[Rd] Problem testing R version 1.3.1 (PR#1143)

Peter Dalgaard BSA p.dalgaard@biostat.ku.dk
26 Oct 2001 18:54:51 +0200


Gordon Lack <gml4410@ggr.co.uk> writes:

> > >    I have compiled R version 1.3.1 on a Linux system.
> >...
> > Which one? In particular, which compiler version?
> 
>    Redhat6.2 with:
> 
> gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
> 
> > Unfortunately, the output is buffered so that one on catastrophic
> > error will lose the last bit. It might be more informative if you ran the
> > command directly, without the redirection to base-Ex.Rout.
> 
>   Good point.  The error at the end is now:
> 
> =====
> > ##--- An example for the use of  `asp' :
> > library(mva)
> > data(eurodist)
> > loc <- cmdscale(eurodist)
> Segmentation fault
> =====
> 
> 
> Martin Maechler wrote:
> 
> > No,
> > but since you're on Linux you can find much more yourself :
> > Run R inside the debugger:
> >...
> >  but you'll see more
> 
>    Yes, thanks.  strace also showed a SIGSEGV on loading lapack.so, but
> gdb gives a little more info.
> 
> 
> 
> Program received signal SIGSEGV, Segmentation fault.
> _dl_lookup_versioned_symbol (
>     undef_name=0xa4ad55dc <Address 0xa4ad55dc out of bounds>,
> ref=0xbfffd2c8, 
>     symbol_scope=0x934f71c, 
>     reference_name=0x8ae7ce0
> "/proj/informatics-build/GML/R/R-1.3.1/modules/lapack.so",
> version=0x8e46c90, reloc_type=112) at ../sysdeps/generic/dl-hash.h:31
> 31      ../sysdeps/generic/dl-hash.h: No such file or directory.
> 

Odd. There are known issues with stock RH 6.2, but base R usually
compiles and passes make check on those systems. What usually happens
is that some packages refuse to dynload, with the error you're
observing, essentially.

The culprit is somewhere in the compiler suite - 2.95.x is known to
work, and I believe 3.0.x too (whereas the unofficial 2.96 from RH 7.0
caused trouble although later ones for 7.1 and 7.2 seem OK). You might
as well upgrade since R-1.4.0 will definitely not work with that
setup.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._