On Tue, Oct 18, 2011 at 7:50 AM, Alexandre Oliva <aol...@redhat.com> wrote:
> On Oct 17, 2011, Jakub Jelinek <ja...@redhat.com> wrote:
>
>> On Tue, Oct 11, 2011 at 04:19:52PM -0300, Alexandre Oliva wrote:
>>> Here's what I've got so far. Regstrapped on x86_64-linux-gnu and
>>> i686-linux-gnu. Ok to install?
>
>> I see
>> +FAIL: gcc.c-torture/compile/pr19080.c -O3 -g (internal compiler error)
>> +FAIL: gcc.c-torture/compile/pr19080.c -O3 -g (test for excess errors)
>> regression with this patch on x86_64-linux (--enable-checking=yes,rtl), ICE
>> in var-tracking. Can you please look at that?
>
> Hey, I'd already fixed that! It turns out that I posted an outdated
> version of the patch, from before moving an incorrect assertion check
> into a test.
>
>> Just a style nit, isn't { supposed to go on the next line?
>
> Not sure. Nearly all multi-line enums in gcc/ have brackets at the end
> of the line rather than in the beginning of the subsequent line; I could
> only find one that followed what we both think the conventions require.
> I changed it, thanks.
>
>> Otherwise looks ok.
>
> Thanks, here's what I'm going to install as soon as I complete a new
> regression test with the corrected patch, just to be sure I didn't
> accidentally drop any other fixes.
>
This caused:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50799
--
H.J.
