Zapping the Xorg server

[Wolfgang Draxinger]

On Wed, 25 Aug 2010 09:07:16 +1000
Peter Hutterer <peter.hutterer at who-t.net> wrote:

> re:xmodmap, the following may be interesting reading
> http://who-t.blogspot.com/2010/06/keyboard-configuration-its-complicated.html

Educational indeed. However I hate relying on desktop environments and
their developers. What if I don't use one of the big two (Gnome, KDE)?
What if I'm using XMonad and must do everything using CLI tools called
by scripts? (BTW: The CLI tools tend to be more reliable than their
GUI counterparts, e.g. I use xrandr exclusively because any GUI to
XRandR has issues in some way, and I've become tired of bug reorting).

Also it's again a prime example of fancy features given higher priority
than the really important stuff. The guy who actually has multiple
keyboards all with different layouts connected to the same X session
please stand up. Yes, it's a nice "to have", but I think there are other
things, that really should be brought forward. Like Gallium and other
parts of a unified graphics system for Linux.

Or working in the stability and code quality of X.org as a whole.

> re:zapping, the XKB option was _added_ in response to the initial
> default of DontZap on (which now defaults to off again, like it
> always has) http://who-t.blogspot.com/2009/04/zapping-server.html

Allright. But then I propose, that there should be some way to
configure a Zap that can't be overridden by the user.

> Either way, it's not going to be removed by "some smartass".

We'll see about that. The problem here (and with some of other more or
less recent developments) is/was, that I tend to have a pretty good
gut feeling about what's going to happen. So let's cross fingers, that
I'm wrong this time around.

If I had to decide anything about this, I'd put development of features
to a full stop and focus on getting rid of bugs and improve stability.
I mean, X must not crash if a client does something nasty. Yet it
happens quite often to me, that X crashes due to a client's action. For
example if I install MinGW tools with(in) Wine, the mapping of one of
the summary window crashes X if there's no reparenting WM (I figured
this out only last week, so I didn't file a bug report yet, because I
don't know yet, what's actually going on).

Stability and code quality, that's what IMHO X.org should focus on. All
the rest, it works good enough to be way ahead of the competition, and
it's mainly the DE's developers who must get their act together. IMHO
KDE4 is an Epic Fail, and Gnome, well, the only good thing about is
GLib and GTK, the rest su*** - I liked KDE3 though and had so much
hope for KDE4 and got hugely disappointed.


Wolfgang