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 #include <string>
23 
24 #define DEBUG_TYPE "format-formatter"
25 
26 namespace clang {
27 namespace format {
28 
29 // Returns the length of everything up to the first possible line break after
30 // the ), ], } or > matching \c Tok.
getLengthToMatchingParen(const FormatToken & Tok)31 static unsigned getLengthToMatchingParen(const FormatToken &Tok) {
32   if (!Tok.MatchingParen)
33     return 0;
34   FormatToken *End = Tok.MatchingParen;
35   while (End->Next && !End->Next->CanBreakBefore) {
36     End = End->Next;
37   }
38   return End->TotalLength - Tok.TotalLength + 1;
39 }
40 
41 // Returns \c true if \c Tok is the "." or "->" of a call and starts the next
42 // segment of a builder type call.
startsSegmentOfBuilderTypeCall(const FormatToken & Tok)43 static bool startsSegmentOfBuilderTypeCall(const FormatToken &Tok) {
44   return Tok.isMemberAccess() && Tok.Previous && Tok.Previous->closesScope();
45 }
46 
47 // Returns \c true if \c Current starts a new parameter.
startsNextParameter(const FormatToken & Current,const FormatStyle & Style)48 static bool startsNextParameter(const FormatToken &Current,
49                                 const FormatStyle &Style) {
50   const FormatToken &Previous = *Current.Previous;
51   if (Current.is(TT_CtorInitializerComma) &&
52       Style.BreakConstructorInitializersBeforeComma)
53     return true;
54   return Previous.is(tok::comma) && !Current.isTrailingComment() &&
55          (Previous.isNot(TT_CtorInitializerComma) ||
56           !Style.BreakConstructorInitializersBeforeComma);
57 }
58 
ContinuationIndenter(const FormatStyle & Style,const AdditionalKeywords & Keywords,SourceManager & SourceMgr,WhitespaceManager & Whitespaces,encoding::Encoding Encoding,bool BinPackInconclusiveFunctions)59 ContinuationIndenter::ContinuationIndenter(const FormatStyle &Style,
60                                            const AdditionalKeywords &Keywords,
61                                            SourceManager &SourceMgr,
62                                            WhitespaceManager &Whitespaces,
63                                            encoding::Encoding Encoding,
64                                            bool BinPackInconclusiveFunctions)
65     : Style(Style), Keywords(Keywords), SourceMgr(SourceMgr),
66       Whitespaces(Whitespaces), Encoding(Encoding),
67       BinPackInconclusiveFunctions(BinPackInconclusiveFunctions),
68       CommentPragmasRegex(Style.CommentPragmas) {}
69 
getInitialState(unsigned FirstIndent,const AnnotatedLine * Line,bool DryRun)70 LineState ContinuationIndenter::getInitialState(unsigned FirstIndent,
71                                                 const AnnotatedLine *Line,
72                                                 bool DryRun) {
73   LineState State;
74   State.FirstIndent = FirstIndent;
75   State.Column = FirstIndent;
76   State.Line = Line;
77   State.NextToken = Line->First;
78   State.Stack.push_back(ParenState(FirstIndent, Line->Level, FirstIndent,
79                                    /*AvoidBinPacking=*/false,
80                                    /*NoLineBreak=*/false));
81   State.LineContainsContinuedForLoopSection = false;
82   State.StartOfStringLiteral = 0;
83   State.StartOfLineLevel = 0;
84   State.LowestLevelOnLine = 0;
85   State.IgnoreStackForComparison = false;
86 
87   // The first token has already been indented and thus consumed.
88   moveStateToNextToken(State, DryRun, /*Newline=*/false);
89   return State;
90 }
91 
canBreak(const LineState & State)92 bool ContinuationIndenter::canBreak(const LineState &State) {
93   const FormatToken &Current = *State.NextToken;
94   const FormatToken &Previous = *Current.Previous;
95   assert(&Previous == Current.Previous);
96   if (!Current.CanBreakBefore &&
97       !(State.Stack.back().BreakBeforeClosingBrace &&
98         Current.closesBlockOrBlockTypeList(Style)))
99     return false;
100   // The opening "{" of a braced list has to be on the same line as the first
101   // element if it is nested in another braced init list or function call.
102   if (!Current.MustBreakBefore && Previous.is(tok::l_brace) &&
103       Previous.isNot(TT_DictLiteral) && Previous.BlockKind == BK_BracedInit &&
104       Previous.Previous &&
105       Previous.Previous->isOneOf(tok::l_brace, tok::l_paren, tok::comma))
106     return false;
107   // This prevents breaks like:
108   //   ...
109   //   SomeParameter, OtherParameter).DoSomething(
110   //   ...
111   // As they hide "DoSomething" and are generally bad for readability.
112   if (Previous.opensScope() && Previous.isNot(tok::l_brace) &&
113       State.LowestLevelOnLine < State.StartOfLineLevel &&
114       State.LowestLevelOnLine < Current.NestingLevel)
115     return false;
116   if (Current.isMemberAccess() && State.Stack.back().ContainsUnwrappedBuilder)
117     return false;
118 
119   // Don't create a 'hanging' indent if there are multiple blocks in a single
120   // statement.
121   if (Previous.is(tok::l_brace) && State.Stack.size() > 1 &&
122       State.Stack[State.Stack.size() - 2].NestedBlockInlined &&
123       State.Stack[State.Stack.size() - 2].HasMultipleNestedBlocks)
124     return false;
125 
126   // Don't break after very short return types (e.g. "void") as that is often
127   // unexpected.
128   if (Current.is(TT_FunctionDeclarationName) && State.Column < 6) {
129     if (Style.AlwaysBreakAfterReturnType == FormatStyle::RTBS_None)
130       return false;
131   }
132 
133   return !State.Stack.back().NoLineBreak;
134 }
135 
mustBreak(const LineState & State)136 bool ContinuationIndenter::mustBreak(const LineState &State) {
137   const FormatToken &Current = *State.NextToken;
138   const FormatToken &Previous = *Current.Previous;
139   if (Current.MustBreakBefore || Current.is(TT_InlineASMColon))
140     return true;
141   if (State.Stack.back().BreakBeforeClosingBrace &&
142       Current.closesBlockOrBlockTypeList(Style))
143     return true;
144   if (Previous.is(tok::semi) && State.LineContainsContinuedForLoopSection)
145     return true;
146   if ((startsNextParameter(Current, Style) || Previous.is(tok::semi) ||
147        (Previous.is(TT_TemplateCloser) && Current.is(TT_StartOfName)) ||
148        (Style.BreakBeforeTernaryOperators && Current.is(TT_ConditionalExpr) &&
149         Previous.isNot(tok::question)) ||
150        (!Style.BreakBeforeTernaryOperators &&
151         Previous.is(TT_ConditionalExpr))) &&
152       State.Stack.back().BreakBeforeParameter && !Current.isTrailingComment() &&
153       !Current.isOneOf(tok::r_paren, tok::r_brace))
154     return true;
155   if (((Previous.is(TT_DictLiteral) && Previous.is(tok::l_brace)) ||
156        Previous.is(TT_ArrayInitializerLSquare)) &&
157       Style.ColumnLimit > 0 &&
158       getLengthToMatchingParen(Previous) + State.Column - 1 >
159           getColumnLimit(State))
160     return true;
161   if (Current.is(TT_CtorInitializerColon) &&
162       (State.Column + State.Line->Last->TotalLength - Current.TotalLength + 2 >
163            getColumnLimit(State) ||
164        State.Stack.back().BreakBeforeParameter) &&
165       ((Style.AllowShortFunctionsOnASingleLine != FormatStyle::SFS_All) ||
166        Style.BreakConstructorInitializersBeforeComma || Style.ColumnLimit != 0))
167     return true;
168   if (Current.is(TT_SelectorName) && State.Stack.back().ObjCSelectorNameFound &&
169       State.Stack.back().BreakBeforeParameter)
170     return true;
171 
172   unsigned NewLineColumn = getNewLineColumn(State);
173   if (State.Column < NewLineColumn)
174     return false;
175 
176   if (Style.AlwaysBreakBeforeMultilineStrings &&
177       (NewLineColumn == State.FirstIndent + Style.ContinuationIndentWidth ||
178        Previous.is(tok::comma) || Current.NestingLevel < 2) &&
179       !Previous.isOneOf(tok::kw_return, tok::lessless, tok::at) &&
180       !Previous.isOneOf(TT_InlineASMColon, TT_ConditionalExpr) &&
181       nextIsMultilineString(State))
182     return true;
183 
184   // Using CanBreakBefore here and below takes care of the decision whether the
185   // current style uses wrapping before or after operators for the given
186   // operator.
