a strange logical bug (PR#162)

[maechler@stat.ethz.ch]

>>>>> "PD" == Peter Dalgaard BSA <p.dalgaard@biostat.ku.dk> writes:

    PD> ripley@stats.ox.ac.uk writes:
    >> On RH 5.2
    >> 
    >> The isnan() function returns a non-zero value if value is
    >> "not-a-number" (NaN), and 0 otherwise.
    >> 
    >> and that non-zero is not defined!

    PD> Uh-oh...

    >> and isnan is used in Arith.c for ISNAN and that is used in coerce.c
    >> for do_isna (why?)

    PD> is.na(0/0) # Splus compat.

    >> So ISNAN needs to be redefined on Linux at least.  Using ISNA not
    >> ISNAN solves this one for me.

    PD> But...(see above)

    PD> A better fix is

    PD> #define ISNAN(x) (isnan(x)!=0)

oh dear;  and we had thought we would clean up the ISNAN definition and
do_isna[n] functions....
Are we all sure that this is *IT* for now...

Could we improve on the tests in  configure.in	to make sure beforehand
that  ISNAN will work?


    PD> Next problem: Messing with Arith.h and make'ing doesn't do anything
    PD> (dependency missing, obviously)

I've added ``a few of these'' to src/appl/Makefile.in and src/nmath/Makefile.in

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