1 //===--- ContinuationIndenter.cpp - Format C++ code -----------------------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 ///
10 /// \file
11 /// \brief This file implements the continuation indenter.
12 ///
13 //===----------------------------------------------------------------------===//
14 
15 #include "BreakableToken.h"
16 #include "ContinuationIndenter.h"
17 #include "WhitespaceManager.h"
18 #include "clang/Basic/OperatorPrecedence.h"
19 #include "clang/Basic/SourceManager.h"
20 #include "clang/Format/Format.h"
21 #include "llvm/Support/Debug.h"
22 
23 #define DEBUG_TYPE "format-formatter"
24 
25 namespace clang {
26 namespace format {
27 
28 // Returns the length of everything up to the first possible line break after
29 // the ), ], } or > matching \c Tok.
30 static unsigned getLengthToMatchingParen(const FormatToken &Tok) {
31   if (!Tok.MatchingParen)
32     return 0;
33   FormatToken *End = Tok.MatchingParen;
34   while (End->Next && !End->Next->CanBreakBefore) {
35     End = End->Next;
36   }
37   return End->TotalLength - Tok.TotalLength + 1;
38 }
39 
40 static unsigned getLengthToNextOperator(const FormatToken &Tok) {
41   if (!Tok.NextOperator)
42     return 0;
43   return Tok.NextOperator->TotalLength - Tok.TotalLength;
44 }
45 
46 // Returns \c true if \c Tok is the "." or "->" of a call and starts the next
47 // segment of a builder type call.
48 static bool startsSegmentOfBuilderTypeCall(const FormatToken &Tok) {
49   return Tok.isMemberAccess() && Tok.Previous && Tok.Previous->closesScope();
50 }
51 
52 // Returns \c true if \c Current starts a new parameter.
53 static bool startsNextParameter(const FormatToken &Current,
54                                 const FormatStyle &Style) {
55   const FormatToken &Previous = *Current.Previous;
56   if (Current.is(TT_CtorInitializerComma) &&
57       Style.BreakConstructorInitializersBeforeComma)
58     return true;
59   return Previous.is(tok::comma) && !Current.isTrailingComment() &&
60          (Previous.isNot(TT_CtorInitializerComma) ||
61           !Style.BreakConstructorInitializersBeforeComma);
62 }
63 
64 ContinuationIndenter::ContinuationIndenter(const FormatStyle &Style,
65                                            const AdditionalKeywords &Keywords,
66                                            const SourceManager &SourceMgr,
67                                            WhitespaceManager &Whitespaces,
68                                            encoding::Encoding Encoding,
69                                            bool BinPackInconclusiveFunctions)
70     : Style(Style), Keywords(Keywords), SourceMgr(SourceMgr),
71       Whitespaces(Whitespaces), Encoding(Encoding),
72       BinPackInconclusiveFunctions(BinPackInconclusiveFunctions),
73       CommentPragmasRegex(Style.CommentPragmas) {}
74 
75 LineState ContinuationIndenter::getInitialState(unsigned FirstIndent,
76                                                 const AnnotatedLine *Line,
77                                                 bool DryRun) {
78   LineState State;
79   State.FirstIndent = FirstIndent;
80   State.Column = FirstIndent;
81   State.Line = Line;
82   State.NextToken = Line->First;
83   State.Stack.push_back(ParenState(FirstIndent, Line->Level, FirstIndent,
84                                    /*AvoidBinPacking=*/false,
85                                    /*NoLineBreak=*/false));
86   State.LineContainsContinuedForLoopSection = false;
87   State.StartOfStringLiteral = 0;
88   State.StartOfLineLevel = 0;
89   State.LowestLevelOnLine = 0;
90   State.IgnoreStackForComparison = false;
91 
92   // The first token has already been indented and thus consumed.
93   moveStateToNextToken(State, DryRun, /*Newline=*/false);
94   return State;
95 }
96 
97 bool ContinuationIndenter::canBreak(const LineState &State) {
98   const FormatToken &Current = *State.NextToken;
99   const FormatToken &Previous = *Current.Previous;
100   assert(&Previous == Current.Previous);
101   if (!Current.CanBreakBefore &&
102       !(State.Stack.back().BreakBeforeClosingBrace &&
103         Current.closesBlockOrBlockTypeList(Style)))
104     return false;
105   // The opening "{" of a braced list has to be on the same line as the first
106   // element if it is nested in another braced init list or function call.
107   if (!Current.MustBreakBefore && Previous.is(tok::l_brace) &&
108       Previous.isNot(TT_DictLiteral) && Previous.BlockKind == BK_BracedInit &&
109       Previous.Previous &&
110       Previous.Previous->isOneOf(tok::l_brace, tok::l_paren, tok::comma))
111     return false;
112   // This prevents breaks like:
113   //   ...
114   //   SomeParameter, OtherParameter).DoSomething(
115   //   ...
116   // As they hide "DoSomething" and are generally bad for readability.
117   if (Previous.opensScope() && Previous.isNot(tok::l_brace) &&
118       State.LowestLevelOnLine < State.StartOfLineLevel &&
119       State.LowestLevelOnLine < Current.NestingLevel)
120     return false;
121   if (Current.isMemberAccess() && State.Stack.back().ContainsUnwrappedBuilder)
122     return false;
123 
124   // Don't create a 'hanging' indent if there are multiple blocks in a single
125   // statement.
126   if (Previous.is(tok::l_brace) && State.Stack.size() > 1 &&
127       State.Stack[State.Stack.size() - 2].NestedBlockInlined &&
128       State.Stack[State.Stack.size() - 2].HasMultipleNestedBlocks)
129     return false;
130 
131   // Don't break after very short return types (e.g. "void") as that is often
132   // unexpected.
133   if (Current.is(TT_FunctionDeclarationName) && State.Column < 6) {
134     if (Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_None)
135       return false;
136   }
137 
138   return !State.Stack.back().NoLineBreak;
139 }
140 
141 bool ContinuationIndenter::mustBreak(const LineState &State) {
142   const FormatToken &Current = *State.NextToken;
143   const FormatToken &Previous = *Current.Previous;
144   if (Current.MustBreakBefore || Current.is(TT_InlineASMColon))
145     return true;
146   if (State.Stack.back().BreakBeforeClosingBrace &&
147       Current.closesBlockOrBlockTypeList(Style))
148     return true;
149   if (Previous.is(tok::semi) && State.LineContainsContinuedForLoopSection)
150     return true;
151   if ((startsNextParameter(Current, Style) || Previous.is(tok::semi) ||
152        (Previous.is(TT_TemplateCloser) && Current.is(TT_StartOfName) &&
153         Style.Language == FormatStyle::LK_Cpp &&
154         // FIXME: This is a temporary workaround for the case where clang-format
155         // sets BreakBeforeParameter to avoid bin packing and this creates a
156         // completely unnecessary line break after a template type that isn't
157         // line-wrapped.
