[Rd] (PR#1008) SIGSEGV began at 1.2.0

Roger.Bivand@nhh.no Roger.Bivand@nhh.no
Tue, 3 Jul 2001 10:42:30 +0200 (MET DST)

This behaviour is confirmed for RH70, RH62, Debian 2.2 on 3 different
Pentium-class i386. I've checked back and package ann_0.2-2 works on
1.1.1, but fails with seg. faults on 1.2.0 (ann_0.2-2.pre3.tar.gz to
accommodate the change in Makeconf) and subsequent (1.2.0 checked today on
RH62 with gcc 2.95.3 19991030 (prerelease)).

It is most likely that 1.1.1 was too forgiving, and that better memory
management from 1.2.0 is uncovering bad code in the *.so, but could there
be an interaction between dynamic memory management in R >= 1.2 and C++
(g++)? (namespace???). It strikes me as odd that the SIGSEGV sometimes
comes from the C++ side (annDeallocPts - __builtin_vec_delete), and
sometimes from the R side (ReleaseLargeFreeVectors), with no pattern that
I can see.


Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93
e-mail: Roger.Bivand@nhh.no
and: Department of Geography and Regional Development, University of
Gdansk, al. Mar. J. Pilsudskiego 46, PL-81 378 Gdynia, Poland.

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