New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
rust-cbindgen: 0.14.4 -> 0.15.0 #101326
rust-cbindgen: 0.14.4 -> 0.15.0 #101326
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM: I've built Thunderbird successfully (since that failed for the bot)
This is most likely the change that now "causes" build failures on Hydra for The several logs I've seen all stop somewhere when compiling @andir: perhaps we can revert for now, but overall I'm not sure what to do next (e.g. how to debug it). |
Interesting. I am pretty sure those build worked for me as well (as stated above). We can give a revert a try but that would be a very odd behavior.. |
My tip is that some property of the machine/builder is triggering this, as it was pretty well reproducible on Hydra but always scheduled on the "hash-named" boxes which are probably very uniform. |
I guess the best step forward would be to reproduce it on some interactive machine. So far I've done these builds only on one (and successfully), now I'm trying another one. |
I've tried three machines now (a few times) but I haven't reproduced it. |
In a few cases we were lucky on Hydra, getting it scheduled to "wendy" machine where it succeeds, e.g. https://hydra.nixos.org/build/128742086#tabs-buildsteps ... but over long term the current rate of failure doesn't seem sustainable. |
Looking at https://hydra.nixos.org/build/128739018 it only timed out at the the install checks:
According to our expression this looks like what I'd expect. Maybe Nix on those "hash" machines (which are probably all managed the same way) is a newer/different version with some bug? |
No, you're too late. Further attempt on the same derivation already succeeded (triggered from a different Hydra job), and the log got overwritten as it's by design one per derivation... |
Luckily I have a few log instances for the very same derivation (logs downloaded from those Hydra.nixos.org links), so for reference here's one: 45aby83kdjlsphifd91zwbns52k87vrq-firefox-esr-unwrapped-78.4.0esr.log |
Shall we try reverting this PR then? The current state isn't good. |
I suppose so. But I see it just as a way to buy more time. |
On 15:44 25.10.20, Vladimír Čunát wrote:
I suppose so. But I see it just as a way to buy more time.
Agreed. An alternative would be to create a new jobset and revert it
there just to see if it actually fixes the problem. I'm not yet sold on
that fact.
|
Apparently this update wasn't the issue. For example, all the timeouts here looked the same and it was with 0.14.4. I hope this will improve with big-parallel (PR #105345). |
Automatic update generated by nixpkgs-update tools. This update was made based on information from https://repology.org/metapackage/rust-cbindgen/versions.
meta.description for rust-cbindgen is: "A project for generating C bindings from Rust code"
meta.homepage for rust-cbindgen is: "https://github.com/eqrion/cbindgen"
meta.changelog for rust-cbindgen is: ""
Updates performed
To inspect upstream changes
Release on GitHub
Compare changes on GitHub
Impact
Checks done (click to expand)
built on NixOS
The tests defined in
passthru.tests
, if any, passed0 of 0 passed binary check by having a zero exit code.
0 of 0 passed binary check by having the new version present in output.
found 0.15.0 with grep in /nix/store/3074537pd7skdl3fbmfxbsdb18qblcrq-rust-cbindgen-0.15.0
directory tree listing: https://gist.github.com/590d5fa332bde0629c6610a87fe63464
du listing: https://gist.github.com/04a17e591f6529659067be3ad2b24480
Rebuild report (if merged into master) (click to expand)
Instructions to test this update (click to expand)
Either download from Cachix:
(The Cachix cache is only trusted for this store-path realization.)
For the Cachix download to work, your user must be in the
trusted-users
list or you can usesudo
since root is effectively trusted.Or, build yourself:
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.
Result of
nixpkgs-review
11 package failed to build:
13 packages built:
Maintainer pings
cc @jtojnar @andir for testing.