g++ 2.95.2 on powerpc
Geoff Keating
geoffk@geoffk.org
Mon Nov 13 13:12:00 GMT 2000
Franz Sirl <Franz.Sirl-kernel@lauterbach.com> writes:
> On Thursday 27 April 2000 02:07, Geoff Keating wrote:
> > Anyway, I will check in a modified version of the testcase as
> > gcc.c-torture/execute/loop-5.c, since obviously this code doesn't get
> > tickled enough.
> >
> > 2000-04-26 Geoff Keating <geoffk@cygnus.com>
> >
> > * loop.c (basic_induction_var): If a REG is set from something
> > that is not a biv, then the REG is not a biv. Even if it is
> > earlier set from something that is a biv.
>
> Geoff,
>
> Bernd noticed me that I accidentally committed this patch with my checkin of
>
> 2000-09-29 Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
>
> * loop.c (check_final_value): A GIV is not replaceable if used
> before set.
>
> Any reason why your patch was never committed? Do we still need it? If the
> patch is still OK, I'll just add your ChangeLog entry, otherwise I will
> revert my accidental checkin.
Since loop-5 seemed to pass before, I think the patch may not be
needed. There was substantial work fixing things like this recently.
If you happen to know that the patch does fix a bug, then leave it in;
otherwise take it out.
--
- Geoffrey Keating <geoffk@geoffk.org>
More information about the Gcc-patches
mailing list