Inconsistent results in SBCL

Question

Consider the following Common Lisp code.

(print (let* ((x nil) (y x)) (setq x t) (and y x)))  ; T
(print (let* ((x nil) (y x)) (setq x t) (and x y)))  ; NIL
(print (let* ((x nil) (y t)) (setq y x) (setq x t) (and y x)))  ; NIL
(defun noop (x) (when x (princ "")))
(print (let* ((x nil) (y x)) (noop y) (setq x t) (and y x)))  ; NIL
(print (let* ((x nil) (y x)) (setq x t) (noop y) (and y x)))  ; NIL

Running the above code in SBCL 2.4.9 in linux, the first line prints T while the remaining lines print NIL. This online SBCL interpreter shows the same results. This is not the case in CLISP, for instance, where all the lines print NIL.

Why am I getting this seemingly strange result?

UPDATE 1: I reported this to the SBCL bug tracker: https://bugs.launchpad.net/sbcl/+bug/2092588

UPDATE 2: Apparently, one of the project's maintainers committed a fix for this bug in response to my report. That was really fast!

@RainerJoswig, Does that mean that I should report this to the SBCL bug tracker? — single-question-123
– single-question-123, Commented Dec 27, 2024 at 15:28
That would be a good idea! That would be very helpful. Thanks! — Rainer Joswig
– Rainer Joswig, Commented Dec 27, 2024 at 15:34
Most likely an optimizer bug. Does changing (DECLARE (OPTIMIZE ...)) settings affect the result? — Barmar
– Barmar, Commented Dec 27, 2024 at 15:53
Happens with DEBUG qualities 0 and 1. Here with macOS 15.2 SBCL 2.4.11 on an M4 Pro. — Rainer Joswig
– Rainer Joswig, Commented Dec 27, 2024 at 16:08

Zeta · Accepted Answer · 2025-08-27 11:13:29Z

2

As you said yourself, this was a compiler bug in versions up to 2.4.11. It was fixed in 2.5.0:

bug fix: fix miscompilation of IF where the consequent and alternative would have the same value but for an intervening side-effect. (#2092588, reported by JA)

It was apparently introduced in 1.4.14:

zeta@machine:$ cat bad.lisp
(let* ((x nil) (y x)) (setq x t) (if (and y x)
                                   (format t "BUG~%")
                                   (format t "Fine~%")))
zeta@machine:$ podman run --quiet --rm -it -v `pwd`:/tmp fukamachi/sbcl:1.4.13-alpine --script /tmp/bad.lisp
Fine
zeta@machine:$ podman run --quiet --rm -it -v `pwd`:/tmp fukamachi/sbcl:1.4.14-alpine --script /tmp/bad.lisp
BUG

If you use Debian, then all versions up to Debian 12 are affected.

answered Aug 27 at 11:13

Zeta

106k13 gold badges205 silver badges249 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Inconsistent results in SBCL

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related