158         (Previous.NestingLevel == 1 || Style.BinPackParameters)) ||
159        (Style.BreakBeforeTernaryOperators && Current.is(TT_ConditionalExpr) &&
160         Previous.isNot(tok::question)) ||
161        (!Style.BreakBeforeTernaryOperators &&
162         Previous.is(TT_ConditionalExpr))) &&
163       State.Stack.back().BreakBeforeParameter && !Current.isTrailingComment() &&
164       !Current.isOneOf(tok::r_paren, tok::r_brace))
165     return true;
166   if (((Previous.is(TT_DictLiteral) && Previous.is(tok::l_brace)) ||
167        (Previous.is(TT_ArrayInitializerLSquare) &&
168         Previous.ParameterCount > 1)) &&
169       Style.ColumnLimit > 0 &&
170       getLengthToMatchingParen(Previous) + State.Column - 1 >
171           getColumnLimit(State))
172     return true;
173   if (Current.is(TT_CtorInitializerColon) &&
174       (State.Column + State.Line->Last->TotalLength - Current.TotalLength + 2 >
175            getColumnLimit(State) ||
176        State.Stack.back().BreakBeforeParameter) &&
177       ((Style.AllowShortFunctionsOnASingleLine != FormatStyle::SFS_All) ||
178        Style.BreakConstructorInitializersBeforeComma || Style.ColumnLimit != 0))
179     return true;
180   if (Current.is(TT_SelectorName) && State.Stack.back().ObjCSelectorNameFound &&
181       State.Stack.back().BreakBeforeParameter)
182     return true;
183 
184   unsigned NewLineColumn = getNewLineColumn(State);
185   if (Current.isMemberAccess() && Style.ColumnLimit != 0 &&
186       State.Column + getLengthToNextOperator(Current) > Style.ColumnLimit &&
187       (State.Column > NewLineColumn ||
188        Current.NestingLevel < State.StartOfLineLevel))
189     return true;
190 
191   if (State.Column <= NewLineColumn)
192     return false;
193 
194   if (Style.AlwaysBreakBeforeMultilineStrings &&
195       (NewLineColumn == State.FirstIndent + Style.ContinuationIndentWidth ||
196        Previous.is(tok::comma) || Current.NestingLevel < 2) &&
197       !Previous.isOneOf(tok::kw_return, tok::lessless, tok::at) &&
198       !Previous.isOneOf(TT_InlineASMColon, TT_ConditionalExpr) &&
199       nextIsMultilineString(State))
200     return true;
201 
202   // Using CanBreakBefore here and below takes care of the decision whether the
203   // current style uses wrapping before or after operators for the given
204   // operator.
205   if (Previous.is(TT_BinaryOperator) && Current.CanBreakBefore) {
206     // If we need to break somewhere inside the LHS of a binary expression, we
207     // should also break after the operator. Otherwise, the formatting would
208     // hide the operator precedence, e.g. in:
209     //   if (aaaaaaaaaaaaaa ==
210     //           bbbbbbbbbbbbbb && c) {..
211     // For comparisons, we only apply this rule, if the LHS is a binary
212     // expression itself as otherwise, the line breaks seem superfluous.
213     // We need special cases for ">>" which we have split into two ">" while
214     // lexing in order to make template parsing easier.
215     bool IsComparison = (Previous.getPrecedence() == prec::Relational ||
216                          Previous.getPrecedence() == prec::Equality) &&
217                         Previous.Previous &&
218                         Previous.Previous->isNot(TT_BinaryOperator); // For >>.
219     bool LHSIsBinaryExpr =
220         Previous.Previous && Previous.Previous->EndsBinaryExpression;
221     if ((!IsComparison || LHSIsBinaryExpr) && !Current.isTrailingComment() &&
222         Previous.getPrecedence() != prec::Assignment &&
223         State.Stack.back().BreakBeforeParameter)
224       return true;
225   } else if (Current.is(TT_BinaryOperator) && Current.CanBreakBefore &&
226              State.Stack.back().BreakBeforeParameter) {
227     return true;
228   }
229 
230   // Same as above, but for the first "<<" operator.
231   if (Current.is(tok::lessless) && Current.isNot(TT_OverloadedOperator) &&
232       State.Stack.back().BreakBeforeParameter &&
233       State.Stack.back().FirstLessLess == 0)
234     return true;
235 
236   if (Current.NestingLevel == 0 && !Current.isTrailingComment()) {
237     // Always break after "template <...>" and leading annotations. This is only
238     // for cases where the entire line does not fit on a single line as a
239     // different LineFormatter would be used otherwise.
240     if (Previous.ClosesTemplateDeclaration)
241       return true;
242     if (Previous.is(TT_FunctionAnnotationRParen))
243       return true;
244     if (Previous.is(TT_LeadingJavaAnnotation) && Current.isNot(tok::l_paren) &&
245         Current.isNot(TT_LeadingJavaAnnotation))
246       return true;
247   }
248 
249   // If the return type spans multiple lines, wrap before the function name.
250   if ((Current.is(TT_FunctionDeclarationName) ||
251        (Current.is(tok::kw_operator) && !Previous.is(tok::coloncolon))) &&
252       !Previous.is(tok::kw_template) && State.Stack.back().BreakBeforeParameter)
253     return true;
254 
255   if (startsSegmentOfBuilderTypeCall(Current) &&
256       (State.Stack.back().CallContinuation != 0 ||
257        State.Stack.back().BreakBeforeParameter))
258     return true;
259 
260   // The following could be precomputed as they do not depend on the state.
261   // However, as they should take effect only if the UnwrappedLine does not fit
262   // into the ColumnLimit, they are checked here in the ContinuationIndenter.
263   if (Style.ColumnLimit != 0 && Previous.BlockKind == BK_Block &&
264       Previous.is(tok::l_brace) && !Current.isOneOf(tok::r_brace, tok::comment))
265     return true;
266 
267   if (Current.is(tok::lessless) &&
268       ((Previous.is(tok::identifier) && Previous.TokenText == "endl") ||
269        (Previous.Tok.isLiteral() && (Previous.TokenText.endswith("\\n\"") ||
270                                      Previous.TokenText == "\'\\n\'"))))
271     return true;
272 
273   return false;
274 }
275 
276 unsigned ContinuationIndenter::addTokenToState(LineState &State, bool Newline,
277                                                bool DryRun,
278                                                unsigned ExtraSpaces) {
279   const FormatToken &Current = *State.NextToken;
280 
281   assert(!State.Stack.empty());
282   if ((Current.is(TT_ImplicitStringLiteral) &&
283        (Current.Previous->Tok.getIdentifierInfo() == nullptr ||
284         Current.Previous->Tok.getIdentifierInfo()->getPPKeywordID() ==
285             tok::pp_not_keyword))) {
286     unsigned EndColumn =
287         SourceMgr.getSpellingColumnNumber(Current.WhitespaceRange.getEnd());
288     if (Current.LastNewlineOffset != 0) {
289       // If there is a newline within this token, the final column will solely
290       // determined by the current end column.
291       State.Column = EndColumn;
292     } else {
293       unsigned StartColumn =
294           SourceMgr.getSpellingColumnNumber(Current.WhitespaceRange.getBegin());
295       assert(EndColumn >= StartColumn);
296       State.Column += EndColumn - StartColumn;
297     }
298     moveStateToNextToken(State, DryRun, /*Newline=*/false);
299     return 0;
300   }
301 
302   unsigned Penalty = 0;
303   if (Newline)
304     Penalty = addTokenOnNewLine(State, DryRun);
305   else
306     addTokenOnCurrentLine(State, DryRun, ExtraSpaces);
307 
308   return moveStateToNextToken(State, DryRun, Newline) + Penalty;
309 }
310 
311 void ContinuationIndenter::addTokenOnCurrentLine(LineState &State, bool DryRun,
312                                                  unsigned ExtraSpaces) {
313   FormatToken &Current = *State.NextToken;
314   const FormatToken &Previous = *State.NextToken->Previous;
315   if (Current.is(tok::equal) &&
316       (State.Line->First->is(tok::kw_for) || Current.NestingLevel == 0) &&
317       State.Stack.back().VariablePos == 0) {
318     State.Stack.back().VariablePos = State.Column;
319     // Move over * and & if they are bound to the variable name.
