Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 1 | //===- TokensTest.cpp -----------------------------------------------------===// |
| 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/Tooling/Syntax/Tokens.h" |
| 10 | #include "clang/AST/ASTConsumer.h" |
| 11 | #include "clang/AST/Expr.h" |
| 12 | #include "clang/Basic/Diagnostic.h" |
| 13 | #include "clang/Basic/DiagnosticIDs.h" |
| 14 | #include "clang/Basic/DiagnosticOptions.h" |
| 15 | #include "clang/Basic/FileManager.h" |
| 16 | #include "clang/Basic/FileSystemOptions.h" |
| 17 | #include "clang/Basic/LLVM.h" |
| 18 | #include "clang/Basic/LangOptions.h" |
| 19 | #include "clang/Basic/SourceLocation.h" |
| 20 | #include "clang/Basic/SourceManager.h" |
| 21 | #include "clang/Basic/TokenKinds.def" |
| 22 | #include "clang/Basic/TokenKinds.h" |
| 23 | #include "clang/Frontend/CompilerInstance.h" |
| 24 | #include "clang/Frontend/FrontendAction.h" |
| 25 | #include "clang/Frontend/Utils.h" |
| 26 | #include "clang/Lex/Lexer.h" |
| 27 | #include "clang/Lex/PreprocessorOptions.h" |
| 28 | #include "clang/Lex/Token.h" |
| 29 | #include "clang/Tooling/Tooling.h" |
| 30 | #include "llvm/ADT/ArrayRef.h" |
| 31 | #include "llvm/ADT/IntrusiveRefCntPtr.h" |
| 32 | #include "llvm/ADT/None.h" |
| 33 | #include "llvm/ADT/Optional.h" |
| 34 | #include "llvm/ADT/STLExtras.h" |
| 35 | #include "llvm/ADT/StringRef.h" |
| 36 | #include "llvm/Support/FormatVariadic.h" |
| 37 | #include "llvm/Support/MemoryBuffer.h" |
| 38 | #include "llvm/Support/VirtualFileSystem.h" |
| 39 | #include "llvm/Support/raw_os_ostream.h" |
| 40 | #include "llvm/Support/raw_ostream.h" |
| 41 | #include "llvm/Testing/Support/Annotations.h" |
| 42 | #include "llvm/Testing/Support/SupportHelpers.h" |
| 43 | #include <cassert> |
| 44 | #include <cstdlib> |
| 45 | #include <gmock/gmock.h> |
| 46 | #include <gtest/gtest.h> |
| 47 | #include <memory> |
| 48 | #include <ostream> |
| 49 | #include <string> |
| 50 | |
| 51 | using namespace clang; |
| 52 | using namespace clang::syntax; |
| 53 | |
| 54 | using llvm::ValueIs; |
| 55 | using ::testing::AllOf; |
| 56 | using ::testing::Contains; |
| 57 | using ::testing::ElementsAre; |
Ilya Biryukov | 5aed309 | 2019-06-18 16:27:27 +0000 | [diff] [blame] | 58 | using ::testing::Field; |
Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 59 | using ::testing::Matcher; |
| 60 | using ::testing::Not; |
| 61 | using ::testing::StartsWith; |
| 62 | |
| 63 | namespace { |
| 64 | // Checks the passed ArrayRef<T> has the same begin() and end() iterators as the |
| 65 | // argument. |
| 66 | MATCHER_P(SameRange, A, "") { |
| 67 | return A.begin() == arg.begin() && A.end() == arg.end(); |
| 68 | } |
Ilya Biryukov | 5aed309 | 2019-06-18 16:27:27 +0000 | [diff] [blame] | 69 | |
| 70 | Matcher<TokenBuffer::Expansion> |
| 71 | IsExpansion(Matcher<llvm::ArrayRef<syntax::Token>> Spelled, |
| 72 | Matcher<llvm::ArrayRef<syntax::Token>> Expanded) { |
| 73 | return AllOf(Field(&TokenBuffer::Expansion::Spelled, Spelled), |
| 74 | Field(&TokenBuffer::Expansion::Expanded, Expanded)); |
| 75 | } |
Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 76 | // Matchers for syntax::Token. |
| 77 | MATCHER_P(Kind, K, "") { return arg.kind() == K; } |
| 78 | MATCHER_P2(HasText, Text, SourceMgr, "") { |
| 79 | return arg.text(*SourceMgr) == Text; |
| 80 | } |
| 81 | /// Checks the start and end location of a token are equal to SourceRng. |
| 82 | MATCHER_P(RangeIs, SourceRng, "") { |
| 83 | return arg.location() == SourceRng.first && |
| 84 | arg.endLocation() == SourceRng.second; |
| 85 | } |
| 86 | |
| 87 | class TokenCollectorTest : public ::testing::Test { |
| 88 | public: |
| 89 | /// Run the clang frontend, collect the preprocessed tokens from the frontend |
| 90 | /// invocation and store them in this->Buffer. |
