cg update failing

I'm getting this on my clone of linus' tree:

,----
| cg-merge: merge blocked: seeked from master
`----

I've not found a way past it.

Is there an easy way, or do I have still yet another borked lk clone?

(Borked clones would be much less of an issue if I had more bandwidth than disk space, but at the moment I'm stuck behind a straw. Cloning something like lk takes hours.)

-JimC

On Thu, 2005-07-14 at 14:04 -0400, James Cloos wrote:

cg-seek master ?

-- Darrin

JimC> cg-merge: merge blocked: seeked from master `----

Darrin> cg-seek master ?

:; cg seek master On commit 514fd7fd01d378a7b5584c657d9807fc28f22079

-JimC

On Thu, 2005-07-14 at 16:06 -0400, James Cloos wrote:

Now your merge should not be blocked. No?

-- Darrin

Darrin> Now your merge should not be blocked. No?

I figured out what you meant after I sent that first reply, but waited for update to finish to be sure it worked before replying again.

It took about thirty minutes, and claimed that I had local changes (it was a pristine clone) that it needed to bring forward, but it seems OK.

Now I just need to remember how to get the checked-out copy updated from the 2.6.12 tag to current. ;)

(I'd been using hg, but www.kernel.org/hg is fubar, so I had to go back to my cg clone.)

-JimC

Well, it wasn't as Ok as I first thought. There were several .rej and backup files as left behind by patch(1). cg update HEAD says: Branch already fully merged but Makefile still says 2.6.12.

I'm cloning now to a remote uml and will try and rsync from there to the laptop. Perhaps *that* will get me a working repo.

[SIGH]

-JimC

Dear diary, on Fri, Jul 15, 2005 at 06:12:48AM CEST, I got a letter where James Cloos <cloos+vger-git@jhcloos.com> told me that...

Huh. Well, you had to do uberweird things with your repository - why did you cg-seek at all? What Cogito version is it?

Anyway, cg-reset (or cg-cancel in "older" Cogito) should bring everything in shape.

Petr> Huh. Well, you had to do uberweird things with your repository - Petr> why did you cg-seek at all? What Cogito version is it?

All I ever did was track and compile linus’ tree. IIRC, when the first new tag was added after I cloned the tree I had to seek to that tag to see the new versions of the files. But obviously that was long enough before hg’s repo broke that I’d forgotten how I’d done that....

I forget what version I started with. I’ve been tracking gentoo’s ebuild since it entered portage.

-JimC

Dear diary, on Wed, Jul 20, 2005 at 06:34:43AM CEST, I got a letter where James Cloos <cloos+vger-git@jhcloos.com> told me that...

Ah so. Well, yes, the first tag was IIRC weird since it was a tree tag, not a commit tag. That confused things, but it shouldn't be necessary anymore since all new tags are hopefully commits. That might somehow create the local modifications as well, although I don't have a clear idea how could be that possible. Well, there were some bugs like this in Cogito in the past, but that should be long fixed.

Thanks anyway,