Tips on Writing Shell Scripts

1. Use "stepwise refinement", which means to work on your program in steps. Make sure each part is well tested before moving on to the next step.
2. Us command-line editing.
   • Since most of you use vi as your editor, I recommend using it to access your history and edit your commands. (You may use emacs or pico if you prefer.)
   • After logging in, type the following:

     set -o vi
     

     Note: this is the only way to set up in some versions of bash. You must type this in at the command-line each time you log in. Having it in a shell script (for instance your profile) does not work in all versions of Linux. We will try it in yours and see what happens.
   • Using command-line editing will let you bring up your previous commands with <ESC>- or <ESC>k Keep pressing k (or -) to move back through your commands.
   • You can also search for a previous command with <ESC>/regex
   • Once you have a command up, you can edit it using vi commands.
   • You can use any commands that work on the current line. For instance use w to move to the next word, fp to move to the next p in the line, xp to reverse two characters, dw to delete a word, ...
   • Once you are ready to execute the new command, simple press Enter.
3. You can also execute a command while you are in vi with !.
   • In vi, the % is used to mean "the current file".
   • So, if you forgot to set the permissions on the shell script you are editing, you can set it without getting out of vi.
     Type:

     :!chmod 700 %
     

     (then press Enter) to give it rwx for the owner (you).
   • To run the shell script from inside of vi, type:

     :!%
     

   • If you want to pass in any arguments, add them after the %. For instance, this:

     :!% arg1 "argument 2"
     

     runs the current file with two arguments.
   • To re-run the previous command, type:

     :!!