matrix-tuwunel: 1.6.1 -> 1.7.0#525790
Conversation
|
This fails to build for me with: Weirdly enough, if I rebase this commit on top of the latest master, the expected checksum changes: I have never seen this before. |
|
I think we're hitting the same problem as the Cosmic DE version bump in #524670 and this should hopefully be fixed in #525255. What I think is happening is:
Copying the patch from the #525255 on top of this PR makes the build succeed locally for me. This leads me to think the hash in this PR is the correct one and my machine had the non-deterministically bad clone. I'm going to deploy this to my server and approve afterwards. |
scvalex
left a comment
There was a problem hiding this comment.
Changes look good. I deployed to my server and everything seems to be working (including the database migrations that were broken in 1.6.2).
I'm going to backport this to 26.05 tomorrow. I want to wait a day because I don't want to test whether backports work on release day.
|
@NixOS/nixpkgs-merge-bot merge |
|
@scvalex wants to merge this PR. Requirements to merge this PR with
|
|
Successfully created backport PR for |
|
I was able to reproduce Hydra's result: https://hydra.nixos.org/build/330696187/nixlog/1 Probably not worth escalating it if I'm the only one affected. |
|
Huh. I built it locally on aarch64-linux without issue issue when I merged this. This is the step that always takes the longest in the build. On Hetzner's smallest ARM64 VM, it takes about 4h to complete. I am traveling at the moment, so I won't have a chance to do any debugging until this weekend. |
|
Ok, yes, it took some hours, but if you watch the processes you see some tests running taking more than 10min each, eventually it finished. |
|
I'm looking at this now and Hydra hasn't tried to rebuild. We can see that it built 1.7.0 successfully on x86-64, but it timed out after 4h on aarch64. It was building 1.6.1 successfully in about 30min and I have no reason to expect it would need significantly longer for 1.7.0. I am mildly tempted to disable the tests on aarch64. Presumably the thing that's hanging is tuwunel's built-in smoke tests, but we already exercise that with NixOS tests that run the server. I also don't know why the smoke tests would be behaving differently on hydra than on my laptop. |
Automatic update generated by nixpkgs-update tools. This update was made based on information from passthru.updateScript.
meta.description for matrix-tuwunel is: Matrix homeserver written in Rust, official successor to conduwuit
meta.homepage for matrix-tuwunel is: https://github.com/matrix-construct/tuwunel
meta.changelog for matrix-tuwunel is: https://github.com/matrix-construct/tuwunel/releases/tag/v1.7.0
Updates performed
To inspect upstream changes
Impact
Checks done
passthru.tests, if any, passedRebuild report (if merged into master) (click to expand)
Instructions to test this update (click to expand)
Either download from the cache:
(The nixpkgs-update cache is only trusted for this store-path realization.)
For the cached download to work, your user must be in the
trusted-userslist or you can usesudosince root is effectively trusted.Or, build yourself:
Or:
After you've downloaded or built it, look at the files and if there are any, run the binaries:
Pre-merge build results
We have automatically built all packages that will get rebuilt due to
this change.
This gives evidence on whether the upgrade will break dependent packages.
Note sometimes packages show up as failed to build independent of the
change, simply because they are already broken on the target branch.
nixpkgs-reviewresultGenerated using
nixpkgs-review.Command:
nixpkgs-review --extra-nixpkgs-config '{ allowBroken = false; }'Commit:
556d06edc30595359d51064822b2378a5ed4bc32x86_64-linux✅ 1 package built:
Maintainer pings
cc @scvalex for testing.
Tip
As a maintainer, if your package is located under
pkgs/by-name/*, you can comment@NixOS/nixpkgs-merge-bot mergeto automatically merge this update using thenixpkgs-merge-bot.