Calling programs from different directories

Shell Programming and Scripting
1

I have some csh and awk scripts on path

/nethome/chrisd/HSeis/TommyCD/TommyCD-1101/Scripts

Usually I have a working directory from where I run the scripts by having a soft link to the Scripts directory
in the directory above the working one

For example from

/nethome/chrisd/HSeis/TimeDistance/TimeDistance-1101/B-rgdt0p25/A-Synt/A-jcdint

I can run a csh program using

../Scripts/prog.csh

Now, this is where I get confused. prog.csh calls two awk programs like this

  if (($nAnomaly == 1) && ) then
      awk -v anomaly=$Anom -v zloc="$zmin/$zmax" -v dz=$dz  \
          -v cols=$cols -v maxdepth=$maxdepth               \
          -f $spath/create-anomaly-zc.awk $fref.zc > $fmod.zc
  else if ($nAnomaly == 2) then
      awk -v anomaly="$nAnom/$pAnom" -v zloc="$zmin/$ztran/$zmax"  \
          -v dz=$dz -v cols=$cols -v maxdepth=$maxdepth            \
          -f $spath/create-anomaly-zc.awk $fref.zc > $fmod.zc
  endif

  awk -v xi="$xix/$xiz" -v xf="$xfx/$xfz"  \
      -f $spath/convert-zc-cmod.awk $fmod.zc > $fmod.cmod

As you can see I have included the variable spath which I set to the full path to the awk scipt

set spath = /nethome/chrisd/HSeis/TommyCD/TommyCD-1101/Scripts

I'm not sure what would be the best way to call the two awk scripts, they are found in the same directory as the .csh file

2

It is not beautiful, but it works. Better is the enemy of good enough. Do you have a specific additional need?

You could use $PATH if the awk files are good executables with first line #!/bin/awk -f and the executable chmod/ls -l permissions.

http://www.unix.com/shell-programming-scripting/146732-calling-perl-script-shell-program.html\#post302464588

The .cshrc is usually in $HOME so login sees it, and scripts and data files should not be!

http://www.unix.com/shell-programming-scripting/146652-ftp-mget-exclusion-question.html\#post302464505

3

Apologies, but I am not sure what you mean exactly.

Can you give some example?

4

Any script whose interpreter can find the script with at most one extra argument, ike ksh or sed -f or awk -f, can be made into a barefoot, freestanding executable with one line and file mode bits.

Any directory in your $PATH is searched for executables by entry name.

Make your awk scripts executable, put them in a application-dedicated bin directory, put that directory in your $PATH, and just call them without the awk, -f, path.

Suppose I was doing something similar:

I type: mkdir ~/bin
and in there I put a file davidx, and in that file I have:

#!/usr/bin/awk -f

 -- awk commands --

I type: chmod u+x ~/bin/davidx
to add execut for me.

I type: PATH=$PATH:~/bin
and add that line to my: ~/.profile

Now, I can just say this in my shell script david_sh: davidx option1 option2 ...

I should put my shell script david_sh in ~/bin, too. Then, I can call it by just typing entry name david_sh when I test. If it is a ksh script, it should look like:

#!/user/bin/ksh

 . ~/.profile

 -- ksh script commands --

 davidx option1 option2 ...

 -- ksh script commands --

Your interpreter paths (/usr/bin for ksh, awk) may vary.

5

There's one problem. I actually go to a different machine, which can see the Scripts and working directories and run everything from there. The thing is that I cannot change stuff on them. Can't use my machine as it's too slow.