oce 0.17 tests fail on mac os X #576

samitbasu · 2015-09-09T16:57:55Z

Building OCE v 0.17 on Mac OS X, with clang (claimed version is Apple LLVM Version 6.1.0). The cake and build steps work fine, however "make test" fails the following tests:

Test project /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld
Start 1: BRepAlgoAPITestSuite.testCutBox
1/48 Test #1: BRepAlgoAPITestSuite.testCutBox ................. Passed 0.02 sec
Start 2: BRepAlgoAPITestSuite.testCutCylSphere
2/48 Test #2: BRepAlgoAPITestSuite.testCutCylSphere ........... Passed 0.02 sec
Start 3: BRepExtremaTestSuite.testDistSphereSphere
3/48 Test #3: BRepExtremaTestSuite.testDistSphereSphere ....... Passed 0.01 sec
Start 4: BRepExtremaTestSuite.testDistSphereVertex
4/48 Test #4: BRepExtremaTestSuite.testDistSphereVertex ....... Passed 0.01 sec
Start 5: BRepExtremaTestSuite.testDistVertexEdge
5/48 Test #5: BRepExtremaTestSuite.testDistVertexEdge ......... Passed 0.01 sec
Start 6: BRepExtremaTestSuite.testDistEdgeVertex
6/48 Test #6: BRepExtremaTestSuite.testDistEdgeVertex ......... Passed 0.01 sec
Start 7: BRepExtremaTestSuite.testDistVertexFace
7/48 Test #7: BRepExtremaTestSuite.testDistVertexFace ......... Passed 0.01 sec
Start 8: BRepExtremaTestSuite.testDistFaceVertex
8/48 Test #8: BRepExtremaTestSuite.testDistFaceVertex ......... Passed 0.01 sec
Start 9: BRepGPropTestSuite.testComputeBoxVolume
9/48 Test #9: BRepGPropTestSuite.testComputeBoxVolume ......... Passed 0.01 sec
Start 10: BRepGPropTestSuite.testComputeBoxSurface
10/48 Test #10: BRepGPropTestSuite.testComputeBoxSurface ........ Passed 0.01 sec
Start 11: BRepGPropTestSuite.testComputeSphereVolume
11/48 Test #11: BRepGPropTestSuite.testComputeSphereVolume ...... Passed 0.01 sec
Start 12: BRepImportTestSuite.testImportBRep1
12/48 Test #12: BRepImportTestSuite.testImportBRep1 ............._Exception: SegFault 0.58 sec
Start 13: BRepImportTestSuite.testImportBRep2
13/48 Test #13: BRepImportTestSuite.testImportBRep2 ............._Exception: SegFault 0.28 sec
Start 14: BRepImportTestSuite.testImportBRep3
14/48 Test #14: BRepImportTestSuite.testImportBRep3 ............._Exception: SegFault 0.28 sec
Start 15: BRepImportTestSuite.testImportBRep4
15/48 Test #15: BRepImportTestSuite.testImportBRep4 ............._Exception: SegFault 0.30 sec
Start 16: BRepMeshTestSuite.testMeshBox
16/48 Test #16: BRepMeshTestSuite.testMeshBox ................... Passed 0.01 sec
Start 17: BRepMeshTestSuite.testMeshSphere
17/48 Test #17: BRepMeshTestSuite.testMeshSphere ................ Passed 0.18 sec
Start 18: BRepMeshTestSuite.testMeshTorus
18/48 Test #18: BRepMeshTestSuite.testMeshTorus ................._Exception: SegFault 0.29 sec
Start 19: BRepOffsetAPITestSuite.testEvolvedShape
19/48 Test #19: BRepOffsetAPITestSuite.testEvolvedShape ........._Exception: SegFault 0.22 sec
Start 20: BRepOffsetAPITestSuite.testOffsetBox
20/48 Test #20: BRepOffsetAPITestSuite.testOffsetBox ............_Exception: SegFault 0.24 sec
Start 21: BRepPrimAPITestSuite.testCreateBox
21/48 Test #21: BRepPrimAPITestSuite.testCreateBox .............. Passed 0.01 sec
Start 22: BRepPrimAPITestSuite.testCreateFlatBox
22/48 Test #22: BRepPrimAPITestSuite.testCreateFlatBox .......... Passed 0.01 sec
Start 23: BRepPrimAPITestSuite.testCreateSmallestBox
23/48 Test #23: BRepPrimAPITestSuite.testCreateSmallestBox ...... Passed 0.01 sec
Start 24: gpTestSuite.testgp_PntConstruct
24/48 Test #24: gpTestSuite.testgp_PntConstruct ................. Passed 0.00 sec
Start 25: gpTestSuite.testgp_PntDistance
25/48 Test #25: gpTestSuite.testgp_PntDistance .................. Passed 0.00 sec
Start 26: gpTestSuite.testBuildThousandsPoints
26/48 Test #26: gpTestSuite.testBuildThousandsPoints ............ Passed 0.00 sec
Start 27: gpTestSuite.testGCC42gp_Trsf
27/48 Test #27: gpTestSuite.testGCC42gp_Trsf .................... Passed 0.00 sec
Start 28: IGESImportTestSuite.testImportIGES_1
28/48 Test #28: IGESImportTestSuite.testImportIGES_1 ............_Exception: SegFault 0.30 sec
Start 29: OCAFExportTestSuite.testExportAscii
29/48 Test #29: OCAFExportTestSuite.testExportAscii ............._Exception: SegFault 0.47 sec
Start 30: OCAFExportTestSuite.testExportNonAscii
30/48 Test #30: OCAFExportTestSuite.testExportNonAscii .........._Exception: SegFault 0.37 sec
Start 31: OCAFExportTestSuite.testOverflow
31/48 Test #31: OCAFExportTestSuite.testOverflow ................ Passed 0.02 sec
Start 32: SortToolsTestSuite.testQuickSortOfInteger
32/48 Test #32: SortToolsTestSuite.testQuickSortOfInteger ....... Passed 0.00 sec
Start 33: Standard_StdAllocatorTestSuite.testTraits
33/48 Test #33: Standard_StdAllocatorTestSuite.testTraits ....... Passed 0.01 sec
Start 34: Standard_StdAllocatorTestSuite.testContainers
34/48 Test #34: Standard_StdAllocatorTestSuite.testContainers ... Passed 0.01 sec
Start 35: STEPImportTestSuite.testImportAP203_1
35/48 Test #35: STEPImportTestSuite.testImportAP203_1 ..........._Exception: SegFault 0.30 sec
Start 36: STEPImportTestSuite.testImportAP203_2
36/48 Test #36: STEPImportTestSuite.testImportAP203_2 ..........._Exception: SegFault 0.29 sec
Start 37: STEPImportTestSuite.testImportAP214_1
37/48 Test #37: STEPImportTestSuite.testImportAP214_1 ..........._Exception: SegFault 0.31 sec
Start 38: STEPImportTestSuite.testImportAP214_2
38/48 Test #38: STEPImportTestSuite.testImportAP214_2 ..........._Exception: SegFault 0.29 sec
Start 39: STEPImportTestSuite.testImportAP214_3
39/48 Test #39: STEPImportTestSuite.testImportAP214_3 ...........***Exception: SegFault 0.29 sec
Start 40: TCollectionTestSuite.testSequenceOfPnt
40/48 Test #40: TCollectionTestSuite.testSequenceOfPnt .......... Passed 0.09 sec
Start 41: TDataXtdTestSuite.testPattern
41/48 Test #41: TDataXtdTestSuite.testPattern ................... Passed 0.02 sec
Start 42: UnitsAPITestSuite.testCurrentUnits
42/48 Test #42: UnitsAPITestSuite.testCurrentUnits .............. Passed 0.01 sec
Start 43: UnitsAPITestSuite.testMDTVCurrentUnits
43/48 Test #43: UnitsAPITestSuite.testMDTVCurrentUnits .......... Passed 0.02 sec
Start 44: TestSuite.testNullPointer
44/48 Test #44: TestSuite.testNullPointer ....................... Passed 0.00 sec
Start 45: TestSuite.testFloatEq
45/48 Test #45: TestSuite.testFloatEq ........................... Passed 0.00 sec
Start 46: TestSuite.testFloatNeq
46/48 Test #46: TestSuite.testFloatNeq .......................... Passed 0.00 sec
Start 47: TestSuite.testBoolean
47/48 Test #47: TestSuite.testBoolean ........................... Passed 0.00 sec
Start 48: TestSuite.testIntegerLighter
48/48 Test #48: TestSuite.testIntegerLighter .................... Passed 0.00 sec

