On Tue, Dec 15, 2009 at 10:59:52AM +0000, Richard W.M. Jones wrote:
On Tue, Dec 15, 2009 at 05:50:05AM -0500, Josh Boyer wrote:
> On Tue, Dec 15, 2009 at 10:34:04AM +0000, Richard W.M. Jones wrote:
> >On Tue, Dec 15, 2009 at 10:06:42AM +0000, Tim Waugh wrote:
> >> On Tue, 2009-12-15 at 09:49 +0000, Richard W.M. Jones wrote:
> >> > Why not put everything in a single git repository?
> >>
> >> That would require every packager to check out the entire package set,
> >> all revisions, all branches. No thanks.
> >
> >Jesse can probably estimate for us how large this will be.
> >
> >I've found that git deals very well with large repositories that have
> >lots of files and lots of history (kernel, qemu). And you only ever
> >have to download it once, since you can use "git fetch" to make local
> >working copies.
>
> A full git repo was 5.7G. I sure as hell don't want to pull that down
> when I'm only interested in a few packages.
>
> (The CVS repo is 16G on the server side if you are wondering.)
Fair enough - it doesn't make sense since the combined repo would
be so large.
I was wondering if you could set up some meta repository, which had a
GIT sub-module for each package, but it seems sub-modules always have
to specify an explicit commit hash so they wouldn't seemlessly follow
changes.
Daniel
--
|: Red Hat, Engineering, London -o-
http://people.redhat.com/berrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org -o-
http://ovirt.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|