Builtin-clone?

Builtin-clone?

[Daniel Barkalow]

I remember some discussion of what was still needed in order to make clone 
a trivial caller of init, config, fetch, and (optionally) checkout, and 
that there was still something. But has anyone actually written the bulk 
of builtin-clone.c? 

	-Daniel
*This .sig left intentionally blank*

Re: Builtin-clone?

[Johan Herland]

On Tuesday 19 February 2008, Daniel Barkalow wrote:
> I remember some discussion of what was still needed in order to make clone 
> a trivial caller of init, config, fetch, and (optionally) checkout, and 
> that there was still something. But has anyone actually written the bulk 
> of builtin-clone.c?

I remember Kristian Høgsberg posted some WIP back in December:
http://thread.gmane.org/gmane.comp.version-control.git/67915/

I recently sent him (and the list) an email asking about his progress (and - 
once his work arrives - whether it's a good idea to make 'git clone' 
produce packed refs), but I have yet to hear from him...


Have fun! :)

...Johan

-- 
Johan Herland, <johan@herland.net>
www.herland.net

Re: Builtin-clone?

[Kristian Høgsberg]

On Tue, 2008-02-19 at 08:08 +0100, Johan Herland wrote:
> On Tuesday 19 February 2008, Daniel Barkalow wrote:
> > I remember some discussion of what was still needed in order to make clone 
> > a trivial caller of init, config, fetch, and (optionally) checkout, and 
> > that there was still something. But has anyone actually written the bulk 
> > of builtin-clone.c?
> 
> I remember Kristian Høgsberg posted some WIP back in December:
> http://thread.gmane.org/gmane.comp.version-control.git/67915/

What Johan said :)  I posted the patch in progress during the 1.5.4
freeze, and threatened to pick it up again in January after 1.5.4 was
released.  I haven't worked on it since, unfortunately.  It does most of
the option parsing and the local clone optimizations, but doesn't use
fetch correctly (as Daniel points out in the thread above).

Daniel, I'm pretty sure you have a better understanding of the fetch and
checkout side of clone, so if you want to take the patch and finish
those parts, I'm fine with that.

> I recently sent him (and the list) an email asking about his progress (and - 
> once his work arrives - whether it's a good idea to make 'git clone' 
> produce packed refs), but I have yet to hear from him...

I think packed refs in clone sounds like a good idea, but my primary
goal was really to just port it to C and avoid changes in behavior
during that work.

cheers,
Kristian