[Rd] R 1.4.0 build fails on AIX (PR#1289)

[Dan Million]

On Wed, 30 Jan 2002, Thomas Lumley wrote:

> I've just compiled R 1.4.1 on AIX 4.3 with
>
> bash% CC=cc FC=f77 ./configure
> bash% make
>
> It all seems to work (I'm running make check, but the computer is very
> slow).

I've just built R 1.4.1 on AIX 4.3 myself, and while I've made progress,
I still have problems.  If I could bother you a bit more, I'd like to
compare some notes.

My environment is as follows:

    RS/6000 model H70 (PowerPC_RS64-II) running AIX 4.3.3 ML09; configure
	calls it "powerpc-ibm-aix4.3.3.0"
    IBM C for AIX Compiler 5.0.2.2
    IBM XL Fortran Compiler 7.1.1.1
    Perl 5.005 patch 03
    makeinfo 4.0
    readline 4.2
    zlib 1.1.3
    libjpeg 6b
    libpng 1.2.0
    tcl/tk 8.3.3
    GNU make 3.79.1

This is what I started with in my config.site file:

    R_PAPERSIZE=letter
    CC=cc
    CFLAGS="-g"
    FC=f77
    CXX=g++
    MAKE=/opt/freeware/bin/make

I originally configured with "configure --enable-R-shlib" (because my
R user thought he needed it), and I gradually concluded that this was
causing lots of my problems, so I dropped it.  Now I just use "configure"
with no options.

R now runs (sort of).  It still coredumps on exit with segmentation
faults.  Sometimes it also coredumps during operation with an illegal
instruction fault.  I did run through "make check" using the expedient
of hacking the Makefiles to ignore exit errors, so it would plow ahead
even after getting the exit coredumps, and it looks like it got through
all the tests.

My user also wanted to install the Rggobi package into R, which I've
done.  Ggobi itself seems to work fine, but when I type the "ggobi"
command in R, I get an illegal instruction coredump.  Not sure if that's
an R problem or a Ggobi problem.

Do you see anything in my environment that is "wrong"?  What level of
AIX (version and maintenance level) are you building on?  What model of
RS/6000?

Thanks for your time.

Dan


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._