I volunteered to do a migration of the code to github.
I will do the following steps:
1. Do a SVN to Git migration for every component on its own. Every
component will have its own repository with issue tracker on Github.
2. Rewrite history to add composer.json files to all stable tags
3. Push to Github
Derick already opened an Organization on Github and I guess all current
committers will get write access to it.
Adding composer helps us with the deployment issue in the short term and in
the medium/long term we could maybe setup a pirum channel on github pages
to enable PEAR installation again. Composer can work with classmaps so that
the file/class layout is no problem for autoloading and components become
instantly usable by third parties.
The next step would then be that committers claim components and we
determine the deprecated/abandoned/maintained state of each of them. Every
maintainer is then responsible for getting the tests back running with the
new schema and adding this to travis-ci. I will attempt to do this for one
component as a demonstration.
I can't give a timeframe on this but will highly prioritize this and hope
to get it done within a week.
On Fri, Apr 20, 2012 at 11:36 AM, Henri Bergius <henri.berg...@iki.fi>wrote:
> On Apr 20, 2012 11:26 AM, "Tobias Schlitt" <tob...@schlitt.info> wrote:
> > sadly, I did not find time to look into all this stuff. This also seems
> > the reason that Zeta is dying: Time lack.
> > I'm a bit sad, that Zeta is being closed down, but I think that it's the
> > logical way to go.
> I think the important part is that PHP is finally starting to have a
> reasonable infrastructure for having a proper ecosystem:
> * Composer for dependency handling and easy library publishing
> * PSR-0 for standard autoloading
> * GitHub for collaborative development
> * Travis for continuous integration
> These four combined make it much easier to publish simpler, more decoupled
> libraries that "just works". Node.js has a very similar infrastructure with
> NPM, and there indeed this has produced hundreds of compatible modules that
> are widely used, and that don't need any sort of monolithic project
> governance structure.
> I think there is still a place for an effort advocating library
> interoperability, quality and coding standards. Zeta could be that, but it
> needs to be part of this new ecosystem, not something standing outside of
> > Toby