On Mar 30, 2009, at 2:20 PM, Martin Atkins wrote:
Will Norris wrote:
A few of us in the community (myself, Chris Messina, David
Recordon, Brian Ellin, et al) have been working to get the OpenID
project on Google Code setup. Right now it is in a reserved status
to ensure that no one other than the OpenID Foundation gets ahold
of it. I think everything is ready to go at this point, we just
need the green light from the board. Our immediate plans for the
Google Code project are:
- house the source code for the openid.net website in a version
control system, both for transparency and security
- as Brian mentioned in this post[0] to the new OpenID Code
mailing list, JanRain has committed to moving their libraries
(currently hosted at openidenabled.com) to this new project space
so that we have a single home for all of the development.
- I don't actually recall if we discussed moving the content
currently housed at svn.openid.net to the new project space. I
assume that we will at some point, but we can certainly wait on
that if there is objection or concern about that.
Initially, we'd like to have the four of us (myself, Chris, David,
and Brian) set as admins for the Google Code project at least while
we get things setup. This can certainly be adjusted down the road
as necessary.
Are there any objections to moving forward with this?
I don't really like the idea of having all of these things in one
source code repository.
For one thing, it'll be difficult to give someone commit access to
the python library without giving them commit access to the website,
etc.
It also makes it easier to pull things out into git/svk/whatever
repositories if they're split out.
Does Google Code support multiple repositories for a single project?
If not, I would strongly encourage either making a separate project
for each distinct library and the website or finding another solution.
Sorry... I guess I didn't finish explaining about the libraries.
We're starting with the PHP library only. The code itself is housed
at github, and the Google Code project will be used for bug tracking
and releases.
I'm honestly not terribly concerned about granular access. We've used
Google Code for numerous other projects like this without incident.
We'll also setup a separate -commits mailing list, so commits to the
repository don't go unnoticed. Another thing to keep in mind with
regards to the OpenID website, just committing to the repository
doesn't actually push anything live to the website. Someone has to
SSH to the openid.net server and checkout the changes. The list of
people who have that access will certainly be much more limited.
Certainly I think it's very likely that the website and the
libraries will be maintained by different people, and it also seems
very likely that the libraries for each language will be maintained
by different folks in future even if they aren't already. There's
also the Net::OpenID libraries for Perl which currently live in a
repository at Six Apart.
(As a historical note also has the original OpenID website in it
which has been an annoyance to me since it comes along for the ride
whenever I make a branch. I intend to delete the website it at some
point. I'd hate to see the mistake of putting the website in the
same repository as one of more of the libraries repeated.)
certainly, if you put both libraries and the website under a single
"trunk" then that's a problem with branches. The plan is to do
something more along the lines of how the EAUT repository was setup... http://code.google.com/p/eaut/source/browse/
-will
_______________________________________________
board mailing list
[email protected]
http://openid.net/mailman/listinfo/board
