no, we have not moved past this, and the error is absolutely because you take too much for granted about lazy behavior
(defn divides? [x y] (zero? (mod x y))) (defn prime-ub [x] (/ x (if (even? x) 2 3))) (defn lower-primes [primes x] (let [ub (prime-ub x)] (println "primes" primes "x" x) (take-while #(<= % ub) primes))) (defn prime? [primes x] (not-any? #(divides? x %)(lower-primes primes x))) user=> (def primes (cons 2 (lazy-seq (filter #(prime? primes %) (drop 3 (range)))))) #'user/primes user=> (take 20 primes) (primes (2) x 3 primes (2) x 4 primes (2) x 5 primes (2) x 6 primes (2) x 7 primes (2) x 8 primes (2) x 9 primes (2) x 10 primes (2) x 11 primes (2) x 12 primes (2) x 13 primes (2) x 14 primes (2) x 15 primes (2) x 16 primes (2) x 17 primes (2) x 18 primes (2) x 19 primes (2) x 20 primes (2) x 21 primes (2) x 22 primes (2) x 23 primes (2) x 24 primes (2) x 25 primes (2) x 26 primes (2) x 27 primes (2) x 28 primes (2) x 29 primes (2) x 30 primes (2) x 31 2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 32 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 33 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 34 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 35 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 36 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 37 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 38 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 39 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 40 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 41 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 42 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 43 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 44 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 45 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 46 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 47 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 48 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 49 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 50 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 51 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 52 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 53 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 54 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 55 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 56 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 57 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 58 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 59 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 60 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 61 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 62 primes (2 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31) x 63 31 37 41 43 47) On Thursday, February 12, 2015 at 6:47:34 PM UTC-8, Jorge Marques Pelizzoni wrote: > > > So why isn't 4 and any even numbers in the result list? Empty primes' > allows everything to pass. We are already beyond this. I've already posted > that even this does not work: > > (def primes (cons 2 (lazy-seq (filter #(prime? primes %) (drop 3 > (range)))))) > > > Em sexta-feira, 13 de fevereiro de 2015 00:39:03 UTC-2, Justin Smith > escreveu: >> >> Considering for the sake of argument the possibility that it is a >> legitimate bug, and not a result of misusing the language features, it is a >> family of bug that will be more common than most, because it reflects a >> style of programming that is rare in real Clojure code. >> >> But it isn't a bug. >> >> Lazy-seqs don't promise to be "as lazy as possible", and when we step >> through the logic of the original code in Michael Blume's post, we can see >> why chunking breaks it. Taking for granted that something lazy is realized >> with a specific granularity is an error in Clojure (see also lazy-seqs used >> inside with-open blocks or db transactions - the opposite error is common, >> not being strict enough). >> >> defn primes [] (let [primes' (atom nil)] >> (reset! primes' (cons 2 (filter #(prime? @primes' %) >> (drop 3 (range))))))) >> >> The first time through the above code, primes' will be empty, so the >> first chunk of results will all be calculated with an empty list of >> candidate divisors. >> >>> -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
