Follow up on https://bugzilla.redhat.com/show_bug.cgi?id=1484086 and ensure it has been addressed in a reasonable manner
Note: this is the second time in the recent past that an SQLite update has broken searching for files. Since this problem is basically identical to the last time this happened, we should review to make sure it's not going to happen again going forward.
Metadata Update from @catanzaro: - Issue tagged with: meeting
I've submitted a tracker+sqlite update to bodhi which fixes this for F26, https://bodhi.fedoraproject.org/updates/FEDORA-2017-31aabe2be0
Quick karma/+1 in the bodhi update ticket is appreciated, if anyone could do that.
why was this not reverted ?
I don't know, that's a question for mcatanzaro who was pushing for that. I went for the least path of resistance here to get a quick fix out to users.
Because the SQLite maintainers did not respond to the bug report. Since this same problem has happened twice in as many years, it seems safe to say it's just going to happen again if we don't receive confirmation that the SQLite maintainers understand the Fedora update policy and also that they will test the desktop experience before releasing future updates.
isn't bodhi supposed to catch this sort of breakage before an update goes out ?
I hope you're kidding. bodhi apparently failed to detect the soname bump in the broken imagemagick update that's causing so much grief on devel@ right now. (Probably it did actually detect that, but nobody ever looks at bodhi warnings as they're hidden away.) I don't think there was any soname bump here, so bodhi would have to run some sort of OpenQA-style test in order to determine that desktop search has been broken. Right now we only test new releases, not updates. It's a free-for-all.
Twice-monthly tested update packs would help mitigate this, and would be independently valuable.
I added a comment in the bug, but it seems to me this is exactly the sort of breakage the CI efforts in Fedora would prevent. The key is to automate this effort, which may be orthogonal to the issue of regular/slower update packs. https://fedoraproject.org/wiki/CI
why was this not reverted ? Because the SQLite maintainers did not respond to the bug report. Since this same problem has happened twice in as many years, it seems safe to say it's just going to happen again if we don't receive confirmation that the SQLite maintainers understand the Fedora update policy and also that they will test the desktop experience before releasing future updates.
They apologized and promised to be more careful in the future. I don't think any follow-up is required for SQLite specifically.
We still have a problem with updates not receiving as much testing as our GA releases, which is worth discussing, but we don't need to do that in this issue.
Metadata Update from @catanzaro: - Issue untagged with: meeting - Issue status updated to: Closed (was: Open)
Log in to comment on this ticket.