;
2602     return Line.Type == LT_ObjCDecl || Left.is(tok::semi) ||
2603            (Style.SpaceBeforeParens != FormatStyle::SBPO_Never &&
2604             (Left.isOneOf(tok::kw_if, tok::pp_elif, tok::kw_for, tok::kw_while,
2605                           tok::kw_switch, tok::kw_case, TT_ForEachMacro,
2606                           TT_ObjCForIn) ||
2607              Left.endsSequence(tok::kw_constexpr, tok::kw_if) ||
2608              (Left.isOneOf(tok::kw_try, Keywords.kw___except, tok::kw_catch,
2609                            tok::kw_new, tok::kw_delete) &&
2610               (!Left.Previous || Left.Previous->isNot(tok::period))))) ||
2611            (spaceRequiredBeforeParens(Right) &&
2612             (Left.is(tok::identifier) || Left.isFunctionLikeKeyword() ||
2613              Left.is(tok::r_paren) || Left.isSimpleTypeSpecifier() ||
2614              (Left.is(tok::r_square) && Left.MatchingParen &&
2615               Left.MatchingParen->is(TT_LambdaLSquare))) &&
2616             Line.Type != LT_PreprocessorDirective);
2617   }
2618   if (Left.is(tok::at) && Right.Tok.getObjCKeywordID() != tok::objc_not_keyword)
2619     return false;
2620   if (Right.is(TT_UnaryOperator))
2621     return !Left.isOneOf(tok::l_paren, tok::l_square, tok::at) &&
2622            (Left.isNot(tok::colon) || Left.isNot(TT_ObjCMethodExpr));
2623   if ((Left.isOneOf(tok::identifier, tok::greater, tok::r_square,
2624                     tok::r_paren) ||
2625        Left.isSimpleTypeSpecifier()) &&
2626       Right.is(tok::l_brace) && Right.getNextNonComment() &&
2627       Right.BlockKind != BK_Block)
2628     return false;
2629   if (Left.is(tok::period) || Right.is(tok::period))
2630     return false;
2631   if (Right.is(tok::hash) && Left.is(tok::identifier) && Left.TokenText == "L")
2632     return false;
2633   if (Left.is(TT_TemplateCloser) && Left.MatchingParen &&
2634       Left.MatchingParen->Previous &&
2635       (Left.MatchingParen->Previous->is(tok::period) ||
2636        Left.MatchingParen->Previous->is(tok::coloncolon)))
2637     // Java call to generic function with explicit type:
2638     // A.<B<C<...>>>DoSomething();
2639     // A::<B<C<...>>>DoSomething();  // With a Java 8 method reference.
2640     return false;
2641   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_square))
2642     return false;
2643   if (Left.is(tok::l_brace) && Left.endsSequence(TT_DictLiteral, tok::at))
2644     // Objective-C dictionary literal -> no space after opening brace.
2645     return false;
2646   if (Right.is(tok::r_brace) && Right.MatchingParen &&
2647       Right.MatchingParen->endsSequence(TT_DictLiteral, tok::at))
2648     // Objective-C dictionary literal -> no space before closing brace.
2649     return false;
2650   return true;
2651 }
2652 
2653 bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
2654                                          const FormatToken &Right) {
2655   const FormatToken &Left = *Right.Previous;
2656   if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
2657     return true; // Never ever merge two identifiers.
2658   if (Style.isCpp()) {
2659     if (Left.is(tok::kw_operator))
2660       return Right.is(tok::coloncolon);
2661     if (Right.is(tok::l_brace) && Right.BlockKind == BK_BracedInit &&
2662         !Left.opensScope() && Style.SpaceBeforeCpp11BracedList)
2663       return true;
2664   } else if (Style.Language == FormatStyle::LK_Proto ||
2665              Style.Language == FormatStyle::LK_TextProto) {
2666     if (Right.is(tok::period) &&
2667         Left.isOneOf(Keywords.kw_optional, Keywords.kw_required,
2668                      Keywords.kw_repeated, Keywords.kw_extend))
2669       return true;
2670     if (Right.is(tok::l_paren) &&
2671         Left.isOneOf(Keywords.kw_returns, Keywords.kw_option))
2672       return true;
2673     if (Right.isOneOf(tok::l_brace, tok::less) && Left.is(TT_SelectorName))
2674       return true;
2675     // Slashes occur in text protocol extension syntax: [type/type] { ... }.
2676     if (Left.is(tok::slash) || Right.is(tok::slash))
2677       return false;
2678     if (Left.MatchingParen &&
2679         Left.MatchingParen->is(TT_ProtoExtensionLSquare) &&
2680         Right.isOneOf(tok::l_brace, tok::less))
2681       return !Style.Cpp11BracedListStyle;
2682     // A percent is probably part of a formatting specification, such as %lld.
2683     if (Left.is(tok::percent))
2684       return false;
2685     // Preserve the existence of a space before a percent for cases like 0x%04x
2686     // and "%d %d"
2687     if (Left.is(tok::numeric_constant) && Right.is(tok::percent))
2688       return Right.WhitespaceRange.getEnd() != Right.WhitespaceRange.getBegin();
2689   } else if (Style.Language == FormatStyle::LK_JavaScript || Style.isCSharp()) {
2690     if (Left.is(TT_JsFatArrow))
2691       return true;
2692     // for await ( ...
2693     if (Right.is(tok::l_paren) && Left.is(Keywords.kw_await) && Left.Previous &&
2694         Left.Previous->is(tok::kw_for))
2695       return true;
2696     if (Left.is(Keywords.kw_async) && Right.is(tok::l_paren) &&
2697         Right.MatchingParen) {
2698       const FormatToken *Next = Right.MatchingParen->getNextNonComment();
2699       // An async arrow function, for example: `x = async () => foo();`,
2700       // as opposed to calling a function called async: `x = async();`
2701       if (Next && Next->is(TT_JsFatArrow))
2702         return true;
2703     }
2704     if ((Left.is(TT_TemplateString) && Left.TokenText.endswith("${")) ||
2705         (Right.is(TT_TemplateString) && Right.TokenText.startswith("}")))
2706       return false;
2707     // In tagged template literals ("html`bar baz`"), there is no space between
2708     // the tag identifier and the template string. getIdentifierInfo makes sure
2709     // that the identifier is not a pseudo keyword like `yield`, either.
