https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119241
--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> --- The *power_of_ten stuff is purely integral, so in the PR119242 patch I've used FIXED_WIDE_INT(128) type for it. So it is just about the case when that needs to be converted to floating point, and we have real_from_integer for that.
