Ensure det for 0x0 matrices always returns 1 (and not 0)

[fingolfin]

Should get some tests... later

[codecov]

Codecov Report

❌ Patch coverage is 89.74359% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 88.68%. Comparing base (b548c14) to head (a4e789b).
⚠️ Report is 16 commits behind head on master.

Files with missing lines	Patch %	Lines
ext/TestExt/Rings-conformance-tests.jl	86.66%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2288      +/-   ##
==========================================
+ Coverage   88.22%   88.68%   +0.45%     
==========================================
  Files         126      126              
  Lines       32000    33859    +1859     
==========================================
+ Hits        28233    30027    +1794     
- Misses       3767     3832      +65     

☔ View full report in Codecov by Sentry.
📢 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.

[thofma]

Later as in "I will add it to this PR later" or as in later later?

[fingolfin]

My plan was "Later as in 'this PR'". But it is actually difficult to add a "natural" test because all the det methods actually were fine, the problem was in det_fflu which is internal (and the calls to it all catch the problematic case first), and in det_interpolation (which is not used at all!).

The only thing I could think of was to add some direct tests for these. In the end I did it by adding them conformance test suite. That is perhaps not the ideal way, athe new tests are focused more on testing det_fflu etc. than the ring interfaces. Then again, this allows us to add general check for det...

The only risk I can think of is that this might break for some rings using the conformance tests. But that's something we can deal with.

[fingolfin]

For reference: this was bad as it returned zero.

[fingolfin]

... and this also returned zero

[lgoettgens]

Lgtm, but please wait until we have working downstream tests again

[lgoettgens]

It looks like some det methods in Nemo also have the same bug, so introducing the tests here breaks CI there. I would suggest to first fix the bug in Nemo, and then re-run CI here.

[fingolfin]

CI passes now. Shall we merge, @lgoettgens ?