| 91 | /// This also clears SourceManager before running the compiler. |
| 92 | void recordTokens(llvm::StringRef Code) { |
| 93 | class RecordTokens : public ASTFrontendAction { |
| 94 | public: |
| 95 | explicit RecordTokens(TokenBuffer &Result) : Result(Result) {} |
| 96 | |
| 97 | bool BeginSourceFileAction(CompilerInstance &CI) override { |
| 98 | assert(!Collector && "expected only a single call to BeginSourceFile"); |
| 99 | Collector.emplace(CI.getPreprocessor()); |
| 100 | return true; |
| 101 | } |
| 102 | void EndSourceFileAction() override { |
| 103 | assert(Collector && "BeginSourceFileAction was never called"); |
| 104 | Result = std::move(*Collector).consume(); |
| 105 | } |
| 106 | |
| 107 | std::unique_ptr<ASTConsumer> |
| 108 | CreateASTConsumer(CompilerInstance &CI, StringRef InFile) override { |
| 109 | return llvm::make_unique<ASTConsumer>(); |
| 110 | } |
| 111 | |
| 112 | private: |
| 113 | TokenBuffer &Result; |
| 114 | llvm::Optional<TokenCollector> Collector; |
| 115 | }; |
| 116 | |
| 117 | constexpr const char *FileName = "./input.cpp"; |
| 118 | FS->addFile(FileName, time_t(), llvm::MemoryBuffer::getMemBufferCopy("")); |
| 119 | // Prepare to run a compiler. |
| 120 | if (!Diags->getClient()) |
| 121 | Diags->setClient(new IgnoringDiagConsumer); |
| 122 | std::vector<const char *> Args = {"tok-test", "-std=c++03", "-fsyntax-only", |
| 123 | FileName}; |
| 124 | auto CI = createInvocationFromCommandLine(Args, Diags, FS); |
| 125 | assert(CI); |
| 126 | CI->getFrontendOpts().DisableFree = false; |
| 127 | CI->getPreprocessorOpts().addRemappedFile( |
| 128 | FileName, llvm::MemoryBuffer::getMemBufferCopy(Code).release()); |
| 129 | CompilerInstance Compiler; |
| 130 | Compiler.setInvocation(std::move(CI)); |
| 131 | Compiler.setDiagnostics(Diags.get()); |
| 132 | Compiler.setFileManager(FileMgr.get()); |
| 133 | Compiler.setSourceManager(SourceMgr.get()); |
| 134 | |
| 135 | this->Buffer = TokenBuffer(*SourceMgr); |
| 136 | RecordTokens Recorder(this->Buffer); |
| 137 | ASSERT_TRUE(Compiler.ExecuteAction(Recorder)) |
| 138 | << "failed to run the frontend"; |
| 139 | } |
| 140 | |
| 141 | /// Record the tokens and return a test dump of the resulting buffer. |
| 142 | std::string collectAndDump(llvm::StringRef Code) { |
| 143 | recordTokens(Code); |
| 144 | return Buffer.dumpForTests(); |
| 145 | } |
| 146 | |
| 147 | // Adds a file to the test VFS. |
| 148 | void addFile(llvm::StringRef Path, llvm::StringRef Contents) { |
| 149 | if (!FS->addFile(Path, time_t(), |
| 150 | llvm::MemoryBuffer::getMemBufferCopy(Contents))) { |
| 151 | ADD_FAILURE() << "could not add a file to VFS: " << Path; |
| 152 | } |
| 153 | } |
| 154 | |
| 155 | /// Add a new file, run syntax::tokenize() on it and return the results. |
| 156 | std::vector<syntax::Token> tokenize(llvm::StringRef Text) { |
| 157 | // FIXME: pass proper LangOptions. |
| 158 | return syntax::tokenize( |
| 159 | SourceMgr->createFileID(llvm::MemoryBuffer::getMemBufferCopy(Text)), |
| 160 | *SourceMgr, LangOptions()); |
| 161 | } |
| 162 | |
| 163 | // Specialized versions of matchers that hide the SourceManager from clients. |
| 164 | Matcher<syntax::Token> HasText(std::string Text) const { |
| 165 | return ::HasText(Text, SourceMgr.get()); |
| 166 | } |
| 167 | Matcher<syntax::Token> RangeIs(llvm::Annotations::Range R) const { |
| 168 | std::pair<SourceLocation, SourceLocation> Ls; |
| 169 | Ls.first = SourceMgr->getLocForStartOfFile(SourceMgr->getMainFileID()) |
| 170 | .getLocWithOffset(R.Begin); |
| 171 | Ls.second = SourceMgr->getLocForStartOfFile(SourceMgr->getMainFileID()) |
| 172 | .getLocWithOffset(R.End); |
| 173 | return ::RangeIs(Ls); |
| 174 | } |
| 175 | |
| 176 | /// Finds a subrange in O(n * m). |
| 177 | template <class T, class U, class Eq> |
