This is going to be a smaller release than the others so far. I've been working primarily on bug 726904 - VideoFrame::GetVideoIntrinsicSize setting wrong size.
This bug has been proving quite difficult to resolve. I have the fix pretty much solid, butt I break a whole bunch of existing video reftests. This is happening only because my fix changes how video elements size themselves. A lot of the tests depend on the old behaviour in their comparisons, which is causing the video to appear a different size. I spent a large part of my time modifying and running these reftests, trying to make them test what they needed, but without relying on the old behaviour. In doing so, I needed to create 2 new ref images for the tests. I posted the patch up to the bug, and Chris Pearce pushed it to the try servers. Unfortunately, they all still failed, which confused me. Then I looked at the patch and realized that the images I had created were curiously missing. I asked around and did some google searching to find out why. Turns out the alias I used in Git didn't have a --binary option >.< which then included the files.
Besides the failing reftests, there are also fails being reported for my mochi-tests. Frustratingly, they're only failing on Mac OS X. I'm going to have to now set up a firefox build on the Macbook I've borrowed from CDOT so I can ensure they work on Mac, as well as Linux (which I'm currently developing and testing on).
I've also applied for level one commit access, which will enable me to push to the try server before I post my patches on bugzilla. hopefully I'll be able to be more confident with my proposed fixes now that I have this tool available to me.
Moving forward, I'm teaming up with David Seifried and Matthew Schranz. We're going to begin implementing media playback statistics in Firefox. Looking forward to that!