#clojure log - Apr 05 2008

The Joy of Clojure
Main Clojure site
Google Group
IRC
List of all logged dates

16:38 jgracin: I see that var *agent* is used in the ants demo. Is this var documented somewhere?

16:39 rhickey: not yet

16:39 it is a reference to the current running agent

16:40 since actions are functions of the state, but sometimes actions need to send messages to self or communicate agent identity to someone else

16:41 still experimental, but working out ok...

16:41 jgracin: Thanks!

18:56 rhickey: enhanced host syntax: http://groups.google.com/group/clojure/browse_frm/thread/dd64bd49191f8062

19:11 Chouser: Interesting.

19:12 The current syntax hadn't bothered me at all, but I guess I'll be happy to drop the extra parens on method calls.

20:36 I thought I had this solved already, but apparently I don't.

20:37 rhickey: Is there a right way for me to determine if an object is a zipper location?

20:38 rhickey: not at present

20:39 why do you need to?

20:43 Chouser: my zip-filter thingy generally expects each function to return a sequence of locs, but it's very convenient to use things like zip/up directly.

20:43 If I can detect that I'm being given a loc instead of a sequence, I can wrap a list around it automagically.

20:44 We talked about this before, and I was trying to tack my own meta-data key on to the root of the zip tree, and then look for it in the child nodes.

20:44 I don't know why I thought that would work.

20:44 rhickey: if the fns are really returning seqs, that is distinguishable from the loc vectors

20:44 seq? vector?

20:45 still relying on the implementation of locs though...

20:46 does the metadata not get carried through?

20:50 Chouser: why would it?

20:50 but no, it appears to not be, assuming there's not bug in my test.

20:51 rhickey: (def data '[[a * b] + [c * d]])

20:51 (def dz1 (vector-zip data))

20:51 (def dz (with-meta dz1 (assoc ^dz1 :my-meta "foo")))

20:52 (:my-meta ^(zip/right (zip/down (zip/right (zip/right (zip/down dz))))))

20:52 -> "foo"

20:53 Chouser: hmph

20:54 rhickey: there might be an unhandled case, but the intention is propagation, as is for all data structures

20:56 Chouser: indeed it appears to work fine now. I guess I'll blame my failing test on gremlins.

20:59 sorry for the false alarm

Logging service provided by n01se.net