Re: [wmii] Re: IntelliJ 5.1 and wmii-2.5.2

From: Steffen Liebergeld <perl_AT_gmx.org>
Date: Fri, 10 Feb 2006 11:10:51 +0100

On Fri, 10 Feb 2006 09:46:36 +0100, "Anselm R. Garbe" <garbeam_AT_wmii.de> said:
> On Fri, Feb 10, 2006 at 08:52:59AM +0100, Steffen Liebergeld wrote:
>> Wmii should not try to handle broken applications if that would
>> mean to break the good design. Nobody really wants to have lots of
>> kludges and special cases in the code to handle different behaviour
>> of programs. An example for software developed to keep application
>> compatibility over ages is windows. And nobody would want to have
>> crap like this.

> Actually wmii should be that robust to not crash if a broken app
> runs amok or does weird things.

That should be clear.

> Another question is about MWMH and EWMH support. wmii already has
> partial EWMH support for setting the desktop number correctly. But
> it is questionable how far one could and should support those
> specifications, because they contain arbitrary crap invented by the
> KDE/Gnome/Xfce front to aim Windows and OSX behavior completely (and
> even extending it). I think if we can live without special hints,
> we should not support them, as long as apps work in floating area at
> least. Otherwise we implement unnecessary code which is executed
> very rare.

I think is is a go or don't go option. If you start to implement a
standard, you should not stop after implementing part of it because
that would introduce another dimension of indeterminate behaviour. I
think that would clearly break the rule of predictability you
expressed on the wmii.de site.

After all, it is because most programs don't fully follow the icccm,
that you have all those problems writing a good usable wm.

>> @garbeam: Will there be a max layout in next snap again? I'm only
>> asking because I didn't hear a clear statement about it.

> Dunno if it will be in next snap, but at least within one of next
> snaps. The 9P integration is finished already and I merged wmiikeys
> and wmiibar into the WM - this means we have only one 9P server -
> wmii itself. To the 9P client world this is transparent, to the WM
> this has some advantages (lesser code and use of symbiotic effects
> such as key handling in one area, knowledge about geometry of
> wmii-related windows like the bar without doing nasty
> synchronization stuff).

Good to hear that.

> In the end we are around 6.5kSLOC again (inlcuding all libs).

About the libs: I know the code is readable (at least those parts I
read were), but it would be good if a) there would be a short
introduction to each library (what does it do) and b) a oneline
comment in front of each methode to describe it. It takes _lots_ of
time getting into a project that does not comment its methods (even
when the code is really clean and readable). It would also be good, if
there was a beginners guide for developers, telling the main data
structures and where they are defined (fuck, that should be my part of
the story).

-- 
Microsoft is simply one example of a proprietary software developer, a
software developer that tries to subjugate users to keep them divided
and helpless. -- Richard M. Stallman
Received on Fri Feb 10 2006 - 11:11:23 UTC

This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:00:00 UTC