3:02 p.m.
Under X, At least one window manager, namely openbox, doesn't see the
basic frame icons because they are not there when a frame window is
first opened. The openbox mantainers claim this is an XEmacs bug, not
their problem.
The attached patch fixes this, but I am not a sufficient (xemacs/)X
wisard to be confident it is otherwise harmless -- make of it what you
will.
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
Half-time member of W3C Team
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 651-1426, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
7:42 a.m.
Henry S. Thompson writes:
The openbox mantainers claim this is an XEmacs bug, not their
problem.
Did they cite a standard for this, by any chance?
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
3:10 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stephen J. Turnbull writes:
Henry S. Thompson writes:
> The openbox mantainers claim this is an XEmacs bug, not their
> problem.
Did they cite a standard for this, by any chance?
Here are a sequence of extracts from the discussion from the openbox
folk:
A few weeks ago we did an investigation with Mikael on IRC and this
behavior happens because XEmacs sets its icon property too late - after
OB sets its own and caches it, thus never rediscovers the correct one.
AFAIK, one possible workaround is to disable icon cache (code change).
Last time we talked, it wasn't known who abuses ICCCM here, OB
or
XEmacs :-(
----------
Reminder:
Run XEmacs - blah, openbox icon.
xprop -remove _NET_WM_ICON
Ta-da, here it is.
----
Ah right, now I remember. So yeah, it is most likely an XEmacs bug. :)
----
It's probably a matter of setting the icon hint before mapping the
window instead of afterward.
"Window managers are free to assume convenient values for all
fields
of the WM_HINTS property if a window is mapped without one." And in
this case, Openbox is assuming it doesn't provide an icon through this
mechanism when it is mapped without one. Thus, it gives it an icon
through the EWMH hints, which overrides the later-appearing icon from
xemacs. There's some race condition possibilities when a window maps
without an icon, and then quickly sets one afterwards.
The quote above comes from ICCCM Section 4 [1].
I could see how you might read all this either way, but I guess I will
run with the patch rather than the post-facto workaround. . .
ht
[1] http://tronche.com/gui/x/icccm/sec-4.html
- --
Henry S. Thompson, School of Informatics, University of Edinburgh
Half-time member of W3C Team
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 651-1426, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
iD8DBQFJhbtkkjnJixAXWBoRArsTAJ49I6pyCcJTlXf1kQfXFU25SufOLQCfZuTj
Z28Lmyp5N3sscocoG0+zIvE=
=h4Fi
-----END PGP SIGNATURE-----
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
1:42 p.m.
Henry S. Thompson writes:
Here are a sequence of extracts from the discussion from the openbox
folk:
> It's probably a matter of setting the icon hint before
mapping the
> window instead of afterward.
This may be standard-conformant, but it's really weak. In particular,
modern applications will often provide dynamic icons that reflect the
state of the window in question. openbox can't deal with that,
apparently.
Could you try loading the frame-icon library, and then find a TeX
file, iconify the frame, then find an awk file, and iconify the frame?
The intent of frame-icon is that those will be different. I suspect
that OpenBox won't notice.
I could see how you might read all this either way, but I guess I
will
run with the patch rather than the post-facto workaround. . .
I see no harm in the change you suggest, but I'm going to document
OpenBox's inabilities in case somebody decides to provide dynamic
icons ....
Further traffic regarding the patch itself will be on the
xemacs-patches list.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
1:57 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stephen J. Turnbull writes:
This may be standard-conformant, but it's really weak. In
particular,
modern applications will often provide dynamic icons that reflect the
state of the window in question. openbox can't deal with that,
apparently.
Hmm. . .
Could you try loading the frame-icon library, and then find a TeX
file, iconify the frame, then find an awk file, and iconify the frame?
The intent of frame-icon is that those will be different. I suspect
that OpenBox won't notice.
Works just fine.
So, maybe the original problem is a race condition. . .
ht
- --
Henry S. Thompson, School of Informatics, University of Edinburgh
Half-time member of W3C Team
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 651-1426, e-mail: ht(a)inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
iD8DBQFJhvvMkjnJixAXWBoRAo3FAJ9zZmzJapPimRhkta3y4J7zhQ4sXwCfXziY
uA7b0jB04KpMFOuoWotrtGY=
=bluf
-----END PGP SIGNATURE-----
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5772
days inactive
5775
days old
4 comments
3 participants
participants (3)
-
Henry S. Thompson -
Stephen J. Turnbull -
Stephen J. Turnbull