Increase warning level for Conway polynomial

[ChrisJefferson]

This just increases the warning level of calculating conway polynomials. I don't feel this message is necessary for most users, and it is causing problems with some new tests I'm writing (and annoys me generally)

[codecov]

Codecov Report

Merging #1363 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1363      +/-   ##
==========================================
+ Coverage   61.77%   61.77%   +<.01%     
==========================================
  Files        1035     1035              
  Lines      356668   356668              
  Branches    14292    14291       -1     
==========================================
+ Hits       220335   220338       +3     
+ Misses     132671   132670       -1     
+ Partials     3662     3660       -2

Impacted Files	Coverage Δ
hpcgap/lib/polyconw.gi	53.79% <100%> (ø)	⬆️
lib/polyconw.gi	77.14% <100%> (ø)	⬆️
src/hpc/traverse.c	79.45% <0%> (-0.39%)	⬇️
src/funcs.c	63.78% <0%> (-0.28%)	⬇️
src/hpc/threadapi.c	30.55% <0%> (+0.09%)	⬆️
src/objset.c	39.88% <0%> (+1.12%)	⬆️
src/hpc/tls.h	67.5% <0%> (+2.5%)	⬆️

[fingolfin]

For me, this warning is in so far useful as it usually tells me I need to try something a bit different, as I am leaving the "fast path".

That said, I certainly will survive if this is changed. shrug

Perhaps @hulpke has a strong opinion on this, though? (given that this also affects e.g. matrix groups)

[ChrisJefferson]

My feeling is there any many, many algorithms in GAP which are "slow", we could warn whenever partition backtrack isn't performing well (or is invoked at all), when we have to verify a stabilizer chain, etc. etc.

Of course, there might be some deeper reason that the conway polynomials are different -- in which case we should at least attach some kind of help message, to tell me (and others) why we should care :)

[fingolfin]

Well, as I said, I am fine with increasing the level here anyway :-)

[hulpke]

As there is a higher level warning if the computation is really hard (e.g. for GF(1013^10)) I cannot see a need for warning about these basic calculations.

[olexandr-konovalov]

Maybe hard cases could be indicated with InfoPerformance, introduced not so far and still rarely used.