2710     if (Left.is(tok::identifier) && Keywords.IsJavaScriptIdentifier(Left) &&
2711         Right.is(TT_TemplateString))
2712       return false;
2713     if (Right.is(tok::star) &&
2714         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield))
2715       return false;
2716     if (Right.isOneOf(tok::l_brace, tok::l_square) &&
2717         Left.isOneOf(Keywords.kw_function, Keywords.kw_yield,
2718                      Keywords.kw_extends, Keywords.kw_implements))
2719       return true;
2720     if (Right.is(tok::l_paren)) {
2721       // JS methods can use some keywords as names (e.g. `delete()`).
2722       if (Line.MustBeDeclaration && Left.Tok.getIdentifierInfo())
2723         return false;
2724       // Valid JS method names can include keywords, e.g. `foo.delete()` or
2725       // `bar.instanceof()`. Recognize call positions by preceding period.
2726       if (Left.Previous && Left.Previous->is(tok::period) &&
2727           Left.Tok.getIdentifierInfo())
2728         return false;
2729       // Additional unary JavaScript operators that need a space after.
2730       if (Left.isOneOf(tok::kw_throw, Keywords.kw_await, Keywords.kw_typeof,
2731                        tok::kw_void))
2732         return true;
2733     }
2734     if ((Left.isOneOf(Keywords.kw_let, Keywords.kw_var, Keywords.kw_in,
2735                       tok::kw_const) ||
2736          // "of" is only a keyword if it appears after another identifier
2737          // (e.g. as "const x of y" in a for loop), or after a destructuring
2738          // operation (const [x, y] of z, const {a, b} of c).
2739          (Left.is(Keywords.kw_of) && Left.Previous &&
2740           (Left.Previous->Tok.is(tok::identifier) ||
2741            Left.Previous->isOneOf(tok::r_square, tok::r_brace)))) &&
2742         (!Left.Previous || !Left.Previous->is(tok::period)))
2743       return true;
2744     if (Left.isOneOf(tok::kw_for, Keywords.kw_as) && Left.Previous &&
2745         Left.Previous->is(tok::period) && Right.is(tok::l_paren))
2746       return false;
2747     if (Left.is(Keywords.kw_as) &&
2748         Right.isOneOf(tok::l_square, tok::l_brace, tok::l_paren))
2749       return true;
2750     if (Left.is(tok::kw_default) && Left.Previous &&
2751         Left.Previous->is(tok::kw_export))
2752       return true;
2753     if (Left.is(Keywords.kw_is) && Right.is(tok::l_brace))
2754       return true;
2755     if (Right.isOneOf(TT_JsTypeColon, TT_JsTypeOptionalQuestion))
2756       return false;
2757     if (Left.is(TT_JsTypeOperator) || Right.is(TT_JsTypeOperator))
2758       return false;
2759     if ((Left.is(tok::l_brace) || Right.is(tok::r_brace)) &&
2760         Line.First->isOneOf(Keywords.kw_import, tok::kw_export))
2761       return false;
2762     if (Left.is(tok::ellipsis))
2763       return false;
2764     if (Left.is(TT_TemplateCloser) &&
2765         !Right.isOneOf(tok::equal, tok::l_brace, tok::comma, tok::l_square,
2766                        Keywords.kw_implements, Keywords.kw_extends))
2767       // Type assertions ('<type>expr') are not followed by whitespace. Other
2768       // locations that should have whitespace following are identified by the
2769       // above set of follower tokens.
2770       return false;
2771     if (Right.is(TT_JsNonNullAssertion))
2772       return false;
2773     if (Left.is(TT_JsNonNullAssertion) &&
2774         Right.isOneOf(Keywords.kw_as, Keywords.kw_in))
2775       return true; // "x! as string", "x! in y"
2776   } else if (Style.Language == FormatStyle::LK_Java) {
2777     if (Left.is(tok::r_square) && Right.is(tok::l_brace))
2778       return true;
2779     if (Left.is(Keywords.kw_synchronized) && Right.is(tok::l_paren))
2780       return Style.SpaceBeforeParens != FormatStyle::SBPO_Never;
2781     if ((Left.isOneOf(tok::kw_static, tok::kw_public, tok::kw_private,
2782                       tok::kw_protected) ||
2783          Left.isOneOf(Keywords.kw_final, Keywords.kw_abstract,
2784                       Keywords.kw_native)) &&
2785         Right.is(TT_TemplateOpener))
2786       return true;
2787   }
2788   if (Left.is(TT_ImplicitStringLiteral))
2789     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2790   if (Line.Type == LT_ObjCMethodDecl) {
2791     if (Left.is(TT_ObjCMethodSpecifier))
2792       return true;
2793     if (Left.is(tok::r_paren) && canBeObjCSelectorComponent(Right))
2794       // Don't space between ')' and <id> or ')' and 'new'. 'new' is not a
2795       // keyword in Objective-C, and '+ (instancetype)new;' is a standard class
2796       // method declaration.
