This seems to be a different problem.
First, I don't have an .Rprofile in the relevant Rstudio workspace. This anomaly occurs in all of my workspaces, none of which set the error option.
Second, the value of options()$error is not changed in the course of the commands listed in my original message above -- as soon as I open RStudio, after loading the libraries, and after the command creating "junk":
> options()$error
(function ()
{
.rs.recordTraceback(TRUE, 5, .rs.enqueueError)
})()
>
Third, the anomaly seems only to occur after using a command from the dplyr package, not after any other data commands. (It occurs after using any of group_by(), arrange(), select(), filter(), or mutate().) I think that there may be a bug in the dplyr package that was activated with the above change in R with 3.0.6. Should I report this to the dplyr bug URL?
Larry Hunsicker