69% tests passed, 15 tests failed out of 48

Total Test time (real) = 5.36 sec

The following tests FAILED:
12 - BRepImportTestSuite.testImportBRep1 (SEGFAULT)
13 - BRepImportTestSuite.testImportBRep2 (SEGFAULT)
14 - BRepImportTestSuite.testImportBRep3 (SEGFAULT)
15 - BRepImportTestSuite.testImportBRep4 (SEGFAULT)
18 - BRepMeshTestSuite.testMeshTorus (SEGFAULT)
19 - BRepOffsetAPITestSuite.testEvolvedShape (SEGFAULT)
20 - BRepOffsetAPITestSuite.testOffsetBox (SEGFAULT)
28 - IGESImportTestSuite.testImportIGES_1 (SEGFAULT)
29 - OCAFExportTestSuite.testExportAscii (SEGFAULT)
30 - OCAFExportTestSuite.testExportNonAscii (SEGFAULT)
35 - STEPImportTestSuite.testImportAP203_1 (SEGFAULT)
36 - STEPImportTestSuite.testImportAP203_2 (SEGFAULT)
37 - STEPImportTestSuite.testImportAP214_1 (SEGFAULT)
38 - STEPImportTestSuite.testImportAP214_2 (SEGFAULT)
39 - STEPImportTestSuite.testImportAP214_3 (SEGFAULT)

Furthermore, running STEPImport_test through valgrind shows the following:
==49147== Invalid read of size 4
==49147== at 0x101BEA3CE: BSplCLib::LocateParameter(int, TColStd_Array1OfReal const&, TColStd_Array1OfInteger const&, double, bool, int&, double&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKMath.10.0.0.dylib)
==49147== by 0x101A445DC: Geom_BSplineCurve::ValidateCache(double) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKG3d.10.0.0.dylib)
==49147== by 0x101A44AD2: Geom_BSplineCurve::D0(double, gp_Pnt&) const (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKG3d.10.0.0.dylib)
==49147== by 0x101A591CE: Geom_Curve::Value(double) const (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKG3d.10.0.0.dylib)
==49147== by 0x100A96AB6: ShapeAnalysis_Curve::Project(Handle_Geom_Curve const&, gp_Pnt const&, double, gp_Pnt&, double&, double, double, bool) const (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKShHealing.10.0.0.dylib)
==49147== by 0x10012E730: StepToTopoDS_TranslateEdge::MakeFromCurve3D(Handle_StepGeom_Curve const&, Handle_StepShape_EdgeCurve const&, Handle_StepShape_Vertex const&, double, TopoDS_Edge&, TopoDS_Vertex&, TopoDS_Vertex&, StepToTopoDS_Tool&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== by 0x10012DD02: StepToTopoDS_TranslateEdge::Init(Handle_StepShape_Edge const&, StepToTopoDS_Tool&, StepToTopoDS_NMTool&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== by 0x100130A5F: StepToTopoDS_TranslateEdgeLoop::Init(Handle_StepShape_FaceBound const&, TopoDS_Face const&, Handle_Geom_Surface const&, Handle_StepGeom_Surface const&, bool, StepToTopoDS_Tool&, StepToTopoDS_NMTool&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== by 0x100133D2E: StepToTopoDS_TranslateFace::Init(Handle_StepShape_FaceSurface const&, StepToTopoDS_Tool&, StepToTopoDS_NMTool&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== by 0x10013642B: StepToTopoDS_TranslateShell::Init(Handle_StepShape_ConnectedFaceSet const&, StepToTopoDS_Tool&, StepToTopoDS_NMTool&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== by 0x100120BF6: StepToTopoDS_Builder::Init(Handle_StepShape_ManifoldSolidBrep const&, Handle_Transfer_TransientProcess const&) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== by 0x100156176: STEPControl_ActorRead::TransferEntity(Handle_StepGeom_GeometricRepresentationItem const&, Handle_Transfer_TransientProcess const&, bool) (in /Users/sbasu/Devel/trimesh/oce-OCE-0.17/bld/Darwin/x86_64-Release-64/libTKSTEP.10.0.0.dylib)
==49147== Address 0x0 is not stack'd, malloc'd or (recently) free'd

