The roads I take...

I've been on a really dense schedule the last days, I hope I'll be able to clear up lots of things now that some bigs steps have been made.

I hoped we would have some time between new releases and the suiterunner switch, but reality wanted it otherwise. This way, both happened only a day apart, and on days where I was studying chemistry with a friend yesterday afternoon (and the day before) as she had an exam today, going out for my weekly Monday/Tuesday Karaoke nights, and having a meeting with my online gaming friends today evening, as well as the Mozpad meeting earlier this evening.

Still, I needed and could find the time to land the suiterunner switch yesterday (which included tree watching and fixups as well as eating dinner on my computer desk while doing that) and get the 1.1.2 (and 1.0.9) release out to the world today.

Ah, yes, right. That's what I wanted to tell you originally. SeaMonkey 1.1.2 is here.

We did a more eloquent announcement for it today, including a paragraph that once again should get the message of SeaMonkey as an upgrade path to Mozilla suite and Netscape 4/6/7 across.
I also did a German version of that message and mailed it out to more IT-related press contacts (in Germany and Austria) than I knew even a day ago. Thanks to #mozilla.de folks for helping me on finding those.

I also got some new download numbers of SeaMonkey, I will probably tell you about those tomorrow, I'm a bit tired right now (3am here in Europe). I probably should already be sleeping, but I still got to look if everything regarding the releases is working as it should and provide Camino folks with some pointer how they can switch over to toolkit as well (they're the last xpfe consumer left on trunk).

After establishing the new project for preserving the original Internet communication suite and good success with SeaMonkey 1.0.x and 1.1.x series, the SeaMonkey team is working very hard on creating a completely new generation of this season product - and bumping its version number to 2.0 along with that.

Using the internal codename of "suiterunner", our development team, has been working for almost 1.5 years now on a project officially named "Turn on the MOZ_XUL_APP flag for SeaMonkey", and esp. Mark Banner (Standard8 on IRC) has invested uncounted hours of work into that (sub)project. While it sounds easy on the surface - after all it's just turning on one simple build flag - this means completely replacing a significant layer in SeaMonkey's codebase with different, newer code, which is to a big part derived from our old code, but has evolved a lot to support Firefox and a growing base of other toolkit-based applications.
Looking at a diagram of the Mozilla Platform recently posted by Mike Schroepfer in a recent blog post, one can get a rough impression of what we're doing here, as we're replacing the layer titled "Toolkit" in this graphic. The changes range from the application startup process via extension management to available/supported/different UI "widgets" and touches a really big range of code areas in between.

The good news is that after having experimental builds of such a SeaMonkey version available for some time, we got them working well enough so that we can retire the old xpfe-based suite and replace it with that new toolkit-based variant on the development trunk. Those are still versions intended for developers and may not work as expected, it's trunk nightlies after all - they just make all your data get lost and your computer to explode. Well, even if this might probably not happen, they are still not though for production use and will have many bugs and regressions - but they should be basically usable. Note that we're changing the base toolkit, but we still haven't managed to get everything working with that new toolkit, so we're still not using the same code as stock XULRunner everywhere, we're excluding some toolkit parts and using our old code instead - until we figure out how to get all that stuff done in the new world (wallet and download manager are examples of that).

Still, pretty major changes are happening throughout the code - for users, the new extension management and new profile location are probably the most notable changes caused directly by that toolkit change. But then, we also have the new icon set as a very visible change, and cairo-based Gecko gives us an additional palette of changes. The more we were thinking about version numbers for a release that will result of this work, we realized that such major changes call for a full major version bump, just upping the minor version number one or more steps from the previous 1.1 version wouldn't fit what's actually happening here.
Because of that, we decided that the suite based on Gecko 1.9 will be called SeaMonkey 2.0.

Of course, we are not anywhere near a real release yet, so when the switch to "suiterunner" will actually land on trunk this week, we'll change the version number to clearly tell that this code is still preceding a first alpha of 2.0, or to make this description fit into a version code, that this is 2.0a1pre code.

Still, this will be a quite important step on the road to a SeaMonkey 2.0 - thanks for everyone helping us to get here, and everyone still helping us continue on that path!

On the road the upcoming SeaMonkey 1.1.2 and 1.0.9 releases, which are due in sync with Firefox 2.0.0.4 and 1.5.0.12 and currently targeted for May 22nd, we once again need people to help testing the candidate builds.

Candidates for 1.0.9 are up on ftp.m.o already, 1.1.2 candidates are being mirrored to those machines at the moment and should be available there within the hour.

If you wonder what changes there have been since 1.1.1/1.0.8, you can get a rough buglist for 1.1.2 from Bugzilla, as well as a rough buglist for 1.0.9 - note that those lists may exclude security bugs that may stay hidden from the public at least until those versions get officially released.
Even without those, we have well over 200 fixes in the codebase in the 1.1.2 timeframe and 100 fixes in the 1.0.9 timeframe (not all listed fixes are actually in the SeaMonkey code), so we should test those builds quite well.

If you have some time left, please help doing smoketests on all 3 primary platforms, so we can be sure to have a chance to fix regressions and keep delivering high-quality software to our users!