Avoid some integer factoring; use PrimeDivisors

[fingolfin]

Also remove the commented out YndexSymmetricGroup (it contained a call to FactorsInt), and replace Length(Filtered(...)) by Number(...) (motivation: such code was close to code calling FactorsInt).

This is motivated by PR #2087. Note that PrimeDivisors currently calls FactorsInt, but now this is in a single place.

[hulpke]

Very good. The PrimeDivisors was on my to do list, even better its done.

The weird Yndex (which was written by Heiko Theissen) routine predates the quick recognition of S_n. Index would have called Size which would have called a stabilizer chain , which for a large S_n is cumbersome. There is no need for it any longer.

[codecov]

Codecov Report

Merging #2094 into master will increase coverage by 0.05%.
The diff coverage is 68.36%.

@@            Coverage Diff             @@
##           master    #2094      +/-   ##
==========================================
+ Coverage   69.27%   69.32%   +0.05%     
==========================================
  Files         491      491              
  Lines      257326   258098     +772     
==========================================
+ Hits       178269   178937     +668     
- Misses      79057    79161     +104

Impacted Files	Coverage Δ
lib/stbcbckt.gi	84.41% <ø> (+0.05%)	⬆️
lib/autsr.gi	0% <0%> (ø)	⬆️
lib/randiso.gi	26.11% <0%> (ø)	⬆️
lib/upoly.gi	57.84% <0%> (ø)	⬆️
lib/ctblfuns.gi	60.21% <0%> (ø)	⬆️
lib/gprd.gi	49.62% <0%> (ø)	⬆️
lib/schur.gi	19.33% <0%> (ø)	⬆️
lib/grppcaut.gi	44.51% <0%> (+0.23%)	⬆️
lib/gaussian.gi	69.31% <100%> (ø)	⬆️
lib/polyfinf.gi	86.63% <100%> (ø)	⬆️
... and 101 more

[olexandr-konovalov]

@frankluebeck said in #2087:

A call of Factors(Integers, x) instead of just Factors(x) can give a significant speedup if this is called frequently with small x.

Should this PR use Factors(Integers, x) instead of Factors(x) where possible then?

[fingolfin]

This PR does not use Factors, so: no.

[stevelinton]

@alex-konovalov @fingolfin

Should this PR use Factors(Integers, x) instead of Factors(x) where possible then?

This PR does not use Factors, so: no.

Aside from this, admittedly pertinent, point though, the answer is yes. If you know the ring you want to factor over, it is always better to provide it. Apart from performance concerns, GAP will sometimes not default to the ring you expect.