repeat.sh is a bash script that will run arbitrary commands in a loop.

It is a replacement (or supplement?) for tools like xargs and parallel,
given a specific set of circumstances:

  - You want to run the same command, over and over and over (for example:
    load testing)

  - You don't need to limit the number of parallel processes being run.

  - You MAY want:

    - A log file with the output of each command and its return status.
    - A lock file to determine if a command is still running.
    - To wait a few seconds before re-running a command.
    - To keep the commands running in the background indefinitely.
    - To feed a list of arguments to a command or read commands from a file.

xargs and parallel don't necessarily make all that easy. Those programs are
designed to run a command once, not loop over it. They also aren't designed
to run them in the background (though that's easy enough to script around).

 - You can call xargs or parallel from repeat.sh, enabling you to re-run
   commands once xargs or parallel are done. The downside to this is that
   repeat.sh will have to wait until the very last command is done running
   from xargs/parallel before everything is restarted again.

 - You can call repeat.sh from xargs or parallel, which enables you to do things
   like keep a certain number of commands looping in parallel. But since
   repeat.sh processes keep repeating forever, that means xargs/parallel will
   just reach whatever maximum parallel process limit you configured and never
   run new commands after that.

Other programs you could use instead of the above:

 - 'watch' will re-run a command over and over. You could probably get the 
   behavior of repeat.sh by scripting 'watch' with bash and xargs, but the
   complexity would get non-trivial once you tried to add all the other
   features of this script.