This is the mail archive of the
mailing list for the glibc project.
Re: still building soft-fp on ppclinux
- From: Jack Howarth <howarth at bromo dot msbb dot uc dot edu>
- To: Ulrich Drepper <drepper at redhat dot com>
- Cc: libc-alpha at sources dot redhat dot com
- Date: Sat, 19 Oct 2002 16:14:27 -0400 (EDT)
- Subject: Re: still building soft-fp on ppclinux
- References: <200210191832.g9JIWtn06069@magilla.sf.frob.com>
Is there a convention for how a sysdep/<arch>/fpu interacts
with a sysdeps/<arch>/fpu? The problem appears to be that
Aldy's patches introduce duplication of files in both of these
as well as a soft-fp specific libm-test-ulps to test against.
If soft-fp is to be built in exclusion of the normal fpu code,
one would think configure.in needs to be re-crafted to insure
that sysdep/<arch>/fpu doesn't appear on the config-sysdirs
list in config.make.
What if we add something to configure.in that (for ppc)
sed's config-sysdirs removing sysdeps/powerpc/fpu if
completely_soft is set otherwise it removes
sysdeps/powerpc/soft-fp if completely_soft is unset?
If Aldy is going as far as having a soft-fp specific
libm-test-ulps, perhaps those two directories should be
detangled so that they are completely exclusive of each