| File: | clang/lib/Format/FormatToken.cpp |
| Warning: | line 193, column 12 Access to field 'HasUnescapedNewline' results in a dereference of a null pointer (loaded from variable 'ItemBegin') |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
| 1 | //===--- FormatToken.cpp - Format C++ code --------------------------------===// | ||||||||
| 2 | // | ||||||||
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. | ||||||||
| 4 | // See https://llvm.org/LICENSE.txt for license information. | ||||||||
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception | ||||||||
| 6 | // | ||||||||
| 7 | //===----------------------------------------------------------------------===// | ||||||||
| 8 | /// | ||||||||
| 9 | /// \file | ||||||||
| 10 | /// This file implements specific functions of \c FormatTokens and their | ||||||||
| 11 | /// roles. | ||||||||
| 12 | /// | ||||||||
| 13 | //===----------------------------------------------------------------------===// | ||||||||
| 14 | |||||||||
| 15 | #include "FormatToken.h" | ||||||||
| 16 | #include "ContinuationIndenter.h" | ||||||||
| 17 | #include "llvm/ADT/SmallVector.h" | ||||||||
| 18 | #include "llvm/Support/Debug.h" | ||||||||
| 19 | #include <climits> | ||||||||
| 20 | |||||||||
| 21 | namespace clang { | ||||||||
| 22 | namespace format { | ||||||||
| 23 | |||||||||
| 24 | const char *getTokenTypeName(TokenType Type) { | ||||||||
| 25 | static const char *const TokNames[] = { | ||||||||
| 26 | #define TYPE(X) #X, | ||||||||
| 27 | LIST_TOKEN_TYPESTYPE(ArrayInitializerLSquare) TYPE(ArraySubscriptLSquare) TYPE (AttributeColon) TYPE(AttributeMacro) TYPE(AttributeParen) TYPE (AttributeSquare) TYPE(BinaryOperator) TYPE(BitFieldColon) TYPE (BlockComment) TYPE(CastRParen) TYPE(ConditionalExpr) TYPE(ConflictAlternative ) TYPE(ConflictEnd) TYPE(ConflictStart) TYPE(ConstraintJunctions ) TYPE(CtorInitializerColon) TYPE(CtorInitializerComma) TYPE( DesignatedInitializerLSquare) TYPE(DesignatedInitializerPeriod ) TYPE(DictLiteral) TYPE(FatArrow) TYPE(ForEachMacro) TYPE(FunctionAnnotationRParen ) TYPE(FunctionDeclarationName) TYPE(FunctionLBrace) TYPE(FunctionTypeLParen ) TYPE(IfMacro) TYPE(ImplicitStringLiteral) TYPE(InheritanceColon ) TYPE(InheritanceComma) TYPE(InlineASMBrace) TYPE(InlineASMColon ) TYPE(InlineASMSymbolicNameLSquare) TYPE(JavaAnnotation) TYPE (JsComputedPropertyName) TYPE(JsExponentiation) TYPE(JsExponentiationEqual ) TYPE(JsPipePipeEqual) TYPE(JsPrivateIdentifier) TYPE(JsTypeColon ) TYPE(JsTypeOperator) TYPE(JsTypeOptionalQuestion) TYPE(JsAndAndEqual ) TYPE(LambdaArrow) TYPE(LambdaLBrace) TYPE(LambdaLSquare) TYPE (LeadingJavaAnnotation) TYPE(LineComment) TYPE(MacroBlockBegin ) TYPE(MacroBlockEnd) TYPE(ModulePartitionColon) TYPE(NamespaceMacro ) TYPE(NonNullAssertion) TYPE(NullCoalescingEqual) TYPE(NullCoalescingOperator ) TYPE(NullPropagatingOperator) TYPE(ObjCBlockLBrace) TYPE(ObjCBlockLParen ) TYPE(ObjCDecl) TYPE(ObjCForIn) TYPE(ObjCMethodExpr) TYPE(ObjCMethodSpecifier ) TYPE(ObjCProperty) TYPE(ObjCStringLiteral) TYPE(OverloadedOperator ) TYPE(OverloadedOperatorLParen) TYPE(PointerOrReference) TYPE (PureVirtualSpecifier) TYPE(RangeBasedForLoopColon) TYPE(RegexLiteral ) TYPE(SelectorName) TYPE(StartOfName) TYPE(StatementAttributeLikeMacro ) TYPE(StatementMacro) TYPE(StructuredBindingLSquare) TYPE(TemplateCloser ) TYPE(TemplateOpener) TYPE(TemplateString) TYPE(ProtoExtensionLSquare ) TYPE(TrailingAnnotation) TYPE(TrailingReturnArrow) TYPE(TrailingUnaryOperator ) TYPE(TypeDeclarationParen) TYPE(TypenameMacro) TYPE(UnaryOperator ) TYPE(UntouchableMacroFunc) TYPE(CSharpStringLiteral) TYPE(CSharpNamedArgumentColon ) TYPE(CSharpNullable) TYPE(CSharpNullConditionalLSquare) TYPE (CSharpGenericTypeConstraint) TYPE(CSharpGenericTypeConstraintColon ) TYPE(CSharpGenericTypeConstraintComma) TYPE(Unknown) | ||||||||
| 28 | #undef TYPE | ||||||||
| 29 | nullptr}; | ||||||||
| 30 | |||||||||
| 31 | if (Type < NUM_TOKEN_TYPES) | ||||||||
| 32 | return TokNames[Type]; | ||||||||
| 33 | llvm_unreachable("unknown TokenType")::llvm::llvm_unreachable_internal("unknown TokenType", "clang/lib/Format/FormatToken.cpp" , 33); | ||||||||
| 34 | return nullptr; | ||||||||
| 35 | } | ||||||||
| 36 | |||||||||
| 37 | // FIXME: This is copy&pasted from Sema. Put it in a common place and remove | ||||||||
| 38 | // duplication. | ||||||||
| 39 | bool FormatToken::isSimpleTypeSpecifier() const { | ||||||||
| 40 | switch (Tok.getKind()) { | ||||||||
| 41 | case tok::kw_short: | ||||||||
| 42 | case tok::kw_long: | ||||||||
| 43 | case tok::kw___int64: | ||||||||
| 44 | case tok::kw___int128: | ||||||||
| 45 | case tok::kw_signed: | ||||||||
| 46 | case tok::kw_unsigned: | ||||||||
| 47 | case tok::kw_void: | ||||||||
| 48 | case tok::kw_char: | ||||||||
| 49 | case tok::kw_int: | ||||||||
| 50 | case tok::kw_half: | ||||||||
| 51 | case tok::kw_float: | ||||||||
| 52 | case tok::kw_double: | ||||||||
| 53 | case tok::kw___bf16: | ||||||||
| 54 | case tok::kw__Float16: | ||||||||
| 55 | case tok::kw___float128: | ||||||||
| 56 | case tok::kw___ibm128: | ||||||||
| 57 | case tok::kw_wchar_t: | ||||||||
| 58 | case tok::kw_bool: | ||||||||
| 59 | case tok::kw___underlying_type: | ||||||||
| 60 | case tok::annot_typename: | ||||||||
| 61 | case tok::kw_char8_t: | ||||||||
| 62 | case tok::kw_char16_t: | ||||||||
| 63 | case tok::kw_char32_t: | ||||||||
| 64 | case tok::kw_typeof: | ||||||||
| 65 | case tok::kw_decltype: | ||||||||
| 66 | case tok::kw__Atomic: | ||||||||
| 67 | return true; | ||||||||
| 68 | default: | ||||||||
| 69 | return false; | ||||||||
| 70 | } | ||||||||
| 71 | } | ||||||||
| 72 | |||||||||
| 73 | bool FormatToken::isTypeOrIdentifier() const { | ||||||||
| 74 | return isSimpleTypeSpecifier() || Tok.isOneOf(tok::kw_auto, tok::identifier); | ||||||||
| 75 | } | ||||||||
| 76 | |||||||||
| 77 | TokenRole::~TokenRole() {} | ||||||||
| 78 | |||||||||
| 79 | void TokenRole::precomputeFormattingInfos(const FormatToken *Token) {} | ||||||||
| 80 | |||||||||
| 81 | unsigned CommaSeparatedList::formatAfterToken(LineState &State, | ||||||||
| 82 | ContinuationIndenter *Indenter, | ||||||||
| 83 | bool DryRun) { | ||||||||
| 84 | if (State.NextToken == nullptr || !State.NextToken->Previous) | ||||||||
| 85 | return 0; | ||||||||
| 86 | |||||||||
| 87 | if (Formats.size() == 1) | ||||||||
| 88 | return 0; // Handled by formatFromToken | ||||||||
| 89 | |||||||||
| 90 | // Ensure that we start on the opening brace. | ||||||||
| 91 | const FormatToken *LBrace = | ||||||||
| 92 | State.NextToken->Previous->getPreviousNonComment(); | ||||||||
| 93 | if (!LBrace || !LBrace->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) || | ||||||||
| 94 | LBrace->is(BK_Block) || LBrace->is(TT_DictLiteral) || | ||||||||
| 95 | LBrace->Next->is(TT_DesignatedInitializerPeriod)) | ||||||||
| 96 | return 0; | ||||||||
| 97 | |||||||||
| 98 | // Calculate the number of code points we have to format this list. As the | ||||||||
| 99 | // first token is already placed, we have to subtract it. | ||||||||
| 100 | unsigned RemainingCodePoints = | ||||||||
| 101 | Style.ColumnLimit - State.Column + State.NextToken->Previous->ColumnWidth; | ||||||||
| 102 | |||||||||
| 103 | // Find the best ColumnFormat, i.e. the best number of columns to use. | ||||||||
| 104 | const ColumnFormat *Format = getColumnFormat(RemainingCodePoints); | ||||||||
| 105 | |||||||||
| 106 | // If no ColumnFormat can be used, the braced list would generally be | ||||||||
| 107 | // bin-packed. Add a severe penalty to this so that column layouts are | ||||||||
| 108 | // preferred if possible. | ||||||||
| 109 | if (!Format) | ||||||||
| 110 | return 10000; | ||||||||
| 111 | |||||||||
| 112 | // Format the entire list. | ||||||||
| 113 | unsigned Penalty = 0; | ||||||||
| 114 | unsigned Column = 0; | ||||||||
| 115 | unsigned Item = 0; | ||||||||
| 116 | while (State.NextToken != LBrace->MatchingParen) { | ||||||||
| 117 | bool NewLine = false; | ||||||||
| 118 | unsigned ExtraSpaces = 0; | ||||||||
| 119 | |||||||||
| 120 | // If the previous token was one of our commas, we are now on the next item. | ||||||||
| 121 | if (Item < Commas.size() && State.NextToken->Previous == Commas[Item]) { | ||||||||
| 122 | if (!State.NextToken->isTrailingComment()) { | ||||||||
| 123 | ExtraSpaces += Format->ColumnSizes[Column] - ItemLengths[Item]; | ||||||||
| 124 | ++Column; | ||||||||
| 125 | } | ||||||||
| 126 | ++Item; | ||||||||
| 127 | } | ||||||||
| 128 | |||||||||
| 129 | if (Column == Format->Columns || State.NextToken->MustBreakBefore) { | ||||||||
| 130 | Column = 0; | ||||||||
| 131 | NewLine = true; | ||||||||
| 132 | } | ||||||||
| 133 | |||||||||
| 134 | // Place token using the continuation indenter and store the penalty. | ||||||||
| 135 | Penalty += Indenter->addTokenToState(State, NewLine, DryRun, ExtraSpaces); | ||||||||
| 136 | } | ||||||||
| 137 | return Penalty; | ||||||||
| 138 | } | ||||||||
| 139 | |||||||||
| 140 | unsigned CommaSeparatedList::formatFromToken(LineState &State, | ||||||||
