Weekly Status Report, W35/2007

This has been a really tightly filled week once again - see for yourself in the following summary of SeaMonkey/Mozilla-related items I worked on in week 35/2007 (August 27 - September 2):

• Website work:
  I did some further work for the future SeaMonkey website, as you can see on the preview site. Still help on content and feedback is appreciated in the development newsgroup.
  We now have all changes in that are needed for possibly introducing localization in the future, the download box on the main page is dynamic if the user has JS activated in his browser, all links should be correct and working, I hope all pages validate, all old breadcrumbs should be removed, and many other such changes.
  A version of the site has been set up on a production-ready system, if the change has been already propagated to your DNS, you should be able to access that one at www-stage.seamonkeyproject.org for now.
• Source L10n:
  I got source L10n for ChatZilla going and have patches read for enabling it for venkman as well.
  This means that we will be able to host localizations for those extensions in mozilla.org CVS now and will be able to build localized versions with the same tools the the original US English ones - for the XPIs as well as for the versions included in (trunk) SeaMonkey builds.
• German L10n:
  Checked in the big update for German security L10n, did some further work on the DOM part of this effort, landed some updates for other shared L10n and kept suite up-to-date with trunk.
  Oh, and I have landed German ChatZilla L10n so it's already usable with the source L10n described above.
  Once the DOM patch and one other toolkit update I added in the same bug as the other shared one are in, we have a complete German L10n of trunk Core again, and I think Firefox is also up-to-date, which should make a German trunk tinderbox go green once again
• SeaMonkey vendor ID and profile location change:
  Led a bigger discussion around the Change SeaMonkey Vendor ID to "Mozilla" bug.
  We will be doing that, which means that all trunk testers will need to move their profiles from the mozilla.org path to the Mozilla one, but this should be fairly easy, as a simple move is enough.
• Code cleanups:
  After Thunderbird and SeaMonkey switched to using toolkit's printUtils instead of an older version from xpfe times, I could remove xpfe/communicator from Thunderbird, which makes Camino-like-Firefox the last potential user of both the xpfe/global and xpfe/communicator directories.
  I also filed a bug and patch on moving SeaMonkey-specific theming to suite/, which should add the toplevel themes/ directory to that almost-orphaned list.
  Additionally, I found out that extensions/help should be completely unused but still around, and filed a bug for removing it from CVS.
• mozStorage Explorer:
  I filed a bug and patch for supporting SeaMonkey in the mozStorage Explorer, which is "a simple little addon [...] that lets you run queries on databases found in your profile", as its creator Shawn Wilsher explains.
  Due to the ability of the new chrome registry .manifest files to do app-specific overlays, this was really easy to do.
  Watch out for a future version of the tool that runs in SeaMonkey!
• Air Mozilla:
  As mentioned earlier in this blog, I met the guy doing the Air Mozilla show, and was interviewed about Mozilla for this broadcast. Watch out for the next episodes!
• Crashes, compiling issues and SeaMonkey startup problems:
  Here's this week's section for nasty things.
  I encountered a printing crash while testing the printUtils fixes mentioned above, filed it and discovered I'm not the only one seeing it, as there were already some nice breakpad reports out there - even from Firefox.
  The I ran into a debug build compile error introduced by making the build process more picky - and some investigation (by me, as some someone who doesn't know much about C/C++) actually made us realize that the compiler is correct to choke on this problem - someone unexpectedly made a mRefPtr be a bitfield...
  (You don't understand what this is or why it matters? Don't worry, you're in good company, I don't either.)
  
  And then, there's this Windows startup issue. I filed a bug on a tinderbox startup test (Ts) slowdown and then realized, it was because of an old friend - the patch for VC8 CRT and XULRunner landed again, and we already had similar problems when it laded the first time. Even worse, the startup failure/crash caused by it back as well
  I hope a solution for it can be found soon, as the future target of building with libxul, which could probably also help solving it, is still too far out for now.
• Various Discussions:
  notification bars, toolbar customization, venkman revival, search component move, prefwindow rewrite, login manager, urlbar history, storage mechanisms, AMO dictionaries, mozpad meeting etc.

I probably spent too much time with work this week, something I shouldn't always do. But then, work on the Mozilla project is so much fun that I just can't stop