187   if (Previous.is(TT_BinaryOperator) && Current.CanBreakBefore) {
188     // If we need to break somewhere inside the LHS of a binary expression, we
189     // should also break after the operator. Otherwise, the formatting would
190     // hide the operator precedence, e.g. in:
191     //   if (aaaaaaaaaaaaaa ==
192     //           bbbbbbbbbbbbbb && c) {..
193     // For comparisons, we only apply this rule, if the LHS is a binary
194     // expression itself as otherwise, the line breaks seem superfluous.
195     // We need special cases for ">>" which we have split into two ">" while
196     // lexing in order to make template parsing easier.
197     bool IsComparison = (Previous.getPrecedence() == prec::Relational ||
198                          Previous.getPrecedence() == prec::Equality) &&
199                         Previous.Previous &&
200                         Previous.Previous->isNot(TT_BinaryOperator); // For >>.
201     bool LHSIsBinaryExpr =
202         Previous.Previous && Previous.Previous->EndsBinaryExpression;
203     if ((!IsComparison || LHSIsBinaryExpr) && !Current.isTrailingComment() &&
204         Previous.getPrecedence() != prec::Assignment &&
205         State.Stack.back().BreakBeforeParameter)
206       return true;
207   } else if (Current.is(TT_BinaryOperator) && Current.CanBreakBefore &&
208              State.Stack.back().BreakBeforeParameter) {
209     return true;
210   }
211 
212   // Same as above, but for the first "<<" operator.
213   if (Current.is(tok::lessless) && Current.isNot(TT_OverloadedOperator) &&
214       State.Stack.back().BreakBeforeParameter &&
215       State.Stack.back().FirstLessLess == 0)
216     return true;
217 
218   if (Current.NestingLevel == 0 && !Current.isTrailingComment()) {
219     // Always break after "template <...>" and leading annotations. This is only
220     // for cases where the entire line does not fit on a single line as a
221     // different LineFormatter would be used otherwise.
222     if (Previous.ClosesTemplateDeclaration)
223       return true;
224     if (Previous.is(TT_FunctionAnnotationRParen))
225       return true;
226     if (Previous.is(TT_LeadingJavaAnnotation) && Current.isNot(tok::l_paren) &&
227         Current.isNot(TT_LeadingJavaAnnotation))
228       return true;
229   }
230 
231   // If the return type spans multiple lines, wrap before the function name.
232   if ((Current.is(TT_FunctionDeclarationName) ||
233        (Current.is(tok::kw_operator) && !Previous.is(tok::coloncolon))) &&
234       State.Stack.back().BreakBeforeParameter)
235     return true;
236 
237   if (startsSegmentOfBuilderTypeCall(Current) &&
238       (State.Stack.back().CallContinuation != 0 ||
239        State.Stack.back().BreakBeforeParameter))
240     return true;
241 
242   // The following could be precomputed as they do not depend on the state.
243   // However, as they should take effect only if the UnwrappedLine does not fit
244   // into the ColumnLimit, they are checked here in the ContinuationIndenter.
245   if (Style.ColumnLimit != 0 && Previous.BlockKind == BK_Block &&
246       Previous.is(tok::l_brace) && !Current.isOneOf(tok::r_brace, tok::comment))
247     return true;
248 
249   if (Current.is(tok::lessless) && Previous.is(tok::identifier) &&
250       Previous.TokenText == "endl")
251     return true;
252 
253   return false;
254 }
255 
addTokenToState(LineState & State,bool Newline,bool DryRun,unsigned ExtraSpaces)256 unsigned ContinuationIndenter::addTokenToState(LineState &State, bool Newline,
257                                                bool DryRun,
258                                                unsigned ExtraSpaces) {
259   const FormatToken &Current = *State.NextToken;
260 
261   assert(!State.Stack.empty());
262   if ((Current.is(TT_ImplicitStringLiteral) &&
263        (Current.Previous->Tok.getIdentifierInfo() == nullptr ||
264         Current.Previous->Tok.getIdentifierInfo()->getPPKeywordID() ==
265             tok::pp_not_keyword))) {
266     unsigned EndColumn =
267         SourceMgr.getSpellingColumnNumber(Current.WhitespaceRange.getEnd());
268     if (Current.LastNewlineOffset != 0) {
269       // If there is a newline within this token, the final column will solely
270       // determined by the current end column.
271       State.Column = EndColumn;
272     } else {
273       unsigned StartColumn =
274           SourceMgr.getSpellingColumnNumber(Current.WhitespaceRange.getBegin());
275       assert(EndColumn >= StartColumn);
276       State.Column += EndColumn - StartColumn;
277     }
278     moveStateToNextToken(State, DryRun, /*Newline=*/false);
279     return 0;
280   }
281 
282   unsigned Penalty = 0;
283   if (Newline)
284     Penalty = addTokenOnNewLine(State, DryRun);
285   else
286     addTokenOnCurrentLine(State, DryRun, ExtraSpaces);
287 
288   return moveStateToNextToken(State, DryRun, Newline) + Penalty;
289 }
290 
addTokenOnCurrentLine(LineState & State,bool DryRun,unsigned ExtraSpaces)291 void ContinuationIndenter::addTokenOnCurrentLine(LineState &State, bool DryRun,
292                                                  unsigned ExtraSpaces) {
293   FormatToken &Current = *State.NextToken;
294   const FormatToken &Previous = *State.NextToken->Previous;
295   if (Current.is(tok::equal) &&
296       (State.Line->First->is(tok::kw_for) || Current.NestingLevel == 0) &&
297       State.Stack.back().VariablePos == 0) {
298     State.Stack.back().VariablePos = State.Column;
299     // Move over * and & if they are bound to the variable name.
300     const FormatToken *Tok = &Previous;
301     while (Tok && State.Stack.back().VariablePos >= Tok->ColumnWidth) {
302       State.Stack.back().VariablePos -= Tok->ColumnWidth;
303       if (Tok->SpacesRequiredBefore != 0)
304         break;
305       Tok = Tok->Previous;
306     }
307     if (Previous.PartOfMultiVariableDeclStmt)
308       State.Stack.back().LastSpace = State.Stack.back().VariablePos;
309   }
310 
311   unsigned Spaces = Current.SpacesRequiredBefore + ExtraSpaces;
312 
313   if (!DryRun)
314     Whitespaces.replaceWhitespace(Current, /*Newlines=*/0, /*IndentLevel=*/0,
315                                   Spaces, State.Column + Spaces);
316 
317   if (Current.is(TT_SelectorName) &&
318       !State.Stack.back().ObjCSelectorNameFound) {
319     if (Current.LongestObjCSelectorName == 0)
320       State.Stack.back().AlignColons = false;
321     else if (State.Stack.back().Indent + Current.LongestObjCSelectorName >
322              State.Column + Spaces + Current.ColumnWidth)
323       State.Stack.back().ColonPos =
324           std::max(State.FirstIndent + Style.ContinuationIndentWidth,
325                    State.Stack.back().Indent) +
326           Current.LongestObjCSelectorName;
327     else
328       State.Stack.back().ColonPos = State.Column + Spaces + Current.ColumnWidth;
329   }
330 
331   // In "AlwaysBreak" mode, enforce wrapping directly after the parenthesis by
332   // disallowing any further line breaks if there is no line break after the
333   // opening parenthesis. Don't break if it doesn't conserve columns.
