LMDE BREAKAGES - tracking TESTING

[karashata]

This is Turpial. It's a microblogging client (eg. Twitter) written in Python.

As for the issue kruijf is having, I use Turpial and I haven't run into issues with it logging me out when I open links, so I wouldn't be able to say what could be causing something like that.

[kruijf]

I'm not sure that the problem is whit links only in Turpial, maybe also in Thunderbird or some other program.
I only had it for now whit Turpial.
But it is whit opening a link, the browser i use is Chrome stable.

[zerozero]

I only had it for now whit Turpial.

hi kruijf,
i see from your sig that you're using latest (and as we all know it hasn't been updated in quite some time, so i suspect that this issue is either from turpial or chrome)
what v. of turpial are you using? where did you get it from? the standard repos or elsewhere?
just for the record the v. available in latest/incoming is

Code: Select all

apt-cache policy turpial
turpial:
  Installed: (none)
  Candidate: 1.5.0-1
  Version table:
     1.5.0-1 0
        500 http://debian.linuxmint.com/incoming/ testing/main amd64 Packages

more, did that started to happen after you update either one of those (turpial/chrome) or is unrelated with any update?

[kruijf]

Thats my fault, a had a switch back to testing.
I will change it in my sig, i forgot that

[zerozero]

kruijf,
no probs
in testing the v. you are running should be

Code: Select all

apt-cache policy turpial
turpial:
  Installed: (none)
  Candidate: 1.6.7-1+ds1-1
  Version table:
     1.6.7-1+ds1-1 0
        700 http://ftp.debian.org/debian/ testing/main amd64 Packages

and for this v. i find a couple bugs (none similar to yours though )
http://bugs.debian.org/cgi-bin/pkgrepor ... t=unstable

these crashes only started after the 16.01.2012 (when this new v. migrated to testing?)
http://packages.qa.debian.org/t/turpial.html
or can you related them with a chrome update?

[kruijf]

whit me it looks like this

Code: Select all

apt-cache policy turpial
turpial:
  Installed: 1.6.7-1+ds1-1
  Candidate: 1.6.7-1+ds1-1
  Version table:
 *** 1.6.7-1+ds1-1 0
        500 http://ftp.us.debian.org/debian/ testing/main i386 Packages
        100 /var/lib/dpkg/status

But it looks like the problem is gone, a few day's now whit out the problem.
So what it was i have no idea.
Thnx for thinking whit me zerozero.

[zerozero]

you're welcome

[kruijf]

you're welcome

Its back, and now even when i use a link in Chrome.
So i think the problem is whit chrome and not whit turpial, but now to find what the problem is.

[nimbvs]

After today's kernel update to 3.2, I got the error:

Code: Select all

Examining /etc/kernel/header_postinst.d.
run-parts: executing /etc/kernel/header_postinst.d/dkms 3.2.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.

I traced it back to a missing BUILT_MODULE_NAME[0]="$PACKAGE_NAME" in /usr/src/ndiswrapper-1.57~rc1/dkms.conf. But now, after adding BUILT_MODULE_NAME to the dkms.conf file, I found that ndiswrapper doesn't build anymore. It seems the version currently in testing doesn't work with kernel 3.2, see Debian Bug #655223. It fails with the following:

Code: Select all

/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c: In function ‘setup_tx_sg_list’:
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:472:10: error: incompatible type for argument 2 of ‘pci_map_page’
/usr/src/linux-headers-3.2.0-1-common/include/asm-generic/pci-dma-compat.h:43:1: note: expected ‘struct page *’ but argument is of type ‘struct <anonymous>’
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c: At top level:
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:1780:2: error: unknown field ‘ndo_set_multicast_list’ specified in initializer
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:1780:2: warning: initialization from incompatible pointer type [enabled by default]
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:1780:2: warning: (near initialization for ‘ndis_netdev_ops.ndo_neigh_setup’) [enabled by default]
make[3]: *** [/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.o] Error 1
make[2]: *** [_module_/var/lib/dkms/ndiswrapper/1.57~rc1/build] Error 2
make[1]: *** [sub-make] Error 2
make: *** [all] Error 2
make: Leaving directory `/usr/src/linux-headers-3.2.0-1-686-pae'

Not needing ndiswrapper, I removed ndiswrapper-dkms, ndiswrapper-utils-1.9, ndiswrapper-common and ndisgtk packages.

Also, broadcom-sta-source doesn't compile on the new 3.2 kernel, see Debian Bug #656600. For now, until the updated package lands in testing, I had to extract the archive /usr/src/broadcom-sta.tar.bz2, modify the file src/wl/sys/wl_linux.c following the patch, recreate the archive, and then m-a -l 3.2.0-1-686-pae a-i broadcom-sta worked fine.

Moved here by moderator.

[gozer]

Just updated and the 3.2 kernel showed up and proceeded to break nvidia.

Booted up 3.1 until I find a fix or one comes down from the mothership...

UPDATE:
fixed by reinstalling the nvidia kernel packages:

Code: Select all

aptitude reinstall nvidia-kernel-common nvidia-kernel-dkms

[Gerd50]

Ahh, same problem here with kernel 3.2 and nvidia gozer. Thanks for the fix information

[Lendo]

that didn't work for me.. just broke kernel 3.0.0-1-amd64

[zerozero]

strange, Lendo, it worked for me as well

just upgrade to 3.2 and as gozer reported nvidia broke, with that fix alls well again

[Lendo]

I tried it .. Guess I'm not supposed to have 3.2

[zerozero]

Lendo, one thing i don't understand, how did it broke 3.0.0-1? didn't you have the 3.1 installed as well?

