Fix IsNilpotentGroup: infinite does not imply "not p-group"

[fingolfin]

This should be backported to 4.10.

[fingolfin]

@stevelinton does this seem sensible to you?

[hulpke]

This makes sense. Is it clear that nothing else assumes that IsPGroup implies finite?

[fingolfin]

Backported to 4.10 in 7388170

@hulpke It is "clear" in so far as some time ago, we (well, mostly we) audited the library and all deposited packages for uses of IsPGroup, and added explicit IsFinite checks when relevant. But I neglected to check for code which explicitly sets IsPGroup to false. When creating this PR, I check for other places doing it, but the on modified in this PR was the only I found. So, to the best of knowledge and ability, no public code is left that assumes IsPGroup implies finite; but of course I still may have missed some, and then there is non-public code... But on the upside, I am not yet aware of any code in GAP providing infinite p-groups, so until that happens, no problems should arise, giving us some more time for the transition.

[hulpke]

@fingolfin Excellent. Thanks.