Re: Runaway darcs process
Em Qua, 2009-09-23 às 17:17 +0200, Joachim Breitner escreveu:
> Am Mittwoch, den 23.09.2009, 16:14 +0100 schrieb Iain Lane:
> > On Wed, Sep 23, 2009 at 05:04:12PM +0200, Joachim Breitner wrote:
> > >Hi,
> > >
> > >you have a runaway darcs process on alioth:
> > >226351 24670 90.7 0.0 99700 7996 ? Rsl Sep22 1279:08 darcs apply --all --repodir /darcs/pkg-haskell/haskell-cabal-install
> > >
> > >please run "killall -KILL darcs" on alioth.
> > Thanks, I've done this. This seems to happen about 75% of the time I
> > push. I usually notice and kill the process but I guess it slipped by
> > this time. Do we have any idea what causes it? Perhaps we can do a
> > stable update of darcs to get it fixed.
> Yes, it’s a know bug caused by ghc-6.8. So it’s actually not the darcs
> version that matters, but the compiler version it is compiled with. The
> “proper” solution is to get ghc-6.10 into backports.org, which I think
> Marco attempted. I’m not sure what the state is, maybe he wanted to wait
> for 6.12 to be released as it will be easer to bootstrap.
I have searched a little bit about this issue, as is described in:
and I guess the better solution is to upload a patched GHC 6.8.2 to
lenny. I asked about how to do this in:
but as no one replied and I wasn't sure about the direction to go, I
stoped thinking about it for a time. As it's said on these e-mails, I
have the package here ready for going to lenny, needing only some
adjustments in debian/changelog.
Do you think we should do this upload to stable?