Is this a known issue? I haven't been able to find anyone else reporting this issue, but I may have missed it.

chrisidefix · 2015-09-15T08:43:16Z

I am having similarly severe issues under Mac OS X. Which OS version are you running on?
Even running some basic example code SEGFAULT (I listed some stack traces here: https://gist.github.com/chrisidefix/f35c3fb0572e6bfc9ba1).

These problems all appear to be related to code execute inside libTKMath.10.0.0.dylib, especially in BSplCLib::LocateParameter BSplCLib::BuildKnots BSplCLib::IncreaseDegree (calling PLib::SetPoles). Have you tried to debug the issue?

chrisidefix · 2015-09-15T08:49:35Z

side note: @tpaviot appears to use OS X himself (if you look through other issues) I can't imagine such basic problems would have not been discovered. Since I am running on OS X 10.11 (GM release), I am starting to think it might be related to some recent changes in OS X.

chrisidefix · 2015-09-15T14:31:59Z

I tested everything under Linux Mint (17.2) first and didn't see any problems occurring with the same basic example / test code.

I then also compiled everything in DEBUG mode to see what is really going on. Of course now all issues are gone. I suspect this might be some race-condition in the code or the compiler optimising something that causes issues, but so far I couldn't re-create the issues in DEBUG mode. Obviously this isn't very satisfactory, but at least there is a work-around for now.

To compile your code in debug mode first add the following flag to your cmake setup:
CMAKE_BUILD_TYPE:STRING=Debug

I also added the -ggdb flag to the compiler flags: CMAKE_CXX_FLAGS_DEBUG and CMAKE_C_FLAGS_DEBUG

Apparently there used to be a variable OCE_BUILD_TYPE:STRING=Debug that had to be set for debug compilation in the past, but from what I can see it has been removed.

samitbasu · 2015-09-20T19:00:54Z

Sorry - was off the grid for a while. I will try and debug the issue. I was using OS X 10.10.4 - now on 10.10.5 and Xcode 7. I also get the same issues when building open cascade 6.9 direct from sources.

samitbasu · 2015-09-20T21:42:10Z

Update: building master in debug mode (i.e., manually setting the CMAKE_BUILD_TYPE to Debug) on OS X 10.10.5 w/ Xcode 7, almost all tests pass:

