ContinuationIndenter.cpp revision 966e6d3a174856ff3fe3d6cfe294ebae832f6c09
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#define DEBUG_TYPE "format-formatter" 16 17#include "BreakableToken.h" 18#include "ContinuationIndenter.h" 19#include "WhitespaceManager.h" 20#include "clang/Basic/OperatorPrecedence.h" 21#include "clang/Basic/SourceManager.h" 22#include "clang/Format/Format.h" 23#include "llvm/Support/Debug.h" 24#include <string> 25 26namespace clang { 27namespace format { 28 29// Returns the length of everything up to the first possible line break after 30// the ), ], } or > matching \c Tok. 31static unsigned getLengthToMatchingParen(const FormatToken &Tok) { 32 if (Tok.MatchingParen == NULL) 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. 43static 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. 48static bool startsNextParameter(const FormatToken &Current, 49 const FormatStyle &Style) { 50 const FormatToken &Previous = *Current.Previous; 51 if (Current.Type == TT_CtorInitializerComma && 52 Style.BreakConstructorInitializersBeforeComma) 53 return true; 54 return Previous.is(tok::comma) && !Current.isTrailingComment() && 55 (Previous.Type != TT_CtorInitializerComma || 56 !Style.BreakConstructorInitializersBeforeComma); 57} 58 59ContinuationIndenter::ContinuationIndenter(const FormatStyle &Style, 60 SourceManager &SourceMgr, 61 WhitespaceManager &Whitespaces, 62 encoding::Encoding Encoding, 63 bool BinPackInconclusiveFunctions) 64 : Style(Style), SourceMgr(SourceMgr), Whitespaces(Whitespaces), 65 Encoding(Encoding), 66 BinPackInconclusiveFunctions(BinPackInconclusiveFunctions) {} 67 68LineState ContinuationIndenter::getInitialState(unsigned FirstIndent, 69 const AnnotatedLine *Line, 70 bool DryRun) { 71 LineState State; 72 State.FirstIndent = FirstIndent; 73 State.Column = FirstIndent; 74 State.Line = Line; 75 State.NextToken = Line->First; 76 State.Stack.push_back(ParenState(FirstIndent, Line->Level, FirstIndent, 77 /*AvoidBinPacking=*/false, 78 /*NoLineBreak=*/false)); 79 State.LineContainsContinuedForLoopSection = false; 80 State.ParenLevel = 0; 81 State.StartOfStringLiteral = 0; 82 State.StartOfLineLevel = State.ParenLevel; 83 State.LowestLevelOnLine = State.ParenLevel; 84 State.IgnoreStackForComparison = false; 85 86 // The first token has already been indented and thus consumed. 87 moveStateToNextToken(State, DryRun, /*Newline=*/false); 88 return State; 89} 90 91bool ContinuationIndenter::canBreak(const LineState &State) { 92 const FormatToken &Current = *State.NextToken; 93 const FormatToken &Previous = *Current.Previous; 94 assert(&Previous == Current.Previous); 95 if (!Current.CanBreakBefore && !(State.Stack.back().BreakBeforeClosingBrace && 96 Current.closesBlockTypeList(Style))) 97 return false; 98 // The opening "{" of a braced list has to be on the same line as the first 99 // element if it is nested in another braced init list or function call. 100 if (!Current.MustBreakBefore && Previous.is(tok::l_brace) && 101 Previous.Type != TT_DictLiteral && 102 Previous.BlockKind == BK_BracedInit && Previous.Previous && 103 Previous.Previous->isOneOf(tok::l_brace, tok::l_paren, tok::comma)) 104 return false; 105 // This prevents breaks like: 106 // ... 107 // SomeParameter, OtherParameter).DoSomething( 108 // ... 109 // As they hide "DoSomething" and are generally bad for readability. 110 if (Previous.opensScope() && State.LowestLevelOnLine < State.StartOfLineLevel) 111 return false; 112 if (Current.isMemberAccess() && State.Stack.back().ContainsUnwrappedBuilder) 113 return false; 114 return !State.Stack.back().NoLineBreak; 115} 116 117bool ContinuationIndenter::mustBreak(const LineState &State) { 118 const FormatToken &Current = *State.NextToken; 119 const FormatToken &Previous = *Current.Previous; 120 if (Current.MustBreakBefore || Current.Type == TT_InlineASMColon) 121 return true; 122 if (State.Stack.back().BreakBeforeClosingBrace && 123 Current.closesBlockTypeList(Style)) 124 return true; 125 if (Previous.is(tok::semi) && State.LineContainsContinuedForLoopSection) 126 return true; 127 if ((startsNextParameter(Current, Style) || Previous.is(tok::semi) || 128 Current.is(tok::question) || 129 (Current.Type == TT_ConditionalExpr && Previous.isNot(tok::question))) && 130 State.Stack.back().BreakBeforeParameter && !Current.isTrailingComment() && 131 !Current.isOneOf(tok::r_paren, tok::r_brace)) 132 return true; 133 if (Style.AlwaysBreakBeforeMultilineStrings && 134 State.Column > State.Stack.back().Indent && // Breaking saves columns. 