| 141 | ContinuationIndenter *Indenter, | ||||||||
| 142 | bool DryRun) { | ||||||||
| 143 | // Formatting with 1 Column isn't really a column layout, so we don't need the | ||||||||
| 144 | // special logic here. We can just avoid bin packing any of the parameters. | ||||||||
| 145 | if (Formats.size() == 1 || HasNestedBracedList) | ||||||||
| 146 | State.Stack.back().AvoidBinPacking = true; | ||||||||
| 147 | return 0; | ||||||||
| 148 | } | ||||||||
| 149 | |||||||||
| 150 | // Returns the lengths in code points between Begin and End (both included), | ||||||||
| 151 | // assuming that the entire sequence is put on a single line. | ||||||||
| 152 | static unsigned CodePointsBetween(const FormatToken *Begin, | ||||||||
| 153 | const FormatToken *End) { | ||||||||
| 154 | assert(End->TotalLength >= Begin->TotalLength)(static_cast <bool> (End->TotalLength >= Begin-> TotalLength) ? void (0) : __assert_fail ("End->TotalLength >= Begin->TotalLength" , "clang/lib/Format/FormatToken.cpp", 154, __extension__ __PRETTY_FUNCTION__ )); | ||||||||
| 155 | return End->TotalLength - Begin->TotalLength + Begin->ColumnWidth; | ||||||||
| 156 | } | ||||||||
| 157 | |||||||||
| 158 | void CommaSeparatedList::precomputeFormattingInfos(const FormatToken *Token) { | ||||||||
| 159 | // FIXME: At some point we might want to do this for other lists, too. | ||||||||
| 160 | if (!Token->MatchingParen || | ||||||||
| |||||||||
| 161 | !Token->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare)) | ||||||||
| 162 | return; | ||||||||
| 163 | |||||||||
| 164 | // In C++11 braced list style, we should not format in columns unless they | ||||||||
| 165 | // have many items (20 or more) or we allow bin-packing of function call | ||||||||
| 166 | // arguments. | ||||||||
| 167 | if (Style.Cpp11BracedListStyle && !Style.BinPackArguments && | ||||||||
| 168 | Commas.size() < 19) | ||||||||
| 169 | return; | ||||||||
| 170 | |||||||||
| 171 | // Limit column layout for JavaScript array initializers to 20 or more items | ||||||||
| 172 | // for now to introduce it carefully. We can become more aggressive if this | ||||||||
| 173 | // necessary. | ||||||||
| 174 | if (Token->is(TT_ArrayInitializerLSquare) && Commas.size() < 19) | ||||||||
| 175 | return; | ||||||||
| 176 | |||||||||
| 177 | // Column format doesn't really make sense if we don't align after brackets. | ||||||||
| 178 | if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign) | ||||||||
| 179 | return; | ||||||||
| 180 | |||||||||
| 181 | FormatToken *ItemBegin = Token->Next; | ||||||||
| 182 | while (ItemBegin->isTrailingComment()) | ||||||||
| 183 | ItemBegin = ItemBegin->Next; | ||||||||
| 184 | SmallVector<bool, 8> MustBreakBeforeItem; | ||||||||
| 185 | |||||||||
| 186 | // The lengths of an item if it is put at the end of the line. This includes | ||||||||
| 187 | // trailing comments which are otherwise ignored for column alignment. | ||||||||
| 188 | SmallVector<unsigned, 8> EndOfLineItemLength; | ||||||||
| 189 | |||||||||
| 190 | bool HasSeparatingComment = false; | ||||||||
| 191 | for (unsigned i = 0, e = Commas.size() + 1; i
| ||||||||
| 192 | // Skip comments on their own line. | ||||||||
| 193 | while (ItemBegin->HasUnescapedNewline && ItemBegin->isTrailingComment()) { | ||||||||
| |||||||||
| 194 | ItemBegin = ItemBegin->Next; | ||||||||
| 195 | HasSeparatingComment = i > 0; | ||||||||
| 196 | } | ||||||||
| 197 | |||||||||
| 198 | MustBreakBeforeItem.push_back(ItemBegin->MustBreakBefore); | ||||||||
| 199 | if (ItemBegin->is(tok::l_brace)) | ||||||||
| 200 | HasNestedBracedList = true; | ||||||||
| 201 | const FormatToken *ItemEnd = nullptr; | ||||||||
| 202 | if (i == Commas.size()) { | ||||||||
| 203 | ItemEnd = Token->MatchingParen; | ||||||||
| 204 | const FormatToken *NonCommentEnd = ItemEnd->getPreviousNonComment(); | ||||||||
| 205 | ItemLengths.push_back(CodePointsBetween(ItemBegin, NonCommentEnd)); | ||||||||
| 206 | if (Style.Cpp11BracedListStyle && | ||||||||
| 207 | !ItemEnd->Previous->isTrailingComment()) { | ||||||||
| 208 | // In Cpp11 braced list style, the } and possibly other subsequent | ||||||||
| 209 | // tokens will need to stay on a line with the last element. | ||||||||
| 210 | while (ItemEnd->Next && !ItemEnd->Next->CanBreakBefore) | ||||||||
| 211 | ItemEnd = ItemEnd->Next; | ||||||||
| 212 | } else { | ||||||||
| 213 | // In other braced lists styles, the "}" can be wrapped to the new line. | ||||||||
| 214 | ItemEnd = Token->MatchingParen->Previous; | ||||||||
| 215 | } | ||||||||
| 216 | } else { | ||||||||
| 217 | ItemEnd = Commas[i]; | ||||||||
| 218 | // The comma is counted as part of the item when calculating the length. | ||||||||
| 219 | ItemLengths.push_back(CodePointsBetween(ItemBegin, ItemEnd)); | ||||||||
| 220 | |||||||||
| 221 | // Consume trailing comments so the are included in EndOfLineItemLength. | ||||||||
| 222 | if (ItemEnd->Next && !ItemEnd->Next->HasUnescapedNewline && | ||||||||
| 223 | ItemEnd->Next->isTrailingComment()) | ||||||||
| 224 | ItemEnd = ItemEnd->Next; | ||||||||
| 225 | } | ||||||||
| 226 | EndOfLineItemLength.push_back(CodePointsBetween(ItemBegin, ItemEnd)); | ||||||||
| 227 | // If there is a trailing comma in the list, the next item will start at the | ||||||||
| 228 | // closing brace. Don't create an extra item for this. | ||||||||
| 229 | if (ItemEnd->getNextNonComment() == Token->MatchingParen) | ||||||||
| 230 | break; | ||||||||
| 231 | ItemBegin = ItemEnd->Next; | ||||||||
| 232 | } | ||||||||
| 233 | |||||||||
| 234 | // Don't use column layout for lists with few elements and in presence of | ||||||||
| 235 | // separating comments. | ||||||||
| 236 | if (Commas.size() < 5 || HasSeparatingComment) | ||||||||
| 237 | return; | ||||||||
| 238 | |||||||||
| 239 | if (Token->NestingLevel != 0 && Token->is(tok::l_brace) && Commas.size() < 19) | ||||||||
| 240 | return; | ||||||||
| 241 | |||||||||
| 242 | // We can never place more than ColumnLimit / 3 items in a row (because of the | ||||||||
| 243 | // spaces and the comma). | ||||||||
| 244 | unsigned MaxItems = Style.ColumnLimit / 3; | ||||||||
| 245 | std::vector<unsigned> MinSizeInColumn; | ||||||||
| 246 | MinSizeInColumn.reserve(MaxItems); | ||||||||
| 247 | for (unsigned Columns = 1; Columns <= MaxItems; ++Columns) { | ||||||||
| 248 | ColumnFormat Format; | ||||||||
| 249 | Format.Columns = Columns; | ||||||||
| 250 | Format.ColumnSizes.resize(Columns); | ||||||||
| 251 | MinSizeInColumn.assign(Columns, UINT_MAX(2147483647 *2U +1U)); | ||||||||
| 252 | Format.LineCount = 1; | ||||||||
| 253 | bool HasRowWithSufficientColumns = false; | ||||||||
| 254 | unsigned Column = 0; | ||||||||
| 255 | for (unsigned i = 0, e = ItemLengths.size(); i != e; ++i) { | ||||||||
| 256 | assert(i < MustBreakBeforeItem.size())(static_cast <bool> (i < MustBreakBeforeItem.size()) ? void (0) : __assert_fail ("i < MustBreakBeforeItem.size()" , "clang/lib/Format/FormatToken.cpp", 256, __extension__ __PRETTY_FUNCTION__ )); | ||||||||
| 257 | if (MustBreakBeforeItem[i] || Column == Columns) { | ||||||||
| 258 | ++Format.LineCount; | ||||||||
| 259 | Column = 0; | ||||||||
| 260 | } | ||||||||
| 261 | if (Column == Columns - 1) | ||||||||
| 262 | HasRowWithSufficientColumns = true; | ||||||||
| 263 | unsigned Length = | ||||||||
| 264 | (Column == Columns - 1) ? EndOfLineItemLength[i] : ItemLengths[i]; | ||||||||
| 265 | Format.ColumnSizes[Column] = std::max(Format.ColumnSizes[Column], Length); | ||||||||
| 266 | MinSizeInColumn[Column] = std::min(MinSizeInColumn[Column], Length); | ||||||||
| 267 | ++Column; | ||||||||
| 268 | } | ||||||||
| 269 | // If all rows are terminated early (e.g. by trailing comments), we don't | ||||||||
| 270 | // need to look further. | ||||||||
| 271 | if (!HasRowWithSufficientColumns) | ||||||||
| 272 | break; | ||||||||
| 273 | Format.TotalWidth = Columns - 1; // Width of the N-1 spaces. | ||||||||
| 274 | |||||||||
| 275 | for (unsigned i = 0; i < Columns; ++i) | ||||||||
| 276 | Format.TotalWidth += Format.ColumnSizes[i]; | ||||||||
| 277 | |||||||||
| 278 | // Don't use this Format, if the difference between the longest and shortest | ||||||||
| 279 | // element in a column exceeds a threshold to avoid excessive spaces. | ||||||||
| 280 | if ([&] { | ||||||||
| 281 | for (unsigned i = 0; i < Columns - 1; ++i) | ||||||||
| 282 | if (Format.ColumnSizes[i] - MinSizeInColumn[i] > 10) | ||||||||
| 283 | return true; | ||||||||
| 284 | return false; | ||||||||
| 285 | }()) | ||||||||
| 286 | continue; | ||||||||
| 287 | |||||||||
| 288 | // Ignore layouts that are bound to violate the column limit. | ||||||||
| 289 | if (Format.TotalWidth > Style.ColumnLimit && Columns > 1) | ||||||||
| 290 | continue; | ||||||||
| 291 | |||||||||
| 292 | Formats.push_back(Format); | ||||||||
| 293 | } | ||||||||
| 294 | } | ||||||||
| 295 | |||||||||
| 296 | const CommaSeparatedList::ColumnFormat * | ||||||||
| 297 | CommaSeparatedList::getColumnFormat(unsigned RemainingCharacters) const { | ||||||||
