drakx as libraries / drakx doc / drakx bindings

[Sfiet_Konstantin]

This idea is a long term idea.
It does not target any version of Mageia.

I wish to see all the drakx tools as libraries, that can be used by others programs. These libraries can be C / C++ shared libs or even perl / python / putyourfavouritelanguagehere.

These libraries allows separation of GUI and core compoenents, and also allows easy designing of new GUI and usage in others applications. For example, we can imagine integration of MCC in Gnome / KDE control centers.

A lib for urpmi can also allow an easier design for an "application store" like app. This program should link to liburpmi and libappdb for example to retrieve information from appdb. Liburpmi can also be used for example in qt4urpm like apps that are specialized for finding a package or managing uninstall ...

Then I also suggests to make the drakx tools more accessible, by providing a documentation to them, and to allow others distros (for example) to take inspiration to design cool configuration centers and allows developers to improve and write others drakx tools.

Finally, I may be an heretic , but I think that drakx may be rewritten in other languages, just because of perl, because perl is not a very attractive language and sometimes perl scripts are quite difficult to maintain. Maybe a switch to python ? Well, this point is open to long discussions

PS : don't throw me stones because I'm pointing some "problems" of drakx: I like MCC, and MCC keeps me using Mdk / Mdv / Mga. I'm just trying to improve them, and it might be needed to rewrite from scratch

[misc]

Having c/c++ library would requires to rewrite them from scratch, as we currently have them in perl. Regarding rewriting them to something else, the question is "who will then develop the tools" ?

After all, we do have tools and people to maintain them, and I see no one starting to redo them in python. So if no one is writing anything, dropping them would not help much. Once someone has a working tool in another language ( and without regression ), maybe we can first start to think switching.

Regarding something to install package, there is already packagekit, which is used on gnome ( a lot ) and that work accross all packages managers. This is also the basis of the AppStream project that was discussed in Nuremberg. Search for it on the web, you should find the video of the presentation, and a student should work on it for summer of code.

[Sfiet_Konstantin]

Regarding something to install package, there is already packagekit, which is used on gnome ( a lot ) and that work accross all packages managers. This is also the basis of the AppStream project that was discussed in Nuremberg. Search for it on the web, you should find the video of the presentation, and a student should work on it for summer of code.

Good news for the GSOC

Well, packagekit upder KDE works quite badly ...
I understand your arguments about rewriting, that is why my post is more speculative than a project. If I find the courage and the time, I even feel write them myself. The only problem is about some lack of doc from drakx. Do you know where to find them (if there are some of them ...) ?

EDIT : well, I think that I will wait for Mandriva to develop their MAM and then check their work and maybe use their Qt bindings. Because MAM ... is ugly ...
Maybe these "drakx libraries" will come from Mdv

[Sfiet_Konstantin]

I have started (just started) some tries of "porting" drakxtools from perl to python.
Well, it is not a real port but some mimic of drakxtools.
I'm focusing on "simple components" like drakboot now.

Please note that I do not have much time (as a student) but I really want the drakxtools to be easier to use / maintain / enhance, especially if it is well documented.
I will setup a gitorious if I have progresses in my work.