|
|
This week I learned just how well Rakudo development is able to continue even when I'm preoccupied with things not directly related to making commits to the Rakudo repository. As you probably know, I spent this past week in San Jose at OSCON, which was as usual was a fantastic conference. I was able to give some presentations about Rakudo and Parrot (recruiting!), meet lots of new people, visit and make plans with others working on Perl 6, and otherwise just have a good time. Prior to coming to San Jose I had been planning that I would have some time in the evenings to work directly on Rakudo and prepare for its #19 release on Thursday (Jul 23). In past years at OSCON I've been able to get a fair amount of hacking done in the evenings... but that didn't happen this year. First of all, I found myself having far more hallway discussions and other sorts of meetings with folks than I have in past OSCONs. Then, on Tuesday night I decided that I really wanted to do a significant refactor of the "Hacking Rakudo Perl" talk that I would be giving on Wednesday. I knew this would take away from my tuits for preparing the Thursday Rakudo release, but I felt the benefit of having a better talk (which I will also be giving at YAPC::EU in Lisbon) was worth the risk of having trouble getting the release out on Thursday. Fortunately I never had to run that risk. A couple of weeks ago I posted an article about getting release managers for Rakudo. On Tuesday (while I was busy at OSCON), Moritz Lenz was able to use the release guide to produce a "practice release" as described in the article. As a result, on early Wednesday morning I was able to see that Moritz had produced a practice release, and since I wanted to focus on my presentation I immediately asked him if he would like to do the official release on Thursday. He accepted. (I also asked Jonathan if he could work on eliminating a Null PMC error that was showing up in backtraces and didn't look good for the release, which he of course took care of very quickly.) I am extremely pleased with how everything turned out. First of course is the release itself, which is equal or better to anything I would've produced. Moritz, along with help from others on IRC #perl6, assembled the ChangeLog, drafted the release announcement (adding a new section for deprecations), and as far as I can tell executed the release flawlessly. As it turns out, my role in the release consisted only of reviewing the release announcement on Wednesday night, adding a couple of statistics to the announcement, and telling Moritz that he could issue the release whenever he felt it was ready. I didn't have any further involvement after that -- the next I heard about it was someone coming up to me at OSCON on Thursday afternoon to congratulate me on the Chicago release. This past week saw more than just the release; during the course of the week several people added new features and bugfixes to Rakudo (with Moritz and others applying contributed patches). Rakudo's passing spectest count increased by over 100 during the course of the week. So, my thanks go out to Moritz, Jonathan, Carl, Kyle, Scott, and everyone else who worked on the release and Rakudo this week. You all did a fantastic job, and tremendously simplified my life while I was at OSCON. Of equal importance, we've now demonstrated that others besides myself can do Rakudo releases (although I had little doubt of this). As a result, I'm now issuing an official "call for release managers" for the remaining Rakudo releases in 2009. Ideally I'd like to see that the release responsibility is regularly rotated across multiple managers (as Parrot does), so that no single person ends up being burdened with producing releases. I'm also planning that I personally won't do any more Rakudo releases myself in 2009, unless we're completely unable to find release managers. If you're interested in becoming a Rakudo release manager -- take a look at the release guide, make sure you've filed a Contributor License Agreement, and we'll put together a schedule for the 2009 releases. Pm |