| 178 | llvm::ArrayRef<T> findSubrange(llvm::ArrayRef<U> Subrange, |
| 179 | llvm::ArrayRef<T> Range, Eq F) { |
| 180 | for (auto Begin = Range.begin(); Begin < Range.end(); ++Begin) { |
| 181 | auto It = Begin; |
| 182 | for (auto ItSub = Subrange.begin(); |
| 183 | ItSub != Subrange.end() && It != Range.end(); ++ItSub, ++It) { |
| 184 | if (!F(*ItSub, *It)) |
| 185 | goto continue_outer; |
| 186 | } |
| 187 | return llvm::makeArrayRef(Begin, It); |
| 188 | continue_outer:; |
| 189 | } |
| 190 | return llvm::makeArrayRef(Range.end(), Range.end()); |
| 191 | } |
| 192 | |
| 193 | /// Finds a subrange in \p Tokens that match the tokens specified in \p Query. |
| 194 | /// The match should be unique. \p Query is a whitespace-separated list of |
| 195 | /// tokens to search for. |
| 196 | llvm::ArrayRef<syntax::Token> |
| 197 | findTokenRange(llvm::StringRef Query, llvm::ArrayRef<syntax::Token> Tokens) { |
| 198 | llvm::SmallVector<llvm::StringRef, 8> QueryTokens; |
| 199 | Query.split(QueryTokens, ' ', /*MaxSplit=*/-1, /*KeepEmpty=*/false); |
| 200 | if (QueryTokens.empty()) { |
| 201 | ADD_FAILURE() << "will not look for an empty list of tokens"; |
| 202 | std::abort(); |
| 203 | } |
| 204 | // An equality test for search. |
| 205 | auto TextMatches = [this](llvm::StringRef Q, const syntax::Token &T) { |
| 206 | return Q == T.text(*SourceMgr); |
| 207 | }; |
| 208 | // Find a match. |
| 209 | auto Found = |
| 210 | findSubrange(llvm::makeArrayRef(QueryTokens), Tokens, TextMatches); |
| 211 | if (Found.begin() == Tokens.end()) { |
| 212 | ADD_FAILURE() << "could not find the subrange for " << Query; |
| 213 | std::abort(); |
| 214 | } |
| 215 | // Check that the match is unique. |
| 216 | if (findSubrange(llvm::makeArrayRef(QueryTokens), |
| 217 | llvm::makeArrayRef(Found.end(), Tokens.end()), TextMatches) |
| 218 | .begin() != Tokens.end()) { |
| 219 | ADD_FAILURE() << "match is not unique for " << Query; |
| 220 | std::abort(); |
| 221 | } |
| 222 | return Found; |
| 223 | }; |
| 224 | |
| 225 | // Specialized versions of findTokenRange for expanded and spelled tokens. |
| 226 | llvm::ArrayRef<syntax::Token> findExpanded(llvm::StringRef Query) { |
| 227 | return findTokenRange(Query, Buffer.expandedTokens()); |
| 228 | } |
| 229 | llvm::ArrayRef<syntax::Token> findSpelled(llvm::StringRef Query, |
| 230 | FileID File = FileID()) { |
| 231 | if (!File.isValid()) |
| 232 | File = SourceMgr->getMainFileID(); |
| 233 | return findTokenRange(Query, Buffer.spelledTokens(File)); |
| 234 | } |
| 235 | |
| 236 | // Data fields. |
| 237 | llvm::IntrusiveRefCntPtr<DiagnosticsEngine> Diags = |
| 238 | new DiagnosticsEngine(new DiagnosticIDs, new DiagnosticOptions); |
| 239 | IntrusiveRefCntPtr<llvm::vfs::InMemoryFileSystem> FS = |
| 240 | new llvm::vfs::InMemoryFileSystem; |
| 241 | llvm::IntrusiveRefCntPtr<FileManager> FileMgr = |
| 242 | new FileManager(FileSystemOptions(), FS); |
| 243 | llvm::IntrusiveRefCntPtr<SourceManager> SourceMgr = |
| 244 | new SourceManager(*Diags, *FileMgr); |
| 245 | /// Contains last result of calling recordTokens(). |
| 246 | TokenBuffer Buffer = TokenBuffer(*SourceMgr); |
| 247 | }; |
| 248 | |
| 249 | TEST_F(TokenCollectorTest, RawMode) { |
| 250 | EXPECT_THAT(tokenize("int main() {}"), |
| 251 | ElementsAre(Kind(tok::kw_int), |
| 252 | AllOf(HasText("main"), Kind(tok::identifier)), |
| 253 | Kind(tok::l_paren), Kind(tok::r_paren), |
| 254 | Kind(tok::l_brace), Kind(tok::r_brace))); |
| 255 | // Comments are ignored for now. |
| 256 | EXPECT_THAT(tokenize("/* foo */int a; // more comments"), |
| 257 | ElementsAre(Kind(tok::kw_int), |
| 258 | AllOf(HasText("a"), Kind(tok::identifier)), |
| 259 | Kind(tok::semi))); |
| 260 | } |
| 261 | |
| 262 | TEST_F(TokenCollectorTest, Basic) { |
| 263 | std::pair</*Input*/ std::string, /*Expected*/ std::string> TestCases[] = { |
