Also fixing the abi issues that started this, and probably making an argument for why it makes sense for all of cpython's built-in float operations to check the fpu flags, and to do so using a weird longjmp-based mechanism that only some platforms support. The fact that it's disabled by default and has been broken for a decade+ without anyone noticing might be working against you here...

You might get the impression that I think this is a bad idea. I do :-). But I am genuinely trying to helpful; I'm sure people would be willing to listen to an argument, and if you want to make one then those are the sorts of issues you're likely to need some answer for.