Running tests...
Test project /Users/sbasu/Devel/trimesh/oce/bld/test
Start 1: BRepAlgoAPITestSuite.testCutBox
1/48 Test #1: BRepAlgoAPITestSuite.testCutBox ................. Passed 0.06 sec
Start 2: BRepAlgoAPITestSuite.testCutCylSphere
2/48 Test #2: BRepAlgoAPITestSuite.testCutCylSphere ........... Passed 0.05 sec
Start 3: BRepExtremaTestSuite.testDistSphereSphere
3/48 Test #3: BRepExtremaTestSuite.testDistSphereSphere ....... Passed 0.02 sec
Start 4: BRepExtremaTestSuite.testDistSphereVertex
4/48 Test #4: BRepExtremaTestSuite.testDistSphereVertex ....... Passed 0.01 sec
Start 5: BRepExtremaTestSuite.testDistVertexEdge
5/48 Test #5: BRepExtremaTestSuite.testDistVertexEdge ......... Passed 0.01 sec
Start 6: BRepExtremaTestSuite.testDistEdgeVertex
6/48 Test #6: BRepExtremaTestSuite.testDistEdgeVertex ......... Passed 0.01 sec
Start 7: BRepExtremaTestSuite.testDistVertexFace
7/48 Test #7: BRepExtremaTestSuite.testDistVertexFace ......... Passed 0.01 sec
Start 8: BRepExtremaTestSuite.testDistFaceVertex
8/48 Test #8: BRepExtremaTestSuite.testDistFaceVertex ......... Passed 0.01 sec
Start 9: BRepGPropTestSuite.testComputeBoxVolume
9/48 Test #9: BRepGPropTestSuite.testComputeBoxVolume ......... Passed 0.02 sec
Start 10: BRepGPropTestSuite.testComputeBoxSurface
10/48 Test #10: BRepGPropTestSuite.testComputeBoxSurface ........ Passed 0.01 sec
Start 11: BRepGPropTestSuite.testComputeSphereVolume
11/48 Test #11: BRepGPropTestSuite.testComputeSphereVolume ...... Passed 0.01 sec
Start 12: BRepImportTestSuite.testImportBRep1
12/48 Test #12: BRepImportTestSuite.testImportBRep1 ............. Passed 0.56 sec
Start 13: BRepImportTestSuite.testImportBRep2
13/48 Test #13: BRepImportTestSuite.testImportBRep2 ............. Passed 0.04 sec
Start 14: BRepImportTestSuite.testImportBRep3
14/48 Test #14: BRepImportTestSuite.testImportBRep3 ............. Passed 0.03 sec
Start 15: BRepImportTestSuite.testImportBRep4
15/48 Test #15: BRepImportTestSuite.testImportBRep4 ............. Passed 0.07 sec
Start 16: BRepMeshTestSuite.testMeshBox
16/48 Test #16: BRepMeshTestSuite.testMeshBox ................... Passed 0.02 sec
Start 17: BRepMeshTestSuite.testMeshSphere
17/48 Test #17: BRepMeshTestSuite.testMeshSphere ................ Passed 0.55 sec
Start 18: BRepMeshTestSuite.testMeshTorus
18/48 Test #18: BRepMeshTestSuite.testMeshTorus ................. Passed 9.17 sec
Start 19: BRepOffsetAPITestSuite.testEvolvedShape
19/48 Test #19: BRepOffsetAPITestSuite.testEvolvedShape ......... Passed 0.05 sec
Start 20: BRepOffsetAPITestSuite.testOffsetBox
20/48 Test #20: BRepOffsetAPITestSuite.testOffsetBox ............***Exception: Other 3.84 sec
Start 21: BRepPrimAPITestSuite.testCreateBox
21/48 Test #21: BRepPrimAPITestSuite.testCreateBox .............. Passed 0.02 sec
Start 22: BRepPrimAPITestSuite.testCreateFlatBox
22/48 Test #22: BRepPrimAPITestSuite.testCreateFlatBox .......... Passed 0.01 sec
Start 23: BRepPrimAPITestSuite.testCreateSmallestBox
23/48 Test #23: BRepPrimAPITestSuite.testCreateSmallestBox ...... Passed 0.01 sec
Start 24: gpTestSuite.testgp_PntConstruct
24/48 Test #24: gpTestSuite.testgp_PntConstruct ................. Passed 0.01 sec
Start 25: gpTestSuite.testgp_PntDistance
25/48 Test #25: gpTestSuite.testgp_PntDistance .................. Passed 0.00 sec
Start 26: gpTestSuite.testBuildThousandsPoints
26/48 Test #26: gpTestSuite.testBuildThousandsPoints ............ Passed 0.00 sec
Start 27: gpTestSuite.testGCC42gp_Trsf
27/48 Test #27: gpTestSuite.testGCC42gp_Trsf .................... Passed 0.01 sec
Start 28: IGESImportTestSuite.testImportIGES_1
28/48 Test #28: IGESImportTestSuite.testImportIGES_1 ............ Passed 0.37 sec
Start 29: OCAFExportTestSuite.testExportAscii
29/48 Test #29: OCAFExportTestSuite.testExportAscii ............. Passed 0.04 sec
Start 30: OCAFExportTestSuite.testExportNonAscii
30/48 Test #30: OCAFExportTestSuite.testExportNonAscii .......... Passed 0.03 sec
Start 31: OCAFExportTestSuite.testOverflow
31/48 Test #31: OCAFExportTestSuite.testOverflow ................ Passed 0.02 sec
Start 32: SortToolsTestSuite.testQuickSortOfInteger
32/48 Test #32: SortToolsTestSuite.testQuickSortOfInteger ....... Passed 0.01 sec
Start 33: Standard_StdAllocatorTestSuite.testTraits
33/48 Test #33: Standard_StdAllocatorTestSuite.testTraits ....... Passed 0.02 sec
Start 34: Standard_StdAllocatorTestSuite.testContainers
34/48 Test #34: Standard_StdAllocatorTestSuite.testContainers ... Passed 0.01 sec
Start 35: STEPImportTestSuite.testImportAP203_1
35/48 Test #35: STEPImportTestSuite.testImportAP203_1 ........... Passed 0.38 sec
Start 36: STEPImportTestSuite.testImportAP203_2
36/48 Test #36: STEPImportTestSuite.testImportAP203_2 ........... Passed 0.33 sec
Start 37: STEPImportTestSuite.testImportAP214_1
37/48 Test #37: STEPImportTestSuite.testImportAP214_1 ........... Passed 0.50 sec
Start 38: STEPImportTestSuite.testImportAP214_2
38/48 Test #38: STEPImportTestSuite.testImportAP214_2 ........... Passed 0.21 sec
Start 39: STEPImportTestSuite.testImportAP214_3
39/48 Test #39: STEPImportTestSuite.testImportAP214_3 ........... Passed 0.12 sec
Start 40: TCollectionTestSuite.testSequenceOfPnt
40/48 Test #40: TCollectionTestSuite.testSequenceOfPnt .......... Passed 0.17 sec
Start 41: TDataXtdTestSuite.testPattern
41/48 Test #41: TDataXtdTestSuite.testPattern ................... Passed 0.03 sec
Start 42: UnitsAPITestSuite.testCurrentUnits
42/48 Test #42: UnitsAPITestSuite.testCurrentUnits .............. Passed 0.01 sec
Start 43: UnitsAPITestSuite.testMDTVCurrentUnits
43/48 Test #43: UnitsAPITestSuite.testMDTVCurrentUnits .......... Passed 0.04 sec
Start 44: TestSuite.testNullPointer
44/48 Test #44: TestSuite.testNullPointer ....................... Passed 0.00 sec
Start 45: TestSuite.testFloatEq
45/48 Test #45: TestSuite.testFloatEq ........................... Passed 0.00 sec
Start 46: TestSuite.testFloatNeq
46/48 Test #46: TestSuite.testFloatNeq .......................... Passed 0.00 sec
Start 47: TestSuite.testBoolean
47/48 Test #47: TestSuite.testBoolean ........................... Passed 0.00 sec
Start 48: TestSuite.testIntegerLighter
48/48 Test #48: TestSuite.testIntegerLighter .................... Passed 0.00 sec

98% tests passed, 1 tests failed out of 48

Total Test time (real) = 16.93 sec

The following tests FAILED:
20 - BRepOffsetAPITestSuite.testOffsetBox (OTHER_FAULT)
Errors while running CTest

Will try rebuilding in release mode to see if the problem reappears.

