Linux
Slackware

xclock -digital and geometry

xclock -d (digital) does not work well with -geometry.

In Openbox (my windowmanager) I was able to get around this by adding to ~/.config/openbox/rc.xml

<application class="XClock">
  <position>
    <x>3515</x>
    <y>1038</y>
    <monitor>1</monitor>  <!-- 0 is primary, 1 is secondary -->
  </position>
  <focus>no</focus>
  <decor>no</decor>
  <desktop>all</desktop>
  <layer>below</layer>
</application>

With settings in rc.xml starting in .xinitrc doesn’t work. Start in ~/.config/openbox/autostart.

xclock -d -chime -update 1 &

Unable to log in my server without password

After following the usual steps (a search for ssh login no password will give plenty of tutorials) I was still unable to login to my second server (which runs ubuntu 22.04).

At first I thought the problem was that RSA was deprecated. So I switched to ecdsa

ssh-keygen -t ecdsa

That did not allow me to login either.

However, when I debugged the login in with:

ssh user@server.com -vv

I noticed that ssh was looking for id_ecdsa in ~/.ssh for the private key. So I copied the file with the private key to id_ecdsa and was able to login subsequently.

I was under the impression that the file with the private key on the local computer could have any name. Perhaps that assumption was wrong.

reload kernel ethernet card module

I don’t know if this works for all modules.

I have a problem with my ethernet card module r8169 for

04:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 15)

It seems to stall or crash. So I do (as root)

rmmod r8169
modprobe r8169

And subsequently restart my network. In Slackware that is (as root)

/etc/rc.d/inet1 restart

Network up again. For now. What’s wrong with the driver/module? I tried a more recent one, same problem. Sigh.

openbox in slackware 15 or current

It seems I need this in .xinitrc:

Reinvoke within a dbus session:

if [ -z "$DBUS_SESSION_BUS_ADDRESS" ] && [ -x /usr/bin/dbus-run-session ]; then
   exec dbus-run-session -- $0 "$@"
fi

https://www.linuxquestions.org/questions/slackware-14/current-notify-send-with-fvwm-or-any-wm-4175706394/

libgtk-3.so: error adding symbols: File in wrong format

I wanted to install Glade (GTK+3 User Interface Builder). However, I’ve got a multilib environment (sort of). Compiling Glade I got this error:

libgtk-3.so: error adding symbols: File in wrong format

This problem has a solution on this page: Glade compile error, could not read symbols: File in wrong format.

The solution from Keith Hedger being:

I have noticed that the odd .pc file or generated configure/makefiles seem to ignore the libdir variable and use /usr/lib ( the usual default ) instead of /usr/lib64, as a quick and dirty workaround you can try

sudo mv /usr/lib /usr/lib.BAK

compile your code and then do

sudo mv /usr/lib.BAK /usr/lib

You shouldn't have to do this very often its just because of some sloppy written makefiles. This works with the glad3 slackbuild by the way.

Worked for me as well.

Patreon account for Slackware

I just found out that Patrick Volkerding wasn’t receiving any money from the Slackware store (store.slackware.com)... for two years?

Via a twitter search I found this post by Eric Hameleers. A post by Patrick himself is here.

Long story short, Patrick now opened a Patreon account:

• https://www.patreon.com/slackwarelinux/overview

I’m not a big fan of Patreon, but for now I’ve joined to be Slackware’s patron number 348.

My wish-list for Slackware 15 is very short:

• please make it easier to add 32bit compatibility
  (I’ve been unable to make it work in 14.2).

...can’t find the file at .../phpmyadmin/export.php

To export use this command:
mysqldump -u USER -p PASSWORD DATABASE > filename.sql

To import database from dump file (in this case called filename.sql) use:
mysql -u username -p password database_name < filename.sql

https://www.knownhost.com/forums/threads/phpmyadmin-database-export.2649/

SSH slow but HTTP not (hint: DNS records)

After my domain name provider switched to a new system, ssh'ing into new subdomains that I had created and edited in the (new) dns editor became very very slow.

Turns out, that while the old system did not create new ipv6 fields for a new sub domain, the new system does - and adds the address of the local server.

Apparently ssh first tries ipv6 if available. (I discovered this using the verbose option ssh -vvv). Only after looking up ipv6 and timing out, it tries ipv4.

So, fill in those AAAA records with the proper ipv6 address!

Note: the record with only @ is domain name proper, without www or another subdomain.

apr_sockaddr_info_get() failed for server

About this error of Apache2:

apr_sockaddr_info_get() failed for server
Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName

Read the documentation about ServerName:

https://httpd.apache.org/docs/2.4/mod/core.html#servername

The ServerName directive may appear anywhere within the definition of a server. However, each appearance overrides the previous appearance (within that server).

If no ServerName is specified, the server attempts to deduce the client visible hostname by first asking the operating system for the system hostname, and if that fails, performing a reverse lookup on an IP address present on the system.

If you run the command hostname, you will possibly see the address Apache is using. If it’s a name that cannot be looked up (check with host your-hostname) this will likely be the reason Apache complains. If you do have a valid server name, one that can be looked up, use that. (Otherwise you should probably be using localhost.)

To change it without rebooting:

hostname new-valid-hostname

Change it permanently: edit /etc/hostname.

I wrote this as a reminder that if you can use a valid domain name, there is no need to edit /etc/hosts (something that is often suggested as a solution).

Pick default device for Pulseaudio for Telegram’s desktop client.

Currently (April 2018) the Telegram Desktop Client only seems to use the default sound device. Perhaps a bug.

Anyway, I had no default device but Telegram picked the wrong one nevertheless, and I could not change that in pavucontrol, nor could I set a default device there (the relevant button seems to be stuck on Set as fallback no option for default).

What worked for me:

1. find the device to make the default (you can use other filters in the second grep of course):

   pactl list | grep alsa_output | grep -i head

2. use your own device with the following command:

   echo "set-default-sink alsa_output.Headset" | pacmd

3. end with:

   pulseaudio --kill ; pulseaudio --start

Most of it from this page.

Edit: the above is not permanent. How to do that is perhaps somewhere in the following text:

• https://www.freedesktop.org/wiki/Software/PulseAudio/Documentation/User/DefaultDevice/