
YUM metadata parser written in C.

* Why? 
The biggest complaint people have with YUM is often the performance of parsing
the metadata. This implementation should be ~10 times faster, parsing the
primary.xml file under 1 second usually, filelists.xml under 3 seconds and
other.xml under 4 seconds. It uses a lot less memory as well, some testings
I have done show it uses ~4mb instead of 40mb standard YUM uses.

* How?
Should be really easy:
python setup.py build
sudo python setup.py install --prefix=/usr

(Assuming you python prefix is /usr).

The next time you use yum, it regenerates the sqlitecache because the database
schema is slightly different.