Samit

samitbasu · 2015-09-20T22:09:45Z

Building in release mode causes the seg faults again:

Test project /Users/sbasu/Devel/trimesh/oce/bld/test
Start 1: BRepAlgoAPITestSuite.testCutBox
1/48 Test #1: BRepAlgoAPITestSuite.testCutBox ................. Passed 0.03 sec
Start 2: BRepAlgoAPITestSuite.testCutCylSphere
2/48 Test #2: BRepAlgoAPITestSuite.testCutCylSphere ........... Passed 0.02 sec
Start 3: BRepExtremaTestSuite.testDistSphereSphere
3/48 Test #3: BRepExtremaTestSuite.testDistSphereSphere ....... Passed 0.02 sec
Start 4: BRepExtremaTestSuite.testDistSphereVertex
4/48 Test #4: BRepExtremaTestSuite.testDistSphereVertex ....... Passed 0.01 sec
Start 5: BRepExtremaTestSuite.testDistVertexEdge
5/48 Test #5: BRepExtremaTestSuite.testDistVertexEdge ......... Passed 0.01 sec
Start 6: BRepExtremaTestSuite.testDistEdgeVertex
6/48 Test #6: BRepExtremaTestSuite.testDistEdgeVertex ......... Passed 0.01 sec
Start 7: BRepExtremaTestSuite.testDistVertexFace
7/48 Test #7: BRepExtremaTestSuite.testDistVertexFace ......... Passed 0.01 sec
Start 8: BRepExtremaTestSuite.testDistFaceVertex
8/48 Test #8: BRepExtremaTestSuite.testDistFaceVertex ......... Passed 0.01 sec
Start 9: BRepGPropTestSuite.testComputeBoxVolume
9/48 Test #9: BRepGPropTestSuite.testComputeBoxVolume ......... Passed 0.01 sec
Start 10: BRepGPropTestSuite.testComputeBoxSurface
10/48 Test #10: BRepGPropTestSuite.testComputeBoxSurface ........ Passed 0.01 sec
Start 11: BRepGPropTestSuite.testComputeSphereVolume
11/48 Test #11: BRepGPropTestSuite.testComputeSphereVolume ...... Passed 0.01 sec
Start 12: BRepImportTestSuite.testImportBRep1
12/48 Test #12: BRepImportTestSuite.testImportBRep1 ............._Exception: SegFault 0.56 sec
Start 13: BRepImportTestSuite.testImportBRep2
13/48 Test #13: BRepImportTestSuite.testImportBRep2 ............._Exception: SegFault 0.29 sec
Start 14: BRepImportTestSuite.testImportBRep3
14/48 Test #14: BRepImportTestSuite.testImportBRep3 ............._Exception: SegFault 0.29 sec
Start 15: BRepImportTestSuite.testImportBRep4
15/48 Test #15: BRepImportTestSuite.testImportBRep4 ............._Exception: SegFault 0.31 sec
Start 16: BRepMeshTestSuite.testMeshBox
16/48 Test #16: BRepMeshTestSuite.testMeshBox ................... Passed 0.02 sec
Start 17: BRepMeshTestSuite.testMeshSphere
17/48 Test #17: BRepMeshTestSuite.testMeshSphere ................ Passed 0.18 sec
Start 18: BRepMeshTestSuite.testMeshTorus
18/48 Test #18: BRepMeshTestSuite.testMeshTorus ................._Exception: SegFault 0.31 sec
Start 19: BRepOffsetAPITestSuite.testEvolvedShape
19/48 Test #19: BRepOffsetAPITestSuite.testEvolvedShape ........._Exception: SegFault 0.25 sec
Start 20: BRepOffsetAPITestSuite.testOffsetBox
20/48 Test #20: BRepOffsetAPITestSuite.testOffsetBox ............_Exception: SegFault 0.24 sec
Start 21: BRepPrimAPITestSuite.testCreateBox
21/48 Test #21: BRepPrimAPITestSuite.testCreateBox .............. Passed 0.01 sec
Start 22: BRepPrimAPITestSuite.testCreateFlatBox
22/48 Test #22: BRepPrimAPITestSuite.testCreateFlatBox .......... Passed 0.01 sec
Start 23: BRepPrimAPITestSuite.testCreateSmallestBox
23/48 Test #23: BRepPrimAPITestSuite.testCreateSmallestBox ...... Passed 0.01 sec
Start 24: gpTestSuite.testgp_PntConstruct
24/48 Test #24: gpTestSuite.testgp_PntConstruct ................. Passed 0.01 sec
Start 25: gpTestSuite.testgp_PntDistance
25/48 Test #25: gpTestSuite.testgp_PntDistance .................. Passed 0.00 sec
Start 26: gpTestSuite.testBuildThousandsPoints
26/48 Test #26: gpTestSuite.testBuildThousandsPoints ............ Passed 0.00 sec
Start 27: gpTestSuite.testGCC42gp_Trsf
27/48 Test #27: gpTestSuite.testGCC42gp_Trsf .................... Passed 0.00 sec
Start 28: IGESImportTestSuite.testImportIGES_1
28/48 Test #28: IGESImportTestSuite.testImportIGES_1 ............_Exception: SegFault 0.30 sec
Start 29: OCAFExportTestSuite.testExportAscii
29/48 Test #29: OCAFExportTestSuite.testExportAscii ............. Passed 0.04 sec
Start 30: OCAFExportTestSuite.testExportNonAscii
30/48 Test #30: OCAFExportTestSuite.testExportNonAscii .......... Passed 0.02 sec
Start 31: OCAFExportTestSuite.testOverflow
31/48 Test #31: OCAFExportTestSuite.testOverflow ................ Passed 0.02 sec
Start 32: SortToolsTestSuite.testQuickSortOfInteger
32/48 Test #32: SortToolsTestSuite.testQuickSortOfInteger ....... Passed 0.01 sec
Start 33: Standard_StdAllocatorTestSuite.testTraits
33/48 Test #33: Standard_StdAllocatorTestSuite.testTraits ....... Passed 0.01 sec
Start 34: Standard_StdAllocatorTestSuite.testContainers
34/48 Test #34: Standard_StdAllocatorTestSuite.testContainers ... Passed 0.01 sec
Start 35: STEPImportTestSuite.testImportAP203_1
35/48 Test #35: STEPImportTestSuite.testImportAP203_1 ..........._Exception: SegFault 0.31 sec
Start 36: STEPImportTestSuite.testImportAP203_2
36/48 Test #36: STEPImportTestSuite.testImportAP203_2 ..........._Exception: SegFault 0.30 sec
Start 37: STEPImportTestSuite.testImportAP214_1
37/48 Test #37: STEPImportTestSuite.testImportAP214_1 ..........._Exception: SegFault 0.33 sec
Start 38: STEPImportTestSuite.testImportAP214_2
38/48 Test #38: STEPImportTestSuite.testImportAP214_2 ..........._Exception: SegFault 0.29 sec
Start 39: STEPImportTestSuite.testImportAP214_3
39/48 Test #39: STEPImportTestSuite.testImportAP214_3 ...........***Exception: SegFault 0.29 sec
Start 40: TCollectionTestSuite.testSequenceOfPnt
40/48 Test #40: TCollectionTestSuite.testSequenceOfPnt .......... Passed 0.09 sec
Start 41: TDataXtdTestSuite.testPattern
41/48 Test #41: TDataXtdTestSuite.testPattern ................... Passed 0.03 sec
Start 42: UnitsAPITestSuite.testCurrentUnits
42/48 Test #42: UnitsAPITestSuite.testCurrentUnits .............. Passed 0.01 sec
Start 43: UnitsAPITestSuite.testMDTVCurrentUnits
43/48 Test #43: UnitsAPITestSuite.testMDTVCurrentUnits .......... Passed 0.03 sec
Start 44: TestSuite.testNullPointer
44/48 Test #44: TestSuite.testNullPointer ....................... Passed 0.01 sec
Start 45: TestSuite.testFloatEq
45/48 Test #45: TestSuite.testFloatEq ........................... Passed 0.00 sec
Start 46: TestSuite.testFloatNeq
46/48 Test #46: TestSuite.testFloatNeq .......................... Passed 0.00 sec
Start 47: TestSuite.testBoolean
47/48 Test #47: TestSuite.testBoolean ........................... Passed 0.00 sec
Start 48: TestSuite.testIntegerLighter
48/48 Test #48: TestSuite.testIntegerLighter .................... Passed 0.00 sec

