R devel - Oct 2005 - R_MakeExternalPtr

Hi,

I'm using R_MakeExternalPtr() to store handles to (COM) objects in a SEXP.

The code basically is

	sexp = R_MakeExternalPtr(handle,R_NilValue,R_NilValue);
	R_RegisterCFinalizerEx(...);

After creating the sexp, LENGTH(sexp) returns some quite large integer
value. It seems like an "unitialized" value.

Can I safely assume, that an SEXP of type EXTPTRSXP can only contain a
single pointer value and never represent a vector (an array) of pointers? Or
should I maybe explicitly add something like

	SETLENGTH(sexp,1);

Best wishes from sunny Austria,

Thomas Baier

This is one of those undocumented things (AFAIK): LENGTH applies only to
'vectors'.  So don't use t and definitely don't set it for
others.  The
macro expands to

#define LENGTH(x)       (((VECSEXP) (x))->vecsxp.length)

What is a vector here?  Something which gets actually allocated by 
allocVector() AFAICS, that is an atomic or generic vector or an 
expression.  OTOH, an EXTPTRSXP is stored in a cons cell.

There is also a function length(), which is safer.

On Mon, 24 Oct 2005, TB wrote:
> Hi,
>
> I'm using R_MakeExternalPtr() to store handles to (COM) objects in a
SEXP.
>
> The code basically is
>
> 	sexp = R_MakeExternalPtr(handle,R_NilValue,R_NilValue);
> 	R_RegisterCFinalizerEx(...);
>
> After creating the sexp, LENGTH(sexp) returns some quite large integer
> value. It seems like an "unitialized" value.
>
> Can I safely assume, that an SEXP of type EXTPTRSXP can only contain a
> single pointer value
It stores a pointer plus other things including a tag.
> and never represent a vector (an array) of pointers? Or
> should I maybe explicitly add something like
>
> 	SETLENGTH(sexp,1);
>
> Best wishes from sunny Austria,
>
> Thomas Baier
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
>
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

Brian,

Prof Brian Ripley <mailto:ripley at stats.ox.ac.uk>
wrote:
> This is one of those undocumented things (AFAIK): LENGTH applies only
> to 'vectors'.  So don't use t and definitely don't set it
for others.
> The macro expands to
> 
> #define LENGTH(x)       (((VECSEXP) (x))->vecsxp.length)
> 
> What is a vector here?  Something which gets actually allocated by
> allocVector() AFAICS, that is an atomic or generic vector or an
> expression.  OTOH, an EXTPTRSXP is stored in a cons cell.
> 
> There is also a function length(), which is safer.
Thanks for your message. Things are much clearer now to me.

Can I safely assume that INTSXP, REALSXP, LGLSXP, STRSXP, CPLXSXP and VECSXP
are
vectors (while EXTPTRSXP is not).

Thomas