October 2009 Archives
Dymo 400 Label Writer in Ubuntu (Jaunty/Karmic)
Apparently the cups driver (ppd) for Ubuntu (in both Karmic and
Jaunty) is not installed by default, see
this and
this for
more information.
The strange thing is that if you look at openprinting.org, it says
Dymo LabelWriter 400
Supplied with CUPS Works perfectly.
Supplied with cups??? Works perfectly?? How, where, when??
Well:
% apt-get source cups
% ls -l cups-1.3.9/ppd/dymo.ppd
-rw-rw-r-- 1 miekg miekg 24K Nov 27 2007 cups-1.3.9/ppd/dymo.ppd
And that's your ppd file. For good measure I've put this up
for download here too.
I'm using this myself and it indeed works correct. When configuring
your printer in cups you need to click use custom ppd file or
something like that.
rdup 11dev
During the past few weeks I been (also) busy updating rdup. There were a few things I wanted to change (also because of users requesting these).
Major changes
- Added username and groupname to the output of rdup
- I've added two new output modifiers:
%Uand%Gwhich hold the user- and groupname. rdup-upwill use the username when creating the file. If the username does not match theuidon the system the file is created, the (numeric)uidis used.
- I've added two new output modifiers:
- Added the
m-timeto the output of rdup- The modifier
%twas already implemented, but now it is used by default. rdup-upuses this to set the last modification time on the files it creates.
- The modifier
- The output of rdup now defaults to with contents (the -c
switch).
- Actually this is the only output still supported
rdup-tralso understands this new output.- The -P switch has been moved from
rdup-trtordup- Insert long story about deadlocks here...
- All in all this has led to simpler and less code.
Smaller changes
- Default install path is
/usr/local - The timestamp file is now created before the backup.
- If
rduptouches this file after the backup there is a time window in which newly created files might not been seen byrdup.
- If
- libnettle has been removed in favor of OpenSSL
How can you help?
Check out rdup from git:
git clone http://www.miek.nl/git/rdup.git/
Pull the 11dev-branch were these changes are taking place:
cd rdup
git checkout -b 11dev origin/11dev
You can can also download a tar file containing the latest
git code:
rdup-11dev.tar.gz.
With sha1: 604832be7c96abdea0ef14f827e3b5b81a233e9c
Test the code, check the documentation (still needs work) and provide feedback.
Fedora 11 vs Ubuntu 9.10
My small and very personal comparison between Fedora 11 (which I run on my laptop) and Ubuntu 9.10 which is my main Linux distribution.
yumvsapt-getyumis a lot slower thanapt-get- why do I need connectivity when using
yum search - Why does
yumdefaults to 'N' (no) when I ask it to install software?
All in all I like
apt-geta lot better.Default resolution
- Ubuntu 9.10 configured my EeeBox and monitor out of the box, Fedora 11 didn't.
Plymouth graphical boot
Very nice to see this in Fedora 11, I want this too for Ubuntu. Too bad Ubuntu does not use this in 9.10. It will happen in 10.04 (I heard).
Laptop
My new EeePC laptop has a Poulsbo video chip (thanks Intel for screwing me!) which Fedora 11 supports wonderful thanks to AdamW.
Ubuntu 9.10 does not support this chip, period.
So I'm happily running Fedora on my laptop and Ubuntu on my other systems.
KPN redux
I installed my new laptop with Fedora 11, and I must say that it is a very nice distribution. I'm even contemplating leaving PulseAudio enabled, 'cause it just works. Unlike ubuntu.
To update my story on the KPN dongle (dongel). I just used it on this laptop and it worked out of the box. The only thing you need to remember is that you need to insert the stick before booting you machine. It will not work if you insert it afterwards. If you observe this rule it will just pop up in NetworkManager.
If it fails you get the dreaded: ignoring due to lack of mobile
broadband capabilties and it will not work in NetworkManager. You can
still use umtsmon then.
Use upstart to replace rc.local
On my new netbook
I wanted to get rid of gdm and
just start X right away. I use auto-login anyway so it
is a bit stupid to first start gdm and then immediately start X.
So I removed gdm and edited /etc/rc.local to start X:
su - miekg -c xinit xterm
But this sort of does not work anymore in Ubuntu Karmic. Karmic
now uses upstart as an init replacement.
So I figured why
not write an upstart job that starts X?
The upstart jobs in Ubuntu are stored in /etc/init.
Gdm's upstart script was a good starting point so I copied that.
