Re: [dev] [dwm] mapping wm state, and other stories

From: Anselm R Garbe <garbeam_AT_gmail.com>
Date: Tue, 1 Nov 2011 12:27:14 +0100

On 1 November 2011 12:18, Connor Lane Smith <cls_AT_lubutu.com> wrote:
> On 1 November 2011 11:11, Anselm R Garbe <garbeam_AT_gmail.com> wrote:
>> I disagree, setting a particular layout implicitly because some random
>> window says it is in fullscreen state is not a great idea imho.
>
> By making one client fullscreen we're already effectively doing this.
> We're just making the other windows invisible, too. That seems to me
> to be even worse.

Well applying monocle will lead to other problems. Some fullscreen
windows will destroy themselves when they loose the input focus (for
example flash). Also bare in mind that you'd treat perfectly fine
positioned clients in a tile layout now as monocles for no really good
reason.

The nature of fullscreen windows is that they are transient and that
there should only be one at a time. Otherwise you have monocle
already. But implicit layout switching is definitely a no go, as
monocle might not exist at all (at least dwm is designed that layouts
could be easily removed).

>> I don't understand this. Cycling through any kind of windows is
>> monitor specific already, if not we have a bug.
>
> What I mean is, if a window is between monitors, you should be able to
> cycle through it on either monitor, its canonical monitor being the
> one with the largest area of intersection. At the moment the window is
> arbitrarily bound to one of the monitors (you cannot tell which
> without focusing it), and can only be cycled through on that monitor.
> This is a symptom of the stack-per-monitor approach, which imo is a
> design flaw.

I don't think so. A window belongs to the monitor where its center is.
If the center is very close to a boundary, well I don't get what the
purpose is with this approach, why would I want some parts of a client
on different screens?

Cheers,
Anselm
Received on Tue Nov 01 2011 - 12:27:14 CET