| 264 | {"int main() {}", |
| 265 | R"(expanded tokens: |
| 266 | int main ( ) { } |
| 267 | file './input.cpp' |
| 268 | spelled tokens: |
| 269 | int main ( ) { } |
| 270 | no mappings. |
| 271 | )"}, |
| 272 | // All kinds of whitespace are ignored. |
| 273 | {"\t\n int\t\n main\t\n (\t\n )\t\n{\t\n }\t\n", |
| 274 | R"(expanded tokens: |
| 275 | int main ( ) { } |
| 276 | file './input.cpp' |
| 277 | spelled tokens: |
| 278 | int main ( ) { } |
| 279 | no mappings. |
| 280 | )"}, |
| 281 | // Annotation tokens are ignored. |
| 282 | {R"cpp( |
| 283 | #pragma GCC visibility push (public) |
| 284 | #pragma GCC visibility pop |
| 285 | )cpp", |
| 286 | R"(expanded tokens: |
| 287 | <empty> |
| 288 | file './input.cpp' |
| 289 | spelled tokens: |
| 290 | # pragma GCC visibility push ( public ) # pragma GCC visibility pop |
| 291 | mappings: |
| 292 | ['#'_0, '<eof>'_13) => ['<eof>'_0, '<eof>'_0) |
Ilya Biryukov | 26c066d | 2019-06-19 13:56:36 +0000 | [diff] [blame] | 293 | )"}, |
| 294 | // Empty files should not crash. |
| 295 | {R"cpp()cpp", R"(expanded tokens: |
| 296 | <empty> |
| 297 | file './input.cpp' |
| 298 | spelled tokens: |
| 299 | <empty> |
| 300 | no mappings. |
Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 301 | )"}}; |
| 302 | for (auto &Test : TestCases) |
| 303 | EXPECT_EQ(collectAndDump(Test.first), Test.second) |
| 304 | << collectAndDump(Test.first); |
| 305 | } |
| 306 | |
| 307 | TEST_F(TokenCollectorTest, Locations) { |
| 308 | // Check locations of the tokens. |
| 309 | llvm::Annotations Code(R"cpp( |
| 310 | $r1[[int]] $r2[[a]] $r3[[=]] $r4[["foo bar baz"]] $r5[[;]] |
| 311 | )cpp"); |
| 312 | recordTokens(Code.code()); |
| 313 | // Check expanded tokens. |
| 314 | EXPECT_THAT( |
| 315 | Buffer.expandedTokens(), |
| 316 | ElementsAre(AllOf(Kind(tok::kw_int), RangeIs(Code.range("r1"))), |
| 317 | AllOf(Kind(tok::identifier), RangeIs(Code.range("r2"))), |
| 318 | AllOf(Kind(tok::equal), RangeIs(Code.range("r3"))), |
| 319 | AllOf(Kind(tok::string_literal), RangeIs(Code.range("r4"))), |
| 320 | AllOf(Kind(tok::semi), RangeIs(Code.range("r5"))), |
| 321 | Kind(tok::eof))); |
| 322 | // Check spelled tokens. |
| 323 | EXPECT_THAT( |
| 324 | Buffer.spelledTokens(SourceMgr->getMainFileID()), |
| 325 | ElementsAre(AllOf(Kind(tok::kw_int), RangeIs(Code.range("r1"))), |
| 326 | AllOf(Kind(tok::identifier), RangeIs(Code.range("r2"))), |
| 327 | AllOf(Kind(tok::equal), RangeIs(Code.range("r3"))), |
| 328 | AllOf(Kind(tok::string_literal), RangeIs(Code.range("r4"))), |
| 329 | AllOf(Kind(tok::semi), RangeIs(Code.range("r5"))))); |
| 330 | } |
| 331 | |
| 332 | TEST_F(TokenCollectorTest, MacroDirectives) { |
| 333 | // Macro directives are not stored anywhere at the moment. |
| 334 | std::string Code = R"cpp( |
| 335 | #define FOO a |
| 336 | #include "unresolved_file.h" |
| 337 | #undef FOO |
| 338 | #ifdef X |
| 339 | #else |
| 340 | #endif |
| 341 | #ifndef Y |
| 342 | #endif |
| 343 | #if 1 |
| 344 | #elif 2 |
| 345 | #else |
| 346 | #endif |
| 347 | #pragma once |
| 348 | #pragma something lalala |
| 349 | |
| 350 | int a; |
| 351 | )cpp"; |
| 352 | std::string Expected = |
| 353 | "expanded tokens:\n" |
| 354 | " int a ;\n" |
| 355 | "file './input.cpp'\n" |
| 356 | " spelled tokens:\n" |
| 357 | " # define FOO a # include \"unresolved_file.h\" # undef FOO " |
| 358 | "# ifdef X # else # endif # ifndef Y # endif # if 1 # elif 2 # else " |
| 359 | "# endif # pragma once # pragma something lalala int a ;\n" |
| 360 | " mappings:\n" |
| 361 | " ['#'_0, 'int'_39) => ['int'_0, 'int'_0)\n"; |
| 362 | EXPECT_EQ(collectAndDump(Code), Expected); |
| 363 | } |
| 364 | |
| 365 | TEST_F(TokenCollectorTest, MacroReplacements) { |
| 366 | std::pair</*Input*/ std::string, /*Expected*/ std::string> TestCases[] = { |
