R help - Jul 1997 - R-beta: dyn.load: incompatability between R and S

I have  C routines in a collection of files

"/homef/jonm/postdocs/GLMM_project/Cprogs/nn/misc.o"
"/homef/jonm/postdocs/GLMM_project/Cprogs/nn/random.o"
"/homef/jonm/postdocs/GLMM_project/Cprogs/nn/ars.o"

... etc , where the third file includes functions which call
functions in the second.

In S, the function
> gload
function()
{
	dyn.load("/homef/jonm/postdocs/GLMM_project/Cprogs/nn/misc.o")
	dyn.load("/homef/jonm/postdocs/GLMM_project/Cprogs/nn/random.o")
	dyn.load("/homef/jonm/postdocs/GLMM_project/Cprogs/nn/ars.o")
	

....

works fine.

However, in R although the first two statements execute OK, the third
produces 
> dyn.load("/homef/jonm/postdocs/GLMM_project/Cprogs/nn/ars.o")
Error in dyn.load(x) : ld.so.1: /scratch0/jonm/R-0.50-a1/bin/R.binary: fatal: re
location error: symbol not found: unitexp: referenced in /homef/jonm/postdocs/GL
 _project/Cprogs/nn/ars.o>

"unitexp" is a program in 


"/homef/jonm/postdocs/GLMM_project/Cprogs/nn/random.o"

so presumably the problem is that S "knows" about functions already
loaded
by earlier dyn.load statements but R doesn't.

Is there any answer other than putting everything into one
huge file?

Thanks

Jonathan Myles



-- 
Dr. Jonathan Myles         e-mail:jonathan.myles at mrc-bsu.cam.ac.uk
MRC Biostatistics Unit     Tel. 01223 330371
Institute of Public Health FAX  01223 330388
University Forvie Site
Robinson Way
CAMBRIDGE
CB2 2SR                   
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-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
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Jonathan Myles writes:

 > However, in R although the first two statements execute OK, the third
 > produces 
 > 
 > >
dyn.load("/homef/jonm/postdocs/GLMM_project/Cprogs/nn/ars.o")
 > Error in dyn.load(x) : ld.so.1: /scratch0/jonm/R-0.50-a1/bin/R.binary:
fatal: re
 > location error: symbol not found: unitexp: referenced in
/homef/jonm/postdocs/GL

The problem is that our "dyn.load" expects to get a shared library
rather than just a compiled binary file.  You could take a look at
the code in demos/dynload to see how this works.  Under SunOS you
would do
	gcc -g -fpic misc.c random.c ars.c
	ld -assert pure-text misc.o random.o ars.o -o mylib.so
and then in R
	dyn.load("mylib.so")
The Makefile in demos/dynload may contain flags that will work
(no promises though).

We do it this way because linking with shared libraries is a much more
portable solution.  I suppose we should change the name to
dyn.load.lib or something.
	Ross
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-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
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=