On Thu, Feb 25, 2021 at 02:23:01AM +0000, Joseph Myers wrote: > On Wed, 24 Feb 2021, Segher Boessenkool wrote: > > > The other option, which we need for correctness *anyway*, is to have > > libgcc do the conversion without using any specific libc. > > Which ends up being quite complicated (see e.g. > https://www.cl.cam.ac.uk/~jrh13/papers/decimal.pdf for a more detailed > discussion of how to do such conversions and how to determine the worst > cases for correct rounding of them).
It may be tricky to get right, but since we need to have it, it is not like there is much choice but to implement it. Segher
