Monthly Archives: March 2010

Slackware-current: the road to 13.1?

Here are some issues I’ve had with Slackware -current (will probably become 13.1).
I’m using this post as a reminder of what happened to my box, and what I did to fix what I didn’t like; it will be updated until slack-13.1 is released; some points are entirely my (or my hardware’s) fault (see kernel and boot sections below); anyway, this is actually a praise to Slackware-current, since it is stable, and is getting better and better.

Gvim [fixed]

$ gvim
** (gvim:21033): CRITICAL **: gtk_form_set_static_gravity: assertion `static_gravity_supported' failed

Gvim needed upstream patch #257

Boot hiccups [fixed]

Boot is slowed down because something (udev? hal?) tries to mount /dev/sr0 (my dvd-rw). It somehow happened in Slackware 13.0 too, but it was not interfering with the rest of the startup process. Now it tries to mount it 3 times, and everytime it stops for I think 10 seconds.

UPDATE: This may be caused by my custom kernel (see below). I have to try this (needed by udev 151):


…I’ll do as soon as is released (yes, I am lazy).

UPDATE 2: kernel with my own config (following udev instructions) didn’t do the trick. The bootup process stills stops for ~15 seconds twice, trying to mount my dvd-rw.
This is the diffconfig:

./diffconfig config-generic-smp- <(zcat /proc/config.gz)
-X86_CMPXCHG64 y
EXT2_FS m -> y
EXT3_FS m -> y
JBD m -> y
LOCALVERSION "-smp" -> "-custom"
M686 y -> n
MK8 n -> y
UEVENT_HELPER_PATH "/sbin/hotplug" -> ""
X86_GENERIC y -> n

UPDATE 3: I must be pretty dumb: the errors at boot were something like:
mount: no medium found on /dev/sr0 ...trying again
To fix the delay, I just had to change /etc/fstab from
/dev/cdrom /mnt/cdrom auto auto,owner,ro 0 0
/dev/cdrom /mnt/cdrom auto noauto,owner,ro 0 0
… now I really wonder why /etc/fstab got that line wrong (and why the problem didn’t show up in Slackware-13.0)…

A new kernel [fixed]

When the kernel loads the module of my audio device, the pc stalls: reboot needed. A patch is queued somewhere. Hope it lands in stable soon.

UPDATE: The fix should be in

UPDATE 2: The fix is in and it works.

Kde4 autostart [fixed]

Kde4 is really getting better, but at startup it loads wicd-gui, even if the daemon is not running and I don’t need it. Ok, I’ll probably begin using wicd sometime in the future anyway, but right now it is annoying. Obviously, there’s nothing in ~/.kde/Autostart, so I’m at a loss, don’t know what to do.

UPDATE: found a solution here.

Graphic stack

The current xorg-mesa stack is great for my radeon. The desktop effects work well, with no crashes, while on -stable it was better having them turned off. I still don’t know if I’m supposed to run a KMS-enabled kernel or not. Currently, I’m not. My guess is: use it for HD3XXX and later.

To enable kernel modesetting, pass radeon.modeset=1 at LILO’s prompt, or add append="radeon.modeset=1" to the proper LILO section.

Even if it gets better everyday, there’s still something missing with some GLX extension. That would probably need to wait for xorg-server-1.8.x, though.

Oh, and there’s a visual glitch: it is noticeable if you open a long page in Firefox and then scroll it up and down fast with the mouse: if you have a dark background on the right of the window, you’ll see some white horizontal lines coming and going. Perhaps a mesa 7.7.1 and an xf86-video-ati 6.13.0 upgrade would solve this ugly effect.

UPDATE: The updates on April 27, 2010 fixed some glitches. Sometimes there still are visual artifacts, especially when using kwrite, but the situation has improved.

Kvm [fixed]

Boot message with Kvm compiled from

syslog:Mar 4 14:08:11 darkstar udevd[786]: NAME="%k" is superfluous and breaks kernel supplied names, please remove it from /lib/udev/rules.d/65-kvm.rules:1

Not really Slackware-current’s fault ;^)
(And, for what I can see, does not affect my simple use case).

UPDATE: This is now fixed at

No worthy mechs around? [fixed?]

Don’t know if this is new or was present in 13.0 (or older) as well, but here it is.
When I download a file from my web server, I get this in /var/adm/messages:

Mar 24 17:58:25 darkstar perl: No worthy mechs found

It means: do a

chmod +x /etc/rc.d/rc.saslauthd

… or probably some configuration option in /etc/httpd needs some love. I’d really like not to have that warning in the logs, nor running an otherwise unused service.

mplayer [fixed]

mplayer changed his behaviour between 13.0 and -current in at least a couple of ways.

For example, the screensaver wakes up while I’m seeing a movie, never did that
before. Solution: just add

# no screensaver

in ~/.mplayer/config

Also, it seems there are some .flv files that mplayer plays badly. Solution: add this in ~/.mplayer/config:

# mplayer in -current (2010-03-10) plays badly some .flv

…and don’t forget the final comma. That should either not play the video at all, or use another codec. If the video does not play at all with this configuration, you may want to use a SlackBuild script to install other codecs.

Akonadi/KMail [fixed]

I was really pissed when I started KMail and an error popped up saying that Akonadi couldn’t run because of something like “no resources found” and telling me that I needed to properly set the XDG_DATA_DIRS environment variable.

I actually spent a good lot of time into this since the software stack needed to run KMail is something like: KMail -> KAddressBook -> Akonadi -> Nepomuk -> MySQL. If anything goes wrong, you are left with the short stick and are on your own.

The thing is, Slackware can’t be wrong. And neither Pat can. :^)

I don’t know why, but my system was missing one package: kdepim-runtime-4.4.3-i486-1. I guess slackpkg didn’t pick it up. To solve the problem, I just did:

# slackpkg install kdepim-runtime

UPDATE: Probably slackpkg was tricked because the changelog said “Updated” instead of “Added” for the kdepim-runtime-4.4.3-i486-1 package entry.