show app icon in the title bar

[Brian J. Tarricone]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Olivier Fourdan wrote:
> You may want to take a look at NVidia forums here
> 
> http://www.nvnews.net/vbulletin/forumdisplay.php?s=&forumid=14
> 
> Next version are said to be usable (as usual), but I doubt it.
> 
> The only driver that worked was 6629. Anything after crash (and by
> crash, I mean 100% CPU utilization)

Hmm, I might give 6629 a try, though I'm not sure the eye-candy is worth
the effort (probably not).

> Some very similar reports in that thread:
> 
> http://www.nvnews.net/vbulletin/showthread.php?t=54863&page=9

Yeah, it's a mess.  I suppose it's a bit much to expect a
known-not-quite-stable binary driver to work properly with an
experimental X extension that has problems even under the best of
conditions.

> * xfwm4 works w/out shadows enabled. The problem seems to come from the
> shadow.
> 
> I've changed to code so that shadows get disabled by default.

Interesting.  I'll have to try it, but tomorrow.  I should be in bed
already...

> BTW you said "that's unfortunate, since it's mostly impossible to prove
> it either way." 
> 
> I strongly disagree with you here. On one side, you have xfwm4, a fully
> open source project with open SVN and developpers avaiable to reply,
> while in the other side, you have a proprietary driver with no source,
> and developpers totally unreachable.
> 
> That's the strngth of open source...

Well, I'll submit that yes, it's impossible to prove that the nvidia
driver is *not* the problem, as we don't have the ability to verify its
correctness.  (Amusingly, it's hard to prove that it *is* the problem,
at least not without verifying everything else is correct.)
<sarcasm>Yay closed source.</sarcasm>

However, simply saying "I don't think anything is wrong with my code"
doesn't necessarily prove that it's not xfwm4's fault either.

Then again, the 100% CPU util is in X, and not xfwm4, so it seems more
likely that something in the X server has gone nuts, unless of course
xfwm4 is executing a bit of code over and over that causes X to eat CPU
for breakfast (lunch, AND dinner).  Whether it's xfwm4/X exposing a bug
in nvidia, or nvidia exposing a bug in XComposite, who knows.  I
(personally) still wouldn't rule out a problem with xfwm4, but I rate
that as several orders of magnitude less likely than a problem with
nvidia or XComposite, which I figure either is just as likely.

	-brian

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)

iD8DBQFDZ0MK6XyW6VEeAnsRApmgAKDJ1445Jim3NGvYz/qzMntSrczAhgCfVGZA
rXAfp6kDnMbvYjQIv3bwUJI=
=Qy8S
-----END PGP SIGNATURE-----