| 367 | // A simple object-like macro. |
| 368 | {R"cpp( |
| 369 | #define INT int const |
| 370 | INT a; |
| 371 | )cpp", |
| 372 | R"(expanded tokens: |
| 373 | int const a ; |
| 374 | file './input.cpp' |
| 375 | spelled tokens: |
| 376 | # define INT int const INT a ; |
| 377 | mappings: |
| 378 | ['#'_0, 'INT'_5) => ['int'_0, 'int'_0) |
| 379 | ['INT'_5, 'a'_6) => ['int'_0, 'a'_2) |
| 380 | )"}, |
| 381 | // A simple function-like macro. |
| 382 | {R"cpp( |
| 383 | #define INT(a) const int |
| 384 | INT(10+10) a; |
| 385 | )cpp", |
| 386 | R"(expanded tokens: |
| 387 | const int a ; |
| 388 | file './input.cpp' |
| 389 | spelled tokens: |
| 390 | # define INT ( a ) const int INT ( 10 + 10 ) a ; |
| 391 | mappings: |
| 392 | ['#'_0, 'INT'_8) => ['const'_0, 'const'_0) |
| 393 | ['INT'_8, 'a'_14) => ['const'_0, 'a'_2) |
| 394 | )"}, |
| 395 | // Recursive macro replacements. |
| 396 | {R"cpp( |
| 397 | #define ID(X) X |
| 398 | #define INT int const |
| 399 | ID(ID(INT)) a; |
| 400 | )cpp", |
| 401 | R"(expanded tokens: |
| 402 | int const a ; |
| 403 | file './input.cpp' |
| 404 | spelled tokens: |
| 405 | # define ID ( X ) X # define INT int const ID ( ID ( INT ) ) a ; |
| 406 | mappings: |
| 407 | ['#'_0, 'ID'_12) => ['int'_0, 'int'_0) |
| 408 | ['ID'_12, 'a'_19) => ['int'_0, 'a'_2) |
| 409 | )"}, |
| 410 | // A little more complicated recursive macro replacements. |
| 411 | {R"cpp( |
| 412 | #define ADD(X, Y) X+Y |
| 413 | #define MULT(X, Y) X*Y |
| 414 | |
| 415 | int a = ADD(MULT(1,2), MULT(3,ADD(4,5))); |
| 416 | )cpp", |
| 417 | "expanded tokens:\n" |
| 418 | " int a = 1 * 2 + 3 * 4 + 5 ;\n" |
| 419 | "file './input.cpp'\n" |
| 420 | " spelled tokens:\n" |
| 421 | " # define ADD ( X , Y ) X + Y # define MULT ( X , Y ) X * Y int " |
| 422 | "a = ADD ( MULT ( 1 , 2 ) , MULT ( 3 , ADD ( 4 , 5 ) ) ) ;\n" |
| 423 | " mappings:\n" |
| 424 | " ['#'_0, 'int'_22) => ['int'_0, 'int'_0)\n" |
| 425 | " ['ADD'_25, ';'_46) => ['1'_3, ';'_12)\n"}, |
| 426 | // Empty macro replacement. |
Ilya Biryukov | 5e69f27 | 2019-06-24 21:39:51 +0000 | [diff] [blame] | 427 | // FIXME: the #define directives should not be glued together. |
Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 428 | {R"cpp( |
| 429 | #define EMPTY |
| 430 | #define EMPTY_FUNC(X) |
| 431 | EMPTY |
| 432 | EMPTY_FUNC(1+2+3) |
| 433 | )cpp", |
| 434 | R"(expanded tokens: |
| 435 | <empty> |
| 436 | file './input.cpp' |
| 437 | spelled tokens: |
| 438 | # define EMPTY # define EMPTY_FUNC ( X ) EMPTY EMPTY_FUNC ( 1 + 2 + 3 ) |
| 439 | mappings: |
Ilya Biryukov | 5e69f27 | 2019-06-24 21:39:51 +0000 | [diff] [blame] | 440 | ['#'_0, 'EMPTY'_9) => ['<eof>'_0, '<eof>'_0) |
| 441 | ['EMPTY'_9, 'EMPTY_FUNC'_10) => ['<eof>'_0, '<eof>'_0) |
| 442 | ['EMPTY_FUNC'_10, '<eof>'_18) => ['<eof>'_0, '<eof>'_0) |
Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 443 | )"}, |
| 444 | // File ends with a macro replacement. |
| 445 | {R"cpp( |
| 446 | #define FOO 10+10; |
| 447 | int a = FOO |
| 448 | )cpp", |
| 449 | R"(expanded tokens: |
| 450 | int a = 10 + 10 ; |
| 451 | file './input.cpp' |
| 452 | spelled tokens: |
| 453 | # define FOO 10 + 10 ; int a = FOO |
| 454 | mappings: |
| 455 | ['#'_0, 'int'_7) => ['int'_0, 'int'_0) |
| 456 | ['FOO'_10, '<eof>'_11) => ['10'_3, '<eof>'_7) |
| 457 | )"}}; |
| 458 | |
| 459 | for (auto &Test : TestCases) |
| 460 | EXPECT_EQ(Test.second, collectAndDump(Test.first)) |
| 461 | << collectAndDump(Test.first); |
| 462 | } |
| 463 | |
| 464 | TEST_F(TokenCollectorTest, SpecialTokens) { |
| 465 | // Tokens coming from concatenations. |
| 466 | recordTokens(R"cpp( |
| 467 | #define CONCAT(a, b) a ## b |
| 468 | int a = CONCAT(1, 2); |
| 469 | )cpp"); |
| 470 | EXPECT_THAT(std::vector<syntax::Token>(Buffer.expandedTokens()), |
| 471 | Contains(HasText("12"))); |
