.NAME in .Fortran
On 12/09/12 10:50, Duncan Murdoch wrote:
On 12-09-11 4:44 PM, Rolf Turner wrote:
Duncan: I experimented by using ".NAME =" in a call to .Fortran() in one of my packages and got the same error as Peter. I am definitely (!) using R version 2.15.1.
Yes, Peter told me the same.
Doing args(.Fortran) gives:
function (.NAME, ..., NAOK = FALSE, DUP = TRUE, PACKAGE, ENCODING) NULL <bytecode: 0x9e7762c>
If I replace ".NAME" by "name" in my call to .Fortran() everything works
smoothly.
Consequently: It looks to me like there is a bug here. An
inconsistency
between .Fortran() and ".Primitive(".Fortran")" perhaps.
There was a bug; it was fixed in August (also without a NEWS entry). You shouldn't see it in R-patched, or 2.15.2, when that is released.
<SNIP>
I installed R-patched and experimented again. Now no error is thrown
when .Fortran()
is called with ".NAME=". However, neither is any error thrown when
..Fortran is called
with "name=".
Is this as it should be, or is there still something not quite right?
cheers,
Rolf