135 !Previous.isOneOf(tok::kw_return, tok::lessless) && 136 Previous.Type != TT_InlineASMColon && NextIsMultilineString(State)) 137 return true; 138 if (((Previous.Type == TT_DictLiteral && Previous.is(tok::l_brace)) || 139 Previous.Type == TT_ArrayInitializerLSquare) && 140 getLengthToMatchingParen(Previous) + State.Column > getColumnLimit(State)) 141 return true; 142 143 if (!Style.BreakBeforeBinaryOperators) { 144 // If we need to break somewhere inside the LHS of a binary expression, we 145 // should also break after the operator. Otherwise, the formatting would 146 // hide the operator precedence, e.g. in: 147 // if (aaaaaaaaaaaaaa == 148 // bbbbbbbbbbbbbb && c) {.. 149 // For comparisons, we only apply this rule, if the LHS is a binary 150 // expression itself as otherwise, the line breaks seem superfluous. 151 // We need special cases for ">>" which we have split into two ">" while 152 // lexing in order to make template parsing easier. 153 // 154 // FIXME: We'll need something similar for styles that break before binary 155 // operators. 156 bool IsComparison = (Previous.getPrecedence() == prec::Relational || 157 Previous.getPrecedence() == prec::Equality) && 158 Previous.Previous && 159 Previous.Previous->Type != TT_BinaryOperator; // For >>. 160 bool LHSIsBinaryExpr = 161 Previous.Previous && Previous.Previous->EndsBinaryExpression; 162 if (Previous.Type == TT_BinaryOperator && 163 (!IsComparison || LHSIsBinaryExpr) && 164 Current.Type != TT_BinaryOperator && // For >>. 165 !Current.isTrailingComment() && 166 !Previous.isOneOf(tok::lessless, tok::question) && 167 Previous.getPrecedence() != prec::Assignment && 168 State.Stack.back().BreakBeforeParameter) 169 return true; 170 } 171 172 // Same as above, but for the first "<<" operator. 173 if (Current.is(tok::lessless) && State.Stack.back().BreakBeforeParameter && 174 State.Stack.back().FirstLessLess == 0) 175 return true; 176 177 // FIXME: Comparing LongestObjCSelectorName to 0 is a hacky way of finding 178 // out whether it is the first parameter. Clean this up. 179 if (Current.Type == TT_ObjCSelectorName && 180 Current.LongestObjCSelectorName == 0 && 181 State.Stack.back().BreakBeforeParameter) 182 return true; 183 if ((Current.Type == TT_CtorInitializerColon || 184 (Previous.ClosesTemplateDeclaration && State.ParenLevel == 0 && 185 !Current.isTrailingComment()))) 186 return true; 187 188 if ((Current.Type == TT_StartOfName || Current.is(tok::kw_operator)) && 189 State.Line->MightBeFunctionDecl && 190 State.Stack.back().BreakBeforeParameter && State.ParenLevel == 0) 191 return true; 192 if (startsSegmentOfBuilderTypeCall(Current) && 193 (State.Stack.back().CallContinuation != 0 || 194 (State.Stack.back().BreakBeforeParameter && 195 State.Stack.back().ContainsUnwrappedBuilder))) 196 return true; 197 return false; 198} 199 200unsigned ContinuationIndenter::addTokenToState(LineState &State, bool Newline, 201 bool DryRun, 202 unsigned ExtraSpaces) { 203 const FormatToken &Current = *State.NextToken; 204 205 if (State.Stack.size() == 0 || 206 (Current.Type == TT_ImplicitStringLiteral && 207 (Current.Previous->Tok.getIdentifierInfo() == NULL || 208 Current.Previous->Tok.getIdentifierInfo()->getPPKeywordID() == 209 tok::pp_not_keyword))) { 210 // FIXME: Is this correct? 211 int WhitespaceLength = SourceMgr.getSpellingColumnNumber( 212 State.NextToken->WhitespaceRange.getEnd()) - 213 SourceMgr.getSpellingColumnNumber( 214 State.NextToken->WhitespaceRange.getBegin()); 215 State.Column += WhitespaceLength + State.NextToken->ColumnWidth; 216 State.NextToken = State.NextToken->Next; 217 return 0; 218 } 219 220 unsigned Penalty = 0; 221 if (Newline) 222 Penalty = addTokenOnNewLine(State, DryRun); 223 else 224 addTokenOnCurrentLine(State, DryRun, ExtraSpaces); 225 226 return moveStateToNextToken(State, DryRun, Newline) + Penalty; 227} 228 229void ContinuationIndenter::addTokenOnCurrentLine(LineState &State, bool DryRun, 230 unsigned ExtraSpaces) { 231 FormatToken &Current = *State.NextToken; 232 const FormatToken &Previous = *State.NextToken->Previous; 233 if (Current.is(tok::equal) && 234 (State.Line->First->is(tok::kw_for) || State.ParenLevel == 0) && 235 State.Stack.back().VariablePos == 0) { 236 State.Stack.back().VariablePos = State.Column; 237 // Move over * and & if they are bound