320     const FormatToken *Tok = &Previous;
321     while (Tok && State.Stack.back().VariablePos >= Tok->ColumnWidth) {
322       State.Stack.back().VariablePos -= Tok->ColumnWidth;
323       if (Tok->SpacesRequiredBefore != 0)
324         break;
325       Tok = Tok->Previous;
326     }
327     if (Previous.PartOfMultiVariableDeclStmt)
328       State.Stack.back().LastSpace = State.Stack.back().VariablePos;
329   }
330 
331   unsigned Spaces = Current.SpacesRequiredBefore + ExtraSpaces;
332 
333   if (!DryRun)
334     Whitespaces.replaceWhitespace(Current, /*Newlines=*/0, /*IndentLevel=*/0,
335                                   Spaces, State.Column + Spaces);
336 
337   if (Current.is(TT_SelectorName) &&
338       !State.Stack.back().ObjCSelectorNameFound) {
339     unsigned MinIndent =
340         std::max(State.FirstIndent + Style.ContinuationIndentWidth,
341                  State.Stack.back().Indent);
342     unsigned FirstColonPos = State.Column + Spaces + Current.ColumnWidth;
343     if (Current.LongestObjCSelectorName == 0)
344       State.Stack.back().AlignColons = false;
345     else if (MinIndent + Current.LongestObjCSelectorName > FirstColonPos)
346       State.Stack.back().ColonPos = MinIndent + Current.LongestObjCSelectorName;
347     else
348       State.Stack.back().ColonPos = FirstColonPos;
349   }
350 
351   // In "AlwaysBreak" mode, enforce wrapping directly after the parenthesis by
352   // disallowing any further line breaks if there is no line break after the
353   // opening parenthesis. Don't break if it doesn't conserve columns.
354   if (Style.AlignAfterOpenBracket == FormatStyle::BAS_AlwaysBreak &&
355       Previous.isOneOf(tok::l_paren, TT_TemplateOpener, tok::l_square) &&
356       State.Column > getNewLineColumn(State) &&
357       (!Previous.Previous ||
358        !Previous.Previous->isOneOf(tok::kw_for, tok::kw_while,
359                                    tok::kw_switch)) &&
360       // Don't do this for simple (no expressions) one-argument function calls
361       // as that feels like needlessly wasting whitespace, e.g.:
362       //
363       //   caaaaaaaaaaaall(
364       //       caaaaaaaaaaaall(
365       //           caaaaaaaaaaaall(
366       //               caaaaaaaaaaaaaaaaaaaaaaall(aaaaaaaaaaaaaa, aaaaaaaaa))));
367       Current.FakeLParens.size() > 0 &&
368       Current.FakeLParens.back() > prec::Unknown)
369     State.Stack.back().NoLineBreak = true;
370 
371   if (Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign &&
372       Previous.opensScope() && Previous.isNot(TT_ObjCMethodExpr) &&
373       (Current.isNot(TT_LineComment) || Previous.BlockKind == BK_BracedInit))
374     State.Stack.back().Indent = State.Column + Spaces;
375   if (State.Stack.back().AvoidBinPacking && startsNextParameter(Current, Style))
376     State.Stack.back().NoLineBreak = true;
377   if (startsSegmentOfBuilderTypeCall(Current) &&
378       State.Column > getNewLineColumn(State))
379     State.Stack.back().ContainsUnwrappedBuilder = true;
380 
381   if (Current.is(TT_LambdaArrow) && Style.Language == FormatStyle::LK_Java)
382     State.Stack.back().NoLineBreak = true;
383   if (Current.isMemberAccess() && Previous.is(tok::r_paren) &&
384       (Previous.MatchingParen &&
385        (Previous.TotalLength - Previous.MatchingParen->TotalLength > 10))) {
386     // If there is a function call with long parameters, break before trailing
387     // calls. This prevents things like:
388     //   EXPECT_CALL(SomeLongParameter).Times(
389     //       2);
390     // We don't want to do this for short parameters as they can just be
391     // indexes.
392     State.Stack.back().NoLineBreak = true;
393   }
394 
395   State.Column += Spaces;
396   if (Current.isNot(tok::comment) && Previous.is(tok::l_paren) &&
397       Previous.Previous &&
398       Previous.Previous->isOneOf(tok::kw_if, tok::kw_for)) {
399     // Treat the condition inside an if as if it was a second function
400     // parameter, i.e. let nested calls have a continuation indent.
401     State.Stack.back().LastSpace = State.Column;
402     State.Stack.back().NestedBlockIndent = State.Column;
403   } else if (!Current.isOneOf(tok::comment, tok::caret) &&
404              ((Previous.is(tok::comma) &&
405                !Previous.is(TT_OverloadedOperator)) ||
406               (Previous.is(tok::colon) && Previous.is(TT_ObjCMethodExpr)))) {
407     State.Stack.back().LastSpace = State.Column;
408   } else if ((Previous.isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
409                                TT_CtorInitializerColon)) &&
410              ((Previous.getPrecedence() != prec::Assignment &&
411                (Previous.isNot(tok::lessless) || Previous.OperatorIndex != 0 ||
412                 Previous.NextOperator)) ||
413               Current.StartsBinaryExpression)) {
414     // Indent relative to the RHS of the expression unless this is a simple
415     // assignment without binary expression on the RHS. Also indent relative to
416     // unary operators and the colons of constructor initializers.
417     State.Stack.back().LastSpace = State.Column;
418   } else if (Previous.is(TT_InheritanceColon)) {
419     State.Stack.back().Indent = State.Column;
420     State.Stack.back().LastSpace = State.Column;
421   } else if (Previous.opensScope()) {
422     // If a function has a trailing call, indent all parameters from the
423     // opening parenthesis. This avoids confusing indents like:
424     //   OuterFunction(InnerFunctionCall( // break
425     //       ParameterToInnerFunction))   // break
426     //       .SecondInnerFunctionCall();
427     bool HasTrailingCall = false;
428     if (Previous.MatchingParen) {
429       const FormatToken *Next = Previous.MatchingParen->getNextNonComment();
430       HasTrailingCall = Next && Next->isMemberAccess();
431     }
432     if (HasTrailingCall && State.Stack.size() > 1 &&
433         State.Stack[State.Stack.size() - 2].CallContinuation == 0)
434       State.Stack.back().LastSpace = State.Column;
435   }
436 }
437 
438 unsigned ContinuationIndenter::addTokenOnNewLine(LineState &State,
439                                                  bool DryRun) {
440   FormatToken &Current = *State.NextToken;
441   const FormatToken &Previous = *State.NextToken->Previous;
442 
443   // Extra penalty that needs to be added because of the way certain line
444   // breaks are chosen.
445   unsigned Penalty = 0;
446 
447   const FormatToken *PreviousNonComment = Current.getPreviousNonComment();
448   const FormatToken *NextNonComment = Previous.getNextNonComment();
449   if (!NextNonComment)
450     NextNonComment = &Current;
451   // The first line break on any NestingLevel causes an extra penalty in order
452   // prefer similar line breaks.
453   if (!State.Stack.back().ContainsLineBreak)
454     Penalty += 15;
455   State.Stack.back().ContainsLineBreak = true;
456 
457   Penalty += State.NextToken->SplitPenalty;
458 
459   // Breaking before the first "<<" is generally not desirable if the LHS is
460   // short. Also always add the penalty if the LHS is split over mutliple lines
461   // to avoid unnecessary line breaks that just work around this penalty.
462   if (NextNonComment->is(tok::lessless) &&
463       State.Stack.back().FirstLessLess == 0 &&
464       (State.Column <= Style.ColumnLimit / 3 ||
465        State.Stack.back().BreakBeforeParameter))
466     Penalty += Style.PenaltyBreakFirstLessLess;
467 
468   State.Column = getNewLineColumn(State);
469 
470   // Indent nested blocks relative to this column, unless in a very specific
471   // JavaScript special case where:
472   //
473   //   var loooooong_name =
474   //       function() {
475   //     // code
476   //   }
477   //
478   // is common and should be formatted like a free-standing function. The same
479   // goes for wrapping before the lambda return type arrow.
