One of the approaches could be to use xmlrpc for the establishment of
jobs at the server. Other than this a log mentioning the complete
build of the final image will be useful so as to avoid the repetition
of the entire process in case some one wants to build a similar image
they built earlier.

On Fri, Mar 13, 2009 at 2:46 AM, Ned Ludd <solar@g.o> wrote:
> On Fri, 2009-03-13 at 01:40 +0530, kartik rustagi wrote:
>> > Did you delete it from the wiki, then? Or should I?
>>  I have deleted it.
>>
>> Their is one another idea listed at the wiki about Creating a Web
>> based image builder for Gentoo similar to SUSE studio. I discussed
>> this with some people and they find it to be interesting and useful to
>> increase Gentoo's adopt-ability further.  The back end can be done
>> mostly in Python and thus suits my skill set. The backend with quite a
>> lot of functionalities plus a basic front end might be just
>> appropriate for the time span of Gsoc.
>>
>> Request for comments ....
>>
>> Regards
>> Kartik Rustagi
>
> Backend being in python would be fine. Frontend should work more or less something like this mock up.
>
> http://tinderbox.dev.gentoo.org/image/
>
> Note these are binary feeds of packages that have already been built and
> not a frontend to run some random emerge for pkgs. However for pkgs that
> don't exist. We probably would want to make some sort of log so that we
> know which types of package jobs to dispatch to the misc bin builders
> that are out there now.
>
> --
> Ned Ludd <solar@g.o>
> Gentoo Linux
>
>