Comment 50 for bug 122703

I guess after more thought, I don't understand why we can't just package the madwifi/ath5k trunk and only use it for known-incompatible chipsets with the current package (or further, known compatible chipsets with the trunk package). "Oh, this is a [5008|5212|5418|etc], I'll use the trunk package, it can't be worse!"

This of course assumes 1) it is easy to detect a chipset, 2) it is easy to have package logic based on chipset and 3) it actually can't be worse. Point 3 however can be mitigated by only using it for whitelisted chipsets.

Is something like this possible?