Re: [dev] clients with large min size

From: Antoni Grzymala <antoni_AT_chopin.edu.pl>
Date: Wed, 14 Oct 2009 09:33:07 +0200

Anselm R Garbe dixit (2009-10-13, 18:05):

> 2009/10/5 Szabolcs Nagy <nszabolcs_AT_gmail.com>:
> > there are some clients with large minimum size (eg gnome programs:
> > gedit, evince (with side plane),..) which might cause problems:
> > * tiling algorithm does not handle the case when the clients on the
> > stack are larger than configured (it handles the case when they are
> > smaller though, like xterm)
> > * large clients on the stack might span across another monitor
> > covering it partly (and mod+left click makes the client jump to the
> > other monitor if its center is on the other monitor)
> >
> > i'm not sure if/how these cases should be handled gracefully
>
> I agree these things are a problem, but more a problem through the
> brokeness in the list of assumptions of the underlying apps. We
> require that clients are not broken and must accept any size in order
> to support tiling layout algorithms. For tile() in particular I always
> preferred to keep it simple instead of trying to support these
> corner-cases. Though it might be worth thinking if there is any
> sensible solution -- a quick thought concludes we could pull clients
> into floating when the resize() operation detects a size that isn't
> acceptable (which will also reject the attempt to resize). What do you
> think?

Makes sense (and is by intuition – at least mine – the expected behaviour).

-- 
[a]
Received on Wed Oct 14 2009 - 07:33:07 UTC

This archive was generated by hypermail 2.2.0 : Wed Oct 14 2009 - 07:36:03 UTC