to the variable name. 238 const FormatToken *Tok = &Previous; 239 while (Tok && State.Stack.back().VariablePos >= Tok->ColumnWidth) { 240 State.Stack.back().VariablePos -= Tok->ColumnWidth; 241 if (Tok->SpacesRequiredBefore != 0) 242 break; 243 Tok = Tok->Previous; 244 } 245 if (Previous.PartOfMultiVariableDeclStmt) 246 State.Stack.back().LastSpace = State.Stack.back().VariablePos; 247 } 248 249 unsigned Spaces = Current.SpacesRequiredBefore + ExtraSpaces; 250 251 if (!DryRun) 252 Whitespaces.replaceWhitespace(Current, /*Newlines=*/0, /*IndentLevel=*/0, 253 Spaces, State.Column + Spaces); 254 255 if (Current.Type == TT_ObjCSelectorName && State.Stack.back().ColonPos == 0) { 256 if (State.Stack.back().Indent + Current.LongestObjCSelectorName > 257 State.Column + Spaces + Current.ColumnWidth) 258 State.Stack.back().ColonPos = 259 State.Stack.back().Indent + Current.LongestObjCSelectorName; 260 else 261 State.Stack.back().ColonPos = State.Column + Spaces + Current.ColumnWidth; 262 } 263 264 if (Previous.opensScope() && Previous.Type != TT_ObjCMethodExpr && 265 Current.Type != TT_LineComment) 266 State.Stack.back().Indent = State.Column + Spaces; 267 if (State.Stack.back().AvoidBinPacking && startsNextParameter(Current, Style)) 268 State.Stack.back().NoLineBreak = true; 269 if (startsSegmentOfBuilderTypeCall(Current)) 270 State.Stack.back().ContainsUnwrappedBuilder = true; 271 272 State.Column += Spaces; 273 if (Current.is(tok::l_paren) && Previous.isOneOf(tok::kw_if, tok::kw_for)) 274 // Treat the condition inside an if as if it was a second function 275 // parameter, i.e. let nested calls have a continuation indent. 276 State.Stack.back().LastSpace = State.Column + 1; // 1 is length of "(". 277 else if (Previous.is(tok::comma)) 278 State.Stack.back().LastSpace = State.Column; 279 else if ((Previous.Type == TT_BinaryOperator || 280 Previous.Type == TT_ConditionalExpr || 281 Previous.Type == TT_UnaryOperator || 282 Previous.Type == TT_CtorInitializerColon) && 283 (Previous.getPrecedence() != prec::Assignment || 284 Current.StartsBinaryExpression)) 285 // Always indent relative to the RHS of the expression unless this is a 286 // simple assignment without binary expression on the RHS. Also indent 287 // relative to unary operators and the colons of constructor initializers. 288 State.Stack.back().LastSpace = State.Column; 289 else if (Previous.Type == TT_InheritanceColon) { 290 State.Stack.back().Indent = State.Column; 291 State.Stack.back().LastSpace = State.Column; 292 } else if (Previous.opensScope()) { 293 // If a function has a trailing call, indent all parameters from the 294 // opening parenthesis. This avoids confusing indents like: 295 // OuterFunction(InnerFunctionCall( // break 296 // ParameterToInnerFunction)) // break 297 // .SecondInnerFunctionCall(); 298 bool HasTrailingCall = false; 299 if (Previous.MatchingParen) { 300 const FormatToken *Next = Previous.MatchingParen->getNextNonComment(); 301 HasTrailingCall = Next && Next->isMemberAccess(); 302 } 303 if (HasTrailingCall && 304 State.Stack[State.Stack.size() - 2].CallContinuation == 0) 305 State.Stack.back().LastSpace = State.Column; 306 } 307} 308 309unsigned ContinuationIndenter::addTokenOnNewLine(LineState &State, 310 bool DryRun) { 311 FormatToken &Current = *State.NextToken; 312 const FormatToken &Previous = *State.NextToken->Previous; 313 // If we are continuing an expression, we want to use the continuation indent. 314 unsigned ContinuationIndent = 315 std::max(State.Stack.back().LastSpace, State.Stack.back().Indent) + 316 Style.ContinuationIndentWidth; 317 // Extra penalty that needs to be added because of the way certain line 318 // breaks are chosen. 319 unsigned Penalty = 0; 320 321 const FormatToken *PreviousNonComment = 322 State.NextToken->getPreviousNonComment(); 323 // The first line break on any ParenLevel causes an extra penalty in order 324 // prefer similar line breaks. 325 if (!State.Stack.back().ContainsLineBreak) 326 Penalty += 15; 327 State.Stack.back().ContainsLineBreak = true; 328 329 Penalty += State.NextToken->SplitPenalty; 330 331 // Breaking before the first "<<" is generally not desirable if the LHS is 332 // short. 