I finally settled on the following job:
% cat /etc/init/X.conf
# X - start X
description "Start X"
author "Miek Gieben <miek@miek.nl>"
start on (filesystem and started hal)
stop on runlevel [016]
task
console output
emits starting-x
script
su - miekg -c xinit xterm
end script
So now I can enjoy a really fast boot and fast X startup.
You'll need a ~/.xinitrc to start up XFCE or GNOME or
whatever.
Note; after reading some docs I still don't know what the keyword task means in
the above job.
Goodbye charm, welcome up
I've bought a new 11.1" netbook from Asus, this is going to replace my aging 4G Surf (named charm). I've named the new one up, so I'm hoping the Large Hadron Collider is up and running soon and discovers a new flavor of quarks - 'cause I'm running out of names. (strange is already allocated if I ever buy a Mac and run Linux on that, top and bottom just don't sound right).
This is hopefully the last time I buy a Intel only machine, turns out
this machine has a GMA500 (aka poulsbo) video chipset which
is not
really
working
well.
This chip is not supported by the 100% open source driver Intel is promoting so much. Even Moblin 2.0 (A Intel sponsored distro) does not support it.
There is a git repo containing a driver which was created for Intel by a company named Tungsten Graphics.
I'm pondering getting into kernel development and also trying to code for X, so this may be a good opportunity to delve in. However, my guess is, that this kernel-X-graphics interaction is one of the more difficult areas...
Microsoft rules . ?
Okay, I could not find this in the specs, but I do find this fishy. When querying a Windows DNS server it will give out an authoritative answer (aa bit set), but without an AUTHORITY section.
dig +nocmd +noidentify +multiline @ns5.msft.net. soa hotmail.com
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1
;; WARNING: recursion requested but not available
;; QUESTION SECTION:
;hotmail.com. IN SOA
;; ANSWER SECTION:
hotmail.com. 86400 IN SOA ns1.msft.net. msnhst.microsoft.com. (
2009100802 ; serial
1800 ; refresh (30 minutes)
900 ; retry (15 minutes)
2419200 ; expire (4 weeks)
3600 ; minimum (1 hour)
)
;; ADDITIONAL SECTION:
ns1.msft.net. 3600 IN A 65.55.37.62
And it gets worse:
dig +nocmd +noidentify +multiline @ns5.msft.net. soa miek.nl
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35647
; flags: qr aa rd; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; WARNING: recursion requested but not available
;; QUESTION SECTION:
;miek.nl. IN SOA
;; AUTHORITY SECTION:
. 3600 IN SOA ns1.msft.net. msnhst.microsoft.com. (
2009071743 ; serial
900 ; refresh (15 minutes)
600 ; retry (10 minutes)
86400 ; expire (1 day)
3600 ; minimum (1 hour)
)
;; ADDITIONAL SECTION:
ns1.msft.net. 3600 IN A 65.55.37.62
WTF?? MS rules the internet and is authoritative for . ?
iotop
iotop is a very neat tool showing the processes which do the most
i/o in a top-like manner.
Again having fun with SLES:
SLES-10:
# rpm -i /tmp/iotop-0.3.2-1.1.x86_64.rpm
warning: iotop-0.3.2-1.1.x86_64.rpm: Header V3 DSA signature: NOKEY, key ID ee454f98
error: Failed dependencies:
python >= 2.5 is needed by iotop-0.3.2-1.1.x86_64
Goes off and installs SLES-11
SLES-11:
# rpm -i
warning: /tmp/iotop-0.3.2-1.1.x86_64.rpm: Header V3 DSA signature: NOKEY, key ID ee454f98
error: Failed dependencies:
python < 2.6 is needed by iotop-0.3.2-1.1.x86_64
Nooooo!
I finally made a symlink from /usr/lib64/python2.6/site-packages/iotop
to /usr/lib64/python2.5/site-packages/iotop which did the trick.
Oh, and you will need a kernel version > 2.6.20 to have it work.
Spot the problem
An unofficial rpm of e2fsprogs installed on a SLES 10 system:
# ldd /sbin/e2fsck
libdb-4.3.so => /usr/lib64/libdb-4.3.so (0x00002b5e004e2000)
libpthread.so.0 => /lib64/libpthread.so.0 (0x00002b5e006d6000)
libc.so.6 => /lib64/libc.so.6 (0x00002b5e007ef000)
/lib64/ld-linux-x86-64.so.2 (0x00002b5e003c6000)
Now guess what happens when you reboot?