| 298 | const ColumnFormat *BestFormat = nullptr; | ||||||||
| 299 | for (const ColumnFormat &Format : llvm::reverse(Formats)) { | ||||||||
| 300 | if (Format.TotalWidth <= RemainingCharacters || Format.Columns == 1) { | ||||||||
| 301 | if (BestFormat && Format.LineCount > BestFormat->LineCount) | ||||||||
| 302 | break; | ||||||||
| 303 | BestFormat = &Format; | ||||||||
| 304 | } | ||||||||
| 305 | } | ||||||||
| 306 | return BestFormat; | ||||||||
| 307 | } | ||||||||
| 308 | |||||||||
| 309 | } // namespace format | ||||||||
| 310 | } // namespace clang |
| 1 | //===--- FormatToken.h - Format C++ code ------------------------*- C++ -*-===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | /// |
| 9 | /// \file |
| 10 | /// This file contains the declaration of the FormatToken, a wrapper |
| 11 | /// around Token with additional information related to formatting. |
| 12 | /// |
| 13 | //===----------------------------------------------------------------------===// |
| 14 | |
| 15 | #ifndef LLVM_CLANG_LIB_FORMAT_FORMATTOKEN_H |
| 16 | #define LLVM_CLANG_LIB_FORMAT_FORMATTOKEN_H |
| 17 | |
| 18 | #include "clang/Basic/IdentifierTable.h" |
| 19 | #include "clang/Basic/OperatorPrecedence.h" |
| 20 | #include "clang/Format/Format.h" |
| 21 | #include "clang/Lex/Lexer.h" |
| 22 | #include <memory> |
| 23 | #include <unordered_set> |
| 24 | |
| 25 | namespace clang { |
| 26 | namespace format { |
| 27 | |
| 28 | #define LIST_TOKEN_TYPESTYPE(ArrayInitializerLSquare) TYPE(ArraySubscriptLSquare) TYPE (AttributeColon) TYPE(AttributeMacro) TYPE(AttributeParen) TYPE (AttributeSquare) TYPE(BinaryOperator) TYPE(BitFieldColon) TYPE (BlockComment) TYPE(CastRParen) TYPE(ConditionalExpr) TYPE(ConflictAlternative ) TYPE(ConflictEnd) TYPE(ConflictStart) TYPE(ConstraintJunctions ) TYPE(CtorInitializerColon) TYPE(CtorInitializerComma) TYPE( DesignatedInitializerLSquare) TYPE(DesignatedInitializerPeriod ) TYPE(DictLiteral) TYPE(FatArrow) TYPE(ForEachMacro) TYPE(FunctionAnnotationRParen ) TYPE(FunctionDeclarationName) TYPE(FunctionLBrace) TYPE(FunctionTypeLParen ) TYPE(IfMacro) TYPE(ImplicitStringLiteral) TYPE(InheritanceColon ) TYPE(InheritanceComma) TYPE(InlineASMBrace) TYPE(InlineASMColon ) TYPE(InlineASMSymbolicNameLSquare) TYPE(JavaAnnotation) TYPE (JsComputedPropertyName) TYPE(JsExponentiation) TYPE(JsExponentiationEqual ) TYPE(JsPipePipeEqual) TYPE(JsPrivateIdentifier) TYPE(JsTypeColon ) TYPE(JsTypeOperator) TYPE(JsTypeOptionalQuestion) TYPE(JsAndAndEqual ) TYPE(LambdaArrow) TYPE(LambdaLBrace) TYPE(LambdaLSquare) TYPE (LeadingJavaAnnotation) TYPE(LineComment) TYPE(MacroBlockBegin ) TYPE(MacroBlockEnd) TYPE(ModulePartitionColon) TYPE(NamespaceMacro ) TYPE(NonNullAssertion) TYPE(NullCoalescingEqual) TYPE(NullCoalescingOperator ) TYPE(NullPropagatingOperator) TYPE(ObjCBlockLBrace) TYPE(ObjCBlockLParen ) TYPE(ObjCDecl) TYPE(ObjCForIn) TYPE(ObjCMethodExpr) TYPE(ObjCMethodSpecifier ) TYPE(ObjCProperty) TYPE(ObjCStringLiteral) TYPE(OverloadedOperator ) TYPE(OverloadedOperatorLParen) TYPE(PointerOrReference) TYPE (PureVirtualSpecifier) TYPE(RangeBasedForLoopColon) TYPE(RegexLiteral ) TYPE(SelectorName) TYPE(StartOfName) TYPE(StatementAttributeLikeMacro ) TYPE(StatementMacro) TYPE(StructuredBindingLSquare) TYPE(TemplateCloser ) TYPE(TemplateOpener) TYPE(TemplateString) TYPE(ProtoExtensionLSquare ) TYPE(TrailingAnnotation) TYPE(TrailingReturnArrow) TYPE(TrailingUnaryOperator ) TYPE(TypeDeclarationParen) TYPE(TypenameMacro) TYPE(UnaryOperator ) TYPE(UntouchableMacroFunc) TYPE(CSharpStringLiteral) TYPE(CSharpNamedArgumentColon ) TYPE(CSharpNullable) TYPE(CSharpNullConditionalLSquare) TYPE (CSharpGenericTypeConstraint) TYPE(CSharpGenericTypeConstraintColon ) TYPE(CSharpGenericTypeConstraintComma) TYPE(Unknown) \ |
| 29 | TYPE(ArrayInitializerLSquare) \ |
| 30 | TYPE(ArraySubscriptLSquare) \ |
| 31 | TYPE(AttributeColon) \ |
| 32 | TYPE(AttributeMacro) \ |
| 33 | TYPE(AttributeParen) \ |
| 34 | TYPE(AttributeSquare) \ |
| 35 | TYPE(BinaryOperator) \ |
| 36 | TYPE(BitFieldColon) \ |
| 37 | TYPE(BlockComment) \ |
| 38 | TYPE(CastRParen) \ |
| 39 | TYPE(ConditionalExpr) \ |
| 40 | TYPE(ConflictAlternative) \ |
| 41 | TYPE(ConflictEnd) \ |
| 42 | TYPE(ConflictStart) \ |
| 43 | TYPE(ConstraintJunctions) \ |
| 44 | TYPE(CtorInitializerColon) \ |
| 45 | TYPE(CtorInitializerComma) \ |
| 46 | TYPE(DesignatedInitializerLSquare) \ |
| 47 | TYPE(DesignatedInitializerPeriod) \ |
| 48 | TYPE(DictLiteral) \ |
| 49 | TYPE(FatArrow) \ |
| 50 | TYPE(ForEachMacro) \ |
| 51 | TYPE(FunctionAnnotationRParen) \ |
| 52 | TYPE(FunctionDeclarationName) \ |
| 53 | TYPE(FunctionLBrace) \ |
| 54 | TYPE(FunctionTypeLParen) \ |
| 55 | TYPE(IfMacro) \ |
| 56 | TYPE(ImplicitStringLiteral) \ |
| 57 | TYPE(InheritanceColon) \ |
| 58 | TYPE(InheritanceComma) \ |
| 59 | TYPE(InlineASMBrace) \ |
| 60 | TYPE(InlineASMColon) \ |
| 61 | TYPE(InlineASMSymbolicNameLSquare) \ |
| 62 | TYPE(JavaAnnotation) \ |
| 63 | TYPE(JsComputedPropertyName) \ |
| 64 | TYPE(JsExponentiation) \ |
| 65 | TYPE(JsExponentiationEqual) \ |
| 66 | TYPE(JsPipePipeEqual) \ |
| 67 | TYPE(JsPrivateIdentifier) \ |
| 68 | TYPE(JsTypeColon) \ |
| 69 | TYPE(JsTypeOperator) \ |
| 70 | TYPE(JsTypeOptionalQuestion) \ |
| 71 | TYPE(JsAndAndEqual) \ |
| 72 | TYPE(LambdaArrow) \ |
| 73 | TYPE(LambdaLBrace) \ |
| 74 | TYPE(LambdaLSquare) \ |
| 75 | TYPE(LeadingJavaAnnotation) \ |
| 76 | TYPE(LineComment) \ |
| 77 | TYPE(MacroBlockBegin) \ |
| 78 | TYPE(MacroBlockEnd) \ |
| 79 | TYPE(ModulePartitionColon) \ |
| 80 | TYPE(NamespaceMacro) \ |
| 81 | TYPE(NonNullAssertion) \ |
| 82 | TYPE(NullCoalescingEqual) \ |
| 83 | TYPE(NullCoalescingOperator) \ |
| 84 | TYPE(NullPropagatingOperator) \ |
| 85 | TYPE(ObjCBlockLBrace) \ |
| 86 | TYPE(ObjCBlockLParen) \ |
| 87 | TYPE(ObjCDecl) \ |
| 88 | TYPE(ObjCForIn) \ |
| 89 | TYPE(ObjCMethodExpr) \ |
| 90 | TYPE(ObjCMethodSpecifier) \ |
| 91 | TYPE(ObjCProperty) \ |
| 92 | TYPE(ObjCStringLiteral) \ |
| 93 | TYPE(OverloadedOperator) \ |
| 94 | TYPE(OverloadedOperatorLParen) \ |
| 95 | TYPE(PointerOrReference) \ |
| 96 | TYPE(PureVirtualSpecifier) \ |
| 97 | TYPE(RangeBasedForLoopColon) \ |
| 98 | TYPE(RegexLiteral) \ |
| 99 | TYPE(SelectorName) \ |
| 100 | TYPE(StartOfName) \ |
| 101 | TYPE(StatementAttributeLikeMacro) \ |
| 102 | TYPE(StatementMacro) \ |
| 103 | TYPE(StructuredBindingLSquare) \ |
| 104 | TYPE(TemplateCloser) \ |
| 105 | TYPE(TemplateOpener) \ |
| 106 | TYPE(TemplateString) \ |
| 107 | TYPE(ProtoExtensionLSquare) \ |
| 108 | TYPE(TrailingAnnotation) \ |
| 109 | TYPE(TrailingReturnArrow) \ |
| 110 | TYPE(TrailingUnaryOperator) \ |
| 111 | TYPE(TypeDeclarationParen) \ |
| 112 | TYPE(TypenameMacro) \ |
| 113 | TYPE(UnaryOperator) \ |
| 114 | TYPE(UntouchableMacroFunc) \ |
| 115 | TYPE(CSharpStringLiteral) \ |
| 116 | TYPE(CSharpNamedArgumentColon) \ |
| 117 | TYPE(CSharpNullable) \ |
| 118 | TYPE(CSharpNullConditionalLSquare) \ |
| 119 | TYPE(CSharpGenericTypeConstraint) \ |
| 120 | TYPE(CSharpGenericTypeConstraintColon) \ |
| 121 | TYPE(CSharpGenericTypeConstraintComma) \ |
| 122 | TYPE(Unknown) |
| 123 | |
| 124 | /// Determines the semantic type of a syntactic token, e.g. whether "<" is a |
| 125 | /// template opener or binary operator. |
| 126 | enum TokenType : uint8_t { |
| 127 | #define TYPE(X) TT_##X, |
| 128 | LIST_TOKEN_TYPESTYPE(ArrayInitializerLSquare) TYPE(ArraySubscriptLSquare) TYPE (AttributeColon) TYPE(AttributeMacro) TYPE(AttributeParen) TYPE (AttributeSquare) TYPE(BinaryOperator) TYPE(BitFieldColon) TYPE (BlockComment) TYPE(CastRParen) TYPE(ConditionalExpr) TYPE(ConflictAlternative ) TYPE(ConflictEnd) TYPE(ConflictStart) TYPE(ConstraintJunctions ) TYPE(CtorInitializerColon) TYPE(CtorInitializerComma) TYPE( DesignatedInitializerLSquare) TYPE(DesignatedInitializerPeriod ) TYPE(DictLiteral) TYPE(FatArrow) TYPE(ForEachMacro) TYPE(FunctionAnnotationRParen ) TYPE(FunctionDeclarationName) TYPE(FunctionLBrace) TYPE(FunctionTypeLParen ) TYPE(IfMacro) TYPE(ImplicitStringLiteral) TYPE(InheritanceColon ) TYPE(InheritanceComma) TYPE(InlineASMBrace) TYPE(InlineASMColon ) TYPE(InlineASMSymbolicNameLSquare) TYPE(JavaAnnotation) TYPE (JsComputedPropertyName) TYPE(JsExponentiation) TYPE(JsExponentiationEqual ) TYPE(JsPipePipeEqual) TYPE(JsPrivateIdentifier) TYPE(JsTypeColon ) TYPE(JsTypeOperator) TYPE(JsTypeOptionalQuestion) TYPE(JsAndAndEqual ) TYPE(LambdaArrow) TYPE(LambdaLBrace) TYPE(LambdaLSquare) TYPE (LeadingJavaAnnotation) TYPE(LineComment) TYPE(MacroBlockBegin ) TYPE(MacroBlockEnd) TYPE(ModulePartitionColon) TYPE(NamespaceMacro ) TYPE(NonNullAssertion) TYPE(NullCoalescingEqual) TYPE(NullCoalescingOperator ) TYPE(NullPropagatingOperator) TYPE(ObjCBlockLBrace) TYPE(ObjCBlockLParen ) TYPE(ObjCDecl) TYPE(ObjCForIn) TYPE(ObjCMethodExpr) TYPE(ObjCMethodSpecifier ) TYPE(ObjCProperty) TYPE(ObjCStringLiteral) TYPE(OverloadedOperator ) TYPE(OverloadedOperatorLParen) TYPE(PointerOrReference) TYPE (PureVirtualSpecifier) TYPE(RangeBasedForLoopColon) TYPE(RegexLiteral ) TYPE(SelectorName) TYPE(StartOfName) TYPE(StatementAttributeLikeMacro ) TYPE(StatementMacro) TYPE(StructuredBindingLSquare) TYPE(TemplateCloser ) TYPE(TemplateOpener) TYPE(TemplateString) TYPE(ProtoExtensionLSquare ) TYPE(TrailingAnnotation) TYPE(TrailingReturnArrow) TYPE(TrailingUnaryOperator ) TYPE(TypeDeclarationParen) TYPE(TypenameMacro) TYPE(UnaryOperator ) TYPE(UntouchableMacroFunc) TYPE(CSharpStringLiteral) TYPE(CSharpNamedArgumentColon ) TYPE(CSharpNullable) TYPE(CSharpNullConditionalLSquare) TYPE (CSharpGenericTypeConstraint) TYPE(CSharpGenericTypeConstraintColon ) TYPE(CSharpGenericTypeConstraintComma) TYPE(Unknown) |
