| 1 |
//===- unittests/Lex/LexerTest.cpp ------ Lexer tests ---------------------===// |
| 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 |
#include "clang/Lex/Lexer.h" |
| 10 |
#include "clang/Basic/Diagnostic.h" |
| 11 |
#include "clang/Basic/DiagnosticOptions.h" |
| 12 |
#include "clang/Basic/FileManager.h" |
| 13 |
#include "clang/Basic/LangOptions.h" |
| 14 |
#include "clang/Basic/SourceLocation.h" |
| 15 |
#include "clang/Basic/SourceManager.h" |
| 16 |
#include "clang/Basic/TargetInfo.h" |
| 17 |
#include "clang/Basic/TargetOptions.h" |
| 18 |
#include "clang/Basic/TokenKinds.h" |
| 19 |
#include "clang/Lex/HeaderSearch.h" |
| 20 |
#include "clang/Lex/HeaderSearchOptions.h" |
| 21 |
#include "clang/Lex/LiteralSupport.h" |
| 22 |
#include "clang/Lex/MacroArgs.h" |
| 23 |
#include "clang/Lex/MacroInfo.h" |
| 24 |
#include "clang/Lex/ModuleLoader.h" |
| 25 |
#include "clang/Lex/Preprocessor.h" |
| 26 |
#include "clang/Lex/PreprocessorOptions.h" |
| 27 |
#include "llvm/ADT/ArrayRef.h" |
| 28 |
#include "llvm/ADT/StringRef.h" |
| 29 |
#include "gmock/gmock.h" |
| 30 |
#include "gtest/gtest.h" |
| 31 |
#include <memory> |
| 32 |
#include <vector> |
| 33 |
|
| 34 |
namespace { |
| 35 |
using namespace clang; |
| 36 |
using testing::ElementsAre; |
| 37 |
|
| 38 |
// The test fixture. |
| 39 |
class LexerTest : public ::testing::Test { |
| 40 |
protected: |
| 41 |
LexerTest() |
| 42 |
: FileMgr(FileMgrOpts), |
| 43 |
DiagID(new DiagnosticIDs()), |
| 44 |
Diags(DiagID, new DiagnosticOptions, new IgnoringDiagConsumer()), |
| 45 |
SourceMgr(Diags, FileMgr), |
| 46 |
TargetOpts(new TargetOptions) |
| 47 |
{ |
| 48 |
TargetOpts->Triple = "x86_64-apple-darwin11.1.0"; |
| 49 |
Target = TargetInfo::CreateTargetInfo(Diags, TargetOpts); |
| 50 |
} |
| 51 |
|
| 52 |
std::unique_ptr<Preprocessor> CreatePP(StringRef Source, |
| 53 |
TrivialModuleLoader &ModLoader) { |
| 54 |
std::unique_ptr<llvm::MemoryBuffer> Buf = |
| 55 |
llvm::MemoryBuffer::getMemBuffer(Source); |
| 56 |
SourceMgr.setMainFileID(SourceMgr.createFileID(std::move(Buf))); |
| 57 |
|
| 58 |
HeaderSearch HeaderInfo(std::make_shared<HeaderSearchOptions>(), SourceMgr, |
| 59 |
Diags, LangOpts, Target.get()); |
| 60 |
std::unique_ptr<Preprocessor> PP = std::make_unique<Preprocessor>( |
| 61 |
std::make_shared<PreprocessorOptions>(), Diags, LangOpts, SourceMgr, |
| 62 |
HeaderInfo, ModLoader, |
| 63 |
/*IILookup =*/nullptr, |
| 64 |
/*OwnsHeaderSearch =*/false); |
| 65 |
PP->Initialize(*Target); |
| 66 |
PP->EnterMainSourceFile(); |
| 67 |
return PP; |
| 68 |
} |
| 69 |
|
| 70 |
std::vector<Token> Lex(StringRef Source) { |
| 71 |
TrivialModuleLoader ModLoader; |
| 72 |
PP = CreatePP(Source, ModLoader); |
| 73 |
|
| 74 |
std::vector<Token> toks; |
| 75 |
while (1) { |
| 76 |
Token tok; |
| 77 |
PP->Lex(tok); |
| 78 |
if (tok.is(tok::eof)) |
| 79 |
break; |
| 80 |
toks.push_back(tok); |
| 81 |
} |
| 82 |
|
| 83 |
return toks; |
| 84 |
} |
| 85 |
|
| 86 |
std::vector<Token> CheckLex(StringRef Source, |
| 87 |
ArrayRef<tok::TokenKind> ExpectedTokens) { |
| 88 |
auto toks = Lex(Source); |
| 89 |
EXPECT_EQ(ExpectedTokens.size(), toks.size()); |
| 90 |
for (unsigned i = 0, e = ExpectedTokens.size(); i != e; ++i) { |
| 91 |
EXPECT_EQ(ExpectedTokens[i], toks[i].getKind()); |
| 92 |
} |
| 93 |
|
| 94 |
return toks; |
| 95 |
} |
| 96 |
|
| 97 |
std::string getSourceText(Token Begin, Token End) { |
| 98 |
bool Invalid; |