[Lendo]

Nope, last kernel I got was 3.0.0-1-amd64

[nimbvs]

This got posted in the wrong thread; I posted it in the Sid thread, but I should have posted it in here, in the testing thread. So here it is:

After today's kernel update to 3.2, I got the error:

Code: Select all

Examining /etc/kernel/header_postinst.d.
run-parts: executing /etc/kernel/header_postinst.d/dkms 3.2.0-1-686-pae
dkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0.
Error! Bad conf file.
File: 
does not represent a valid dkms.conf file.

I traced it back to a missing BUILT_MODULE_NAME[0]="$PACKAGE_NAME" in /usr/src/ndiswrapper-1.57~rc1/dkms.conf. But now, after adding BUILT_MODULE_NAME to the dkms.conf file, I found that ndiswrapper doesn't build anymore. It seems the version currently in testing doesn't work with kernel 3.2, see Debian Bug #655223. It fails with the following:

Code: Select all

/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c: In function ‘setup_tx_sg_list’:
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:472:10: error: incompatible type for argument 2 of ‘pci_map_page’
/usr/src/linux-headers-3.2.0-1-common/include/asm-generic/pci-dma-compat.h:43:1: note: expected ‘struct page *’ but argument is of type ‘struct <anonymous>’
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c: At top level:
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:1780:2: error: unknown field ‘ndo_set_multicast_list’ specified in initializer
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:1780:2: warning: initialization from incompatible pointer type [enabled by default]
/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.c:1780:2: warning: (near initialization for ‘ndis_netdev_ops.ndo_neigh_setup’) [enabled by default]
make[3]: *** [/var/lib/dkms/ndiswrapper/1.57~rc1/build/wrapndis.o] Error 1
make[2]: *** [_module_/var/lib/dkms/ndiswrapper/1.57~rc1/build] Error 2
make[1]: *** [sub-make] Error 2
make: *** [all] Error 2
make: Leaving directory `/usr/src/linux-headers-3.2.0-1-686-pae'

Not needing ndiswrapper, I removed ndiswrapper-dkms, ndiswrapper-utils-1.9, ndiswrapper-common and ndisgtk packages.

Also, broadcom-sta-source doesn't compile on the new 3.2 kernel, see Debian Bug #656600. For now, until the updated package lands in testing, I had to extract the archive /usr/src/broadcom-sta.tar.bz2, modify the file src/wl/sys/wl_linux.c following the patch, recreate the archive, and then m-a -l 3.2.0-1-686-pae a-i broadcom-sta worked fine.

[viking777]

Just updated and the 3.2 kernel showed up and proceeded to break nvidia.

Booted up 3.1 until I find a fix or one comes down from the mothership...

UPDATE:
fixed by reinstalling the nvidia kernel packages:

Code: Select all

aptitude reinstall nvidia-kernel-common nvidia-kernel-dkms

Thanks for the warning gozer - same problem/solution here.

[Chris M]

Just updated and the 3.2 kernel showed up and proceeded to break nvidia.

Booted up 3.1 until I find a fix or one comes down from the mothership...

UPDATE:
fixed by reinstalling the nvidia kernel packages:

Code: Select all

aptitude reinstall nvidia-kernel-common nvidia-kernel-dkms

Thanks for the warning gozer - same problem/solution here.

viking, are you still using smxi-sgfxi?

When I went from nouveau to the nvidia driver a while back, I used smxi-sgfxi, and it worked like a charm. But nothing nvidia showed up in Synaptic. So when I tried to do a preemptive strike before re-booting into the 3.2 kernel with the aptitude reinstall nvidia-kernel-common nvidia-kernel-dkms, the following made perfect sense:

Code: Select all

chris@9150:/$ sudo aptitude reinstall nvidia-kernel-common nvidia-kernel-dkms
[sudo] password for chris: 
nvidia-kernel-common is not currently installed, so it will not be reinstalled.
nvidia-kernel-dkms is not currently installed, so it will not be reinstalled.
nvidia-kernel-common is not currently installed, so it will not be reinstalled.
nvidia-kernel-dkms is not currently installed, so it will not be reinstalled.
No packages will be installed, upgraded, or removed.
0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 0 B will be used.
                                         
chris@9150:/$  

When I re-booted after that, I booted into a login prompt. Maybe I should have re-booted again, but what I did was to run smxi-sgfxi, and reinstall the current nvidia driver. It worked.

So again, are you still using smxi-sgfxi, because I don't even have nvidia-kernel-common and nvidia-kernel-dkms installed.

Also, when I originally ran apt-get update && sudo apt-get upgrade && sudo apt-get dist-upgrade, it did not give me the new 3.2 kernel. I had to specifically install it through Synaptic. Then I ran smxi-sgfxi. Is not having anything nvidia show up as installed in Synaptic normal? I didn't think so at the time, but it's never been an issue.

Edit

Perhaps a meaningless edit. But if I've confused anyone, I just remembered that when I went to the nvidia driver from nouveau, I actually did the standard dkms nvidia install first. That failed, so I removed all non-free nvidia files, and then used sgfxi. That's why there was no nvidia-kernel to reinstall/upgrade.

[viking777]

So again, are you still using smxi-sgfxi,

Well remembered Chris! I did use sgfxi for quite a while, but I switched back to dkms a couple of days before this problem occurred - ironically to prevent me having to run sgfxi every time I got an xorg, nvidia or kernel update