Workaround issue send_key 'ret' for desktop_runner.pm #5275
Conversation
- verification run: http://e13.suse.de/tests/2983 - see poo#37354 for details
There was a problem hiding this comment.
Unfortunately your change does not make any sense and would break all tests except desktop_runner. Besides, for a sporadic issue I suggest to not just provide a single job for verification. We already discussed this in the last review meeting: First reproduce with good statistics, then fix and verify with good statistics. If we can not reproduce the original issue or do not yet understand what the problem is we should also not try to change any test code
| send_key 'ret'; | ||
| # sometimes send_key 'ret' doesn't work because of low performance | ||
| # so workaround this with repeat send_key can help | ||
| send_key_until_needlematch 'generic-desktop', "ret", 5, 5; |
There was a problem hiding this comment.
This will not work in general as 'generic-desktop' is only the very specific target we want to reach in case of the desktop runner test module. And also I do not think we should recheck the final screen we want to screen more than needed. As every post_run_hook of x11test based test modules would check for generic-desktop anyway we would do it twice. So besides that this idea in general does not make sense and can not work I would ensure that in the post_run_hook the desktop is only looked for if the last match was not already exactly that
There was a problem hiding this comment.
okay, then no changes here. I don't know a better to workaround issue related on openQA production server. The statistic shows only problem with low performance on server.
|
close this PR for now. |
http://e13.suse.de/tests/2983