2797       return false;
2798   }
2799   if (Line.Type == LT_ObjCProperty &&
2800       (Right.is(tok::equal) || Left.is(tok::equal)))
2801     return false;
2802 
2803   if (Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow) ||
2804       Left.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow))
2805     return true;
2806   if (Right.is(TT_OverloadedOperatorLParen))
2807     return spaceRequiredBeforeParens(Right);
2808   if (Left.is(tok::comma))
2809     return true;
2810   if (Right.is(tok::comma))
2811     return false;
2812   if (Right.is(TT_ObjCBlockLParen))
2813     return true;
2814   if (Right.is(TT_CtorInitializerColon))
2815     return Style.SpaceBeforeCtorInitializerColon;
2816   if (Right.is(TT_InheritanceColon) && !Style.SpaceBeforeInheritanceColon)
2817     return false;
2818   if (Right.is(TT_RangeBasedForLoopColon) &&
2819       !Style.SpaceBeforeRangeBasedForLoopColon)
2820     return false;
2821   if (Right.is(tok::colon)) {
2822     if (Line.First->isOneOf(tok::kw_case, tok::kw_default) ||
2823         !Right.getNextNonComment() || Right.getNextNonComment()->is(tok::semi))
2824       return false;
2825     if (Right.is(TT_ObjCMethodExpr))
2826       return false;
2827     if (Left.is(tok::question))
2828       return false;
2829     if (Right.is(TT_InlineASMColon) && Left.is(tok::coloncolon))
2830       return false;
2831     if (Right.is(TT_DictLiteral))
2832       return Style.SpacesInContainerLiterals;
2833     if (Right.is(TT_AttributeColon))
2834       return false;
2835     return true;
2836   }
2837   if (Left.is(TT_UnaryOperator))
2838     return (Style.SpaceAfterLogicalNot && Left.is(tok::exclaim)) ||
2839            Right.is(TT_BinaryOperator);
2840 
2841   // If the next token is a binary operator or a selector name, we have
2842   // incorrectly classified the parenthesis as a cast. FIXME: Detect correctly.
2843   if (Left.is(TT_CastRParen))
2844     return Style.SpaceAfterCStyleCast ||
2845            Right.isOneOf(TT_BinaryOperator, TT_SelectorName);
2846 
2847   if (Left.is(tok::greater) && Right.is(tok::greater)) {
2848     if (Style.Language == FormatStyle::LK_TextProto ||
2849         (Style.Language == FormatStyle::LK_Proto && Left.is(TT_DictLiteral)))
2850       return !Style.Cpp11BracedListStyle;
2851     return Right.is(TT_TemplateCloser) && Left.is(TT_TemplateCloser) &&
2852            (Style.Standard != FormatStyle::LS_Cpp11 || Style.SpacesInAngles);
2853   }
2854   if (Right.isOneOf(tok::arrow, tok::arrowstar, tok::periodstar) ||
2855       Left.isOneOf(tok::arrow, tok::period, tok::arrowstar, tok::periodstar) ||
2856       (Right.is(tok::period) && Right.isNot(TT_DesignatedInitializerPeriod)))
2857     return false;
2858   if (!Style.SpaceBeforeAssignmentOperators &&
2859       Right.getPrecedence() == prec::Assignment)
2860     return false;
2861   if (Style.Language == FormatStyle::LK_Java && Right.is(tok::coloncolon) &&
2862       (Left.is(tok::identifier) || Left.is(tok::kw_this)))
2863     return false;
2864   if (Right.is(tok::coloncolon) && Left.is(tok::identifier))
2865     // Generally don't remove existing spaces between an identifier and "::".
2866     // The identifier might actually be a macro name such as ALWAYS_INLINE. If
2867     // this turns out to be too lenient, add analysis of the identifier itself.
2868     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
2869   if (Right.is(tok::coloncolon) && !Left.isOneOf(tok::l_brace, tok::comment))
2870     return (Left.is(TT_TemplateOpener) &&
2871             Style.Standard == FormatStyle::LS_Cpp03) ||
2872            !(Left.isOneOf(tok::l_paren, tok::r_paren, tok::l_square,
2873                           tok::kw___super, TT_TemplateCloser,
2874                           TT_TemplateOpener)) ||
2875            (Left.is(tok ::l_paren) && Style.SpacesInParentheses);
2876   if ((Left.is(TT_TemplateOpener)) != (Right.is(TT_TemplateCloser)))
2877     return Style.SpacesInAngles;
2878   // Space before TT_StructuredBindingLSquare.
2879   if (Right.is(TT_StructuredBindingLSquare))
2880     return !Left.isOneOf(tok::amp, tok::ampamp) ||
2881            Style.PointerAlignment != FormatStyle::PAS_Right;
2882   // Space before & or && following a TT_StructuredBindingLSquare.
2883   if (Right.Next && Right.Next->is(TT_StructuredBindingLSquare) &&
2884       Right.isOneOf(tok::amp, tok::ampamp))
2885     return Style.PointerAlignment != FormatStyle::PAS_Left;
2886   if ((Right.is(TT_BinaryOperator) && !Left.is(tok::l_paren)) ||
2887       (Left.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
2888        !Right.is(tok::r_paren)))
2889     return true;
2890   if (Left.is(TT_TemplateCloser) && Right.is(tok::l_paren) &&
2891       Right.isNot(TT_FunctionTypeLParen))
2892     return spaceRequiredBeforeParens(Right);
2893   if (Right.is(TT_TemplateOpener) && Left.is(tok::r_paren) &&
2894       Left.MatchingParen && Left.MatchingParen->is(TT_OverloadedOperatorLParen))
2895     return false;
2896   if (Right.is(tok::less) && Left.isNot(tok::l_paren) &&
2897       Line.startsWith(tok::hash))
2898     return true;
2899   if (Right.is(TT_TrailingUnaryOperator))
2900     return false;
2901   if (Left.is(TT_RegexLiteral))
2902     return false;
2903   return spaceRequiredBetween(Line, Left, Right);
2904 }
2905 
2906 // Returns 'true' if 'Tok' is a brace we'd want to break before in Allman style.