480   if (!Current.is(TT_LambdaArrow) &&
481       (Style.Language != FormatStyle::LK_JavaScript ||
482        Current.NestingLevel != 0 || !PreviousNonComment ||
483        !PreviousNonComment->is(tok::equal) ||
484        !Current.isOneOf(Keywords.kw_async, Keywords.kw_function)))
485     State.Stack.back().NestedBlockIndent = State.Column;
486 
487   if (NextNonComment->isMemberAccess()) {
488     if (State.Stack.back().CallContinuation == 0)
489       State.Stack.back().CallContinuation = State.Column;
490   } else if (NextNonComment->is(TT_SelectorName)) {
491     if (!State.Stack.back().ObjCSelectorNameFound) {
492       if (NextNonComment->LongestObjCSelectorName == 0) {
493         State.Stack.back().AlignColons = false;
494       } else {
495         State.Stack.back().ColonPos =
496             (Style.IndentWrappedFunctionNames
497                  ? std::max(State.Stack.back().Indent,
498                             State.FirstIndent + Style.ContinuationIndentWidth)
499                  : State.Stack.back().Indent) +
500             NextNonComment->LongestObjCSelectorName;
501       }
502     } else if (State.Stack.back().AlignColons &&
503                State.Stack.back().ColonPos <= NextNonComment->ColumnWidth) {
504       State.Stack.back().ColonPos = State.Column + NextNonComment->ColumnWidth;
505     }
506   } else if (PreviousNonComment && PreviousNonComment->is(tok::colon) &&
507              PreviousNonComment->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)) {
508     // FIXME: This is hacky, find a better way. The problem is that in an ObjC
509     // method expression, the block should be aligned to the line starting it,
510     // e.g.:
511     //   [aaaaaaaaaaaaaaa aaaaaaaaa: \\ break for some reason
512     //                        ^(int *i) {
513     //                            // ...
514     //                        }];
515     // Thus, we set LastSpace of the next higher NestingLevel, to which we move
516     // when we consume all of the "}"'s FakeRParens at the "{".
517     if (State.Stack.size() > 1)
518       State.Stack[State.Stack.size() - 2].LastSpace =
519           std::max(State.Stack.back().LastSpace, State.Stack.back().Indent) +
520           Style.ContinuationIndentWidth;
521   }
522 
523   if ((Previous.isOneOf(tok::comma, tok::semi) &&
524        !State.Stack.back().AvoidBinPacking) ||
525       Previous.is(TT_BinaryOperator))
526     State.Stack.back().BreakBeforeParameter = false;
527   if (Previous.isOneOf(TT_TemplateCloser, TT_JavaAnnotation) &&
528       Current.NestingLevel == 0)
529     State.Stack.back().BreakBeforeParameter = false;
530   if (NextNonComment->is(tok::question) ||
531       (PreviousNonComment && PreviousNonComment->is(tok::question)))
532     State.Stack.back().BreakBeforeParameter = true;
533   if (Current.is(TT_BinaryOperator) && Current.CanBreakBefore)
534     State.Stack.back().BreakBeforeParameter = false;
535 
536   if (!DryRun) {
537     unsigned Newlines = std::max(
538         1u, std::min(Current.NewlinesBefore, Style.MaxEmptyLinesToKeep + 1));
539     Whitespaces.replaceWhitespace(Current, Newlines,
540                                   State.Stack.back().IndentLevel, State.Column,
541                                   State.Column, State.Line->InPPDirective);
542   }
543 
544   if (!Current.isTrailingComment())
545     State.Stack.back().LastSpace = State.Column;
546   if (Current.is(tok::lessless))
547     // If we are breaking before a "<<", we always want to indent relative to
548     // RHS. This is necessary only for "<<", as we special-case it and don't
549     // always indent relative to the RHS.
550     State.Stack.back().LastSpace += 3; // 3 -> width of "<< ".
551 
552   State.StartOfLineLevel = Current.NestingLevel;
553   State.LowestLevelOnLine = Current.NestingLevel;
554 
555   // Any break on this level means that the parent level has been broken
556   // and we need to avoid bin packing there.
557   bool NestedBlockSpecialCase =
558       Style.Language != FormatStyle::LK_Cpp &&
559       Current.is(tok::r_brace) && State.Stack.size() > 1 &&
560       State.Stack[State.Stack.size() - 2].NestedBlockInlined;
561   if (!NestedBlockSpecialCase)
562     for (unsigned i = 0, e = State.Stack.size() - 1; i != e; ++i)
563       State.Stack[i].BreakBeforeParameter = true;
564 
565   if (PreviousNonComment &&
566       !PreviousNonComment->isOneOf(tok::comma, tok::semi) &&
567       (PreviousNonComment->isNot(TT_TemplateCloser) ||
568        Current.NestingLevel != 0) &&
569       !PreviousNonComment->isOneOf(
570           TT_BinaryOperator, TT_FunctionAnnotationRParen, TT_JavaAnnotation,
571           TT_LeadingJavaAnnotation) &&
572       Current.isNot(TT_BinaryOperator) && !PreviousNonComment->opensScope())
573     State.Stack.back().BreakBeforeParameter = true;
574 
575   // If we break after { or the [ of an array initializer, we should also break
576   // before the corresponding } or ].
577   if (PreviousNonComment &&
578       (PreviousNonComment->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare)))
579     State.Stack.back().BreakBeforeClosingBrace = true;
580 
581   if (State.Stack.back().AvoidBinPacking) {
582     // If we are breaking after '(', '{', '<', this is not bin packing
583     // unless AllowAllParametersOfDeclarationOnNextLine is false or this is a
584     // dict/object literal.
585     if (!Previous.isOneOf(tok::l_paren, tok::l_brace, TT_BinaryOperator) ||
586         (!Style.AllowAllParametersOfDeclarationOnNextLine &&
587          State.Line->MustBeDeclaration) ||
588         Previous.is(TT_DictLiteral))
589       State.Stack.back().BreakBeforeParameter = true;
590   }
591 
592   return Penalty;
593 }
594 
595 unsigned ContinuationIndenter::getNewLineColumn(const LineState &State) {
596   if (!State.NextToken || !State.NextToken->Previous)
597     return 0;
598   FormatToken &Current = *State.NextToken;
599   const FormatToken &Previous = *Current.Previous;
600   // If we are continuing an expression, we want to use the continuation indent.
601   unsigned ContinuationIndent =
602       std::max(State.Stack.back().LastSpace, State.Stack.back().Indent) +
603       Style.ContinuationIndentWidth;
604   const FormatToken *PreviousNonComment = Current.getPreviousNonComment();
605   const FormatToken *NextNonComment = Previous.getNextNonComment();
606   if (!NextNonComment)
607     NextNonComment = &Current;
608 
609   // Java specific bits.
