A view from the hill

When I got a Handspring Visor (a Palm Pilot offspring) roughly ten years ago, I was amazed about all the nice things you could do with the device. I even wrote a German version of the visor linux howto, although I never really finished it. What I especially liked was the possibility to install a ssh client and use it to connect on the push of a button to my workstation -- which I took to good use several imes when my Xserver went beserk. Unfortunately, over time the major incentive to use the thing, syncing calender and address book vanished, as it turned out to be a major hassle when the main calendar to sync with was Lotus Domino. So, since some years my trusty Visor has only gathered dust. But even back in 2000, it was clear to me that the Palm Pilot would only be the starting point and that at some time in the then-unknown future, we would have small devices that would allow issuing phone calls, taking pictures, listening to music as well as providing the "usual" features of handhelds.

Back then, I didn't really think about when exactly that device would arrive but I think the time has come finally. Some days ago I bought a Nokia N900, which is a mobile device running Maemo on top of a Linux kernel on an ARM platform (see this overview of the hardware for more details). This is really a neat little gadget which is really as close to my ten year old dream as you can get: It comes with mobile phone capabilitiy, UMTS and wlan connection, 5MP camera, media player, RSS feed reader, etc. and also has a quite usable sliding keyboard. It's not really a mobile phone but a pocket computer which you can also use to call people with, which shows for instance in that most applications require the device to held in landscape position.

Of course, what's most interesting to me is the software on the device. Basically, Maemo 5 is a linux distribution based on Debian technology, e.g. apt/dpkg and friends work as expected. You don't get all of your favourite linux applications of course, but still there are quite a lot of the regular stuff you would expect. This includes the mobile version of Firefox, Fennec, vim, OpenSSH (client and server) and even Emacs. Being the geek that I am, nearly the first thing I did was a remote login via ssh from my workstation on the N900 (also see this overview of USB networking on the N900). Native maemo applications are typically developed with Python and GTK, although that is probably going to change for Maemo 6 (Nokia has announced it favours Qt over GTK). The community seems quite active, too: just shortly, a port of Clisp (a Common Lisp compiling to byte code) was announced (including a running Swank backend, so you can connect from your development Emacs remotely), although I imagine that the development environment for Python is, generally speaking, more advanced (for Maemo development, this is). I haven't looked any closer into developing for Maemo up to now, but it surely looks interesting.

ObTitle: Gossip, "Music for men"

When I upgraded lately, my old fbpanel configuration was broken. I basically stumbled over Debian bug #288594, my old configuration as well as the system default pointed to a no longer existing /etc/fbpanel/menu file. So, my fbpanel no longer showed a menu. Sure enough, my WindowMaker has its own menu, so it was not much of a hassle. The broken menu entry (I just edited out the reference in my .fbpanel/default) annoyed me nethertheless, so I gtfn[*] today and found the missing info: in order to include the system menu, you just have to include a systemmenu entry. My menu referencing section now looks like this and I have a working menu again.

Plugin {
    type = menu
    config {
	image = /usr/share/icons/gnome/16x16/places/gnome-main-menu.png
        systemmenu {
        }

	separator {
	}
        
        item {
            name = terminal
            image = /usr/share/fbpanel/images/gnome_terminal.svg
            action = /usr/bin/x-terminal-emulator
        }

    }
}

ObTitle: The Mighty Mighty Bosstones, "Let's face it"

[*] gtfn: googled the friendly net

November is here and Ubuntu 9.10 Karmic Koala is there, too. As one of its promises is faster start-up time, I used the weekend to update my laptop (a Dell 610) which was still running Hardy (8.04). The upgrade path (hardy->intrepid->jaunty->karmic) took quite some time and I'm not entirely sure it was worth it:

- Bootup time doesn't seem to have improved, quite to the contrary. I'm mounting ntfs volumes on startup and it seems that
the new parallel bootup procedure seems to run into trouble with that.
- I lost support for fglrx, the proprietary driver for the ATI graphics. It's possible to reinstall the package, but modprobing the module gives a "failed to allocate memory" error. Hence, no "graphics effects" (aka) compiz for me.
- Gnome-Shell isn't installed by default for me.
- The xemacs21-mule package postinstall script fails. This has the not exactly nice side effect that the upgrade procedure decided to quit after installing new packages, which means that it didn't clean up the cruft left behind from the old installation and also didn't boot into the new kernel.
- The update to grub2 didn't honour my old selection for which OS to boot by default (the laptop is mostly used by family members who prefer alternative systems).

