🚀 go-pugleaf

In article <JGREELY.90May10033656@oz.cis.ohio-state.edu> jgreely@oz.cis.ohio-state.edu (J Greely) writes:
   > Ran into what looks like a bug in the byte-compiling functions.  The
   > following function returns t unless it's byte-compiled (I think this
   > is a minimal test-case :-)):
   > 
   > 	(defun generate-bad-byte-code ()
   > 	  (cond (t)))
   > 
   > Disassembling it returns the following:
   > 
   > 	byte code for generate-bad-byte-code:
   > 	 args: nil
   > 
   > 	0   constant nil
   > 	1   return

For 18.55, the following fixes this & gives the behaviour documented in
Steele (I don't have the Elisp ref manual which is what it *should*
agree with!).

[diff -c bytecomp.el-orig bytecomp.el]

*** bytecomp.el-orig	Fri May 11 02:50:31 1990
--- bytecomp.el	Fri May 11 03:03:47 1990
***************
*** 944,950 ****
    (if (or (eq (car (car clauses)) t)
  	  (eq (car-safe (car (car clauses))) 'quote))
        ;; Unconditional clause
!       (byte-compile-body (cdr (car clauses)))
      (if (null (cdr clauses))
  	;; Only one clause
  	(let ((donetag (byte-compile-make-tag)))
--- 944,951 ----
    (if (or (eq (car (car clauses)) t)
  	  (eq (car-safe (car (car clauses))) 'quote))
        ;; Unconditional clause
!       ;; Handle case of (t) correctly
!       (byte-compile-body (or (cdr (car clauses)) (car clauses)))
      (if (null (cdr clauses))
  	;; Only one clause
  	(let ((donetag (byte-compile-make-tag)))

--tim
Tim Bradshaw.  ARPA: tim%ed.cstr@nsfnet-relay.ac.uk
UUCP: ...!uunet!mcvax!ukc!cstr!tim  JANET: tim@uk.ac.ed.cstr
"Quis custodiet ipsos custodes?"