Re: [EWMH] _NET_WM_WINDOW_TYPE_AUXILIARY

[Tuomo Valkonen <tuomov iki fi>]

On 2007-10-16, Lubos Lunak <l lunak suse cz> wrote:
>  I did, but maybe I didn't get it. Are you talking maybe about "crappy 
> toolkits or such forcing a window type being specified"? AFAIK it's a rather 
> common practice to detect client support for the spec by a presence of any 
> _NET_WM_WINDOW_TYPE hint. 

We're talking about __override-redirect__ windows here. It's rather 
common for them to have no hints. Why should programs specifically
support EWMH to not have their override-redirect windows arbitrarily
messed with? Why should they support it at all? It's largely highly
WIMPshit-specific crap designed by people with a tunnel vision and 
little regard for abstraction and alternatives. The ICCCM, by contrast,
mostly provides basic "how can we get along without imposing too much 
policy" guidelines. If the EWMH or other FDO crap goes against the spirit
of the ICCCM, instead of simply providing additional _hints_, why should
one choose it over the former or an interpretation thereof? Because the
desktop herd that is the source of present sordid state of *nix/FOS 
wants that?

I also mentioned that AUXILIARY is a better more abstract hint to 
override crap that wants to set the TYPE hint on override-redirects,
than the highly-specific EFFECT. Why should the CM or WM care what
is contained in an override-redirect that it doesn't know how to
handle in a special manner, unlike e.g. menus? Can you think of 
any actually useful case where it should know that it's dealing
with an EFFECT specifically? And isn't that likely to be a very
marginal class? Wasn't even the original post about some very ugly
overlay hack or something? (Can't be arsed to find it now.)
Let's keep the complexity down if it achieves little.

But do whatever you will. I've lost all hope in FOSS and am likely
to switch to Windows once it comes time to upgrade.

-- 
Tuomo