610   if (Style.Language == FormatStyle::LK_Java &&
611       Current.isOneOf(Keywords.kw_implements, Keywords.kw_extends))
612     return std::max(State.Stack.back().LastSpace,
613                     State.Stack.back().Indent + Style.ContinuationIndentWidth);
614 
615   if (NextNonComment->is(tok::l_brace) && NextNonComment->BlockKind == BK_Block)
616     return Current.NestingLevel == 0 ? State.FirstIndent
617                                      : State.Stack.back().Indent;
618   if (Current.isOneOf(tok::r_brace, tok::r_square) && State.Stack.size() > 1) {
619     if (Current.closesBlockOrBlockTypeList(Style))
620       return State.Stack[State.Stack.size() - 2].NestedBlockIndent;
621     if (Current.MatchingParen &&
622         Current.MatchingParen->BlockKind == BK_BracedInit)
623       return State.Stack[State.Stack.size() - 2].LastSpace;
624     return State.FirstIndent;
625   }
626   if (Current.is(tok::identifier) && Current.Next &&
627       Current.Next->is(TT_DictLiteral))
628     return State.Stack.back().Indent;
629   if (NextNonComment->isStringLiteral() && State.StartOfStringLiteral != 0)
630     return State.StartOfStringLiteral;
631   if (NextNonComment->is(TT_ObjCStringLiteral) &&
632       State.StartOfStringLiteral != 0)
633     return State.StartOfStringLiteral - 1;
634   if (NextNonComment->is(tok::lessless) &&
635       State.Stack.back().FirstLessLess != 0)
636     return State.Stack.back().FirstLessLess;
637   if (NextNonComment->isMemberAccess()) {
638     if (State.Stack.back().CallContinuation == 0)
639       return ContinuationIndent;
640     return State.Stack.back().CallContinuation;
641   }
642   if (State.Stack.back().QuestionColumn != 0 &&
643       ((NextNonComment->is(tok::colon) &&
644         NextNonComment->is(TT_ConditionalExpr)) ||
645        Previous.is(TT_ConditionalExpr)))
646     return State.Stack.back().QuestionColumn;
647   if (Previous.is(tok::comma) && State.Stack.back().VariablePos != 0)
648     return State.Stack.back().VariablePos;
649   if ((PreviousNonComment &&
650        (PreviousNonComment->ClosesTemplateDeclaration ||
651         PreviousNonComment->isOneOf(
652             TT_AttributeParen, TT_FunctionAnnotationRParen, TT_JavaAnnotation,
653             TT_LeadingJavaAnnotation))) ||
654       (!Style.IndentWrappedFunctionNames &&
655        NextNonComment->isOneOf(tok::kw_operator, TT_FunctionDeclarationName)))
656     return std::max(State.Stack.back().LastSpace, State.Stack.back().Indent);
657   if (NextNonComment->is(TT_SelectorName)) {
658     if (!State.Stack.back().ObjCSelectorNameFound) {
659       if (NextNonComment->LongestObjCSelectorName == 0)
660         return State.Stack.back().Indent;
661       return (Style.IndentWrappedFunctionNames
662                   ? std::max(State.Stack.back().Indent,
663                              State.FirstIndent + Style.ContinuationIndentWidth)
664                   : State.Stack.back().Indent) +
665              NextNonComment->LongestObjCSelectorName -
666              NextNonComment->ColumnWidth;
667     }
668     if (!State.Stack.back().AlignColons)
669       return State.Stack.back().Indent;
670     if (State.Stack.back().ColonPos > NextNonComment->ColumnWidth)
671       return State.Stack.back().ColonPos - NextNonComment->ColumnWidth;
672     return State.Stack.back().Indent;
673   }
674   if (NextNonComment->is(TT_ArraySubscriptLSquare)) {
675     if (State.Stack.back().StartOfArraySubscripts != 0)
676       return State.Stack.back().StartOfArraySubscripts;
677     return ContinuationIndent;
678   }
679 
680   // This ensure that we correctly format ObjC methods calls without inputs,
681   // i.e. where the last element isn't selector like: [callee method];
682   if (NextNonComment->is(tok::identifier) && NextNonComment->FakeRParens == 0 &&
683       NextNonComment->Next && NextNonComment->Next->is(TT_ObjCMethodExpr))
684     return State.Stack.back().Indent;
685 
686   if (NextNonComment->isOneOf(TT_StartOfName, TT_PointerOrReference) ||
687       Previous.isOneOf(tok::coloncolon, tok::equal, TT_JsTypeColon))
688     return ContinuationIndent;
689   if (PreviousNonComment && PreviousNonComment->is(tok::colon) &&
690       PreviousNonComment->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))
691     return ContinuationIndent;
692   if (NextNonComment->is(TT_CtorInitializerColon))
693     return State.FirstIndent + Style.ConstructorInitializerIndentWidth;
694   if (NextNonComment->is(TT_CtorInitializerComma))
695     return State.Stack.back().Indent;
696   if (Previous.is(tok::r_paren) && !Current.isBinaryOperator() &&
697       !Current.isOneOf(tok::colon, tok::comment))
698     return ContinuationIndent;
699   if (State.Stack.back().Indent == State.FirstIndent && PreviousNonComment &&
700       PreviousNonComment->isNot(tok::r_brace))
701     // Ensure that we fall back to the continuation indent width instead of
702     // just flushing continuations left.
703     return State.Stack.back().Indent + Style.ContinuationIndentWidth;
704   return State.Stack.back().Indent;
705 }
706 
707 unsigned ContinuationIndenter::moveStateToNextToken(LineState &State,
708                                                     bool DryRun, bool Newline) {
709   assert(State.Stack.size());
710   const FormatToken &Current = *State.NextToken;
711 
712   if (Current.is(TT_InheritanceColon))
713     State.Stack.back().AvoidBinPacking = true;
714   if (Current.is(tok::lessless) && Current.isNot(TT_OverloadedOperator)) {
715     if (State.Stack.back().FirstLessLess == 0)
716       State.Stack.back().FirstLessLess = State.Column;
717     else
718       State.Stack.back().LastOperatorWrapped = Newline;
719   }
720   if ((Current.is(TT_BinaryOperator) && Current.isNot(tok::lessless)) ||
721       Current.is(TT_ConditionalExpr))
722     State.Stack.back().LastOperatorWrapped = Newline;
723   if (Current.is(TT_ArraySubscriptLSquare) &&
724       State.Stack.back().StartOfArraySubscripts == 0)
725     State.Stack.back().StartOfArraySubscripts = State.Column;
726   if (Style.BreakBeforeTernaryOperators && Current.is(tok::question))
727     State.Stack.back().QuestionColumn = State.Column;
728   if (!Style.BreakBeforeTernaryOperators && Current.isNot(tok::colon)) {
729     const FormatToken *Previous = Current.Previous;
730     while (Previous && Previous->isTrailingComment())
731       Previous = Previous->Previous;
732     if (Previous && Previous->is(tok::question))
733       State.Stack.back().QuestionColumn = State.Column;
734   }
735   if (!Current.opensScope() && !Current.closesScope())
736     State.LowestLevelOnLine =
737         std::min(State.LowestLevelOnLine, Current.NestingLevel);
738   if (Current.isMemberAccess())
739     State.Stack.back().StartOfFunctionCall =
740         !Current.NextOperator ? 0 : State.Column;
741   if (Current.is(TT_SelectorName)) {
742     State.Stack.back().ObjCSelectorNameFound = true;
743     if (Style.IndentWrappedFunctionNames) {
744       State.Stack.back().Indent =
745           State.FirstIndent + Style.ContinuationIndentWidth;
746     }
747   }
748   if (Current.is(TT_CtorInitializerColon)) {
749     // Indent 2 from the column, so:
750     // SomeClass::SomeClass()
751     //     : First(...), ...
752     //       Next(...)
753     //       ^ line up here.
754     State.Stack.back().Indent =
755         State.Column + (Style.BreakConstructorInitializersBeforeComma ? 0 : 2);
756     State.Stack.back().NestedBlockIndent = State.Stack.back().Indent;
757     if (Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
758       State.Stack.back().AvoidBinPacking = true;
759     State.Stack.back().BreakBeforeParameter = false;
760   }
761   if (Current.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && Newline)
762     State.Stack.back().NestedBlockIndent =
763         State.Column + Current.ColumnWidth + 1;
764 
765   // Insert scopes created by fake parenthesis.