| 129 | #undef TYPE |
| 130 | NUM_TOKEN_TYPES |
| 131 | }; |
| 132 | |
| 133 | /// Determines the name of a token type. |
| 134 | const char *getTokenTypeName(TokenType Type); |
| 135 | |
| 136 | // Represents what type of block a set of braces open. |
| 137 | enum BraceBlockKind { BK_Unknown, BK_Block, BK_BracedInit }; |
| 138 | |
| 139 | // The packing kind of a function's parameters. |
| 140 | enum ParameterPackingKind { PPK_BinPacked, PPK_OnePerLine, PPK_Inconclusive }; |
| 141 | |
| 142 | enum FormatDecision { FD_Unformatted, FD_Continue, FD_Break }; |
| 143 | |
| 144 | /// Roles a token can take in a configured macro expansion. |
| 145 | enum MacroRole { |
| 146 | /// The token was expanded from a macro argument when formatting the expanded |
| 147 | /// token sequence. |
| 148 | MR_ExpandedArg, |
| 149 | /// The token is part of a macro argument that was previously formatted as |
| 150 | /// expansion when formatting the unexpanded macro call. |
| 151 | MR_UnexpandedArg, |
| 152 | /// The token was expanded from a macro definition, and is not visible as part |
| 153 | /// of the macro call. |
| 154 | MR_Hidden, |
| 155 | }; |
| 156 | |
| 157 | struct FormatToken; |
| 158 | |
| 159 | /// Contains information on the token's role in a macro expansion. |
| 160 | /// |
| 161 | /// Given the following definitions: |
| 162 | /// A(X) = [ X ] |
| 163 | /// B(X) = < X > |
| 164 | /// C(X) = X |
| 165 | /// |
| 166 | /// Consider the macro call: |
| 167 | /// A({B(C(C(x)))}) -> [{<x>}] |
| 168 | /// |
| 169 | /// In this case, the tokens of the unexpanded macro call will have the |
| 170 | /// following relevant entries in their macro context (note that formatting |
| 171 | /// the unexpanded macro call happens *after* formatting the expanded macro |
| 172 | /// call): |
| 173 | /// A( { B( C( C(x) ) ) } ) |
| 174 | /// Role: NN U NN NN NNUN N N U N (N=None, U=UnexpandedArg) |
| 175 | /// |
| 176 | /// [ { < x > } ] |
| 177 | /// Role: H E H E H E H (H=Hidden, E=ExpandedArg) |
| 178 | /// ExpandedFrom[0]: A A A A A A A |
| 179 | /// ExpandedFrom[1]: B B B |
| 180 | /// ExpandedFrom[2]: C |
| 181 | /// ExpandedFrom[3]: C |
| 182 | /// StartOfExpansion: 1 0 1 2 0 0 0 |
| 183 | /// EndOfExpansion: 0 0 0 2 1 0 1 |
| 184 | struct MacroExpansion { |
| 185 | MacroExpansion(MacroRole Role) : Role(Role) {} |
| 186 | |
| 187 | /// The token's role in the macro expansion. |
| 188 | /// When formatting an expanded macro, all tokens that are part of macro |
| 189 | /// arguments will be MR_ExpandedArg, while all tokens that are not visible in |
| 190 | /// the macro call will be MR_Hidden. |
| 191 | /// When formatting an unexpanded macro call, all tokens that are part of |
| 192 | /// macro arguments will be MR_UnexpandedArg. |
| 193 | MacroRole Role; |
| 194 | |
| 195 | /// The stack of macro call identifier tokens this token was expanded from. |
| 196 | llvm::SmallVector<FormatToken *, 1> ExpandedFrom; |
| 197 | |
| 198 | /// The number of expansions of which this macro is the first entry. |
| 199 | unsigned StartOfExpansion = 0; |
| 200 | |
| 201 | /// The number of currently open expansions in \c ExpandedFrom this macro is |
| 202 | /// the last token in. |
| 203 | unsigned EndOfExpansion = 0; |
| 204 | }; |
| 205 | |
| 206 | class TokenRole; |
| 207 | class AnnotatedLine; |
| 208 | |
| 209 | /// A wrapper around a \c Token storing information about the |
| 210 | /// whitespace characters preceding it. |
| 211 | struct FormatToken { |
| 212 | FormatToken() |
| 213 | : HasUnescapedNewline(false), IsMultiline(false), IsFirst(false), |
| 214 | MustBreakBefore(false), IsUnterminatedLiteral(false), |
| 215 | CanBreakBefore(false), ClosesTemplateDeclaration(false), |
| 216 | StartsBinaryExpression(false), EndsBinaryExpression(false), |
| 217 | PartOfMultiVariableDeclStmt(false), ContinuesLineCommentSection(false), |
| 218 | Finalized(false), BlockKind(BK_Unknown), Decision(FD_Unformatted), |
| 219 | PackingKind(PPK_Inconclusive), Type(TT_Unknown) {} |
| 220 | |
| 221 | /// The \c Token. |
| 222 | Token Tok; |
| 223 | |
| 224 | /// The raw text of the token. |
| 225 | /// |
| 226 | /// Contains the raw token text without leading whitespace and without leading |
| 227 | /// escaped newlines. |
| 228 | StringRef TokenText; |
| 229 | |
| 230 | /// A token can have a special role that can carry extra information |
| 231 | /// about the token's formatting. |
| 232 | /// FIXME: Make FormatToken for parsing and AnnotatedToken two different |
| 233 | /// classes and make this a unique_ptr in the AnnotatedToken class. |
| 234 | std::shared_ptr<TokenRole> Role; |
| 235 | |
| 236 | /// The range of the whitespace immediately preceding the \c Token. |
| 237 | SourceRange WhitespaceRange; |
| 238 | |
| 239 | /// Whether there is at least one unescaped newline before the \c |
| 240 | /// Token. |
| 241 | unsigned HasUnescapedNewline : 1; |
| 242 | |
| 243 | /// Whether the token text contains newlines (escaped or not). |
| 244 | unsigned IsMultiline : 1; |
| 245 | |
| 246 | /// Indicates that this is the first token of the file. |
| 247 | unsigned IsFirst : 1; |
| 248 | |
| 249 | /// Whether there must be a line break before this token. |
| 250 | /// |
| 251 | /// This happens for example when a preprocessor directive ended directly |
| 252 | /// before the token. |
| 253 | unsigned MustBreakBefore : 1; |
| 254 | |
| 255 | /// Set to \c true if this token is an unterminated literal. |
| 256 | unsigned IsUnterminatedLiteral : 1; |
| 257 | |
| 258 | /// \c true if it is allowed to break before this token. |
| 259 | unsigned CanBreakBefore : 1; |
| 260 | |
| 261 | /// \c true if this is the ">" of "template<..>". |
| 262 | unsigned ClosesTemplateDeclaration : 1; |
| 263 | |
| 264 | /// \c true if this token starts a binary expression, i.e. has at least |
| 265 | /// one fake l_paren with a precedence greater than prec::Unknown. |
| 266 | unsigned StartsBinaryExpression : 1; |
| 267 | /// \c true if this token ends a binary expression. |
| 268 | unsigned EndsBinaryExpression : 1; |
| 269 | |
| 270 | /// Is this token part of a \c DeclStmt defining multiple variables? |
| 271 | /// |
| 272 | /// Only set if \c Type == \c TT_StartOfName. |
| 273 | unsigned PartOfMultiVariableDeclStmt : 1; |
| 274 | |
| 275 | /// Does this line comment continue a line comment section? |
| 276 | /// |
| 277 | /// Only set to true if \c Type == \c TT_LineComment. |
| 278 | unsigned ContinuesLineCommentSection : 1; |
| 279 | |
| 280 | /// If \c true, this token has been fully formatted (indented and |
| 281 | /// potentially re-formatted inside), and we do not allow further formatting |
| 282 | /// changes. |
| 283 | unsigned Finalized : 1; |
| 284 | |
| 285 | private: |
| 286 | /// Contains the kind of block if this token is a brace. |
| 287 | unsigned BlockKind : 2; |
| 288 | |
| 289 | public: |
| 290 | BraceBlockKind getBlockKind() const { |
| 291 | return static_cast<BraceBlockKind>(BlockKind); |
| 292 | } |
| 293 | void setBlockKind(BraceBlockKind BBK) { |
| 294 | BlockKind = BBK; |
| 295 | assert(getBlockKind() == BBK && "BraceBlockKind overflow!")(static_cast <bool> (getBlockKind() == BBK && "BraceBlockKind overflow!" ) ? void (0) : __assert_fail ("getBlockKind() == BBK && \"BraceBlockKind overflow!\"" , "clang/lib/Format/FormatToken.h", 295, __extension__ __PRETTY_FUNCTION__ )); |
| 296 | } |
| 297 | |
| 298 | private: |
| 299 | /// Stores the formatting decision for the token once it was made. |
| 300 | unsigned Decision : 2; |
| 301 | |
| 302 | public: |
| 303 | FormatDecision getDecision() const { |
| 304 | return static_cast<FormatDecision>(Decision); |
| 305 | } |
| 306 | void setDecision(FormatDecision D) { |
| 307 | Decision = D; |
| 308 | assert(getDecision() == D && "FormatDecision overflow!")(static_cast <bool> (getDecision() == D && "FormatDecision overflow!" ) ? void (0) : __assert_fail ("getDecision() == D && \"FormatDecision overflow!\"" , "clang/lib/Format/FormatToken.h", 308, __extension__ __PRETTY_FUNCTION__ )); |
| 309 | } |
| 310 | |
| 311 | private: |
| 312 | /// If this is an opening parenthesis, how are the parameters packed? |
| 313 | unsigned PackingKind : 2; |
| 314 | |
| 315 | public: |
| 316 | ParameterPackingKind getPackingKind() const { |
| 317 | return static_cast<ParameterPackingKind>(PackingKind); |
| 318 | } |
| 319 | void setPackingKind(ParameterPackingKind K) { |
| 320 | PackingKind = K; |
