[R] R-Embedding and error messages

[Jorge Cardoso]

I tried both solutions but I always get
"undefined symbol"  when executing

I have checked my Makefile and I'm sure
that -lR is present when compiling

I did a grep for the R-sources and verified
that both functions exists

I will try to post it in r-devel list.

Thank for yours responses dear Martin and Luke.

Jorge


El Mié 20 Ago 2008, Luke Tierney escribió:
> On Wed, 20 Aug 2008, Martin Morgan wrote:
> > Hi Jorge --
> >
> > I hesitate to suggest this unsupported solution (as these public
> > suggestions tend to mean that the solution disappears!), but in C
>
> Don't use this -- it is subject to change/removal with no notice.
>
> I do not believe there is a C level methanism for catching the R error
> conditon at this point.  Thinking about this is on the list but it
> hasn't happened yet.  One thing you can do is use the geterrmessage R
> function to retrieve the most recent message (so if R_tryEval
> indicates an error occurred then build and execute a call to
> geterrmessage and extranct the message from the returned STRSXP).
>
> luke
>
> > extern void R_SetErrorHook(void (*)(SEXP, char *));
> >
> > void
> > my_ErrorHook(SEXP call, char *message)
> > {
> >
> >   /*
> >        etc; e.g., cache message, perhaps reset error hook (hook is
> >        removed after error; nested R-level try() may mean multiple
> >        errors per R_tryEval?
> >   */
> >
> > }
> >
> > and then
> >
> > ...
> >
> >    R_SetErrorHook( my_ErrorHook );
> >    PROTECT(value = R_tryEval( expr, env, errorOccurred ));
> >    if( *errorOccurred ) {
> >        /* etc, e.g., retrieve error message(s) from cache */
> >    }
> >    /*
> >        hmm, maybe R_SetErrorHook needs to be cleared here, or at
> >        least replaced with a no-op?
> >    */
> > ...
> >
> > This is an R-devel question, not R-help.
> >
> > Martin
> >
> > "Jorge W. Cardoso" <cardoso en montevideo.com.uy> writes:
> >> I'm writing a C++ application using R-embedding to do some
> >> forecast process.
> >>
> >> I also use R_tryEval instead of R_Eval to run my R-script,
> >> so that in case of error I know exactly in which line number
> >> was the last error.
> >>
> >> In particular, from time in time y get some error messages
> >> refering an exceptional condition in the arima model.
> >>
> >> I want to catch these error messages from de C code
> >> (or at least the code error number) to take some
> >> corrective actions.
> >>
> >> Is there any way to do that ?
> >>
> >> Regards,
> >>
> >> Jorge W. Cardoso
> >>
> >> ______________________________________________
> >> R-help en r-project.org mailing list
> >> https://stat.ethz.ch/mailman/listinfo/r-help
> >> PLEASE do read the posting guide
> >> http://www.R-project.org/posting-guide.html and provide commented,
> >> minimal, self-contained, reproducible code.