[R] question on R's makefile and configure system

Christopher Marshall christopherlmarshall at yahoo.com
Tue Mar 23 17:07:19 CET 2004


I should have cc'd the list in my first response and forgot.

This exchange is about which path variables to set in the R make process so you can make a
compiled package for a linux distribution.

Dirk gave me the missing piece and I am able to compile proper slackware packages now.

--- Dirk Eddelbuettel <edd at debian.org> wrote:
> On Tue, Mar 23, 2004 at 07:14:27AM -0800, Christopher Marshall wrote:
> > So you mean I should do this:
> > 
> > ./configure --prefix=/usr
> > make
> > mkdir /tmp/r-package
> 
> Not needed, install will create the directory
> 
> > make install prefix=/tmp/r-package/usr
> 
> Yes.
> 
> > I have not used debian and don't know my way around its package management system.
> 
> This has nothing to do with Debian.  Configure, make, install, ... are all
> GNU tools that behave identically irrespective of the system.
> 
> I suggest you try to familiarize yourself with the building process by
> studying a simpler, smaller package for slackware, and then carry over what
> worked for it.  
> 

That's exactly what I did.  I took the gawk.SlackBuild script from
slackware-9.1/sources/a/gawk/gawk/SlackBuild.  I also looked at a lot of other slackware source
packages and the pattern in the build scripts was usually this

./configure --prefix=/usr
make
make install DESTDIR=/tmp/some-package

After you gave me the hint that the variable I needed at the "make install" step was prefix, I got
R to build this way

./configure --prefix=/usr
make
make install prefix=/tmp/some-package/usr

So my problem is solved, and I can now create proper slackware packages.

I realize that the autoconf system and the rest of the build tools are used by all distributions,
but there is a lot of variation in how you set up the variables used in your scripts and it seems
that R is an oddball in this respect.

Thanks for responding so quickly and solving my problem.

Chris Marshall




More information about the R-help mailing list