openblas: disable zblas3 test on i686

[Nivalux]

Fixes: #534670 (same behaviour on current staging-next)

Alternative to: #534770

Things done

• Built on platform:
  • x86_64-linux
  • aarch64-linux
  • aarch64-darwin
• Tested, as applicable:
  • NixOS tests in nixos/tests.
  • Package tests at passthru.tests.
  • Tests in lib/tests or pkgs/test for functions and "core" functionality.
• Ran nixpkgs-review on this PR. See nixpkgs-review usage.
• Tested basic functionality of all binary files, usually in ./result/bin/.
• Nixpkgs Release Notes
  • Package update: when the change is major or breaking.
• NixOS Release Notes
  • Module addition: when adding a new NixOS module.
  • Module update: when the change is significant.
• Fits CONTRIBUTING.md, pkgs/README.md, maintainers/README.md and other READMEs.
• Follows the automation/AI policy.

[Nivalux]

Running tests...
/nix/store/mgvw77gvbc5kmv5rbrxhqfhqiq2qdpvk-cmake-4.1.2/bin/ctest -E zblas3
Test project /build/source/build
      Start  1: openblas_utest
      Start  2: openblas_utest_ext
      Start  3: sblas1
      Start  4: sblas2
      Start  5: sblas3
      Start  6: dblas1
      Start  7: dblas2
      Start  8: dblas3
      Start  9: cblas1
      Start 10: cblas2
      Start 11: cblas3
      Start 12: cblas3_3m
      Start 13: zblas1
      Start 14: zblas2
      Start 15: xscblat1
      Start 16: xscblat2
      Start 17: xscblat3
      Start 18: xdcblat1
      Start 19: xdcblat2
      Start 20: xdcblat3
      Start 21: xccblat1
      Start 22: xccblat2
      Start 23: xccblat3
      Start 24: xccblat3_3m
 1/28 Test  #1: openblas_utest ...................   Passed    0.04 sec
 2/28 Test  #2: openblas_utest_ext ...............   Passed    0.04 sec
 3/28 Test  #3: sblas1 ...........................   Passed    0.03 sec
 4/28 Test  #6: dblas1 ...........................   Passed    0.03 sec
 5/28 Test  #9: cblas1 ...........................   Passed    0.03 sec
 6/28 Test #13: zblas1 ...........................   Passed    0.03 sec
 7/28 Test #15: xscblat1 .........................   Passed    0.02 sec
 8/28 Test #18: xdcblat1 .........................   Passed    0.01 sec
 9/28 Test #21: xccblat1 .........................   Passed    0.01 sec
      Start 25: xzcblat1
      Start 26: xzcblat2
      Start 27: xzcblat3
      Start 28: xzcblat3_3m
10/28 Test #25: xzcblat1 .........................   Passed    0.00 sec
11/28 Test #24: xccblat3_3m ......................   Passed    0.05 sec
12/28 Test  #8: dblas3 ...........................   Passed    0.09 sec
13/28 Test #23: xccblat3 .........................   Passed    0.07 sec
14/28 Test #12: cblas3_3m ........................   Passed    0.13 sec
15/28 Test  #5: sblas3 ...........................   Passed    0.30 sec
16/28 Test #28: xzcblat3_3m ......................   Passed    0.27 sec
17/28 Test #11: cblas3 ...........................   Passed   16.27 sec
18/28 Test #20: xdcblat3 .........................   Passed   38.21 sec
19/28 Test #17: xscblat3 .........................   Passed   38.22 sec
20/28 Test #27: xzcblat3 .........................   Passed   56.08 sec
21/28 Test  #7: dblas2 ...........................   Passed  109.20 sec
22/28 Test  #4: sblas2 ...........................   Passed  137.32 sec
23/28 Test #10: cblas2 ...........................   Passed  152.58 sec
24/28 Test #14: zblas2 ...........................   Passed  168.40 sec
25/28 Test #22: xccblat2 .........................   Passed  223.40 sec
26/28 Test #16: xscblat2 .........................   Passed  223.44 sec
27/28 Test #19: xdcblat2 .........................   Passed  225.45 sec
28/28 Test #26: xzcblat2 .........................   Passed  225.66 sec

100% tests passed, 0 tests failed out of 28

Total Test time (real) = 225.70 sec

[Nivalux]

The backport to 26.05, should it target release- or staging-? Not sure how many packages this will try to rebuild on 26.05.

Edit: in the past backport PRs it used staging. Changed the bp label to staging.

[Nivalux]

Is the target branch "staging-next" ok in this case?
I targeted it, because it has the openblas update, which has this issue.
Or should the PR target "staging"?

[Atemu]

I think we should figure out why this broke. Tests exist for a reason and this implies to me that there is some kind of performance regression?

OTOH we could also simply decide to not care about i686-linux. In that case I'd prefer to explicitly disable all tests to signify this non-support.

[vcunat]

This is too large a rebuild for staging-next at this moment, so I picked the workaround for now as fef886e.

[Atemu]

Front-porting 🙈️

It would have made sense to send the patch to staging-next first I suppose; I was too focussed on unbreaking stable.

[Nivalux]

Fix has been included in:

#534570