SimpleGroup: some fixes; better input validation; tests

[fingolfin]

• adjust case of error messages
• catch various invalid parameters classical groups, instead of
  returning non-simple groups with IsSimpleGroup forcibly set
• fix error message for 2E(n,q) with n<>6 (referred to 3D with n<>4)
• add tests

[codecov]

Codecov Report

Merging #2326 into master will increase coverage by 0.05%.
The diff coverage is 96.15%.

@@            Coverage Diff             @@
##           master    #2326      +/-   ##
==========================================
+ Coverage   72.52%   72.58%   +0.05%     
==========================================
  Files         478      478              
  Lines      246799   246812      +13     
==========================================
+ Hits       178996   179151     +155     
+ Misses      67803    67661     -142

Impacted Files	Coverage Δ
grp/simple.gi	77.39% <96.15%> (+16.77%)	⬆️
src/hpc/traverse.c	95.75% <0%> (-0.59%)	⬇️
src/hpc/threadapi.c	37.22% <0%> (-0.2%)	⬇️
src/funcs.c	80.19% <0%> (-0.14%)	⬇️
src/gap.c	80.29% <0%> (-0.09%)	⬇️
src/objset.c	81.71% <0%> (+0.22%)	⬆️
lib/grp.gi	85.81% <0%> (+0.34%)	⬆️

[hulpke]

I wouldn't have called it a bug, but yes, these are good additions.

[fingolfin]

hmm, so what do you call groups of order 1, 6 or 18 labelled by GAP as simple groups? :-). They lead to all kinds of interesting interactions subsequently.

[hulpke]

They are not simple, of course. If the name corresponds to a simple group an isomorphic group is returned. Otherwise the result is undefined :-)