And as always on Ubuntu update, a lot of crap that I got rid off previously, gets reinstalled. Like fspot, tomboy and wvdial, to name but a few. I can't help being reminded of adware.

Update:Turns out that newer fglrx modules don't support my ATI Radeon Mobility 1400 any longer (since jaunty, that is). To get GLX working, I had to un-install everything related to fglrx and re-install mesa. What I learned again the hard way during me fiddling with my old xorg.conf is that gdm is a real PITA if you have no working X configuration. What really annoyed me, though, was that status messages from dosfsck hosed the recovery (boot) mode in that the keyboard became unusable. Worse, my hard configured wlan setup at boot time is currently not working anymore, I can connect to my WiFi only through network manager. But without X I couldn't startup network manager, of course, so I also could't do a remote login. And the Wifi connection is really flakey whereas it had been stable ever since with the older hardy installation. Probably a driver issue. Update 2: Various forum entries suggested using backported modules for wifi, including a newer version of the iwl3945 driver. However, this didn't solve my problem. Another entry suggested switching to wicd instead of NetworkManager. And yes, indeed, my wifi problems are gone now.

ObTitle: Gorillaz, "Demon Days"

So I finally bought a new desktop computer in order to replace my old and loud one. The new one should fulfill two more or less conflicting requirements: it should be fast and quiet. "Fast" mainly refers to cpu and disk speed, for the gpu I only had the minimal requirement that it would be possible to see videos (non-HD) and show some basic 3d effects. I left most of the details to a local dealer, but I verified that the components would work with Linux, which actually means I only needed to ensure that the board, an Asus P5Q-VM, which is based on the intel G45 chipset, would work. A little digging showed that there might be some problems with X11 if the distribution didn't provide new enough drivers but I found enough success reports to be optimistic.

Well, I finally got the new workstation. It's an impressive black box with a Silentmaxx ST-11 tower which is really very silent -- it's so quiet that they provide large blue leds on the box to that you have a reminder that the system is on -- this was one of the details I wasn't aware of when ordering the system. I started installing Debian Lenny on the box, as I always followed the rule that buying a new computer is a good point in time to install from scratch and get rid of old software cruft. So I did nothing to carry over my old installation which I had regularly updated at least since Sarge. That was actually a quite pleasant way for me as it was the first time I've got to see the "new" graphical installer. There were some minor obstacles, though. First, this time I wanted to have both an LVM setup and encryption running. I was pleasantly suprised to see the installer providing an option for such a setup, but quite less suprised to see it fail with a segmentation fault during wiping the (500GB sata) disk. Fortunately, the manual partitioning option provided enough easy guidelines to quickly rebuild the setup of the automated partitioning scheme while allowing to leave out the unneeded step to clean the disk, so this was not a big issue. Another point that I think is interesting is that hibernation worked out of the box even with the LVM/crypto setup -- I had thought that hibernate or better said, suspend to disk, would requre the swap to be unencrypted but apparently I was wrong. Very nice.

I went on to install a very minimal system without any desktop setup. I went away from Gnome back to WindowMaker some time ago, but I had seen the LXDE as an install option which is a much smaller "desktop environment" with OpenBox2 as the central window manager. Now, compare the list of packages for Gnome with the list of list of packages for lxde -- that's quite a difference. But soon enough I realized how difficult it has become to avoid the larger environments. There are quite a lot of applications which I am used to using right now (having installed Gnome on my older desktop as well as running Ubuntu with Gnome on several other installations) which depend on Gnome. Evolution, for instance, which I privately mainly use as a glorified calender, depends on Gnome just as Rhythmbox does. This may hardly come as a suprise, but I've simpy not thought about it in advance. So I'm now looking for replacements for some applications.

