This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC] Quad-float support in Fortran
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: FX <fxcoudert at gmail dot com>
- Cc: gcc-patches at gcc dot gnu dot org, gfortran List <fortran at gcc dot gnu dot org>, Jakub Jelinek <jakub at redhat dot com>, Michael Meissner <meissner at linux dot vnet dot ibm dot com>
- Date: Thu, 28 Oct 2010 11:52:08 +0000 (UTC)
- Subject: Re: [RFC] Quad-float support in Fortran
- References: <E8FAAF3E-9E15-461F-BBCF-CD49DC41DEFB@gmail.com>
On Wed, 27 Oct 2010, FX wrote:
> -- The patch breaks -static-libgfortran, which needs to be
> reimplemented inside the spec file (probably a one-liner)
I don't know what you mean by breaks, but as per yesterday's IRC
discussion -static-libgfortran:
* should *not* link with static libquad, since that's LGPL rather than
GPL+exception;
* should avoid a shared libquad dependency unless the program actually
uses the __float128 facilities.
> -- Namespace separation in libquad (math symbols will be exported as,
> e.g., acoshq, but the rest should be liquad_-prefixed)
The shared library should also use symbol versioning with an explicit map
showing the exact set of exported symbols. Does it?
Also, I think a name such as libquadmath was preferred over just libquad
(for the library at least, if not for the symbols).
> +# Create a spec file, so that compile/link tests don't fail
> +test -f libgcj.spec || touch libgcj.spec
libgfortran.spec? Or maybe this code isn't needed at all, if things
worked with the wrong name used?
--
Joseph S. Myers
joseph@codesourcery.com