A bug in CL:EVERY with list (rest list).

[pjb@…]

Here is what is expected:

[pjb@kuiper :0 swig]$ clisp -norc -ansi -q [1]> (defun diff1p (lon)

(every (lambda (a b)

(print (list a b (= 1 (- a b)))) (= 1 (- a b)))

lon (rest lon)))

DIFF1P [2]> (diff1p '(6 5 4 3 2 1))

(6 5 T) (5 4 T) (4 3 T) (3 2 T) (2 1 T) T [3]> (quit)

Here is what we get with ccl:

[pjb@kuiper :0 swig]$ ccl -norc Welcome to Clozure Common Lisp Version 1.6-RC1-r14432M (LinuxX8664)! ? (defun diff1p (lon)

(every (lambda (a b)

(print (list a b (= 1 (- a b)))) (= 1 (- a b)))

lon (rest lon)))

DIFF1P ? (diff1p '(6 5 4 3 2 1))

(6 5 T) (4 5 NIL) NIL ? (quit) [pjb@kuiper :0 swig]$ ccl --version Version 1.6-RC1-r14432M (LinuxX8664)

[pjb@…]

The same problem occurs with some, notany and notevery, only when the two sequences are related (eg. lon and (rest lon)), and only in this order. 


CL-USER> (every (lambda (a b) (print (list a b))) '(1 2 3 4) '(a b c d))

(1 A) 
(2 B) 
(3 C) 
(4 D) 
T
CL-USER> (let ((lon '(1 2 3 4))) (every (lambda (a b) (print (list a b))) lon (rest lon)))

(1 2) 
(3 2) 
(4 3) 
T
CL-USER> (let ((lon '(1 2 3 4))) (some (lambda (a b) (print (list a b))) lon (rest lon)))

(1 2) 
(1 2)
CL-USER> (let ((lon '(1 2 3 4))) (some (lambda (a b) (print (list a b)) nil) lon (rest lon)))

(1 2) 
(3 2) 
(4 3) 
NIL
CL-USER> (let ((lon '(1 2 3 4))) (notany (lambda (a b) (print (list a b)) nil) lon (rest lon)))

(1 2) 
(3 2) 
(4 3) 
T
CL-USER> (let ((lon '(1 2 3 4))) (notevery (lambda (a b) (print (list a b)) nil) lon (rest lon)))

(1 2) 
T
CL-USER> (let ((lon '(1 2 3 4))) (notevery (lambda (a b) (print (list a b)) t) lon (rest lon)))

(1 2) 
(3 2) 
(4 3) 
NIL
CL-USER> (let ((lon '(1 2 3 4))) (notevery (lambda (a b) (print (list a b)) t) lon '(a b c d)))

(1 A) 
(2 B) 
(3 C) 
(4 D) 
NIL

[gb]

(In [14788]) In SOME-XX-MULTI, traverse the list of sequences in a way that doesn't require us to try to find the current sequence via MEMQ; this could affect the wrong sequence if some sequences shared structure (and wasn't the smartest thing to do even if they didn't.) Fixes ticket:859.