Of course, in order to run any desktop environment you have to have a running X11 installation. This failed at first: I had connected my 19"-TFT (an old Benq one) via VGA to the on-board graphic chipset from the G45, which resulted in the monitor proclaiming to get out-of sync timings. After replacing the VGA connection with a DVI one, the vesa driver provided a working X11 setup. However, I soon discovered that I would not want to stick with the vesa setup, as for instance there is no xvideo extension. Unfortunately, the xserver-xorg-video-intel package in Lenny is version 2.3.2, which is too old to support the G45, so even though I specified the intel driver in the device section of /etc/X11/xorg.conf, I didn't get a usable screen. However, packages for Xorg in version 2.8 had entered unstable some time ago -- as I had quickly installed kernel 2.6.30 from backports.org and the packages doesn't have to broad version requirements on other packages, this seemed like a viable option. So I know have an up and running intel driver for X11, but it remains to be seen how stable it really is over time. And, of course, I've got to keep an eye on security updates.

ObTitle: Björk, "Post", (more info)

Some time ago I updated my trusty old workstation from Debian Etch to Lenny. As always, there were some minor glitches, not necessarily all of them were due to Lenny itself, though. Here's a rough list.

One really annoying issue was that my trusty old Matrox G400 suddenly behaved unbelievably slow under X11. I could nearly watch the pixels going by. I could tell by the X server log that some things had changed (like Xrandr now running) but I couldn't tell what was responsible for the problem. Explicitly setting the option "NoAccel" to "false" and "NoHal" to "true" (I have a single screen setup) finally settled the problem. I also tried "UseFBDev" "true" but this only works for me with a Depth of 24, which in turn screws my screen whenever I switch to the console (for instance on hibernate), so I'm back to not using the framebuffer device (and a depth of 16).

The Postgres update didn't work for me. I guess that this was mainly due to a lack of disc space. It took me some time to figure out that no default cluster had been created, which was the reason that the manual restore of the DB backup I had made failed.

I took the chance and finally updated to Iceweasel (firefox) 3. Some of my extensions were lost this way, but I managed to replace or update nearly all of them. Only Hit-a-Hint required a manual intervention directly in the install.rdf of the extension.

One typical annoyance of updates is that user settings are often not updated. This time I experienced that mainly with FBpanel in which all icon settings were broken.

Of course, the update broke some of my proprietary programs because of library problems. Korrektor, BMM, all are history now. As are some trusty old programs I still used, for instance xmms is gone, too. I used to stick with xmms mainly due to the Windowmaker mini app wmusic which allowed control of xmms via the dock, but as I've been using Rhythmbox for quite a while now, which has a NETWM compliant panel control, I could finally let it go.

ObTitle: Amy Whinehouse, "Back to black"

I've got tired of writing perl snippets just for simple shell tasks [*]. One such typical task is summing up the occurences of a particular pattern in a set of files to get a total of matches. Of course, this is trivial, but it's so trivial I tend to forget the use of 'expr'.

# code snippet GUID: 3E731D3F-51A3-43C7-B567-26760636DBCD
result=0
for i in `grep -c '[[:digit:]]:|>>>' .local*| cut -f2 -d: `
do
  result=`expr $result + $i` 
done
echo $result

As a one liner for copy&paste:

result=0; for i in `grep -c '[[:digit:]]:|>>>' .local*| cut -f2 -d: `; do result=`expr $result + $i`; done; echo $result

[*] The following quote is due to Kristian Köhntopp: "Use perl. It's necessary to know shell programming but not to use it."

ObTitle: Love is all, "Nine times that same song"

One of the nicer extensions of mozilla I use is It's all text. It allows you to call an external editor to edit text fields which is a really nice thing if you're editing longer wiki pages, for instance.
Not surprisingly, my external editor of choice is XEmacs. But my XEmacs is heavily configured and loads a lot of stuff. Even worse, I usually run a beta version with debugging turned on, so that it runs extra slow, so the time gap between clicking on that little "edit" button below the text field and the moment when I could finally start typing started to annoy me quickly.