766   const FormatToken *Previous = Current.getPreviousNonComment();
767 
768   // Add special behavior to support a format commonly used for JavaScript
769   // closures:
770   //   SomeFunction(function() {
771   //     foo();
772   //     bar();
773   //   }, a, b, c);
774   if (Current.isNot(tok::comment) && Previous &&
775       Previous->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) &&
776       !Previous->is(TT_DictLiteral) && State.Stack.size() > 1) {
777     if (State.Stack[State.Stack.size() - 2].NestedBlockInlined && Newline)
778       for (unsigned i = 0, e = State.Stack.size() - 1; i != e; ++i)
779         State.Stack[i].NoLineBreak = true;
780     State.Stack[State.Stack.size() - 2].NestedBlockInlined = false;
781   }
782   if (Previous && (Previous->isOneOf(tok::l_paren, tok::comma, tok::colon) ||
783                    Previous->isOneOf(TT_BinaryOperator, TT_ConditionalExpr)) &&
784       !Previous->isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
785     State.Stack.back().NestedBlockInlined =
786         !Newline &&
787         (Previous->isNot(tok::l_paren) || Previous->ParameterCount > 1);
788   }
789 
790   moveStatePastFakeLParens(State, Newline);
791   moveStatePastScopeOpener(State, Newline);
792   moveStatePastScopeCloser(State);
793   moveStatePastFakeRParens(State);
794 
795   if (Current.isStringLiteral() && State.StartOfStringLiteral == 0)
796     State.StartOfStringLiteral = State.Column;
797   if (Current.is(TT_ObjCStringLiteral) && State.StartOfStringLiteral == 0)
798     State.StartOfStringLiteral = State.Column + 1;
799   else if (!Current.isOneOf(tok::comment, tok::identifier, tok::hash) &&
800            !Current.isStringLiteral())
801     State.StartOfStringLiteral = 0;
802 
803   State.Column += Current.ColumnWidth;
804   State.NextToken = State.NextToken->Next;
805   unsigned Penalty = breakProtrudingToken(Current, State, DryRun);
806   if (State.Column > getColumnLimit(State)) {
807     unsigned ExcessCharacters = State.Column - getColumnLimit(State);
808     Penalty += Style.PenaltyExcessCharacter * ExcessCharacters;
809   }
810 
811   if (Current.Role)
812     Current.Role->formatFromToken(State, this, DryRun);
813   // If the previous has a special role, let it consume tokens as appropriate.
814   // It is necessary to start at the previous token for the only implemented
815   // role (comma separated list). That way, the decision whether or not to break
816   // after the "{" is already done and both options are tried and evaluated.
817   // FIXME: This is ugly, find a better way.
818   if (Previous && Previous->Role)
819     Penalty += Previous->Role->formatAfterToken(State, this, DryRun);
820 
821   return Penalty;
822 }
823 
824 void ContinuationIndenter::moveStatePastFakeLParens(LineState &State,
825                                                     bool Newline) {
826   const FormatToken &Current = *State.NextToken;
827   const FormatToken *Previous = Current.getPreviousNonComment();
828 
829   // Don't add extra indentation for the first fake parenthesis after
830   // 'return', assignments or opening <({[. The indentation for these cases
831   // is special cased.
832   bool SkipFirstExtraIndent =
833       (Previous && (Previous->opensScope() ||
834                     Previous->isOneOf(tok::semi, tok::kw_return) ||
835                     (Previous->getPrecedence() == prec::Assignment &&
836                      Style.AlignOperands) ||
837                     Previous->is(TT_ObjCMethodExpr)));
838   for (SmallVectorImpl<prec::Level>::const_reverse_iterator
839            I = Current.FakeLParens.rbegin(),
840            E = Current.FakeLParens.rend();
841        I != E; ++I) {
842     ParenState NewParenState = State.Stack.back();
843     NewParenState.ContainsLineBreak = false;
844 
845     // Indent from 'LastSpace' unless these are fake parentheses encapsulating
846     // a builder type call after 'return' or, if the alignment after opening
847     // brackets is disabled.
848     if (!Current.isTrailingComment() &&
849         (Style.AlignOperands || *I < prec::Assignment) &&
850         (!Previous || Previous->isNot(tok::kw_return) ||
851          (Style.Language != FormatStyle::LK_Java && *I > 0)) &&
852         (Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign ||
853          *I != prec::Comma || Current.NestingLevel == 0))
854       NewParenState.Indent =
855           std::max(std::max(State.Column, NewParenState.Indent),
856                    State.Stack.back().LastSpace);
857 
858     // Don't allow the RHS of an operator to be split over multiple lines unless
859     // there is a line-break right after the operator.
860     // Exclude relational operators, as there, it is always more desirable to
861     // have the LHS 'left' of the RHS.
862     if (Previous && Previous->getPrecedence() != prec::Assignment &&
863         Previous->isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
864         Previous->getPrecedence() != prec::Relational) {
865       bool BreakBeforeOperator =
866           Previous->is(tok::lessless) ||
867           (Previous->is(TT_BinaryOperator) &&
868            Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None) ||
869           (Previous->is(TT_ConditionalExpr) &&
870            Style.BreakBeforeTernaryOperators);
871       if ((!Newline && !BreakBeforeOperator) ||
872           (!State.Stack.back().LastOperatorWrapped && BreakBeforeOperator))
873         NewParenState.NoLineBreak = true;
874     }
875 
876     // Do not indent relative to the fake parentheses inserted for "." or "->".
877     // This is a special case to make the following to statements consistent:
878     //   OuterFunction(InnerFunctionCall( // break
879     //       ParameterToInnerFunction));
880     //   OuterFunction(SomeObject.InnerFunctionCall( // break
881     //       ParameterToInnerFunction));
882     if (*I > prec::Unknown)
883       NewParenState.LastSpace = std::max(NewParenState.LastSpace, State.Column);
884     if (*I != prec::Conditional && !Current.is(TT_UnaryOperator) &&
885         Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign)
886       NewParenState.StartOfFunctionCall = State.Column;
887 
888     // Always indent conditional expressions. Never indent expression where
889     // the 'operator' is ',', ';' or an assignment (i.e. *I <=
890     // prec::Assignment) as those have different indentation rules. Indent
891     // other expression, unless the indentation needs to be skipped.
892     if (*I == prec::Conditional ||
893         (!SkipFirstExtraIndent && *I > prec::Assignment &&
894          !Current.isTrailingComment()))
895       NewParenState.Indent += Style.ContinuationIndentWidth;
896     if ((Previous && !Previous->opensScope()) || *I != prec::Comma)
897       NewParenState.BreakBeforeParameter = false;
898     State.Stack.push_back(NewParenState);
899     SkipFirstExtraIndent = false;
900   }
901 }
902 
903 void ContinuationIndenter::moveStatePastFakeRParens(LineState &State) {
904   for (unsigned i = 0, e = State.NextToken->FakeRParens; i != e; ++i) {
905     unsigned VariablePos = State.Stack.back().VariablePos;
906     if (State.Stack.size() == 1) {
907       // Do not pop the last element.
908       break;
909     }
910     State.Stack.pop_back();
911     State.Stack.back().VariablePos = VariablePos;
912   }
913 }
914 
915 void ContinuationIndenter::moveStatePastScopeOpener(LineState &State,
916                                                     bool Newline) {
917   const FormatToken &Current = *State.NextToken;
918   if (!Current.opensScope())
919     return;
920 
921   if (Current.MatchingParen && Current.BlockKind == BK_Block) {
922     moveStateToNewBlock(State);
923     return;
924   }
925 
926   unsigned NewIndent;
927   unsigned NewIndentLevel = State.Stack.back().IndentLevel;
928   unsigned LastSpace = State.Stack.back().LastSpace;
929   bool AvoidBinPacking;
930   bool BreakBeforeParameter = false;
931   unsigned NestedBlockIndent = std::max(State.Stack.back().StartOfFunctionCall,
932                                         State.Stack.back().NestedBlockIndent);
933   if (Current.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare)) {
934     if (Current.opensBlockOrBlockTypeList(Style)) {
935       NewIndent = State.Stack.back().NestedBlockIndent + Style.IndentWidth;
936       NewIndent = std::min(State.Column + 2, NewIndent);
937       ++NewIndentLevel;
938     } else {
939       NewIndent = State.Stack.back().LastSpace + Style.ContinuationIndentWidth;
940     }
941     const FormatToken *NextNoComment = Current.getNextNonComment();
942     bool EndsInComma = Current.MatchingParen &&
943                        Current.MatchingParen->Previous &&
944                        Current.MatchingParen->Previous->is(tok::comma);
945     AvoidBinPacking =
946         (Current.is(TT_ArrayInitializerLSquare) && EndsInComma) ||
947         Current.is(TT_DictLiteral) ||
948         Style.Language == FormatStyle::LK_Proto || !Style.BinPackArguments ||
949         (NextNoComment && NextNoComment->is(TT_DesignatedInitializerPeriod));
950     if (Current.ParameterCount > 1)
951       NestedBlockIndent = std::max(NestedBlockIndent, State.Column + 1);
952   } else {
953     NewIndent = Style.ContinuationIndentWidth +
954                 std::max(State.Stack.back().LastSpace,
955                          State.Stack.back().StartOfFunctionCall);
956 
957     // Ensure that different different brackets force relative alignment, e.g.:
958     // void SomeFunction(vector<  // break
959     //                       int> v);
960     // FIXME: We likely want to do this for more combinations of brackets.
