[R] Rf_error crashes entire program.

[Prof Brian Ripley]

Rf_error is used so often in R itself and packages that it is almost 
certainly not the problem -- rather something else your program has done 
has damaged R's internals (e.g. overrrun an array).

'Writing R Extensions' discusses how to debug R code, including foreign 
code.  For example, I would run under valgrind to look for incorrect 
memory accesses.

This was an appropriate topic for R-devel -- please DO study the posting 
guide.

On Mon, 28 Jul 2008, Andrew Redd wrote:

> I'm having a problem with the error and warning functions.  I've tried this
> on multiple machine so I'm fairly sure it's not machine dependent and I've
> tried it on the latest versions 2.6.0-2.7.1.  Whenever my program gets to an
> error or warning it crashes the entire program rather than throwing the
> error like it should.
>
> Here are the relevant bits.
>
> ...
>    char const * const ExeedsMinVarianceError = "PFDA ERR: Near zero
> variance encountered.  Estimation Unstable. Terminating Estimation.";
> ....
>        if(debug){printf("Da:\n");printmat(DaOld,1,*ka);fflush(stdout);}
>        daxpy_(ka, &mOne, Da, &one, DaOld, &one);
>        for(i=0;i<*ka;i++)convergenceCriteria+=fabs(DaOld[i]);
>        if(Da[*ka] < MinVariance){
>            printf("PING");fflush(stdout);
>            warning(ExeedsMinVarianceError);
>            break;
>        }
>
> and the output from running in batch mode that I get is this:
> Da:
> 0.1803    4.988e-017
>
> PING
> and here the program crashes.   I've tried this in multiple places and
> sometimes the error is thrown sometimes not.  Does anyone have an idea of
> what is going on.  I could not find any discussion of this error yet.
> Thank you,
> Andrew Redd
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at 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.
>

-- 
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