2907 static bool isAllmanBrace(const FormatToken &Tok) {
2908   return Tok.is(tok::l_brace) && Tok.BlockKind == BK_Block &&
2909          !Tok.isOneOf(TT_ObjCBlockLBrace, TT_LambdaLBrace, TT_DictLiteral);
2910 }
2911 
2912 bool TokenAnnotator::mustBreakBefore(const AnnotatedLine &Line,
2913                                      const FormatToken &Right) {
2914   const FormatToken &Left = *Right.Previous;
2915   if (Right.NewlinesBefore > 1 && Style.MaxEmptyLinesToKeep > 0)
2916     return true;
2917 
2918   if (Style.Language == FormatStyle::LK_JavaScript) {
2919     // FIXME: This might apply to other languages and token kinds.
2920     if (Right.is(tok::string_literal) && Left.is(tok::plus) && Left.Previous &&
2921         Left.Previous->is(tok::string_literal))
2922       return true;
2923     if (Left.is(TT_DictLiteral) && Left.is(tok::l_brace) && Line.Level == 0 &&
2924         Left.Previous && Left.Previous->is(tok::equal) &&
2925         Line.First->isOneOf(tok::identifier, Keywords.kw_import, tok::kw_export,
2926                             tok::kw_const) &&
2927         // kw_var/kw_let are pseudo-tokens that are tok::identifier, so match
2928         // above.
2929         !Line.First->isOneOf(Keywords.kw_var, Keywords.kw_let))
2930       // Object literals on the top level of a file are treated as "enum-style".
2931       // Each key/value pair is put on a separate line, instead of bin-packing.
2932       return true;
2933     if (Left.is(tok::l_brace) && Line.Level == 0 &&
2934         (Line.startsWith(tok::kw_enum) ||
2935          Line.startsWith(tok::kw_const, tok::kw_enum) ||
2936          Line.startsWith(tok::kw_export, tok::kw_enum) ||
2937          Line.startsWith(tok::kw_export, tok::kw_const, tok::kw_enum)))
2938       // JavaScript top-level enum key/value pairs are put on separate lines
2939       // instead of bin-packing.
2940       return true;
2941     if (Right.is(tok::r_brace) && Left.is(tok::l_brace) &&
2942         !Left.Children.empty())
2943       // Support AllowShortFunctionsOnASingleLine for JavaScript.
2944       return Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_None ||
2945              Style.AllowShortFunctionsOnASingleLine == FormatStyle::SFS_Empty ||
2946              (Left.NestingLevel == 0 && Line.Level == 0 &&
2947               Style.AllowShortFunctionsOnASingleLine &
2948                   FormatStyle::SFS_InlineOnly);
2949   } else if (Style.Language == FormatStyle::LK_Java) {
2950     if (Right.is(tok::plus) && Left.is(tok::string_literal) && Right.Next &&
2951         Right.Next->is(tok::string_literal))
2952       return true;
2953   } else if (Style.Language == FormatStyle::LK_Cpp ||
2954              Style.Language == FormatStyle::LK_ObjC ||
2955              Style.Language == FormatStyle::LK_Proto ||
2956              Style.Language == FormatStyle::LK_TableGen ||
2957              Style.Language == FormatStyle::LK_TextProto) {
2958     if (Left.isStringLiteral() && Right.isStringLiteral())
2959       return true;
2960   }
2961 
2962   // If the last token before a '}', ']', or ')' is a comma or a trailing
2963   // comment, the intention is to insert a line break after it in order to make
2964   // shuffling around entries easier. Import statements, especially in
2965   // JavaScript, can be an exception to this rule.
2966   if (Style.JavaScriptWrapImports || Line.Type != LT_ImportStatement) {
2967     const FormatToken *BeforeClosingBrace = nullptr;
2968     if ((Left.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
2969          (Style.Language == FormatStyle::LK_JavaScript &&
2970           Left.is(tok::l_paren))) &&
2971         Left.BlockKind != BK_Block && Left.MatchingParen)
2972       BeforeClosingBrace = Left.MatchingParen->Previous;
2973     else if (Right.MatchingParen &&
2974              (Right.MatchingParen->isOneOf(tok::l_brace,
2975                                            TT_ArrayInitializerLSquare) ||
2976               (Style.Language == FormatStyle::LK_JavaScript &&
2977                Right.MatchingParen->is(tok::l_paren))))
2978       BeforeClosingBrace = &Left;
2979     if (BeforeClosingBrace && (BeforeClosingBrace->is(tok::comma) ||
2980                                BeforeClosingBrace->isTrailingComment()))
2981       return true;
2982   }
2983 
2984   if (Right.is(tok::comment))
2985     return Left.BlockKind != BK_BracedInit &&
2986            Left.isNot(TT_CtorInitializerColon) &&
2987            (Right.NewlinesBefore > 0 && Right.HasUnescapedNewline);
2988   if (Left.isTrailingComment())
2989     return true;
2990   if (Right.Previous->IsUnterminatedLiteral)
2991     return true;
2992   if (Right.is(tok::lessless) && Right.Next &&
2993       Right.Previous->is(tok::string_literal) &&
2994       Right.Next->is(tok::string_literal))
2995     return true;
2996   if (Right.Previous->ClosesTemplateDeclaration &&
2997       Right.Previous->MatchingParen &&
2998       Right.Previous->MatchingParen->NestingLevel == 0 &&
2999       Style.AlwaysBreakTemplateDeclarations == FormatStyle::BTDS_Yes)
3000     return true;
3001   if (Right.is(TT_CtorInitializerComma) &&
3002       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
3003       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
3004     return true;
3005   if (Right.is(TT_CtorInitializerColon) &&
3006       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma &&
3007       !Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
3008     return true;
3009   // Break only if we have multiple inheritance.
3010   if (Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma &&
3011       Right.is(TT_InheritanceComma))
3012     return true;
3013   if (Right.is(tok::string_literal) && Right.TokenText.startswith("R\""))
3014     // Multiline raw string literals are special wrt. line breaks. The author
3015     // has made a deliberate choice and might have aligned the contents of the
3016     // string literal accordingly. Thus, we try keep existing line breaks.
