> > To close out this issue, I propose that the TLS spec forbid > negotiating to NULL_WITH_NULL_NULL. I understand > the argument for testing, but I suspect the risks of this in > practical deployment make it dangerous. > > Win Treese I concur.