1st draft: PRs fixed in 3.1.1

[Joe Buck]

Gaby writes:
> I'm having trouble to qualify a hole that got fixed as a regression.  

I can see why, and I didn't mean to offend anyone.  But since there is
developer interest in supporting multiple allocation models without
requiring a library rebuild (see Phil's message), it should be clear
that the current situation is not ideal.

I will add something to the writeup explaining the issue.