You just forgot to use “cd” in front of the keryx directory.

The tutorial looks great, but as a classic user, I am hitting some snags. I get as far as step 2. Ive been doing some reading trying figure it out, maybe I need the wxPython thing.

I run /media/kingston/keryx-0.92.1/linux and get:
Bash: /media/kingston/keryx-0.92.1/linux: is a directory

user@user-desktop:-$ python keryx.py –create jake-desktop debian
python: can’t open file ‘keryx.py’: [Errno 2] No such file or directory.

I downloaded keryx 0.92.1 on my kingston thumbdrive while using my windows xp online pc. I have tried removing the extracted files I originally extracted in while using the windows pc, and extracting the files while on my ubuntu offline pc.

Cool program, can’t wait to figure it out, my bad.

The package lists should only download the latest packages when you select to install something. The version comparison is actually semi-complicated and could very well be flawed and that would be what is causing your problem. This should all be resolved in the next version of Keryx as well. Remember, its still pretty much an Alpha.

Keryx is quite a very nice piece of software. Congratulations!

Playing a bit with it I noticed the following things:

- I ran the ‘Get updates’ procedure twice in a row and it looks like the download is performed at each time, even if all the updates were successfully downloaded during the first run. Does the update procedure check if the updates are already present in the project directory? Maybe, this could avoid some extra downloads and save time.

- some repositories of my sources list contain the same packages but with different versions (eg. Ubuntu official repos with some version of R and CRAN repository with the latest version of R, usually newer than the Ubuntu one).
The packages installed on my remote machine are the latest ones (from CRAN). When I run Keryx to get the updates, all these packages appear in red, flaged as ‘Newer than Repository’. I tried putting the CRAN repository at the top and end of the sources list but it did not change anything.
Is it a known behaviour? Does the list loader take into account multiple available versions of packages?

Thanks and well done again.

@Chiflado
For now you need python installed as well as wxPython (http://www.wxpython.org/download.php). You can then run Keryx from keryx.py and if you’ve got wxPython installed it will bring up the interface just like it would if you ran it on Windows or on Linux that had wxPython installed.

@Apoorv
Since Keryx is still in development I have just been able to finish up the bare minimum, it does the absolute basics. Over the summer that is going to be one of the main features I will be adding. It will take care of installing the packages, copying over the list files and the packages over to the apt cache folder. I’ll try to keep everyone posted on the development/release status!

I have a suggestion. If I want to download 2 softwares X and Y but I want to install only X and if X doesn’t work then I’ll install Y.

Why don’t you just copy the package lists to /var/lib/apt/lists and deb files to the /var/cache/apt/archives so that installation can be done by apt-get or by synaptic or by any other tool like restricted driver manager etc.
This will let the beginners and GUI lovers use Synaptic and will make your tool usable with restricted-driver-manager etc

Thanks

However I have a question, whenever I click on get updates, it downloads hundreds of files. does it download all the new softwares from repository or just updates for installed packages?

Thanks

In my case, what was most important was not to have an estimation of installation time, but rather an estimation of download time, which is easier to estimate after the first few seconds…
For installation, maybe you could just add something like “installing package X of Y”, or make an estimation based upon the size of the packages themselves… (ok, it’s not very accurate, but it’s usually quite verified…)

We have also been planning on improving the progress bar, and the statistics idea (how many packages being installed, etc) is a very good idea. I’m not sure how feasible the installation time estimate is though, as that his highly dependent upon the hardware of the machine running the installations.