333 if (Current.is(tok::lessless) && State.Stack.back().FirstLessLess == 0 && 334 State.Column <= Style.ColumnLimit / 2) 335 Penalty += Style.PenaltyBreakFirstLessLess; 336 337 if (Current.is(tok::l_brace) && Current.BlockKind == BK_Block) { 338 State.Column = State.FirstIndent; 339 } else if (Current.isOneOf(tok::r_brace, tok::r_square)) { 340 if (Current.closesBlockTypeList(Style) || 341 (Current.MatchingParen && 342 Current.MatchingParen->BlockKind == BK_BracedInit)) 343 State.Column = State.Stack[State.Stack.size() - 2].LastSpace; 344 else 345 State.Column = State.FirstIndent; 346 } else if (Current.is(tok::string_literal) && 347 State.StartOfStringLiteral != 0) { 348 State.Column = State.StartOfStringLiteral; 349 State.Stack.back().BreakBeforeParameter = true; 350 } else if (Current.is(tok::lessless) && 351 State.Stack.back().FirstLessLess != 0) { 352 State.Column = State.Stack.back().FirstLessLess; 353 } else if (Current.isMemberAccess()) { 354 if (State.Stack.back().CallContinuation == 0) { 355 State.Column = ContinuationIndent; 356 State.Stack.back().CallContinuation = State.Column; 357 } else { 358 State.Column = State.Stack.back().CallContinuation; 359 } 360 } else if (Current.Type == TT_ConditionalExpr) { 361 State.Column = State.Stack.back().QuestionColumn; 362 } else if (Previous.is(tok::comma) && State.Stack.back().VariablePos != 0) { 363 State.Column = State.Stack.back().VariablePos; 364 } else if ((PreviousNonComment && 365 PreviousNonComment->ClosesTemplateDeclaration) || 366 ((Current.Type == TT_StartOfName || 367 Current.is(tok::kw_operator)) && 368 State.ParenLevel == 0 && 369 (!Style.IndentFunctionDeclarationAfterType || 370 State.Line->StartsDefinition))) { 371 State.Column = State.Stack.back().Indent; 372 } else if (Current.Type == TT_ObjCSelectorName) { 373 if (State.Stack.back().ColonPos > Current.ColumnWidth) { 374 State.Column = State.Stack.back().ColonPos - Current.ColumnWidth; 375 } else { 376 State.Column = State.Stack.back().Indent; 377 State.Stack.back().ColonPos = State.Column + Current.ColumnWidth; 378 } 379 } else if (Current.Type == TT_ArraySubscriptLSquare) { 380 if (State.Stack.back().StartOfArraySubscripts != 0) 381 State.Column = State.Stack.back().StartOfArraySubscripts; 382 else 383 State.Column = ContinuationIndent; 384 } else if (Current.Type == TT_StartOfName || 385 Previous.isOneOf(tok::coloncolon, tok::equal) || 386 Previous.Type == TT_ObjCMethodExpr) { 387 State.Column = ContinuationIndent; 388 } else if (Current.Type == TT_CtorInitializerColon) { 389 State.Column = State.FirstIndent + Style.ConstructorInitializerIndentWidth; 390 } else if (Current.Type == TT_CtorInitializerComma) { 391 State.Column = State.Stack.back().Indent; 392 } else { 393 State.Column = State.Stack.back().Indent; 394 // Ensure that we fall back to the continuation indent width instead of just 395 // flushing continuations left. 396 if (State.Column == State.FirstIndent && 397 PreviousNonComment->isNot(tok::r_brace)) 398 State.Column += Style.ContinuationIndentWidth; 399 } 400 401 if (Current.is(tok::question)) 402 State.Stack.back().BreakBeforeParameter = true; 403 if ((Previous.isOneOf(tok::comma, tok::semi) && 404 !State.Stack.back().AvoidBinPacking) || 405 Previous.Type == TT_BinaryOperator) 406 State.Stack.back().BreakBeforeParameter = false; 407 if (Previous.Type == TT_TemplateCloser && State.ParenLevel == 0) 408 State.Stack.back().BreakBeforeParameter = false; 409 410 if (!DryRun) { 411 unsigned Newlines = 1; 412 if (Current.is(tok::comment)) 413 Newlines = std::max(Newlines, std::min(Current.NewlinesBefore, 414 Style.MaxEmptyLinesToKeep + 1)); 415 Whitespaces.replaceWhitespace(Current, Newlines, 416 State.Stack.back().IndentLevel, State.Column, 417 State.Column, State.Line->InPPDirective); 418 } 419 420 if (!Current.isTrailingComment()) 421 State.Stack.back().LastSpace = State.Column; 422 if (Current.isMemberAccess()) 423 State.Stack.back().LastSpace += Current.ColumnWidth; 424 State.StartOfLineLevel = State.ParenLevel; 425 State.LowestLevelOnLine = State.ParenLevel; 426 427 // Any break on this level means that the parent level has been broken 428 // and we need to avoid bin packing there. 429 for (unsigned i = 0, e = State.Stack.size() - 1; i != e; ++i) { 430 State.Stack[i].BreakBeforeParameter = true; 431 } 432 const FormatToken *TokenBefore = Current.getPreviousNonComment(); 433 if (TokenBefore && !TokenBefore->isOneOf(tok::comma, tok::semi) && 434 TokenBefore->Type != TT_TemplateCloser && 435 TokenBefore->Type != TT_BinaryOperator && !TokenBefore->opensScope()) 436 State.Stack.back().BreakBeforeParameter = true; 437 438 // If we break after { or the [ of an array initializer, we should also break 439 // before the corresponding } or ]. 