blog.foosion.org

Iotop

Posted by bene - 30/06/08 at 11:06:55 pm

One of the drawbacks of top is that it often can’t help to spot processes which push up system load so high.
High system loads are often caused by very I/O intensive tasks.
And as I/O intensive needn’t mean CPU intensive, those tasks may not even show up in top.

This is where Iotop comes into play. It is just a python script which evaluates the per-task disk I/O accounting statistics exported by the Linux kernel (2.6.20+).

This is what you need to enable in your kernel configuration:

General setup
[*] Export task/process statistics through netlink (EXPERIMENTAL)
[ ]   Enable per-task delay accounting (EXPERIMENTAL)
[*]   Enable extended accounting over taskstats (EXPERIMENTAL)
[*]     Enable per-task storage I/O accounting (EXPERIMENTAL)

Now THAT feels ever more like home…

Posted by Nils - 18/06/08 at 06:06:22 am

Firefox 3 is finally ready to download (though all servers are at the moment more or less overwhelmed, due to the worldrecord attempt).

Since it’s first release it integrates very well into the various systems because someone at Mozilla (after all) had the look & feel in mind. I think thats a big achievement though I stilled wished a support for the various password managers like the keychain or the kde wallet would have made it into the final release.

To top it all off, the UI gets even more “mac” with Aroonax’s GrApple Theme from takebacktheweb.org. (Which is an awesome domainname by the way )

So folks, have a nice day and light a candle, hoping for people around the world to finally abandon this so called “internet explorer” and get their copy of web-freedom.

word!

Songbird

Posted by micele - 17/06/08 at 11:06:21 am

Did anyone try Songbird? As a Linux user I’m always in search of a good audio player. Of course there is Amarok and of course there is Audacious but in my eyes there is not much in between. Amarok is quite large and didn’t really work out at all on my 600MHz P3. On the other hand Audacious is quite tiny but not made for handling large music archives like, .. let’s say.. mine
More accidentally I discovered a new audio player called Songbird. It somehow uses mozillas xul and seems to be a good compromise. To be honest, I didn’t test it too much so far. This might relate to my love-hate relationship to some other media player. Obviously the design is somehow inspired by this specific product
The current version 0.6 is still alpha, so please be patient. Maybe at some day it becomes a real alternative to the mentioned audio players. At least I already love it for it’s logo

Linux kernel 2.6.25.6

Posted by bene - 09/06/08 at 09:06:15 pm

Chris Wright just pushed version 2.6.25.6 to the current stable kernel tree.

It’s definitely worth to check it out: There are 50 commits since 2.6.25.5, many of them backports of bugfixes in 2.6.26-rcx…

Happy compiling

Parse HTML the Groovy way

Posted by thomas - 09/06/08 at 05:06:53 pm

In the last couple of weeks I often had to download a lot of files, submitted to a web-based teaching platform. Downloading all these files by hand is very annoying so I implemented a short Groovy script. Since Groovy has a great support for parsing well-formed XML-like information it fails if you want to parse unstructured and nasty HTML code.

At last I searched for a Java library containing an HTML-parsers and I found TagSoup. This is a SAX-compliant HTML-parser specialized in re-formating and cleaning up faulty HTML code.

This is <B>bold, <I>bold italic, </b>italic, </i>normal text

will be rewritten to

This is <b>bold, <i>bold italic, </i></b><i>italic, </i>normal text.

One advantage of TagSoup is the Xpath-like query mechanism. It parses the HTML code and generates an object structure representing this content. Now the user can access the single elements. One possible example could be:

def slurper = new XmlSlurper(new org.ccil.cowan.tagsoup.Parser())
html = slurper.parse("an_example_file.html")
table = html.body.div.find{ it.@id == "content" }.form.table.
find{ it.@id == "attempts" }

This retrieves the table “attempts” placed inside a form in the div “content”. The method findAll() will retrieve all elements for a given attribute or with given child elements.

After all I fell in love with TagSoup. It saves a lot of work when you have to access HTML content of websites, portals or similar, which are not able to send a XHTML 1.x compliant responses. But this is an other topic .

Big Buck Bunny

Posted by joern - 06/06/08 at 02:06:49 pm

Shortly after Blender 2.46, Big Buck Bunny was released recently. The new open movie (released under CC 3.0) resulted from the “Peach” project, which was already the second open movie project after project “Orange” (resulting in Elephants Dream).

Both movies demonstrate the numerous astonishing features of Blender and caused loads of valuable feedback for improvements for the Blender developers. In order to keep up with the young tradition, Blender already started the follow-up project called Apricot in December 2007. This time the main focus lies not only on modeling and rendering, but also on the included 3D-Engine that Blender comes with.

Colored Manpages

Posted by micele - 05/06/08 at 10:06:16 pm

Manpages are a very appropriate instrument for documentation on linux und unix based systems. Unfortunately manpages have been invented in times of monochrome displays. Todays colored displays offer much more clearness and readability. Unfortunately manpages don’t - at least not by default…
I really tried hard to give my linux manpages more readability and finally - it worked. There are a lot of different (and also difficult) approaches. In my eyes the most easiest way is to simply add some less capabilities (less is used as the default man page browser) to enforce coloring. This can be done by using your favorite shell .rc for setting those capabilities. In my case i just added

export LESS_TERMCAP_mb=$'\E[01;31m'
export LESS_TERMCAP_md=$'\E[01;31m'
export LESS_TERMCAP_me=$'\E[0m'
export LESS_TERMCAP_se=$'\E[0m'
export LESS_TERMCAP_so=$'\E[01;44;33m'
export LESS_TERMCAP_ue=$'\E[0m'
export LESS_TERMCAP_us=$'\E[01;32m'

to my .zshrc. The result is quite satisfying and looks like this:

I really fell in love with man pages again. Ok, I confess: I sometimes really care about appearance

Update: While gentoo has some problems with the mentioned method it even works fine for mac.