🚀 go-pugleaf

>I'm writing an elisp function that calls a subprocess and waits for
>completion.  The obvious way to do this is to use 'call-process', but
>I need the process exit status before the elisp can continue with
>confidence.  In all the tests I've run, 'call-process' returns 'nil'
>regardless of the exit status.
>
>Is there a way to obtain the exit status besides resorting to
>'start-process' and using a process-sentinal?  The function I'm
>writing shouldn't require such hair, but I *must* have the exit
>status.



This is most assuredly a hack.  Nonetheless I offer it because it has a certain
beauty to it - similar to that of well-done "objet-trouvee" art.

(let ((buf (get-buffer-create "*Turnin Error*")))
      (set-buffer buf)
      (erase-buffer)
      (call-process "csh" nil buf nil
                    "-fc" (concat "compress" " -f" "< " source-file ">!
" target
-file ";if ($status) then\necho ERROR MESSAGE from TURNIN\nelse\necho
OK\nendif"
 ))
      (if (equal (buffer-string) "OK\n")
          (progn (kill-buffer buf)
                 (message "Your File Was Turned In Successfully."))
        (progn
          (goto-char (point-max))
          (princ "

                         **** E R R O R ****

Your File Was NOT Turned In.

Try Again.  But if that doesn't work, check with a TA.


**** BE SURE TO WRITE DOWN THE ERROR MESSAGE AT THE TOP OF THE BUFFER!!! ****"
                 buf)
          (pop-to-buffer buf)
          (other-window 1)
          (beep)
          (message "C-x 1 to remove *Turnin Error* buffer")
          ))

The idea is simply to force the shell to test the error status and to write to 
standard output the result of the test.  call-process will place the standard
output in a buffer where the full power of emacs is available to test the
error status.