334   if (Style.AlignAfterOpenBracket == FormatStyle::BAS_AlwaysBreak &&
335       Previous.is(tok::l_paren) && State.Column > getNewLineColumn(State) &&
336       (!Previous.Previous ||
337        !Previous.Previous->isOneOf(tok::kw_for, tok::kw_while, tok::kw_switch)))
338     State.Stack.back().NoLineBreak = true;
339 
340   if (Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign &&
341       Previous.opensScope() && Previous.isNot(TT_ObjCMethodExpr) &&
342       (Current.isNot(TT_LineComment) || Previous.BlockKind == BK_BracedInit))
343     State.Stack.back().Indent = State.Column + Spaces;
344   if (State.Stack.back().AvoidBinPacking && startsNextParameter(Current, Style))
345     State.Stack.back().NoLineBreak = true;
346   if (startsSegmentOfBuilderTypeCall(Current) &&
347       State.Column > getNewLineColumn(State))
348     State.Stack.back().ContainsUnwrappedBuilder = true;
349 
350   if (Current.is(TT_LambdaArrow) && Style.Language == FormatStyle::LK_Java)
351     State.Stack.back().NoLineBreak = true;
352   if (Current.isMemberAccess() && Previous.is(tok::r_paren) &&
353       (Previous.MatchingParen &&
354        (Previous.TotalLength - Previous.MatchingParen->TotalLength > 10))) {
355     // If there is a function call with long parameters, break before trailing
356     // calls. This prevents things like:
357     //   EXPECT_CALL(SomeLongParameter).Times(
358     //       2);
359     // We don't want to do this for short parameters as they can just be
360     // indexes.
361     State.Stack.back().NoLineBreak = true;
362   }
363 
364   State.Column += Spaces;
365   if (Current.isNot(tok::comment) && Previous.is(tok::l_paren) &&
366       Previous.Previous &&
367       Previous.Previous->isOneOf(tok::kw_if, tok::kw_for)) {
368     // Treat the condition inside an if as if it was a second function
369     // parameter, i.e. let nested calls have a continuation indent.
370     State.Stack.back().LastSpace = State.Column;
371     State.Stack.back().NestedBlockIndent = State.Column;
372   } else if (!Current.isOneOf(tok::comment, tok::caret) &&
373              (Previous.is(tok::comma) ||
374               (Previous.is(tok::colon) && Previous.is(TT_ObjCMethodExpr)))) {
375     State.Stack.back().LastSpace = State.Column;
376   } else if ((Previous.isOneOf(TT_BinaryOperator, TT_ConditionalExpr,
377                                TT_CtorInitializerColon)) &&
378              ((Previous.getPrecedence() != prec::Assignment &&
379                (Previous.isNot(tok::lessless) || Previous.OperatorIndex != 0 ||
380                 !Previous.LastOperator)) ||
381               Current.StartsBinaryExpression)) {
382     // Always indent relative to the RHS of the expression unless this is a
383     // simple assignment without binary expression on the RHS. Also indent
384     // relative to unary operators and the colons of constructor initializers.
385     State.Stack.back().LastSpace = State.Column;
386   } else if (Previous.is(TT_InheritanceColon)) {
387     State.Stack.back().Indent = State.Column;
388     State.Stack.back().LastSpace = State.Column;
389   } else if (Previous.opensScope()) {
390     // If a function has a trailing call, indent all parameters from the
391     // opening parenthesis. This avoids confusing indents like:
392     //   OuterFunction(InnerFunctionCall( // break
393     //       ParameterToInnerFunction))   // break
394     //       .SecondInnerFunctionCall();
395     bool HasTrailingCall = false;
396     if (Previous.MatchingParen) {
397       const FormatToken *Next = Previous.MatchingParen->getNextNonComment();
398       HasTrailingCall = Next && Next->isMemberAccess();
399     }
400     if (HasTrailingCall && State.Stack.size() > 1 &&
401         State.Stack[State.Stack.size() - 2].CallContinuation == 0)
402       State.Stack.back().LastSpace = State.Column;
403   }
404 }
405 
addTokenOnNewLine(LineState & State,bool DryRun)406 unsigned ContinuationIndenter::addTokenOnNewLine(LineState &State,
407                                                  bool DryRun) {
408   FormatToken &Current = *State.NextToken;
409   const FormatToken &Previous = *State.NextToken->Previous;
410 
411   // Extra penalty that needs to be added because of the way certain line
412   // breaks are chosen.
413   unsigned Penalty = 0;
414 
415   const FormatToken *PreviousNonComment = Current.getPreviousNonComment();
416   const FormatToken *NextNonComment = Previous.getNextNonComment();
417   if (!NextNonComment)
418     NextNonComment = &Current;
419   // The first line break on any NestingLevel causes an extra penalty in order
420   // prefer similar line breaks.
421   if (!State.Stack.back().ContainsLineBreak)
422     Penalty += 15;
423   State.Stack.back().ContainsLineBreak = true;
424 
425   Penalty += State.NextToken->SplitPenalty;
426 
427   // Breaking before the first "<<" is generally not desirable if the LHS is
428   // short. Also always add the penalty if the LHS is split over mutliple lines
429   // to avoid unnecessary line breaks that just work around this penalty.
430   if (NextNonComment->is(tok::lessless) &&
431       State.Stack.back().FirstLessLess == 0 &&
432       (State.Column <= Style.ColumnLimit / 3 ||
433        State.Stack.back().BreakBeforeParameter))
434     Penalty += Style.PenaltyBreakFirstLessLess;
435 
436   State.Column = getNewLineColumn(State);
437 
438   // Indent nested blocks relative to this column, unless in a very specific
439   // JavaScript special case where:
440   //
441   //   var loooooong_name =
442   //       function() {
443   //     // code
444   //   }
445   //
446   // is common and should be formatted like a free-standing function.
447   if (Style.Language != FormatStyle::LK_JavaScript ||
448       Current.NestingLevel != 0 || !PreviousNonComment->is(tok::equal) ||
449       !Current.is(Keywords.kw_function))
450     State.Stack.back().NestedBlockIndent = State.Column;
451 
452   if (NextNonComment->isMemberAccess()) {
453     if (State.Stack.back().CallContinuation == 0)
454       State.Stack.back().CallContinuation = State.Column;
455   } else if (NextNonComment->is(TT_SelectorName)) {
456     if (!State.Stack.back().ObjCSelectorNameFound) {
457       if (NextNonComment->LongestObjCSelectorName == 0) {
458         State.Stack.back().AlignColons = false;
459       } else {
460         State.Stack.back().ColonPos =
461             (Style.IndentWrappedFunctionNames
462                  ? std::max(State.Stack.back().Indent,
463                             State.FirstIndent + Style.ContinuationIndentWidth)
464                  : State.Stack.back().Indent) +
465             NextNonComment->LongestObjCSelectorName;
466       }
467     } else if (State.Stack.back().AlignColons &&
468                State.Stack.back().ColonPos <= NextNonComment->ColumnWidth) {
469       State.Stack.back().ColonPos = State.Column + NextNonComment->ColumnWidth;
470     }
471   } else if (PreviousNonComment && PreviousNonComment->is(tok::colon) &&
472              PreviousNonComment->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral)) {
473     // FIXME: This is hacky, find a better way. The problem is that in an ObjC
474     // method expression, the block should be aligned to the line starting it,
475     // e.g.:
476     //   [aaaaaaaaaaaaaaa aaaaaaaaa: \\ break for some reason
477     //                        ^(int *i) {
478     //                            // ...
