Since I’m quite the python fan, and also indirectly a fan of bazaar, I’m really liking the bazaar explorer. Especially on windows. Since I wanted the same application on linux, tried to install the explorer.

Somehow, somewhere something went wrong. Obviously not a good thing. After installing the bazaar explorer (first through downloading the tarball here, secondly by executing “bzr lp:bzr-explorer explorer” in the ~/.bazaar/plugins folder) I stumbled upon an error:

Unable to load plugin 'explorer' from '/usr/lib/python2.5/site-packages/bzrlib/plugins'

and later the same from my ~/.bazaar/plugins folder…

The solution

Actually deceivingly simple, the source installation page from bazaar explorer said:

If this fails to start, ensure that you have compatible versions of dependent products installed, namely: QBzr, PyQt, Qt and bzr.

Making sure you’ve got these requirements fixes the whole ordeal.

Qbzr on debian

My solution was simply to install qbzr on debian, since I already had bzr, qt and pyqt installed.

To install qbzr on debian, one needs to add the following to /etc/apt/sources.list (following these guidelines):

deb http://ppa.launchpad.net/qbzr-dev/ppa/ubuntu jaunty main

Updating apt and installing qbzr (as root of course, or sudo):

# aptitude update
# aptitude install qbzr

Now, enjoy your bazaar explorer!

PS: when you have the error on the key, here’s an explanation on how to add it to apt

The Haiku OS project site has published the first version ISO so I immediatelly went to the download page of the Haiku OS site, since I’m a big pro of the haiku idea. Downloading the iso (with sourceforge mirror) only took 10 minutes, but it isn’t nice of them to zip it (like an iso in itself isn’t good enough)

Preparing Virtualbox with pretty much everything standard (for an OS not known by Virtualbox that is):

• 256 MB ram — I assume it will be enough
• 2 GB HDD — Said in the installation instructions to be sufficient

The iso mounted in Virtualbox really did a nice bootup time, installer could use some work, especially since exiting (square box left-above) the installer / setup actually makes the total system do nothing anymore (it should reboot, or failsafe to desktop imho). Total setup time seems to have come down to 10 minutes, extremely fast!

In the end, there’s got to be large credits for the installer being so straight-forward. In the end, if one reads the installer instructions as they pop-up in the process, one can install with ease. Most files seemed to be html files, which is quite funny ^^

Also, having an OS with 16835 files is seriously nice. Don’t really know if I should say this, but at the moment, it’s the smallest OS I’ve got in my entire Virtualbox setup. And I will most certainly play around with this one a _lot_ considering I hope it will become what I always wished for in the Desktop Linux experience.

What is also quite amazing imho is the fact that python, perl, and more are installed by default. Because I believe scripting will become more and more common in the real world, this can go the correct way!
Having played around with this some, taking some (loaded) screenshots, etc… I think Haiku OS has gone a long way, but still has very long to go. Positive points (for now):

• Extremely fast boot-time, and a no-nonsense approach to the whole deal
• Having the BeOS ideas and not trying to re-invent things seems to have effect

Negative points (points to work on):

• alt-tab feature, or something more obvious than the Deskbar to switch between applications

These points and conclusions are what I saw on first-impression. And I like what I see at the moment! Although the project has a long way to go still, it seems to gain momentum every time I try it, and I hope to start developing for it soon myself (although I prefer developing for servers :P)

Share this post, help Haiku gain even more momentum, try things out and comment here (or directly on the haiku site)

This person has been very busy, and still little work seems to be done.

In a nutshell, we’ve been busy getting the main website for pySM online. pySM is me and Tim’s project into system management (There are some posts about it in this blog)

However, soon after getting quite the lot of url’s and basic apps online, I decided to change the url structure to become a littlebit more simple. Fact is that we were trying to overachieve, something that seems to happen to us a lot :(

Anyhow, pySM is currently online on 2 (yes, 2 url’s, instead of the douzen-or-so earlier) links:

• pysm.be << Main site + Blog, makes sense that people interested in the project will directly see the progress. However we will probably in the future try to maintain a proper website with introduction, links, screenshots, documentation, world domination plans, etc…
• trac.pysm.be << Trac site, contains documentation (raw material basically) and a webview for the bazaar repository. Tickets can be created over there too, although I don’t really see the use for the moment.

I hope that having this simpler structure will enable us to work more on pySM to get our first _decent_ version out. After that, we can try to overachieve again :)

Best of regards,
cpf

Today all pysm sites (which used to be only trac) has been moved.

The “old” url (www.pysm.be) has been moved to the server hosting codercpf.be (Yes, this site)

We’ve changed the old only trac to seperated sites:

Update!

I put some url’s here last time, but those changed. For sake of having consistency, and not sending people to the wrong places, here are the updated ones :) (And the post announcing it on this blog)

• pysm.be << Main site + Blog, makes sense that people interested in the project will directly see the progress. However we will probably in the future try to maintain a proper website with introduction, links, screenshots, documentation, world domination plans, etc…
• trac.pysm.be << Trac site, contains documentation (raw material basically) and a webview for the bazaar repository. Tickets can be created over there too, although I don’t really see the use for the moment.

