> On Sept. 28, 2012, 5:42 p.m., David Faure wrote:
> > dolphin/src/kitemviews/kitemlistview.cpp, line 871
> > <http://git.reviewboard.kde.org/r/106569/diff/2/?file=87267#file87267line871>
> >
> >     This strikes me as wrong. Why should this widget's event(QEvent*) 
> > change the default state of ALL events ?!? No widgets do that, anywhere.
> 
> David Faure wrote:
>     For instance, this would break "closing" such a widget (if it was a 
> toplevel).
>     
>     ignore() all key events, OK (like QWidget::keyPressEvent does). But not 
> *all* events.
> 
> Dawit Alemayehu wrote:
>     Right. That is why I just fixed it. See 
> http://commits.kde.org/kde-baseapps/b3789357335cbbb100b4a089ee7723078e5d219f.
> 
> David Faure wrote:
>     Much better, thanks.
>     
>     (Sorry for the race condition between reading k-c-d and kde-commits ;) 
> Didn't happen in the old days before reviewboard ;)
> 
> Frank Reininghaus wrote:
>     Thanks Dawit and David for noticing that this must be fixed in a better 
> way! It seems that my knowledge of Qt's event handling is even more limited 
> than I thought. I thought that ignoring an event is the way to go if we don't 
> handle it inside the class, and I can't see right now why the first commit 
> broke things like drag&drop, but maybe I should think about it again tomorrow 
> after getting some sleep :-)
>     
>     Anyway, the 4.9.2 packages unfortunately contain the bad commit. I'll 
> send a message to the release-team and packagers mailing lists to inform them 
> about the issue.
>     
>     Thanks again for your help. That shows me that I should probably add a 
> couple more unit tests to cover the basic things...
> 
> David Faure wrote:
>     Some events are accepted by default, some are ignored by default (= 
> before being sent to the widget).
>     
>     Propagation happens [for the type of events that do propagate] if the 
> event is ignored and event() returns true.

Thanks for explaining, David! Just for the record: the 4.9.2 packages were 
updated with Dawit's fix, so nobody will suffer from the temporary regression.


- Frank


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/106569/#review19524
-----------------------------------------------------------


On Sept. 26, 2012, 12:53 p.m., Dawit Alemayehu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/106569/
> -----------------------------------------------------------
> 
> (Updated Sept. 26, 2012, 12:53 p.m.)
> 
> 
> Review request for Dolphin and KDE Base Apps.
> 
> 
> Description
> -------
> 
> This patch fixes a bug where pressing CTRL+Tab does not switch tabs when 
> active tab in Konqueror is a Dolphin's filemanagement part. It happens 
> because DolphinView installs an event filter and does not call 
> event->ignore() or event->setAccepted(false) to allow those events to be 
> propagated up the event chain.
> 
> 
> This addresses bug 302329.
>     http://bugs.kde.org/show_bug.cgi?id=302329
> 
> 
> Diffs
> -----
> 
>   dolphin/src/kitemviews/kitemlistview.cpp 54a8cbf 
> 
> Diff: http://git.reviewboard.kde.org/r/106569/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dawit Alemayehu
> 
>

Reply via email to