Nathanial Sloss <nathanialsl...@yahoo.com.au> writes: > Hi > > There are issues with m68k softfloat (fpgnulib.c) not perserving NaNs and Inf > when converting from (long) double and float types. > > This issue has been documented in a NetBSD PR: > > http://gnats.netbsd.org/59616 > > Addressed by the attached patch which sets the value correctly for NaN and > +/- > Infinity when converting between different types. >
Please see https://gcc.gnu.org/contribute.html for how to submit patches. In addition to the other guidance there, please make sure it's based correctly on the trunk sources so it applies cleanly. > > Best regards, > > Nat > > [2. text/x-patch; m68k-soffloat-ints-nans-v5-gcc.diff]...
