mercurial --> plain text --> mercurial

Matt Mackall mpm at selenic.com
Thu Mar 27 15:08:30 CDT 2008


On Thu, 2008-03-27 at 19:54 +0000, didier deshommes wrote:
> Matt Mackall <mpm <at> selenic.com> writes:
> 
> > 
> > Well you didn't expect it to do a checkout too, did you?
> 
> I wish:
> dfdeshom at sage:~/hg-test$ hg co
> resolving manifests
> 0 files updated, 0 files merged, 0 files removed, 0 files unresolved 

It doesn't attempt to checkout. But nonetheless, you have no changelog
data so it couldn't.

> > 
> > Did you dump the .i files in .hg? Or just the ones in .hg/data?
> 
> They went into the .hg/ directory when I ran the undump script.

Are you sure? The behaviour you're reporting is consistent with having a
missing or empty 00changelog.i (and probably no 00manifest.i).

Oh. I know what's wrong. New hg init is creating a dummy 00changelog.i
in the old location to make old clients expecting the old layout error
out. That's probably making a mess of your undump. I would have expected
undumping to give you a big fat error though.

> > BTW, modern hg uses a slightly different layout by default which
> > addresses some issues like filesystem case-folding. Which is why the
> > comments mention .hg/store/. Any time you do a clone over the wire or
> > with --pull, you'll get this new layout.
> > 
> 
> That must be it. I'll try that and report. Thanks

That'll probably fix it, yes.

-- 
Mathematics is the supreme nostalgia of our time.



More information about the Mercurial mailing list