I've been having problems with gcc-3.1 and R-devel on one of our Solaris 7 systems (not the one I usually use). We have seen a report of this before (Chuck Berry, 17 July 2002). The symptom is that make check fails with var(1) being NaN not NA. The underlying cause is that as.double(NA) is NaN. It can be solved by compiling arithmetic.c without -O2. Another installation with a single-processor UltraSparc 1 works correctly. My guess is that gcc-3.1 and multiprocessor/Sparc IIIs are not mixing. -- Brian D. Ripley, ripley@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 272860 (secr) Oxford OX1 3TG, UK Fax: +44 1865 272595 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
On Sat, 3 Aug 2002 ripley@stats.ox.ac.uk wrote:> I've been having problems with gcc-3.1 and R-devel on one of our Solaris 7 > systems (not the one I usually use). We have seen a report of > this before (Chuck Berry, 17 July 2002). > > The symptom is that make check fails with var(1) being NaN not NA. The > underlying cause is that as.double(NA) is NaN. > > It can be solved by compiling arithmetic.c without -O2. > > Another installation with a single-processor UltraSparc 1 works correctly. > My guess is that gcc-3.1 and multiprocessor/Sparc IIIs are not mixing.FWIW, the specific problem seems to be with static const int hw = 0; static const int lw = 1; and their use in array style indexing like this: y.word[lw] == 1954 y.word[lw] != 1954 R passes 'make check' on my Solaris 8 Blade 100 with -O2 when *either* 'const' is dropped or ' *(y.word + lw) == 1954 ' style indexing is used. Chuck Charles C. Berry (858) 534-2098 Dept of Family/Preventive Medicine E mailto:cberry@tajo.ucsd.edu UC San Diego http://hacuna.ucsd.edu/members/ccb.html La Jolla, San Diego 92093-0645 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- 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 _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._