Message-ID: <49796.209.195.160.60.1123902023.squirrel@webmail.pitt.edu>
Date: 2005-08-13T03:00:23Z
From: Rick Bilonick
Subject: g77 and R
In-Reply-To: <42FD43FA.1080105@verizon.net>
> rab45+ at pitt.edu wrote:
>> Jerry,
>>
>> Here are the two R packages.
>>
>> Thanks.
>>
>> Rick
> All the fortran codes compile OK so the problem is invoking the compiler
> you
> want. One approach which is a work around is to rename the gfortran
> executable
> to somthing else like gfortran-save and create a link called gfortran that
> points to g77. That might work, but its a hack.
>
> From looking at these it appears that they are intended to be built from
> the R
> environment. Unfortunately I am not an R user and I am not familiar with
> it.
>
> What is the usual build process you go through?
>
> Jerry
>
Installing R packages (sets of statistical analysis functions for specific
types of problems) is usually very simple. If the R package is in a
repository, then something like:
> install.packages(c("foreign","chron"))
will automatically find, download, compile c and fortran code (if
necessary), and put the package functions from the foreign and chron
packages into the library. It's usually painless and simple - all the
details automatically handled. Since gfortran was installed, a number of
packages have failed to compile properly and weren't installed. I have
never had a problem installing R packages until very recently - I suspect
when gfortran was foisted on us unawares.
Before FC changed to gfortran, g77 was always used. All I want to do is to
make sure that if fortran code needs to be compiled, g77 will be used. I
don't think R knows anything about gfortran or g77 - it just uses what is
available on a given system - so this is something I need to change on FC4
- not R.
Rick