After a fair amount of macro writing in CL, I think I got the backquote expansion mechanism pretty much nailed down (after all, it is quite clear in the HyperSpec). Now, I thought I understood the Clojure algorithm just as well. It looked very similar (concat instead of append and so on). Apart from the namespace resolution business, and a few other irrelevant details, the essence looked the same.
But I did the following just for fun, and I got a surprising result: > (def x `a) > (def a 5) > `(list `(list ,,x)) What I theoretically was expecting was the outcome of this: (concat (list `list) (list (concat (list `concat) (list (concat (list `list) (list ``list))) (list (concat (list `list) (list x)))))) In other words, behind the curtains, the latter was supposed to be evaluated. Now, as Guy Steele points out, with a k-level nested backquote, you are guaranteed to get the same result only after k evaluations. So evaluating twice the result is: ((clojure.core/list 5)) I checked this in Common Lisp and the result was sure enough exactly the equivalent, as I expected. Yet, in clojure this is what happens (of course you see the first evaluation directly in this case): > `(list `(list ,,x)) (clojure.core/list (clojure.core/concat (clojure.core/list (quote clojure.core/list)) (clojure.core/list (quote user/x)))) Another evaluation gives: ((clojure.core/list user/x)) That's ALMOST the same but not quite. See the "user/x"? It came from a very strange (quote user/x) above which wasn't supposed to be there. I mean user/x should not have been quoted. Is this a bug? Rock --~--~---------~--~----~------------~-------~--~----~ 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 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 -~----------~----~----~----~------~----~------~--~---
