Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 2 Jun 1999 23:50:40 -0400
From:      "Sean O'Connell" <sean@stat.Duke.EDU>
To:        rdmurphy@vt.edu
Cc:        FreeBSD emulation <freebsd-emulation@FreeBSD.ORG>
Subject:   Re: Matlab 5.3
Message-ID:  <19990602235040.B5623@stat.Duke.EDU>
In-Reply-To: <14165.38910.206452.281058@knock.econ.vt.edu>; from Russell D. Murphy Jr. on Wed, Jun 02, 1999 at 05:14:52PM -0400
References:  <14165.38910.206452.281058@knock.econ.vt.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On 1999 Jun 02, Russell D. Murphy Jr. (aka rdmurphy@neale.econ.vt.edu) wrote:
> 
> I'm trying (again) without success to install Matlab.  I'm using:

There are two main ways of doing this.

1) Use a supported platform (a linux box or actually a digital unix or
   sun or whatever will also unpack the linux distribution too) to unpack 
   it onto an nfs mount or a local directory, tar it up and change the 
   MATLABDIR stuff, and modify a few other key scripts that use the
   /bin/uname command and look for linux (Linux?) and lie to them :).

2) A very, very ugly and tedious process.  The installation scripts
   for Matlab are pure evil.  I stupidly did not do version one :)

   It bascially boils down to unpacking the cdrom onto a hardrive,
   making a LOT of changes including untarring and gzipping several
   files making the relevant modifications listed above and repacking
   everything and then installing.

   I haven't done this in a while, but I will be soon.  I have some
   notes that I will include from Matlab5.2; be prepared to pull chunks 
   of hair out.  You will probably need to do steps involving the
   pd/tar.cmp for each of the toolboxes too.  This was for a base install. 

   Good Luck

   S

Here are my install notes, more or less :)   These were aided by
discussion with Tom Bartol from the Salk Institute (thanks Tom).

1) cd /bin
2) ln pax tar

*this is critical;  /bin/tar is hard coded into the tsetup and xsetup
linux executables...grrr, this caused the majority of my troubles*

3) copy all of the cdrom to a disk somewhere, $CDROM  (in my case it
   was on another machine and then I did an nfs mount ... see below).

4) apply the arch.sh patch that you recommended
   > $CDROM/install
   > $CDROM/update/install/arch.sh
   > $CDROM/update/install/linstall
   > $CDROM/update/install/install/arch.sh
   > $CDROM/update/install/etc/util/arch.sh

*** arch.sh     Tue Jul  8 09:58:10 1997
--- arch.sh.freebsd     Tue Jul  8 09:58:42 1997
***************
*** 70,75 ****
--- 70,81 ----
                :
                ;;
          esac
+    elif [ -f /usr/bin/uname ]; then                     # FreeBSD?
+         case "`/usr/bin/uname`" in
+             FreeBSD)
+                 Arch="lnx86"
+                 ;; 
+         esac
     elif [ -f /bin/arch ]; then                                # early sun4
        Arch="`/bin/arch`"
     fi

5) Since the good folks at the Mathworks worked feverishly to make
   it difficult, you also have to:  upack 3 different tar.cmp (tar.tar.Z)
   and add the arch.sh patch

   > $CDROM/upate/pd/lm/tar.cmp
     etc/util/arch.sh  (note: this is what the install leaves in $MATLAB/etc/util)

   > $CDROM/upate/pd/matlab/tar.cmp
     bin/util/arch.sh  (note: this is what the install leaves in $MATLAB/bin/util)

   > $CDROM/upate/pd/install/tar.cmp
     install/arch.sh
     install_matlab    (final version of the master install script)

6) while you have these tar files unpacked, there is one more little hack that
   has to get put in in the osversion check business:  (sorry I don't have
   a diff convenient)

   add:

   elif [ -f /usr/bin/uname ]; then
      case $Arch in
         lnx86)
            ver='2.0.18'
            ;;
      esac

   This needs to end up in:

   $CDROM/install
   $CDROM/update/install/linstall
   $CDROM/update/pd/install/tar.cmp (in install_matlab)
   $CDROM/update/pd/matlab/tar.cmp  (in bin/util/oscheck.sh)

7) I then nfs mounted this on /cdrom on my freebsd and did
   /cdrom/install &

-- 
-------------------------------------------------------------------------
Sean O'Connell                                  Email: sean@stat.Duke.EDU
Institute of Statistics and Decision Sciences   Phone: (919) 684-5419
Duke University                                 Fax:   (919) 684-8594


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-emulation" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990602235040.B5623>