Created attachment 195929 [details] py-buildbot-assign-to-sqlalchemy12.patch Hello, attached is the patch that assigns devel/py-buildbot from databases/py-sqlalchemy10 to databases/py-sqlalchemy12. This is a prophylactic patch to prevent build failures when PR #230379 has been committed. QA: ~~~ - poudriere (10.4-, 11.1-, 11.2-RELEASE, 12.0-CURRENT @r337160) for each amd64 + i386 arch and each py27 + py36 flavor -> OK - portlint -> OK Note: make test yields following results: > Ran 5230 tests in 83.367s > > FAILED (skips=101, failures=15, errors=4, successes=5113) > *** Error code 1 OTOH the same results are there when databases/py-sqlalchemy10 is used.
This then (must be committed before) blocks bug 230379, rather than depending on it
Thank you Kai. buildbots setup.py does not pin sqlalchemy* to < anything [1], so this change looks OK to me. 'sqlalchemy>=0.8.0', 'sqlalchemy-migrate>=0.9', Is this related to the 'conflicting (sqlalchemy) dependencies' BROKEN/fallout that occurred recently?
Mmm, on second view, it appears both this and bug 230379 need to be committed (atomically) together/at the same time. @Kai Is this interpretation correct?
This bug wasn't assigned to me (maintainer), because summary contained two category/port references (maintainers are CC'd, not assigned in this case)
(In reply to Kubilay Kocak from comment #2) (In reply to Kubilay Kocak from comment #3) Hello Koobs, thank you for your quick info. Yes, this patch is related to the conflicting sqlalchemy dependencies and is meant as prophylactic measure to avoid further build failures. This would be the best option when this patch will be committed (atomically) together with bug #230379 at the same time.
(In reply to Kai from comment #5) If this and bug 230379 need to land together, then presumably that is also true for all the other ports (and bugs that have been created for them) that also depend on sqlalchemy-migrate, yes? If so, I would recommend combining (closing) all separate bugs/issues into a single one (bug 230379), assigned to the sqlalchemy-migrate maintainer, as this qualifies as a 'sweeping commit' where a single dependency is updated, and all the dependents must be checked/updated as well. The patch there should include changing the dependency for dependent ports too.
(In reply to Kubilay Kocak from comment #6) Your suggested approach makes sense. In that case I would attach the related patches from this bug and bug #230395 to bug #230379 . When done so I would close this bug and bug #230395 with resolution "DUPLICATE" if it's apt in that case?
Yep, marking as duplicate is fine, adding the comment 'Superseded by bug 230379' to clarify and be explicit. There's no special syntax/attributes in Bugzilla for 'supersedes' Keep up the good work :)
(In reply to Kubilay Kocak from comment #8) Thank you for the info and feedback. :) Superseded by bug #230379 - the patch were added there and the further progress can be tracked there. *** This bug has been marked as a duplicate of bug 230379 ***