| 321 | assert(getPackingKind() == K && "ParameterPackingKind overflow!")(static_cast <bool> (getPackingKind() == K && "ParameterPackingKind overflow!" ) ? void (0) : __assert_fail ("getPackingKind() == K && \"ParameterPackingKind overflow!\"" , "clang/lib/Format/FormatToken.h", 321, __extension__ __PRETTY_FUNCTION__ )); |
| 322 | } |
| 323 | |
| 324 | private: |
| 325 | TokenType Type; |
| 326 | |
| 327 | public: |
| 328 | /// Returns the token's type, e.g. whether "<" is a template opener or |
| 329 | /// binary operator. |
| 330 | TokenType getType() const { return Type; } |
| 331 | void setType(TokenType T) { Type = T; } |
| 332 | |
| 333 | /// The number of newlines immediately before the \c Token. |
| 334 | /// |
| 335 | /// This can be used to determine what the user wrote in the original code |
| 336 | /// and thereby e.g. leave an empty line between two function definitions. |
| 337 | unsigned NewlinesBefore = 0; |
| 338 | |
| 339 | /// The offset just past the last '\n' in this token's leading |
| 340 | /// whitespace (relative to \c WhiteSpaceStart). 0 if there is no '\n'. |
| 341 | unsigned LastNewlineOffset = 0; |
| 342 | |
| 343 | /// The width of the non-whitespace parts of the token (or its first |
| 344 | /// line for multi-line tokens) in columns. |
| 345 | /// We need this to correctly measure number of columns a token spans. |
| 346 | unsigned ColumnWidth = 0; |
| 347 | |
| 348 | /// Contains the width in columns of the last line of a multi-line |
| 349 | /// token. |
| 350 | unsigned LastLineColumnWidth = 0; |
| 351 | |
| 352 | /// The number of spaces that should be inserted before this token. |
| 353 | unsigned SpacesRequiredBefore = 0; |
| 354 | |
| 355 | /// Number of parameters, if this is "(", "[" or "<". |
| 356 | unsigned ParameterCount = 0; |
| 357 | |
| 358 | /// Number of parameters that are nested blocks, |
| 359 | /// if this is "(", "[" or "<". |
| 360 | unsigned BlockParameterCount = 0; |
| 361 | |
| 362 | /// If this is a bracket ("<", "(", "[" or "{"), contains the kind of |
| 363 | /// the surrounding bracket. |
| 364 | tok::TokenKind ParentBracket = tok::unknown; |
| 365 | |
| 366 | /// The total length of the unwrapped line up to and including this |
| 367 | /// token. |
| 368 | unsigned TotalLength = 0; |
| 369 | |
| 370 | /// The original 0-based column of this token, including expanded tabs. |
| 371 | /// The configured TabWidth is used as tab width. |
| 372 | unsigned OriginalColumn = 0; |
| 373 | |
| 374 | /// The length of following tokens until the next natural split point, |
| 375 | /// or the next token that can be broken. |
| 376 | unsigned UnbreakableTailLength = 0; |
| 377 | |
| 378 | // FIXME: Come up with a 'cleaner' concept. |
| 379 | /// The binding strength of a token. This is a combined value of |
| 380 | /// operator precedence, parenthesis nesting, etc. |
| 381 | unsigned BindingStrength = 0; |
| 382 | |
| 383 | /// The nesting level of this token, i.e. the number of surrounding (), |
| 384 | /// [], {} or <>. |
| 385 | unsigned NestingLevel = 0; |
| 386 | |
| 387 | /// The indent level of this token. Copied from the surrounding line. |
| 388 | unsigned IndentLevel = 0; |
| 389 | |
| 390 | /// Penalty for inserting a line break before this token. |
| 391 | unsigned SplitPenalty = 0; |
| 392 | |
| 393 | /// If this is the first ObjC selector name in an ObjC method |
| 394 | /// definition or call, this contains the length of the longest name. |
| 395 | /// |
| 396 | /// This being set to 0 means that the selectors should not be colon-aligned, |
| 397 | /// e.g. because several of them are block-type. |
| 398 | unsigned LongestObjCSelectorName = 0; |
| 399 | |
| 400 | /// If this is the first ObjC selector name in an ObjC method |
| 401 | /// definition or call, this contains the number of parts that the whole |
| 402 | /// selector consist of. |
| 403 | unsigned ObjCSelectorNameParts = 0; |
| 404 | |
| 405 | /// The 0-based index of the parameter/argument. For ObjC it is set |
| 406 | /// for the selector name token. |
| 407 | /// For now calculated only for ObjC. |
| 408 | unsigned ParameterIndex = 0; |
| 409 | |
| 410 | /// Stores the number of required fake parentheses and the |
| 411 | /// corresponding operator precedence. |
| 412 | /// |
| 413 | /// If multiple fake parentheses start at a token, this vector stores them in |
| 414 | /// reverse order, i.e. inner fake parenthesis first. |
| 415 | SmallVector<prec::Level, 4> FakeLParens; |
| 416 | /// Insert this many fake ) after this token for correct indentation. |
| 417 | unsigned FakeRParens = 0; |
| 418 | |
| 419 | /// If this is an operator (or "."/"->") in a sequence of operators |
| 420 | /// with the same precedence, contains the 0-based operator index. |
| 421 | unsigned OperatorIndex = 0; |
| 422 | |
| 423 | /// If this is an operator (or "."/"->") in a sequence of operators |
| 424 | /// with the same precedence, points to the next operator. |
| 425 | FormatToken *NextOperator = nullptr; |
| 426 | |
| 427 | /// If this is a bracket, this points to the matching one. |
| 428 | FormatToken *MatchingParen = nullptr; |
| 429 | |
| 430 | /// The previous token in the unwrapped line. |
| 431 | FormatToken *Previous = nullptr; |
| 432 | |
| 433 | /// The next token in the unwrapped line. |
| 434 | FormatToken *Next = nullptr; |
| 435 | |
| 436 | /// The first token in set of column elements. |
| 437 | bool StartsColumn = false; |
| 438 | |
| 439 | /// This notes the start of the line of an array initializer. |
| 440 | bool ArrayInitializerLineStart = false; |
| 441 | |
| 442 | /// This starts an array initializer. |
| 443 | bool IsArrayInitializer = false; |
| 444 | |
| 445 | /// Is optional and can be removed. |
| 446 | bool Optional = false; |
| 447 | |
| 448 | /// If this token starts a block, this contains all the unwrapped lines |
| 449 | /// in it. |
| 450 | SmallVector<AnnotatedLine *, 1> Children; |
| 451 | |
| 452 | // Contains all attributes related to how this token takes part |
| 453 | // in a configured macro expansion. |
| 454 | llvm::Optional<MacroExpansion> MacroCtx; |
| 455 | |
| 456 | bool is(tok::TokenKind Kind) const { return Tok.is(Kind); } |
| 457 | bool is(TokenType TT) const { return getType() == TT; } |
| 458 | bool is(const IdentifierInfo *II) const { |
| 459 | return II && II == Tok.getIdentifierInfo(); |
| 460 | } |
| 461 | bool is(tok::PPKeywordKind Kind) const { |
| 462 | return Tok.getIdentifierInfo() && |
| 463 | Tok.getIdentifierInfo()->getPPKeywordID() == Kind; |
| 464 | } |
| 465 | bool is(BraceBlockKind BBK) const { return getBlockKind() == BBK; } |
| 466 | bool is(ParameterPackingKind PPK) const { return getPackingKind() == PPK; } |
| 467 | |
| 468 | template <typename A, typename B> bool isOneOf(A K1, B K2) const { |
| 469 | return is(K1) || is(K2); |
| 470 | } |
| 471 | template <typename A, typename B, typename... Ts> |
| 472 | bool isOneOf(A K1, B K2, Ts... Ks) const { |
| 473 | return is(K1) || isOneOf(K2, Ks...); |
| 474 | } |
| 475 | template <typename T> bool isNot(T Kind) const { return !is(Kind); } |
| 476 | |
| 477 | bool isIf(bool AllowConstexprMacro = true) const { |
| 478 | return is(tok::kw_if) || endsSequence(tok::kw_constexpr, tok::kw_if) || |
| 479 | (endsSequence(tok::identifier, tok::kw_if) && AllowConstexprMacro); |
| 480 | } |
| 481 | |
| 482 | bool closesScopeAfterBlock() const { |
| 483 | if (getBlockKind() == BK_Block) |
| 484 | return true; |
| 485 | if (closesScope()) |
| 486 | return Previous->closesScopeAfterBlock(); |
| 487 | return false; |
| 488 | } |
| 489 | |
| 490 | /// \c true if this token starts a sequence with the given tokens in order, |
| 491 | /// following the ``Next`` pointers, ignoring comments. |
| 492 | template <typename A, typename... Ts> |
| 493 | bool startsSequence(A K1, Ts... Tokens) const { |
| 494 | return startsSequenceInternal(K1, Tokens...); |
| 495 | } |
| 496 | |
| 497 | /// \c true if this token ends a sequence with the given tokens in order, |
| 498 | /// following the ``Previous`` pointers, ignoring comments. |
| 499 | /// For example, given tokens [T1, T2, T3], the function returns true if |
| 500 | /// 3 tokens ending at this (ignoring comments) are [T3, T2, T1]. In other |
| 501 | /// words, the tokens passed to this function need to the reverse of the |
| 502 | /// order the tokens appear in code. |
| 503 | template <typename A, typename... Ts> |
| 504 | bool endsSequence(A K1, Ts... Tokens) const { |
| 505 | return endsSequenceInternal(K1, Tokens...); |
| 506 | } |
| 507 | |
| 508 | bool isStringLiteral() const { return tok::isStringLiteral(Tok.getKind()); } |
| 509 | |
| 510 | bool isObjCAtKeyword(tok::ObjCKeywordKind Kind) const { |
| 511 | return Tok.isObjCAtKeyword(Kind); |
| 512 | } |
| 513 | |
| 514 | bool isAccessSpecifier(bool ColonRequired = true) const { |
| 515 | return isOneOf(tok::kw_public, tok::kw_protected, tok::kw_private) && |
| 516 | (!ColonRequired || (Next && Next->is(tok::colon))); |
| 517 | } |
| 518 | |
| 519 | bool canBePointerOrReferenceQualifier() const { |