| 99 |
StringRef Str = |
| 100 |
Lexer::getSourceText(CharSourceRange::getTokenRange(SourceRange( |
| 101 |
Begin.getLocation(), End.getLocation())), |
| 102 |
SourceMgr, LangOpts, &Invalid); |
| 103 |
if (Invalid) |
| 104 |
return "<INVALID>"; |
| 105 |
return std::string(Str); |
| 106 |
} |
| 107 |
|
| 108 |
FileSystemOptions FileMgrOpts; |
| 109 |
FileManager FileMgr; |
| 110 |
IntrusiveRefCntPtr<DiagnosticIDs> DiagID; |
| 111 |
DiagnosticsEngine Diags; |
| 112 |
SourceManager SourceMgr; |
| 113 |
LangOptions LangOpts; |
| 114 |
std::shared_ptr<TargetOptions> TargetOpts; |
| 115 |
IntrusiveRefCntPtr<TargetInfo> Target; |
| 116 |
std::unique_ptr<Preprocessor> PP; |
| 117 |
}; |
| 118 |
|
| 119 |
TEST_F(LexerTest, GetSourceTextExpandsToMaximumInMacroArgument) { |
| 120 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 121 |
ExpectedTokens.push_back(tok::identifier); |
| 122 |
ExpectedTokens.push_back(tok::l_paren); |
| 123 |
ExpectedTokens.push_back(tok::identifier); |
| 124 |
ExpectedTokens.push_back(tok::r_paren); |
| 125 |
|
| 126 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 127 |
"M(f(M(i)))", |
| 128 |
ExpectedTokens); |
| 129 |
|
| 130 |
EXPECT_EQ("M(i)", getSourceText(toks[2], toks[2])); |
| 131 |
} |
| 132 |
|
| 133 |
TEST_F(LexerTest, GetSourceTextExpandsToMaximumInMacroArgumentForEndOfMacro) { |
| 134 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 135 |
ExpectedTokens.push_back(tok::identifier); |
| 136 |
ExpectedTokens.push_back(tok::identifier); |
| 137 |
|
| 138 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 139 |
"M(M(i) c)", |
| 140 |
ExpectedTokens); |
| 141 |
|
| 142 |
EXPECT_EQ("M(i)", getSourceText(toks[0], toks[0])); |
| 143 |
} |
| 144 |
|
| 145 |
TEST_F(LexerTest, GetSourceTextExpandsInMacroArgumentForBeginOfMacro) { |
| 146 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 147 |
ExpectedTokens.push_back(tok::identifier); |
| 148 |
ExpectedTokens.push_back(tok::identifier); |
| 149 |
ExpectedTokens.push_back(tok::identifier); |
| 150 |
|
| 151 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 152 |
"M(c c M(i))", |
| 153 |
ExpectedTokens); |
| 154 |
|
| 155 |
EXPECT_EQ("c M(i)", getSourceText(toks[1], toks[2])); |
| 156 |
} |
| 157 |
|
| 158 |
TEST_F(LexerTest, GetSourceTextExpandsInMacroArgumentForEndOfMacro) { |
| 159 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 160 |
ExpectedTokens.push_back(tok::identifier); |
| 161 |
ExpectedTokens.push_back(tok::identifier); |
| 162 |
ExpectedTokens.push_back(tok::identifier); |
| 163 |
|
| 164 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 165 |
"M(M(i) c c)", |
| 166 |
ExpectedTokens); |
| 167 |
|
| 168 |
EXPECT_EQ("M(i) c", getSourceText(toks[0], toks[1])); |
| 169 |
} |
| 170 |
|
| 171 |
TEST_F(LexerTest, GetSourceTextInSeparateFnMacros) { |
| 172 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 173 |
ExpectedTokens.push_back(tok::identifier); |
| 174 |
ExpectedTokens.push_back(tok::identifier); |
| 175 |
ExpectedTokens.push_back(tok::identifier); |
| 176 |
ExpectedTokens.push_back(tok::identifier); |
| 177 |
|
| 178 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 179 |
"M(c M(i)) M(M(i) c)", |
| 180 |
ExpectedTokens); |
| 181 |
|
| 182 |
EXPECT_EQ("<INVALID>", getSourceText(toks[1], toks[2])); |
| 183 |
} |
| 184 |
|
| 185 |
TEST_F(LexerTest, GetSourceTextWorksAcrossTokenPastes) { |
| 186 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 187 |
