[Webkit-unassigned] [Bug 22624] [SOUP][GTK] Need API to get SoupSession from WebKit.

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Jan 19 17:49:11 PST 2009


------- Comment #13 from xan.lopez at gmail.com  2009-01-19 17:49 PDT -------
(In reply to comment #12)
> > +    SoupSession* session =
> > static_cast<FrameLoaderClient*>(loader->client())->session();
> > +    return session ? session : default_session;
> > +}
> > 
> > Also, just wondering in general, if the sessions are async maybe some checks
> > wold be in order when swapping them (making sure they have no pending work?).
> > Maybe that's related to your error...
> My error came during browsing while keeping the default session, and while
> using a non-default session, so it doesn't look like swapping is the problem. I
> added an if(!session) to ensureSession now, and it looks like it prevents that
> wanring. I'm still unsure however under what condition startHttp would have an
> empty session...

That still does not cover changing the session for a view while the view is
using it. It seems to me that the way it's done in the patch can't be too safe.
At the very least we should check the queued messages and cancel them...?

> > And as an optimization, I think a flag saying the session is initialized
> > instead of checking for the features every time we go through startHttp would
> > be good (I think I had this in previous patches).
> I wonder about that, I was thinking you intended to make it dynamic later on.
> Do you think the feature test is that slow? If we agree on which we prefer and
> how to document that API wise, I can change it accordingly.

It might be a soptimization, but I think it's pretty sane to just set a flag
after you are done (and reset it when we swap sessions) so you just don't check
the features over and over again...

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

More information about the webkit-unassigned mailing list