3017     return Right.IsMultiline && Right.NewlinesBefore > 0;
3018   if ((Right.Previous->is(tok::l_brace) ||
3019        (Right.Previous->is(tok::less) && Right.Previous->Previous &&
3020         Right.Previous->Previous->is(tok::equal))) &&
3021       Right.NestingLevel == 1 && Style.Language == FormatStyle::LK_Proto) {
3022     // Don't put enums or option definitions onto single lines in protocol
3023     // buffers.
3024     return true;
3025   }
3026   if (Right.is(TT_InlineASMBrace))
3027     return Right.HasUnescapedNewline;
3028   if (isAllmanBrace(Left) || isAllmanBrace(Right))
3029     return (Line.startsWith(tok::kw_enum) && Style.BraceWrapping.AfterEnum) ||
3030            (Line.startsWith(tok::kw_typedef, tok::kw_enum) &&
3031             Style.BraceWrapping.AfterEnum) ||
3032            (Line.startsWith(tok::kw_class) && Style.BraceWrapping.AfterClass) ||
3033            (Line.startsWith(tok::kw_struct) && Style.BraceWrapping.AfterStruct);
3034   if (Left.is(TT_ObjCBlockLBrace) && !Style.AllowShortBlocksOnASingleLine)
3035     return true;
3036 
3037   if (Left.is(TT_LambdaLBrace)) {
3038     if (Left.MatchingParen && Left.MatchingParen->Next &&
3039         Left.MatchingParen->Next->isOneOf(tok::comma, tok::r_paren) &&
3040         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline)
3041       return false;
3042 
3043     if (Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_None ||
3044         Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Inline ||
3045         (!Left.Children.empty() &&
3046          Style.AllowShortLambdasOnASingleLine == FormatStyle::SLS_Empty))
3047       return true;
3048   }
3049 
3050   // Put multiple C# attributes on a new line.
3051   if (Style.isCSharp() &&
3052       ((Left.is(TT_AttributeSquare) && Left.is(tok::r_square)) ||
3053        (Left.is(tok::r_square) && Right.is(TT_AttributeSquare) &&
3054         Right.is(tok::l_square))))
3055     return true;
3056 
3057   // Put multiple Java annotation on a new line.
3058   if ((Style.Language == FormatStyle::LK_Java ||
3059        Style.Language == FormatStyle::LK_JavaScript) &&
3060       Left.is(TT_LeadingJavaAnnotation) &&
3061       Right.isNot(TT_LeadingJavaAnnotation) && Right.isNot(tok::l_paren) &&
3062       (Line.Last->is(tok::l_brace) || Style.BreakAfterJavaFieldAnnotations))
3063     return true;
3064 
3065   if (Right.is(TT_ProtoExtensionLSquare))
3066     return true;
3067 
3068   // In text proto instances if a submessage contains at least 2 entries and at
3069   // least one of them is a submessage, like A { ... B { ... } ... },
3070   // put all of the entries of A on separate lines by forcing the selector of
3071   // the submessage B to be put on a newline.
3072   //
3073   // Example: these can stay on one line:
3074   // a { scalar_1: 1 scalar_2: 2 }
3075   // a { b { key: value } }
3076   //
3077   // and these entries need to be on a new line even if putting them all in one
3078   // line is under the column limit:
3079   // a {
3080   //   scalar: 1
3081   //   b { key: value }
3082   // }
3083   //
3084   // We enforce this by breaking before a submessage field that has previous
3085   // siblings, *and* breaking before a field that follows a submessage field.
3086   //
3087   // Be careful to exclude the case  [proto.ext] { ... } since the `]` is
3088   // the TT_SelectorName there, but we don't want to break inside the brackets.
3089   //
3090   // Another edge case is @submessage { key: value }, which is a common
3091   // substitution placeholder. In this case we want to keep `@` and `submessage`
3092   // together.
3093   //
3094   // We ensure elsewhere that extensions are always on their own line.
3095   if ((Style.Language == FormatStyle::LK_Proto ||
3096        Style.Language == FormatStyle::LK_TextProto) &&
3097       Right.is(TT_SelectorName) && !Right.is(tok::r_square) && Right.Next) {
3098     // Keep `@submessage` together in:
3099     // @submessage { key: value }
3100     if (Right.Previous && Right.Previous->is(tok::at))
3101       return false;
3102     // Look for the scope opener after selector in cases like:
3103     // selector { ...
3104     // selector: { ...
3105     // selector: @base { ...
3106     FormatToken *LBrace = Right.Next;
3107     if (LBrace && LBrace->is(tok::colon)) {
3108       LBrace = LBrace->Next;
3109       if (LBrace && LBrace->is(tok::at)) {
3110         LBrace = LBrace->Next;
3111         if (LBrace)
3112           LBrace = LBrace->Next;
3113       }
3114     }
3115     if (LBrace &&
3116         // The scope opener is one of {, [, <:
3117         // selector { ... }
3118         // selector [ ... ]
3119         // selector < ... >
3120         //
3121         // In case of selector { ... }, the l_brace is TT_DictLiteral.
3122         // In case of an empty selector {}, the l_brace is not TT_DictLiteral,
3123         // so we check for immediately following r_brace.
3124         ((LBrace->is(tok::l_brace) &&
3125           (LBrace->is(TT_DictLiteral) ||
3126            (LBrace->Next && LBrace->Next->is(tok::r_brace)))) ||
3127          LBrace->is(TT_ArrayInitializerLSquare) || LBrace->is(tok::less))) {
3128       // If Left.ParameterCount is 0, then this submessage entry is not the
3129       // first in its parent submessage, and we want to break before this entry.
3130       // If Left.ParameterCount is greater than 0, then its parent submessage
3131       // might contain 1 or more entries and we want to break before this entry
3132       // if it contains at least 2 entries. We deal with this case later by
3133       // detecting and breaking before the next entry in the parent submessage.