479     //                        }];
480     // Thus, we set LastSpace of the next higher NestingLevel, to which we move
481     // when we consume all of the "}"'s FakeRParens at the "{".
482     if (State.Stack.size() > 1)
483       State.Stack[State.Stack.size() - 2].LastSpace =
484           std::max(State.Stack.back().LastSpace, State.Stack.back().Indent) +
485           Style.ContinuationIndentWidth;
486   }
487 
488   if ((Previous.isOneOf(tok::comma, tok::semi) &&
489        !State.Stack.back().AvoidBinPacking) ||
490       Previous.is(TT_BinaryOperator))
491     State.Stack.back().BreakBeforeParameter = false;
492   if (Previous.isOneOf(TT_TemplateCloser, TT_JavaAnnotation) &&
493       Current.NestingLevel == 0)
494     State.Stack.back().BreakBeforeParameter = false;
495   if (NextNonComment->is(tok::question) ||
496       (PreviousNonComment && PreviousNonComment->is(tok::question)))
497     State.Stack.back().BreakBeforeParameter = true;
498   if (Current.is(TT_BinaryOperator) && Current.CanBreakBefore)
499     State.Stack.back().BreakBeforeParameter = false;
500 
501   if (!DryRun) {
502     unsigned Newlines = std::max(
503         1u, std::min(Current.NewlinesBefore, Style.MaxEmptyLinesToKeep + 1));
504     Whitespaces.replaceWhitespace(Current, Newlines,
505                                   State.Stack.back().IndentLevel, State.Column,
506                                   State.Column, State.Line->InPPDirective);
507   }
508 
509   if (!Current.isTrailingComment())
510     State.Stack.back().LastSpace = State.Column;
511   State.StartOfLineLevel = Current.NestingLevel;
512   State.LowestLevelOnLine = Current.NestingLevel;
513 
514   // Any break on this level means that the parent level has been broken
515   // and we need to avoid bin packing there.
516   bool NestedBlockSpecialCase =
517       Style.Language != FormatStyle::LK_Cpp &&
518       Current.is(tok::r_brace) && State.Stack.size() > 1 &&
519       State.Stack[State.Stack.size() - 2].NestedBlockInlined;
520   if (!NestedBlockSpecialCase)
521     for (unsigned i = 0, e = State.Stack.size() - 1; i != e; ++i)
522       State.Stack[i].BreakBeforeParameter = true;
523 
524   if (PreviousNonComment &&
525       !PreviousNonComment->isOneOf(tok::comma, tok::semi) &&
526       (PreviousNonComment->isNot(TT_TemplateCloser) ||
527        Current.NestingLevel != 0) &&
528       !PreviousNonComment->isOneOf(
529           TT_BinaryOperator, TT_FunctionAnnotationRParen, TT_JavaAnnotation,
530           TT_LeadingJavaAnnotation) &&
531       Current.isNot(TT_BinaryOperator) && !PreviousNonComment->opensScope())
532     State.Stack.back().BreakBeforeParameter = true;
533 
534   // If we break after { or the [ of an array initializer, we should also break
535   // before the corresponding } or ].
536   if (PreviousNonComment &&
537       (PreviousNonComment->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare)))
538     State.Stack.back().BreakBeforeClosingBrace = true;
539 
540   if (State.Stack.back().AvoidBinPacking) {
541     // If we are breaking after '(', '{', '<', this is not bin packing
542     // unless AllowAllParametersOfDeclarationOnNextLine is false or this is a
543     // dict/object literal.
544     if (!Previous.isOneOf(tok::l_paren, tok::l_brace, TT_BinaryOperator) ||
545         (!Style.AllowAllParametersOfDeclarationOnNextLine &&
546          State.Line->MustBeDeclaration) ||
547         Previous.is(TT_DictLiteral))
548       State.Stack.back().BreakBeforeParameter = true;
549   }
550 
551   return Penalty;
552 }
553 
getNewLineColumn(const LineState & State)554 unsigned ContinuationIndenter::getNewLineColumn(const LineState &State) {
555   if (!State.NextToken || !State.NextToken->Previous)
556     return 0;
557   FormatToken &Current = *State.NextToken;
558   const FormatToken &Previous = *Current.Previous;
559   // If we are continuing an expression, we want to use the continuation indent.
560   unsigned ContinuationIndent =
561       std::max(State.Stack.back().LastSpace, State.Stack.back().Indent) +
562       Style.ContinuationIndentWidth;
563   const FormatToken *PreviousNonComment = Current.getPreviousNonComment();
564   const FormatToken *NextNonComment = Previous.getNextNonComment();
565   if (!NextNonComment)
566     NextNonComment = &Current;
567 
568   // Java specific bits.
569   if (Style.Language == FormatStyle::LK_Java &&
570       Current.isOneOf(Keywords.kw_implements, Keywords.kw_extends))
571     return std::max(State.Stack.back().LastSpace,
572                     State.Stack.back().Indent + Style.ContinuationIndentWidth);
573 
574   if (NextNonComment->is(tok::l_brace) && NextNonComment->BlockKind == BK_Block)
575     return Current.NestingLevel == 0 ? State.FirstIndent
576                                      : State.Stack.back().Indent;
577   if (Current.isOneOf(tok::r_brace, tok::r_square) && State.Stack.size() > 1) {
578     if (Current.closesBlockOrBlockTypeList(Style))
579       return State.Stack[State.Stack.size() - 2].NestedBlockIndent;
580     if (Current.MatchingParen &&
581         Current.MatchingParen->BlockKind == BK_BracedInit)
582       return State.Stack[State.Stack.size() - 2].LastSpace;
583     return State.FirstIndent;
584   }
585   if (Current.is(tok::identifier) && Current.Next &&
586       Current.Next->is(TT_DictLiteral))
587     return State.Stack.back().Indent;
588   if (NextNonComment->isStringLiteral() && State.StartOfStringLiteral != 0)
589     return State.StartOfStringLiteral;
590   if (NextNonComment->is(TT_ObjCStringLiteral) &&
591       State.StartOfStringLiteral != 0)
592     return State.StartOfStringLiteral - 1;
593   if (NextNonComment->is(tok::lessless) &&
594       State.Stack.back().FirstLessLess != 0)
595     return State.Stack.back().FirstLessLess;
596   if (NextNonComment->isMemberAccess()) {
597     if (State.Stack.back().CallContinuation == 0)
598       return ContinuationIndent;
599     return State.Stack.back().CallContinuation;
600   }
601   if (State.Stack.back().QuestionColumn != 0 &&
602       ((NextNonComment->is(tok::colon) &&
603         NextNonComment->is(TT_ConditionalExpr)) ||
604        Previous.is(TT_ConditionalExpr)))
605     return State.Stack.back().QuestionColumn;
606   if (Previous.is(tok::comma) && State.Stack.back().VariablePos != 0)
607     return State.Stack.back().VariablePos;
608   if ((PreviousNonComment &&
609        (PreviousNonComment->ClosesTemplateDeclaration ||
610         PreviousNonComment->isOneOf(
611             TT_AttributeParen, TT_FunctionAnnotationRParen, TT_JavaAnnotation,
612             TT_LeadingJavaAnnotation))) ||
613       (!Style.IndentWrappedFunctionNames &&
614        NextNonComment->isOneOf(tok::kw_operator, TT_FunctionDeclarationName)))
615     return std::max(State.Stack.back().LastSpace, State.Stack.back().Indent);
616   if (NextNonComment->is(TT_SelectorName)) {
617     if (!State.Stack.back().ObjCSelectorNameFound) {
618       if (NextNonComment->LongestObjCSelectorName == 0)
619         return State.Stack.back().Indent;
620       return (Style.IndentWrappedFunctionNames
621                   ? std::max(State.Stack.back().Indent,