ExpectedTokens.push_back(tok::identifier); |
| 188 |
ExpectedTokens.push_back(tok::l_paren); |
| 189 |
ExpectedTokens.push_back(tok::identifier); |
| 190 |
ExpectedTokens.push_back(tok::r_paren); |
| 191 |
|
| 192 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 193 |
"#define C(x) M(x##c)\n" |
| 194 |
"M(f(C(i)))", |
| 195 |
ExpectedTokens); |
| 196 |
|
| 197 |
EXPECT_EQ("C(i)", getSourceText(toks[2], toks[2])); |
| 198 |
} |
| 199 |
|
| 200 |
TEST_F(LexerTest, GetSourceTextExpandsAcrossMultipleMacroCalls) { |
| 201 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 202 |
ExpectedTokens.push_back(tok::identifier); |
| 203 |
ExpectedTokens.push_back(tok::l_paren); |
| 204 |
ExpectedTokens.push_back(tok::identifier); |
| 205 |
ExpectedTokens.push_back(tok::r_paren); |
| 206 |
|
| 207 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 208 |
"f(M(M(i)))", |
| 209 |
ExpectedTokens); |
| 210 |
EXPECT_EQ("M(M(i))", getSourceText(toks[2], toks[2])); |
| 211 |
} |
| 212 |
|
| 213 |
TEST_F(LexerTest, GetSourceTextInMiddleOfMacroArgument) { |
| 214 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 215 |
ExpectedTokens.push_back(tok::identifier); |
| 216 |
ExpectedTokens.push_back(tok::l_paren); |
| 217 |
ExpectedTokens.push_back(tok::identifier); |
| 218 |
ExpectedTokens.push_back(tok::r_paren); |
| 219 |
|
| 220 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 221 |
"M(f(i))", |
| 222 |
ExpectedTokens); |
| 223 |
EXPECT_EQ("i", getSourceText(toks[2], toks[2])); |
| 224 |
} |
| 225 |
|
| 226 |
TEST_F(LexerTest, GetSourceTextExpandsAroundDifferentMacroCalls) { |
| 227 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 228 |
ExpectedTokens.push_back(tok::identifier); |
| 229 |
ExpectedTokens.push_back(tok::l_paren); |
| 230 |
ExpectedTokens.push_back(tok::identifier); |
| 231 |
ExpectedTokens.push_back(tok::r_paren); |
| 232 |
|
| 233 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 234 |
"#define C(x) x\n" |
| 235 |
"f(C(M(i)))", |
| 236 |
ExpectedTokens); |
| 237 |
EXPECT_EQ("C(M(i))", getSourceText(toks[2], toks[2])); |
| 238 |
} |
| 239 |
|
| 240 |
TEST_F(LexerTest, GetSourceTextOnlyExpandsIfFirstTokenInMacro) { |
| 241 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 242 |
ExpectedTokens.push_back(tok::identifier); |
| 243 |
ExpectedTokens.push_back(tok::l_paren); |
| 244 |
ExpectedTokens.push_back(tok::identifier); |
| 245 |
ExpectedTokens.push_back(tok::identifier); |
| 246 |
ExpectedTokens.push_back(tok::r_paren); |
| 247 |
|
| 248 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 249 |
"#define C(x) c x\n" |
| 250 |
"f(C(M(i)))", |
| 251 |
ExpectedTokens); |
| 252 |
EXPECT_EQ("M(i)", getSourceText(toks[3], toks[3])); |
| 253 |
} |
| 254 |
|
| 255 |
TEST_F(LexerTest, GetSourceTextExpandsRecursively) { |
| 256 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 257 |
ExpectedTokens.push_back(tok::identifier); |
| 258 |
ExpectedTokens.push_back(tok::identifier); |
| 259 |
ExpectedTokens.push_back(tok::l_paren); |
| 260 |
ExpectedTokens.push_back(tok::identifier); |
| 261 |
ExpectedTokens.push_back(tok::r_paren); |
| 262 |
|
| 263 |
std::vector<Token> toks = CheckLex("#define M(x) x\n" |
| 264 |
"#define C(x) c M(x)\n" |
| 265 |
"C(f(M(i)))", |
| 266 |
ExpectedTokens); |
| 267 |
EXPECT_EQ("M(i)", getSourceText(toks[3], toks[3])); |
| 268 |
} |
| 269 |
|
| 270 |
TEST_F(LexerTest, LexAPI) { |
| 271 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 272 |
