Bug #7779
closed
The "Stop decision" button does not work
Added by Ilja Zakharov over 7 years ago.
Updated over 7 years ago.
Description
In the branch "better-unsafes-and-marks" the button "Stop decision" has stopped working. Nothing happens after pressing the button. But after pressing N times the button and refreshing the page I see N messages "Ajax request error. Maybe connection refused." (Look at a screenshot). It is also strange that I cannot see any errors in Bridge log files.
Files
- File deleted (
Снимок экрана от 2016-12-06 16-41-24.png)
- Priority changed from High to Urgent
Likely this is a consequence of #7658. It would be great to have a corresponding test case.
- Due date set to 12/06/2016
- Status changed from New to Feedback
- % Done changed from 0 to 90
Evgeny Novikov wrote:
Likely this is a consequence of #7658. It would be great to have a corresponding test case.
I can't simulate server timeout or disabling in tests.
Concerning the bug I think the problem was a lot of "task.delete()". It was very long so I've optimized it in commit '08e5c7c'. Please check that 'cancel job' works (I can't do it with real jobs).
Vladimir Gratinskiy wrote:
Evgeny Novikov wrote:
Likely this is a consequence of #7658. It would be great to have a corresponding test case.
I can't simulate server timeout or disabling in tests.
Concerning the bug I think the problem was a lot of "task.delete()". It was very long so I've optimized it in commit '08e5c7c'. Please check that 'cancel job' works (I can't do it with real jobs).
Does this mean that theoretically this should always work? I don't believe that optimizations can solve this issue in general.
Evgeny Novikov wrote:
Vladimir Gratinskiy wrote:
Evgeny Novikov wrote:
Likely this is a consequence of #7658. It would be great to have a corresponding test case.
I can't simulate server timeout or disabling in tests.
Concerning the bug I think the problem was a lot of "task.delete()". It was very long so I've optimized it in commit '08e5c7c'. Please check that 'cancel job' works (I can't do it with real jobs).
Does this mean that theoretically this should always work? I don't believe that optimizations can solve this issue in general.
I didn't find any other place where error can happen. I don't have explanations why there are "ajax errors", but I believe that requests were too long and timeout or something happened.
- Status changed from Feedback to Closed
Let's think that it was fixed. Reopen the issue if this isn't the case.
Also available in: Atom
PDF