622                              State.FirstIndent + Style.ContinuationIndentWidth)
623                   : State.Stack.back().Indent) +
624              NextNonComment->LongestObjCSelectorName -
625              NextNonComment->ColumnWidth;
626     }
627     if (!State.Stack.back().AlignColons)
628       return State.Stack.back().Indent;
629     if (State.Stack.back().ColonPos > NextNonComment->ColumnWidth)
630       return State.Stack.back().ColonPos - NextNonComment->ColumnWidth;
631     return State.Stack.back().Indent;
632   }
633   if (NextNonComment->is(TT_ArraySubscriptLSquare)) {
634     if (State.Stack.back().StartOfArraySubscripts != 0)
635       return State.Stack.back().StartOfArraySubscripts;
636     return ContinuationIndent;
637   }
638 
639   // This ensure that we correctly format ObjC methods calls without inputs,
640   // i.e. where the last element isn't selector like: [callee method];
641   if (NextNonComment->is(tok::identifier) && NextNonComment->FakeRParens == 0 &&
642       NextNonComment->Next && NextNonComment->Next->is(TT_ObjCMethodExpr))
643     return State.Stack.back().Indent;
644 
645   if (NextNonComment->isOneOf(TT_StartOfName, TT_PointerOrReference) ||
646       Previous.isOneOf(tok::coloncolon, tok::equal, TT_JsTypeColon))
647     return ContinuationIndent;
648   if (PreviousNonComment && PreviousNonComment->is(tok::colon) &&
649       PreviousNonComment->isOneOf(TT_ObjCMethodExpr, TT_DictLiteral))
650     return ContinuationIndent;
651   if (NextNonComment->is(TT_CtorInitializerColon))
652     return State.FirstIndent + Style.ConstructorInitializerIndentWidth;
653   if (NextNonComment->is(TT_CtorInitializerComma))
654     return State.Stack.back().Indent;
655   if (Previous.is(tok::r_paren) && !Current.isBinaryOperator() &&
656       !Current.isOneOf(tok::colon, tok::comment))
657     return ContinuationIndent;
658   if (State.Stack.back().Indent == State.FirstIndent && PreviousNonComment &&
659       PreviousNonComment->isNot(tok::r_brace))
660     // Ensure that we fall back to the continuation indent width instead of
661     // just flushing continuations left.
662     return State.Stack.back().Indent + Style.ContinuationIndentWidth;
663   return State.Stack.back().Indent;
664 }
665 
moveStateToNextToken(LineState & State,bool DryRun,bool Newline)666 unsigned ContinuationIndenter::moveStateToNextToken(LineState &State,
667                                                     bool DryRun, bool Newline) {
668   assert(State.Stack.size());
669   const FormatToken &Current = *State.NextToken;
670 
671   if (Current.is(TT_InheritanceColon))
672     State.Stack.back().AvoidBinPacking = true;
673   if (Current.is(tok::lessless) && Current.isNot(TT_OverloadedOperator)) {
674     if (State.Stack.back().FirstLessLess == 0)
675       State.Stack.back().FirstLessLess = State.Column;
676     else
677       State.Stack.back().LastOperatorWrapped = Newline;
678   }
679   if ((Current.is(TT_BinaryOperator) && Current.isNot(tok::lessless)) ||
680       Current.is(TT_ConditionalExpr))
681     State.Stack.back().LastOperatorWrapped = Newline;
682   if (Current.is(TT_ArraySubscriptLSquare) &&
683       State.Stack.back().StartOfArraySubscripts == 0)
684     State.Stack.back().StartOfArraySubscripts = State.Column;
685   if ((Current.is(tok::question) && Style.BreakBeforeTernaryOperators) ||
686       (Current.getPreviousNonComment() && Current.isNot(tok::colon) &&
687        Current.getPreviousNonComment()->is(tok::question) &&
688        !Style.BreakBeforeTernaryOperators))
689     State.Stack.back().QuestionColumn = State.Column;
690   if (!Current.opensScope() && !Current.closesScope())
691     State.LowestLevelOnLine =
692         std::min(State.LowestLevelOnLine, Current.NestingLevel);
693   if (Current.isMemberAccess())
694     State.Stack.back().StartOfFunctionCall =
695         Current.LastOperator ? 0 : State.Column;
696   if (Current.is(TT_SelectorName)) {
697     State.Stack.back().ObjCSelectorNameFound = true;
698     if (Style.IndentWrappedFunctionNames) {
699       State.Stack.back().Indent =
700           State.FirstIndent + Style.ContinuationIndentWidth;
701     }
702   }
703   if (Current.is(TT_CtorInitializerColon)) {
704     // Indent 2 from the column, so:
705     // SomeClass::SomeClass()
706     //     : First(...), ...
707     //       Next(...)
708     //       ^ line up here.
709     State.Stack.back().Indent =
710         State.Column + (Style.BreakConstructorInitializersBeforeComma ? 0 : 2);
711     State.Stack.back().NestedBlockIndent = State.Stack.back().Indent;
712     if (Style.ConstructorInitializerAllOnOneLineOrOnePerLine)
713       State.Stack.back().AvoidBinPacking = true;
714     State.Stack.back().BreakBeforeParameter = false;
715   }
716   if (Current.isOneOf(TT_BinaryOperator, TT_ConditionalExpr) && Newline)
717     State.Stack.back().NestedBlockIndent =
718         State.Column + Current.ColumnWidth + 1;
719 
720   // Insert scopes created by fake parenthesis.
721   const FormatToken *Previous = Current.getPreviousNonComment();
722 
723   // Add special behavior to support a format commonly used for JavaScript
724   // closures:
725   //   SomeFunction(function() {
726   //     foo();
727   //     bar();
728   //   }, a, b, c);
729   if (Current.isNot(tok::comment) && Previous &&
730       Previous->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) &&
731       State.Stack.size() > 1) {
732     if (State.Stack[State.Stack.size() - 2].NestedBlockInlined && Newline)
733       for (unsigned i = 0, e = State.Stack.size() - 1; i != e; ++i)
734         State.Stack[i].NoLineBreak = true;
735     State.Stack[State.Stack.size() - 2].NestedBlockInlined = false;
736   }
737   if (Previous && (Previous->isOneOf(tok::l_paren, tok::comma, tok::colon) ||
738                    Previous->isOneOf(TT_BinaryOperator, TT_ConditionalExpr)) &&
739       !Previous->isOneOf(TT_DictLiteral, TT_ObjCMethodExpr)) {
740     State.Stack.back().NestedBlockInlined =
741         !Newline &&
742         (Previous->isNot(tok::l_paren) || Previous->ParameterCount > 1);
743   }
744 
745   moveStatePastFakeLParens(State, Newline);
746   moveStatePastScopeOpener(State, Newline);
747   moveStatePastScopeCloser(State);
748   moveStatePastFakeRParens(State);
749 
750   if (Current.isStringLiteral() && State.StartOfStringLiteral == 0)
751     State.StartOfStringLiteral = State.Column;
752   if (Current.is(TT_ObjCStringLiteral) && State.StartOfStringLiteral == 0)
753     State.StartOfStringLiteral = State.Column + 1;
754   else if (!Current.isOneOf(tok::comment, tok::identifier, tok::hash) &&
755            !Current.isStringLiteral())
756     State.StartOfStringLiteral = 0;
757 
758   State.Column += Current.ColumnWidth;
759   State.NextToken = State.NextToken->Next;
760   unsigned Penalty = breakProtrudingToken(Current, State, DryRun);
761   if (State.Column > getColumnLimit(State)) {
762     unsigned ExcessCharacters = State.Column - getColumnLimit(State);
763     Penalty += Style.PenaltyExcessCharacter * ExcessCharacters;
764   }
765 
766   if (Current.Role)
767     Current.Role->formatFromToken(State, this, DryRun);
768   // If the previous has a special role, let it consume tokens as appropriate.