| 520 | return isOneOf(tok::kw_const, tok::kw_restrict, tok::kw_volatile, |
| 521 | tok::kw___attribute, tok::kw__Nonnull, tok::kw__Nullable, |
| 522 | tok::kw__Null_unspecified, tok::kw___ptr32, tok::kw___ptr64, |
| 523 | TT_AttributeMacro); |
| 524 | } |
| 525 | |
| 526 | /// Determine whether the token is a simple-type-specifier. |
| 527 | LLVM_NODISCARD[[clang::warn_unused_result]] bool isSimpleTypeSpecifier() const; |
| 528 | |
| 529 | LLVM_NODISCARD[[clang::warn_unused_result]] bool isTypeOrIdentifier() const; |
| 530 | |
| 531 | bool isObjCAccessSpecifier() const { |
| 532 | return is(tok::at) && Next && |
| 533 | (Next->isObjCAtKeyword(tok::objc_public) || |
| 534 | Next->isObjCAtKeyword(tok::objc_protected) || |
| 535 | Next->isObjCAtKeyword(tok::objc_package) || |
| 536 | Next->isObjCAtKeyword(tok::objc_private)); |
| 537 | } |
| 538 | |
| 539 | /// Returns whether \p Tok is ([{ or an opening < of a template or in |
| 540 | /// protos. |
| 541 | bool opensScope() const { |
| 542 | if (is(TT_TemplateString) && TokenText.endswith("${")) |
| 543 | return true; |
| 544 | if (is(TT_DictLiteral) && is(tok::less)) |
| 545 | return true; |
| 546 | return isOneOf(tok::l_paren, tok::l_brace, tok::l_square, |
| 547 | TT_TemplateOpener); |
| 548 | } |
| 549 | /// Returns whether \p Tok is )]} or a closing > of a template or in |
| 550 | /// protos. |
| 551 | bool closesScope() const { |
| 552 | if (is(TT_TemplateString) && TokenText.startswith("}")) |
| 553 | return true; |
| 554 | if (is(TT_DictLiteral) && is(tok::greater)) |
| 555 | return true; |
| 556 | return isOneOf(tok::r_paren, tok::r_brace, tok::r_square, |
| 557 | TT_TemplateCloser); |
| 558 | } |
| 559 | |
| 560 | /// Returns \c true if this is a "." or "->" accessing a member. |
| 561 | bool isMemberAccess() const { |
| 562 | return isOneOf(tok::arrow, tok::period, tok::arrowstar) && |
| 563 | !isOneOf(TT_DesignatedInitializerPeriod, TT_TrailingReturnArrow, |
| 564 | TT_LambdaArrow, TT_LeadingJavaAnnotation); |
| 565 | } |
| 566 | |
| 567 | bool isUnaryOperator() const { |
| 568 | switch (Tok.getKind()) { |
| 569 | case tok::plus: |
| 570 | case tok::plusplus: |
| 571 | case tok::minus: |
| 572 | case tok::minusminus: |
| 573 | case tok::exclaim: |
| 574 | case tok::tilde: |
| 575 | case tok::kw_sizeof: |
| 576 | case tok::kw_alignof: |
| 577 | return true; |
| 578 | default: |
| 579 | return false; |
| 580 | } |
| 581 | } |
| 582 | |
| 583 | bool isBinaryOperator() const { |
| 584 | // Comma is a binary operator, but does not behave as such wrt. formatting. |
| 585 | return getPrecedence() > prec::Comma; |
| 586 | } |
| 587 | |
| 588 | bool isTrailingComment() const { |
| 589 | return is(tok::comment) && |
| 590 | (is(TT_LineComment) || !Next || Next->NewlinesBefore > 0); |
| 591 | } |
| 592 | |
| 593 | /// Returns \c true if this is a keyword that can be used |
| 594 | /// like a function call (e.g. sizeof, typeid, ...). |
| 595 | bool isFunctionLikeKeyword() const { |
| 596 | switch (Tok.getKind()) { |
| 597 | case tok::kw_throw: |
| 598 | case tok::kw_typeid: |
| 599 | case tok::kw_return: |
| 600 | case tok::kw_sizeof: |
| 601 | case tok::kw_alignof: |
| 602 | case tok::kw_alignas: |
| 603 | case tok::kw_decltype: |
| 604 | case tok::kw_noexcept: |
| 605 | case tok::kw_static_assert: |
| 606 | case tok::kw__Atomic: |
| 607 | case tok::kw___attribute: |
| 608 | case tok::kw___underlying_type: |
| 609 | case tok::kw_requires: |
| 610 | return true; |
| 611 | default: |
| 612 | return false; |
| 613 | } |
| 614 | } |
| 615 | |
| 616 | /// Returns \c true if this is a string literal that's like a label, |
| 617 | /// e.g. ends with "=" or ":". |
| 618 | bool isLabelString() const { |
| 619 | if (!is(tok::string_literal)) |
| 620 | return false; |
| 621 | StringRef Content = TokenText; |
| 622 | if (Content.startswith("\"") || Content.startswith("'")) |
| 623 | Content = Content.drop_front(1); |
| 624 | if (Content.endswith("\"") || Content.endswith("'")) |
| 625 | Content = Content.drop_back(1); |
| 626 | Content = Content.trim(); |
| 627 | return Content.size() > 1 && |
| 628 | (Content.back() == ':' || Content.back() == '='); |
| 629 | } |
| 630 | |
| 631 | /// Returns actual token start location without leading escaped |
| 632 | /// newlines and whitespace. |
| 633 | /// |
| 634 | /// This can be different to Tok.getLocation(), which includes leading escaped |
| 635 | /// newlines. |
| 636 | SourceLocation getStartOfNonWhitespace() const { |
| 637 | return WhitespaceRange.getEnd(); |
| 638 | } |
| 639 | |
| 640 | prec::Level getPrecedence() const { |
| 641 | return getBinOpPrecedence(Tok.getKind(), /*GreaterThanIsOperator=*/true, |
| 642 | /*CPlusPlus11=*/true); |
| 643 | } |
| 644 | |
| 645 | /// Returns the previous token ignoring comments. |
| 646 | FormatToken *getPreviousNonComment() const { |
| 647 | FormatToken *Tok = Previous; |
| 648 | while (Tok && Tok->is(tok::comment)) |
| 649 | Tok = Tok->Previous; |
| 650 | return Tok; |
| 651 | } |
| 652 | |
| 653 | /// Returns the next token ignoring comments. |
| 654 | const FormatToken *getNextNonComment() const { |
| 655 | const FormatToken *Tok = Next; |
| 656 | while (Tok && Tok->is(tok::comment)) |
| 657 | Tok = Tok->Next; |
| 658 | return Tok; |
| 659 | } |
| 660 | |
| 661 | /// Returns \c true if this tokens starts a block-type list, i.e. a |
| 662 | /// list that should be indented with a block indent. |
| 663 | bool opensBlockOrBlockTypeList(const FormatStyle &Style) const { |
| 664 | // C# Does not indent object initialisers as continuations. |
| 665 | if (is(tok::l_brace) && getBlockKind() == BK_BracedInit && Style.isCSharp()) |
| 666 | return true; |
| 667 | if (is(TT_TemplateString) && opensScope()) |
| 668 | return true; |
| 669 | return is(TT_ArrayInitializerLSquare) || is(TT_ProtoExtensionLSquare) || |
| 670 | (is(tok::l_brace) && |
| 671 | (getBlockKind() == BK_Block || is(TT_DictLiteral) || |
| 672 | (!Style.Cpp11BracedListStyle && NestingLevel == 0))) || |
| 673 | (is(tok::less) && (Style.Language == FormatStyle::LK_Proto || |
| 674 | Style.Language == FormatStyle::LK_TextProto)); |
| 675 | } |
| 676 | |
| 677 | /// Returns whether the token is the left square bracket of a C++ |
| 678 | /// structured binding declaration. |
| 679 | bool isCppStructuredBinding(const FormatStyle &Style) const { |
| 680 | if (!Style.isCpp() || isNot(tok::l_square)) |
| 681 | return false; |
| 682 | const FormatToken *T = this; |
| 683 | do { |
| 684 | T = T->getPreviousNonComment(); |
| 685 | } while (T && T->isOneOf(tok::kw_const, tok::kw_volatile, tok::amp, |
| 686 | tok::ampamp)); |
| 687 | return T && T->is(tok::kw_auto); |
| 688 | } |
| 689 | |
| 690 | /// Same as opensBlockOrBlockTypeList, but for the closing token. |
| 691 | bool closesBlockOrBlockTypeList(const FormatStyle &Style) const { |
| 692 | if (is(TT_TemplateString) && closesScope()) |
| 693 | return true; |
| 694 | return MatchingParen && MatchingParen->opensBlockOrBlockTypeList(Style); |
| 695 | } |
| 696 | |
| 697 | /// Return the actual namespace token, if this token starts a namespace |
| 698 | /// block. |
| 699 | const FormatToken *getNamespaceToken() const { |
| 700 | const FormatToken *NamespaceTok = this; |
| 701 | if (is(tok::comment)) |
| 702 | NamespaceTok = NamespaceTok->getNextNonComment(); |
| 703 | // Detect "(inline|export)? namespace" in the beginning of a line. |
| 704 | if (NamespaceTok && NamespaceTok->isOneOf(tok::kw_inline, tok::kw_export)) |
| 705 | NamespaceTok = NamespaceTok->getNextNonComment(); |
| 706 | return NamespaceTok && |
| 707 | NamespaceTok->isOneOf(tok::kw_namespace, TT_NamespaceMacro) |
| 708 | ? NamespaceTok |
| 709 | : nullptr; |
| 710 | } |
| 711 | |
| 712 | void copyFrom(const FormatToken &Tok) { *this = Tok; } |
| 713 | |
| 714 | private: |
| 715 | // Only allow copying via the explicit copyFrom method. |
| 716 | FormatToken(const FormatToken &) = delete; |
| 717 | FormatToken &operator=(const FormatToken &) = default; |
| 718 | |
| 719 | template <typename A, typename... Ts> |
| 720 | bool startsSequenceInternal(A K1, Ts... Tokens) const { |
| 721 | if (is(tok::comment) && Next) |
| 722 | return Next->startsSequenceInternal(K1, Tokens...); |
| 723 | return is(K1) && Next && Next->startsSequenceInternal(Tokens...); |
| 724 | } |
| 725 | |
| 726 | template <typename A> bool startsSequenceInternal(A K1) const { |
| 727 | if (is(tok::comment) && Next) |
| 728 | return Next->startsSequenceInternal(K1); |
| 729 | return is(K1); |
| 730 | } |
| 731 | |
| 732 | template <typename A, typename... Ts> bool endsSequenceInternal(A K1) const { |
| 733 | if (is(tok::comment) && Previous) |
| 734 | return Previous->endsSequenceInternal(K1); |
| 735 | return is(K1); |
| 736 | } |
| 737 | |
| 738 | template <typename A, typename... Ts> |
| 739 | bool endsSequenceInternal(A K1, Ts... Tokens) const { |
| 740 | if (is(tok::comment) && Previous) |
| 741 | return Previous->endsSequenceInternal(K1, Tokens...); |
| 742 | return is(K1) && Previous && Previous->endsSequenceInternal(Tokens...); |