// Line 1 (after the #defines) |
| 273 |
ExpectedTokens.push_back(tok::l_square); |
| 274 |
ExpectedTokens.push_back(tok::identifier); |
| 275 |
ExpectedTokens.push_back(tok::r_square); |
| 276 |
ExpectedTokens.push_back(tok::l_square); |
| 277 |
ExpectedTokens.push_back(tok::identifier); |
| 278 |
ExpectedTokens.push_back(tok::r_square); |
| 279 |
// Line 2 |
| 280 |
ExpectedTokens.push_back(tok::identifier); |
| 281 |
ExpectedTokens.push_back(tok::identifier); |
| 282 |
ExpectedTokens.push_back(tok::identifier); |
| 283 |
ExpectedTokens.push_back(tok::identifier); |
| 284 |
|
| 285 |
std::vector<Token> toks = CheckLex("#define M(x) [x]\n" |
| 286 |
"#define N(x) x\n" |
| 287 |
"#define INN(x) x\n" |
| 288 |
"#define NOF1 INN(val)\n" |
| 289 |
"#define NOF2 val\n" |
| 290 |
"M(foo) N([bar])\n" |
| 291 |
"N(INN(val)) N(NOF1) N(NOF2) N(val)", |
| 292 |
ExpectedTokens); |
| 293 |
|
| 294 |
SourceLocation lsqrLoc = toks[0].getLocation(); |
| 295 |
SourceLocation idLoc = toks[1].getLocation(); |
| 296 |
SourceLocation rsqrLoc = toks[2].getLocation(); |
| 297 |
CharSourceRange macroRange = SourceMgr.getExpansionRange(lsqrLoc); |
| 298 |
|
| 299 |
SourceLocation Loc; |
| 300 |
EXPECT_TRUE(Lexer::isAtStartOfMacroExpansion(lsqrLoc, SourceMgr, LangOpts, &Loc)); |
| 301 |
EXPECT_EQ(Loc, macroRange.getBegin()); |
| 302 |
EXPECT_FALSE(Lexer::isAtStartOfMacroExpansion(idLoc, SourceMgr, LangOpts)); |
| 303 |
EXPECT_FALSE(Lexer::isAtEndOfMacroExpansion(idLoc, SourceMgr, LangOpts)); |
| 304 |
EXPECT_TRUE(Lexer::isAtEndOfMacroExpansion(rsqrLoc, SourceMgr, LangOpts, &Loc)); |
| 305 |
EXPECT_EQ(Loc, macroRange.getEnd()); |
| 306 |
EXPECT_TRUE(macroRange.isTokenRange()); |
| 307 |
|
| 308 |
CharSourceRange range = Lexer::makeFileCharRange( |
| 309 |
CharSourceRange::getTokenRange(lsqrLoc, idLoc), SourceMgr, LangOpts); |
| 310 |
EXPECT_TRUE(range.isInvalid()); |
| 311 |
range = Lexer::makeFileCharRange(CharSourceRange::getTokenRange(idLoc, rsqrLoc), |
| 312 |
SourceMgr, LangOpts); |
| 313 |
EXPECT_TRUE(range.isInvalid()); |
| 314 |
range = Lexer::makeFileCharRange(CharSourceRange::getTokenRange(lsqrLoc, rsqrLoc), |
| 315 |
SourceMgr, LangOpts); |
| 316 |
EXPECT_TRUE(!range.isTokenRange()); |
| 317 |
EXPECT_EQ(range.getAsRange(), |
| 318 |
SourceRange(macroRange.getBegin(), |
| 319 |
macroRange.getEnd().getLocWithOffset(1))); |
| 320 |
|
| 321 |
StringRef text = Lexer::getSourceText( |
| 322 |
CharSourceRange::getTokenRange(lsqrLoc, rsqrLoc), |
| 323 |
SourceMgr, LangOpts); |
| 324 |
EXPECT_EQ(text, "M(foo)"); |
| 325 |
|
| 326 |
SourceLocation macroLsqrLoc = toks[3].getLocation(); |
| 327 |
SourceLocation macroIdLoc = toks[4].getLocation(); |
| 328 |
SourceLocation macroRsqrLoc = toks[5].getLocation(); |
| 329 |
SourceLocation fileLsqrLoc = SourceMgr.getSpellingLoc(macroLsqrLoc); |
| 330 |
SourceLocation fileIdLoc = SourceMgr.getSpellingLoc(macroIdLoc); |
| 331 |
SourceLocation fileRsqrLoc = SourceMgr.getSpellingLoc(macroRsqrLoc); |
| 332 |
|
| 333 |
range = Lexer::makeFileCharRange( |
| 334 |
CharSourceRange::getTokenRange(macroLsqrLoc, macroIdLoc), |
| 335 |
SourceMgr, LangOpts); |
| 336 |
EXPECT_EQ(SourceRange(fileLsqrLoc, fileIdLoc.getLocWithOffset(3)), |
| 337 |
range.getAsRange()); |
| 338 |
|
| 339 |
range = Lexer::makeFileCharRange(CharSourceRange::getTokenRange(macroIdLoc, macroRsqrLoc), |
| 340 |
SourceMgr, LangOpts); |
| 341 |
