I have no idea what the failing checks mean. I assume it's just a hickup.
Can someone manually trigger the build again? (I think, I don't have the required permissions)

@GrahamcOfBorg build sqlite sqlite-analyzer

I ran the compilation for both on NixOS, and both succeeded. Also tested that this indeed fixes the issue using the snippet from https://www.sqlite.org/src/info/e4598ecbdd18bd82945f6029013296690e719a62 .

Pinging @edolstra as @eelco is non-existent.

@cko: given that nobody has come around yet, I'd suggest to rebase and push force again, that should retrigger ofBorg.

@d-goldin that's a good idea, thank you! Done.

Poke security team: @fpletz @domenkozar @grahamc

@GrahamcOfBorg build pythonPackages.apsw python3Packages.sqlalchemy

Testing these two because I saw build failures when I applied this to release-19.09, not sure if master has the same problem.

I tried to fix those two by bumping to the latest versions, but no luck:

sqlalchemy test failure

apsw test failure / issue: rogerbinns/apsw#275

@ivan: Thanks for pointing that out. I briefly looked around and it also does not seem like there is an issue filed with sqlalchemy yet, so it might take a little while. Also because the failure actually looks like some more significant change has happened.

I don't think this issue is too high severity, but since we're at it, maybe we can go with a backported fix like in: d-goldin@3958e02

I ran sqlalchemy and apsw against it, and it seemed to pass the tests fine.

One commit per package please.

This seems somewhat stalled a bit. Should I open a PR just for SQLite with the backported fix instead?

Actually, 6101751 went in so closing.

@FRidh: Alright, missed that one. Keep in mind that it might have breakages with sqlalchemy, as outlined above.

Yep, I pushed a fix for sqlalchemy.

@FRidh: Ah, great that there is a fix a vailable upstream by now. Is the plan to backport the bumped version + sqlalchemy fix to release-19.09 too, or might a backport be more suitable for that?

I don't know. I kind of hoped we would not have to backport a sqlite version bump

@FRidh: I agree, that would be not overly stable releasy. I propose the backport PR listed above my comment.

FWIW sqlite 3.30.1 has been released! \o/
In my own testing/usage there were no new failures and the existing problems re:sqlalchemy and APSW are still present.

FWIW sqlite 3.30.1 has been released! \o/
In my own testing/usage there were no new failures and the existing problems re:sqlalchemy and APSW are still present.

Could you open a PR? I think we can keep that one for the next iteration. Most things have already been built now with 3.30.0.