[Webkit-unassigned] [Bug 70955] Add FakeCCThread that supports multiple in-flight tasks

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=70955





--- Comment #5 from Iain Merrick <husky at google.com>  2011-10-27 04:07:18 PST ---
(From update of attachment 112570)
View in context: https://bugs.webkit.org/attachment.cgi?id=112570&action=review

>> Source/WebKit/chromium/tests/FakeCCThread.h:39
>> +    long long getTime() const;
> 
> We use double elsewhere, so we can get precision. The postTask takes a long long because it only understand that, but queuing could introduce 
> 
> I support it being on ccthread, but we would need to be able to write code saying "GetTheCurrentTime" that would work on main thread too. We need that anyway because I'd like the same CCDelayBasedTimeSource code to work regardless of running on mainthread or ccthread. I think a good way to do this is do away with CCMainThread class and instead have have a global somewhere in the compositor which is s_mainThread. The embeddor of the compositor, e.g WebKit/src/chromium could set up CCMainThread just like it does for the impl thread.

Hmm, there are a bunch of issues here, let's discuss it in person.

I'm suspicious of new globals, but I need to think it through. I like the idea of the main thread and the cc thread just being instances of some common Thread class, not different types. We'd be introducing more Chrome-style threading into WebKit, which feels like the right direction.

>> Source/WebKit/chromium/tests/FakeCCThread.h:45
>> +    void incrementTime(long long);
> 
> advanceTime

Done

>> Source/WebKit/chromium/tests/FakeCCThread.h:49
>> +    bool runNextTask(long long);
> 
> also a runall?

Good idea, done.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.