On Fri, 20 Mar 2020 at 13:08, Ben Craig via Lib <lib_at_[hidden]>
wrote:

> P0829 (the old omnibus paper) recommends making the integer overloads
> freestanding. I haven't made a smaller, split out paper for charconv yet.
>
> For the floating point overloads, there are two major challenges.
>
> First, is the floating point part. It is my (weakly held) opinion that
> floating point functions shouldn't be required in freestanding,


I'm not even sure the FP charconv functions should be required in a hosted
implementation :-P