I am using a mobile broadband device and created a ad-hoc connection in NM to share it. The device works out of the box and ad-hoc connection also works very fine. But some times if I disconnect from the broadband device, NerworkManager crashes, and after restart when I start that ad-hoc connection once again, it automatically creates too many ad-hoc connections with strange names and the list continues to grow.

You can find the bug here. I am on rawhide.

pyexiv2 is the python extension of Exiv2 library. It helps to find the EXIF information from an image.

The current rawhide version is broken, it just can not be imported.

[kdas@d80 ~]$ python
Python 2.6 (r26:66714, Mar 17 2009, 11:44:21)
[GCC 4.4.0 20090313 (Red Hat 4.4.0-0.26)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import pyexiv2
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.6/site-packages/pyexiv2.py", line 60, in <module>
    import libexiv2python
ImportError: /usr/lib/python2.6/site-packages/libexiv2python.so: undefined

Tried to find what that symbol is using c++filt

[kdas@d80 ~]$ c++filt _ZNK5Exiv29Exifdatum6typeIdEv
Exiv2::Exifdatum::typeId() const

Now is this a very new pyexiv2 which contains symbols from latest exiv2-lib or a problem itself ? The bug is here.

Update: It seems to be a local symbol. It comes from exiv2-lib only , but it shows as local with "notype" from the eu-readelf output :(

From today morning my Dell Latitude D630 is not booting. Called service center , the guy asked me to press different kind of key combos for troubleshoot. Then he asaked me to open the keyboard which is a real bad idea as I can damage the system. After a big argument on this finally they agreed to send an engineer with required parts.

Asking a user to work/open the hardware is not at all a good practice, don't know why they keep asking the same.

Meanwhile view this (less than 1MB), I hope that code is safe in my broken box :)

This Sunday I went to OLPC Pune meet to give a talk on activity development. The students seem be very enthusiastic. The talk lasted for around 2 hours. Photos can be found here.

On the other side, I was looking for a photo catalogue software which can catalogue my photos on DVD or any other removable media, it should  also allow me to search them based on EXIF data and tags and should show thumbnail(s) in the result. I generally use digikam and gthumb too manage and view the photos. In digikam I never being properly able to catalogue photos on DVD (never tried on the last few updates) and gthumb is too slow in showing the thumbnail or full screen views for the large size photos I take. I prefer gthumb over over digikam to view the photos due to the simpler UI.

Finally decided to write my own app for this, which will not have any editing features, Gimp is the tool I use for editing. Watch this video (3.2MB) and please comment if the  thumbnail creation / viewing is fast or not ?

Dear lazyweb,

How one can start learning PyKDE ? For PyQt , reading a few very basic examples given with the source and then reading Qt-help docs seems to be fine. But the KDE libs are much bigger, though one can find a suitable class to work with, how to use that is still not so easy for a newbie.

When it comes to the GUI, it is again not clear to me that which approach one should take , design everything in code ? or one can use qt-designer for that.

Kushal

You can find the wiki page here. This year we want to make groups such that there are two students in a group and make them work on any upstream project or  a new project. They should start the work of finding the project ideas within first two weeks of the program.  Comments/suggestions are welcome.

We have added a few buzz words in the poster to attract students. Generally students prefer to go to different small institutions to do summer project where most of the times they pay a good amount to get a certificate. Trying to change the situation slowly.

Thanks to Nicu for the poster.

Soas (Sugar on a Stick) is a LiveUSB system with sugar on it. It is based on Fedora. You can carry it and boot anywhere as you want like any other live system. You can download the ISO from here.

I was using it on a Eee PC (4GB model), suddenly thought of installing it on the system.

The other Fedora based systems have an option on desktop to install on hard disk, but this system is running sugar so no desktop means no such option. From the #fedora-devel I found the command is "liveinst" which is part of anaconda.

So I installed anaconda ( #yum install anaconda ) and then used the command "liveinst" ( without quotes) , this will fire up the installer for you. Then go through like any other installation. After finishing the installation I found X was crashing, tried with many different config but nothing helped. Finally decided to install gdm and gnome on that box ( #yum install @gnome-desktop gdm ),  also installed system-switch-displaymanager using which I changed the display manager to gdm. Next task was to change the init to 5 in inittab. Rebooted and found gdm , chose "sugar" as the desktop option and logged in :)

If you want any help come to #sugar on freenode.

Where I can find videos of Linux Plumbers Conf ?

A very bad expression:

23462376427 ^ 2132323

Question is which computer languages can give me the answer without much trouble ? Haskell did.

Update: For Python the correct expression is 23462376427 ** 2132323

I just released translation-filter 0.3. You can download it from here.

The scope of development is to create a rudimentary tool that can allow comparing the quality of string translations in an offline mode. For example, the KDE Project uses a web based tool that allows strings to be searched in translated content.

The current set of requirements for this offline tool are:

• to be able to take as input a directory path containing multiple .po files
• to be able to take as input an English string or word to be searched within the .po files collection
• to be able to display every occurring instance of the English string or word along with equivalent translated string
• to be able to display the name of the files being searched
• to be able to save the search result set in an .html file

It uses PyQt4 for the gui, used QWebView to render the result. You will also require translate-toolkit (at least 1.0.1 version, available in Fedora 9).

Among the new features (mostly in the command line version) , using -t or --term you can search for any single term.  You can even check for unattached dependent vowel signs of Indic languages using -d option.