[Webkit-unassigned] [Bug 30979] [Qt] Expose QWebPageClient through the QtWebKit API.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=30979





--- Comment #4 from Simon Hausmann <hausmann at webkit.org>  2009-11-01 19:34:27 PDT ---
(In reply to comment #3)
> > We should extend the existing public API (QWebPage, QWebView, QGraphicsWebView)
> > instead of introducing yet another public interface for developers to learn.
> Qt 4.5 allows replacing QWebView with another QWidget, and there are no side
> effects. In Qt 4.6, applications that replace QWebView with their own QWidget
> will crash when loading plugins. IMO, this is regression.

That is true, it is a regression. Even though it's a use-case that we shouldn't
encourage we shouldn't break it either, especially as it looks like it's easy
to fix:

QWebViewPrivate is the implementation of the QWebPageClient interface that
operates on the QWebView. However at closer inspection it appears that there's
nothing in that implementation that makes it really depend on QWebView. AFAICS
all the calls and queries to the view would work on any QWidget.

It appears the fix would be to rip out the QWebPageClient implementation from
QWebViewPrivate, turn into into a private standalone class (say
QWebPageWidgetClient) and use it independently from QWebView. I guess QWebPage
should own/allocate the instance when QWebPage::setView() is called.

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