| 743 | } |
| 744 | }; |
| 745 | |
| 746 | class ContinuationIndenter; |
| 747 | struct LineState; |
| 748 | |
| 749 | class TokenRole { |
| 750 | public: |
| 751 | TokenRole(const FormatStyle &Style) : Style(Style) {} |
| 752 | virtual ~TokenRole(); |
| 753 | |
| 754 | /// After the \c TokenAnnotator has finished annotating all the tokens, |
| 755 | /// this function precomputes required information for formatting. |
| 756 | virtual void precomputeFormattingInfos(const FormatToken *Token); |
| 757 | |
| 758 | /// Apply the special formatting that the given role demands. |
| 759 | /// |
| 760 | /// Assumes that the token having this role is already formatted. |
| 761 | /// |
| 762 | /// Continues formatting from \p State leaving indentation to \p Indenter and |
| 763 | /// returns the total penalty that this formatting incurs. |
| 764 | virtual unsigned formatFromToken(LineState &State, |
| 765 | ContinuationIndenter *Indenter, |
| 766 | bool DryRun) { |
| 767 | return 0; |
| 768 | } |
| 769 | |
| 770 | /// Same as \c formatFromToken, but assumes that the first token has |
| 771 | /// already been set thereby deciding on the first line break. |
| 772 | virtual unsigned formatAfterToken(LineState &State, |
| 773 | ContinuationIndenter *Indenter, |
| 774 | bool DryRun) { |
| 775 | return 0; |
| 776 | } |
| 777 | |
| 778 | /// Notifies the \c Role that a comma was found. |
| 779 | virtual void CommaFound(const FormatToken *Token) {} |
| 780 | |
| 781 | virtual const FormatToken *lastComma() { return nullptr; } |
| 782 | |
| 783 | protected: |
| 784 | const FormatStyle &Style; |
| 785 | }; |
| 786 | |
| 787 | class CommaSeparatedList : public TokenRole { |
| 788 | public: |
| 789 | CommaSeparatedList(const FormatStyle &Style) |
| 790 | : TokenRole(Style), HasNestedBracedList(false) {} |
| 791 | |
| 792 | void precomputeFormattingInfos(const FormatToken *Token) override; |
| 793 | |
| 794 | unsigned formatAfterToken(LineState &State, ContinuationIndenter *Indenter, |
| 795 | bool DryRun) override; |
| 796 | |
| 797 | unsigned formatFromToken(LineState &State, ContinuationIndenter *Indenter, |
| 798 | bool DryRun) override; |
| 799 | |
| 800 | /// Adds \p Token as the next comma to the \c CommaSeparated list. |
| 801 | void CommaFound(const FormatToken *Token) override { |
| 802 | Commas.push_back(Token); |
| 803 | } |
| 804 | |
| 805 | const FormatToken *lastComma() override { |
| 806 | if (Commas.empty()) |
| 807 | return nullptr; |
| 808 | return Commas.back(); |
| 809 | } |
| 810 | |
| 811 | private: |
| 812 | /// A struct that holds information on how to format a given list with |
| 813 | /// a specific number of columns. |
| 814 | struct ColumnFormat { |
| 815 | /// The number of columns to use. |
| 816 | unsigned Columns; |
| 817 | |
| 818 | /// The total width in characters. |
| 819 | unsigned TotalWidth; |
| 820 | |
| 821 | /// The number of lines required for this format. |
| 822 | unsigned LineCount; |
| 823 | |
| 824 | /// The size of each column in characters. |
| 825 | SmallVector<unsigned, 8> ColumnSizes; |
| 826 | }; |
| 827 | |
| 828 | /// Calculate which \c ColumnFormat fits best into |
| 829 | /// \p RemainingCharacters. |
| 830 | const ColumnFormat *getColumnFormat(unsigned RemainingCharacters) const; |
| 831 | |
| 832 | /// The ordered \c FormatTokens making up the commas of this list. |
| 833 | SmallVector<const FormatToken *, 8> Commas; |
| 834 | |
| 835 | /// The length of each of the list's items in characters including the |
| 836 | /// trailing comma. |
| 837 | SmallVector<unsigned, 8> ItemLengths; |
| 838 | |
| 839 | /// Precomputed formats that can be used for this list. |
| 840 | SmallVector<ColumnFormat, 4> Formats; |
| 841 | |
| 842 | bool HasNestedBracedList; |
| 843 | }; |
| 844 | |
| 845 | /// Encapsulates keywords that are context sensitive or for languages not |
| 846 | /// properly supported by Clang's lexer. |
| 847 | struct AdditionalKeywords { |
| 848 | AdditionalKeywords(IdentifierTable &IdentTable) { |
| 849 | kw_final = &IdentTable.get("final"); |
| 850 | kw_override = &IdentTable.get("override"); |
| 851 | kw_in = &IdentTable.get("in"); |
| 852 | kw_of = &IdentTable.get("of"); |
| 853 | kw_CF_CLOSED_ENUM = &IdentTable.get("CF_CLOSED_ENUM"); |
| 854 | kw_CF_ENUM = &IdentTable.get("CF_ENUM"); |
| 855 | kw_CF_OPTIONS = &IdentTable.get("CF_OPTIONS"); |
| 856 | kw_NS_CLOSED_ENUM = &IdentTable.get("NS_CLOSED_ENUM"); |
| 857 | kw_NS_ENUM = &IdentTable.get("NS_ENUM"); |
| 858 | kw_NS_OPTIONS = &IdentTable.get("NS_OPTIONS"); |
| 859 | |
| 860 | kw_as = &IdentTable.get("as"); |
| 861 | kw_async = &IdentTable.get("async"); |
| 862 | kw_await = &IdentTable.get("await"); |
| 863 | kw_declare = &IdentTable.get("declare"); |
| 864 | kw_finally = &IdentTable.get("finally"); |
| 865 | kw_from = &IdentTable.get("from"); |
| 866 | kw_function = &IdentTable.get("function"); |
| 867 | kw_get = &IdentTable.get("get"); |
| 868 | kw_import = &IdentTable.get("import"); |
| 869 | kw_infer = &IdentTable.get("infer"); |
| 870 | kw_is = &IdentTable.get("is"); |
| 871 | kw_let = &IdentTable.get("let"); |
| 872 | kw_module = &IdentTable.get("module"); |
| 873 | kw_readonly = &IdentTable.get("readonly"); |
| 874 | kw_set = &IdentTable.get("set"); |
| 875 | kw_type = &IdentTable.get("type"); |
| 876 | kw_typeof = &IdentTable.get("typeof"); |
| 877 | kw_var = &IdentTable.get("var"); |
| 878 | kw_yield = &IdentTable.get("yield"); |
| 879 | |
| 880 | kw_abstract = &IdentTable.get("abstract"); |
| 881 | kw_assert = &IdentTable.get("assert"); |
| 882 | kw_extends = &IdentTable.get("extends"); |
| 883 | kw_implements = &IdentTable.get("implements"); |
| 884 | kw_instanceof = &IdentTable.get("instanceof"); |
| 885 | kw_interface = &IdentTable.get("interface"); |
| 886 | kw_native = &IdentTable.get("native"); |
| 887 | kw_package = &IdentTable.get("package"); |
| 888 | kw_synchronized = &IdentTable.get("synchronized"); |
| 889 | kw_throws = &IdentTable.get("throws"); |
| 890 | kw___except = &IdentTable.get("__except"); |
| 891 | kw___has_include = &IdentTable.get("__has_include"); |
| 892 | kw___has_include_next = &IdentTable.get("__has_include_next"); |
| 893 | |
| 894 | kw_mark = &IdentTable.get("mark"); |
| 895 | |
| 896 | kw_extend = &IdentTable.get("extend"); |
| 897 | kw_option = &IdentTable.get("option"); |
| 898 | kw_optional = &IdentTable.get("optional"); |
| 899 | kw_repeated = &IdentTable.get("repeated"); |
| 900 | kw_required = &IdentTable.get("required"); |
| 901 | kw_returns = &IdentTable.get("returns"); |
| 902 | |
| 903 | kw_signals = &IdentTable.get("signals"); |
| 904 | kw_qsignals = &IdentTable.get("Q_SIGNALS"); |
| 905 | kw_slots = &IdentTable.get("slots"); |
| 906 | kw_qslots = &IdentTable.get("Q_SLOTS"); |
| 907 | |
| 908 | // C# keywords |
| 909 | kw_dollar = &IdentTable.get("dollar"); |
| 910 | kw_base = &IdentTable.get("base"); |
| 911 | kw_byte = &IdentTable.get("byte"); |
| 912 | kw_checked = &IdentTable.get("checked"); |
| 913 | kw_decimal = &IdentTable.get("decimal"); |
| 914 | kw_delegate = &IdentTable.get("delegate"); |
| 915 | kw_event = &IdentTable.get("event"); |
| 916 | kw_fixed = &IdentTable.get("fixed"); |
| 917 | kw_foreach = &IdentTable.get("foreach"); |
| 918 | kw_implicit = &IdentTable.get("implicit"); |
| 919 | kw_internal = &IdentTable.get("internal"); |
| 920 | kw_lock = &IdentTable.get("lock"); |
| 921 | kw_null = &IdentTable.get("null"); |
| 922 | kw_object = &IdentTable.get("object"); |
| 923 | kw_out = &IdentTable.get("out"); |
| 924 | kw_params = &IdentTable.get("params"); |
| 925 | kw_ref = &IdentTable.get("ref"); |
| 926 | kw_string = &IdentTable.get("string"); |
| 927 | kw_stackalloc = &IdentTable.get("stackalloc"); |
| 928 | kw_sbyte = &IdentTable.get("sbyte"); |
| 929 | kw_sealed = &IdentTable.get("sealed"); |
| 930 | kw_uint = &IdentTable.get("uint"); |
| 931 | kw_ulong = &IdentTable.get("ulong"); |
| 932 | kw_unchecked = &IdentTable.get("unchecked"); |
| 933 | kw_unsafe = &IdentTable.get("unsafe"); |
| 934 | kw_ushort = &IdentTable.get("ushort"); |
| 935 | kw_when = &IdentTable.get("when"); |
| 936 | kw_where = &IdentTable.get("where"); |
| 937 | |
| 938 | // Keep this at the end of the constructor to make sure everything here |
| 939 | // is |
| 940 | // already initialized. |
| 941 | JsExtraKeywords = std::unordered_set<IdentifierInfo *>( |
| 942 | {kw_as, kw_async, kw_await, kw_declare, kw_finally, kw_from, |
| 943 | kw_function, kw_get, kw_import, kw_is, kw_let, kw_module, kw_override, |
| 944 | kw_readonly, kw_set, kw_type, kw_typeof, kw_var, kw_yield, |
| 945 | // Keywords from the Java section. |
| 946 | kw_abstract, kw_extends, kw_implements, kw_instanceof, kw_interface}); |
| 947 | |
| 948 | CSharpExtraKeywords = std::unordered_set<IdentifierInfo *>( |
| 949 | {kw_base, kw_byte, kw_checked, kw_decimal, kw_delegate, kw_event, |
| 950 | kw_fixed, kw_foreach, kw_implicit, kw_in, kw_interface, kw_internal, |
| 951 | kw_is, kw_lock, kw_null, kw_object, kw_out, kw_override, kw_params, |
