[Patch] Fixing the Access Key system

[bj at altern.org]

>  I think you're leaking QLabel's, close() doesn't delete the widget unless
> it has WDestructiveClose set (or you could connect to deleteLater() instead
> of close() ). The labels also don't update position if the page relayouts
> and the position of the element with accesskey changes, not sure how to fix
> that, but that's just a minor glitch. 
Right, I will add WDestructiveClose. Regarding the relayout problem, I think 
the easiest thing is to delete the labels when the page is relayouted.

> Other than that I like it and I'd support committing it.
Cool. I will wait until cvs is open for non-urgent commits (18th), and try to 
commit then.

>  I also noticed another problem with accesskeys after you changed the
> activation to just Ctrl: It doesn't work if the focus is e.g. in a
> lineedit, and it used to with Ctrl+Alt. You can try it e.g. in
> linuxtoday.com discussions.

Yes, that's where the grabKeybord() problem. I made a few tests reintroducing 
it, and I must say that since we only grab one key and the releaseKeyboard 
occurs before re-dispatching the event, I couldn't find any side affect. 
Global shortcuts work well and you can use the accesskey from inside a 
textarea. (see patch attached).

Next biggest problem is to find a way to make all this work with the sticky 
keys feature... 

jean-baptiste
-------------- next part --------------
A non-text attachment was scrubbed...
Name: accesskeys_grab.diff
Type: text/x-diff
Size: 1172 bytes
Desc: not available
URL: <https://mail.kde.org/mailman/private/kfm-devel/attachments/20040716/dd136dd9/attachment.diff>