jaunty: Resizing mpg using tcrequant gives bad quality output with lots of artifacts

December 19, 2009 by linos

Running Hardy for a year or so I used tcrequant without any problems. I use the tool to resize mpg files so that they fit onto a dvd. My wife prefers watching our home recorded videos this way …

After upgrading to Jaunty I realized that resized records came with artifacts. It turned out that tcrequant caused it.

In Hardy, I had:


root@jojo:~# /usr/bin/tcrequant -v
tcrequant (transcode v1.0.2) (C) 2003 Antoine Missout

With Jaunty came:


root@jojo:~# /usr/bin/tcrequant -v
tcrequant (transcode v1.0.7) (C) 2003 Antoine Missout

I found no way around and finally ended up with replacing the Jaunty binary by the old Hardy tcrequant binary. Yes – this is ugly …

If you are asking: Why do you use tcrequant anyway ? It’s said to be deprecated in Transcode ! May be – but have you ever compared the run time of avidemux or mencoder with tcrequant ? The latter is far quicker. Obviously quite different algorithms are used. But I’m lucky with the output of tcrequant 1.02 …

Karmic: Happy with vmplayer, no longer searching the web for a any-any patch to get vmware-server running

December 19, 2009 by linos

For a couple of years I used vmware-server on my laptop. This was a hassle with all new kernels, as the vmware-drivers never built out of the box. It was always necessary to patch the driver sources to get them going.

I’m no kernel hacker and sometimes it was difficult to find the right patches on the web. At least once after a Ubuntu security kernel update the vmware-config.pl failed until I again managed to find a better (newer) patch.

Forget about this trouble. Use VMware Player 3.0.0. After a kernel update starting vmplayer gave me:

Hitting “Install” I got another screen informing me nicely about what was going on:

And I was done ! It simply worked !

But I forgot one important new feature of vmplayer 3.0.0: You can create new machines:

Karmic: ApacheDirectorStudio no longer usable and Workaround

December 2, 2009 by linos

I used ApacheDirectorStudio with sun-java5 without problems for at least a year. After changing to Karmic I installed sun-java6 and did a short test of ApacheDirectorStudio. I did not notice any problems.

Today I again ran ApacheDirectorStudio and this time I ran into trouble. No error message, but a button simply did not appear to work.

Being in trouble to do the real work done, I rebooted into my old Intrepid and did what had to be done.

Now I got a second to analyse the problem and finally found an hint on the Download page of ApacheDirectorStudio

Set

export GDK_NATIVE_WINDOWS=1

and then start ApacheDirectorStudio. Now the application works as expected. So the rootcause was not java6 vs. java5, but Gnome !

Karmic: Keyring + Figaro’s Password Manager 2 fpm2

December 1, 2009 by linos

IMHO gnome-keyring has all components for a password manager application. But AFAIK there is no such application using libgnome-keyring0 and gnome-keyring.

But I can’t live without. So I recently checked the Karmic repositories for suitable applications and finally decided to give fpm2 Figaro’s Password Manager 2 a try.

Since then I use fpm2 on a regular basis. It’s fast, it’s straightforward and there are some features I really like.

For each entry you can define a so called Starter-Application. If e.g. you want to put your Launchpad-account into fpm2, you would add https://launchpad.net/ as URL and Web as starter-app. BTW, the default gnome-moz-remote does not really make sense in Ubuntu, because this command is unknown. Simply change the Web starter to


/usr/bin/firefox $a

Click on User and the defined username will be inserted into the cut buffer. Paste it into the webform. Same procedure for password and you’re done.

Xfce 4.6: Change default settings to use workspaces more efficient

November 29, 2009 by linos

I use 10 workspaces to virtually enhance my screen’s size. My personal best practice is to have certain important applications on dedicated workspaces, e.g. thunderbird on no 2, firefox on no 3 and emacs on no 4.

If I get an email with an URL, clicking on that URL will bring up a new TAB in my running firefox. Unfortunately standard Xfce 4.6 in Karmic will switch the firefox window from workspace no 3 to the thunderbird workspace no 2.

I do not like this. I want the apps keep running on the workspace I started them on.

Finally I found this to make things work:


xfconf-query -c xfwm4 -p /general/activate_action -s switch

I really can recommend this – give it a try !

Karmic: Keyboard customizing w/o xmodmap

November 28, 2009 by linos

Using a german keyboard most of the time I have to press <Shift><7> to get a slash. Working in Unix this is not really convenient so I created an suitable .Xmodmap ages ago.

This worked rock solid until evdev was rolled out (e.g. in Hardy Heron if I remember correctly).

From that point of time my settings in .Xmodmap were gone after a Supend / Resume of my laptop.

As usual with Google’s help it turned out, that I was not the only one with this issue. But I could not find a solution.

Finally I changed the systemwide configuration file /usr/share/X11/xkb/symbols/de.

In the section xkb_symbols “basic” I changed the keys AB08 and AB09


key <AB08> { [ comma, semicolon, slash, multiply ] };
key <AB09> { [ period, colon, slash, division ] };

(On my current laptop keyboard AB08 is the best choice, but on a standard keybord its AB09, so I changed both).

