[comp.unix.questions] running a program test in a shell script

russell@ccu1.aukuni.ac.nz (Russell J Fulton;ccc032u) (03/14/91)

We have two Unix machines, a Sun4 and an SGI 4D. Users log on to the SGI and
do most of their work there. There are however several bits of software that
are not available for the SGI but do have SUN4 versions. So we have set up
scripts for these packages that do an rlogin to the sun to invoke the 
package concerned. (The users files are all share via NFS.) In general this
works well except for occasional problems caused by delays in updating
the directories of NFS mounted disks.

We recently started running Simscript in this manner. Simscript is a compiler
so it was necessary to provide scripts both for compiling and running the
programs. All was well until somebody named their executable 'test'.
This results in a script being generated and executed on the sun that includes
the line 

test <parameters>

The symptoms were that the program exited without executing the first 
statement of the program test. The reason (I assume) is that the test command
in the shell is being invoked and not the program test as found by following
PATH. 

I have had a look through the man pages for csh but can't find anything that
sheds light on what to do about this problem. I need a method of protecting
the command invoked in the script from being interpreted as shell commands.

Cheers Russell.
-- 
Russell Fulton, Computer Center, University of Auckland, New Zealand.
<rj_fulton@aukuni.ac.nz>