73% tests passed, 13 tests failed out of 48

Total Test time (real) = 4.74 sec

The following tests FAILED:
12 - BRepImportTestSuite.testImportBRep1 (SEGFAULT)
13 - BRepImportTestSuite.testImportBRep2 (SEGFAULT)
14 - BRepImportTestSuite.testImportBRep3 (SEGFAULT)
15 - BRepImportTestSuite.testImportBRep4 (SEGFAULT)
18 - BRepMeshTestSuite.testMeshTorus (SEGFAULT)
19 - BRepOffsetAPITestSuite.testEvolvedShape (SEGFAULT)
20 - BRepOffsetAPITestSuite.testOffsetBox (SEGFAULT)
28 - IGESImportTestSuite.testImportIGES_1 (SEGFAULT)
35 - STEPImportTestSuite.testImportAP203_1 (SEGFAULT)
36 - STEPImportTestSuite.testImportAP203_2 (SEGFAULT)
37 - STEPImportTestSuite.testImportAP214_1 (SEGFAULT)
38 - STEPImportTestSuite.testImportAP214_2 (SEGFAULT)
39 - STEPImportTestSuite.testImportAP214_3 (SEGFAULT)
Errors while running CTest

That will make debugging more difficult. Will post anything I can find.

Samit

chrisidefix · 2015-09-21T11:27:55Z

Yes, thanks for also checking in debug mode - I think it would be worth extending the travis testing options to include a build under OS X (apparently they can generate VMs for that now) to make issues like this one surface more prominently. I am starting to test this here (https://github.com/heartvalve/oce/blob/osx-test/.travis.yml).

samitbasu · 2015-09-21T15:45:25Z

I've narrowed it down (I believe) to the following idiom. See in BSplCLib.cxx, on line 301:

 const TColStd_Array1OfInteger& Mults,
 const Standard_Real            U,
 const Standard_Boolean         Periodic,
 Standard_Integer&              KnotIndex,
 Standard_Real&                 NewU) 
{
  Standard_Integer first,last;
  if (&Mults) { // <-- This test is always true, even when Mults is null.
    if (Periodic) {
      first = Knots.Lower();

During compilation, clang complains that because Mults is a reference, taking it's address must yield something that is not-NULL. Here is the warning during build:

[ 5%] Building CXX object >adm/cmake/TKMath/CMakeFiles/TKMath.dir///__/src/BSplCLib/BSplCLib.cxx.o
/Users/sbasu/Devel/trimesh/oce/src/BSplCLib/BSplCLib.cxx:301:8: warning: reference cannot be bound >to dereferenced null
pointer in well-defined C++ code; pointer may be assumed to always convert to true [-Wundefined->bool-conversion]
if (&Mults) {
~~ ^~~~~
/Users/sbasu/Devel/trimesh/oce/src/BSplCLib/BSplCLib.cxx:1437:8: warning: reference cannot be bound to dereferenced null
pointer in well-defined C++ code; comparison may be assumed to always evaluate to false
[-Wtautological-undefined-compare]
if (&Mults == NULL) {
^~~~~ ~~~~

