[Rd] 3.5 on Tiger

Martin Maechler maechler at stat.math.ethz.ch
Mon Aug 2 10:15:21 CEST 2004


>>>>> "Jan" == Jan de Leeuw <deleeuw at stat.ucla.edu>
>>>>>     on Sun, 1 Aug 2004 10:25:24 -0700 writes:

    Jan> setenv CC gcc-3.5
    Jan> setenv F77 gfortran
    Jan> setenv CXX g++-3.5

    Jan> What I had to do.

(for what version of R exactly)

    Jan> 1. Build a local zlib-1.2.1

    Jan> 2. Compile ppr.f in stats with g77-3.4 and not with gfortran-3.5

(why -- a bug in gfortran-3.5 ? )

    Jan> 3. Change line 1581 in plot.c to

    Jan> double *aa = REAL(cex);
    Jan> if (R_FINITE(aa[i % ncex])

line 1582 {R-devel} is

		if (R_FINITE(thiscex = REAL(cex)[i % ncex])

would this work?

    Jan> i.e. get the assignment out of the macro

    Jan> 4. Disable the DO_STR_DIM macro in plot.c (and thus the
    Jan> functions do_strheight and do_strwidth). This could be fixed by
    Jan> expanding the macros and pasting into the source.

??? why that??

I wrote those quite carefully IIRC.


Regards,
Martin Maechler



More information about the R-devel mailing list