[Rd] Install R-patched_2006-07-13 on i386-pc-solaris2.10 with Sun Studio 11

Latchezar Dimitrov ldimitro at wfubmc.edu
Wed Jul 26 22:48:17 CEST 2006


Dear Prof. Ripley and R-developers:

Thank you very much for the reply. Please see bellow 

> -----Original Message-----
> From: Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk] 
> Sent: Wednesday, July 26, 2006 2:37 AM
> To: Latchezar Dimitrov
> Cc: r-devel at stat.math.ethz.ch
> Subject: Re: [Rd] Install R-patched_2006-07-13 on 
> i386-pc-solaris2.10 with Sun Studio 11
> 
> On Wed, 26 Jul 2006, Latchezar Dimitrov wrote:
> 
> > Dear R-developers:
> > 
> > I'm trying to build a 64-bit R-patched_2006-07-24 on 
> SunFire V40z with 
> > on Solaris OS 10 64-bit kernel and using Sun Studio 11 compilers.
> > Everything runs OK until it gets to building package tools (all.R) 
> > where it fails. Bellow is how I tried it (I can provide any other 
> > additional info if needed). Any help please?
> 
> You seem to have gcc 4.1.1 in your paths, so can you try that instead.

True (that's why I gave the env) however that is c/c++ only. I cannot
build gcc fortran right way and since R is on top of my list I switched
to Sun Studio 11. I included the libraries and the path in order to use
makeinfo and readline which I compiled with that gcc (I build most
essential gnu utils with it too as well as some applications and
everything seems ok). To avoid misuse of it I explicitly specified all
the programs that I knew could be mistaken (CC,CXX,etc), gnu ld was
renamed. I as sure as one can that there was not any (mis)use of
/usr/local/ except for readline and makeinfo (I believe the build did
not go that far though). One additional thing that I missed in my prev.
e-mail is that R binary was build seemingly properly, i.e., it starts ok
however with complaints about missing base parts (obviously).

Now to make it clear it has nothing to do with that gcc I removed all
the paths from the environment and to make sure nothing leaks through I
renamed /usr/local so it stays out of the way. Then I set
--with-readline=no and started clean (as you may have already noticed I
build in a separate from src dir (obj-R) completely empty at the
beginning. It failed the very same way.

Further, I decided to give 32-bit build a try. To make things more
interesting I restore /usr/local and the path to /usr/local/bin in order
to be able to use makeinfo (since the only readline I had was a 64-bit
one I gave it up). And voila, it built like a charm. The only tests that
I noticed failed were those involving tcltk (I had only 64-bit ones
installed).

The only problem still remaining is I cannot care less about 32-bit
version.

Please find attached the log of 64-bit try and a little bit of 32
success in the end.

So would someone please help me find out what is wrong and build my
favorite 64-bit R?

Thank you very much,
Latchezar Dimitrov


> 
> This looks like a fairly fundamental problem with your 
> current build, possibly a mis-compile.
> 
> > 
> > Thank you very much
> > 
> > Latchezar Dimitrov
> > Wake Forest Univ. School of Medicine
> > 
> > 
> > 
> > root at Huba #   echo   $PATH
> > 
> /opt/SUNWspro/bin:/usr/sbin:/usr/bin:/usr/openwin/bin:/usr/ccs/bin:/us
> > r/ 
> > 
> openwin/bin:/usr/dt/bin:/usr/platform/i86pc/sbin:/opt/SUNWvts/bin:/opt
> > /S
> > UNWexplo/bin:/usr/local/bin
> > root at Huba #   echo   $CC
> > cc
> > root at Huba #   echo   $CXX
> > CC
> > root at Huba #   echo   $CFLAGS
> > -xarch=amd64 -xmodel=medium
> > root at Huba #   echo   $CXXFLAGS
> > -xarch=amd64 -xmodel=medium
> > root at Huba #   echo   $FCFLAGS
> > -xarch=amd64 -xmodel=medium
> > root at Huba #   echo   $FFLAGS
> > -xarch=amd64 -xmodel=medium
> > root at Huba #   echo   $LDFLAGS
> > -xarch=amd64 -xmodel=medium
> > root at Huba #   echo   $R_BROWSER
> > /usr/sfw/bin/mozilla
> > root at Huba #   echo   $r_arch
> > amd64
> > root at Huba #   echo   $LD_LIBRARY_PATH
> > 
> /usr/openwin/lib:/usr/local/gcc-4.1.1-x86-bootstrap/lib/gcc/i386-pc-so
> > la 
> > 
> ris2.10/4.1.1/amd64:/usr/local/lib:/usr/local/gcc-4.1.1-x86-bootstrap/
> > li 
> > 
> b/gcc/i386-pc-solaris2.10/4.1.1/amd64:/usr/local/lib:/usr/loca
> l/gcc-4.1.
> > 
> 1-x86-bootstrap/lib/gcc/i386-pc-solaris2.10/4.1.1/amd64:/usr/local/lib
> > root at Huba #
> > 
> > 
> > ../src/R-patched_2006-07-24/configure
> > --prefix=/opt/R-2.3.1-patched_2006-07-24-Sun_Studio_11 
> --with-readline 
> > --disable-mbcs R_PAPERSIZE=letter --disable-rpath --with-bzlib 
> > --with-zlib --with-spcre --with-tcltk --disable-R-profiling 
> > --disable-nls
> > 
> > 
> > 
> > Error in parseNamespaceFile(package, package.lib, mustExist 
> = FALSE) : 
> >         subscript out of bounds
> > Execution halted
> > *** Error code 1
> > The following command caused the error:
> > (cat
> > 
> ../../../../src/R-patched_2006-07-24/src/library/tools/R/makeLazyLoad.
> > R;
> > \
> >   echo "makeLazyLoading(\"tools\")") | \
> >   R_DEFAULT_PACKAGES=NULL LC_COLLATE=C ../../../bin/R --vanilla 
> > --slave
> > > /dev/null
> > dmake: Fatal error: Command failed for target `all'
> > Current working directory /export/data/obj-R/src/library/tools
> > *** Error code 1
> > The following command caused the error:
> > for pkg in profile base tools utils grDevices graphics 
> stats datasets 
> > methods grid splines stats4 tcltk; do \
> >   (cd ${pkg} && dmake) || exit 1; \
> > done
> > dmake: Fatal error: Command failed for target `R'
> > Current working directory /export/data/obj-R/src/library
> > *** Error code 1
> > The following command caused the error:
> > for d in scripts include extra appl nmath unix main modules 
> library; 
> > do \
> >   (cd ${d} && dmake R) || exit 1; \
> > done
> > dmake: Fatal error: Command failed for target `R'
> > Current working directory /export/data/obj-R/src
> > *** Error code 1
> > The following command caused the error:
> > for d in m4 tools doc etc share src tests; do \
> >   (cd ${d} && dmake R) || exit 1; \
> > done
> > dmake: Fatal error: Command failed for target `R'
> > root at Huba #
> > 
> > ______________________________________________
> > R-devel at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
> > 
> > 
> 
> -- 
> 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
> 


More information about the R-devel mailing list