[R] dyn.load'ing dll's created by Scompile causes R to crash

Thomas Lumley tlumley at u.washington.edu
Thu Nov 8 17:38:04 CET 2001

On Thu, 8 Nov 2001, [iso-8859-1] Søren Højsgaard wrote:

> Dear all,
> I have used the Scompile package (by Mathhew Calder) on previous releases of
> R (on Windows), and it has worked well. However, R.1.3.1 crashes when
> dyn.load'ing C-functions created by Scompile.
> From the last R-newsletter it seems that there has been some changes to how
> external routines are being included in version 1.3.1. I won't claim to
> understand the details, but I was wondering whether these changes are the
> source of my problem? If so, I would like to know if there is a way of
> "getting back" to the old way of doing things such that Scompile can still
> be used?

The old way is still the default. Duncan Temple Lang's article describes
an optional new way.

Looking at the CHANGES file for R 1.3.1 under Windows, the recommended
compiler system has been changed. Possibly the new system isn't
quite compatible with lcc, though DLLs should be fairly standard.
Unfortunately I think debugging under Windows is tricky.

A long-term fix would be to patch Scompile so that it uses Rcmd SHLIB to
compile things.  This would mean that the recommended compilers for R
could always be used in compiling. The patch to Scompile would be fairly
easy, but it would also mean downloading Perl, gcc and the other things
needed to compile R. I did something like this a few years ago to get
Scompile to work with R on Solaris (probably about R 0.50), but I
didn't find it useful enough to keep updated.


Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle

r-help 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-help-request at stat.math.ethz.ch

More information about the R-help mailing list