769   // It is necessary to start at the previous token for the only implemented
770   // role (comma separated list). That way, the decision whether or not to break
771   // after the "{" is already done and both options are tried and evaluated.
772   // FIXME: This is ugly, find a better way.
773   if (Previous && Previous->Role)
774     Penalty += Previous->Role->formatAfterToken(State, this, DryRun);
775 
776   return Penalty;
777 }
778 
moveStatePastFakeLParens(LineState & State,bool Newline)779 void ContinuationIndenter::moveStatePastFakeLParens(LineState &State,
780                                                     bool Newline) {
781   const FormatToken &Current = *State.NextToken;
782   const FormatToken *Previous = Current.getPreviousNonComment();
783 
784   // Don't add extra indentation for the first fake parenthesis after
785   // 'return', assignments or opening <({[. The indentation for these cases
786   // is special cased.
787   bool SkipFirstExtraIndent =
788       (Previous && (Previous->opensScope() ||
789                     Previous->isOneOf(tok::semi, tok::kw_return) ||
790                     (Previous->getPrecedence() == prec::Assignment &&
791                      Style.AlignOperands) ||
792                     Previous->is(TT_ObjCMethodExpr)));
793   for (SmallVectorImpl<prec::Level>::const_reverse_iterator
794            I = Current.FakeLParens.rbegin(),
795            E = Current.FakeLParens.rend();
796        I != E; ++I) {
797     ParenState NewParenState = State.Stack.back();
798     NewParenState.ContainsLineBreak = false;
799 
800     // Indent from 'LastSpace' unless these are fake parentheses encapsulating
801     // a builder type call after 'return' or, if the alignment after opening
802     // brackets is disabled.
803     if (!Current.isTrailingComment() &&
804         (Style.AlignOperands || *I < prec::Assignment) &&
805         (!Previous || Previous->isNot(tok::kw_return) ||
806          (Style.Language != FormatStyle::LK_Java && *I > 0)) &&
807         (Style.AlignAfterOpenBracket != FormatStyle::BAS_DontAlign ||
808          *I != prec::Comma || Current.NestingLevel == 0))
809       NewParenState.Indent =
810           std::max(std::max(State.Column, NewParenState.Indent),
811                    State.Stack.back().LastSpace);
812 
813     // Don't allow the RHS of an operator to be split over multiple lines unless
814     // there is a line-break right after the operator.
815     // Exclude relational operators, as there, it is always more desirable to
816     // have the LHS 'left' of the RHS.
817     if (Previous && Previous->getPrecedence() > prec::Assignment &&
818         Previous->isOneOf(TT_BinaryOperator, TT_ConditionalExpr) &&
819         Previous->getPrecedence() != prec::Relational) {
820       bool BreakBeforeOperator =
821           Previous->is(tok::lessless) ||
822           (Previous->is(TT_BinaryOperator) &&
823            Style.BreakBeforeBinaryOperators != FormatStyle::BOS_None) ||
824           (Previous->is(TT_ConditionalExpr) &&
825            Style.BreakBeforeTernaryOperators);
826       if ((!Newline && !BreakBeforeOperator) ||
827           (!State.Stack.back().LastOperatorWrapped && BreakBeforeOperator))
828         NewParenState.NoLineBreak = true;
829     }
830 
831     // Do not indent relative to the fake parentheses inserted for "." or "->".
832     // This is a special case to make the following to statements consistent:
833     //   OuterFunction(InnerFunctionCall( // break
834     //       ParameterToInnerFunction));
835     //   OuterFunction(SomeObject.InnerFunctionCall( // break
836     //       ParameterToInnerFunction));
837     if (*I > prec::Unknown)
838       NewParenState.LastSpace = std::max(NewParenState.LastSpace, State.Column);
839     if (*I != prec::Conditional && !Current.is(TT_UnaryOperator))
840       NewParenState.StartOfFunctionCall = State.Column;
841 
842     // Always indent conditional expressions. Never indent expression where
843     // the 'operator' is ',', ';' or an assignment (i.e. *I <=
844     // prec::Assignment) as those have different indentation rules. Indent
845     // other expression, unless the indentation needs to be skipped.
846     if (*I == prec::Conditional ||
847         (!SkipFirstExtraIndent && *I > prec::Assignment &&
848          !Current.isTrailingComment()))
849       NewParenState.Indent += Style.ContinuationIndentWidth;
850     if ((Previous && !Previous->opensScope()) || *I > prec::Comma)
851       NewParenState.BreakBeforeParameter = false;
852     State.Stack.push_back(NewParenState);
853     SkipFirstExtraIndent = false;
854   }
855 }
856 
moveStatePastFakeRParens(LineState & State)857 void ContinuationIndenter::moveStatePastFakeRParens(LineState &State) {
858   for (unsigned i = 0, e = State.NextToken->FakeRParens; i != e; ++i) {
859     unsigned VariablePos = State.Stack.back().VariablePos;
860     if (State.Stack.size() == 1) {
861       // Do not pop the last element.
862       break;
863     }
864     State.Stack.pop_back();
865     State.Stack.back().VariablePos = VariablePos;
866   }
867 }
868 
moveStatePastScopeOpener(LineState & State,bool Newline)869 void ContinuationIndenter::moveStatePastScopeOpener(LineState &State,
870                                                     bool Newline) {
871   const FormatToken &Current = *State.NextToken;
872   if (!Current.opensScope())
873     return;
874 
875   if (Current.MatchingParen && Current.BlockKind == BK_Block) {
876     moveStateToNewBlock(State);
877     return;
878   }
879 
880   unsigned NewIndent;
881   unsigned NewIndentLevel = State.Stack.back().IndentLevel;
882   unsigned LastSpace = State.Stack.back().LastSpace;
883   bool AvoidBinPacking;
884   bool BreakBeforeParameter = false;
885   unsigned NestedBlockIndent = std::max(State.Stack.back().StartOfFunctionCall,
886                                         State.Stack.back().NestedBlockIndent);
887   if (Current.isOneOf(tok::l_brace, TT_ArrayInitializerLSquare)) {
888     if (Current.opensBlockOrBlockTypeList(Style)) {
889       NewIndent = State.Stack.back().NestedBlockIndent + Style.IndentWidth;
890       NewIndent = std::min(State.Column + 2, NewIndent);
891       ++NewIndentLevel;
892     } else {
893       NewIndent = State.Stack.back().LastSpace + Style.ContinuationIndentWidth;
894     }
895     const FormatToken *NextNoComment = Current.getNextNonComment();
896     AvoidBinPacking =
897         Current.isOneOf(TT_ArrayInitializerLSquare, TT_DictLiteral) ||
898         Style.Language == FormatStyle::LK_Proto || !Style.BinPackArguments ||
899         (NextNoComment && NextNoComment->is(TT_DesignatedInitializerPeriod));
900     if (Current.ParameterCount > 1)
901       NestedBlockIndent = std::max(NestedBlockIndent, State.Column + 1);
902   } else {
903     NewIndent = Style.ContinuationIndentWidth +
904                 std::max(State.Stack.back().LastSpace,
905                          State.Stack.back().StartOfFunctionCall);
906 
907     // Ensure that different different brackets force relative alignment, e.g.:
908     // void SomeFunction(vector<  // break
909     //                       int> v);
910     // FIXME: We likely want to do this for more combinations of brackets.