Furthermore, the calling code is a bit odd to me -

-> 1223 BSplCLib::LocateParameter(deg,
1224 (flatknots->Array1()),
1225 (BSplCLib::NoMults()),
1226 Parameter,

From the backtrace in LLDB, note the call to BSplCLib::NoMults(), which I gather is supposed to be the knot multiplicity vector when the spline has no define multiplicities (equivalent to all unit-multiplicities, I guess). This is an interesting function. It is defined in BSplCLib.hxx (around line 380)

  //! Used as argument for a non rational curve.
    static   TColStd_Array1OfReal& NoWeights() ;

  //! Used as argument for a flatknots evaluation.
    static   TColStd_Array1OfInteger& NoMults() ;

But the function itself is actually undefined - or at least I could not find the definition anywhere (including searching through the symbols in the libraries). As a result, at run time, you get a unique phenomenon - a const reference bound to a null value. I don't know if that's the intent, but Clang doesn't like it (and it seems to violate the requirements on references never pointing to NULL addresses). I suspect that GCC accepts the condition (for now). But I would be surprised if this doesn't crop up in a future version of GCC as well.

Anyone have a suggestion for how this should be fixed? This particular idiom (of checking a const-ref for NULL-ness) happens in a bunch of places. I checked the code for OpenCascade 6.9.0, and the issue appears there as well.

a-betenev · 2015-09-21T15:51:14Z

Hello,

This is a known issue, see
http://tracker.dev.opencascade.org/view.php?id=26042
There is a work-around there; the correct fix naturally is to avoid this
kind of code; to be done in OCCT 7.0.
Hope this helps.

Andrey

On 21.09.2015 18:45, samitbasu wrote:

I've narrowed it down (I believe) to the following idiom. See in
BSplCLib.cxx, on line 301:

| const TColStd_Array1OfInteger& Mults,
const Standard_Real U,
const Standard_Boolean Periodic,
Standard_Integer& KnotIndex,
Standard_Real& NewU)
{
Standard_Integer first,last;
if (&Mults) { // <-- This test is always true, even when Mults is null.
if (Periodic) {
first = Knots.Lower();
|

During compilation, clang complains that because |Mults| is a
reference, taking it's address must yield something that is not-NULL.
Here is the warning during build:
[ 5%] Building CXX object
>adm/cmake/TKMath/CMakeFiles/TKMath.dir/*/*/__/src/BSplCLib/BSplCLib.cxx.o
/Users/sbasu/Devel/trimesh/oce/src/BSplCLib/BSplCLib.cxx:301:8:
warning: reference cannot be bound >to dereferenced null
pointer in well-defined C++ code; pointer may be assumed to always
convert to true [-Wundefined->bool-conversion]
if (&Mults) {
~~ ^~~~~
/Users/sbasu/Devel/trimesh/oce/src/BSplCLib/BSplCLib.cxx:1437:8:
warning: reference cannot be bound to dereferenced null
pointer in well-defined C++ code; comparison may be assumed to
always evaluate to false
[-Wtautological-undefined-compare]
if (&Mults == NULL) {
^~~~~ ~~~~
Furthermore, the calling code is a bit odd to me -
-> 1223 BSplCLib::LocateParameter(deg,
1224 (flatknots->Array1()),
1225 (BSplCLib::NoMults()),
1226 Parameter,
From the backtrace in LLDB, note the call to |BSplCLib::NoMults()|,
which I gather is supposed to be the knot multiplicity vector when the
spline has no define multiplicities (equivalent to all
unit-multiplicities, I guess). This is an interesting function. It is
defined in BSplCLib.hxx (around line 380)

| //! Used as argument for a non rational curve.
static TColStd_Array1OfReal& NoWeights() ;

//! Used as argument for a flatknots evaluation.
static TColStd_Array1OfInteger& NoMults() ;
|

But the function itself is actually undefined - or at least I could
not find the definition anywhere (including searching through the
symbols in the libraries). As a result, at run time, you get a unique
phenomenon - a const reference bound to a null value. I don't know if
that's the intent, but Clang doesn't like it (and it seems to violate
the requirements on references never pointing to NULL addresses). I
suspect that GCC accepts the condition (for now). But I would be
surprised if this doesn't crop up in a future version of GCC as well.

Anyone have a suggestion for how this /should/ be fixed? This
particular idiom (of checking a const-ref for NULL-ness) happens in a
bunch of places. I checked the code for OpenCascade 6.9.0, and the
issue appears there as well.

—
Reply to this email directly or view it on GitHub
#576 (comment).

samitbasu · 2015-09-21T15:57:44Z

Yes - thanks for pointing out the issue, Andrey. Would it make sense to patch the CMake build system to disable optimized builds for Clang until the issue is fixed?

a-betenev · 2015-09-22T07:49:20Z

On 21.09.2015 18:57, samitbasu wrote:

Yes - thanks for pointing out the issue, Andrey. Would it make sense
to patch the CMake build system to disable optimized builds for Clang
until the issue is fixed?

I believe that all fixes (including local disabling of optimization)
should already be present in relevant commits in OCCT repo.

ghost · 2015-11-19T23:19:57Z

http://tracker.dev.opencascade.org/view.php?id=26042

rainman110 · 2016-01-20T14:39:29Z

I don't think it is an option to build only in debug mode due to performance reasons.

I backported the OCCT 7.0 patch for this fix. This patch is currently OCE-0.16.1 only but can be easily adapted to newer oce releases. The patch can be downloaded here:
https://github.com/DLR-SC/tigl-conda/blob/master/oce/oce-nullrefs-fix.patch

jf--- · 2016-01-25T09:41:17Z

@rainman110 , wow, that backport of the Clang compiler fix from upstream OCC is an impressive effort!
That's an important contribution 👍 , I think this also affects builds on linux with Clang 3.4+ .

