[patch] Enable fmove-loop-invariants by default
Joseph S. Myers
joseph@codesourcery.com
Sun Feb 26 22:57:00 GMT 2006
On Sun, 26 Feb 2006, Mark Mitchell wrote:
> Richard Guenther wrote:
>
> > We should at least deprecate them (and maybe warn, if not -Werror) so we
> > can get rid of them in 4.3.
>
> I think we should probably accept them forever more. As Daniel says,
> there is a ton of code out there passing -fno-strength-reduce, and it
> does us no harm just to keep accepting those options.
Indeed, these are options with no intended effect on code semantics or
diagnostics, so accepting forever is harmless. It's options with semantic
effect (e.g. enabling non-standard extensions we've now removed) that
should give errors for being no longer valid. The code using
-fno-strength-reduce is probably using it because of superstition about an
old GCC 2.7 bug (see <http://gcc.gnu.org/ml/gcc/1997-09/msg00920.html> for
the testcase), not for any semantic effect other than "avoid this old
bug".
--
Joseph S. Myers http://www.srcf.ucam.org/~jsm28/gcc/
jsm@polyomino.org.uk (personal mail)
joseph@codesourcery.com (CodeSourcery mail)
jsm28@gcc.gnu.org (Bugzilla assignments and CCs)
More information about the Gcc-patches
mailing list