fatal exception in Group_WriteGrid

[jbordner]

lcatest runs of the following test problems crash in various ways in Group_WriteGrid (links to output)

• HD-Sedov-Taylor-BlastWave
• MHD2DBlastWave

FIXED IN changeset 1337:

• RampedKelvinHelmholtz2D
• HDMHD2DCheckOddEvenCouplingOfRiemannSolver

[jbordner]

Enabling "new grid io" does not resolve this issue.

New Grid I/O disabled:

[gedeckt:17825] *** Process received signal ***
[gedeckt:17825] Signal: Segmentation fault (11)
[gedeckt:17825] Signal code: Address not mapped (1)
[gedeckt:17825] Failing at address: 0x18
[gedeckt:17825] [ 0] /lib/libpthread.so.0(+0xf8f0) [0x7f1c6e2c28f0]
[gedeckt:17825] [ 1] /usr/lib/libhdf5.so.6(H5Dcreate1+0x1fa) [0x7f1c701c64d8]
[gedeckt:17825] [ 2] enzo(_ZN4grid15Group_WriteGridEP8_IO_FILEPcxi+0x22ca) [0x548cea]
[gedeckt:17825] [ 3] enzo(_Z24Group_WriteDataHierarchyP8_IO_FILER11TopGridDataP14HierarchyEntryPcRxdix+0xd0) [0x5ca844]
[gedeckt:17825] [ 4] enzo(_Z18Group_WriteAllDataPcxP14HierarchyEntryR11TopGridDataP16ExternalBoundarydx+0x1607) [0x5c9bc3]
[gedeckt:17825] [ 5] enzo(_Z14CheckForOutputP14HierarchyEntryR11TopGridDataP16ExternalBoundaryRx+0x37d) [0x44ea71]
[gedeckt:17825] [ 6] enzo(_Z15EvolveHierarchyR14HierarchyEntryR11TopGridDataP16ExternalBoundaryPP19LevelHierarchyEntryd+0x5a6) [0x49e8b6]
[gedeckt:17825] [ 7] enzo(main+0xb42) [0x434ca6]
[gedeckt:17825] [ 8] /lib/libc.so.6(__libc_start_main+0xfd) [0x7f1c6df4fc4d]
[gedeckt:17825] [ 9] enzo() [0x4340a9]
[gedeckt:17825] *** End of error message ***

New Grid I/O enabled:

[gedeckt:17842] *** Process received signal ***
[gedeckt:17842] Signal: Segmentation fault (11)
[gedeckt:17842] Signal code: Address not mapped (1)
[gedeckt:17842] Failing at address: 0x7fffe7d2d000
[gedeckt:17842] [ 0] /lib/libpthread.so.0(+0xf8f0) [0x7ff89f4ee8f0]
[gedeckt:17842] [ 1] enzo(_ZN4grid13write_datasetExPyPciiPvxPd+0x1d4) [0x5490f0]
[gedeckt:17842] [ 2] enzo(_ZN4grid15Group_WriteGridEP8_IO_FILEPcxix+0x112e) [0x5479ba]
[gedeckt:17842] [ 3] enzo(_Z24Group_WriteDataHierarchyP8_IO_FILER11TopGridDataP14HierarchyEntryPcRxdix+0xa5) [0x5c6199]
[gedeckt:17842] [ 4] enzo(_Z18Group_WriteAllDataPcxP14HierarchyEntryR11TopGridDataP16ExternalBoundarydx+0x1607) [0x5c5543]
[gedeckt:17842] [ 5] enzo(_Z14CheckForOutputP14HierarchyEntryR11TopGridDataP16ExternalBoundaryRx+0x37d) [0x44ea81]
[gedeckt:17842] [ 6] enzo(_Z15EvolveHierarchyR14HierarchyEntryR11TopGridDataP16ExternalBoundaryPP19LevelHierarchyEntryd+0x5a6) [0x49e8c6]
[gedeckt:17842] [ 7] enzo(main+0xb42) [0x434cb6]
[gedeckt:17842] [ 8] /lib/libc.so.6(__libc_start_main+0xfd) [0x7ff89f17bc4d]
[gedeckt:17842] [ 9] enzo() [0x4340b9]
[gedeckt:17842] *** End of error message ***

[jwise]

Fixed in 1337:09009e6f3f50. Problem was in VelAnyl?, which deleted the temp variable. With the new grid I/O, the allocated pointers weren't being returned by Grid_ComputeVectorAnalysisFields.C.

[jbordner]

Thanks John! 1337 fixes HDMHD2DCheckOddEvenCouplingOfRiemannSolver and RampedKelvinHelmholtz2D, but MHD2DBlastWave and HD-Sedov-Taylor-BlastWave still crash in Group_WriteGrid(), so I'm reopening this. Link to lcatest 1337 runs