911     // Verify that it is wanted for ObjC, too.
912     if (Current.Tok.getKind() == tok::less &&
913         Current.ParentBracket == tok::l_paren) {
914       NewIndent = std::max(NewIndent, State.Stack.back().Indent);
915       LastSpace = std::max(LastSpace, State.Stack.back().Indent);
916     }
917 
918     AvoidBinPacking =
919         (State.Line->MustBeDeclaration && !Style.BinPackParameters) ||
920         (!State.Line->MustBeDeclaration && !Style.BinPackArguments) ||
921         (Style.ExperimentalAutoDetectBinPacking &&
922          (Current.PackingKind == PPK_OnePerLine ||
923           (!BinPackInconclusiveFunctions &&
924            Current.PackingKind == PPK_Inconclusive)));
925     if (Current.is(TT_ObjCMethodExpr) && Current.MatchingParen) {
926       if (Style.ColumnLimit) {
927         // If this '[' opens an ObjC call, determine whether all parameters fit
928         // into one line and put one per line if they don't.
929         if (getLengthToMatchingParen(Current) + State.Column >
930             getColumnLimit(State))
931           BreakBeforeParameter = true;
932       } else {
933         // For ColumnLimit = 0, we have to figure out whether there is or has to
934         // be a line break within this call.
935         for (const FormatToken *Tok = &Current;
936              Tok && Tok != Current.MatchingParen; Tok = Tok->Next) {
937           if (Tok->MustBreakBefore ||
938               (Tok->CanBreakBefore && Tok->NewlinesBefore > 0)) {
939             BreakBeforeParameter = true;
940             break;
941           }
942         }
943       }
944     }
945   }
946   // Generally inherit NoLineBreak from the current scope to nested scope.
947   // However, don't do this for non-empty nested blocks, dict literals and
948   // array literals as these follow different indentation rules.
949   bool NoLineBreak =
950       Current.Children.empty() &&
951       !Current.isOneOf(TT_DictLiteral, TT_ArrayInitializerLSquare) &&
952       (State.Stack.back().NoLineBreak ||
953        (Current.is(TT_TemplateOpener) &&
954         State.Stack.back().ContainsUnwrappedBuilder));
955   State.Stack.push_back(ParenState(NewIndent, NewIndentLevel, LastSpace,
956                                    AvoidBinPacking, NoLineBreak));
957   State.Stack.back().NestedBlockIndent = NestedBlockIndent;
958   State.Stack.back().BreakBeforeParameter = BreakBeforeParameter;
959   State.Stack.back().HasMultipleNestedBlocks = Current.BlockParameterCount > 1;
960 }
961 
moveStatePastScopeCloser(LineState & State)962 void ContinuationIndenter::moveStatePastScopeCloser(LineState &State) {
963   const FormatToken &Current = *State.NextToken;
964   if (!Current.closesScope())
965     return;
966 
967   // If we encounter a closing ), ], } or >, we can remove a level from our
968   // stacks.
969   if (State.Stack.size() > 1 &&
970       (Current.isOneOf(tok::r_paren, tok::r_square) ||
971        (Current.is(tok::r_brace) && State.NextToken != State.Line->First) ||
972        State.NextToken->is(TT_TemplateCloser)))
973     State.Stack.pop_back();
974 
975   if (Current.is(tok::r_square)) {
976     // If this ends the array subscript expr, reset the corresponding value.
977     const FormatToken *NextNonComment = Current.getNextNonComment();
978     if (NextNonComment && NextNonComment->isNot(tok::l_square))
979       State.Stack.back().StartOfArraySubscripts = 0;
980   }
981 }
982 
moveStateToNewBlock(LineState & State)983 void ContinuationIndenter::moveStateToNewBlock(LineState &State) {
984   unsigned NestedBlockIndent = State.Stack.back().NestedBlockIndent;
985   // ObjC block sometimes follow special indentation rules.
986   unsigned NewIndent =
987       NestedBlockIndent + (State.NextToken->is(TT_ObjCBlockLBrace)
988                                ? Style.ObjCBlockIndentWidth
989                                : Style.IndentWidth);
990   State.Stack.push_back(ParenState(
991       NewIndent, /*NewIndentLevel=*/State.Stack.back().IndentLevel + 1,
992       State.Stack.back().LastSpace, /*AvoidBinPacking=*/true,
993       /*NoLineBreak=*/false));
994   State.Stack.back().NestedBlockIndent = NestedBlockIndent;
995   State.Stack.back().BreakBeforeParameter = true;
996 }
997 
addMultilineToken(const FormatToken & Current,LineState & State)998 unsigned ContinuationIndenter::addMultilineToken(const FormatToken &Current,
999                                                  LineState &State) {
1000   // Break before further function parameters on all levels.
1001   for (unsigned i = 0, e = State.Stack.size(); i != e; ++i)
1002     State.Stack[i].BreakBeforeParameter = true;
1003 
1004   unsigned ColumnsUsed = State.Column;
1005   // We can only affect layout of the first and the last line, so the penalty
1006   // for all other lines is constant, and we ignore it.
1007   State.Column = Current.LastLineColumnWidth;
1008 
1009   if (ColumnsUsed > getColumnLimit(State))
1010     return Style.PenaltyExcessCharacter * (ColumnsUsed - getColumnLimit(State));
1011   return 0;
1012 }
1013 
breakProtrudingToken(const FormatToken & Current,LineState & State,bool DryRun)1014 unsigned ContinuationIndenter::breakProtrudingToken(const FormatToken &Current,
1015                                                     LineState &State,
1016                                                     bool DryRun) {
1017   // Don't break multi-line tokens other than block comments. Instead, just
1018   // update the state.
1019   if (Current.isNot(TT_BlockComment) && Current.IsMultiline)
1020     return addMultilineToken(Current, State);
1021 
1022   // Don't break implicit string literals or import statements.
1023   if (Current.is(TT_ImplicitStringLiteral) ||
1024       State.Line->Type == LT_ImportStatement)
1025     return 0;
1026 
1027   if (!Current.isStringLiteral() && !Current.is(tok::comment))
1028     return 0;
1029 
1030   std::unique_ptr<BreakableToken> Token;
1031   unsigned StartColumn = State.Column - Current.ColumnWidth;
1032   unsigned ColumnLimit = getColumnLimit(State);
1033 
1034   if (Current.isStringLiteral()) {
1035     // FIXME: String literal breaking is currently disabled for Java and JS, as
1036     // it requires strings to be merged using "+" which we don't support.
1037     if (Style.Language == FormatStyle::LK_Java ||
1038         Style.Language == FormatStyle::LK_JavaScript)
1039       return 0;
1040 
1041     // Don't break string literals inside preprocessor directives (except for
1042     // #define directives, as their contents are stored in separate lines and
1043     // are not affected by this check).
1044     // This way we avoid breaking code with line directives and unknown
1045     // preprocessor directives that contain long string literals.
1046     if (State.Line->Type == LT_PreprocessorDirective)
1047       return 0;
1048     // Exempts unterminated string literals from line breaking. The user will
1049     // likely want to terminate the string before any line breaking is done.