440 if (Previous.is(tok::l_brace) || Previous.Type == TT_ArrayInitializerLSquare) 441 State.Stack.back().BreakBeforeClosingBrace = true; 442 443 if (State.Stack.back().AvoidBinPacking) { 444 // If we are breaking after '(', '{', '<', this is not bin packing 445 // unless AllowAllParametersOfDeclarationOnNextLine is false. 446 if (!(Previous.isOneOf(tok::l_paren, tok::l_brace) || 447 Previous.Type == TT_BinaryOperator) || 448 (!Style.AllowAllParametersOfDeclarationOnNextLine && 449 State.Line->MustBeDeclaration)) 450 State.Stack.back().BreakBeforeParameter = true; 451 } 452 453 return Penalty; 454} 455 456unsigned ContinuationIndenter::moveStateToNextToken(LineState &State, 457 bool DryRun, bool Newline) { 458 const FormatToken &Current = *State.NextToken; 459 assert(State.Stack.size()); 460 461 if (Current.Type == TT_InheritanceColon) 462 State.Stack.back().AvoidBinPacking = true; 463 if (Current.is(tok::lessless) && State.Stack.back().FirstLessLess == 0) 464 State.Stack.back().FirstLessLess = State.Column; 465 if (Current.Type == TT_ArraySubscriptLSquare && 466 State.Stack.back().StartOfArraySubscripts == 0) 467 State.Stack.back().StartOfArraySubscripts = State.Column; 468 if (Current.is(tok::question)) 469 State.Stack.back().QuestionColumn = State.Column; 470 if (!Current.opensScope() && !Current.closesScope()) 471 State.LowestLevelOnLine = 472 std::min(State.LowestLevelOnLine, State.ParenLevel); 473 if (Current.isMemberAccess()) 474 State.Stack.back().StartOfFunctionCall = 475 Current.LastInChainOfCalls ? 0 : State.Column + Current.ColumnWidth; 476 if (Current.Type == TT_CtorInitializerColon) { 477 // Indent 2 from the column, so: 478 // SomeClass::SomeClass() 479 // : First(...), ... 480 // Next(...) 481 // ^ line up here. 482 State.Stack.back().Indent = 483 State.Column + (Style.BreakConstructorInitializersBeforeComma ? 0 : 2); 484 if (Style.ConstructorInitializerAllOnOneLineOrOnePerLine) 485 State.Stack.back().AvoidBinPacking = true; 486 State.Stack.back().BreakBeforeParameter = false; 487 } 488 489 // In ObjC method declaration we align on the ":" of parameters, but we need 490 // to ensure that we indent parameters on subsequent lines by at least our 491 // continuation indent width. 492 if (Current.Type == TT_ObjCMethodSpecifier) 493 State.Stack.back().Indent += Style.ContinuationIndentWidth; 494 495 // Insert scopes created by fake parenthesis. 496 const FormatToken *Previous = Current.getPreviousNonComment(); 497 // Don't add extra indentation for the first fake parenthesis after 498 // 'return', assignements or opening <({[. The indentation for these cases 499 // is special cased. 500 bool SkipFirstExtraIndent = 501 (Previous && (Previous->opensScope() || Previous->is(tok::kw_return) || 502 Previous->getPrecedence() == prec::Assignment || 503 Previous->Type == TT_ObjCMethodExpr)); 504 for (SmallVectorImpl<prec::Level>::const_reverse_iterator 505 I = Current.FakeLParens.rbegin(), 506 E = Current.FakeLParens.rend(); 507 I != E; ++I) { 508 ParenState NewParenState = State.Stack.back(); 509 NewParenState.ContainsLineBreak = false; 510 511 // Indent from 'LastSpace' unless this the fake parentheses encapsulating a 512 // builder type call after 'return'. If such a call is line-wrapped, we 513 // commonly just want to indent from the start of the line. 514 if (!Previous || Previous->isNot(tok::kw_return) || *I > 0) 515 NewParenState.Indent = 516 std::max(std::max(State.Column, NewParenState.Indent), 517 State.Stack.back().LastSpace); 518 519 // Do not indent relative to the fake parentheses inserted for "." or "->". 520 // This is a special case to make the following to statements consistent: 521 // OuterFunction(InnerFunctionCall( // break 522 // ParameterToInnerFunction)); 523 // OuterFunction(SomeObject.InnerFunctionCall( // break 524 // ParameterToInnerFunction)); 525 if (*I > prec::Unknown) 526 NewParenState.LastSpace = std::max(NewParenState.LastSpace, State.Column); 527 528 // Always indent conditional expressions. Never indent expression where 529 // the 'operator' is ',', ';' or an assignment (i.e. *I <= 530 // prec::Assignment) as those have different indentation rules. Indent 531 // other expression, unless the indentation needs to be skipped. 