Issue running git-svn-mirror script on code
Derek Atkins
warlord at MIT.EDU
Mon Jan 28 12:44:41 EST 2013
John Ralls <jralls at ceridwen.us> writes:
>>> I guess the best way would be svn->gitolite->github, so when we drop svn and start committing to gitolite directly we don't have to do anything.
>> That is how it is planned indeed. For completeness, I'll just point out that the push to github (or any other stand-in repo) is missing still and there's an additional intermediary repo involved for the svn->git sync. So currently we have this:
>> svn -> intermed repo -> gitolite
>
> Yes, the intermediate repo is a working repo while the gitolite one is
> a bare repo. AFAIK git-svn can't work directly with a bare repo, so it
> has to be that way.
Right. So right now we have:
svn -> git-svn -> gitolite
>>> Is that what's in place now (except to a stand-in repo on
>>> code.gnucash.org instead of github)?
>>
>> There is no stand-in repo. The current end point is gitolite.
>>
>> I never imagined a stand-in repo. In retrospect I understand now it
>> can be useful to test our commit hook. We can set up a hook in
>> gitolite and let it push to a temporary (local) repo, just to make
>> sure the hook works. Once it works, we only have to change the origin
>> config parameter in the gitolite repos to point at github instead of
>> the stand-in and we're done. Is that what you had in mind as well ?
>
> Yes, exactly.
Right. I didn't think about adding a local, stand-in repo upstream of
gitolite, but you're right that that would work to test our hooks.
Then, as you say, just point the url to github.
I guess I just need to add a new remote ("origin"?) to the repos that
point to, e.g., some local repos, and we can figure out where to put the
script(s).
-derek
--
Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
Member, MIT Student Information Processing Board (SIPB)
URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH
warlord at MIT.EDU PGP key available
More information about the gnucash-devel
mailing list