Re: Trunk frozen starting tomorrow! No new features!

["Carlos O'Donell" <carlos at systemhalted dot org>]

On Thu, Jun 14, 2012 at 3:30 PM, Chris Metcalf <cmetcalf@tilera.com> wrote:
> On 6/14/2012 10:41 AM, Joseph S. Myers wrote:
>> If there are testsuite failures that are known and expected on your
>> architecture, please document them; information such as "No new testsuite
>> regressions" (currently listed for tilepro/tilegx) is less helpful than
>> having a list of what failures are known and not regressions.
>
> I updated the tilepro/tilegx section for the Release/2.16 wiki
> appropriately. ?I asked our gcc point person to let me know how to
> reference the remaining open issue with -fexception support and I'll update
> the wiki when I hear back from him.
>
> I have some unresolved questions from an email I posted earlier
> (http://sourceware.org/ml/libc-alpha/2012-05/msg01204.html). ?In a very
> vanilla build of the current glibc tip on x86_64 RHEL 6.2, using just
> "configure --prefix=/usr --enable-add-ons=nptl", I get failures in
> elf/tst-unique{3lib,4lib,3lib2}.so of the form:
>
> elf/tst-unique3lib.os:(.data.DW.ref.__gxx_personality_v0[DW.ref.__gxx_personality_v0]+0x0):
> ?undefined reference to `__gxx_personality_v0'
>
> Perhaps I'm missing something, but AJ reports no testsuite failures with
> newer components. ?RHEL 6.2 is kernel 2.6.32, gcc 4.4.6, binutils
> 2.20.51.0.2, and one of those might be responsible for the difference. ?If
> this is a well-known issue it would be nice to track it.
>
> My immediate selfish motivation is that I feel like I need to report this
> as a testsuite failure, but if it's really a generic issue (and it does
> look like it), I'd rather report it that way and not incriminate
> tilepro/tilegx any further :-)
>

+1

I also see the same tst-unique* failures on my RHEL 6.2 box.

It seems restricted to RHEL 6.2, since I don't see them anywhere else.

Cheers,
Carlos.