[Rpm-metadata] more updates - filelists and more! :)

[seth vidal]

Hi all,
 Script posted again:
    http://linux.duke.edu/~skvidal/metadata/rpm-md-dump.py
 
 Changes:
   - made checksum have an id attribute
   - create filelists.xml
   - stuff in place for 'other'
   - just replace non-utf8 characters with a '?' - not noting it in the
     xml, yet, will add shortly.
   - cleaned up a few things in generateXML to make the function less
repetitive in silly places
   - made empty obs/conf/reqs/provides not be listed

new output files:
http://linux.duke.edu/~skvidal/metadata/base-sample.xml
http://linux.duke.edu/~skvidal/metadata/filelists-sample.xml

 - filelists lists package n,e,v,r,a in a comment - I'm of two minds on
this - one is - if it's in a comment we're paying the price for it being
there in terms of data xfer, why not just specify it too. The second is
- eh - don't make people parse it if it is not necessary.

Open to suggestions :)

All of fedora core test 3 (severn) - rpms and src.rpms:
 base:  http://linux.duke.edu/~skvidal/metadata/base.xml.bz2 - 395K
 files: http://linux.duke.edu/~skvidal/metadata/filelists.xml.bz2 - 1.3M

uncompressed they are 5.2M and 20M, respectively.

I'll probably work on the changelog/etc/ file tomorrow night.

How do those look?

Thanks,
-sv