1050     if (Current.IsUnterminatedLiteral)
1051       return 0;
1052 
1053     StringRef Text = Current.TokenText;
1054     StringRef Prefix;
1055     StringRef Postfix;
1056     bool IsNSStringLiteral = false;
1057     // FIXME: Handle whitespace between '_T', '(', '"..."', and ')'.
1058     // FIXME: Store Prefix and Suffix (or PrefixLength and SuffixLength to
1059     // reduce the overhead) for each FormatToken, which is a string, so that we
1060     // don't run multiple checks here on the hot path.
1061     if (Text.startswith("\"") && Current.Previous &&
1062         Current.Previous->is(tok::at)) {
1063       IsNSStringLiteral = true;
1064       Prefix = "@\"";
1065     }
1066     if ((Text.endswith(Postfix = "\"") &&
1067          (IsNSStringLiteral || Text.startswith(Prefix = "\"") ||
1068           Text.startswith(Prefix = "u\"") || Text.startswith(Prefix = "U\"") ||
1069           Text.startswith(Prefix = "u8\"") ||
1070           Text.startswith(Prefix = "L\""))) ||
1071         (Text.startswith(Prefix = "_T(\"") && Text.endswith(Postfix = "\")"))) {
1072       Token.reset(new BreakableStringLiteral(
1073           Current, State.Line->Level, StartColumn, Prefix, Postfix,
1074           State.Line->InPPDirective, Encoding, Style));
1075     } else {
1076       return 0;
1077     }
1078   } else if (Current.is(TT_BlockComment) && Current.isTrailingComment()) {
1079     if (!Style.ReflowComments ||
1080         CommentPragmasRegex.match(Current.TokenText.substr(2)))
1081       return 0;
1082     Token.reset(new BreakableBlockComment(
1083         Current, State.Line->Level, StartColumn, Current.OriginalColumn,
1084         !Current.Previous, State.Line->InPPDirective, Encoding, Style));
1085   } else if (Current.is(TT_LineComment) &&
1086              (Current.Previous == nullptr ||
1087               Current.Previous->isNot(TT_ImplicitStringLiteral))) {
1088     if (!Style.ReflowComments ||
1089         CommentPragmasRegex.match(Current.TokenText.substr(2)))
1090       return 0;
1091     Token.reset(new BreakableLineComment(Current, State.Line->Level,
1092                                          StartColumn, /*InPPDirective=*/false,
1093                                          Encoding, Style));
1094     // We don't insert backslashes when breaking line comments.
1095     ColumnLimit = Style.ColumnLimit;
1096   } else {
1097     return 0;
1098   }
1099   if (Current.UnbreakableTailLength >= ColumnLimit)
1100     return 0;
1101 
1102   unsigned RemainingSpace = ColumnLimit - Current.UnbreakableTailLength;
1103   bool BreakInserted = false;
1104   unsigned Penalty = 0;
1105   unsigned RemainingTokenColumns = 0;
1106   for (unsigned LineIndex = 0, EndIndex = Token->getLineCount();
1107        LineIndex != EndIndex; ++LineIndex) {
1108     if (!DryRun)
1109       Token->replaceWhitespaceBefore(LineIndex, Whitespaces);
1110     unsigned TailOffset = 0;
1111     RemainingTokenColumns =
1112         Token->getLineLengthAfterSplit(LineIndex, TailOffset, StringRef::npos);
1113     while (RemainingTokenColumns > RemainingSpace) {
1114       BreakableToken::Split Split =
1115           Token->getSplit(LineIndex, TailOffset, ColumnLimit);
1116       if (Split.first == StringRef::npos) {
1117         // The last line's penalty is handled in addNextStateToQueue().
1118         if (LineIndex < EndIndex - 1)
1119           Penalty += Style.PenaltyExcessCharacter *
1120                      (RemainingTokenColumns - RemainingSpace);
1121         break;
1122       }
1123       assert(Split.first != 0);
1124       unsigned NewRemainingTokenColumns = Token->getLineLengthAfterSplit(
1125           LineIndex, TailOffset + Split.first + Split.second, StringRef::npos);
1126 
1127       // We can remove extra whitespace instead of breaking the line.
1128       if (RemainingTokenColumns + 1 - Split.second <= RemainingSpace) {
1129         RemainingTokenColumns = 0;
1130         if (!DryRun)
1131           Token->replaceWhitespace(LineIndex, TailOffset, Split, Whitespaces);
1132         break;
1133       }
1134 
1135       // When breaking before a tab character, it may be moved by a few columns,
1136       // but will still be expanded to the next tab stop, so we don't save any
1137       // columns.
1138       if (NewRemainingTokenColumns == RemainingTokenColumns)
1139         break;
1140 
1141       assert(NewRemainingTokenColumns < RemainingTokenColumns);
1142       if (!DryRun)
1143         Token->insertBreak(LineIndex, TailOffset, Split, Whitespaces);
1144       Penalty += Current.SplitPenalty;
1145       unsigned ColumnsUsed =
1146           Token->getLineLengthAfterSplit(LineIndex, TailOffset, Split.first);
1147       if (ColumnsUsed > ColumnLimit) {
1148         Penalty += Style.PenaltyExcessCharacter * (ColumnsUsed - ColumnLimit);
1149       }
1150       TailOffset += Split.first + Split.second;
1151       RemainingTokenColumns = NewRemainingTokenColumns;
1152       BreakInserted = true;
1153     }
1154   }
1155 
1156   State.Column = RemainingTokenColumns;
1157 
1158   if (BreakInserted) {
1159     // If we break the token inside a parameter list, we need to break before
1160     // the next parameter on all levels, so that the next parameter is clearly
1161     // visible. Line comments already introduce a break.
1162     if (Current.isNot(TT_LineComment)) {
1163       for (unsigned i = 0, e = State.Stack.size(); i != e; ++i)
1164         State.Stack[i].BreakBeforeParameter = true;
1165     }
1166 
1167     Penalty += Current.isStringLiteral() ? Style.PenaltyBreakString
1168                                          : Style.PenaltyBreakComment;
1169 
1170     State.Stack.back().LastSpace = StartColumn;
1171   }
1172   return Penalty;
1173 }
1174 
getColumnLimit(const LineState & State) const1175 unsigned ContinuationIndenter::getColumnLimit(const LineState &State) const {
1176   // In preprocessor directives reserve two chars for trailing " \"
1177   return Style.ColumnLimit - (State.Line->InPPDirective ? 2 : 0);
1178 }
1179 
nextIsMultilineString(const LineState & State)1180 bool ContinuationIndenter::nextIsMultilineString(const LineState &State) {
1181   const FormatToken &Current = *State.NextToken;
1182   if (!Current.isStringLiteral() || Current.is(TT_ImplicitStringLiteral))
1183     return false;
1184   // We never consider raw string literals "multiline" for the purpose of
1185   // AlwaysBreakBeforeMultilineStrings implementation as they are special-cased
1186   // (see TokenAnnotator::mustBreakBefore().
1187   if (Current.TokenText.startswith("R\""))
1188     return false;
1189   if (Current.IsMultiline)
1190     return true;
1191   if (Current.getNextNonComment() &&
1192       Current.getNextNonComment()->isStringLiteral())
1193     return true; // Implicit concatenation.
1194   if (Style.ColumnLimit != 0 &&
1195       State.Column + Current.ColumnWidth + Current.UnbreakableTailLength >
1196           Style.ColumnLimit)
1197     return true; // String will be split.
1198   return false;
1199 }
1200 
1201 } // namespace format
1202 } // namespace clang
1203