I remembered that some years ago I used to log in remotely to a running machine, fire up some script and have either my running XEmacs session connected or a new XEmacs process started. Connecting to the running XEmacs happens via gnuclient, so much was clear but gnuclient doesn't start a new XEmacs if there is none running (actually, the server process gnuserv will die when the XEmacs process terminates, so there isn't much gnuclient can do). The emacs wiki page linked to above already contains a number of scripts that eventually do what I need to do, but I've found none of them convincing, so here is my version. It's linux specific in its BSD style call to 'ps' to determine the running processes, but should be portable sh otherwise. I could have sprinkled the fetch_procs function with OS specific variants, but as I'm currently running my XEmacs on linux only, I left that as an exercise to the reader.

#!/bin/sh

emacs=xemacs
gnuclient=gnuclient
gnuserv=gnuserv
user=`id -un`

gnuclientopts=""

fetch_procs () {
 ps xU $user
}

runninggnuserv=`fetch_procs | grep "[^]]$gnuserv"`
runningemacs=`fetch_procs | grep "[^]]$emacs"`
if [ "$?" -eq "0" ]; then
    gnuservpid=`echo $runninggnuserv | cut -f1 -d' '`
    emacspid=`echo $runningemacs | cut -f1 -d' '`
    echo "Using running gnuserv $gnuservpid of emacs $emacspid"
    $gnuclient $gnuclientopts "$@"
else
    echo "Starting new $emacs"
    $emacs "$@"
fi

ObTitle: Panic at the disco, from the album "A fever you can't sweat out"

Some time ago I started having nothing but trouble with my rt2570 based USB wlan stick. After we moved into our new flat, I finally started out and looked for a better solution. Looking through some retail market I found a package claiming it would support linux. After I was ensured I could return the thing in case of trouble (like outdated proprietary drivers), I bought it -- a Dlink pci card. After installing the thing, I quickly discovered that it's again a Ralink based product -- this time a RT61 based one, requiring firmware. I'm not happy with that but at least my workstation has wifi again.

One thing has changed, btw.: the newer kernels (I run 2.6.27 at the time of this writing) happily support the driver out of the box (i.e., after a recompilation and figuring out where to put the firmware), no more manual compiling and patching of drivers is required. And then you can also use wpa_supplicant to configure the network, which makes the entry in /etc/network/interfaces alot simpler:

auto wlan1
iface wlan1 inet dhcp
      pre-up ifconfig wlan1 up
      wpa-conf /etc/wpa_supplicant.conf
      post-up iwconfig wlan1 rate 54M
      post-up /etc/network/if-post-up.d/bauhaus.fw

Unfortunately, as you can probably guess, my connection is rather lousy and setting the rate up doesn't help much. It's wlan1 for me, btw., because I told /etc/udev/rules.d/z25_persistent-net.rules to configure it that way.

Recent fun:
- Working hibernate/suspend without any manual configuration on a brand new Dell desktop using Ubuntu 8.04 (beta).
- Watching a self-generated video CD (probably done under Windows) under Linux without any problem with my wife that we couldn't watch on Windows due to scrambled colours under Windows media player.

Recent less fun:
- Hibernate/suspend on our older Dell laptop with Ubuntu 7.10 and finding it's a known issue with the ipw3945 driver.
- Getting vmware to work on a recent 2.6.24 system.
- Having fixed that finding out the hard way that Aero won't display in the VM, requiring to go back to cloning the (not so small) physical machine a third time.
- Being unable to get a DHCP lease over the rt2x00 driver for my USB wlan stick with kernel 2.6.24, despite being able to get a sucessful connect via wpa_supplicant/wext.

Etch ships with drivers for the Ralink RT2570 wifi/wlan chipset in the package rt2570-source. However, for whatever reason, if I try to use that package (i.e. compile it the usual Debian way via make-kpkg modules-image and install the resulting deb), modprobe rt2570, I only see that the usb subsystem sees the adapter but I get no interface (rausb0). I then just recompiled my local CVS copy from last summer, modprobe/insmod it and everything is fine. Huh?

Update: I updated to 2.6.21.1 today, using a fresh vanilla kernel. Still Debians module compiles flawlessly, can be installed and the module installed, but no working interface shows up. Manual compiling/installing the old CVS version (probably by a month younger) works out of the box.

