A bunch of things have happened over the past few weeks, and if I didn't write about them now, chances are I would've given up and just dropped them on the floor.  Shortly after returning from the Mozilla Summit — and surviving power outages (except for our wireless network, it stayed up, even though the rest of the hotel was out of power!), landslides, bears — I went to SIGGRAPH where I was on a panel as part of the cohosted Web3D Symposium.  There were a few things that happened at the conference that I'm excited about, and I'll be writing about those soon.

After SIGGRAPH, Mozilla hosted a number of core Cairo developers at the first-ever Cairo Summit.  It was great having everyone in the same room together (well, almost — Behdad was omnipresent by phone).  Carl and others took some great notes about the discussions, where we were able to both solve existing problems as well as do some planning for future work.

One particular work item that I was able to start at the summit was to begin setting up a buildbot for Cairo, so that we have constant build and test runs on all the major platforms that Cairo supports.  The initial plan is to set up a few flavours of Linux (Fedora and Ubuntu, as well as an x86-64 variant), Windows XP, and Mac OS X; in addition, Mozilla will be providing some Windows and Mac OS X machines that Cairo developers can use remotely for testing and development work on those platforms.  Having the builders also means that we'll be able to provide nightly (and release) binary packages for those platforms, thus lowering the bar for new developers who wish to start working whit Cairo on non-Linux systems.

We've still got a lot of work ahead of us both in Cairo and in Mozilla's graphics layer, but at this point the main focus is going to be performance.  This was the main outcome from the graphics session at the summit as well — Cairo and Thebes expose the right functionality, but the uncommon cases just need to be faster.  Jeff's been focusing optimizing common paths in pixman for ARM, and we'll soon be able to start focusing on GL acceleration.  We have some things to solve before we can go fully down the GL/DirectX route (plugins, compositor), but all that should come together.

Cairo's turned into a pretty solid cross-platform graphics library, and I'm excited that we've been able to contribute.  Another graphics library, Skia, has recently appeared as part of the Google Chrome code drop.  It's unfortunate that Google felt they needed to develop their own alternative in a closed fashion instead of joining an existing open source project.  The Cairo project, and through it the many open source projects that depend on it, could have benefitted from the work that was done on Skia behind closed doors.  Even worse, unlike most of the rest of the Chrome code, Skia is licensed under the Apache Public License v2.0.  This creates difficulties in being able to reuse the Skia code in most projects.  (From looking at the source, it seems like there's a private "upstream" for Skia as well, and that the Chrome version is just a copy... I guess maybe it'll get thrown over the wall at some point as well.)  I'll certainly be taking a detailed look, though.

I've also been spending time working on some aspects of our platform's mobile story; I picked up a BeagleBoard a while ago, and it's been a fantastic platform for development and profiling of Gecko on ARM.  That work deserves its own post, though, so I'll write up something later on this week.  (That's two things that I've now set myself up for followup posts about; I'll probably blog more this week than I have in the last two months!)