Created attachment 40424 [details]
Created attachment 40425 [details]
Comment on attachment 40425 [details]
Normally in cases like this I start the new file with an svn copy of the large file I am extracting it from.
I am disappointed we had to make so many FrameLoader internals public in order to break this out! (But that was a comment on the earlier patch really, not this one.)
> + void schedule(PassOwnPtr<ScheduledRedirection> redirection);
Should leave our argument here.
I think the RedirectScheduler is not well factored with respect to the frame loader. It contains more of the logic than I would like that has nothing to do with scheduling per se, which is why we end up exporting so many new functions from FrameLoader.
But this is a noble effort, lets keep going and see where it leads.
> I am disappointed we had to make so many FrameLoader internals public in order
> to break this out! (But that was a comment on the earlier patch really, not
> this one.)
Yeah, me too. I'm hoping this will be a somewhat temporary situation. Once I get a better understanding of FrameLoader, I hope to move some of this back (in particular, the parts that care about isComplete).
Committed r49012: <http://trac.webkit.org/changeset/49012>