lxde fails to open properly after upgrade

S

Sparky

I first installed v9.1 on a eeePC 900A a couple of months ago and have been trying to update everything the last few days. I ran [cmd=]portmaster -afD[/cmd]. I added the -f option after many errors as per Handbook 5.6.5.3. After that completed successfully I ran [cmd=]portmaster -L[/cmd] which reported there was nothing further to do. And rebooted.

As usual I ran [cmd=]startx[/cmd] My .xinitrc file is one line: exec /usr/local/bin/startlxde

But instead of the desktop opening as normal, the panel at the bottom was just a blinking white band. I went back to the console and saw this streaming down the screen until I hit ctrl-c:
Code: 
pid 2394 (lxpanel), uid 1001: exited on signal 6
GLib (gthread-posix.c): Unexpected error from C library during 'pthread_mutex_unlock': Operation not permitted. Aborting.

How do I fix this?
 
Oops, neglected xorg log. Here's a grep EE /var/log/Xorg*:

Code: 
(EE) AIGLX error: dlopen of /usr/local/lib/dri/swrast_dri.so failed (Cannot open /usr/local/lib/dri/swrast_dri.so)
(EE) GLX: could not load software renderer
 
It's not an xorg error. First guess is that it was because of skipping something important in /usr/ports/UPDATING in the time between the initial install and the rebuild of everything.
 
Sparky said:
Yow! That's a 1/4M file. Can you suggest a method of searching for the offending port?
Click to expand...

That's not the way to read it. Note the dates. The most recent stuff is first. Normally, you look back through it to the last date you installed or updated a port. Often, you can get an idea of when that was with % ls -ltr /var/db/pkg.

Of course, now all those dates are from when you rebuilt everything. So pick an estimated time, go back in /usr/ports/UPDATING to that point, and work forward.

Only do the steps that apply to your system. If it says to upgrade a port you've never heard of, check whether it has been installed first.
 
Here's the only instance of graphics/dri in UPDATING. But I'm using neither Mesa3D nor x11-drivers/xf86-video-nouveau.

20100207:
AFFECTS: users of Mesa3D libraries and x11-drivers/xf86-video-nouveau
AUTHOR: nork@FreeBSD.org

If you want to use Mesa3D 7.6.1 and libdrm 2.4.17 rather than 7.4.4
and 2.4.12, you must define WITHOUT_NOUVEAU global macro, at least,
enabled on graphics/libGL*, graphics/libglut, graphics/dri,
graphics/mesa-demos, and graphics/libdrm. And please give up using
x11-drivers/xf86-video-nouveau.

At this time, I cannot enable latest Mesa3D and libdrm, because they
break xf86-video-nouveau. But old (current?) Mesa3D and libdrm do not
break any drivers.

AMD Radeon HD 2xxx/3xxx/4xxx users: If you use AMD Radeon HD [234]xxx
series, please define WITHOUT_NOUVEAU global macro. You can then use
OpenGL Hardware Accelerator feature on these series.
Click to expand...

Nevertheless put
Code: 
 WITHOUT_NOUVEAU=YES
into /etc/make.conf and ran # portmaster graphics/dri. This failed at the message
==> pkg-1.0.9_1 is already installed You may wish to make ''make deinstall'' and install this port again by ''make reinstall''
Click to expand...
This lead to more errors as did the suggestion to append
Code: 
FORCE_PKG_REGISTER
with
Code: 
usr/ports/ports-mgmt/pgk #make install
which failed with
Code: 
***[check-vulnerable] Error code 127
 
Please re-read post #6. There is no reason to attempt anything from UPDATING before the date of your install, and 9.1 has only been out for three months.
 
Unfortunately I hadn't read #6 in time and had already tried #5's suggestion. Which seems to have been a step backwards.

The original errors I had with the first try using portmaster had to do with Cairo and Pango. http://pastebin.com/PaBER6VV

Pango isn't mentioned in UPDATING nor % ls /var/db/pkg/pango* but Cairo is. Am I on the right track? Or does the original complaint have nothing to with the later use of the '-f' flag?
 
Sorry to be misleading, I’m just a beginner willing to help. On the other hand, you did more than simply updating graphics/dri… Anyway, from now on, I’ll let the experts help and try to learn in the process.

Best regards,
Juan
 
Now you also have installed pkgng and then it was deleted? There are so many things broken in so many different ways that I have no idea how to start fixing it.
 
I did most of the damage in a terminal while the desktop was still working. I notice the console history is just commands entered in the console. Is there a way to look at the desktop/terminal history?
 
There's .history. That will probably not help. If you can get % portmaster --list-origins to work, that is the list of ports that need to be installed to recreate everything. Then delete all installed ports, remove pkgng changes, and reinstall that list using portmaster.
 
% portmaster --list-origins Just caused a long stream of
Code: 
 ===>>> No ORIGIN in /var/db/pkg/<pkg>
Seems to include everything under /var/db/pkg

Maybe re-boot? IIRC, there's an repair option in the ASCII demon. Thing is, I don't have a clue how to get to the web by hand: (which is what I believe awaits if I use that method).
 
No, there is nothing in the boot menu that will repair ports. I don't know how you managed to get to where you are. It seems you have been using pkgng, but somehow deleted parts of it. Now you can't add packages or build ports, which really cuts down on the options. If it were me, I'd make a backup. Short-term goal: get ports-mgmt/pkg reinstalled. Long-term: deinstall every port and install them all again.
 
wblock@ said:
If it were me, I'd make a backup. Short-term goal: get ports-mgmt/pkg reinstalled. Long-term: deinstall every port and install them all again.
Click to expand...

Studying 19.12 now. It's going to be a while as I'm being called away to other duties. But I will keep the forum posted. Would love to amend this thread to [SOLVED] :)
 
You must log in or register to reply here.
Back
Top