EXPECT_EQ(SourceRange(fileIdLoc, fileRsqrLoc.getLocWithOffset(1)), |
| 342 |
range.getAsRange()); |
| 343 |
|
| 344 |
macroRange = SourceMgr.getExpansionRange(macroLsqrLoc); |
| 345 |
range = Lexer::makeFileCharRange( |
| 346 |
CharSourceRange::getTokenRange(macroLsqrLoc, macroRsqrLoc), |
| 347 |
SourceMgr, LangOpts); |
| 348 |
EXPECT_EQ(SourceRange(macroRange.getBegin(), macroRange.getEnd().getLocWithOffset(1)), |
| 349 |
range.getAsRange()); |
| 350 |
|
| 351 |
text = Lexer::getSourceText( |
| 352 |
CharSourceRange::getTokenRange(SourceRange(macroLsqrLoc, macroIdLoc)), |
| 353 |
SourceMgr, LangOpts); |
| 354 |
EXPECT_EQ(text, "[bar"); |
| 355 |
|
| 356 |
|
| 357 |
SourceLocation idLoc1 = toks[6].getLocation(); |
| 358 |
SourceLocation idLoc2 = toks[7].getLocation(); |
| 359 |
SourceLocation idLoc3 = toks[8].getLocation(); |
| 360 |
SourceLocation idLoc4 = toks[9].getLocation(); |
| 361 |
EXPECT_EQ("INN", Lexer::getImmediateMacroName(idLoc1, SourceMgr, LangOpts)); |
| 362 |
EXPECT_EQ("INN", Lexer::getImmediateMacroName(idLoc2, SourceMgr, LangOpts)); |
| 363 |
EXPECT_EQ("NOF2", Lexer::getImmediateMacroName(idLoc3, SourceMgr, LangOpts)); |
| 364 |
EXPECT_EQ("N", Lexer::getImmediateMacroName(idLoc4, SourceMgr, LangOpts)); |
| 365 |
} |
| 366 |
|
| 367 |
TEST_F(LexerTest, HandlesSplitTokens) { |
| 368 |
std::vector<tok::TokenKind> ExpectedTokens; |
| 369 |
// Line 1 (after the #defines) |
| 370 |
ExpectedTokens.push_back(tok::identifier); |
| 371 |
ExpectedTokens.push_back(tok::less); |
| 372 |
ExpectedTokens.push_back(tok::identifier); |
| 373 |
ExpectedTokens.push_back(tok::less); |
| 374 |
ExpectedTokens.push_back(tok::greatergreater); |
| 375 |
// Line 2 |
| 376 |
ExpectedTokens.push_back(tok::identifier); |
| 377 |
ExpectedTokens.push_back(tok::less); |
| 378 |
ExpectedTokens.push_back(tok::identifier); |
| 379 |
ExpectedTokens.push_back(tok::less); |
| 380 |
ExpectedTokens.push_back(tok::greatergreater); |
| 381 |
|
| 382 |
std::vector<Token> toks = CheckLex("#define TY ty\n" |
| 383 |
"#define RANGLE ty<ty<>>\n" |
| 384 |
"TY<ty<>>\n" |
| 385 |
"RANGLE", |
| 386 |
ExpectedTokens); |
| 387 |
|
| 388 |
SourceLocation outerTyLoc = toks[0].getLocation(); |
| 389 |
SourceLocation innerTyLoc = toks[2].getLocation(); |
| 390 |
SourceLocation gtgtLoc = toks[4].getLocation(); |
| 391 |
// Split the token to simulate the action of the parser and force creation of |
| 392 |
// an `ExpansionTokenRange`. |
| 393 |
SourceLocation rangleLoc = PP->SplitToken(gtgtLoc, 1); |
| 394 |
|
| 395 |
// Verify that it only captures the first greater-then and not the second one. |
| 396 |
CharSourceRange range = Lexer::makeFileCharRange( |
| 397 |
CharSourceRange::getTokenRange(innerTyLoc, rangleLoc), SourceMgr, |
| 398 |
LangOpts); |
| 399 |
EXPECT_TRUE(range.isCharRange()); |
| 400 |
EXPECT_EQ(range.getAsRange(), |
| 401 |
SourceRange(innerTyLoc, gtgtLoc.getLocWithOffset(1))); |
| 402 |
|
| 403 |
// Verify case where range begins in a macro expansion. |
| 404 |
range = Lexer::makeFileCharRange( |
| 405 |
CharSourceRange::getTokenRange(outerTyLoc, rangleLoc), SourceMgr, |
| 406 |
LangOpts); |
| 407 |
EXPECT_TRUE(range.isCharRange()); |
| 408 |
EXPECT_EQ(range.getAsRange(), |
| 409 |
SourceRange(SourceMgr.getExpansionLoc(outerTyLoc), |
| 410 |
gtgtLoc.getLocWithOffset(1))); |
| 411 |
|
| 412 |
SourceLocation macroInnerTyLoc = toks[7].getLocation(); |
| 413 |
SourceLocation macroGtgtLoc = toks[9].getLocation(); |
