FYI: @dotlambda @jbaum98
nixpkgs-review
resultGenerated using nixpkgs-review
.
Command: nixpkgs-review pr 366944
x86_64-linux
x86_64-darwin
aarch64-darwin
@wolfgangwalther I tried to re-create what you saw. And I'm also seeing failures. But, the failures don't really make sense to me in terms of the change I'm proposing here. For example, i'm seeing a failure like this with ax
:
> =========================== short test summary info ============================
> FAILED ax/models/torch/tests/test_acquisition.py::AcquisitionTest::test_optimize - TimeoutError: Test took longer than 60 seconds. To see a profiler output, s...
> FAILED ax/plot/tests/test_contours.py::ContoursTest::test_Contours - TimeoutError: Test took longer than 60 seconds. To see a profiler output, s...
> === 2 failed, 1089 passed, 5 deselected, 1265 warnings in 1472.69s (0:24:32) ===
And, its super unclear to me how upgrading this package would cause another package to experience a timeout while running tests. So, I dunno, I'm at a bit of a loss here.
nixpkgs-review
resultGenerated using nixpkgs-review
.
Command: nixpkgs-review pr 366944
x86_64-linux
x86_64-darwin
aarch64-darwin
Hm, I tried to build some of the failing packages manually on this branch and they all succeeded. Those test failures might just happen because of running multiple builds in parallel with nixpkgs-review.
So while there is a lot of noise in the build results, we should confirm that none of those is a real failure caused by the update.
@wolfgangwalther I tried with --build-args "-j 2"
which may or may not have helped reduce random failures. I got this result:
Command: nixpkgs-review pr 366944
x86_64-darwin
So, 12 failures. Of those, 6 are ".dist" packages that I don't 100% understand what they are. Of the remaining 6, 5 appear to also be failing on Hydra and 1 doesn't seem to be built by hydra?
and 1 doesn't seem to be built by hydra?
Hm, on the release-24.11 branch I do this:
% nix-build -A python312Packages.jupysql --argstr system x86_64-darwin
these 2 paths will be fetched (0.31 MiB download, 1.59 MiB unpacked):
/nix/store/y8wnkdggbqsl6pmsdlz5izlydzv1z5jn-python3.12-jupysql-0.10.13
/nix/store/zn7297sz2g8ray5r48lbbfi1fxhphqqg-python3.12-jupysql-plugin-0.4.5
copying path '/nix/store/zn7297sz2g8ray5r48lbbfi1fxhphqqg-python3.12-jupysql-plugin-0.4.5' from 'https://cache.nixos.org'...
copying path '/nix/store/y8wnkdggbqsl6pmsdlz5izlydzv1z5jn-python3.12-jupysql-0.10.13' from 'https://cache.nixos.org'...
/nix/store/y8wnkdggbqsl6pmsdlz5izlydzv1z5jn-python3.12-jupysql-0.10.13
So jupysql seems to build just fine on that branch. However, I can't build it myself either, so it's very likely a sandboxing issue - not a blocker for me.
I'm running a nixpkgs-review with -j 1
for the 3 platforms I can do now, to confirm the failures on other platforms as well.
Forgot to post my result of running it with "-j1":
nixpkgs-review
resultGenerated using nixpkgs-review
.
Command: nixpkgs-review pr 366944
x86_64-linux
x86_64-darwin
aarch64-darwin
The builds on darwin were run on the community builder, so could still be affected by other builds running in parallel.
I didn't find any blockers for linux, either. So at least x86_64-linux and x86_64-darwin should be good to go.
I wonder whether this should target staging-24.11
instead, given the number of rebuilds? @vcunat
The numbers shown here aren't a big deal. At this moment you could switch to staging-next-24.11
which is in progress and expected to land early next week. That would save a little, but it's normal not to stage these amounts.
Login to write a write a comment.
Changelog: https://pypi.org/project/pkginfo/#pkginfo-changelog (cherry picked from commit 6b32ff9)
pkginfo 1.12.0 supports a new metadata format - verison 2.4. If you run "poetry lock" and it tries to process a package that has updated to use this format, poetry fails with an error. A number of packages on pypi are now using this format - for example Pydantic.
The error that occurs without this change looks something like (edited a bit for clarity):
Things done
nix.conf
? (See Nix manual)sandbox = relaxed
sandbox = true
nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD"
. Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/
)Add a 👍 reaction to pull requests you find important.