R devel - Mar 2008 - (PR#11054) "Writing R Extensions": bad example with CAR /

But it is not taken 'verbatim from src/main/print.c' (at least not in
that
version of R), and the code is not run with USE_RINTERNALS defined when 
write-barrier checking is enabled.

The example has been updated to match the current code in 2.7.0 alpha.


On Sun, 30 Mar 2008, sanders at fs.tum.de wrote:
> Full_Name: Simon Anders
> Version: 2.6.2
> OS: Ubuntu Linux
> Submission from: (NULL) (86.22.75.91)
>
>
> This is a rather minor "documentation bug", certainly not at all
urgent.
>
> The manual "Writing R extensions" explains in section 5.10
("Evaluating R
> expressions from C") how to use the eval function. Without going into
much
> details, an example is given how to build up a LANGSXP pairlist to
represent an
> expression.
>
> This example is taken at verbatim from src/main/print.c, and uses CAR and
CDR as
> l-values, e.g.
>   CAR(t) = install("print")
>
> This has just confused me quite a bit, as my own code, which I wrote
following
> this example, did not compile, complaining about 'CAR(t)' being
illegal as
> l-value.
>
> After a while I figured out the reason, which is (as you probably have
spotted
> immediatly) that the example is taken from a code snipped that imports
> Rinternals.h with defined USE_RINTERNALS, while I followed the manual's
advice
> to not define USE_RINTERNALS for extensions.
>
> Maybe you could add a note in this section of the manual that one should
use
> SETCAR and SETCDR instead. A more illustrative example would be even
better.
>
>
> Just in case you want to use it as example, here is my code. I hope it is
> correct in your expert eyes.
>
>   /* Remove the variable with name s from the environment ev,
>      i.e. construct the call: rm( list=s, envir=ev ) */
>
>   SEXP call;
>   PROTECT( call = Rf_allocList( 3 ) );
>   SET_TYPEOF( call, LANGSXP );
>
>   SETCAR( call, install( "rm" ) );
>
>   SETCAR( CDR(call), allocVector( STRSXP, 1 ) );
>   SET_STRING_ELT( CAR( CDR(call) ), 0, mkChar( s ) );
>   SET_TAG( CDR(call), install("list") );
>
>   SETCAR( CDDR(call), ev );
>   SET_TAG( CDDR(call), install("envir") );
>
>   eval( call, R_GlobalEnv );
>   UNPROTECT( 1 );
>
> ______________________________________________
> 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

On Mon, 31 Mar 2008, Simon Anders wrote:
> Dear Prof Ripley,
>
> Prof Brian Ripley wrote:
>> But it is not taken 'verbatim from src/main/print.c' (at least
not in that
>> version of R), and the code is not run with USE_RINTERNALS defined when
>> write-barrier checking is enabled.
>> 
>> The example has been updated to match the current code in 2.7.0 alpha.
>
> I just assumed that it is copied verbatim without checking because it says
so
> in the sentence above the code ("As a more comprehensive example of 
> constructing an R call in C code and evaluating, consider the following 
> fragment of printAttributes in src/main/print.c.")
That was true once, long ago.
> I don't understand, however, what you mean by stating that
USE_RINTERNALS is
> not needed (and I have to admit that I have no knowledge of this 
> 'write-barrier checking' mechanism.)
I didn't say 'USE_RINTERNALS is not needed' for your code: I said it
was
not used for the code that was run.

...

-- 
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