3134       if (Left.ParameterCount == 0)
3135         return true;
3136       // However, if this submessage is the first entry in its parent
3137       // submessage, Left.ParameterCount might be 1 in some cases.
3138       // We deal with this case later by detecting an entry
3139       // following a closing paren of this submessage.
3140     }
3141 
3142     // If this is an entry immediately following a submessage, it will be
3143     // preceded by a closing paren of that submessage, like in:
3144     //     left---.  .---right
3145     //            v  v
3146     // sub: { ... } key: value
3147     // If there was a comment between `}` an `key` above, then `key` would be
3148     // put on a new line anyways.
3149     if (Left.isOneOf(tok::r_brace, tok::greater, tok::r_square))
3150       return true;
3151   }
3152 
3153   // Deal with lambda arguments in C++ - we want consistent line breaks whether
3154   // they happen to be at arg0, arg1 or argN. The selection is a bit nuanced
3155   // as aggressive line breaks are placed when the lambda is not the last arg.
3156   if ((Style.Language == FormatStyle::LK_Cpp ||
3157        Style.Language == FormatStyle::LK_ObjC) &&
3158       Left.is(tok::l_paren) && Left.BlockParameterCount > 0 &&
3159       !Right.isOneOf(tok::l_paren, TT_LambdaLSquare)) {
3160     // Multiple lambdas in the same function call force line breaks.
3161     if (Left.BlockParameterCount > 1)
3162       return true;
3163 
3164     // A lambda followed by another arg forces a line break.
3165     if (!Left.Role)
3166       return false;
3167     auto Comma = Left.Role->lastComma();
3168     if (!Comma)
3169       return false;
3170     auto Next = Comma->getNextNonComment();
3171     if (!Next)
3172       return false;
3173     if (!Next->isOneOf(TT_LambdaLSquare, tok::l_brace, tok::caret))
3174       return true;
3175   }
3176 
3177   return false;
3178 }
3179 
3180 bool TokenAnnotator::canBreakBefore(const AnnotatedLine &Line,
3181                                     const FormatToken &Right) {
3182   const FormatToken &Left = *Right.Previous;
3183 
3184   // Language-specific stuff.
3185   if (Style.Language == FormatStyle::LK_Java) {
3186     if (Left.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3187                      Keywords.kw_implements))
3188       return false;
3189     if (Right.isOneOf(Keywords.kw_throws, Keywords.kw_extends,
3190                       Keywords.kw_implements))
3191       return true;
3192   } else if (Style.Language == FormatStyle::LK_JavaScript) {
3193     const FormatToken *NonComment = Right.getPreviousNonComment();
3194     if (NonComment &&
3195         NonComment->isOneOf(
3196             tok::kw_return, Keywords.kw_yield, tok::kw_continue, tok::kw_break,
3197             tok::kw_throw, Keywords.kw_interface, Keywords.kw_type,
3198             tok::kw_static, tok::kw_public, tok::kw_private, tok::kw_protected,
3199             Keywords.kw_readonly, Keywords.kw_abstract, Keywords.kw_get,
3200             Keywords.kw_set, Keywords.kw_async, Keywords.kw_await))
3201       return false; // Otherwise automatic semicolon insertion would trigger.
3202     if (Right.NestingLevel == 0 &&
3203         (Left.Tok.getIdentifierInfo() ||
3204          Left.isOneOf(tok::r_square, tok::r_paren)) &&
3205         Right.isOneOf(tok::l_square, tok::l_paren))
3206       return false; // Otherwise automatic semicolon insertion would trigger.
3207     if (Left.is(TT_JsFatArrow) && Right.is(tok::l_brace))
3208       return false;
3209     if (Left.is(TT_JsTypeColon))
3210       return true;
3211     // Don't wrap between ":" and "!" of a strict prop init ("field!: type;").
3212     if (Left.is(tok::exclaim) && Right.is(tok::colon))
3213       return false;
3214     // Look for is type annotations like:
3215     // function f(): a is B { ... }
3216     // Do not break before is in these cases.
3217     if (Right.is(Keywords.kw_is)) {
3218       const FormatToken *Next = Right.getNextNonComment();
3219       // If `is` is followed by a colon, it's likely that it's a dict key, so
3220       // ignore it for this check.
3221       // For example this is common in Polymer:
3222       // Polymer({
3223       //   is: 'name',
3224       //   ...
3225       // });
3226       if (!Next || !Next->is(tok::colon))
3227         return false;
3228     }
3229     if (Left.is(Keywords.kw_in))
3230       return Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None;
3231     if (Right.is(Keywords.kw_in))
3232       return Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None;
3233     if (Right.is(Keywords.kw_as))
3234       return false; // must not break before as in 'x as type' casts
3235     if (Right.isOneOf(Keywords.kw_extends, Keywords.kw_infer)) {
3236       // extends and infer can appear as keywords in conditional types:
3237       //   https://www.typescriptlang.org/docs/handbook/release-notes/typescript-2-8.html#conditional-types
3238       // do not break before them, as the expressions are subject to ASI.
3239       return false;
3240     }
3241     if (Left.is(Keywords.kw_as))
3242       return true;
3243     if (Left.is(TT_JsNonNullAssertion))
3244       return true;
3245     if (Left.is(Keywords.kw_declare) &&
3246         Right.isOneOf(Keywords.kw_module, tok::kw_namespace,
3247                       Keywords.kw_function, tok::kw_class, tok::kw_enum,
3248                       Keywords.kw_interface, Keywords.kw_type, Keywords.kw_var,
3249                       Keywords.kw_let, tok::kw_const))
3250       // See grammar for 'declare' statements at:
3251       // https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#A.10
3252       return false;
3253     if (Left.isOneOf(Keywords.kw_module, tok::kw_namespace) &&
3254         Right.isOneOf(tok::identifier, tok::string_literal))
3255       return false; // must not break in "module foo { ...}"
3256     if (Right.is(TT_TemplateString) && Right.closesScope())
3257       return false;
3258     // Don't split tagged template literal so there is a break between the tag
3259     // identifier and template string.
