The roads I take...

Here's a short summary of SeaMonkey/Mozilla-related work I've done in week 27/2011 (July 4 - 10, 2011):

• Mozilla work / crash-stats:
  Continued calculations of more real crash rates for recent release and beta versions.
  Discussed with the Socorro team on getting improved graphs.
  Followed the source uplift and new version coming out of it, useful data from the new version is only to come in the upcoming week.
  Stayed in the loop on GC crash instrumentation, MethodJIT fixes and MemShrink outcomes that all could or should influence crash data.
  As always, looked into "explosive"/rising crashes with my experimental stats.
• SeaMonkey Build & Release:
  Helped Callek to get the final 2.2 release out the door, including website updates as Jens was on vacation.
  Got Linux64 updates going for 2.0.*->2.0.14 and 2.1*->2.2, major update will follow once it goes public for the other systems as well.
  I also helped to get the major update billboards up on a https website (mozilla.org, in this case).
• SeaMonkey L10n:
  Went through the first round of sign-offs for aurora 2.4 and beta 2.3 for SeaMonkey.
• German L10n:
  Updated the German website for SeaMonkey 2.2, and synched DOMi, SeaMonkey and toolkit localizations for all of -central, -aurora, and -beta.
• Various Discussions/Topics:
  Source uplift, input/feedback sites and non-supported browsers, new release process and "enterprise" users, MeeGo N900 CE, N9/Harmattan/MeeGo and Firefox mobile, Mozilla sponsorship for SotM-EU, etc.

I feel we start moving more and more in directions to improve the situation with crashes on Firefox by attacking some of the larger areas where we are seeing problems - unfortunately, we have three large areas where the "signature" doesn't tell us a lot about what the actual problem was (GC and MethodJIT crashes as well as plugin hangs) and so those areas are hard to work on and need more instrumentation. Still, we are moving there as well. Reducing memory usages in MemShrink should help reducing out-of-memory crashes. And we just uplifted another version to the Beta channel, which should get it finally exposed to enough users to do good crash analysis. If you want to see more interesting work, though, you should check out Aurora, which already contains some more of that work on memory usage, etc. I love working in an area where you see things really moving toward getting better!