04:01:11 <stepkut> #karma
04:01:11 <acid-bot> unexpected "k"; expecting "dice"
04:02:46 <stepkut> #karma
04:02:47 <acid-bot> unexpected "k"; expecting "dice"
04:02:47 <acid-bot> unexpected end of input; expecting white space, "succ", "pred" or "show"
04:03:01 <stepkut> #karma succ synthea
04:03:01 <acid-bot> unexpected "k"; expecting "dice"
04:03:01 <acid-bot> synthea now has 1 karma points.
04:03:05 <stepkut> hmm
17:00:15 <stepcut> donri: regarding IxSet safety.. probably. Check out the API for kdtree, http://darcs.monoid.at/kdtree/
17:03:38 <stepcut> also, http://www.google.com/url?sa=D&q=http://src.seereason.com/haskell-kdmap/examples.hs&usg=AFQjCNEbhypjnQ_-EqyZkmEZ7_PLjt-YDg
17:03:40 <stepcut> gah
17:03:44 <stepcut> http://src.seereason.com/haskell-kdmap/examples.hs
17:06:00 <stepkut> actually the example.hs is not helpful
17:06:32 <stepkut> ACTION is still looking
17:10:22 <stepkut> well I can't find my original post
17:10:25 <stepkut> but if you look at:
17:10:28 <stepkut> darcs get http://darcs.monoid.at/kdtree/
17:10:36 <stepkut> and look at Data.KdTree.Strategy
17:10:46 <stepkut> you will see that the query functions have types like:
17:10:53 <stepkut> (@=) :: (Indices a,Index k a,Typeable a) => KdTree a -> k -> KdTree a
17:11:17 <stepkut> so, you can only look up a key 'k' in 'KdTree a', if 'Index k a' exists
17:11:31 <stepkut> we could do the same thing in IxSet as well
17:12:15 <stepkut> though, an IxSet can (hypothetically) be extended with additional keys at runtimein theory at lea
17:12:44 <stepkut> gah. This new trackpad misfires a lot
17:23:14 <donri> interesting, is kdtree some sort of alternative to ixset?
17:30:43 <donri> doesn't look like type families though?
17:54:08 <stepcut> donri: kdtree is an alternative to ixset that should be more efficient when doing multi-key queries
17:54:28 <stepcut> donri: I don't think it requires type families. just multi parameter type classes
17:54:54 <donri> have you seen https://github.com/lpeterse/HiggsSet ?
17:55:23 <stepcut> no.. looks useful
17:55:34 <donri> i have no idea though how it compares to ixset/kdtree
17:55:57 <stepcut> me neithre
17:56:06 <stepcut> but I am all for more competition in that area
17:57:20 <donri> equals :: (Indexable a, Index i, IndexOf a ~ i) => i -> HiggsQuery a i (Selection i)
17:57:45 <donri> looks like it also adds some type safety here
17:57:48 <stepcut> yeah
17:57:59 <stepcut> gotta run, bbl.
17:58:00 <donri> bbl dinner
17:58:03 <donri> haha, bye
20:21:06 <donri> stepcut: re GET should imply HEAD; also: you currently get a 404 for unmatched methods, should be a 405
20:24:04 <stepkut> ah
20:24:29 <donri> easy enough to add yourself of course
20:24:31 <stepkut> that seems a bit trickier to implement in an automatic way
20:24:39 <donri> nah
20:24:44 <stepkut> ?
20:24:55 <donri> hold
20:26:07 <stepkut> 405 is for a method we don't allow at all?
20:26:21 <stepkut> not just one that is invalid for a specific URL?
20:27:11 <donri> I'm not sure but I *think* 404 is only for GET really
20:27:24 <donri> or at least that would be an improvement: 404 if no serverpart matched and method [GET,HEAD] otherwise 405
20:28:03 <donri> but i'm probably wrong: should only 405 if the resource *does* exist (a handler matched on URL) but method is not allowed (handler failed to match on method)
20:28:17 <donri> which doesn't play well with how filtermonad works as you noted
20:28:38 <donri> uh, not filtermonad -.-
20:29:30 <donri> perhaps method should use webmonad to force a 405 early?
20:37:45 <stepcut> going to gymnastics, bbiab
20:37:50 <donri> have fun