I've been successfully using a dll via dyn.load() with R1.6.0 for Windows, but when I try it under R1.6.1 it manages to crash the program completely. Has there been a change in how R1.6.1 handles dynamic loading? I couldn't spot any such changes in the documentation. This problem occurred on two different machines, and both run the code under R1.6.0 without a problem. Rob Hyndman ___________________________________________________ Rob J Hyndman Associate Professor & Director of Consulting Department of Econometrics & Business Statistics Monash University, VIC 3800, Australia. http://www-personal.buseco.monash.edu.au/~hyndman/
No deliberate change (it _was_ a patch release). However, this is a not uncommon occurence with code that is incorrect and corrupts memory or the stack: sometimes you get away with it on some platforms. We found that with a version of the tseries package: it worked on 1.6.0 but not 1.6.1 on Windows. (The problem was a .C call with one too few arguments.) On Thu, 12 Dec 2002, Rob Hyndman wrote:> I've been successfully using a dll via dyn.load() with R1.6.0 for > Windows, but when I try it under R1.6.1 it manages to crash the program > completely. Has there been a change in how R1.6.1 handles dynamic > loading? I couldn't spot any such changes in the documentation. This > problem occurred on two different machines, and both run the code under > R1.6.0 without a problem.-- Brian D. Ripley, ripley at 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
Rob Hyndman wrote:> > I've been successfully using a dll via dyn.load() with R1.6.0 for > Windows, but when I try it under R1.6.1 it manages to crash the program > completely. Has there been a change in how R1.6.1 handles dynamic > loading? I couldn't spot any such changes in the documentation. This > problem occurred on two different machines, and both run the code under > R1.6.0 without a problem.I don't see any main differences except of bug fixes. So I guess it's a bug in the dll, which has not bit in R-1.6.0, however. Recompiling for the recent version of R is never a bad idea anyway. Uwe Ligges