Description:
This isn't seen on pushbuild1 because, at least for 6.0-maria, pushbuild1 does only non-debug builds and crash_commit_before is skipped with non-debug builds.
I have a debug binary of MySQL 6.0 (6.0-maria) and crash_commit_before fails like this:
C:\Users\Administrateur\bzrrepos\mysql-6.0-maria\mysql-test>perl mysql-test-run.
pl crash_commit_before
MySQL Version 6.0.11
==============================================================================
TEST RESULT TIME (ms)
------------------------------------------------------------
worker[1] Using MTR_BUILD_THREAD 250, with reserved ports 12500..12509
worker[1] mysql-test-run: WARNING: running this script as _root_ will cause some
tests to be skipped
worker[1] mysql-test-run: WARNING: Process [mysqld.1 - pid: 4008, winpid: 4008,
exit: 256] died
Could not open event or process 0, error: 87
Couldn't open the winpid: for pid: 4008, try one more time
Could not open event or process 0, error: 87
Couldn't open the winpid: for pid: 4008, continue and see what happens...
main.crash_commit_before [ fail ]
Test ended at 2009-03-18 14:34:31
CURRENT_TEST: main.crash_commit_before
safe_process[3784]: FATAL ERROR, Unknown option: --nocore
Failed to start mysqld.1
The --nocore is fishy: safe_process_win.cc does not support this option it seems (looking at code), so SafeProcess.pm should not pass this option. If I remove the line passing --nocore, from SafeProcess.pm, the test passes.
How to repeat:
see description
Suggested fix:
Either tell SafeProcess.pm to not use this option under Windows, or tell safe_process_win.cc to ignore it (instead of bailing out).
Contact me if you need me to test a patch.
Description: This isn't seen on pushbuild1 because, at least for 6.0-maria, pushbuild1 does only non-debug builds and crash_commit_before is skipped with non-debug builds. I have a debug binary of MySQL 6.0 (6.0-maria) and crash_commit_before fails like this: C:\Users\Administrateur\bzrrepos\mysql-6.0-maria\mysql-test>perl mysql-test-run. pl crash_commit_before MySQL Version 6.0.11 ============================================================================== TEST RESULT TIME (ms) ------------------------------------------------------------ worker[1] Using MTR_BUILD_THREAD 250, with reserved ports 12500..12509 worker[1] mysql-test-run: WARNING: running this script as _root_ will cause some tests to be skipped worker[1] mysql-test-run: WARNING: Process [mysqld.1 - pid: 4008, winpid: 4008, exit: 256] died Could not open event or process 0, error: 87 Couldn't open the winpid: for pid: 4008, try one more time Could not open event or process 0, error: 87 Couldn't open the winpid: for pid: 4008, continue and see what happens... main.crash_commit_before [ fail ] Test ended at 2009-03-18 14:34:31 CURRENT_TEST: main.crash_commit_before safe_process[3784]: FATAL ERROR, Unknown option: --nocore Failed to start mysqld.1 The --nocore is fishy: safe_process_win.cc does not support this option it seems (looking at code), so SafeProcess.pm should not pass this option. If I remove the line passing --nocore, from SafeProcess.pm, the test passes. How to repeat: see description Suggested fix: Either tell SafeProcess.pm to not use this option under Windows, or tell safe_process_win.cc to ignore it (instead of bailing out). Contact me if you need me to test a patch.