What is up with glibc in rawhide?

[Tom Lane]

Jesse Keating <jkeating at redhat.com> writes:
> On Thu, 2009-04-09 at 13:31 -0400, Tom Lane wrote:
>> And mysql, and perhaps rpm according to Panu, and we know not what else
>> that has not got thorough regression tests executed as part of its build.

> Those aren't broken themselves, instead when used with glibc.i586 you
> get memory corruption from memove which can manifest itself in many
> ways.  These packages won't need to be rebuilt at all, they'll start
> working again with a fixed glibc.i586.

You are assuming that the broken glibc hasn't caused any undetected
errors during building of packages.  Given that it can have odd effects
on the execution of just about any tool used during a build, I find
your optimism 100% unjustified.

Even without that consideration, you are misinformed if you think that
only i586 is broken.  The mysql failures I was seeing manifested on
x86_64 but not the other three platforms.

			regards, tom lane