On 03/08/2013 07:57 PM, Jocelyn Turcotte wrote:
>> But in order to build it, I needed to make use of some of the private
>> headers (such as QWebPageAdaptor).
>> Given that the corresponding symbols are defined (at least in Ubuntu's
>> libQt5WebKit.so), do you think it makes sense to distribute private
>> headers as well?
>
> Those headers are going to change for sure, so personally I think that it
> would make it easier if you would be able to use the widgets/webkitwidgets
> stable API instead.
> The performance loss that you would suffer from the WebKit1 graphics
> limitations are going to be more important than the widgets overhead anyway
> IMHO.
It wouldn't be an optimal solution, but at least for the time being it
would be a good one. Unfortunately QWebPage makes some assumption that
the view is either a QWebView or a QGraphicsWebView, so it cannot be use
as is.
> I think that using QWebPageAdapter is a great idea short term, but it would
> be a mistake to guarantee you that we won't change this module architecture
> detail considerably in the future.
Another problem with my approach is that I need to use exactly the same
configuration used when building QtWebKit, because some classes change
size according to configuration flags:
class QWebFrameAdapter {
...
#if ENABLE(ORIENTATION_EVENTS)
// define a field
#endif
// more fields
}
and deriving from these classes while having different configuration
options will make my code crash.
In fact it does, and I'm struggling to find a way to obtain the same
list of options.
> Please don't hesitate du ping a few people on IRC to discuss the details.
Thanks! I appreciate it, and I'll surely use the opportunity before
banging my head against some wall. :-)
Ciao,
Alberto
--
http://blog.mardy.it <- geek in un lingua international!
_______________________________________________
webkit-qt mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-qt