Finally, are your builds on Conda already built with the Clang patch?

rainman110 · 2016-01-25T12:49:14Z

Hi @jf---

I only build it for OSX for now with the patch. The linux builds on conda are not affected, since I used an old gcc which does not do these compiler optimizations.

So far, all the conda builds on dlr-sc should not include these problems anymore.

jf--- · 2016-01-25T15:00:24Z

OK, thanks for the heads up, that's good to know...
Regarding linux, building with a recent Clang compiler leads to similar issues, right?

rainman110 · 2016-01-25T16:42:04Z

@jf--- I didn't try to compile it with clang on linux, but the problems should be identical.

It would make sense to integrate this patch into the next oce release.

jf--- · 2016-01-25T17:48:23Z

Very much in favor of that, its critical fix actually, so many thanks for your efforts backporting this!

jf--- · 2016-01-27T06:49:10Z

@rainman110 , regarding the Patch, is it OK to ask you for a PR at the OCE repo? Its such an important fix 🏆 🍹

Would you mind how you've gone about creating this patch?
Was there a particular OCC branch you could derive it from?

rainman110 · 2016-01-27T09:07:55Z

@jf--- I just followed the mantis entry by @a-betenev / #576 (comment).

First I backported the proper fix from http://git.dev.opencascade.org/gitweb/?p=occt.git;a=commitdiff;h=0e14656b3052946192bbb08a998e9c23a3be8752. This however changed the public API of some classes and therefore @tpaviot didn't really like it.

Then, I just implemented a work-around inspired by the workaround attached to this mantis issue. I just looked at the clang compiler warnings and figured out, were this patch is required.

This workaround won't be included in OCCT 7.0. Instead they will provide a proper fix including a changed API.

I can do a PR. How do you actually integrate changes to OCE, when a new OCCT release is targeted?

OCCT/OCE includes some evil code that uses NULL references, which are normally not possible. Clang removes code in branches like if(&myNullRef==NULL) as it assumes this can never be true. This fix was inspired from the mantis issue http://tracker.dev.opencascade.org/view.php?id=26042. This code will be fixed in OCCT 7, but we might require the fix for earlier releases as well. Fixes issue tpaviot#576

OCCT/OCE includes some evil code that uses NULL references, which are normally not possible. Clang removes code in branches like if(&myNullRef==NULL) as it assumes this can never be true. This fix was inspired from the mantis issue http://tracker.dev.opencascade.org/view.php?id=26042. This code will be fixed in OCCT 7, but we might require the fix for earlier releases as well. Fixes issue tpaviot#576 (cherry picked from commit 61cb965)

rainman110 · 2016-05-24T13:17:52Z

@tpaviot I think this issue can be closed.

jf--- mentioned this issue Oct 12, 2015

Upgrade to OCCT 6.9 #561

Closed

ghost added the OSX label Nov 19, 2015

ghost added the Reported to OCCT label Dec 15, 2015

ghost self-assigned this Dec 19, 2015

rainman110 mentioned this issue Jan 19, 2016

Failing bezier surface test on OS X tpaviot/pythonocc-core#202

Closed

rainman110 mentioned this issue Jan 21, 2016

Crash with the module libTKMath.9.dylib on examples/core_load_step_ap203.py tpaviot/pythonocc-core#147

Closed

rainman110 mentioned this issue Jan 28, 2016

Workaround clang optimizations for null references #604

Merged

tpaviot mentioned this issue Feb 17, 2016

OCE 0.16.1 Tests segfaulting with GCC 6.0.0 #605

Closed

tpaviot closed this as completed May 24, 2016

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

oce 0.17 tests fail on mac os X #576

oce 0.17 tests fail on mac os X #576

samitbasu commented Sep 9, 2015

chrisidefix commented Sep 15, 2015

chrisidefix commented Sep 15, 2015

chrisidefix commented Sep 15, 2015

samitbasu commented Sep 20, 2015

samitbasu commented Sep 20, 2015

samitbasu commented Sep 20, 2015

chrisidefix commented Sep 21, 2015

samitbasu commented Sep 21, 2015

a-betenev commented Sep 21, 2015

samitbasu commented Sep 21, 2015

a-betenev commented Sep 22, 2015

ghost commented Nov 19, 2015

rainman110 commented Jan 20, 2016

jf--- commented Jan 25, 2016

rainman110 commented Jan 25, 2016

jf--- commented Jan 25, 2016

rainman110 commented Jan 25, 2016

jf--- commented Jan 25, 2016

jf--- commented Jan 27, 2016

rainman110 commented Jan 27, 2016

rainman110 commented May 24, 2016

oce 0.17 tests fail on mac os X #576

oce 0.17 tests fail on mac os X #576

Comments

samitbasu commented Sep 9, 2015

chrisidefix commented Sep 15, 2015

chrisidefix commented Sep 15, 2015

chrisidefix commented Sep 15, 2015

samitbasu commented Sep 20, 2015

samitbasu commented Sep 20, 2015

samitbasu commented Sep 20, 2015

chrisidefix commented Sep 21, 2015

samitbasu commented Sep 21, 2015

a-betenev commented Sep 21, 2015

samitbasu commented Sep 21, 2015

a-betenev commented Sep 22, 2015

ghost commented Nov 19, 2015

rainman110 commented Jan 20, 2016

jf--- commented Jan 25, 2016

rainman110 commented Jan 25, 2016

jf--- commented Jan 25, 2016

rainman110 commented Jan 25, 2016

jf--- commented Jan 25, 2016

jf--- commented Jan 27, 2016

rainman110 commented Jan 27, 2016

rainman110 commented May 24, 2016