[R] make check on Solaris 8 fails due to plot

[Prof Brian Ripley]

On Tue, 16 Mar 2004, Adaikalavan Ramasamy wrote:

> Many thanks to Prof.Ripley and Dr.Dalgaard. Yes, I am using gcc-3.2.2
> (correctly configured for solaris 2.8) and this could be the problem.

It is a problem.

> Is there a way to fix this myself ? My machine specs are 'SunOS enterprise
> 5.8 Generic_108528-23 sun4u sparc SUNW,Ultra-Enterprise'.
> 
> The R-admin seems to suggest that I can solve this by redefining the CC,
> FFLAGS, CXXFLAGS and LDFLAGS in the config.site file but I am not sure if I
> am reading it correctly and am unable to find the config.site file.

You can probably solve it by turning optimization off (via CFLAGS and
FFLAGS), but a better way to solve it is to install/build a working
compiler: gcc-3.2.2 is quite old now, too.

config.site is in the top level of the sources.

> > -----Original Message-----
> > [mailto:r-help-bounces at stat.math.ethz.ch]On Behalf Of Prof Brian Ripley
> > Sent: 16 March 2004 07:13
> >
> > See the R-admin manual: this happens for certain broken versions of
> > Solaris-sparc gcc.  If that is not the cause we would need to know much
> > more about what you used to build R.
> >
> > We do regularly test on Solaris 8, with gcc 3.3.3 and Forte 7.
> >
> > On Mon, 15 Mar 2004, Adaikalavan Ramasamy wrote:
> >
> > > Dear all,
> > >
> > > I am having trouble trying to install R-1.8.1 on a Sun Solaris 8
> > > (Generic_108528-23 version) machine. The configuration was
> > successful but
> > > make check fails. I traced the the problem to the plot() function.
> > >
> > > > 1 + 1
> > > [1] 2
> > > > capabilities()
> > >     jpeg      png    tcltk      X11    GNOME     libz http/ftp  sockets
> > >     TRUE    FALSE     TRUE     TRUE    FALSE     TRUE     TRUE     TRUE
> > >   libxml     fifo   cledit  IEEE754    bzip2     PCRE
> > >     TRUE     TRUE     TRUE     TRUE     TRUE     TRUE
> > > > plot( 1:10 )
> > > Bus Error
> > >
> > > and R terminates to the shell prompt. I have searched the
> > archives with no
> > > success. Can anyone kindly help me fix the problem or at least
> > show me how
> > > to get more information about the error.

-- 
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