This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: iconvdata regression
- From: "Carlos O'Donell" <carlos at systemhalted dot org>
- To: Andreas Jaeger <aj at suse dot com>
- Cc: libc-alpha at sourceware dot org
- Date: Mon, 2 Apr 2012 08:46:32 -0400
- Subject: Re: iconvdata regression
- References: <20120331.053437.2170377486919118624.davem@redhat.com><20120331.173213.411657149270174548.davem@davemloft.net><CADZpyizNRosqEHf_Nn_obD+SV9gHyweSkkxE_ViMNFsTpB3hVg@mail.gmail.com><201204021238.48868.aj@suse.com>
On Mon, Apr 2, 2012 at 6:38 AM, Andreas Jaeger <aj@suse.com> wrote:
>> Do we agree that build breakage is *much* more serious since it
>> impacts reg testing?
>
> Yes, I do.
>
> IMO giving a grace period for fixing broken stuff is fine and we should
> record what the grace period is - dependend on whether it's a build failure
> or a testsuite one. If somebody is actively working on it promptly, I would
> not revert for a testsuite failure. I'm also fine with reverting patches.
>
> Btw. I commited for Tulin and did not test all of glibc directly (just run
> the testsuite for wcsmbs) and commented that it's broken as well in the
> thread.
>
> I can revert the patch if the testsuite fix takes longer,
Thanks for your input Andreas.
I'm fine giving upwards of a week to fix a testsuite failure.
I want 24h revert of unacknowledged build breakage, or at most 48h
revert of acknowledged build breakage.
Does anyone think this is fair?
What does GCC do here?
Cheers,
Carlos.