dulwich-users team mailing list archive

[Augie Fackler <durin42@xxxxxxxxx>]

On May 19, 2010, at 7:59 PM, Jelmer Vernooij wrote:

> On Wed, 2010-05-19 at 19:02 -0500, Augie Fackler wrote:
> > On May 19, 2010, at 6:46 PM, Jelmer Vernooij wrote:
> > > On Wed, 2010-05-19 at 18:27 -0500, Augie Fackler wrote:
> > > > I noticed you pushed only one of my most recent set of changes - is
> > > > there something wrong with the other changesets? If so, can we do
> > > > some
> > > > kind of code review on them so I can fix them up, either by
> > > > installing
> > > > gerrit somewhere or just doing patchbombs on the dulwich-users  
> > > > list?
> > > > Either way, I went ahead and rebased my remaining changes on the  
> > > > one
> > > > you cherry picked - they're on github in the usual place.
> > > I haven't had time to have a look at the other changesets yet, but
> > > this
> > > one was trivial enough to cherrypick.
> > >
> > > I'll look at the others later this week but will need more time for
> > > that, possibly after 0.6.0.
> > I'd really like those cleanups (and the deprecation, so it can come
> > out sooner rather than later) in 0.6.0 rather than after if possible
> > since there's a protocol violation fix in there.
> I'm happy to land the protocol compliancy bits

Please do, it causes errors on push if nothing needed to be pushed  
(6c07a270e99122e722f5a1e56289596156d3a2d4, but you'll need to remove  
the test change since that assumes my other patches, I can split the  
change if you'd like).

> but the GitClient
> refactoring needs more discussion and I don't want to postpone the
> release further because of that.

(disclaimer: this feels like it comes across as pushy - that's not my  
intent - I merely want to try and have a shot at fixing these for the  
release if that's at all possible)
I'm open to suggestions on the code, the only thing I wasn't happy  
with was all protocol objects taking a can_read function, but it was  
the only sane way I could find to make things work. If you can look at  
that patch (bf9d8a90601d2aad0dfc57f0860196d5589e9718) and provide  
feedback, I'll be happy to see what I can do about fixing it in short  
order. I don't think the other ones should be terribly controversial  
(as far as I can see, anyway), they're mostly just organizing and  
cleaning up the client code.

Out of curiosity, why the hurry to get 0.6 out?

> Cheers,
>
> Jelmer

Thanks,
Augie