Improve large-prime SL classical recognition

[fingolfin]

Strengthen the no-classical-forms check by using the current random element's trace and characteristic-polynomial coefficients to rule out dual and Frobenius forms directly.

Add a seeded regression for RecogniseClassical(SL(6,257)) so the previous unknown result is covered by the quick bugfix tests.

Co-authored-by: Codex codex@openai.com

Resolves #492

---

Before this PR, I get this (of course the numbers are randomized):

gap> Collected(List([1..100], i->RecogniseClassical(SL(6,257)).isSLContained));
[ [ true, 48 ], [ "unknown", 52 ] ]
gap> time;
8425

After this PR:

gap> Collected(List([1..100], i->RecogniseClassical(SL(6,257)).isSLContained));
[ [ true, 100 ] ]
gap> time;
518

---

Disclaimer: up to minor tweaking and prompting, this was entirely done by the AI. I am not yet sure this is the right way to "fit it in".

It would be great to discuss this with @aniemeyer to determine if we should merge this; or if not, how else we should handle this.

[codecov]

Codecov Report

❌ Patch coverage is 87.23404% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 88.05%. Comparing base (f792cef) to head (008c261).

Files with missing lines	Patch %	Lines
gap/matrix/classical.gi	87.23%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #496      +/-   ##
==========================================
- Coverage   88.46%   88.05%   -0.42%     
==========================================
  Files          45       45              
  Lines       19019    19066      +47     
==========================================
- Hits        16825    16788      -37     
- Misses       2194     2278      +84     

Files with missing lines	Coverage Δ
gap/matrix/classical.gi	84.25% <87.23%> (-0.62%)	⬇️

... and 8 files with indirect coverage changes

🚀 New features to boost your workflow:

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