Using shell commands in lisp

Slosh can easily be used to launch processes in lisp mode (e.g. in a script). Interacting with the input and output of those programs can be done in a variety of ways.

Two main flavors of starting shell processes

1. The $sh variety which is a little clunky as you directly pass it a string, as seen in the let binding.
2. The $ reader macro which treats everything as a string but allows escaping slosh variables directly in the shell with macro-like syntax ~

(defn parse-git-branch () (let (branch ($sh "git rev-parse --abbrev-ref HEAD 2>/dev/null"))
    (if (= branch "")
            (str "")
                    (do $(export BRANCH_NAME=~branch)
                                (str "(" branch ")")))))

~ under construction ~

instead of (sh "wc" "-l") we have $(wc -l) (def v "foo") $(echo ~foo) $(echo ~v)

calling shell commands in lisp code

3 ways

• $() - reader-macro - convenient because it interprets everything as a string - can be used with macro like ~

• (sh "") - regular function, but you have to pass it a string ** $() & (sh) are very similar.

• ($sh ) - like backticks, call to shell, and get back a slosh string, e.g. run this command and give me the output but we do not have a reader macro yet.

  • this is not implemented w/ a reader macro but could be something like $(( )) except that might be ugly
  • ($sh "git rev-parse --abbrev-ref HEAD 2>/dev/null")
  • this returns a string, does more work for you
  • (sh "git rev-parse --abbrev-ref HEAD 2>/dev/null")
  • Returns the exit status if foreground and the PID if background. Add the '&' to the string to background it.

NOTE:
on the CLI typing $() directly on the REPL, but if you wrapped it in say a (do $()) it would work, REPL isn't smart enough yet to
pass the $() to the LISP repl 

bash precedent

• diff <(echo "1\n2\n3") <(echo "echo 1\n\1\n3")
• https://www.gnu.org/software/bash/manual/html_node/Process-Substitution.html#Process-Substitution