Classical natural tests

[codecov[bot]]

Codecov Report

Merging #200 (425df87) into master (4796d16) will increase coverage by 0.08%. The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #200      +/-   ##
==========================================
+ Coverage   76.37%   76.45%   +0.08%     
==========================================
  Files          37       37              
  Lines       17199    17193       -6     
==========================================
+ Hits        13135    13145      +10     
+ Misses       4064     4048      -16     

Impacted Files	Coverage Δ
gap/matrix/matimpr.gi	62.74% <0.00%> (-1.97%)	:arrow_down:
gap/projective/findnormal.gi	78.96% <0.00%> (-0.92%)	:arrow_down:
gap/matrix/ppd.gi	93.16% <0.00%> (-0.86%)	:arrow_down:
gap/projective/d247.gi	75.24% <0.00%> (-0.66%)	:arrow_down:
gap/matrix/classical.gi	55.49% <0.00%> (-0.55%)	:arrow_down:
gap/projective/c6.gi	61.58% <0.00%> (-0.19%)	:arrow_down:
gap/matrix.gi	91.30% <0.00%> (-0.02%)	:arrow_down:
gap/base/recognition.gd	100.00% <0.00%> (ø)
gap/base/recognition.gi	73.93% <0.00%> (+1.28%)	:arrow_up:
gap/perm/giant.gi	90.33% <0.00%> (+2.02%)	:arrow_up:
... and 1 more

[ssiccha]

Unfortunately on my machine, some of the smaller GLs fail. relatively regularly. I'll try to find out which ones that are. I have seen that there already are related issues, so I'll look there too.

[fingolfin]

We now have tst/working/quick/ClassicalNaturalNaming.tst. Might be better to extend that instead. Perhaps also get rid of tst/testclassicalnatural.g; for that, we maybe should...

• make the number of repetitions in TestNaming an argument??
• teach TestNaming to use randomize generating sets (but really, even without that we already find plenty bugs)
• add more input tests; inspired by tst/testclassicalnatural.g and this PR?