[Webkit-unassigned] [Bug 70289] Add JSC support for delivering mutations when the outermost script context exits

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=70289





--- Comment #4 from Adam Klein <adamk at chromium.org>  2011-11-08 09:48:48 PST ---
(In reply to comment #1)
> We mostly have a bottleneck for call in JSC by way of JSMainThreadExecState::call(), though it is not used everywhere (that we can fix).  I am not clear on how aggressive we have to be however.  For instance, do we need to consider calling a getter while processing an options dictionary passed to an Event constructor a script invocation?

Ideally I'd think we'd want to deliver mutations if we ever exit JS and find that it was the outermost script invocation.  In your example, the event constructor would actually have been called by script itself, so it wouldn't be affected. But I can imagine cases where a JS object's properties are accessed from C++ (maybe structured clone in IndexedDB?). For completeness, I think getters/setters would need to be covered as well, though I am having trouble coming up with a lot of cases where it matters (since most accesses of JS objects are due to calls from script themselves).

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