apport-retrace doesn't check to see if a core file is bad
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Apport |
New
|
Undecided
|
Unassigned |
Bug Description
The Error Tracker retraces do a first check of a core file before setting up sandboxes and downloading packages, however apport itself does not. Here's what the ET retracers do:
650 # confirm that gdb thinks the core file is good
651 gdb_cmd = [self.gdb_path, "--batch", "--ex", "target core %s" %
652 core_file]
653 proc = Popen(gdb_cmd, stdout=PIPE, stderr=PIPE,
654 universal_
655 (out, err) = proc.communicate()
656 if 'is truncated: expected core file size' in err or \
657 'not a core dump' in err:
It'd be good if apport-retrace itself did this check and we avoided ambiguous failures to retrace like bug 1760286. I mean it has a bad core dump but nothing in the bug report indicates that.
summary: |
- No detection of bad core files done + apport-retrace doesn't check to see if a core file is bad |
tags: | added: id-5ac258c846d018c62db1c1bd |