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
sbcl: 1.4.7 -> 1.4.9; tested on nixos #42826
Conversation
@GrahamcOfBorg build sbcl |
Success on x86_64-linux (full log) Attempted: sbcl Partial log (click to expand)
|
Success on x86_64-darwin (full log) Attempted: sbcl Partial log (click to expand)
|
Failure on aarch64-linux (full log) Attempted: sbcl Partial log (click to expand)
|
Any ideas what we need to change for aarch64 to work again? Or is there a reason/way to just disable something related to sb-concurrency? |
It should work as threads are supported upstream on aarch64-linux and
apart from version bump, there were no changes in nixpkgs connected to
sbcl I think.
Unfortunately I can't see the log at
https://logs.nix.ci/?key=nixos/nixpkgs.42826&attempt_id=3369e726-9c3b-4618-9a3f-2c892eb5de88
as it doesn't work without Javascript.
I also don't have aarch64-linux machine to test this anymore.
It is also possible that sb-concurrency failure is transient. Could it
be tested by grahamc just on aarch64-linux again?
Otherwise one could disable threadSupport on aarch64-linux.
|
@GrahamcOfBorg build sbcl |
Success on x86_64-darwin (full log) Attempted: sbcl Partial log (click to expand)
|
Success on x86_64-linux (full log) Attempted: sbcl Partial log (click to expand)
|
Weird results in the module test… |
Failure on aarch64-linux (full log) Attempted: sbcl Partial log (click to expand)
|
Hm, seems to be repeatable. |
Michael Raskin <notifications@github.com> writes:
`sb-concurrency` is a contrib module, it should be possible to nuke
just it with less collateral damage than nuking threads completely.
probably sed make-target-contrib.sh and touch
obj/asdf-cache/sb-concurrency/test-passed.test-report at the right place
1 out of 25 total tests failed: SB-CONCURRENCY-TEST::FRLOCK.1.
maybe a timing problem
asking on irc #sbcl could give some ideas, e.g. if that is known issue
sometimes sbcl tests don't work on a platform and need to be marked
accordingly, but i would be surprised if this was the case here
sbcl 1.4.7 worked fine, but unfortunatelly i can't bisect this
|
Hm. Looks like an actual consistency problem. Need to look at upstream ML at some point… |
1.4.10 seems to lack this problem (and 1.4.11 breaks Maxima…). With 1.4.10 in |
Michael Raskin <notifications@github.com> writes:
1.4.10 seems to lack this problem (and 1.4.11 breaks Maxima…). With
1.4.10 in `master` I close this as «unclear whether there is an
upstream problem»
great, thank you
|
No description provided.