961     // Verify that it is wanted for ObjC, too.
962     if (Current.Tok.getKind() == tok::less &&
963         Current.ParentBracket == tok::l_paren) {
964       NewIndent = std::max(NewIndent, State.Stack.back().Indent);
965       LastSpace = std::max(LastSpace, State.Stack.back().Indent);
966     }
967 
968     AvoidBinPacking =
969         (State.Line->MustBeDeclaration && !Style.BinPackParameters) ||
970         (!State.Line->MustBeDeclaration && !Style.BinPackArguments) ||
971         (Style.ExperimentalAutoDetectBinPacking &&
972          (Current.PackingKind == PPK_OnePerLine ||
973           (!BinPackInconclusiveFunctions &&
974            Current.PackingKind == PPK_Inconclusive)));
975     if (Current.is(TT_ObjCMethodExpr) && Current.MatchingParen) {
976       if (Style.ColumnLimit) {
977         // If this '[' opens an ObjC call, determine whether all parameters fit
978         // into one line and put one per line if they don't.
979         if (getLengthToMatchingParen(Current) + State.Column >
980             getColumnLimit(State))
981           BreakBeforeParameter = true;
982       } else {
983         // For ColumnLimit = 0, we have to figure out whether there is or has to
984         // be a line break within this call.
985         for (const FormatToken *Tok = &Current;
986              Tok && Tok != Current.MatchingParen; Tok = Tok->Next) {
987           if (Tok->MustBreakBefore ||
988               (Tok->CanBreakBefore && Tok->NewlinesBefore > 0)) {
989             BreakBeforeParameter = true;
990             break;
991           }
992         }
993       }
994     }
995   }
996   // Generally inherit NoLineBreak from the current scope to nested scope.
997   // However, don't do this for non-empty nested blocks, dict literals and
998   // array literals as these follow different indentation rules.
999   bool NoLineBreak =
1000       Current.Children.empty() &&
1001       !Current.isOneOf(TT_DictLiteral, TT_ArrayInitializerLSquare) &&
1002       (State.Stack.back().NoLineBreak ||
1003        (Current.is(TT_TemplateOpener) &&
1004         State.Stack.back().ContainsUnwrappedBuilder));
1005   State.Stack.push_back(ParenState(NewIndent, NewIndentLevel, LastSpace,
1006                                    AvoidBinPacking, NoLineBreak));
1007   State.Stack.back().NestedBlockIndent = NestedBlockIndent;
1008   State.Stack.back().BreakBeforeParameter = BreakBeforeParameter;
1009   State.Stack.back().HasMultipleNestedBlocks = Current.BlockParameterCount > 1;
1010 }
1011 
1012 void ContinuationIndenter::moveStatePastScopeCloser(LineState &State) {
1013   const FormatToken &Current = *State.NextToken;
1014   if (!Current.closesScope())
1015     return;
1016 
1017   // If we encounter a closing ), ], } or >, we can remove a level from our
1018   // stacks.
1019   if (State.Stack.size() > 1 &&
1020       (Current.isOneOf(tok::r_paren, tok::r_square) ||
1021        (Current.is(tok::r_brace) && State.NextToken != State.Line->First) ||
1022        State.NextToken->is(TT_TemplateCloser)))
1023     State.Stack.pop_back();
1024 
1025   if (Current.is(tok::r_square)) {
1026     // If this ends the array subscript expr, reset the corresponding value.
1027     const FormatToken *NextNonComment = Current.getNextNonComment();
1028     if (NextNonComment && NextNonComment->isNot(tok::l_square))
1029       State.Stack.back().StartOfArraySubscripts = 0;
1030   }
1031 }
1032 
1033 void ContinuationIndenter::moveStateToNewBlock(LineState &State) {
1034   unsigned NestedBlockIndent = State.Stack.back().NestedBlockIndent;
1035   // ObjC block sometimes follow special indentation rules.
1036   unsigned NewIndent =
1037       NestedBlockIndent + (State.NextToken->is(TT_ObjCBlockLBrace)
1038                                ? Style.ObjCBlockIndentWidth
1039                                : Style.IndentWidth);
1040   State.Stack.push_back(ParenState(
1041       NewIndent, /*NewIndentLevel=*/State.Stack.back().IndentLevel + 1,
1042       State.Stack.back().LastSpace, /*AvoidBinPacking=*/true,
1043       /*NoLineBreak=*/false));
1044   State.Stack.back().NestedBlockIndent = NestedBlockIndent;
1045   State.Stack.back().BreakBeforeParameter = true;
1046 }
1047 
1048 unsigned ContinuationIndenter::addMultilineToken(const FormatToken &Current,
1049                                                  LineState &State) {
1050   if (!Current.IsMultiline)
1051     return 0;
1052 
1053   // Break before further function parameters on all levels.
1054   for (unsigned i = 0, e = State.Stack.size(); i != e; ++i)
1055     State.Stack[i].BreakBeforeParameter = true;
1056 
1057   unsigned ColumnsUsed = State.Column;
1058   // We can only affect layout of the first and the last line, so the penalty
1059   // for all other lines is constant, and we ignore it.
1060   State.Column = Current.LastLineColumnWidth;
1061 
1062   if (ColumnsUsed > getColumnLimit(State))
1063     return Style.PenaltyExcessCharacter * (ColumnsUsed - getColumnLimit(State));
1064   return 0;
1065 }
1066 
1067 unsigned ContinuationIndenter::breakProtrudingToken(const FormatToken &Current,
1068                                                     LineState &State,
1069                                                     bool DryRun) {
1070   // Don't break multi-line tokens other than block comments. Instead, just
1071   // update the state.
1072   if (Current.isNot(TT_BlockComment) && Current.IsMultiline)
1073     return addMultilineToken(Current, State);
1074 
1075   // Don't break implicit string literals or import statements.
1076   if (Current.is(TT_ImplicitStringLiteral) ||
1077       State.Line->Type == LT_ImportStatement)
1078     return 0;
1079 
1080   if (!Current.isStringLiteral() && !Current.is(tok::comment))
1081     return 0;
1082 
1083   std::unique_ptr<BreakableToken> Token;
1084   unsigned StartColumn = State.Column - Current.ColumnWidth;
1085   unsigned ColumnLimit = getColumnLimit(State);
1086 
1087   if (Current.isStringLiteral()) {
1088     // FIXME: String literal breaking is currently disabled for Java and JS, as
1089     // it requires strings to be merged using "+" which we don't support.
1090     if (Style.Language == FormatStyle::LK_Java ||
1091         Style.Language == FormatStyle::LK_JavaScript ||
1092         !Style.BreakStringLiterals)
1093       return 0;
1094 
1095     // Don't break string literals inside preprocessor directives (except for
1096     // #define directives, as their contents are stored in separate lines and
1097     // are not affected by this check).
