Re: RealDeal patch to Matchbox 0.50

Attention: 'dis a Two-In-One Reply (tm)! Now at Walmart for just $2.99!

> Okay, I totally am not in a position to be feeding input here,
It struck me someone who is not _within_ something is _out_ of it and
thus can provide _in_sight which one with_in_ cannot. So never mind, eh?

> Have you tried doing this kind of thing with a custom theme - and just
> setting main window borders to be very big ? If not I think adding
> support for this kind of thing via themes may be the best way.
I have not tried it but given it a thought. Problem is, unless I have
my application write the theme before starting Matchbox it won't work.
Well, it would work _now_ (and require hacking together a Matchbox theme
generator) but not some not-so-far day in the future when I actually
want to resize the Viewport/set it to fullscreen. Actually, with the
patch sent two days ago, the resizing doesn't work either.

> But perhaps there is some demand of Matchbox to have some sort of way
> for a window to hint to Matchbox that it really knows what it's doing,
> Matchbox should honour its configure events..?
It does sound like something that should be thought of from my pov.
(Matthew?)

It does not, however, sound like it would solve my specific problem,
to say it in Greg's words.

Maybe generalizing it would help. Here's a specification:
- Fact is: I have a _NET_WM_WINDOW_DESKTOP window that knows what
  position all the other windows have to have. It does not know their
  size though, which is calculated in conjunction with the screen res,
  which only the window manager knows.
- Fact is: That _DESKTOP window must always be fullscreen.
- Fact is: Each other window must obey to it and must not be granted
  resize and/or move requests of its own.
- Fact is: the _DESKTOP window must be able to ask the window manager
  to resize and/or move all other windows of one or more specific types.
  "Type" is to be understood as defined at freedesktop.org

> Just as your extras to the main 'Wm' struct, most funcs can get at
> this via being passed it directly or the clients ref to it. 
Have I added extras to the Wm struct? Or did you mean to write "Just
_add_ your [...]" instead of "_as_"? Anyway, sounds good :-)

> Hey, please don't be sorry ! Theres nothing wrong with the patch ( and
> thanks for cleaning up the annoying warning ), we just need to figure
> out the best way to support this kind of thing. 
Cool :-). Thanks for the encouragement. There's just one or two mailing
lists out there where you ask a question that could have been resolved
reading TFM and you get flamed to hell. Well, seems this isn't one of
them ;-)

> *sigh* beer.
Well, Greg. Some drink too much of it, but there's also some who should
consider drinking more of it. Heheh.

Now, this was a rather long one. Need to execute water i/o urgently.

- R