Since over a month, I have a Debian Etch DVD lying around. Last night, I finally finished backing up any valuable data and started the upgrade, closely following the careful upgrade instructions given in the release notes. This worked quite well, some minor glitches have happened, though. For instance, wmsensors and wmwifi were removed, although at last wmsensors is included in etch; I haven't checked wmwifi yet. Then my mouse stopped working. Turned out, I never updated gpm.conf from my old serial mouseman configuration to the newer usb mouse I have since quite some time. This was easy to fix, though. Finally, the installation of iceweasel aka Firefox 2.0 resulted in the loss of two of my favourite extensions: SurfKeys, which for instance allows to move between tabs with "o" and "u" and Reveal, which shows mini views of the contents of other tabs. I can live without Reveal but I'm going to look for a replacement for SurfKeys (yes, I use hit-a-hint and keyconfig, but I don't know how I can configure moving between tabs). There is probably more that needs fixing, but nothing spectacular so far.

Update: The surfkeys problem can be solved (see this discussion).

The other issue is that I never got hibernation working on the server and I fixed that today, too. I use a vanilla kernel.org self-compiled kernel on the box, so I could have gone for Suspend2, but as I have sysfs_power_state working on my Ubuntu laptop without much hassle, I wondered what was the problem on the server. As it turned out, suspend to ram (mem) doesn't seem to work and as that always had failed, I think I never really tried suspend to disk. Additionally, I hadn't added a resume line to my grub configuration, so it wouldn't have worked anyway. After a freeze in NetworkStop on my first try with suspend to disk, I added the rt2570 module to the list of blacklisted modules and now suspend to disk (and resume, of course) works. Yay! I'll be able to continue my hacking on UCW without having to boot it all the time, for instance. Here is what is actually doing the work from the config file:

[root@bauhaus->root]grep -v '#' /etc/hibernate/hibernate.conf | sed '/^$/d'
UseSysfsPowerState disk
PowerdownMethod shutdown
Verbosity 1
LogFile /var/log/hibernate.log
LogVerbosity 1
Distribution debian
SaveClock restore-only
UnmountFSTypes smbfs nfs
UnloadBlacklistedModules yes
LoadModules auto
DownInterfaces rausb0 eth0
UpInterfaces auto
XStatus x
XSuspendText Preparing to suspend...
XResumeText Resuming from suspend...

I finally fixed a very annoying bug in the network setup. Bauhaus is connected via an USB wlan stick by Conceptronic, which has a Ralink chipset (rt2570). I made that thing work quite some time ago, although the (legacy) module from
rt2x00.serialmonkey.com required a small patch at the time.
What I couldn't fix though was a very annoying behaviour of the driver: It wouldn't connect on the first try. Only if I re-started /etc/init.d/network, the server would connect to my AP. I had figured out that the driver required the interface to be upped before you would be able to set ESSID etc., but I couldn't get it to work with Debians ifupdown scheme. Some days ago, somebody in de.comp.os.unix.linux.misc posted his configuration and he had a pre-up ifconfig rausb0 two times, so that makes three times the interface gets upped. This actually did the trick. I now have the following in my /etc/network/interfaces:

# The wlan interface
auto rausb0
iface rausb0 inet dhcp
      pre-up ifconfig rausb0 up
      pre-up iwconfig rausb0 essid XXXXX
      pre-up iwconfig rausb0 channel 10
      pre-up sleep 3
      pre-up ifconfig rausb0 up
      pre-up iwpriv rausb0 enc 3
      pre-up iwpriv rausb0 auth 3
      pre-up iwpriv rausb0 wpapsk XXXXX
      pre-up sleep 3
      post-up /etc/network/if-post-up.d/bauhaus.fw

And this finally allows to connect the damned thing at boot time, which is nice for two reasons: It speeds up my boot process (before dhclient would fail after an awful lot of time) and my clock will be much more precise. Now the next thing would be to try the new implementation of the driver, aka rt2x00, and go for WPA2 (which the current legacy driver is not capable of). I'll leave that for another day, though.