This issue was filed in R bugs and addresses by Prof Brian Ripley. I had
neglected to name the data argument, but Brian checked in a fix to R-devel so
folks who make the same mistake I did will know they supplied the wrong object
type to the specials argument in the terms function.
Cheers,
Pat
I stumbled across one of those intermittent bugs where the code
sometimes works and sometimes doesn't. Below is an example run showing
a failure where a CHARSXP object was not properly created (by the sub
function?). I was able to reproduce this error on three different R
builds in fresh R ses
