[R] .NAME in .Fortran

[Rolf Turner]

I don't understand at all what the problem is, but a workaround
is simply not to specify the "pdf" argument in "name=value" form
but rather in positional form.  I.e. just make it the first argument
to .Fortran().  As in:

     tmp <- .Fortran("pdf", <other parameters here>)

Works for me! :-)

     cheers,

         Rolf

On 11/09/12 13:21, Peter Dunn wrote:
> Hi all
>
> I've been getting some emails from the R package maintainers that I need
> to update some code in a CRAN packge that uses FORTRAN, to comply with
> (not so recent) changes.  I've been a little busy...
>
> I'm having trouble adjusting my code.  I hope someone can help.
>
> The package was working fine, and a few R functions in my package had
> lines like this:
>
> tmp <- .Fortran( name="pdf", <other parameters here>)
>
> and similar.  The "pdf" is the name of the FORTRAN subroutine being
> called, of course.
>
> When I replace that line with
>
> tmp <- .Fortran( .NAME="pdf", <other parameters here>)
>
>
> ...the package then does not pass the checks.  The library package builds
> OK with no errors or warnings, and then during "R CMD check" the examples
> give this error:
>
> out <- tweedie.profile( y~1, p.vec=seq(1.1, 1.9, length=9),
> do.plot=TRUE)
> 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
> .Error in .Fortran(.NAME = "pdfun", as.double(power),
> as.double(phi[i]/(y[i]^(2 -  :
>    supplied argument name '.NAME' does not match 'name'
> Calls: tweedie.profile -> dtweedie.inversion -> .Fortran
> Execution halted
>
>
>
>
> I've had a search--including a read of the "Writing R Extensions" manual
> on CRAN (I did read it, but interestingly a search of that document did
> not reveal any instance of .NAME)--and I cannot see the problem.  It's
> probably obvious.
>
> Any help appreciated, even pointers to more info.

     <SNIP>