532 if (*I == prec::Conditional || 533 (!SkipFirstExtraIndent && *I > prec::Assignment && 534 !Style.BreakBeforeBinaryOperators)) 535 NewParenState.Indent += Style.ContinuationIndentWidth; 536 if ((Previous && !Previous->opensScope()) || *I > prec::Comma) 537 NewParenState.BreakBeforeParameter = false; 538 State.Stack.push_back(NewParenState); 539 SkipFirstExtraIndent = false; 540 } 541 542 // If we encounter an opening (, [, { or <, we add a level to our stacks to 543 // prepare for the following tokens. 544 if (Current.opensScope()) { 545 unsigned NewIndent; 546 unsigned NewIndentLevel = State.Stack.back().IndentLevel; 547 bool AvoidBinPacking; 548 bool BreakBeforeParameter = false; 549 if (Current.is(tok::l_brace) || 550 Current.Type == TT_ArrayInitializerLSquare) { 551 if (Current.MatchingParen && Current.BlockKind == BK_Block) { 552 // If this is an l_brace starting a nested block, we pretend (wrt. to 553 // indentation) that we already consumed the corresponding r_brace. 554 // Thus, we remove all ParenStates caused bake fake parentheses that end 555 // at the r_brace. The net effect of this is that we don't indent 556 // relative to the l_brace, if the nested block is the last parameter of 557 // a function. For example, this formats: 558 // 559 // SomeFunction(a, [] { 560 // f(); // break 561 // }); 562 // 563 // instead of: 564 // SomeFunction(a, [] { 565 // f(); // break 566 // }); 567 for (unsigned i = 0; i != Current.MatchingParen->FakeRParens; ++i) 568 State.Stack.pop_back(); 569 NewIndent = State.Stack.back().LastSpace + Style.IndentWidth; 570 ++NewIndentLevel; 571 BreakBeforeParameter = true; 572 } else { 573 NewIndent = State.Stack.back().LastSpace; 574 if (Current.opensBlockTypeList(Style)) { 575 NewIndent += Style.IndentWidth; 576 ++NewIndentLevel; 577 } else { 578 NewIndent += Style.ContinuationIndentWidth; 579 } 580 } 581 const FormatToken *NextNoComment = Current.getNextNonComment(); 582 AvoidBinPacking = Current.BlockKind == BK_Block || 583 Current.Type == TT_ArrayInitializerLSquare || 584 Current.Type == TT_DictLiteral || 585 (NextNoComment && 586 NextNoComment->Type == TT_DesignatedInitializerPeriod); 587 } else { 588 NewIndent = Style.ContinuationIndentWidth + 589 std::max(State.Stack.back().LastSpace, 590 State.Stack.back().StartOfFunctionCall); 591 AvoidBinPacking = !Style.BinPackParameters || 592 (Style.ExperimentalAutoDetectBinPacking && 593 (Current.PackingKind == PPK_OnePerLine || 594 (!BinPackInconclusiveFunctions && 595 Current.PackingKind == PPK_Inconclusive))); 596 // If this '[' opens an ObjC call, determine whether all parameters fit 597 // into one line and put one per line if they don't. 598 if (Current.Type == TT_ObjCMethodExpr && 599 getLengthToMatchingParen(Current) + State.Column > 600 getColumnLimit(State)) 601 BreakBeforeParameter = true; 602 } 603 604 bool NoLineBreak = State.Stack.back().NoLineBreak || 605 (Current.Type == TT_TemplateOpener && 606 State.Stack.back().ContainsUnwrappedBuilder); 607 State.Stack.push_back(ParenState(NewIndent, NewIndentLevel, 608 State.Stack.back().LastSpace, 609 AvoidBinPacking, NoLineBreak)); 610 State.Stack.back().BreakBeforeParameter = BreakBeforeParameter; 611 ++State.ParenLevel; 612 } 613 614 // If we encounter a closing ), ], } or >, we can remove a level from our 615 // stacks. 616 if (State.Stack.size() > 1 && 617 (Current.isOneOf(tok::r_paren, tok::r_square) || 618 (Current.is(tok::r_brace) && State.NextToken != State.Line->First) || 619 State.NextToken->Type == TT_TemplateCloser)) { 620 State.Stack.pop_back(); 621 --State.ParenLevel; 622 } 623 if (Current.is(tok::r_square)) { 624 // If this ends the array subscript expr, reset the corresponding value. 625 const FormatToken *NextNonComment = Current.getNextNonComment(); 626 if (NextNonComment && NextNonComment->isNot(tok::l_square)) 627 State.Stack.back().StartOfArraySubscripts = 0; 628 } 629 630 // Remove scopes created by fake parenthesis. 631 if (Current.isNot(tok::r_brace) || 632 (Current.MatchingParen && Current.MatchingParen->BlockKind != BK_Block)) { 633 // Don't remove FakeRParens attached to r_braces that surround nested blocks 634 // as they will have been removed early (see above). 