fix precompilation on 1.13

[araujoms]

Part of #180

The Hessenberg changes were backported to 1.13, so the compat bound needs to be changed here as well. Unfortunately the bugfix JuliaLang/LinearAlgebra.jl#1602 wasn't backported together with it, so GenericLinearAlgebra will remain broken until that happens.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.73%. Comparing base (2f53d98) to head (8e96376).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #181   +/-   ##
=======================================
  Coverage   95.73%   95.73%           
=======================================
  Files          11       11           
  Lines        1524     1524           
=======================================
  Hits         1459     1459           
  Misses         65       65           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[andreasnoack]

Thanks for looking into this. Hopefully, we can get the backport in before the final release of 1.13.

[araujoms]

I feel it's my responsibility.

Backport is in, so it should work fine when 1.13 is released. Perhaps it's better to wait for that to merge this PR? Because then we can get CI to confirm, and it's anyway useless before it.

[davidanthoff]

It would be great if this could be merged now. GenericLinearAlgebra is deep in the stack, and the current situation prevents any downstream package from getting ready for 1.13. Also, there is a 1.13 release candidate out, and that is exactly the phase when packages ideally should start to fix compat problems, so that we don't have a situation where lots of things are broken when 1.13 is released. Is there any reason to not fix it now? I think in general it is a good idea to add Julia RC candidates to CI, so that could take care of actually running CI on it.

[andreasnoack]

@davidanthoff see the discussion above. We need rc4 before the fix here will pass. This package already tests on prereleases, and that is why the CI is failing here. We also have a scheduled CI run exactly to detect these failures. It only runs monthly, though, which maybe adds too long a delay. The issue here was reported on 14 August and was caused by rc2, which was released on 12 August. The scheduled run on 1 August passed on rc1.

[davidanthoff]

Ah, sorry, I misunderstood.