| 414 |
// Split the token to simulate the action of the parser and force creation of |
| 415 |
// an `ExpansionTokenRange`. |
| 416 |
SourceLocation macroRAngleLoc = PP->SplitToken(macroGtgtLoc, 1); |
| 417 |
|
| 418 |
// Verify that it fails (because it only captures the first greater-then and |
| 419 |
// not the second one, so it doesn't span the entire macro expansion). |
| 420 |
range = Lexer::makeFileCharRange( |
| 421 |
CharSourceRange::getTokenRange(macroInnerTyLoc, macroRAngleLoc), |
| 422 |
SourceMgr, LangOpts); |
| 423 |
EXPECT_TRUE(range.isInvalid()); |
| 424 |
} |
| 425 |
|
| 426 |
TEST_F(LexerTest, DontMergeMacroArgsFromDifferentMacroFiles) { |
| 427 |
std::vector<Token> toks = |
| 428 |
Lex("#define helper1 0\n" |
| 429 |
"void helper2(const char *, ...);\n" |
| 430 |
"#define M1(a, ...) helper2(a, ##__VA_ARGS__)\n" |
| 431 |
"#define M2(a, ...) M1(a, helper1, ##__VA_ARGS__)\n" |
| 432 |
"void f1() { M2(\"a\", \"b\"); }"); |
| 433 |
|
| 434 |
// Check the file corresponding to the "helper1" macro arg in M2. |
| 435 |
// |
| 436 |
// The lexer used to report its size as 31, meaning that the end of the |
| 437 |
// expansion would be on the *next line* (just past `M2("a", "b")`). Make |
| 438 |
// sure that we get the correct end location (the comma after "helper1"). |
| 439 |
SourceLocation helper1ArgLoc = toks[20].getLocation(); |
| 440 |
EXPECT_EQ(SourceMgr.getFileIDSize(SourceMgr.getFileID(helper1ArgLoc)), 8U); |
| 441 |
} |
| 442 |
|
| 443 |
TEST_F(LexerTest, DontOverallocateStringifyArgs) { |
| 444 |
TrivialModuleLoader ModLoader; |
| 445 |
auto PP = CreatePP("\"StrArg\", 5, 'C'", ModLoader); |
| 446 |
|
| 447 |
llvm::BumpPtrAllocator Allocator; |
| 448 |
std::array<IdentifierInfo *, 3> ParamList; |
| 449 |
MacroInfo *MI = PP->AllocateMacroInfo({}); |
| 450 |
MI->setIsFunctionLike(); |
| 451 |
MI->setParameterList(ParamList, Allocator); |
| 452 |
EXPECT_EQ(3u, MI->getNumParams()); |
| 453 |
EXPECT_TRUE(MI->isFunctionLike()); |
| 454 |
|
| 455 |
Token Eof; |
| 456 |
Eof.setKind(tok::eof); |
| 457 |
std::vector<Token> ArgTokens; |
| 458 |
while (1) { |
| 459 |
Token tok; |
| 460 |
PP->Lex(tok); |
| 461 |
if (tok.is(tok::eof)) { |
| 462 |
ArgTokens.push_back(Eof); |
| 463 |
break; |
| 464 |
} |
| 465 |
if (tok.is(tok::comma)) |
| 466 |
ArgTokens.push_back(Eof); |
| 467 |
else |
| 468 |
ArgTokens.push_back(tok); |
| 469 |
} |
| 470 |
|
| 471 |
auto MacroArgsDeleter = [&PP](MacroArgs *M) { M->destroy(*PP); }; |
| 472 |
std::unique_ptr<MacroArgs, decltype(MacroArgsDeleter)> MA( |
| 473 |
MacroArgs::create(MI, ArgTokens, false, *PP), MacroArgsDeleter); |
| 474 |
auto StringifyArg = [&](int ArgNo) { |
| 475 |
return MA->StringifyArgument(MA->getUnexpArgument(ArgNo), *PP, |
| 476 |
/*Charify=*/false, {}, {}); |
| 477 |
}; |
| 478 |
Token Result = StringifyArg(0); |
| 479 |
EXPECT_EQ(tok::string_literal, Result.getKind()); |
| 480 |
EXPECT_STREQ("\"\\\"StrArg\\\"\"", Result.getLiteralData()); |
| 481 |
Result = StringifyArg(1); |
| 482 |
EXPECT_EQ(tok::string_literal, Result.getKind()); |
| 483 |
EXPECT_STREQ("\"5\"", Result.getLiteralData()); |
| 484 |
Result = StringifyArg(2); |
| 485 |
EXPECT_EQ(tok::string_literal, Result.getKind()); |
| 486 |
EXPECT_STREQ("\"'C'\"", Result.getLiteralData()); |
| 487 |
#if !defined(NDEBUG) && GTEST_HAS_DEATH_TEST |
| 488 |
EXPECT_DEATH(StringifyArg(3), "Invalid arg #"); |
| 489 |
#endif |
| 490 |
} |
| 491 |