635 for (unsigned i = 0, e = Current.FakeRParens; i != e; ++i) { 636 unsigned VariablePos = State.Stack.back().VariablePos; 637 State.Stack.pop_back(); 638 State.Stack.back().VariablePos = VariablePos; 639 } 640 } 641 642 if (Current.is(tok::string_literal) && State.StartOfStringLiteral == 0) { 643 State.StartOfStringLiteral = State.Column; 644 } else if (!Current.isOneOf(tok::comment, tok::identifier, tok::hash, 645 tok::string_literal)) { 646 State.StartOfStringLiteral = 0; 647 } 648 649 State.Column += Current.ColumnWidth; 650 State.NextToken = State.NextToken->Next; 651 unsigned Penalty = breakProtrudingToken(Current, State, DryRun); 652 if (State.Column > getColumnLimit(State)) { 653 unsigned ExcessCharacters = State.Column - getColumnLimit(State); 654 Penalty += Style.PenaltyExcessCharacter * ExcessCharacters; 655 } 656 657 // If the previous has a special role, let it consume tokens as appropriate. 658 // It is necessary to start at the previous token for the only implemented 659 // role (comma separated list). That way, the decision whether or not to break 660 // after the "{" is already done and both options are tried and evaluated. 661 // FIXME: This is ugly, find a better way. 662 if (Previous && Previous->Role) 663 Penalty += Previous->Role->format(State, this, DryRun); 664 665 return Penalty; 666} 667 668unsigned ContinuationIndenter::addMultilineToken(const FormatToken &Current, 669 LineState &State) { 670 // Break before further function parameters on all levels. 671 for (unsigned i = 0, e = State.Stack.size(); i != e; ++i) 672 State.Stack[i].BreakBeforeParameter = true; 673 674 unsigned ColumnsUsed = State.Column; 675 // We can only affect layout of the first and the last line, so the penalty 676 // for all other lines is constant, and we ignore it. 677 State.Column = Current.LastLineColumnWidth; 678 679 if (ColumnsUsed > getColumnLimit(State)) 680 return Style.PenaltyExcessCharacter * (ColumnsUsed - getColumnLimit(State)); 681 return 0; 682} 683 684static bool getRawStringLiteralPrefixPostfix(StringRef Text, 685 StringRef &Prefix, 686 StringRef &Postfix) { 687 if (Text.startswith(Prefix = "R\"") || Text.startswith(Prefix = "uR\"") || 688 Text.startswith(Prefix = "UR\"") || Text.startswith(Prefix = "u8R\"") || 689 Text.startswith(Prefix = "LR\"")) { 690 size_t ParenPos = Text.find('('); 691 if (ParenPos != StringRef::npos) { 692 StringRef Delimiter = 693 Text.substr(Prefix.size(), ParenPos - Prefix.size()); 694 Prefix = Text.substr(0, ParenPos + 1); 695 Postfix = Text.substr(Text.size() - 2 - Delimiter.size()); 696 return Postfix.front() == ')' && Postfix.back() == '"' && 697 Postfix.substr(1).startswith(Delimiter); 698 } 699 } 700 return false; 701} 702 703unsigned ContinuationIndenter::breakProtrudingToken(const FormatToken &Current, 704 LineState &State, 705 bool DryRun) { 706 // Don't break multi-line tokens other than block comments. Instead, just 707 // update the state. 708 if (Current.Type != TT_BlockComment && Current.IsMultiline) 709 return addMultilineToken(Current, State); 710 711 // Don't break implicit string literals. 712 if (Current.Type == TT_ImplicitStringLiteral) 713 return 0; 714 715 if (!Current.isOneOf(tok::string_literal, tok::wide_string_literal, 716 tok::utf8_string_literal, tok::utf16_string_literal, 717 tok::utf32_string_literal, tok::comment)) 718 return 0; 719 720 llvm::OwningPtr<BreakableToken> Token; 721 unsigned StartColumn = State.Column - Current.ColumnWidth; 722 723 if (Current.isOneOf(tok::string_literal, tok::wide_string_literal, 724 tok::utf8_string_literal, tok::utf16_string_literal, 725 tok::utf32_string_literal) && 726 Current.Type != TT_ImplicitStringLiteral) { 727 // Don't break string literals inside preprocessor directives (except for 728 // #define directives, as their contents are stored in separate lines and 729 // are not affected by this check). 730 // This way we avoid breaking code with line directives and unknown 731 // preprocessor directives that contain long string literals. 732 if (State.Line->Type == LT_PreprocessorDirective) 733 return 0; 734 // Exempts unterminated string literals from line breaking. The user will 735 // likely want to terminate the string before any line breaking is done. 736 if (Current.IsUnterminatedLiteral) 737 return 0; 738 739 StringRef Text = Current.TokenText; 740 StringRef Prefix; 741 StringRef Postfix; 742 // FIXME: Handle whitespace between '_T', '(', '"..."', and ')'. 