3260     if (Left.is(tok::identifier) && Right.is(TT_TemplateString)) {
3261       return false;
3262     }
3263     if (Left.is(TT_TemplateString) && Left.opensScope())
3264       return true;
3265   }
3266 
3267   if (Left.is(tok::at))
3268     return false;
3269   if (Left.Tok.getObjCKeywordID() == tok::objc_interface)
3270     return false;
3271   if (Left.isOneOf(TT_JavaAnnotation, TT_LeadingJavaAnnotation))
3272     return !Right.is(tok::l_paren);
3273   if (Right.is(TT_PointerOrReference))
3274     return Line.IsMultiVariableDeclStmt ||
3275            (Style.PointerAlignment == FormatStyle::PAS_Right &&
3276             (!Right.Next || Right.Next->isNot(TT_FunctionDeclarationName)));
3277   if (Right.isOneOf(TT_StartOfName, TT_FunctionDeclarationName) ||
3278       Right.is(tok::kw_operator))
3279     return true;
3280   if (Left.is(TT_PointerOrReference))
3281     return false;
3282   if (Right.isTrailingComment())
3283     // We rely on MustBreakBefore being set correctly here as we should not
3284     // change the "binding" behavior of a comment.
3285     // The first comment in a braced lists is always interpreted as belonging to
3286     // the first list element. Otherwise, it should be placed outside of the
3287     // list.
3288     return Left.BlockKind == BK_BracedInit ||
3289            (Left.is(TT_CtorInitializerColon) &&
3290             Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon);
3291   if (Left.is(tok::question) && Right.is(tok::colon))
3292     return false;
3293   if (Right.is(TT_ConditionalExpr) || Right.is(tok::question))
3294     return Style.BreakBeforeTernaryOperators;
3295   if (Left.is(TT_ConditionalExpr) || Left.is(tok::question))
3296     return !Style.BreakBeforeTernaryOperators;
3297   if (Left.is(TT_InheritanceColon))
3298     return Style.BreakInheritanceList == FormatStyle::BILS_AfterColon;
3299   if (Right.is(TT_InheritanceColon))
3300     return Style.BreakInheritanceList != FormatStyle::BILS_AfterColon;
3301   if (Right.is(TT_ObjCMethodExpr) && !Right.is(tok::r_square) &&
3302       Left.isNot(TT_SelectorName))
3303     return true;
3304 
3305   if (Right.is(tok::colon) &&
3306       !Right.isOneOf(TT_CtorInitializerColon, TT_InlineASMColon))
3307     return false;
3308   if (Left.is(tok::colon) && Left.isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
3309     if (Style.Language == FormatStyle::LK_Proto ||
3310         Style.Language == FormatStyle::LK_TextProto) {
3311       if (!Style.AlwaysBreakBeforeMultilineStrings && Right.isStringLiteral())
3312         return false;
3313       // Prevent cases like:
3314       //
3315       // submessage:
3316       //     { key: valueeeeeeeeeeee }
3317       //
3318       // when the snippet does not fit into one line.
3319       // Prefer:
3320       //
3321       // submessage: {
3322       //   key: valueeeeeeeeeeee
3323       // }
3324       //
3325       // instead, even if it is longer by one line.
3326       //
3327       // Note that this allows allows the "{" to go over the column limit
3328       // when the column limit is just between ":" and "{", but that does
3329       // not happen too often and alternative formattings in this case are
3330       // not much better.
3331       //
3332       // The code covers the cases:
3333       //
3334       // submessage: { ... }
3335       // submessage: < ... >
3336       // repeated: [ ... ]
3337       if (((Right.is(tok::l_brace) || Right.is(tok::less)) &&
3338            Right.is(TT_DictLiteral)) ||
3339           Right.is(TT_ArrayInitializerLSquare))
3340         return false;
3341     }
3342     return true;
3343   }
3344   if (Right.is(tok::r_square) && Right.MatchingParen &&
3345       Right.MatchingParen->is(TT_ProtoExtensionLSquare))
3346     return false;
3347   if (Right.is(TT_SelectorName) || (Right.is(tok::identifier) && Right.Next &&
3348                                     Right.Next->is(TT_ObjCMethodExpr)))
3349     return Left.isNot(tok::period); // FIXME: Properly parse ObjC calls.
3350   if (Left.is(tok::r_paren) && Line.Type == LT_ObjCProperty)
3351     return true;
3352   if (Left.ClosesTemplateDeclaration || Left.is(TT_FunctionAnnotationRParen))
3353     return true;
3354   if (Right.isOneOf(TT_RangeBasedForLoopColon, TT_OverloadedOperatorLParen,
3355                     TT_OverloadedOperator))
3356     return false;
3357   if (Left.is(TT_RangeBasedForLoopColon))
3358     return true;
3359   if (Right.is(TT_RangeBasedForLoopColon))
3360     return false;
3361   if (Left.is(TT_TemplateCloser) && Right.is(TT_TemplateOpener))
3362     return true;
3363   if (Left.isOneOf(TT_TemplateCloser, TT_UnaryOperator) ||
3364       Left.is(tok::kw_operator))
3365     return false;
3366   if (Left.is(tok::equal) && !Right.isOneOf(tok::kw_default, tok::kw_delete) &&
3367       Line.Type == LT_VirtualFunctionDecl && Left.NestingLevel == 0)
3368     return false;
3369   if (Left.is(tok::equal) && Right.is(tok::l_brace) &&
3370       !Style.Cpp11BracedListStyle)
3371     return false;
3372   if (Left.is(tok::l_paren) && Left.is(TT_AttributeParen))
3373     return false;
3374   if (Left.is(tok::l_paren) && Left.Previous &&
3375       (Left.Previous->isOneOf(TT_BinaryOperator, TT_CastRParen)))
3376     return false;
3377   if (Right.is(TT_ImplicitStringLiteral))
3378     return false;
3379 
3380   if (Right.is(tok::r_paren) || Right.is(TT_TemplateCloser))
3381     return false;
3382   if (Right.is(tok::r_square) && Right.MatchingParen &&
3383       Right.MatchingParen->is(TT_LambdaLSquare))
3384     return false;
3385 
3386   // We only break before r_brace if there was a corresponding break before
3387   // the l_brace, which is tracked by BreakBeforeClosingBrace.
