Re: [R] Segfault and bad output with fOptions::rnorm.sobol

2011-10-23 Thread Robert McDonald
On Sun, Oct 23, 2011 at 11:48 AM, Robert McDonald wrote: > I think your question is answered by >> >> http://cran.r-project.org/web/packages/fOptions/ChangeLog >> >> 2010-04-23 chalabi >> >>* ChangeLog, DESCRIPTION: updated DESCR and ChangeLog >>* src/085A-LowDiscrepancy.f: fixed

Re: [R] Segfault and bad output with fOptions::rnorm.sobol

2011-10-23 Thread Robert McDonald
> > I think your question is answered by > > http://cran.r-project.org/web/packages/fOptions/ChangeLog > > 2010-04-23 chalabi > >* ChangeLog, DESCRIPTION: updated DESCR and ChangeLog >* src/085A-LowDiscrepancy.f: fixed sobol RVS on 64 bit platform >* ChangeLog, DESCRIPTION

Re: [R] Segfault and bad output with fOptions::rnorm.sobol

2011-10-22 Thread Ben Bolker
Robert McDonald gmail.com> writes: > I have had the following problem with R 2.10, 2.13.1, and 2.13.2, running on > Ubuntu linux 10.04, xubuntu 11.10, and a version of Redhat (I think 5). > rnorm.sobol is producing impossible random values, and occasionally the > routine crashes. Here are samples

Re: [R] Segfault and bad output with fOptions::rnorm.sobol

2011-10-22 Thread Robert McDonald
On Sat, Oct 22, 2011 at 2:45 AM, Patrick Burns wrote: > This should really go to Rmetrics > since it is a bug in their code. > > You could find out more specifically > where the problem is by running it > under valgrind. > > First you have to have valgrind, then > start R with it: > > R -d valgrin

[R] Segfault and bad output with fOptions::rnorm.sobol

2011-10-21 Thread Robert McDonald
I have had the following problem with R 2.10, 2.13.1, and 2.13.2, running on Ubuntu linux 10.04, xubuntu 11.10, and a version of Redhat (I think 5). rnorm.sobol is producing impossible random values, and occasionally the routine crashes. Here are samples of the output and the crash message. librar