Mercurial Linux kernel repository

[Matt Mackall]

On Tue, Feb 06, 2007 at 04:10:11PM -0800, Inaky Perez-Gonzalez wrote:
> On Tuesday 06 February 2007 14:44, Matt Mackall wrote:
> > On Tue, Feb 06, 2007 at 02:37:37PM -0800, Iñaky Pérez-González wrote:
> > > Hi All
> > >
> > > Sorry for the semi-off topic, but:
> > >
> > > Does anyone have an idea of why the mercurial kernel repo
> > > on kernel.org/hg/linux-2.6 was chopped off? History only
> > > goes now back to 2.6.20-rc7 (give or take) and all the
> > > previous history was lost (as well, can't pull and merge into
> > > repos cloned before the change).
> >
> > I still see >46000 commits here:
> >
> > http://www.kernel.org/hg/linux-2.6/
> 
> Huh, so I got it wrong there. Then I wonder why I can't pull 
> into my previously cloned repo (my first thought was a chop had 
> happened) I got:
> 
> $ hg id
> 757c4b54b4e1
> $ hg pull -u
> pulling from http://kernel.org/hg/linux-2.6
> searching for changes
> adding changesets
> adding manifests
> adding file changes
> abort: unknown parent 1e370d8153d5!
> transaction abort!
> rollback completed
>
> After this I ran a verify which found 18 integrity errors
> 
> $ hg verify
> checking changesets
> checking manifests
> crosschecking files in changesets and manifests
> checking files
> node 1e370d8153d5cfa38bdd947a59d4cf51c889a790 in manifests not in drivers/cpufreq/cpufreq.c
> node afc12ddc5499edb11f8821c7b2008b19ef4b704c in manifests not in drivers/hwmon/ams/Makefile

This is a known bug. We're not sure what's causing this problem, but
it appears to be related to the automatic rsync between master.kernel.org and
www.kernel.org occurring during a pull, which breaks various locking
rules. The resulting pull ends up missing some file revisions.

Unfortunately, I don't have any control over the kernel.org rsync
process and very little debugging capability either. I'll try another
hack to work around the problem..

> I retried the pull and got the same error

Yep. Your local repo is corrupt. If you do a local clone -r [some
revision close to tip], you should be able to get rid of the corrupted
bits and pull again. Once you've done that, you can pull any
changesets you added over to your fixed repo.

-- 
Mathematics is the supreme nostalgia of our time.