Lesson 03 · Run an external command

Run an external command

This is the lesson your shell stops being a parser and starts being a shell - it launches a real program. You spawn a child process, run the named executable in it, and wait for it to finish.

The goal

Given a word list, run the named program in a child process and wait for it to complete.

Start here - the target

TO DO

Scenario: Running an external program

Giventhe word list ["/bin/echo", "hi"]

Whenthe shell runs it

Thena child process runs /bin/echo and "hi" appears on standard output

Andthe shell waits for the child to finish before prompting again

Background

To run a program, a Unix shell does something that looks strange the first time: it clones itself. The fork call makes a copy of the shell process; the copy (the child) then execs - it throws away the shell’s code and replaces itself with the program you asked for, keeping the same open files. The original (the parent) stays behind and waits for the child to finish. This fork-then-exec dance is the single most important mechanic in the whole project.

Notice that exec only returns if it failed - on success there is no shell code left in the child to return to. So the line right after exec is your error-and-give-up path. For today, name programs by their full path (/bin/echo); teaching the shell to find echo on its own comes in a few lessons.

Make it work

pid_t pid = fork();
if (pid == 0) {              // child
    execv(words[0], words);  // replace this process with the program
    perror(words[0]);        // only reached if exec fails
    _exit(127);
}
waitpid(pid, &status, 0);    // parent waits

CheckpointDONE

Your shell runs real programs - `/bin/echo hi`, `/bin/ls`, anything you name by full path. Commit and stop here.