| 952 | kw_readonly, kw_ref, kw_string, kw_stackalloc, kw_sbyte, kw_sealed, |
| 953 | kw_uint, kw_ulong, kw_unchecked, kw_unsafe, kw_ushort, kw_when, |
| 954 | kw_where, |
| 955 | // Keywords from the JavaScript section. |
| 956 | kw_as, kw_async, kw_await, kw_declare, kw_finally, kw_from, |
| 957 | kw_function, kw_get, kw_import, kw_is, kw_let, kw_module, kw_readonly, |
| 958 | kw_set, kw_type, kw_typeof, kw_var, kw_yield, |
| 959 | // Keywords from the Java section. |
| 960 | kw_abstract, kw_extends, kw_implements, kw_instanceof, kw_interface}); |
| 961 | } |
| 962 | |
| 963 | // Context sensitive keywords. |
| 964 | IdentifierInfo *kw_final; |
| 965 | IdentifierInfo *kw_override; |
| 966 | IdentifierInfo *kw_in; |
| 967 | IdentifierInfo *kw_of; |
| 968 | IdentifierInfo *kw_CF_CLOSED_ENUM; |
| 969 | IdentifierInfo *kw_CF_ENUM; |
| 970 | IdentifierInfo *kw_CF_OPTIONS; |
| 971 | IdentifierInfo *kw_NS_CLOSED_ENUM; |
| 972 | IdentifierInfo *kw_NS_ENUM; |
| 973 | IdentifierInfo *kw_NS_OPTIONS; |
| 974 | IdentifierInfo *kw___except; |
| 975 | IdentifierInfo *kw___has_include; |
| 976 | IdentifierInfo *kw___has_include_next; |
| 977 | |
| 978 | // JavaScript keywords. |
| 979 | IdentifierInfo *kw_as; |
| 980 | IdentifierInfo *kw_async; |
| 981 | IdentifierInfo *kw_await; |
| 982 | IdentifierInfo *kw_declare; |
| 983 | IdentifierInfo *kw_finally; |
| 984 | IdentifierInfo *kw_from; |
| 985 | IdentifierInfo *kw_function; |
| 986 | IdentifierInfo *kw_get; |
| 987 | IdentifierInfo *kw_import; |
| 988 | IdentifierInfo *kw_infer; |
| 989 | IdentifierInfo *kw_is; |
| 990 | IdentifierInfo *kw_let; |
| 991 | IdentifierInfo *kw_module; |
| 992 | IdentifierInfo *kw_readonly; |
| 993 | IdentifierInfo *kw_set; |
| 994 | IdentifierInfo *kw_type; |
| 995 | IdentifierInfo *kw_typeof; |
| 996 | IdentifierInfo *kw_var; |
| 997 | IdentifierInfo *kw_yield; |
| 998 | |
| 999 | // Java keywords. |
| 1000 | IdentifierInfo *kw_abstract; |
| 1001 | IdentifierInfo *kw_assert; |
| 1002 | IdentifierInfo *kw_extends; |
| 1003 | IdentifierInfo *kw_implements; |
| 1004 | IdentifierInfo *kw_instanceof; |
| 1005 | IdentifierInfo *kw_interface; |
| 1006 | IdentifierInfo *kw_native; |
| 1007 | IdentifierInfo *kw_package; |
| 1008 | IdentifierInfo *kw_synchronized; |
| 1009 | IdentifierInfo *kw_throws; |
| 1010 | |
| 1011 | // Pragma keywords. |
| 1012 | IdentifierInfo *kw_mark; |
| 1013 | |
| 1014 | // Proto keywords. |
| 1015 | IdentifierInfo *kw_extend; |
| 1016 | IdentifierInfo *kw_option; |
| 1017 | IdentifierInfo *kw_optional; |
| 1018 | IdentifierInfo *kw_repeated; |
| 1019 | IdentifierInfo *kw_required; |
| 1020 | IdentifierInfo *kw_returns; |
| 1021 | |
| 1022 | // QT keywords. |
| 1023 | IdentifierInfo *kw_signals; |
| 1024 | IdentifierInfo *kw_qsignals; |
| 1025 | IdentifierInfo *kw_slots; |
| 1026 | IdentifierInfo *kw_qslots; |
| 1027 | |
| 1028 | // C# keywords |
| 1029 | IdentifierInfo *kw_dollar; |
| 1030 | IdentifierInfo *kw_base; |
| 1031 | IdentifierInfo *kw_byte; |
| 1032 | IdentifierInfo *kw_checked; |
| 1033 | IdentifierInfo *kw_decimal; |
| 1034 | IdentifierInfo *kw_delegate; |
| 1035 | IdentifierInfo *kw_event; |
| 1036 | IdentifierInfo *kw_fixed; |
| 1037 | IdentifierInfo *kw_foreach; |
| 1038 | IdentifierInfo *kw_implicit; |
| 1039 | IdentifierInfo *kw_internal; |
| 1040 | |
| 1041 | IdentifierInfo *kw_lock; |
| 1042 | IdentifierInfo *kw_null; |
| 1043 | IdentifierInfo *kw_object; |
| 1044 | IdentifierInfo *kw_out; |
| 1045 | |
| 1046 | IdentifierInfo *kw_params; |
| 1047 | |
| 1048 | IdentifierInfo *kw_ref; |
| 1049 | IdentifierInfo *kw_string; |
| 1050 | IdentifierInfo *kw_stackalloc; |
| 1051 | IdentifierInfo *kw_sbyte; |
| 1052 | IdentifierInfo *kw_sealed; |
| 1053 | IdentifierInfo *kw_uint; |
| 1054 | IdentifierInfo *kw_ulong; |
| 1055 | IdentifierInfo *kw_unchecked; |
| 1056 | IdentifierInfo *kw_unsafe; |
| 1057 | IdentifierInfo *kw_ushort; |
| 1058 | IdentifierInfo *kw_when; |
| 1059 | IdentifierInfo *kw_where; |
| 1060 | |
| 1061 | /// Returns \c true if \p Tok is a true JavaScript identifier, returns |
| 1062 | /// \c false if it is a keyword or a pseudo keyword. |
| 1063 | /// If \c AcceptIdentifierName is true, returns true not only for keywords, |
| 1064 | // but also for IdentifierName tokens (aka pseudo-keywords), such as |
| 1065 | // ``yield``. |
| 1066 | bool IsJavaScriptIdentifier(const FormatToken &Tok, |
| 1067 | bool AcceptIdentifierName = true) const { |
| 1068 | // Based on the list of JavaScript & TypeScript keywords here: |
| 1069 | // https://github.com/microsoft/TypeScript/blob/main/src/compiler/scanner.ts#L74 |
| 1070 | switch (Tok.Tok.getKind()) { |
| 1071 | case tok::kw_break: |
| 1072 | case tok::kw_case: |
| 1073 | case tok::kw_catch: |
| 1074 | case tok::kw_class: |
| 1075 | case tok::kw_continue: |
| 1076 | case tok::kw_const: |
| 1077 | case tok::kw_default: |
| 1078 | case tok::kw_delete: |
| 1079 | case tok::kw_do: |
| 1080 | case tok::kw_else: |
| 1081 | case tok::kw_enum: |
| 1082 | case tok::kw_export: |
| 1083 | case tok::kw_false: |
| 1084 | case tok::kw_for: |
| 1085 | case tok::kw_if: |
| 1086 | case tok::kw_import: |
| 1087 | case tok::kw_module: |
| 1088 | case tok::kw_new: |
| 1089 | case tok::kw_private: |
| 1090 | case tok::kw_protected: |
| 1091 | case tok::kw_public: |
| 1092 | case tok::kw_return: |
| 1093 | case tok::kw_static: |
| 1094 | case tok::kw_switch: |
| 1095 | case tok::kw_this: |
| 1096 | case tok::kw_throw: |
| 1097 | case tok::kw_true: |
| 1098 | case tok::kw_try: |
| 1099 | case tok::kw_typeof: |
| 1100 | case tok::kw_void: |
| 1101 | case tok::kw_while: |
| 1102 | // These are JS keywords that are lexed by LLVM/clang as keywords. |
| 1103 | return false; |
| 1104 | case tok::identifier: { |
| 1105 | // For identifiers, make sure they are true identifiers, excluding the |
| 1106 | // JavaScript pseudo-keywords (not lexed by LLVM/clang as keywords). |
| 1107 | bool IsPseudoKeyword = |
| 1108 | JsExtraKeywords.find(Tok.Tok.getIdentifierInfo()) != |
| 1109 | JsExtraKeywords.end(); |
| 1110 | return AcceptIdentifierName || !IsPseudoKeyword; |
| 1111 | } |
| 1112 | default: |
| 1113 | // Other keywords are handled in the switch below, to avoid problems due |
| 1114 | // to duplicate case labels when using the #include trick. |
| 1115 | break; |
| 1116 | } |
| 1117 | |
| 1118 | switch (Tok.Tok.getKind()) { |
| 1119 | // Handle C++ keywords not included above: these are all JS identifiers. |
| 1120 | #define KEYWORD(X, Y) case tok::kw_##X: |
| 1121 | #include "clang/Basic/TokenKinds.def" |
| 1122 | // #undef KEYWORD is not needed -- it's #undef-ed at the end of |
| 1123 | // TokenKinds.def |
| 1124 | return true; |
| 1125 | default: |
| 1126 | // All other tokens (punctuation etc) are not JS identifiers. |
| 1127 | return false; |
| 1128 | } |
| 1129 | } |
| 1130 | |
| 1131 | /// Returns \c true if \p Tok is a C# keyword, returns |
| 1132 | /// \c false if it is a anything else. |
| 1133 | bool isCSharpKeyword(const FormatToken &Tok) const { |
| 1134 | switch (Tok.Tok.getKind()) { |
| 1135 | case tok::kw_bool: |
| 1136 | case tok::kw_break: |
| 1137 | case tok::kw_case: |
| 1138 | case tok::kw_catch: |
| 1139 | case tok::kw_char: |
| 1140 | case tok::kw_class: |
| 1141 | case tok::kw_const: |
| 1142 | case tok::kw_continue: |
| 1143 | case tok::kw_default: |
| 1144 | case tok::kw_do: |
| 1145 | case tok::kw_double: |
| 1146 | case tok::kw_else: |
| 1147 | case tok::kw_enum: |
| 1148 | case tok::kw_explicit: |
| 1149 | case tok::kw_extern: |
| 1150 | case tok::kw_false: |
| 1151 | case tok::kw_float: |
| 1152 | case tok::kw_for: |
| 1153 | case tok::kw_goto: |
| 1154 | case tok::kw_if: |
| 1155 | case tok::kw_int: |
| 1156 | case tok::kw_long: |
| 1157 | case tok::kw_namespace: |
| 1158 | case tok::kw_new: |
| 1159 | case tok::kw_operator: |
| 1160 | case tok::kw_private: |
| 1161 | case tok::kw_protected: |
| 1162 | case tok::kw_public: |
| 1163 | case tok::kw_return: |
| 1164 | case tok::kw_short: |
| 1165 | case tok::kw_sizeof: |
| 1166 | case tok::kw_static: |
| 1167 | case tok::kw_struct: |
| 1168 | case tok::kw_switch: |
| 1169 | case tok::kw_this: |
| 1170 | case tok::kw_throw: |
| 1171 | case tok::kw_true: |
| 1172 | case tok::kw_try: |
| 1173 | case tok::kw_typeof: |
| 1174 | case tok::kw_using: |
| 1175 | case tok::kw_virtual: |
| 1176 | case tok::kw_void: |
| 1177 | case tok::kw_volatile: |
| 1178 | case tok::kw_while: |
| 1179 | return true; |
| 1180 | default: |
| 1181 | return Tok.is(tok::identifier) && |
| 1182 | CSharpExtraKeywords.find(Tok.Tok.getIdentifierInfo()) == |
| 1183 | CSharpExtraKeywords.end(); |
| 1184 | } |
| 1185 | } |
| 1186 | |
| 1187 | private: |
| 1188 | /// The JavaScript keywords beyond the C++ keyword set. |
| 1189 | std::unordered_set<IdentifierInfo *> JsExtraKeywords; |
| 1190 | |
| 1191 | /// The C# keywords beyond the C++ keyword set |
| 1192 | std::unordered_set<IdentifierInfo *> CSharpExtraKeywords; |
| 1193 | }; |
| 1194 | |
| 1195 | } // namespace format |
| 1196 | } // namespace clang |
| 1197 | |
| 1198 | #endif |