01:40 (quit) eli: Quit: rcirc on GNU Emacs 23.1.1
03:17 (quit) jonrafkind: Ping timeout: 264 seconds
04:32 (quit) carleastlund: Quit: carleastlund
08:32 (quit) sstrickl: Quit: sstrickl
09:30 (join) sstrickl
11:34 (join) samth
12:25 (join) jonrafkind
12:28 (join) carleastlund
12:54 (quit) carleastlund: Quit: carleastlund
13:44 (join) carleastlund
14:46 jay-mccarthy: jonrafkind: mongodb is json based too
14:47 jay-mccarthy: mattmight & jonrafkind: it's so much a whole program compiler as a whole program finder
14:47 jay-mccarthy: it takes a module's bytecode and removes all dependencies by putting the phase 0 import definitions in the module and adjusting the references
14:48 jay-mccarthy: thus it takes a modular program and finds the whole program
14:49 jay-mccarthy: i intend to write a translation between byte-code and un-resolved code (basically what the expand produces and the current optimizer/compiler consumes) so that i can use the existing optimizations again without module boundaries
14:49 jay-mccarthy: but it would be nice to make new optimizations, perhaps at the bytecode level
15:11 samth: jonrafkind, is this channel being logged somewhere?
15:11 jonrafkind: yea by gabot
15:11 jonrafkind: eli said he would make the logs public at some point
15:11 samth: where?
15:12 jonrafkind: eli's server I guess
15:14 (topic) samth: Racket: http://racket-lang.org (logged @ ???)
16:16 jay-mccarthy: i just wrote a totally gnarly macro
16:16 jonrafkind: paste it up yo
16:17 jay-mccarthy: it expands its argument using local-expand with a side-effecting syntax-parameter to record occurrences of the syntax-parameter, then uses that information in its final output which causes the argument to expand again, but this time a little differently
16:18 jay-mccarthy: the point is to allow you to use slideshow's with-steps without having to choose and name the number of steps before-hand
16:20 jay-mccarthy: http://gist.github.com/421425
16:20 jay-mccarthy: there's a snippet
16:30 jonrafkind: hm, anyone understand the semantics of #:at in syntax-parse exactly?
16:30 jonrafkind: i think my understanding of it is flawed somehow
16:30 jay-mccarthy: i've never used it but i can read the docs and double check with you
16:30 jonrafkind: they dont seem to be documented :p
16:30 jonrafkind: ryan explained it to me one day, but hes not in the office today
16:31 jonrafkind: #:phase is documented but im not sure if I want to use that
16:31 jonrafkind: actually #:at might have been documented in the latest version, im on an old git revision
16:32 jay-mccarthy: i am going to guess that the literal is compared using the lexical context that context-id evals to
16:32 jay-mccarthy: perhaps for use with a namespace anchor or another particular pattern variable
16:33 jonrafkind: yea thats more or less it, im getting literal mismatches and I think its because I chose a bad identifier for the context-id
16:35 jonrafkind: oh ryan got rid of #:phase even
16:40 samth: jay-mccarthy, that is frightening
16:41 jay-mccarthy: :)
16:41 jay-mccarthy: i thought YOU would think so
16:42 samth: I think you could do that with continuations
16:42 jay-mccarthy: with-steps needs the steps as syntax
16:43 samth: yeah, you couldn't use `with-steps'
16:43 jay-mccarthy: (btw the first half avoids the non-hygeine of with-steps by making its own stxparams)
16:44 samth: also, you should never use with-steps - it's terrible
16:44 jay-mccarthy: blake asked if it could be done. i said i didn't think it was a good idea but it could be done :P
16:44 samth: check out carl's slideshow step macros
16:44 jay-mccarthy: i've seen them
16:44 samth: they're *so* much nicer
16:44 jay-mccarthy: they should be part of his core contribution
16:45 jay-mccarthy: i had never seen them before that recent email thread though
20:01 (quit) sstrickl: Quit: sstrickl
20:01 (nick) samth -> samth_away
20:05 (quit) carleastlund: Quit: carleastlund
20:31 (quit) jonrafkind: Ping timeout: 276 seconds
20:42 (join) carleastlund
21:37 (join) sstrickl
23:53 (join) jonrafkind