Madwifi/wpa_supplicant packaging

[Philip Prindeville]

I was looking at the way wpa_supplicant was packaged, and wondered
why not make it dependent instead on madwifi-devel or madwifi-headers
instead (and tweak that src.rpm accordingly as well).  Or it could
optionally detect the presence of the madwifi headers, and enable
CONFIG_DRIVER_MADWIFI=y automatically...

I like the functionality of the FC5 wpa_supplicant, but it doesn't play well
with the madwifi rpm's from ATrpms.

That way if you grab madwifi from ATrpms, and rebuild wpa_supplicant
from source it will do the right thing...

Seem reasonable?

-Philip