[R] [Mac OS X 10.4] object R_loess_raw not found

[René J.V. Bertin]

Hello,

Since I (finally) upgraded to 10.4 "Tiger", I've been experiencing
some strange behaviour with dynamic libraries.

When I now launch R, and try to use the loess() function from the
stats package, I get the following:

> R

R version 2.4.0 (2006-10-03)
Copyright (C) 2006 The R Foundation for Statistical Computing
ISBN 3-900051-07-0

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

  Natural language support but running in an English locale

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

[Previously saved workspace restored]

> example(loess)

loess> cars.lo <- loess(dist ~ speed, cars)
Error: object "R_loess_raw" not found


I saw this first with a self-built 2.3.1 version, that worked fine
under 10.3 . Hoping that a Tiger build would solve matters, I
installed the fresh 2.4.0 from the downloaded disk image. As you can
see, the error message persists.

Curiously, the ppc/stats.so file somewhere deep in the 2.4 framework's
bowels does contain a symbol loess_raw, but not R_loess_raw. I wrote a
quick C app that dlopens a .so file and then tries to obtain a given
symbol from it using dlsym. I can't seem to get any of the defined
symbol from stats.so . I'd say my little loader is broken somewhere,
BUT ...

This does remind me of a problem I had with a Python module
(pygame/constants), where the expected init function (initconstants)
was available but invisible to dlsym(). In that case, changing the
module (and init function's) name solved the issue, I still don't
understand why.

This is really annoying, so if anyone has an idea about what might be
going on, I'd appreciate the feedback!

BTW, I did try building my own copy, but for some reason, the
configure/build process ignores the optimised BLAS in the vecLib
framework ...


René Bertin
(PS: please CC me via email if you reply)