Re: RE: rpmUtils cvs repository

"Adam T. Gautier" <adam_gautier-/[email protected]> Tue, 23 Mar 2004 14:27:13 +0000 (UTC)
Newsgroups gmane.linux.redhat.rpm.python
Message-ID <[email protected]>
>Cool - it'd be great if you could discuss your thoughts here.  I'm aware that
>currently rpm-python maps closely onto rpmlib rather than what people might
>expect.  On the other hand rpm-python's use means that I still need to preserve
>that atm.  Making it easy to layer is being done atm.
>  
>
Done, I think the rpm-python should be very closely mapped to rpmlib(And 
probably stay that way), That gets rpmlib into python and then 
abstraction(s) from that can be made rpm-objects or what ever.  I 
actually like the complexity that is currently there becasue the RPM 
system is somewhat complex and should anyone what do anythin directly 
the need to seriously look into the C code and rpmlib itself.  I don't 
thaink that should change.

I need to build for another project higher level concepts and need 
things that are well documented inside of python an are virtually safe 
from my mucking around (I don't want to corrupt my rpmdb).  the current 
mapping with a good abstraction should provide that for me.

>Also taking all three approaches (rpmlib, functions, objects) means we have
>room for discussion.
>
>  
>
Agreed, I lean towards objects for reuse, I also have more experience 
with OOP. and the function sets are somewhat scattered.

>>This is just after a quick glance of SV's rpmUtils 
>>and I plan to dive into that code more.   I think the most important 
>>thing after an easy to user rpm-utility module would be to have simple 
>>example code for Add, Delete, Signature and Searching of the RPM 
>>database.  
>>    
>>
>
>Documentation (docstr and examples) is known issue.  Some of that exists in
>doxygen but I guess that may not be the expected place.  Duplicating \class in
>the class level docstrings is something I've wondered about - particularly if I
>can do via pre-processing.  The other side is getting this stuff into max-rpm
>too.
>
>  
>
Agreed,  This should be far simpler.  I definatly plan on writing some 
documentation/HOWTO as this goes along.

>>Also, what would it take to get ANY rpm-utility module added 
>>to the rpm-python*.rpm used by distros?
>>    
>>
>
>My roadmap is to provide more in python than done presently.   This list and
>discussions here with code and patches will definitely get considered.  
>
>As well as go through what is there and missing from python->rpmlib.  My stuff
>will go into HEAD, sane stuff into rpm 4.3 - but I have no control on that, but
>if it's sane it should happen.  What other distros take is up to them, but I'd
>like to think they'd move with the times.  
>
>  
>
Just wanted to make sure someone was listening to this mailing list or 
if it needed to be posted elsewhere.