Compiling MAPGEN under Linux.

[Gerald I. Evenden]

>From: STEWARTB at CPVA.SAIC.COM
>Date: Tue, 31 May 1994 11:16:22 -0700 (PDT)
>To: grassu-list at max.cecer.army.mil
>Subject: Compiling MAPGEN under Linux.
>
>Has anyone successfully compiled MAPGEN under Linux yet?  I have had some
>limited success but I'm not all the way there yet.  Some background:
> 
>I started with Changyong Cao's XIS collection of Linux, X-Windows, and
>GRASS 4.1.  The installation went fine and the core routines of GRASS work
>as expected (remarkably fast too even compared to our Sparcstation 2!).
>We use MAPGEN a lot in our line of work so compiling that was my next
>mission.  I started with version 4.0 which is what we use on the Sun with
>GRASS 4.1 (since MAPGEN 4.1 had problems).  I was able to compile most of 
>the source code but ran into problems while running "make.mapdef".
> 
>Since we want the latest version of MAPGEN my next step was to try and 
>compile that.  I pulled version 4.11 off of the moon ftp site and tried 
>compiling both on our Sun and on our Linux box.  The Sun compile went well
>until I got to PROJ.4.  Need for an ANSI C compiler stopped me short there
>and I should get my GNU compiler this week to finish that install.  I then
>tried to compile under Linux and got through most of the code until I got 
>to PROJ.4.

PROJ.4.3.1 on charon.er.usgs.gov (128.128.40.24) uses "configure"
followed by "sh intall."  That may be a little more forgiving for
Linux.  Obviously, with gcc on Linux, ANSI C problems go away.
You may want to use "--prefix=" with configure---check README
for reasons.

>PROJ.4 does not have a Gmakefile but there is an "install" routine that 
>apparently works for several machine architectures (Sun, DG, etc.) but 
>not for Linux.  Thus, even though I do have an ANSI C compiler, I do not
>have the proper "install" code to complete the installation.  

I am sure that Gmakefile is a fine system, but I try to stay away
from things like this.  I have spent too many hours cursing the X11
make system to want to get involved with another.  IMHO, the
"configure" system, from FSF (GNU), is a good system.  Not perfect,
but it handles most of the problems I have seen and does so
with normally available software.  I intend to stick with it.

>Has anyone figured this out yet?   
> 
>(Some postscripts---Apparently the topquark ftp site does not have Linux
>binaries for MAPGEN.  I have yet to check the SCS ftp site since I do not
>have their address but I am assuming that they are not dabbling in Linux.)
> 
>Brad Stewart
>stewartb at cpva.saic.com

Related to the Linux problems, if I can figure out how to use a Sony
CDU-6211 drive with Linux, the primary development system for PROJ.4
and MAPGEN (my base distribution) will be Linux.  But I have run into
a snag here.  I'm going to be pissed if I have to by another drive!

Another point, if I can get Linux up and going on Pluto I may be
willing to make a Linux binary distribution available on charon
(mirrored on moon).  But this will require using environment
variables rather than compiled-in settings to point to required
library directory.  Any comments on this are welcome.

MAPGEN poses additional problems, because I am not sure what
value-added work has been made to moon distribution (principly by SCS
folks).  Thus I do not know if my version is useful to GRASS users in
general.  Again, comments welcome.

Gerald (Jerry) I. Evenden   Internet: gie at charon.er.usgs.gov
voice: (508)563-6766          Postal: P.O. Box 1027
  fax: (508)457-2310                  N.Falmouth, MA 02556-1027