This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: simple minded mapping on darwin?
Andrew,
In case my last post was a bit unclear, what I found
interesting from the build where I used both the hack
in gcc/fortran/f95-lang.c as well as your patch...
patched sym = expl$LDBL128$LDBL128
patched sym = fmodl$LDBL128
Here only the math functions touched by the DO_DEFINE_MATH_BUILTIN
macro have a second $LDBL128 appended. So we have the following
results for the combination of patches...
1) your patch => darwin_patch_builtin() shows appended $LDBL128 but actual symbols in binary don't
2) my hack => appended $LDBL128 to symbols in the binaries for only long double math calls.
3) both your patch and my hack => darwin_patch_builtin() shows doubled suffix on long double math routine symbols
but only single suffix for the same symbols in binaries
These results make me wonder if somehow the darwin_patch_builtin() calls are being made too late
in the current compiler. Otherwise I don't quite understand how the appended suffix from f95-lang.c survives
to be used and the second one, made in darwin_patch_builtin(), gets ignored.
Jack