1098     // This way we avoid breaking code with line directives and unknown
1099     // preprocessor directives that contain long string literals.
1100     if (State.Line->Type == LT_PreprocessorDirective)
1101       return 0;
1102     // Exempts unterminated string literals from line breaking. The user will
1103     // likely want to terminate the string before any line breaking is done.
1104     if (Current.IsUnterminatedLiteral)
1105       return 0;
1106 
1107     StringRef Text = Current.TokenText;
1108     StringRef Prefix;
1109     StringRef Postfix;
1110     bool IsNSStringLiteral = false;
1111     // FIXME: Handle whitespace between '_T', '(', '"..."', and ')'.
1112     // FIXME: Store Prefix and Suffix (or PrefixLength and SuffixLength to
1113     // reduce the overhead) for each FormatToken, which is a string, so that we
1114     // don't run multiple checks here on the hot path.
1115     if (Text.startswith("\"") && Current.Previous &&
1116         Current.Previous->is(tok::at)) {
1117       IsNSStringLiteral = true;
1118       Prefix = "@\"";
1119     }
1120     if ((Text.endswith(Postfix = "\"") &&
1121          (IsNSStringLiteral || Text.startswith(Prefix = "\"") ||
1122           Text.startswith(Prefix = "u\"") || Text.startswith(Prefix = "U\"") ||
1123           Text.startswith(Prefix = "u8\"") ||
1124           Text.startswith(Prefix = "L\""))) ||
1125         (Text.startswith(Prefix = "_T(\"") && Text.endswith(Postfix = "\")"))) {
1126       Token.reset(new BreakableStringLiteral(
1127           Current, State.Line->Level, StartColumn, Prefix, Postfix,
1128           State.Line->InPPDirective, Encoding, Style));
1129     } else {
1130       return 0;
1131     }
1132   } else if (Current.is(TT_BlockComment)) {
1133     if (!Current.isTrailingComment() || !Style.ReflowComments ||
1134         CommentPragmasRegex.match(Current.TokenText.substr(2)))
1135       return addMultilineToken(Current, State);
1136     Token.reset(new BreakableBlockComment(
1137         Current, State.Line->Level, StartColumn, Current.OriginalColumn,
1138         !Current.Previous, State.Line->InPPDirective, Encoding, Style));
1139   } else if (Current.is(TT_LineComment) &&
1140              (Current.Previous == nullptr ||
1141               Current.Previous->isNot(TT_ImplicitStringLiteral))) {
1142     if (!Style.ReflowComments ||
1143         CommentPragmasRegex.match(Current.TokenText.substr(2)))
1144       return 0;
1145     Token.reset(new BreakableLineComment(Current, State.Line->Level,
1146                                          StartColumn, /*InPPDirective=*/false,
1147                                          Encoding, Style));
1148     // We don't insert backslashes when breaking line comments.
1149     ColumnLimit = Style.ColumnLimit;
1150   } else {
1151     return 0;
1152   }
1153   if (Current.UnbreakableTailLength >= ColumnLimit)
1154     return 0;
1155 
1156   unsigned RemainingSpace = ColumnLimit - Current.UnbreakableTailLength;
1157   bool BreakInserted = false;
1158   unsigned Penalty = 0;
1159   unsigned RemainingTokenColumns = 0;
1160   for (unsigned LineIndex = 0, EndIndex = Token->getLineCount();
1161        LineIndex != EndIndex; ++LineIndex) {
1162     if (!DryRun)
1163       Token->replaceWhitespaceBefore(LineIndex, Whitespaces);
1164     unsigned TailOffset = 0;
1165     RemainingTokenColumns =
1166         Token->getLineLengthAfterSplit(LineIndex, TailOffset, StringRef::npos);
1167     while (RemainingTokenColumns > RemainingSpace) {
1168       BreakableToken::Split Split =
1169           Token->getSplit(LineIndex, TailOffset, ColumnLimit);
1170       if (Split.first == StringRef::npos) {
1171         // The last line's penalty is handled in addNextStateToQueue().
1172         if (LineIndex < EndIndex - 1)
1173           Penalty += Style.PenaltyExcessCharacter *
1174                      (RemainingTokenColumns - RemainingSpace);
1175         break;
1176       }
1177       assert(Split.first != 0);
1178       unsigned NewRemainingTokenColumns = Token->getLineLengthAfterSplit(
1179           LineIndex, TailOffset + Split.first + Split.second, StringRef::npos);
1180 
1181       // We can remove extra whitespace instead of breaking the line.
1182       if (RemainingTokenColumns + 1 - Split.second <= RemainingSpace) {
1183         RemainingTokenColumns = 0;
1184         if (!DryRun)
1185           Token->replaceWhitespace(LineIndex, TailOffset, Split, Whitespaces);
1186         break;
1187       }
1188 
1189       // When breaking before a tab character, it may be moved by a few columns,
1190       // but will still be expanded to the next tab stop, so we don't save any
1191       // columns.
1192       if (NewRemainingTokenColumns == RemainingTokenColumns)
1193         break;
1194 
1195       assert(NewRemainingTokenColumns < RemainingTokenColumns);
1196       if (!DryRun)
1197         Token->insertBreak(LineIndex, TailOffset, Split, Whitespaces);
1198       Penalty += Current.SplitPenalty;
1199       unsigned ColumnsUsed =
1200           Token->getLineLengthAfterSplit(LineIndex, TailOffset, Split.first);
1201       if (ColumnsUsed > ColumnLimit) {
1202         Penalty += Style.PenaltyExcessCharacter * (ColumnsUsed - ColumnLimit);
1203       }
1204       TailOffset += Split.first + Split.second;
1205       RemainingTokenColumns = NewRemainingTokenColumns;
1206       BreakInserted = true;
1207     }
1208   }
1209 
1210   State.Column = RemainingTokenColumns;
1211 
1212   if (BreakInserted) {
1213     // If we break the token inside a parameter list, we need to break before
1214     // the next parameter on all levels, so that the next parameter is clearly
1215     // visible. Line comments already introduce a break.
1216     if (Current.isNot(TT_LineComment)) {
1217       for (unsigned i = 0, e = State.Stack.size(); i != e; ++i)
1218         State.Stack[i].BreakBeforeParameter = true;
1219     }
1220 
1221     Penalty += Current.isStringLiteral() ? Style.PenaltyBreakString
1222                                          : Style.PenaltyBreakComment;
1223 
1224     State.Stack.back().LastSpace = StartColumn;
1225   }
1226   return Penalty;
1227 }
1228 
1229 unsigned ContinuationIndenter::getColumnLimit(const LineState &State) const {
1230   // In preprocessor directives reserve two chars for trailing " \"
1231   return Style.ColumnLimit - (State.Line->InPPDirective ? 2 : 0);
1232 }
1233 
1234 bool ContinuationIndenter::nextIsMultilineString(const LineState &State) {
1235   const FormatToken &Current = *State.NextToken;
1236   if (!Current.isStringLiteral() || Current.is(TT_ImplicitStringLiteral))
1237     return false;
1238   // We never consider raw string literals "multiline" for the purpose of
1239   // AlwaysBreakBeforeMultilineStrings implementation as they are special-cased
1240   // (see TokenAnnotator::mustBreakBefore().
1241   if (Current.TokenText.startswith("R\""))
1242     return false;
1243   if (Current.IsMultiline)
1244     return true;
1245   if (Current.getNextNonComment() &&
1246       Current.getNextNonComment()->isStringLiteral())
1247     return true; // Implicit concatenation.
1248   if (Style.ColumnLimit != 0 &&
1249       State.Column + Current.ColumnWidth + Current.UnbreakableTailLength >
1250           Style.ColumnLimit)
1251     return true; // String will be split.
1252   return false;
1253 }
1254 
1255 } // namespace format
1256 } // namespace clang
1257