This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Regression for gcc.dg/Wstrict-overflow-11.c between 124216:124219
- From: "Richard Guenther" <richard dot guenther at gmail dot com>
- To: "Hans-Peter Nilsson" <hans-peter dot nilsson at axis dot com>
- Cc: rguenther at suse dot de, gcc-patches at gcc dot gnu dot org, "Ian Lance Taylor" <iant at google dot com>
- Date: Fri, 27 Apr 2007 20:22:46 +0200
- Subject: Re: Regression for gcc.dg/Wstrict-overflow-11.c between 124216:124219
- References: <200704271744.l3RHidZ7024972@ignucius.se.axis.com>
On 4/27/07, Hans-Peter Nilsson <hans-peter.nilsson@axis.com> wrote:
Something in your
+ PR tree-optimization/30965
+ PR tree-optimization/30978
fix caused a regression for gcc.dg/Wstrict-overflow-11.c (for
cris-elf, but that shouldn't matter):
Well, loop header copying introduces a comparison that is always true:
<bb 2>:
D.1630_13 = i_4(D) + 4;
if (i_4(D) <= D.1630_13)
goto <bb 3> (<L0>);
else
goto <bb 4> (<L2>);
and before this patch we were not able to fold this. Now we do and
get this warning.
Which makes me think again that all "compiler generated" code should
have TREE_NO_WARNING set. Or, Ian, any idea on how to fix _this_
warning but not warnings from forwprop in general?
Richard.