Oh, and ruby1.9.1 (was Re: For someone bored: fix glib2.0 and apr)
---------- Forwarded message ----------
From: Antonio Terceiro <terceiro@debian.org>
Message-ID: <20130518152806.GA32300@debian.org>
To: Thorsten Glaser <tg@mirbsd.de>
Cc: lucas@debian.org, daigo@debian.org, akira@debian.org
Date: Sat, 18 May 2013 12:28:06 -0300
Subject: Re: Log for attempted build of ruby1.9.1_1.9.3.194-8.1 on m68k (dist=unstable)
On Sat, May 18, 2013 at 01:03:23AM +0000, Thorsten Glaser wrote:
> Antonio Terceiro dixit:
>
> >I don't like the idea of disabling tests blindly ... do you have a link
> >to build logs so I can check which tests fail?
>
> Sure: http://buildd.debian-ports.org/status/package.php?p=ruby1.9.1
>
> On the regular Debian buildd logs page, the “[6]b.d-ports.o” link
> also points there.
right. I didn't remember that, sorry.
> Note that x32 seems to have different problems; most likely, it’s
> misdetected as amd64, i386 or both of them. But for the other
> minority architectures, a small subset of tests fail.
>
> I’ve manually built ruby1.9.1 with disabled testsuite for now,
> due to the libffi transition, and will porter-upload that, but
> if we can cooperate trying to fix it, sure. There’s no porterbox,
> but https://wiki.debian.org/Aranym/Quick shows how to set up
> a VM quickly.
I think fixing that kind of failure is outside of my area of expertise,
but you can send a patch disabling the test suite on the architectures
you are interested in and we can carry that.
The only problem is that once we disable the test suite, the probability
of us re-enabling them in the future is very low. The architectures in
which we currently have the test suite disabled are already lost cases
to me. :-(
--
Antonio Terceiro <terceiro@debian.org>
[ Part 2, "Digital signature" Application/PGP-SIGNATURE ]
[ 205bytes. ]
[ Unable to print this part. ]
Reply to: