Thank you for your report.
I analyzed this problem and bumped into some issues.
I have used the commit referenced above to solve that bug, but it turned out that it does not cherry-pick cleanly. It means that there have been other changes upstream, that change the context of the patch - this makes it hard to apply and risky as there might be side effects. The fix is too complicated to make - as is - progress on it right now due to all the churn.
According to the SRU policy, we should only fix something which will not regress other aspects.
But backporting this patch rings all alarm bells.
We’d need to modify this and potentially find more patches to make it apply cleanly and with the confidence of not regressing. Given that we can’t reproduce, no one else chimed in and that it is hard to solve as-is we can not make progress right now.
I’d be happy to support if there is a volunteer identifying these changes and providing a reviewable change to the packaging that is supposed to fix this issue - that can be built/tested and passed into the SRU processing.
It would also help to have more exact steps to reproduce (without MacOS). The steps are necessary, because the person investigating the bug may not be as familiar with that case as you are. It is possible that these steps could help us to figure out some workaround to help with that problem.
Sorry, but since I can’t resolve it right now I’ll unassign myself waiting for those steps and/or proposed changes.
Hi Andrew,
Thank you for your report.
I analyzed this problem and bumped into some issues.
I have used the commit referenced above to solve that bug, but it turned out that it does not cherry-pick cleanly. It means that there have been other changes upstream, that change the context of the patch - this makes it hard to apply and risky as there might be side effects. The fix is too complicated to make - as is - progress on it right now due to all the churn.
According to the SRU policy, we should only fix something which will not regress other aspects.
But backporting this patch rings all alarm bells.
We’d need to modify this and potentially find more patches to make it apply cleanly and with the confidence of not regressing. Given that we can’t reproduce, no one else chimed in and that it is hard to solve as-is we can not make progress right now.
I’d be happy to support if there is a volunteer identifying these changes and providing a reviewable change to the packaging that is supposed to fix this issue - that can be built/tested and passed into the SRU processing.
It would also help to have more exact steps to reproduce (without MacOS). The steps are necessary, because the person investigating the bug may not be as familiar with that case as you are. It is possible that these steps could help us to figure out some workaround to help with that problem.
Sorry, but since I can’t resolve it right now I’ll unassign myself waiting for those steps and/or proposed changes.