Re: Strange behaviour: nested def in proxy method body
Ah, thanks for your reply, both of you! That's why I got confused: it seemed to work when I didn't use proxy. I'll stop using nested defs for now. But I will miss them: they allow me to evaluate them top-level in the REPL (using C-x C-e in emacs) and play around with them. K. On Thursday, December 13, 2012 7:56:50 PM UTC+1, Alan Malloy wrote: On Thursday, December 13, 2012 4:14:23 AM UTC-8, Marshall Bockrath-Vandegrift wrote: kristianlm kris...@adellica.com writes: I'm enjoying testing Java code with Clojure and it's been a lot of fun so far. Coming from Scheme, the transit is comfortable. However, I encountered a big surprise when I nested def's and used them with a proxy: This is a common surprise for people with previous exposure to Scheme: `def` in Clojure is always explicitly namespace-scoped. What it does is create a var with a name and intern it into the namespace with that name, not introduce a name in the current lexical scope. Nested `def`s are thus very rarely going to be what one actually want, and the behavior you’re seeing is exactly what one would expect. To further clarify: the mew* version doesn't *actually* behave just like new-let*, although it usually will have the same consequences. The issue is that there's a race condition if two different threads call mew* at the same time: they might both mutate i, then both read it, and thus end up with the same reference. Really, as Marhsall says, a nested def is (almost) never what you want. -- 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
Strange behaviour: nested def in proxy method body
Hello Clojurers! I'm enjoying testing Java code with Clojure and it's been a lot of fun so far. Coming from Scheme, the transit is comfortable. However, I encountered a big surprise when I nested def's and used them with a proxy: (defn new* [] (def i (atom 0)) (proxy [Object] [] (toString [] (str i (defn new*-let [] (let [i (atom 0)] (proxy [Object] [] (toString [] (str i) (defn mew* [] (def i (atom 0)) i) (pprint (list (new*) (new*) (new*) (new*-let) (new*-let) (new*-let) (mew*) (mew*) (mew*))) I expected to see a list of 3 * 3 elements, each element a unique Atmon instance. But from it's printout, the first three instances are identical. Shouldn't all versions of new* produce their own version of a lexically-scoped i (like in Scheme), or am I missing something? Are nested defs and lets not supposed to behave identically? Note that the non-proxied version, mew*, makes nested def work like let. Could this be a proxy-bug? K. -- 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
Re: Strange behaviour: nested def in proxy method body
kristianlm krist...@adellica.com writes: I'm enjoying testing Java code with Clojure and it's been a lot of fun so far. Coming from Scheme, the transit is comfortable. However, I encountered a big surprise when I nested def's and used them with a proxy: This is a common surprise for people with previous exposure to Scheme: `def` in Clojure is always explicitly namespace-scoped. What it does is create a var with a name and intern it into the namespace with that name, not introduce a name in the current lexical scope. Nested `def`s are thus very rarely going to be what one actually want, and the behavior you’re seeing is exactly what one would expect. -Marshall -- 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
Re: Strange behaviour: nested def in proxy method body
On Thursday, December 13, 2012 4:14:23 AM UTC-8, Marshall Bockrath-Vandegrift wrote: kristianlm kris...@adellica.com javascript: writes: I'm enjoying testing Java code with Clojure and it's been a lot of fun so far. Coming from Scheme, the transit is comfortable. However, I encountered a big surprise when I nested def's and used them with a proxy: This is a common surprise for people with previous exposure to Scheme: `def` in Clojure is always explicitly namespace-scoped. What it does is create a var with a name and intern it into the namespace with that name, not introduce a name in the current lexical scope. Nested `def`s are thus very rarely going to be what one actually want, and the behavior you’re seeing is exactly what one would expect. To further clarify: the mew* version doesn't *actually* behave just like new-let*, although it usually will have the same consequences. The issue is that there's a race condition if two different threads call mew* at the same time: they might both mutate i, then both read it, and thus end up with the same reference. Really, as Marhsall says, a nested def is (almost) never what you want. -- 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