|
| 492 |
TEST_F(LexerTest, IsNewLineEscapedValid) { |
| 493 |
auto hasNewLineEscaped = [](const char *S) { |
| 494 |
return Lexer::isNewLineEscaped(S, S + strlen(S) - 1); |
| 495 |
}; |
| 496 |
|
| 497 |
EXPECT_TRUE(hasNewLineEscaped("\\\r")); |
| 498 |
EXPECT_TRUE(hasNewLineEscaped("\\\n")); |
| 499 |
EXPECT_TRUE(hasNewLineEscaped("\\\r\n")); |
| 500 |
EXPECT_TRUE(hasNewLineEscaped("\\\n\r")); |
| 501 |
EXPECT_TRUE(hasNewLineEscaped("\\ \t\v\f\r")); |
| 502 |
EXPECT_TRUE(hasNewLineEscaped("\\ \t\v\f\r\n")); |
| 503 |
|
| 504 |
EXPECT_FALSE(hasNewLineEscaped("\\\r\r")); |
| 505 |
EXPECT_FALSE(hasNewLineEscaped("\\\r\r\n")); |
| 506 |
EXPECT_FALSE(hasNewLineEscaped("\\\n\n")); |
| 507 |
EXPECT_FALSE(hasNewLineEscaped("\r")); |
| 508 |
EXPECT_FALSE(hasNewLineEscaped("\n")); |
| 509 |
EXPECT_FALSE(hasNewLineEscaped("\r\n")); |
| 510 |
EXPECT_FALSE(hasNewLineEscaped("\n\r")); |
| 511 |
EXPECT_FALSE(hasNewLineEscaped("\r\r")); |
| 512 |
EXPECT_FALSE(hasNewLineEscaped("\n\n")); |
| 513 |
} |
| 514 |
|
| 515 |
TEST_F(LexerTest, GetBeginningOfTokenWithEscapedNewLine) { |
| 516 |
// Each line should have the same length for |
| 517 |
// further offset calculation to be more straightforward. |
| 518 |
const unsigned IdentifierLength = 8; |
| 519 |
std::string TextToLex = "rabarbar\n" |
| 520 |
"foo\\\nbar\n" |
| 521 |
"foo\\\rbar\n" |
| 522 |
"fo\\\r\nbar\n" |
| 523 |
"foo\\\n\rba\n"; |
| 524 |
std::vector<tok::TokenKind> ExpectedTokens{5, tok::identifier}; |
| 525 |
std::vector<Token> LexedTokens = CheckLex(TextToLex, ExpectedTokens); |
| 526 |
|
| 527 |
for (const Token &Tok : LexedTokens) { |
| 528 |
std::pair<FileID, unsigned> OriginalLocation = |
| 529 |
SourceMgr.getDecomposedLoc(Tok.getLocation()); |
| 530 |
for (unsigned Offset = 0; Offset < IdentifierLength; ++Offset) { |
| 531 |
SourceLocation LookupLocation = |
| 532 |
Tok.getLocation().getLocWithOffset(Offset); |
| 533 |
|
| 534 |
std::pair<FileID, unsigned> FoundLocation = |
| 535 |
SourceMgr.getDecomposedExpansionLoc( |
| 536 |
Lexer::GetBeginningOfToken(LookupLocation, SourceMgr, LangOpts)); |
| 537 |
|
| 538 |
// Check that location returned by the GetBeginningOfToken |
| 539 |
// is the same as original token location reported by Lexer. |
| 540 |
EXPECT_EQ(FoundLocation.second, OriginalLocation.second); |
| 541 |
} |
| 542 |
} |
| 543 |
} |
| 544 |
|
| 545 |
TEST_F(LexerTest, AvoidPastEndOfStringDereference) { |
| 546 |
EXPECT_TRUE(Lex(" // \\\n").empty()); |
| 547 |
EXPECT_TRUE(Lex("#include <\\\\").empty()); |
| 548 |
EXPECT_TRUE(Lex("#include <\\\\\n").empty()); |
| 549 |
} |
| 550 |
|
| 551 |
TEST_F(LexerTest, StringizingRasString) { |
| 552 |
// For "std::string Lexer::Stringify(StringRef Str, bool Charify)". |
| 553 |
std::string String1 = R"(foo |
| 554 |
{"bar":[]} |
| 555 |
baz)"; |
| 556 |
// For "void Lexer::Stringify(SmallVectorImpl<char> &Str)". |
| 557 |
SmallString<128> String2; |
| 558 |
String2 += String1.c_str(); |
| 559 |
|
| 560 |
// Corner cases. |
| 561 |
std::string String3 = R"(\ |
| 562 |
\n |
| 563 |
\\n |
| 564 |
\\)"; |
| 565 |
SmallString<128> String4; |
| 566 |
String4 += String3.c_str(); |
| 567 |
std::string String5 = R"(a\ |
| 568 |
|
| 569 |
|
| 570 |
\\b)"; |
| 571 |
SmallString<128> String6; |
| 572 |
String6 += String5.c_str(); |
| 573 |
|
| 574 |
String1 = Lexer::Stringify(StringRef(String1)); |
| 575 |
Lexer::Stringify(String2); |
| 576 |
String3 = Lexer::Stringify(StringRef(String3)); |
