After arriving to Boston last Monday, dropped my stuff in J5's house and went together to have some beers with some Sugar SLOB's at the CBC, there I finally met Colin Walters, putting one more face to an IRC nick. Afterwards we went to the OLPC offices to discuss Sugar Labs' trademark policy, and though I had to leave shortly after because of the lack of sleep, turns out they swiftly solved this issue that has been dragging on for so long, congrats!
On Tuesday several of the Sugar SLOB's got together to discuss miscellanous stuff about Sugar, mainly doing some high-bandwidth chatter to share our positions on several issues. Discussions should follow soon on the IAEP mailing list.
Wednesday started the hackfest, met Zach Goldberg, John Ehresman (of Wingware fame) and Red Hat's David Malcolm. I'm still impressed by the professionalism and technical knowledge of these guys, it allowed us to keep a strong focus and to overcome together the roadblocks we found, even when we weren't able to agree on a single position. David partnered with John and J5 on the Python 3.x side, and Colin, Zach and me on the introspection side. Having in the same room people with as a deep knowledge of Python such as John and David also helped a lot to the introspection side of things.
From Thursday to Sunday we just continued hacking on our goals, from 9am to 6pm on the OLPC offices with a small pause to grab food. Was so much fun that we only stopped when our brains grinded to a halt. Check the links below from Zach for more details about each day.
On Saturday night we celebrated the first PyGI release: 0.5. One very important outcome of the hackfest is that Zach has joined the maintenance team of PyGI, assuring its continuity, this will mean that PyGI will have a well triaged list of bugs, patches will be reviewed on time and releases will be made on time. Simon van der Linden (who unfortunately couldn't make it to Boston) and me will be lending a hand on that from time to time.
The Python 3.x patches haven't landed yet because are very invasive and we need to make sure we don't introduce more regressions than what can be fixed timely, we don't want to break all PyGTK software out there! So the plan is to keep developing the Py3k ports of PyGObject and PyGI in separate branches and test them with big codebases such as Sugar until we are confident of their stability, then we'll propose the merge. Both ports are feature complete and ready to be tested, which is a giant step forward.
I'm very grateful to my employer Collabora for having sponsored the time I spent hacking on PyGObject (and traveling!), to the GNOME Foundation for sponsoring my trip, to my OLPC colleagues for lending us their facilities, to Red Hat for taking us out for dinner and to Canonical for the coffee. Also, my special admiration to Red Hat for understanding that a strong downstream requires a strong upstream and sending three of its best hackers to lend us a hand here.You can read more about the hackfest from others:
* http://arstechnica.com/open-source/reviews/2010/04/python-support-in-gnome-gets-a-boost-from-hackfest.ars
* http://blog.verbum.org/2010/04/14/pygtk-performing-engine-maintenance-while-the-car-is-running/
* http://zachgoldberg.com/2010/04/15/gnomepython-hackfest-2010-day-2-or-how-to-actually-coordinate-packages/
* http://zachgoldberg.com/2010/04/17/pygi-hackfest-day-4-a-call-back-to-the-past/
* http://zachgoldberg.com/2010/04/18/pygi-version-0-5-the-watch-out-theres-a-volcano-release/
* http://www.j5live.com/2010/04/12/gnome-python-hackfest-thanks-to-sponsors/
* http://www.j5live.com/2010/04/14/caffeined-up-and-starting-to-hack/
* http://www.j5live.com/2010/04/14/gnome-python-hackfest-day-1/
* http://www.j5live.com/2010/04/15/gnome-python-hackfest-day-2-quick-roundup/