For the sake of completeness: The files in /usr/share/X11/xkb/keycodes map keycodes to symbolic keynames. In my case the file evdev contains:


xkb_keycodes "basic" {
...
<AB08> = 59;
<AB09> = 60;
...
};

You may call the binary xev to get the keycode of any key (among other information).

Now on to another configuration task. I use my laptop frequently with a standard PS/2 keyboard, connected to the laptop via a little electronic device which connects to the laptop via USB. lsusb gives me:


Bus 006 Device 002: ID 074e:0001 Digital Stream Corp. PS/2 Adapter

For reasons completely unknown to me the key “less greater bar” (German keyboard, left to y, would normally be keycode 94) is completely dead (LEDs on the PS/2 Adapter do not even flash when pressing this key). Maybe this device was designed for US keyboards in mind, were no such key exists … In my old .Xmodmap I used the left Windows key (keycode 133) to replace the dead one.

Unfortunately keycode 133 is bound to the key , which is a modifier key in karmic. Run xmodmap -pm to check. I reconfigured the key to emit “less” and “”greater”, but pressing <LWIN> + <AltGr> I never got “bar”.

Finally I changed /usr/share/X11/xkb/keycodes/evdev to read:


// Microsoft keyboard extra keys
// <LWIN> = 133;
<AE13> = 133;
...
//// <AE13> = 132; // Yen

Again in file /usr/share/X11/xkb/symbols/de I added:


default
xkb_symbols "basic" {
...
key <AE13> { [ less, greater, bar ] };
...
};

This configuration works as fine as my old .Xmodmap did also after a Suspend + Resume. Unfortunately it took me quite some time to get this working. From a user’s perspective this stuff is not really well documented. Of course my changes will be eventually overwritten by an update.

If somebody knows how to make my changes on a per user basis, please share.

Xserver issues: 96 dpi vs. 133 dpi, nv vs. nvidia

December 22, 2008 by linos

Since some weeks I occasionally run Intrepid Ibex on my Lifebook (booting from my test partition). I installed the box with defaults only. Having
lulu:/home/bav> lspci | grep -i nvidia
01:00.0 VGA compatible controller: nVidia Corporation GeForce 8400M G (rev a1)

Intrepid used the free nv-Driver:

lulu:/home/bav> less /var/log/Xorg.0.log
...
(II) LoadModule: "nv"
(II) Loading /usr/lib/xorg/modules/drivers//nv_drv.so
...

This worked, but surprisingly I got


(==) NV(0): DPI set to (96, 96)

which is simply wrong. The Lifebook has 133 dpi.

As I could not fix this, I quickly installed the nonfree NVIDIA binary Xorg driver, package nvidia-glx-177.

With this driver, I got correctly 133 dpi. Everything seemed fine …

I then realized, that changing the screen’s brightness no longer worked. It worked fine as long as I used the free nv driver. Now with the nonfree NVIDIA driver, the screen did not change brightness any longer.

Unfortunately there are situations, where the Lifebook’s screen stays dark (e.g. maybe after disconnecting from docking station). As I was not able to switch the screen brighter I had to reboot in this situation :-(

I uninstalled nvidia-glx-177 and started over with nv. Changing values in /proc/acpi/video/VGA/LCD/brightness worked as seen before but I was not able to find the right place to start X with 133 dpi.

Finally I got the tip to use Xfce 4 Settings Manager | User Interface | DPI. It had the value System Default. After I changed it all XFCE and Gnome applications were fine. Firefox and Thunderbird too.
Some other (e.g. emacs) still need some fine tuning.

I have no idea what xfce does here behind the scenes. Obviously the X-Server stills runs with 96dpi …

Complex systems tend to appear full of mystery …

Exploring the Asus EEE PC 900A with Xandros Linux

September 17, 2008 by linos

If you are interested in that topic, you might like to visit

http://eeex.ohost.de

Update: My EEE is broken, so currently no news on that page right now.

Update2: I finally got a new EEE, it’s running now with Intrepid Ibex.

Update3: If you run your EEE 900A with BIOS version 0403 (08/04/2008) you really should consider to update to 0607 (09/24/08). See above site for details.

sudo Trouble: Always asking for password despite NOPASSWD

May 26, 2008 by linos

Yes, I’m too lazy: I love NOPASSWD in my sudoers. Yes, this is bad security practice. Anyway, after installing Hardy I couldn’t make it work.

Finally I found that sudo takes the last appropriate entry that matches, not the first, as other tools do.

So after moving my custom sudoers rule to the end of /etc/sudoers, I now can continue my bad habits …

gnome-screensaver stays always black: lots of segfaults

May 26, 2008 by linos

Running xfce on my new Lifebook, I was completely locked out after I ran xflock4 (which in turn calls gnome-screensaver-command –lock).

Back to the prompt I saw a lot of segfaults reported by the kernel like this one:

May 19 12:54:07 lulu kernel: [ 1166.757433] gnome-screensav[8313]:
segfault at 31 rip 7f8a51b1c8d4 rsp 7fff5b3d9a10 error 4

Fortunately using xscreensaver instead of gnome-screensaver is a work around.