Good day, LLVMers! We are starting on our third (and hopefully last) round of testing for LLVM 3.0. Please visit: http://llvm.org/pre-releases/3.0/rc3/ for the sources. There are also binaries for Darwin up there, with more to come during the week. Please build this release candidate, test it out on your projects, and let us know if you find any regressions from the 2.9 release. Please keep in mind that we are not taking any more fixes except for those issues which are critical for the release. Share and enjoy! -bw
On Nov 7, 2011, at 2:00 PM, Bill Wendling wrote:> Good day, LLVMers! > > We are starting on our third (and hopefully last) round of testing for LLVM 3.0. Please visit: > > http://llvm.org/pre-releases/3.0/rc3/ > > for the sources. There are also binaries for Darwin up there, with more to come during the week. Please build this release candidate, test it out on your projects, and let us know if you find any regressions from the 2.9 release.Especially if you're running Linux, please give Clang in RC3 a spin, even if you previously tested RC2. We've updated the header-search logic to try to make Clang more robust against different system include directory layouts, and need help with more widespread testing. - Doug
Which files are needed to build clang? Is it clang-3.0rc2 and llvm-3.0rc3 or do I need the whole test suite? John Dr John P. Fletcher Tel: (44) 121 204 3389 (direct line), FAX: (44) 121 204 3678 Chemical Engineering and Applied Chemistry (CEAC), Associate Dean - External Relations, School of Engineering and Applied Science (EAS), Aston University, Aston Triangle, BIRMINGHAM B4 7ET U.K. -----Original Message----- From: cfe-dev-bounces at cs.uiuc.edu [mailto:cfe-dev-bounces at cs.uiuc.edu] On Behalf Of Douglas Gregor Sent: 07 November 2011 22:29 To: Bill Wendling Cc: cfe-dev at cs.uiuc.edu Developers; LLVM Developers Mailing List Subject: Re: [cfe-dev] LLVM 3.0rc3 Testing Beginning On Nov 7, 2011, at 2:00 PM, Bill Wendling wrote:> Good day, LLVMers! > > We are starting on our third (and hopefully last) round of testing for LLVM 3.0. Please visit: > > http://llvm.org/pre-releases/3.0/rc3/ > > for the sources. There are also binaries for Darwin up there, with more to come during the week. Please build this release candidate, test it out on your projects, and let us know if you find any regressions from the 2.9 release.Especially if you're running Linux, please give Clang in RC3 a spin, even if you previously tested RC2. We've updated the header-search logic to try to make Clang more robust against different system include directory layouts, and need help with more widespread testing. - Doug _______________________________________________ cfe-dev mailing list cfe-dev at cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
On 7 November 2011 22:00, Bill Wendling <wendling at apple.com> wrote:> We are starting on our third (and hopefully last) round of testing for LLVM 3.0. Please visit: > > http://llvm.org/pre-releases/3.0/rc3/ > > for the sources. There are also binaries for Darwin up there, with more to come during the week. Please build this release candidate, test it out on your projects, and let us know if you find any regressions from the 2.9 release.I've done a "make -j2 all check-all", llvm + clang, Release build, on Ubuntu Linux 11.04 x86_64. I got one test failure: Failing Tests (1): Clang-Unit :: AST/Release/ASTTests/APValue.Diagnostics I've pasted the output of the test below. Jay. FAIL: Clang-Unit :: AST/Release/ASTTests/APValue.Diagnostics (3931 of 9598) ******************** TEST 'Clang-Unit :: AST/Release/ASTTests/APValue.Diagnostics' FAILED ******************** Note: Google Test filter = APValue.Diagnostics [==========] Running 1 test from 1 test case. [----------] Global test environment set-up. [----------] 1 test from APValue [ RUN ] APValue.Diagnostics /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:62: Failure Value of: GetDiagnosticOutput(APValue()) Actual: "" Expected: "Uninitialized" /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:63: Failure Value of: GetDiagnosticOutput(APValue(APSInt(APInt(16, 5)))) Actual: "" Expected: "5" /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:66: Failure Value of: GetDiagnosticOutput(APValue(APFloat(APFloat::IEEEdouble, "3.14159"))) Actual: "" Expected: "3.141590e+00" /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:69: Failure Value of: GetDiagnosticOutput(APValue(APSInt(APInt(16, 3)), APSInt(APInt(16, 4)))) Actual: "" Expected: "3+4i" /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:73: Failure Value of: GetDiagnosticOutput(APValue( APFloat(APFloat::IEEEdouble, "3.2"), APFloat(APFloat::IEEEdouble, "5.7"))) Actual: "" Expected: "3.200000e+00+5.700000e+00i" /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:80: Failure Value of: GetDiagnosticOutput(APValue(V, array_lengthof(V))) Actual: "" Expected: "[3, 4, 5]" [ FAILED ] APValue.Diagnostics (1 ms) [----------] 1 test from APValue (16 ms total) [----------] Global test environment tear-down [==========] 1 test from 1 test case ran. (16 ms total) [ PASSED ] 0 tests. [ FAILED ] 1 test, listed below: [ FAILED ] APValue.Diagnostics 1 FAILED TEST
On Nov 8, 2011, at 7:18 AM, Jay Foad wrote:> On 7 November 2011 22:00, Bill Wendling <wendling at apple.com> wrote: >> We are starting on our third (and hopefully last) round of testing for LLVM 3.0. Please visit: >> >> http://llvm.org/pre-releases/3.0/rc3/ >> >> for the sources. There are also binaries for Darwin up there, with more to come during the week. Please build this release candidate, test it out on your projects, and let us know if you find any regressions from the 2.9 release. > > I've done a "make -j2 all check-all", llvm + clang, Release build, on > Ubuntu Linux 11.04 x86_64. I got one test failure: > > Failing Tests (1): > Clang-Unit :: AST/Release/ASTTests/APValue.Diagnostics > > I've pasted the output of the test below. >Jay, Did you see these failures for rc2? Chandler, Could these be related to your changes? -bw> FAIL: Clang-Unit :: AST/Release/ASTTests/APValue.Diagnostics (3931 of 9598) > ******************** TEST 'Clang-Unit :: > AST/Release/ASTTests/APValue.Diagnostics' FAILED ******************** > Note: Google Test filter = APValue.Diagnostics > [==========] Running 1 test from 1 test case. > [----------] Global test environment set-up. > [----------] 1 test from APValue > [ RUN ] APValue.Diagnostics > /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:62: Failure > Value of: GetDiagnosticOutput(APValue()) > Actual: "" > Expected: "Uninitialized" > /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:63: Failure > Value of: GetDiagnosticOutput(APValue(APSInt(APInt(16, 5)))) > Actual: "" > Expected: "5" > /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:66: Failure > Value of: GetDiagnosticOutput(APValue(APFloat(APFloat::IEEEdouble, "3.14159"))) > Actual: "" > Expected: "3.141590e+00" > /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:69: Failure > Value of: GetDiagnosticOutput(APValue(APSInt(APInt(16, 3)), > APSInt(APInt(16, 4)))) > Actual: "" > Expected: "3+4i" > /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:73: Failure > Value of: GetDiagnosticOutput(APValue( APFloat(APFloat::IEEEdouble, > "3.2"), APFloat(APFloat::IEEEdouble, "5.7"))) > Actual: "" > Expected: "3.200000e+00+5.700000e+00i" > /home/jay/llvm/clang-3.0rc3.src/unittests/AST/APValueTest.cpp:80: Failure > Value of: GetDiagnosticOutput(APValue(V, array_lengthof(V))) > Actual: "" > Expected: "[3, 4, 5]" > [ FAILED ] APValue.Diagnostics (1 ms) > [----------] 1 test from APValue (16 ms total) > > [----------] Global test environment tear-down > [==========] 1 test from 1 test case ran. (16 ms total) > [ PASSED ] 0 tests. > [ FAILED ] 1 test, listed below: > [ FAILED ] APValue.Diagnostics > > 1 FAILED TEST
On 7 November 2011 22:00, Bill Wendling <wendling at apple.com> wrote:> We are starting on our third (and hopefully last) round of testing for LLVM 3.0. Please visit: > > http://llvm.org/pre-releases/3.0/rc3/ > > for the sources. There are also binaries for Darwin up there, with > more to come during the week. Please build this release candidate, > test it out on your projects, and let us know if you find any > regressions from the 2.9 release.On Linux Debian testing, gcc-4.6.1: Built fine, with warnings listed below make check-all passed I couldn't figure out how to run the test-suite. Are there up-to-date instructions somewhere? llvm-gcc doesn't exist anymore... -Dave I get the following warnings building llvm+clang (default configure, default make target): /home/greened/src/llvm-30-rc3/llvm/lib/Support/Mutex.cpp: In constructor ‘llvm::sys::MutexImpl::MutexImpl(bool)’: /home/greened/src/llvm-30-rc3/llvm/lib/Support/Mutex.cpp:69:9: warning: variable ‘errorcode’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp: In member function ‘llvm::SDValue {anonymous}::SelectionDAGLegalize::LegalizeOp(llvm::SDValue)’: /home/greened/src/llvm-30-rc3/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:1652:19: warning: variable ‘EltVT’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:1632:15: warning: variable ‘WideScalarVT’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/llvm/lib/Target/ARM/ARMISelDAGToDAG.cpp: In member function ‘llvm::SDNode* {anonymous}::ARMDAGToDAGISel::SelectABSOp(llvm::SDNode*)’: /home/greened/src/llvm-30-rc3/llvm/lib/Target/ARM/ARMISelDAGToDAG.cpp:2319:12: warning: variable ‘DL’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/llvm/lib/Target/ARM/AsmParser/ARMAsmParser.cpp: In member function ‘bool {anonymous}::ARMAsmParser::tryParseRegisterWithWriteBack(llvm::SmallVectorImpl<llvm::MCParsedAsmOperand*>&)’: /home/greened/src/llvm-30-rc3/llvm/lib/Target/ARM/AsmParser/ARMAsmParser.cpp:1979:11: warning: variable ‘ExprLoc’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/llvm/lib/Target/ARM/Disassembler/ARMDisassembler.cpp: In function ‘DecodeStatus DecodeCopMemInstruction(llvm::MCInst&, unsigned int, uint64_t, const void*)’: /home/greened/src/llvm-30-rc3/llvm/lib/Target/ARM/Disassembler/ARMDisassembler.cpp:1241:12: warning: variable ‘idx_mode’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Lex/Lexer.cpp: In static member function ‘static clang::SourceLocation clang::Lexer::GetBeginningOfToken(clang::SourceLocation, const clang::SourceManager&, const clang::LangOptions&)’: /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Lex/Lexer.cpp:489:30: warning: variable ‘FileLocInfo’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Lex/Lexer.cpp:490:30: warning: variable ‘BeginFileLocInfo’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/AST/Decl.cpp: In member function ‘unsigned int clang::FieldDecl::getFieldIndex() const’: /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/AST/Decl.cpp:2209:53: warning: variable ‘e’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Sema/SemaExpr.cpp: In member function ‘clang::QualType clang::Sema::CheckVectorCompareOperands(clang::ExprResult&, clang::ExprResult&, clang::SourceLocation, bool)’: /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Sema/SemaExpr.cpp:6698:12: warning: variable ‘RHSType’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/StaticAnalyzer/Core/RegionStore.cpp: In member function ‘clang::ento::SVal {anonymous}::RegionStoreManager::RetrieveStruct(clang::ento::Store, const clang::ento::TypedValueRegion*)’: /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/StaticAnalyzer/Core/RegionStore.cpp:1280:12: warning: variable ‘T’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Sema/SemaTemplateInstantiateDecl.cpp: In member function ‘clang::NamedDecl* clang::Sema::FindInstantiatedDecl(clang::SourceLocation, clang::NamedDecl*, const clang::MultiLevelTemplateArgumentList&)’: /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/Sema/SemaTemplateInstantiateDecl.cpp:3170:12: warning: variable ‘SawNonDependentContext’ set but not used [-Wunused-but-set-variable] /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/CodeGen/CGRecordLayoutBuilder.cpp: In member function ‘void {anonymous}::CGRecordLayoutBuilder::LayoutUnion(const clang::RecordDecl*)’: /home/greened/src/llvm-30-rc3/clang-3.0rc3.src/lib/CodeGen/CGRecordLayoutBuilder.cpp:528:8: warning: variable ‘hasOnlyZeroSizedBitFields’ set but not used [-Wunused-but-set-variable]
On 8 November 2011 15:18, Jay Foad <jay.foad at gmail.com> wrote:> I've done a "make -j2 all check-all", llvm + clang, Release build, on > Ubuntu Linux 11.04 x86_64. I got one test failure: > > Failing Tests (1): > Clang-Unit :: AST/Release/ASTTests/APValue.DiagnosticsFor the record, I had a closer look at this. It seems to be a bug in my system compiler on Ubuntu 11.04, gcc (Ubuntu/Linaro 4.5.2-8ubuntu4) 4.5.2. In this code from unittests/AST/APValueTest.cpp: template<typename T> std::string operator()(const T& value) { Diag.Report(diag_just_format) << value; return LastDiagnostic.get().str(); } ... the compiler is failing to call the destructor for the temporary DiagnosticBuilder returned from Diag.Report(diag_just_format). If I compile the test with gcc 4.6.1 (which I built from source) the problem goes away. Jay.
Possibly Parallel Threads
- [LLVMdev] [cfe-dev] LLVM 3.0rc3 Testing Beginning
- [LLVMdev] [cfe-dev] LLVM 3.0rc3 Testing Beginning
- [LLVMdev] [cfe-dev] LLVM 3.0rc3 Testing Beginning
- Why SI.isSigned() is not equals to E->getType()->isSignedIntegerOrEnumerationType()?
- [LLVMdev] APFloat::convertToDouble asserts