Review Request: Fancy-schmancy Kickoff layout tricks for dissident panel placements

Will Stephenson wstephenson at kde.org
Tue Feb 26 21:52:29 CET 2008


On Tuesday 26 February 2008 21:04:28 Aaron J. Seigo wrote:
> On Tuesday 26 February 2008, Celeste Lyn Paul wrote:
> > >     Oh, and as a long term rightedge+northeast user on KDE 3, I assert
> > > that having the leave tab always closest to the mouse on popup is
> > > genuinely tiresome.
> >
> > The menu tab options are ordered in order of importance so it makes sense
> > for the Leave button to always be the "last" option, either left to right
> > or top to bottom ordering.
>
> that's what i thought as well ...

I've followed this up on the review board and asked Celeste to clarify, since 
at the first couple of readings I read '"last"' as agreeing with me :).

> with the tab button hover delay this shouldn't be much of an issue either;
> it's much more of an issue with kickoff in kde3 as the delay is much lower
> (non-existent?)

Mis-tab-switches are annoying, but less annoying than always having to go over 
the less important tabs to get where you want to be.  In the scenarios where 
you have reversed order (5,6,7,8 according to the comment at the top of 
setLauncherOrigin()) your search algorithm goes from O(1) to O(N).

If you get time, have a play with and without the patch with these rather odd 
placements and see if your experience matches mine.

> > Are the top-right bottom-right modes more common in right-to-left
> > languages (I can't really see it being a preference for left-to-right
> > language users)?  The inverse organization would might make sense in that
> > case (because order of importance for those languages are right-to-left
> > instead of left-to-right anyway).  This would probably need to be
> > reviewed by i18n.
>
> the order is already reversed for us in RTL desktops for us by the QTabBar,
> so we don't need to worry about that really.

Yup, if the ordering goes in I would have to check layoutDirection() first.

Will


-- 
Will Stephenson
IRC: Bille


More information about the Panel-devel mailing list