A very general question: what's the high-level plan for Servo regarding Firefox? I see a lot of requests/issues dealing with stuff that doesn't seem related to a layout engine. Is the plan to have browser.html as a temporary incubator for Servo and eventually migrate it to replace FF's layout engine, or is it to "grow" an entire new browser around this new engine?
There are efforts to make various components work in Firefox. The largest effort is "Stylo", which replaces Firefox's selector matching with ours. I'm not sure what the future is on Firefox using Webrender or our layout; though I suspect it depends on how well stylo works.
The plan seems to be to let Servo continue to evolve as a testbed for new ideas (like webrender), and share components with firefox whenever ready. An independent Servo product is not likely in the near future, since there's a lot of work to make it fully web compat. In the far future ... well, we can't tell :)
Admittedly, android-specific efforts have slowed down in the past few months, but I believe it is still a goal :)
I mean, Servo-as-a-product isn't even planned yet, and may never be, so by the time that question is relevant the ios rules may have changed, too :)
As for Firefox, later on this year we should see the integration of Servo's style engine ("Stylo") into Gecko, and going forward we should expect to see the two codebases share even more components. However, wholesale replacement of Gecko with Servo in Firefox isn't on the table at the moment.
Well, it depends on some privileged APIs, of course. But we've tried to keep that API surface minimal.
In addition, the browser needs to act as a sort of broker between sites and the OS. Storing cookies and session data, showing notifications or dialogs, and entering fullscreen mode are just a few examples of this. (Note that most of these things aren't even implemented yet. Browsers are complex applications in their own right, even after accounting for the browser engine.)
We are deliberately trying to keep this to the bare minimum, and will likely try and find someway to standardize those few things we do if possible.
Servo's style engine is parallelized and scales nearly linearly with number of cores, and style time is a non-trivial amount of the total rendering time.