signedness of wireless keys.

[Robert Love <rml novell com>]

Just a thought, as we move to libnm-util, we should settle on a
signedness of the char* we use for wireless keys.  Sometimes we use
char*, sometimes we use unsigned char*.  But the bigger problem is that
we mix the types via e.g. return values.

You can see all of the mismatches if you turn on the appropriate compile
flags.

Signed or unsigned?

	Robert Love