| 472 | // Multi-line tokens with slashes at the end. |
| 473 | recordTokens("i\\\nn\\\nt"); |
| 474 | EXPECT_THAT(Buffer.expandedTokens(), |
| 475 | ElementsAre(AllOf(Kind(tok::kw_int), HasText("i\\\nn\\\nt")), |
| 476 | Kind(tok::eof))); |
| 477 | // FIXME: test tokens with digraphs and UCN identifiers. |
| 478 | } |
| 479 | |
| 480 | TEST_F(TokenCollectorTest, LateBoundTokens) { |
| 481 | // The parser eventually breaks the first '>>' into two tokens ('>' and '>'), |
| 482 | // but we choose to record them as a single token (for now). |
| 483 | llvm::Annotations Code(R"cpp( |
| 484 | template <class T> |
| 485 | struct foo { int a; }; |
| 486 | int bar = foo<foo<int$br[[>>]]().a; |
| 487 | int baz = 10 $op[[>>]] 2; |
| 488 | )cpp"); |
| 489 | recordTokens(Code.code()); |
| 490 | EXPECT_THAT(std::vector<syntax::Token>(Buffer.expandedTokens()), |
| 491 | AllOf(Contains(AllOf(Kind(tok::greatergreater), |
| 492 | RangeIs(Code.range("br")))), |
| 493 | Contains(AllOf(Kind(tok::greatergreater), |
| 494 | RangeIs(Code.range("op")))))); |
| 495 | } |
| 496 | |
| 497 | TEST_F(TokenCollectorTest, DelayedParsing) { |
| 498 | llvm::StringLiteral Code = R"cpp( |
| 499 | struct Foo { |
| 500 | int method() { |
| 501 | // Parser will visit method bodies and initializers multiple times, but |
| 502 | // TokenBuffer should only record the first walk over the tokens; |
| 503 | return 100; |
| 504 | } |
| 505 | int a = 10; |
| 506 | |
| 507 | struct Subclass { |
| 508 | void foo() { |
| 509 | Foo().method(); |
| 510 | } |
| 511 | }; |
| 512 | }; |
| 513 | )cpp"; |
| 514 | std::string ExpectedTokens = |
| 515 | "expanded tokens:\n" |
| 516 | " struct Foo { int method ( ) { return 100 ; } int a = 10 ; struct " |
| 517 | "Subclass { void foo ( ) { Foo ( ) . method ( ) ; } } ; } ;\n"; |
| 518 | EXPECT_THAT(collectAndDump(Code), StartsWith(ExpectedTokens)); |
| 519 | } |
| 520 | |
| 521 | TEST_F(TokenCollectorTest, MultiFile) { |
| 522 | addFile("./foo.h", R"cpp( |
| 523 | #define ADD(X, Y) X+Y |
| 524 | int a = 100; |
| 525 | #include "bar.h" |
| 526 | )cpp"); |
| 527 | addFile("./bar.h", R"cpp( |
| 528 | int b = ADD(1, 2); |
| 529 | #define MULT(X, Y) X*Y |
| 530 | )cpp"); |
| 531 | llvm::StringLiteral Code = R"cpp( |
| 532 | #include "foo.h" |
| 533 | int c = ADD(1, MULT(2,3)); |
| 534 | )cpp"; |
| 535 | |
| 536 | std::string Expected = R"(expanded tokens: |
| 537 | int a = 100 ; int b = 1 + 2 ; int c = 1 + 2 * 3 ; |
| 538 | file './input.cpp' |
| 539 | spelled tokens: |
| 540 | # include "foo.h" int c = ADD ( 1 , MULT ( 2 , 3 ) ) ; |
| 541 | mappings: |
| 542 | ['#'_0, 'int'_3) => ['int'_12, 'int'_12) |
| 543 | ['ADD'_6, ';'_17) => ['1'_15, ';'_20) |
| 544 | file './foo.h' |
| 545 | spelled tokens: |
| 546 | # define ADD ( X , Y ) X + Y int a = 100 ; # include "bar.h" |
| 547 | mappings: |
| 548 | ['#'_0, 'int'_11) => ['int'_0, 'int'_0) |
| 549 | ['#'_16, '<eof>'_19) => ['int'_5, 'int'_5) |
| 550 | file './bar.h' |
| 551 | spelled tokens: |
| 552 | int b = ADD ( 1 , 2 ) ; # define MULT ( X , Y ) X * Y |
| 553 | mappings: |
| 554 | ['ADD'_3, ';'_9) => ['1'_8, ';'_11) |
| 555 | ['#'_10, '<eof>'_21) => ['int'_12, 'int'_12) |
| 556 | )"; |
| 557 | |
| 558 | EXPECT_EQ(Expected, collectAndDump(Code)) |
| 559 | << "input: " << Code << "\nresults: " << collectAndDump(Code); |
| 560 | } |
| 561 | |
| 562 | class TokenBufferTest : public TokenCollectorTest {}; |
| 563 | |
| 564 | TEST_F(TokenBufferTest, SpelledByExpanded) { |
| 565 | recordTokens(R"cpp( |
| 566 | a1 a2 a3 b1 b2 |
| 567 | )cpp"); |
| 568 | |
| 569 | // Sanity check: expanded and spelled tokens are stored separately. |
| 570 | EXPECT_THAT(findExpanded("a1 a2"), Not(SameRange(findSpelled("a1 a2")))); |
| 571 | // Searching for subranges of expanded tokens should give the corresponding |
| 572 | // spelled ones. |
