This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Move mode-switching to its own file
- From: Roger Sayle <roger at eyesopen dot com>
- To: Uros Bizjak <uros dot bizjak at kss-loka dot si>
- Cc: stevenb at suse dot de, <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 6 Jun 2005 09:34:48 -0600 (MDT)
- Subject: Re: [patch] Move mode-switching to its own file
Hi Uros,
On Mon, 6 Jun 2005, Uros Bizjak wrote:
> This bugfix in included in (currenly unreviewed) patch to improve mode
> switching: http://gcc.gnu.org/ml/gcc-patches/2005-06/msg00131.html
>
> (I have added Roger to CC as this is IMO a midlevel bug).
Steven's split of optimize_mode_switching to its own file did the
right thing, by not touching the existing implementation and only
refactored it into its own source file. For these types of changes,
its preferable to avoid functionality changes, which would make it
far harder to trace bugs when real changes are unnecessarily hidden
in a large patch.
I agree completely that there's a real bug in the code that was
moved, but unfortunately your "rounding mode switching improvements"
patch included the n_nexprs bug fix in the middle of a significant
change to the i386 backend. This pretty much limits the number of
possible reviewers to just RTH.
Please could you split out your one line big-fix, and bootstrap and
regression test it in the new mode-switching.c file? This middle-end
change on it's own is much easier to review (not only by me but the
other global maintainers too).
Many thanks in advance,
p.s. Thanks to stevenb for splitting up lcm.c, it needed to be done.
Roger
--