Problem with DTYPE ABI regarding long double vs. __float128

[FX]

> I don't know how you can deal with this at the moment, other than to
> generate specific instances of the library routines for __float128.
> 
> Of course, once the array descriptor business is done life will become
> much easier.

That's part of what I was asking. I mean, having partially-incorrect binary128 support for some time during 4.6 development, if it's fixed before branching, is OK. So, what is the current ETA for array descriptor rework?

Also, what are the changes currently planned? Would they cover this issue?

FX