The missingness of arguments is not preserved after once being captured in `...`.
check <- function(x,y,z) c(missing(x), missing(y), missing(z))
check(one, , three) #FALSE, TRUE, FALSE
check1 <- function(...) check(...)
check1(one, , three) #FALSE, TRUE, FALSE
check2 <- function(...) check1(...)
check2(one, , three) #FALSE, FALSE, FALSE
A bug indeed.
Note that I see the same in R 1.0.1 and R 2.2.1 ... so I guess
this has been with us since "the beginning of time" ;-)
and cannot have been crucial...
Nevertheless worth fixing.. (but probably not be me).
as we (well, ML) are currently working at fixing some '...' problems,
maybe we can address this one, too ?
Checked in a fix with r67166. Pretty new at this stuff, so extra eyes are appreciated.