This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: List of important PRs for 3.3.1, week -3.5
- From: Andrew Pinski <pinskia at physics dot uc dot edu>
- To: Dale Johannesen <dalej at apple dot com>
- Cc: Andrew Pinski <pinskia at physics dot uc dot edu>, Wolfgang Bangerth <bangerth at ices dot utexas dot edu>, gcc at gcc dot gnu dot org, mark at codesourcery dot com
- Date: Mon, 23 Jun 2003 21:14:58 -0400
- Subject: Re: List of important PRs for 3.3.1, week -3.5
On Monday, Jun 23, 2003, at 21:08 US/Eastern, Dale Johannesen wrote:
On Monday, June 23, 2003, at 05:50 PM, Andrew Pinski wrote:
Could you add 3187 as it is a regression from 2.95.3 and should have
been marked as a regression long time ago?
I think someone from apple sent in a patch to fix this but I could be
wrong.
It's only sort of a regresion; 2.95.3 did not follow the standard in
this area, and the extra functions
are a side effect of conformance. At least another entry point is
necessary.
Apple has code that changes one of the duplicate functions to be a
call to the other one, and
lets sibcall change this into a single branch insn. I don't recall
whether this was submitted,
and don't know whether such an approach would be welcomed. Certainly
multiple entry
points seem cleaner to me, but MacOS can't currently handle them.
It was submitted:
<http://gcc.gnu.org/ml/gcc-patches/2002-08/msg00354.html> and there is
support for multiple entry points according to Mark Mitchel in one of
the follow ups:
Note that Zack recently checked in better support for alternate entry
points. What we really want to do is use alternate entry points; that
will save even more space that your approach, and be faster as well.
Thanks,
Andrew Pinski