R-beta: linux: problem installing R-base-0.50.a4-1.i386.rpm
ihaka at stat.auckland.ac.nz
Tue Oct 7 23:14:52 CEST 1997
Martyn Plummer writes:
> When you build R on Linux, the configure script automatically adds f2c to
> the list of libraries. This means you must have the static or shared
> f2c libraries installed to build R. But I'm wondering why it needs to be
> linked to f2c at all. R is not supposed to require a Fortran compiler and,
> indeed, it builds quite happily if you edit the configure script and remove
> the line that adds -lf2c to the list of libraries. Moreover the
> statically linked version of R.binary is, in fact, smaller that the
> linked version, which suggests to me that it hasn't taken anything from
> libf2c.a, but perhaps I'm being naive here.
> Is this a hangover from the time when a Fortran compiler was needed to
> build R?
Fortran has always been a problem. We have tried to keep R so that
only a C compiler is required to build and install it. This was
partly to "be portable" and partly because getting C and Fortran to
talk to each other in a platform independent way has been too hard for
mortals like us.
This is soon going to change. We have permission from the author of
Octave to use his configuration tools to set up R so that Fortran will
be detected and used. If Fortran is not available (or usable) on a
given platform, f2c will be used. I.e. you will have to have a
working Fortran or f2c in place in order to use R.
As far as I can tell, every platform which R currently runs on will be
suported by the Octave scripts.
[ There is a new version of R nearly ready to roll, but it will probably
not have the new configuration scheme in place. All future versions
will have much better Fortran integration. ]
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