Mr. bent tells me that currently they don't.
This patch was pretty easy ... perhaps too easy. I hope I didn't violate any threadsafety rules, please check carefully! I tested it by flipping the pref during long worker computations, and it didn't seem to break anything. (And wow, the JIT makes a huge difference for bz's mandelbrot demo.) Interestingly enough, the JIT-enabledness state change seemed to be applied nearly instantaneously while the workers were still executing. Kudos to the spidermonkey guys for getting that to work. One issue is that this patch sets JIT-enabledness for all Worker JSContexts based on jit.content; it doesn't attempt to determine whether chrome or content JS is executing on the JSContext, and so ignores jit.chrome. Is that going to be an issue? Does browser.js use workers at all?
Assignee: nobody → jones.chris.g
Attachment #414192 - Flags: review?(bent.mozilla)
This would be very useful to have (could have used it in bug 653967 right now, for example). Did this get forgotten, or have we given up on it?
This was a four-day project and I just ran out of time. I'd still like to take this patch if it wasn't fixed in the meantime.
Assignee: jones.chris.g → nobody
This hasn't been fixed in the meantime, and would be useful to have I think.
We need this to enable TI for web workers. We can't hard code it since TI will be disabled by default when it lands on TM. Chris, still interested in finishing this patch?
All this code is gone soon, we'll need a different approach as soon as bug 649537 lands.
Will be fixed in bug 649537.
Attachment #414192 - Flags: review?(bent.mozilla) → review-
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → WORKSFORME
Component: DOM → DOM: Core & HTML
You need to log in before you can comment on or make changes to this bug.