[webkit-reviews] review requested: [Bug 22624] [SOUP][GTK] Need API to get SoupSession from WebKit. : [Attachment 26286] API to retrieve the soup session from the view #3

[bugzilla-daemon at webkit.org]

Christian Dywan <christian at imendio.com> has asked  for review:
Bug 22624: [SOUP][GTK] Need API to get SoupSession from WebKit.
https://bugs.webkit.org/show_bug.cgi?id=22624

Attachment 26286: API to retrieve the soup session from the view #3
https://bugs.webkit.org/attachment.cgi?id=26286&action=review

------- Additional Comments from Christian Dywan <christian at imendio.com>
(In reply to comment #6)
> (In reply to comment #5)
> > Created an attachment (id=26240) [review] [review]
> > API to retrieve the soup session from the view #2
> For the QtWebKit port we have added the necessary Document* pointers to be
> able to get the QNetworkManager of the the QWebView. What I probably tried to

> say/ encourage is that newly created code should try to get the SoupSession
> through the (Document->ChromeClient->WebKitWebPage). Whenever we decide to
> allow to set a per session SoupSession we only have to change WebKitWebPage
> and the rest will be fine. I think it should be little work to do that right
> from the start.

I took an inspirative look at what Qt does and ended up adding a
FrameLoaderClient::session() function, giving ResourceHandle::Session a Frame*
argument and replacing global jar and session variables with Document* or
Session* arguments. FrameLoaderClient seemed the right place for a network
session.
I don't know if this is the best approach, particularly I'm not sure where the
initial session should be set from and if it's good that WebKit now always
creates a cookie jar if there is none.