Re: [Rd] Problem with compiling - Freebsd on AMD64 (PR#7489)

From: Kelly Black <kjblack_at_gmail.com>
Date: Tue 11 Jan 2005 - 02:59:08 EST

Dear Prof's Ripley and Dalgaard ,

You are correct; this is a compiler error. I assumed that since it was an easy fix to the problem by just changing the compiler option to -O that it could save some headaches for you. As far as the version goes, it is gcc v. 3.4.2 built for freebsd on the amd64:

black@compaqlaptop$ gcc -v
Using built-in specs.
Configured with: FreeBSD/amd64 system compiler Thread model: posix
gcc version 3.4.2 [FreeBSD] 20040728

Finally, the output was with "-g -O2". Using -0 compiles and installs with no problems.

Sincerely,
Kel

On Mon, 10 Jan 2005 15:01:52 +0000 (GMT), Prof Brian Ripley <ripley@stats.ox.ac.uk> wrote:
> In what sense is this a bug in R? If changing your compiler optimization
> leads to a segfault, it looks very like a compiler bug. So what was the
> compiler?
>
> [R does work using AMD64 on Linux, but only correctly with gcc >= 3.3.3.]
>
> BTW gdb output with "-g -O2" (the R default) ought to be informative, so
> was this with -O?
>
> On Mon, 10 Jan 2005 kjblack@gmail.com wrote:
>
> > Full_Name: Kelly Black
> > Version: 2.0.1
> > OS: FreeBSD 5.3 on AMD64
> > Submission from: (NULL) (149.106.32.93)
> >
> >
> > I downloaded R-2.0.1 from the R website, ran configure then make and got
> > an error. I will paste a copy of the output from the makefile below. The
> > problem was when the makefile tried to run the R binary. When I came
> > across the error I reran configure, but this time with the following
> > arguments:
> >
> > ./configure FFLAGS="-g" CFLAGS="-g" CXXFLAGS="-g"
> >
> > This time when I ran make it compiled with no problems. I tried the same thing
> > again only with all of the flags set to "-O" and again it compiled with no
> > problems. When I ran it with the flags set to "-g -O2" it had the same error. A
> > copy of the output from gdb can be found at
> > http://blackk.union.edu/~black/freebsd/R-error.txt, but the output isn't very
> > helpful.
> >
> > The system used is a laptop, Compaq Presario (amd64), under FreeBSD 5.3 compiled
> > for the amd64 architecture.
> >
> > Output from make with flags "-O2"
> >
> > black@compaqlaptop$ pwd
> > /tmp/R-2.0.1/src/library/base
> > black@compaqlaptop$ make
> > echo "building package 'base'"
> > building package 'base'
> > /usr/local/bin/bash ../../../src/scripts/mkinstalldirs ../../../library/base/R
> > (f=${TMPDIR:-/tmp}/R$$; cat `LC_COLLATE=C ls ./R/*.R ./R/unix/*.R` >> ${f};
> > /usr/local/bin/bash ../../../tools/move-if-change ${f} all.R)
> > all.R is unchanged
> > /usr/local/bin/bash ../../../tools/copy-if-change all.R
> > ../../../library/base/R/base ${f}
> > ../../../library/base/R/base is unchanged
> > for f in COPYRIGHTS INDEX SOURCES; do if test -f ./${f}; then
> > /usr/bin/install -c -m 644 ./${f} ../../../library/base; fi; done
> > if test -f DESCRIPTION; then /usr/bin/install -c -m 644 DESCRIPTION
> > ../../../library/base; (tmp="Built: R 2.0.1; "; if test -d ./src; then
> > tmp="${tmp}x86_64-unknown-freebsd5.3"; fi; tmp="${tmp}; `date`; unix"; echo
> > "${tmp}") >> ../../../library/base/DESCRIPTION; fi
> > /usr/local/bin/bash ../../../src/scripts/mkinstalldirs
> > ../../../library/base/demo
> > for f in `ls -d ./demo/* | sed -e '/CVS/d' -e '/00Index/d'`; do
> > /usr/bin/install -c -m 644 ${f} ../../../library/base/demo; done
> > /usr/local/bin/bash ../../../src/scripts/mkinstalldirs
> > ../../../library/base/man
> > (f=${TMPDIR:-/tmp}/R$$; for rdfile in ./man/*.Rd ./man/unix/*.Rd; do echo "%
> > --- Source file: ${rdfile} ---"; cat ${rdfile}; echo '\eof'; done >> ${f};
> > /usr/local/bin/bash ../../../tools/move-if-change ${f}
> > ../../../library/base/man/base.Rd)
> > rm -f ../../../library/base/man/base.Rd.gz
> > /usr/bin/gzip ../../../library/base/man/base.Rd
> > /usr/bin/install -c -m 644 ./inst/CITATION ../../../library/base
> > /usr/bin/install -c -m 644 all.R ../../../library/base/R/base
> > cat ./makebasedb.R | R_DEFAULT_PACKAGES=NULL LC_COLLATE=C ../../../bin/R
> > --vanilla --slave > /dev/null
> > Floating point exception (core dumped)
> > *** Error code 136
> >
> > Stop in /tmp/R-2.0.1/src/library/base.
> >
> > ______________________________________________
> > R-devel@stat.math.ethz.ch mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
> >
> >
>
> --
> Brian D. Ripley, ripley@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
>

-- 
_______________________________________________________          __o
Kelly Black                     Phone: (603) 862-3587  \       _`\<,_
Department of Math. & Stat.     FAX:   (603) 862-4096   \_____(_)/ (_)
University of New Hampshire     e-mail: kelly.black@unh.edu
Durham, NH 03824  (USA)         WWW: http://www.math.unh.edu/~black

______________________________________________
R-devel@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Received on Tue Jan 11 02:04:09 2005

This archive was generated by hypermail 2.1.8 : Tue 11 Jan 2005 - 02:23:10 EST