Road map for rpm-python please ...

Jeff Johnson <[email protected]> Wed, 15 Oct 2003 13:38:50 -0400
Newsgroups gmane.linux.redhat.rpm.python
Message-ID <[email protected]>
OK, enough whining. I can bash my bits better than any
of you. ;-)

Let's try to make this list positive and productive please, shall we?

Up until now, I have served the role of "maintainer", but not "architect",
of rpm-python. There are important differences between the two roles, duh.

It's time for a change methinks. Here are specific suggestions:

1) Who wants to be "architect" of rpm-python?

     I assure you, rpm-python maintainership is available for the 
asking. Lord knows
    you can/will be able to do a better job than me.

    Any takers?

    If not, then I claim the role of architect, because I am certainly 
the architect of
    rpm-python n my "RPM Package Manager" OSS fork already, although my 
community
    currently exists of exactly one person, me. I can and will continue 
making up sh*t on
    my  code base no matter what.

2) What road map for rpm-python?

   I can suggest several avenues to the future, including:

       a) forking the existing rpm-python bindinsg into per-application 
(i.e. up2date/beehive/yum/...)
       bindings. This gives each application ultimate control over their 
own Python API. Yes, there
       are negatives, but -- having survived the librpm404 fork -- I 
believe the engineering issues
       ar manageable.

       b) freexing the de facto incarnations of rpm-python forever more. 
This is essentially
       what is happening now anyways because of my inability to release 
rpm in any
        comnprehensible or timely fashion. The only new addition is to 
make "freeze"
       official, whatever that means.

        c) attempting new -- python native -- bindings.

        d) splitting the existing rpm-python bindings into per-subsystem 
conponents

   and probably 5 or 6 other approaches to the engineering problems of 
continued rpm-python
    development.

So, comments please. rpm-python needs an architect and a road map, it's 
impossible
to address any deficiencies in the existing rpm-python bindings without 
those essential items
any more. If you re-read my rants you will find hints at road map issues 
that need to be addressed.

73 de Jeff