#1234 test-alberta-3-3 segfaults due to stack overflow
Metadata
| Property | Value |
|---|---|
| Reported by | Markus Blatt (markus@dr-blatt.de) |
| Reported at | Jan 21, 2013 12:47 |
| Type | Bug Report |
| Version | 2.2 |
| Operating System | Unspecified / All |
| Last edited by | Christoph Grüninger (gruenich@iws.uni-stuttgart.de) |
| Last edited at | May 11, 2013 15:55 |
| Closed by | Christoph Grüninger (gruenich@iws.uni-stuttgart.de) |
| Closed at | May 11, 2013 15:55 |
| Closed in version | 2.3 |
| Resolution | Fixed |
| Comment | Fixed, test-alberta-3-3 uses grid3Y.dgf since commit bb637bc688678bdcab67e04163c0f8d048fe894a |
test-alberta3d-refine is an other story. |
Description
The problem occurs both with the trunk and the release branch and with at least alberta2.1 and alberta3.0.0.
Testing AlbertaGrid< 3, 3 >...
Checking AlbertaReader... macro_test_3d: WARNING in ./../3d/macro_3d.c, line 368 not implemented for 3d yet: no check is performed not implemented for 3d yet: no check is performed check_mesh: checking mesh 'DUNE AlbertaGrid' checking done; no error detected Warning: Skipping level neighbor check. Warning: Skipping level neighbor check. Warning: Skipping index test using LevelIntersectionIterator. Warning: Skipping boundary segment index check (missing level intersection iterator).
AlbertaGrid< 3, 3 > with grid file: ../../../doc/grids/dgf/simplex-testgrid-3-3.dgf
GridParameterBlock: Parameter 'refinementedge' not specified, defaulting to 'ARBITRARY'. macro_test_3d: WARNING in ./../3d/macro_3d.c, line 368 not implemented for 3d yet: no check is performed check_mesh: checking mesh 'DUNE AlbertaGrid' checking done; no error detected
Checking macro grid... Program received signal SIGSEGV, Segmentation fault. 0x000000000068f23c in fill_macro_info_3d (mesh=<error reading variable: Cannot access memory at address 0x7fffff7aefe8>, mel=<error reading variable: Cannot access memory at address 0x7fffff7aefe0>, el_info=<error reading variable: Cannot access memory at address 0x7fffff7aefd8>) at ./../3d/traverse_r_3d.c:61
Valgrind suspected that the problem might be a stack overflow. Indeed printing a backtrace in gdb takes forever. It seems like there is recursion without an end. I will attach the gdb output with the backtrace.