| 573 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 b1 b2")), |
| 574 | ValueIs(SameRange(findSpelled("a1 a2 a3 b1 b2")))); |
| 575 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3")), |
| 576 | ValueIs(SameRange(findSpelled("a1 a2 a3")))); |
| 577 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("b1 b2")), |
| 578 | ValueIs(SameRange(findSpelled("b1 b2")))); |
| 579 | |
| 580 | // Test search on simple macro expansions. |
| 581 | recordTokens(R"cpp( |
| 582 | #define A a1 a2 a3 |
| 583 | #define B b1 b2 |
| 584 | |
| 585 | A split B |
| 586 | )cpp"); |
| 587 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 split b1 b2")), |
| 588 | ValueIs(SameRange(findSpelled("A split B")))); |
| 589 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3")), |
| 590 | ValueIs(SameRange(findSpelled("A split").drop_back()))); |
| 591 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("b1 b2")), |
| 592 | ValueIs(SameRange(findSpelled("split B").drop_front()))); |
| 593 | // Ranges not fully covering macro invocations should fail. |
| 594 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a1 a2")), llvm::None); |
| 595 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("b2")), llvm::None); |
| 596 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a2 a3 split b1 b2")), |
| 597 | llvm::None); |
| 598 | |
| 599 | // Recursive macro invocations. |
| 600 | recordTokens(R"cpp( |
| 601 | #define ID(x) x |
| 602 | #define B b1 b2 |
| 603 | |
| 604 | ID(ID(ID(a1) a2 a3)) split ID(B) |
| 605 | )cpp"); |
| 606 | |
| 607 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3")), |
| 608 | ValueIs(SameRange(findSpelled("ID ( ID ( ID ( a1 ) a2 a3 ) )")))); |
| 609 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("b1 b2")), |
| 610 | ValueIs(SameRange(findSpelled("ID ( B )")))); |
| 611 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 split b1 b2")), |
| 612 | ValueIs(SameRange(findSpelled( |
| 613 | "ID ( ID ( ID ( a1 ) a2 a3 ) ) split ID ( B )")))); |
| 614 | // Ranges crossing macro call boundaries. |
| 615 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a1 a2 a3 split b1")), |
| 616 | llvm::None); |
| 617 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a2 a3 split b1")), |
| 618 | llvm::None); |
| 619 | // FIXME: next two examples should map to macro arguments, but currently they |
| 620 | // fail. |
| 621 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a2")), llvm::None); |
| 622 | EXPECT_EQ(Buffer.spelledForExpanded(findExpanded("a1 a2")), llvm::None); |
| 623 | |
| 624 | // Empty macro expansions. |
| 625 | recordTokens(R"cpp( |
| 626 | #define EMPTY |
| 627 | #define ID(X) X |
| 628 | |
| 629 | EMPTY EMPTY ID(1 2 3) EMPTY EMPTY split1 |
| 630 | EMPTY EMPTY ID(4 5 6) split2 |
| 631 | ID(7 8 9) EMPTY EMPTY |
| 632 | )cpp"); |
| 633 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("1 2 3")), |
| 634 | ValueIs(SameRange(findSpelled("ID ( 1 2 3 )")))); |
| 635 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("4 5 6")), |
| 636 | ValueIs(SameRange(findSpelled("ID ( 4 5 6 )")))); |
| 637 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("7 8 9")), |
| 638 | ValueIs(SameRange(findSpelled("ID ( 7 8 9 )")))); |
| 639 | |
| 640 | // Empty mappings coming from various directives. |
| 641 | recordTokens(R"cpp( |
| 642 | #define ID(X) X |
| 643 | ID(1) |
| 644 | #pragma lalala |
| 645 | not_mapped |
| 646 | )cpp"); |
| 647 | EXPECT_THAT(Buffer.spelledForExpanded(findExpanded("not_mapped")), |
| 648 | ValueIs(SameRange(findSpelled("not_mapped")))); |
| 649 | } |
| 650 | |
Ilya Biryukov | 5aed309 | 2019-06-18 16:27:27 +0000 | [diff] [blame] | 651 | TEST_F(TokenBufferTest, ExpansionStartingAt) { |
| 652 | // Object-like macro expansions. |
| 653 | recordTokens(R"cpp( |
| 654 | #define FOO 3+4 |
| 655 | int a = FOO 1; |