| 577 |
Lexer::Stringify(String4); |
| 578 |
String5 = Lexer::Stringify(StringRef(String5)); |
| 579 |
Lexer::Stringify(String6); |
| 580 |
|
| 581 |
EXPECT_EQ(String1, R"(foo\n {\"bar\":[]}\n baz)"); |
| 582 |
EXPECT_EQ(String2, R"(foo\n {\"bar\":[]}\n baz)"); |
| 583 |
EXPECT_EQ(String3, R"(\\\n \\n\n \\\\n\n \\\\)"); |
| 584 |
EXPECT_EQ(String4, R"(\\\n \\n\n \\\\n\n \\\\)"); |
| 585 |
EXPECT_EQ(String5, R"(a\\\n\n\n \\\\b)"); |
| 586 |
EXPECT_EQ(String6, R"(a\\\n\n\n \\\\b)"); |
| 587 |
} |
| 588 |
|
| 589 |
TEST_F(LexerTest, CharRangeOffByOne) { |
| 590 |
std::vector<Token> toks = Lex(R"(#define MOO 1 |
| 591 |
void foo() { MOO; })"); |
| 592 |
const Token &moo = toks[5]; |
| 593 |
|
| 594 |
EXPECT_EQ(getSourceText(moo, moo), "MOO"); |
| 595 |
|
| 596 |
SourceRange R{moo.getLocation(), moo.getLocation()}; |
| 597 |
|
| 598 |
EXPECT_TRUE( |
| 599 |
Lexer::isAtStartOfMacroExpansion(R.getBegin(), SourceMgr, LangOpts)); |
| 600 |
EXPECT_TRUE( |
| 601 |
Lexer::isAtEndOfMacroExpansion(R.getEnd(), SourceMgr, LangOpts)); |
| 602 |
|
| 603 |
CharSourceRange CR = Lexer::getAsCharRange(R, SourceMgr, LangOpts); |
| 604 |
|
| 605 |
EXPECT_EQ(Lexer::getSourceText(CR, SourceMgr, LangOpts), "MOO"); // Was "MO". |
| 606 |
} |
| 607 |
|
| 608 |
TEST_F(LexerTest, FindNextToken) { |
| 609 |
Lex("int abcd = 0;\n" |
| 610 |
"int xyz = abcd;\n"); |
| 611 |
std::vector<std::string> GeneratedByNextToken; |
| 612 |
SourceLocation Loc = |
| 613 |
SourceMgr.getLocForStartOfFile(SourceMgr.getMainFileID()); |
| 614 |
while (true) { |
| 615 |
auto T = Lexer::findNextToken(Loc, SourceMgr, LangOpts); |
| 616 |
ASSERT_TRUE(T); |
| 617 |
if (T->is(tok::eof)) |
| 618 |
break; |
| 619 |
GeneratedByNextToken.push_back(getSourceText(*T, *T)); |
| 620 |
Loc = T->getLocation(); |
| 621 |
} |
| 622 |
EXPECT_THAT(GeneratedByNextToken, ElementsAre("abcd", "=", "0", ";", "int", |
| 623 |
"xyz", "=", "abcd", ";")); |
| 624 |
} |
| 625 |
|
| 626 |
TEST_F(LexerTest, CreatedFIDCountForPredefinedBuffer) { |
| 627 |
TrivialModuleLoader ModLoader; |
| 628 |
auto PP = CreatePP("", ModLoader); |
| 629 |
while (1) { |
| 630 |
Token tok; |
| 631 |
PP->Lex(tok); |
| 632 |
if (tok.is(tok::eof)) |
| 633 |
break; |
| 634 |
} |
| 635 |
EXPECT_EQ(SourceMgr.getNumCreatedFIDsForFileID(PP->getPredefinesFileID()), |
| 636 |
1U); |
| 637 |
} |
| 638 |
|
| 639 |
TEST_F(LexerTest, RawAndNormalLexSameForLineComments) { |
| 640 |
const llvm::StringLiteral Source = R"cpp( |
| 641 |
// First line comment. |
| 642 |
//* Second line comment which is ambigious. |
| 643 |
; // Have a non-comment token to make sure something is lexed. |
| 644 |
)cpp"; |
| 645 |
LangOpts.LineComment = false; |
| 646 |
auto Toks = Lex(Source); |
| 647 |
auto &SM = PP->getSourceManager(); |
| 648 |
auto SrcBuffer = SM.getBufferData(SM.getMainFileID()); |
| 649 |
Lexer L(SM.getLocForStartOfFile(SM.getMainFileID()), PP->getLangOpts(), |
| 650 |
SrcBuffer.data(), SrcBuffer.data(), |
| 651 |
SrcBuffer.data() + SrcBuffer.size()); |
| 652 |
|
| 653 |
auto ToksView = llvm::ArrayRef(Toks); |
| 654 |
clang::Token T; |
| 655 |
EXPECT_FALSE(ToksView.empty()); |
| 656 |
while (!L.LexFromRawLexer(T)) { |
| 657 |
ASSERT_TRUE(!ToksView.empty()); |
| 658 |
EXPECT_EQ(T.getKind(), ToksView.front().getKind()); |
| 659 |
ToksView = ToksView.drop_front(); |
| 660 |
} |
| 661 |
EXPECT_TRUE(ToksView.empty()); |
| 662 |
} |
| 663 |
} // anonymous namespace |