Makefiles

Paul Gilbert pgilbert@bank-banque-canada.ca
Tue, 10 Feb 1998 13:01:22 -0500


While on the subject of makefiles, I have discovered a few problems:

If a package directory has a src subdirectory with fortran code and a makefile,
and a fortran compiler is available, I think R INSTALL tries to use this rather
than the src-c directory. Is this the generally preferred stategy? Currently g77
seems to have a bug and my fortran code does not compile.  If I were to just
include the fortran code for reference, but not expect anyone to compile it, is
the best way to put it in src-c, or should I put it in src and not put a
Makefile there?

I previously reported that my compiled C code does not find pow_ii and was led
to believe this is a routine in the f2c lib which I need to include in some way.
However, prompted by the above problem, I experimented with compliing using the
Solaris SunPro fortran compiler. Now I don't find __pow_ii and __pow_ri unless I
"hard code" the location of libF77 and libM77 into the .so file using -L. R
INSTALL doesn't do this. I'm not sure how this should be fixed. Also, I would
appreciate it if someone could explain whether it should be my .so file or
R.binary that resolves these references.

Paul Gilbert

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._