3388   if (Right.is(tok::r_brace))
3389     return Right.MatchingParen && Right.MatchingParen->BlockKind == BK_Block;
3390 
3391   // Allow breaking after a trailing annotation, e.g. after a method
3392   // declaration.
3393   if (Left.is(TT_TrailingAnnotation))
3394     return !Right.isOneOf(tok::l_brace, tok::semi, tok::equal, tok::l_paren,
3395                           tok::less, tok::coloncolon);
3396 
3397   if (Right.is(tok::kw___attribute) ||
3398       (Right.is(tok::l_square) && Right.is(TT_AttributeSquare)))
3399     return true;
3400 
3401   if (Left.is(tok::identifier) && Right.is(tok::string_literal))
3402     return true;
3403 
3404   if (Right.is(tok::identifier) && Right.Next && Right.Next->is(TT_DictLiteral))
3405     return true;
3406 
3407   if (Left.is(TT_CtorInitializerColon))
3408     return Style.BreakConstructorInitializers == FormatStyle::BCIS_AfterColon;
3409   if (Right.is(TT_CtorInitializerColon))
3410     return Style.BreakConstructorInitializers != FormatStyle::BCIS_AfterColon;
3411   if (Left.is(TT_CtorInitializerComma) &&
3412       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
3413     return false;
3414   if (Right.is(TT_CtorInitializerComma) &&
3415       Style.BreakConstructorInitializers == FormatStyle::BCIS_BeforeComma)
3416     return true;
3417   if (Left.is(TT_InheritanceComma) &&
3418       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
3419     return false;
3420   if (Right.is(TT_InheritanceComma) &&
3421       Style.BreakInheritanceList == FormatStyle::BILS_BeforeComma)
3422     return true;
3423   if ((Left.is(tok::greater) && Right.is(tok::greater)) ||
3424       (Left.is(tok::less) && Right.is(tok::less)))
3425     return false;
3426   if (Right.is(TT_BinaryOperator) &&
3427       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None &&
3428       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_All ||
3429        Right.getPrecedence() != prec::Assignment))
3430     return true;
3431   if (Left.is(TT_ArrayInitializerLSquare))
3432     return true;
3433   if (Right.is(tok::kw_typename) && Left.isNot(tok::kw_const))
3434     return true;
3435   if ((Left.isBinaryOperator() || Left.is(TT_BinaryOperator)) &&
3436       !Left.isOneOf(tok::arrowstar, tok::lessless) &&
3437       Style.BreakBeforeBinaryOperators != FormatStyle::BOS_All &&
3438       (Style.BreakBeforeBinaryOperators == FormatStyle::BOS_None ||
3439        Left.getPrecedence() == prec::Assignment))
3440     return true;
3441   if ((Left.is(TT_AttributeSquare) && Right.is(tok::l_square)) ||
3442       (Left.is(tok::r_square) && Right.is(TT_AttributeSquare)))
3443     return false;
3444   return Left.isOneOf(tok::comma, tok::coloncolon, tok::semi, tok::l_brace,
3445                       tok::kw_class, tok::kw_struct, tok::comment) ||
3446          Right.isMemberAccess() ||
3447          Right.isOneOf(TT_TrailingReturnArrow, TT_LambdaArrow, tok::lessless,
3448                        tok::colon, tok::l_square, tok::at) ||
3449          (Left.is(tok::r_paren) &&
3450           Right.isOneOf(tok::identifier, tok::kw_const)) ||
3451          (Left.is(tok::l_paren) && !Right.is(tok::r_paren)) ||
3452          (Left.is(TT_TemplateOpener) && !Right.is(TT_TemplateCloser));
3453 }
3454 
3455 void TokenAnnotator::printDebugInfo(const AnnotatedLine &Line) {
3456   llvm::errs() << "AnnotatedTokens(L=" << Line.Level << "):\n";
3457   const FormatToken *Tok = Line.First;
3458   while (Tok) {
3459     llvm::errs() << " M=" << Tok->MustBreakBefore
3460                  << " C=" << Tok->CanBreakBefore
3461                  << " T=" << getTokenTypeName(Tok->Type)
3462                  << " S=" << Tok->SpacesRequiredBefore
3463                  << " B=" << Tok->BlockParameterCount
3464                  << " BK=" << Tok->BlockKind << " P=" << Tok->SplitPenalty
3465                  << " Name=" << Tok->Tok.getName() << " L=" << Tok->TotalLength
3466                  << " PPK=" << Tok->PackingKind << " FakeLParens=";
3467     for (unsigned i = 0, e = Tok->FakeLParens.size(); i != e; ++i)
3468       llvm::errs() << Tok->FakeLParens[i] << "/";
3469     llvm::errs() << " FakeRParens=" << Tok->FakeRParens;
3470     llvm::errs() << " II=" << Tok->Tok.getIdentifierInfo();
3471     llvm::errs() << " Text='" << Tok->TokenText << "'\n";
3472     if (!Tok->Next)
3473       assert(Tok == Line.Last);
3474     Tok = Tok->Next;
3475   }
3476   llvm::errs() << "----\n";
3477 }
3478 
3479 } // namespace format
3480 } // namespace clang
3481