> 	This shows another defficiency in sup... that it could let
> such a thing happen in the first place... and then fail to correct it
> later.
> 
> Dave.
> 

I had the same problem yesterday with the pcvt_* changes that were
commited a few days ago.