While we were at it, we changed the previous Subversion repository to a bazaar repo. Makes it a lot easier for us developers, although on first glance there’ll be little difference between the current bazaar repo and the Subversion way of doing things.

This will only be a small step into our attempt to rule the world using pysm!

Update: The site has removed the logo. Arch is victorious (linky)

As it appears, there is someone (in particular a company named Ace International Tutoring Pty Ltd) as outragious as it might seem, it is true.

As for the original article, it explains how the devs and official archlinux crew have come to the conclusion that this logo is indeed originaly theirs, and it has been “stolen” by someone thinking he can get away with it…

DIGG THE ARTICLE :)

Long title for a short post. I tried to search on google, and since I found nothing decent.

To migrate from thunderbird to icedove in debian linux, one only needs to rename the config directory of thunderbird:

mv .thunderbird .mozilla-thunderbird

Although I think this is stupid (I’d have prefered to see .icedove instead of .mozilla-thunderbird)

Enjoy your path towards ice* enlightenment…

So, since I’m a long and happy yeganesh user, but I recently moved to debian and didn’t want to have the haskell overhead in my system, I wrote my own little implementation in python.

What is it?

PyDmenu tries to clone the yeganesh behavior. However, it has its own ups and downs. Currently, the code is made to work, not made to look good (it will look good in the future though)

It’s a shell above the dmenu application which is basically a simple launcher.

How to use?

Download the current .tar.bz2 file, untar:

$ tar xjfv pydmenu.tar.bz2

Using it like this:

$(/path/to/pydmenu.py)

will automatically execute the selected item. I have this in a script bound to mod4+p.

What else

The only thing handy would be to know that it creates a simple cPickle file in ~/pydmenu_save

To change any behavior you should edit the pydmenu.py file, the font, colors, etc… are in the DMENU array, only requirement there is to keep dmenu as first.

Enjoy, and please leave any comments (/emails) if something went wrong somewhere.

Tried to keep the title short. In this article I’ll be trying to explain how to start a new X screen with some WM in it.

I wanted to use it to launch a new X screen for gaming purposes.

The code

In a nutshell, this is code that resembles the most basic functionality of startx (without the use of xinitrc of course). I found a serious lack of documentation about this… Although it is far from hard :)

Xorg :1 -ac -config xorg-game.conf &
DISPLAY=:1 openbox

This is actually really basic. It basically tells Xorg to start a new screen. Instead of :0 (standard), this will launch on :1 -> next we only need to make a wm (in this case openbox) start in that screen.

The script will take you directly at the new screen. Getting back at your first requires ctrl+alt+f7, your new one is at ctrl+alt+f8, and the next would be at ctrl+alt+f9, and so on.

So, using sleekxmpp to send IQ set / get is quite simple, but requires some manual labor…

This is what I came up with, it’s all quite arbitrary, and can be put in a function easily. But at least now I finally figured out how to send IQ stuff. Receiving some might prove harder…

# Because the backbone works with this... >.>
from xml.etree import cElementTree as cET
[...]
iq = self.makeIqSet()
iq.attrib['from'] = self.fulljid
iq.attrib['to'] = 'test@localhost/bla'
query = self.makeIqQuery(iq, "pysm:iq:shutdown")
shutdown = cET.Element('shutdown')
query.append(shutdown)
shutdown.attrib['when'] = 'now'
self.send(iq)

The shutdown stuff is nothing at the moment, it doesn’t really do anything. The current server just displays the received xml…

The 6th day is already nearly finished, and I haven’t really posted anything about it yet :o

So, as the title says I’ll be (have been) working on pySM and the archer startpage

Startpage

The startpage is now finally finished for as far as I wanted it. However, it’s possible that I’ll be adjusting the theme to a darker one.

The html entities are decoded, so everything in the feeds will be displayed properly in the pages. Links are clickable like they were for archlinux, titles are clickable to go to the original page, etc…

If there’s any comments on the layout, hit me at the forums

PySM

There’s little work been done this morning. However, a lot of work will be done on this from now.

Although I consider this version of the XMPP server complete enough to function in a simple environment, a lot of work will be required to get it at least scaled to a presentable result. Also, currently, I have no idea on how to make 2 clients communicate, so I won’t be able to patch through the most essential things at the moment. But since I can now connect, and test clients, I’ll be focussing on getting clients connected, and having clients send through valid iq requests. This shouldn’t be hard, but since I’d like to go on with sleekxmpp it might require some adjustments to their code to make that easier.

After that, I should still have some time to focus on module management. Since I’d like to have a sort of eventhandler in the server, those events will be linked to a module, which in return will be loaded and maintained by a module manager. So perhaps it’s not really the most difficult part, but it should work good, and be quite pluggable.

When that’s done, I hope to have some time left to figure out some way to have 2 clients communicate through the server. Patching through messages and iq requests and responses.

Anyhow, a lot of work needs to be done, and it’s not going to work itself, so I’ll be starting now :)