(...)reuse the same window as I navigate around the filesystem.
Well, that's not spatial at all. What you want is browser mode without the toolbars. Well, it's very easy to get that: open a browser mode window, go to View->Main Toolbar and View->Location Bar and deactivate them! There you go ...

Thanks, compiz-quinnstorm worked for me! I thought I had tried it, but maybe I was dreaming.

Reported on the xgl-coffee forums at http://forums.xgl-coffee.org/viewtopic.php?p=143#143

I'm using the xgl-coffee overlay, and with the latest updates, compiz (vanilla) keeps crashing for me whenever I run a program on the display managed by compiz. (like gnome-terminal, or even gnome-window-decorator)

Here's the backtrace:
#0 0xb7ebd65b in glXBindTexImageEXT ()
from /usr/lib/opengl ...

Raffo wrote:what should I do?

The latest hanno overlay has a patch to fix that issue.

Belgabor wrote:It works fine on my Geforce4...

It's not Geforce4 MX, is it?

More specifically, is GL_ARB_texture_non_power_of_two present when you do a glxinfo on the Xgl server?

[Edit: and thanks to wintersommer for posting in the nVidia Linux Forum, let's see where that goes ]

redmatrix wrote:For nvidia users, I started a thread in the nvidia linux forum, (...)

redmatrix, could you see this post: http://forums.gentoo.org/viewtopic-p-31 ... ml#3108233
and if you, or anybody else, could do as suggested at the end, I'd be eternally grateful!

A recap of what we know about the black windows problem:

This problem was already raised in the xorg mailing list[1], but with no solution.

To expand a bit on that message, here's the snippet of code that is run by compiz:

From compiz/src/texture.c +218
case GLX_TEXTURE_2D_EXT:
texture->target ...

dllloader must be used by the modular X.org so the use flag does nothing now, dllloader is always enabled.

CoronaLover wrote:Am I the only one who STILL get the black windows bug?

No, I think everybody with nVidia Geforce 4 MX or similar cards have that problem.
Well, at least I do...

Hey just put up the start of a small howto at http://gentoo-wiki.com/HOWTO_XGL
please contribute with your knowledge.
Starting scripts, more troubleshooting tips, anything will do.

(...)so all the fancy effects are done in software not on your gfx card(...)
Correct me if I'm wrong, but the only part of this stack implemented by mesa is the conversion of a pixmap to a GL texture, afterwards all "fancy effects" are simple GL transformations and are in fact accelerated. That's ...

Is anyone still having the infamous black windows problem?

I got the new overlay from hanno boeck, and followed all the steps, Xgl runs fine but still when I try to run compiz, the windows turn black, and it outputs something like this:
compiz: pixmap 0x2000d3 can't be bound to texture
compiz ...

To everyone who gets "GLX_EXT_texture_from_pixmap is missing", please post your emerge info.
Why emerge info? The problem is that you emerged Xgl with the old patches where the functions are called GLX_ MESA _texture_from_pixmap. Either patch Xgl or compiz so that the function names are the same.

Did anyone solve the: dlopen: /usr/lib/xorg/modules/xgl/libglx.so: undefined symbol: mglNewList problem that some people were describing?

I saw this mail: http://lists.freedesktop.org/archives/xorg/2006-February/012790.html (sent by Hanno Boeck, who made the original XGL guide) and was expecting ...

any fix (or cvs update) to glxcompmgr "black windows"?

I've been looking for a list of commits or a recent viewCVS of mesa and XGL, and I didn't find any.. does anyone know if there is such a thing?

I've been updating the mesa, xgl and glxcompmgr by emerging them again, which I didn't want to do ...

For anyone browsing for aisleriot in gnome-games:

The gnome-games ebuild only installs aisleriot if the use flag guile is in use.

I think the POSIX locale "C" is usually in en_US. A quick check sows that:
$ ls /usr/share/locale/en_GB/LC_MESSAGES/* | wc -l
103
ls /usr/share/locale/en_US/LC_MESSAGES/* | wc -l
1
Which means there is a significative en_GB translation, but no significative en_US translation, supporting my theory ...

I think this is http://bugs.gentoo.org/show_bug.cgi?id=109642
which is already solved by a new version in portage

With tex'ed documents, however, scrolling is glitchy and not really comfortable.

Did anyone else experience this?
I have an issue with scrolling pdf documents in which pages are just images (i.e. page scans). Each scroll causes 100% CPU usage for several seconds.

I think this is caused by:
1 ...

What i wish to know, is, where the <new icon & new size for the desktop> are saved ? might be an xml file but i can't put my hands on it.
Nautilus saves metadata in ~/.nautilus/metafiles

Files there are created by replacing / with %2F, i.e.:

.nautilus/metafiles/file:%2F%2F%2Fhome%2Fkrinn.xml ...

Somehow I got his problem too. I created a ~/.inputrc with this: That solved the problem for me.

There is already one line like that in /etc/inputrc, but it's inside an $if mode=emacs.
Copying the line out of the if should make it work for all users.

It seems hald is running, and you're just trying to start another one. you can confirm by running
ps aux | grep hal
The devices not showing is a 2.12 bug, see for example, this thread:
http://forums.gentoo.org/viewtopic-t-388078.html

One workaround discussed on the main gnome 2.12 thread is ...

Same problem here.. there's no solution yet on that thread. Maybe we can get one here.

What we know so far:

Now that gnome-volume-manager uses pmount, there's no need for adding lines to /etc/fstab. Adding oneself to the plugdev group is necessary for g-v-m to (p)mount the devices, and if a device ...

maKKus wrote:Am I the only one with this problem?

Nope, same here.
Adding myself to plugdev group didn't solve the problem, although it must be a necessary step.
Still looking for the solution..

It probably is /dev/psaux