743 // FIXME: Store Prefix and Suffix (or PrefixLength and SuffixLength to 744 // reduce the overhead) for each FormatToken, which is a string, so that we 745 // don't run multiple checks here on the hot path. 746 if ((Text.endswith(Postfix = "\"") && 747 (Text.startswith(Prefix = "\"") || Text.startswith(Prefix = "u\"") || 748 Text.startswith(Prefix = "U\"") || Text.startswith(Prefix = "u8\"") || 749 Text.startswith(Prefix = "L\""))) || 750 (Text.startswith(Prefix = "_T(\"") && Text.endswith(Postfix = "\")")) || 751 getRawStringLiteralPrefixPostfix(Text, Prefix, Postfix)) { 752 Token.reset(new BreakableStringLiteral( 753 Current, State.Line->Level, StartColumn, Prefix, Postfix, 754 State.Line->InPPDirective, Encoding, Style)); 755 } else { 756 return 0; 757 } 758 } else if (Current.Type == TT_BlockComment && Current.isTrailingComment()) { 759 Token.reset(new BreakableBlockComment( 760 Current, State.Line->Level, StartColumn, Current.OriginalColumn, 761 !Current.Previous, State.Line->InPPDirective, Encoding, Style)); 762 } else if (Current.Type == TT_LineComment && 763 (Current.Previous == NULL || 764 Current.Previous->Type != TT_ImplicitStringLiteral)) { 765 Token.reset(new BreakableLineComment(Current, State.Line->Level, 766 StartColumn, State.Line->InPPDirective, 767 Encoding, Style)); 768 } else { 769 return 0; 770 } 771 if (Current.UnbreakableTailLength >= getColumnLimit(State)) 772 return 0; 773 774 unsigned RemainingSpace = 775 getColumnLimit(State) - Current.UnbreakableTailLength; 776 bool BreakInserted = false; 777 unsigned Penalty = 0; 778 unsigned RemainingTokenColumns = 0; 779 for (unsigned LineIndex = 0, EndIndex = Token->getLineCount(); 780 LineIndex != EndIndex; ++LineIndex) { 781 if (!DryRun) 782 Token->replaceWhitespaceBefore(LineIndex, Whitespaces); 783 unsigned TailOffset = 0; 784 RemainingTokenColumns = 785 Token->getLineLengthAfterSplit(LineIndex, TailOffset, StringRef::npos); 786 while (RemainingTokenColumns > RemainingSpace) { 787 BreakableToken::Split Split = 788 Token->getSplit(LineIndex, TailOffset, getColumnLimit(State)); 789 if (Split.first == StringRef::npos) { 790 // The last line's penalty is handled in addNextStateToQueue(). 791 if (LineIndex < EndIndex - 1) 792 Penalty += Style.PenaltyExcessCharacter * 793 (RemainingTokenColumns - RemainingSpace); 794 break; 795 } 796 assert(Split.first != 0); 797 unsigned NewRemainingTokenColumns = Token->getLineLengthAfterSplit( 798 LineIndex, TailOffset + Split.first + Split.second, StringRef::npos); 799 assert(NewRemainingTokenColumns < RemainingTokenColumns); 800 if (!DryRun) 801 Token->insertBreak(LineIndex, TailOffset, Split, Whitespaces); 802 Penalty += Current.SplitPenalty; 803 unsigned ColumnsUsed = 804 Token->getLineLengthAfterSplit(LineIndex, TailOffset, Split.first); 805 if (ColumnsUsed > getColumnLimit(State)) { 806 Penalty += Style.PenaltyExcessCharacter * 807 (ColumnsUsed - getColumnLimit(State)); 808 } 809 TailOffset += Split.first + Split.second; 810 RemainingTokenColumns = NewRemainingTokenColumns; 811 BreakInserted = true; 812 } 813 } 814 815 State.Column = RemainingTokenColumns; 816 817 if (BreakInserted) { 818 // If we break the token inside a parameter list, we need to break before 819 // the next parameter on all levels, so that the next parameter is clearly 820 // visible. Line comments already introduce a break. 821 if (Current.Type != TT_LineComment) { 822 for (unsigned i = 0, e = State.Stack.size(); i != e; ++i) 823 State.Stack[i].BreakBeforeParameter = true; 824 } 825 826 Penalty += Current.is(tok::string_literal) ? Style.PenaltyBreakString 827 : Style.PenaltyBreakComment; 828 829 State.Stack.back().LastSpace = StartColumn; 830 } 831 return Penalty; 832} 833 834unsigned ContinuationIndenter::getColumnLimit(const LineState &State) const { 835 // In preprocessor directives reserve two chars for trailing " \" 836 return Style.ColumnLimit - (State.Line->InPPDirective ? 2 : 0); 837} 838 839bool ContinuationIndenter::NextIsMultilineString(const LineState &State) { 840 const FormatToken &Current = *State.NextToken; 841 if (!Current.is(tok::string_literal)) 842 return false; 843 // We never consider raw string literals "multiline" for the purpose of 844 // AlwaysBreakBeforeMultilineStrings implementation. 845 if (Current.TokenText.startswith("R\"")) 846 return false; 847 if (Current.IsMultiline) 848 return true; 849 if (Current.getNextNonComment() && 850 Current.getNextNonComment()->is(tok::string_literal)) 851 return true; // Implicit concatenation. 852 if (State.Column + Current.ColumnWidth + Current.UnbreakableTailLength > 853 Style.ColumnLimit) 854 return true; // String will be split. 855 return false; 856} 857 858} // namespace format 859} // namespace clang 860