Unfixed bugs (TODO)
Can't iterate over nested Hashtables
Found by Bob Hutchison. It's related to 2.1, "Macrolets in the iterate form":
(iter (for (k1 v1) in-hashtable ht1) (iter (for (k2 v2) in-hashtable ht2) (format t "~A ~A~%" k1 k2)))
This fails with the typical macrolet error. Why? Hah.
It's the fault of the new IN-HASHTABLE code that makes use of WITH-HASHTABLE-ITERATOR. The iterator is specified to be defined by a symbol-macrolet. I think we'll have to make a rudimentary macrolet walker. Argh. Hmpf.
This affects nested package walking, as well. A solution to one will probably be a solution to both.
PREVIOUS and INITIALLY
(iter (for el in '(1 2 3 4)) (for p-el previous el) (for pp-el previous p-el initially 0) (collect pp-el))
gives (0 NIL 1 2)
; should be (0 0 1 2)
according to the manual. I
suspect it's a doc bug. Found by Carlos Ungil.
(iter (for el in '(1 2 3 4)) (for pp-el previous el back 2 initially 0) (collect pp-el))
gives the right result.
So does:
(iter (for el in '(1 2 3 4)) (for p-el previous el initially 0) (for pp-el previous p-el initially 0) (collect pp-el))
Without an INITIALLY 0, p-el gets set to NIL the first time through the loop. So, pp-el's value is NIL the second time through the loop. The for...previous documentation talks about the "first" and "second" value. But it doesn't detail if the initial value counts as a "first" value.
Wishlist
Macrolets in the iterate form
(iterate (for i from 0 to 20) (macrolet ((collect-if-divisible (variable divisor result-var) `(when (zerop (mod ,variable ,divisor)) (collect ,variable into ,result-var)))) (collect-if-divisible i 3 result-3) (collect-if-divisible i 5 result-5)) (finally (return (values result-3 result-5))))
would be nice to have. Currently has to be rewritten:
(macrolet ((collect-if-divisible (variable divisor result-var) `(when (zerop (mod ,variable ,divisor)) (collect ,variable into ,result-var)))) (iterate (for i from 0 to 20) (collect-if-divisible i 3 result-3) (collect-if-divisible i 5 result-5) (finally (return (values result-3 result-5)))))
of course, this does not work in all cases.
Already done (DONE)
synonyms
Another bug found by Carlos Ungil. I have now imported most examples from the manual into the test suite (-:
(iter (generating (key . item) in '((a . 1) (b . 2) (c .3))) (collect (next key)) (collect (next item))) TYPE ERROR : The value ITEM is not of type LIST. in the manual: (a 2 c)
iter:symbol-synonym
was broken.
spurious warnings when using QUOTE, GO or the FOR...NEXT clause
(let ((y '(a b c))) (iter (for x next (if y (pop y) (terminate))) (collect x)))
gave:
; Warning: Unknown special form GO. The lisp environment claims ; that GO is a special operator, but ITERATE doesn't know how to ; handle it. ; While executing: ITERATE::WALK-SPECIAL-FORM (A B C)
that is, the right result plus one or more warnings about unknown special operators. This was a bad check that I had introduced to catch other bugs in implementations that kept some macros as special operators. The check is fixed in versions of iterate from 1.0.5 upwards.
Didn't build on Corman Lisp
reported by Carlos Ungil; fix at http://www.artofprogramming.com/bb/viewtopic.php?t=85
Applied. There's still the question of the #\Nul vs. #\Null character (both of which are not valid characters by the ANSI spec. hmpf)
Multiple-value-prog1 wasn't recognized as a special form
This was a punctuation/space error. Spotted and fixed by Simon Katz.
Administrativa
arch-tag: de16d41a-b240-11d8-a267-000c76244c24
You can find the most recent version of this file at http://common-lisp.net/project/iterate/bugs.html