Bug for backport of Debian 11 bootstrap changes. I think it will be useful to test Samba 4.15 on Debian 11 also.
This bug was referenced in samba master: c901adaa0d4526deff550806e49976d686122674
Andrew, let's hold off on the backport for a day or 2. When I copied the Debian 10 setup, I didn't check whether liburing-dev is available in Debian 11. It is, so I'll add a patch on top, test it out and open a new MR...
No worries, there isn't a rush just keen to be good and open bugs for things before they got into master rather than only for a possible backport.
This bug was referenced in samba master: 04f188f4d573f0138b75f26d1c18d98329a3446e
Created attachment 16854 [details] Patch for v4-15-test Commits from master cherry-pick with conflicts into v4-15-test, so marked as backport. After fixing conflicts and re-generating the hash, re-rendering causes no changes, so I must have got it right...
Have not bothered with v4-14-test. It doesn't have Fedora 34, so I don't suppose we backport to all existing releases?
Just a note that a backport isn't a trivial as I had hoped, the oss-fuzz changes would need to be backported or we would need to re-generate the container images (it creates a different hash). So lets close this for now and someone can re-open if they have time to sort that out. In general I'm keep to keep the bootstrap stuff as a linear series going forward, to avoid having a lot of different image sets. Sorry!
(In reply to Andrew Bartlett from comment #7) Never mind, sorry I didn't actually read your updated comment.
Comment on attachment 16854 [details] Patch for v4-15-test Argh, sorry for the flip-flop. With the new hash the new container images need to be generated, otherwise we can't do CI on 4.15 in GitLab. The images used are based on that hash (this is how we tie container images back to the code). Push a pipeline with the magic options and they will appear.
(or backport the oss-fuzz fixes also, so we get to the same hash as master) Either way, you need to run a CI job with 4.15 in GitLab with this backported to show it works, as an sn-devel build won't be any validation of this change. Point to the successful pipeline here as proof and I can put my review back on the possibly updated patch.
Hi Andrew, Successful CI pipeline here: https://gitlab.com/samba-team/devel/samba/-/pipelines/394312460 Thanks!
Comment on attachment 16854 [details] Patch for v4-15-test Looks good, sha1 matches between the CI and the final patch, so good to go!
Assigning to Jule for next available 4.15.next (not urgent, just helpful).
Pushed to autobuild-v4-15-test.
This bug was referenced in samba v4-15-test: 6ea70022f20b6c03b58cacf628bba1cebedbcf9f 71792ae9886c5605e4207337b12f032e33ee9bc5
Closing out bug report. Thanks!
This bug was referenced in samba v4-15-stable (Release samba-4.15.3): 6ea70022f20b6c03b58cacf628bba1cebedbcf9f 71792ae9886c5605e4207337b12f032e33ee9bc5