| 656 | int b = FOO 2; |
| 657 | )cpp"); |
| 658 | |
| 659 | llvm::ArrayRef<syntax::Token> Foo1 = findSpelled("FOO 1").drop_back(); |
| 660 | EXPECT_THAT( |
| 661 | Buffer.expansionStartingAt(Foo1.data()), |
| 662 | ValueIs(IsExpansion(SameRange(Foo1), |
| 663 | SameRange(findExpanded("3 + 4 1").drop_back())))); |
| 664 | |
| 665 | llvm::ArrayRef<syntax::Token> Foo2 = findSpelled("FOO 2").drop_back(); |
| 666 | EXPECT_THAT( |
| 667 | Buffer.expansionStartingAt(Foo2.data()), |
| 668 | ValueIs(IsExpansion(SameRange(Foo2), |
| 669 | SameRange(findExpanded("3 + 4 2").drop_back())))); |
| 670 | |
| 671 | // Function-like macro expansions. |
| 672 | recordTokens(R"cpp( |
| 673 | #define ID(X) X |
| 674 | int a = ID(1+2+3); |
| 675 | int b = ID(ID(2+3+4)); |
| 676 | )cpp"); |
| 677 | |
| 678 | llvm::ArrayRef<syntax::Token> ID1 = findSpelled("ID ( 1 + 2 + 3 )"); |
| 679 | EXPECT_THAT(Buffer.expansionStartingAt(&ID1.front()), |
| 680 | ValueIs(IsExpansion(SameRange(ID1), |
| 681 | SameRange(findExpanded("1 + 2 + 3"))))); |
| 682 | // Only the first spelled token should be found. |
| 683 | for (const auto &T : ID1.drop_front()) |
| 684 | EXPECT_EQ(Buffer.expansionStartingAt(&T), llvm::None); |
| 685 | |
| 686 | llvm::ArrayRef<syntax::Token> ID2 = findSpelled("ID ( ID ( 2 + 3 + 4 ) )"); |
| 687 | EXPECT_THAT(Buffer.expansionStartingAt(&ID2.front()), |
| 688 | ValueIs(IsExpansion(SameRange(ID2), |
| 689 | SameRange(findExpanded("2 + 3 + 4"))))); |
| 690 | // Only the first spelled token should be found. |
| 691 | for (const auto &T : ID2.drop_front()) |
| 692 | EXPECT_EQ(Buffer.expansionStartingAt(&T), llvm::None); |
| 693 | |
| 694 | // PP directives. |
| 695 | recordTokens(R"cpp( |
| 696 | #define FOO 1 |
| 697 | int a = FOO; |
| 698 | #pragma once |
| 699 | int b = 1; |
| 700 | )cpp"); |
| 701 | |
| 702 | llvm::ArrayRef<syntax::Token> DefineFoo = findSpelled("# define FOO 1"); |
| 703 | EXPECT_THAT( |
| 704 | Buffer.expansionStartingAt(&DefineFoo.front()), |
| 705 | ValueIs(IsExpansion(SameRange(DefineFoo), |
| 706 | SameRange(findExpanded("int a").take_front(0))))); |
| 707 | // Only the first spelled token should be found. |
| 708 | for (const auto &T : DefineFoo.drop_front()) |
| 709 | EXPECT_EQ(Buffer.expansionStartingAt(&T), llvm::None); |
| 710 | |
| 711 | llvm::ArrayRef<syntax::Token> PragmaOnce = findSpelled("# pragma once"); |
| 712 | EXPECT_THAT( |
| 713 | Buffer.expansionStartingAt(&PragmaOnce.front()), |
| 714 | ValueIs(IsExpansion(SameRange(PragmaOnce), |
| 715 | SameRange(findExpanded("int b").take_front(0))))); |
| 716 | // Only the first spelled token should be found. |
| 717 | for (const auto &T : PragmaOnce.drop_front()) |
| 718 | EXPECT_EQ(Buffer.expansionStartingAt(&T), llvm::None); |
| 719 | } |
| 720 | |
Ilya Biryukov | e7230ea | 2019-05-22 14:44:45 +0000 | [diff] [blame] | 721 | TEST_F(TokenBufferTest, TokensToFileRange) { |
| 722 | addFile("./foo.h", "token_from_header"); |
| 723 | llvm::Annotations Code(R"cpp( |
| 724 | #define FOO token_from_expansion |
| 725 | #include "./foo.h" |
| 726 | $all[[$i[[int]] a = FOO;]] |
| 727 | )cpp"); |
| 728 | recordTokens(Code.code()); |
| 729 | |
| 730 | auto &SM = *SourceMgr; |
| 731 | |
| 732 | // Two simple examples. |
| 733 | auto Int = findExpanded("int").front(); |
| 734 | auto Semi = findExpanded(";").front(); |
| 735 | EXPECT_EQ(Int.range(SM), FileRange(SM.getMainFileID(), Code.range("i").Begin, |
| 736 | Code.range("i").End)); |
| 737 | EXPECT_EQ(syntax::Token::range(SM, Int, Semi), |
| 738 | FileRange(SM.getMainFileID(), Code.range("all").Begin, |
| 739 | Code.range("all").End)); |
| 740 | // We don't test assertion failures because death tests are slow. |
| 741 | } |
| 742 | |
Duncan P. N. Exon Smith | d4a9cae | 2019-05-25 22:38:02 +0000 | [diff] [blame] | 743 | } // namespace |