XFWM4 Window Problems, Part Deux

Olivier Fourdan fourdan at xfce.org
Wed Nov 15 22:09:31 CET 2006

Hash: SHA1

Milosz Derezynski wrote:
> On 11/15/06, *Mike Massonnet* <mmassonnet at gmail.com
> <mailto:mmassonnet at gmail.com>> wrote:
>> just reappeared in the taskbar.  If you add gtk_widget_show (GTK_WIDGET
>> (window)) it works... but I know you will refuse this line of code...
> Yeah, i could add this line of code, but anyone dare to explain me why i
> need to add this _only_ for XFWM4?

That's simply plainly false. what happens here is that the window switch
to "Initial state is Iconic State" ICCCM WM_HINTS when you call
gtk_widget_hide() in on_window_state_event()

Look, try yourself, do xprop on your window and look for "Initial state
is Iconic State.".

It happen that metacity is currently broken and do not implents this
properly. Kwin does, xfwm4 does, but metacity does not. It used to in
previous version, but not in current version. Try in an earlier version
of metacity/gnome and you'll see by yourself.

A proof? Try "xterm -iconic" in xfwm4, in kwin and in metacity: In kwin
and xfwm4, the window starts iconic as it should. In metacity it does
not... That's exactly the same that happen with your app.

> Seriously your project _does_ claim high standards compliance and holds
> this up as one of it's major points/benefits/whatever the right for this
> is (sorry non native english speaker here) and please don't come with
> this "I just work on it in my free time" statement.

Seriuously xfwm4 does implement the standard, while metacity does not.
It happen (as already explained earlier) that your program in broken,
but it doesn't show in metacity because metacity does *not* implement
the protocol properly.

> So, given the both above statements, anyone care to explain me because
> of which spec (or perhaps brokenness, after all ?) of XFWM4  i need to
> add gtk_widget_show() explicity just for XFWM4

Really, I'm doing my best, but it sounds like a useless waste of time.
You do not listen not even try to understand what I'm desesperatly
tryingto tell you...

Again it's not just for xfwm4. Did you actually try with kwin, or any
descent good window manager? I guess you did not, otherwise you would
not claim such nonsense.

> Just btw, if you look at the code of gtk_window_present() in the gtk+
> source tree, it does already call _show() so what is the point here
> anyway ?

Because if the gtk_widget_hide() in on_window_state_event() that trigger
the start iconic flag. I already explained that.

> I'm riding so much on a minor issue because you boast all the time with
> standards complicance, yet now you suggest me to use a workaround; and
> if it's not a workaround then i'd seriously like to know by which spec
> this behaviour  is defined, and different from all other WMs (and also
> implying that all other WMs are broken or at least not fully compliant).

Afgai, it's metacity that is broken, you cannot just trash all other
good window managers because metacity is broken.

> Anyone? (again)

Again and again and again, but you really don't seem to try to understand...

Version: GnuPG v1.4.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the Xfce4-dev mailing list