[R] core dump with point.in.polygon

Roger Bivand Roger.Bivand at nhh.no
Fri Sep 1 08:56:48 CEST 2006


On Fri, 1 Sep 2006, Prof Brian Ripley wrote:

> On Thu, 31 Aug 2006, Jeff D. Hamann wrote:
> 
> > I've been trying to get some code to run on a 64-bit FreeBSD machine running
> > 
> > R : Copyright 2006, The R Foundation for Statistical Computing
> > Version 2.3.0 (2006-04-24)
> > ISBN 3-900051-07-0
> > 
> > and keep getting a core dump with the following results:
> > 
> >  *** caught segfault ***
> > address 0x0, cause 'unknown'
> > 
> > Traceback:
> >  1: .Call("R_point_in_polygon_sp", as.numeric(point.x),
> > as.numeric(point.y),     as.numeric(pol.x), as.numeric(pol.y), PACKAGE =
> > "sp")
> >  2: sp::point.in.polygon(point.x = stems$easting, point.y =
> > stems$northing,     pol.x = get.pts(gpc.plot)[[1]]$x, pol.y =
> > get.pts(gpc.plot)[[1]]$y)
> >  3: as.logical(sp::point.in.polygon(point.x = stems$easting, point.y =
> > stems$northing,     pol.x = get.pts(gpc.plot)[[1]]$x, pol.y =
> > get.pts(gpc.plot)[[1]]$y))
> >  4: poi.fp(gpc.perim, plot.grid[i, ], stems, plot.designs[1, ], plot.it =
> > FALSE)
> >  5: eval.with.vis(expr, envir, enclos)
> >  6: eval.with.vis(ei, envir)
> >  7: source("generate_samples.r")
> > 
> > Not sure what to do now. I've been running this on 2.3 Winxp box just fine
> > and will continue to do so.
> 
> There are contributed packages involved here: please identify which and 
> report this to the maintainer with a reproducible example.

Jeff:

Could you please provide a case with the code of your function and the 
data that is going to R_point_in_polygon_sp? Please also consider the 
R-sig-geo list (or the Sourceforge R-spat-dev list on r-spatial) in 
addition to writing to Edzer Pebesma and me as maintainers?

If it helps, we can use the sourceforge CVS repository to try to nail this 
down. Running R -d gdb would be helpful.



> 
> There are 10 or so known packages which seqfault on 64-bit platforms: the 
> authors have been informed (often long ago).  This looks as if it might be 
> 'sp', which is not one.
> 
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no



More information about the R-help mailing list