bug#58203: Cuirass fails to respect (target-x86-32?) in some circumstances

bug#58203: Cuirass fails to respect (target-x86-32?) in some circumstances

[Marius Bakke]

[-- Attachment #1: Type: text/plain, Size: 283 bytes --]

Hello,

I noticed Cuirass attempts to run imath tests on i686-linux:

  https://ci.guix.gnu.org/build/739643/details

Which is odd, because the package definition has:

  (list #:tests? (not (target-x86-32?)))

This behavior does not reproduce when manually building for i686-linux.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

bug#58203: Cuirass fails to respect (target-x86-32?) in some circumstances

[Ludovic Courtès]

Marius Bakke <marius@gnu.org> skribis:

> Hello,
>
> I noticed Cuirass attempts to run imath tests on i686-linux:
>
>   https://ci.guix.gnu.org/build/739643/details

Following the links, this corresponds to
<https://ci.guix.gnu.org/eval/302107>, itself corresponding to commit
f5fe0082abe4547f3fb9f29d8351473cfb3a387b, which dates back to May.

So this is a recent build for an old commit; maybe it had been queued
for some time and just got unlocked today?  Weirdness.

The good news is that with today’s ‘staging’, everything’s fine:

--8<---------------cut here---------------start------------->8---
$ guix time-machine --commit=e0546a11f0b65e301dc9d1d1cfec26e686070029 -- weather imath -s i686-linux
computing 1 package derivations for i686-linux...
looking for 1 store items on https://ci.guix.gnu.org...
https://ci.guix.gnu.org ☀
  100.0% substitutes available (1 out of 1)
  at least 0.2 MiB of nars (compressed)
  0.9 MiB on disk (uncompressed)
  0.230 seconds per request (0.2 seconds in total)
  4.3 requests per second

[...]

$ guix time-machine --commit=e0546a11f0b65e301dc9d1d1cfec26e686070029 -- build imath -s i686-linux -d --no-grafts
/gnu/store/g31xfskliryh10gij1hidydykqk7bb6z-imath-3.1.3.drv
--8<---------------cut here---------------end--------------->8---

Thanks,
Ludo’.

bug#58203: Cuirass fails to respect (target-x86-32?) in some circumstances

[Marius Bakke]

[-- Attachment #1: Type: text/plain, Size: 960 bytes --]

Ludovic Courtès <ludo@gnu.org> skriver:

> Marius Bakke <marius@gnu.org> skribis:
>
>> Hello,
>>
>> I noticed Cuirass attempts to run imath tests on i686-linux:
>>
>>   https://ci.guix.gnu.org/build/739643/details
>
> Following the links, this corresponds to
> <https://ci.guix.gnu.org/eval/302107>, itself corresponding to commit
> f5fe0082abe4547f3fb9f29d8351473cfb3a387b, which dates back to May.
>
> So this is a recent build for an old commit; maybe it had been queued
> for some time and just got unlocked today?  Weirdness.

Hmm.  I was following some "dependency failed" links in Cuirass which
lead to that.  It was recently built because I restarted it.

> The good news is that with today’s ‘staging’, everything’s fine:

Not quite:

  https://ci.guix.gnu.org/search?query=system%3Ai686-linux+imath

Imath has no dependencies apart from CMake, so all of those should be
the same derivation, yet some are failing.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

bug#58203: Cuirass fails to respect (target-x86-32?) in some circumstances

[Ludovic Courtès]

Hi Marius,

(Cc: Mathieu for the search box issue below.)

Marius Bakke <marius@gnu.org> skribis:

> Ludovic Courtès <ludo@gnu.org> skriver:

[...]

>> The good news is that with today’s ‘staging’, everything’s fine:
>
> Not quite:
>
>   https://ci.guix.gnu.org/search?query=system%3Ai686-linux+imath
>
> Imath has no dependencies apart from CMake, so all of those should be
> the same derivation, yet some are failing.

This is for:

  /gnu/store/sdgcjk8wiv4wk69nlxwp43adppmvijpz-imath-3.1.3.drv

which has #:tests? #true… but that’s also from an evaluation from May,
even if the build is from Sep. 30:

  https://ci.guix.gnu.org/build/739643/details
  https://ci.guix.gnu.org/eval/302107

From <https://ci.guix.gnu.org/eval/683442/dashboard?system=i686-linux>,
I found this recent build ✅:

  https://ci.guix.gnu.org/build/1324091/details

It’s a bug that we don’t get it from the search box:

  https://ci.guix.gnu.org/search?query=spec%3Astaging+system%3Ai686-linux+imath

Mathieu, any ideas?

Anyhow, the “ground truth” in general is Guix itself:

  guix time-machine --branch=staging -- \
    build imath -s i686-linux --no-grafts

Thanks,
Ludo’.