Justin Wood (Callek) wrote:
But then this becomes problematic for cases of Extensions/Profiles, does it not?
Depending on the supporting architecture, possibly...
Especially when needing to ensure that one instance doesnt overwrite another; for example, the bookmarks data, even more complex if we add a new bookmark in _both_ instances seperately.
This is indeed an issue. There are reasons we don't implement this either... ;)
I would be curious on how much of a perf hit happens with a threadsafe DOM. Even if it does not get checked in.
Good question. Want to do it and test? For a start, put locking around all modifications of the content tree in the HTML content sink. See how much that affects pageload?
-Boris
I wont have time to undertake a task like this for quite a while, I expect...(especially since I would first need to better understand the content sink stuff); Not to mention my machine would likely be a bad candidate for reliable perf hit checking.
Though if it remains untested by the time I am available for such a patch (likely a year from now, if not more :( ) I will try and do it, and if nothing more provide the diff.
~Justin Wood (Callek on moznet IRC) _______________________________________________ mozilla-layout mailing list [EMAIL PROTECTED] http://mail.mozilla.org/listinfo/mozilla-layout
