[R] Fwd: probably "round()" bug --- in Guido's Win' R 0.62.3

Martin Maechler Martin Maechler <maechler@stat.math.ethz.ch>
Thu, 15 Oct 1998 15:59:03 +0200


Andrea,

this problem does NOT appear on my (Sun Solaris) version of R 0.62.3.
It must be a "architecture dependent" (maybe Windows only) bug, triggered
by bogous  libraries / optimizations, or such things.

Here is a mail from the R core development list, of half a year ago.
I don't really know the details of what it's about...

>> Date: Sun, 15 Feb 1998 09:14:50 +0100
>> From: Kurt Hornik <hornik@ci.tuwien.ac.at>
>> To: R-core@stat.math.ethz.ch
>> Subject: round() problem
>> 
>> The following problem occurs under ix86 Debian GNU/Linux when compiling
>> with optimization turned on:
>> 
>> R> round(0.5)
>> [1] 0.5
>> 
>> etc etc
>> 
>> The problem is eliminated if in src/nmath/fround.c we use the math lib's
>> rint() rather than private_rint(), as in the previous versions ...
>> 
>> Is it o.k. to go back to using rint() whereever available?  (Do we still
>> need to override rint() on hpux?)


Note: This is *NOT* CC'ed to R-help (but to R-devel, the "R code experts")
	with my understanding that finally a SUMMARY 
	or *the* solution will appear on R-help.

Best Regards:

Martin Maechler <maechler@stat.math.ethz.ch>	http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum SOL G1;	Sonneggstr.33
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1086			<><
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._