[00:07:20] * ubolonton [ubolonton!~ubolonton@171.233.75.238] has joined the channel.
[00:43:21] * ubolonton [ubolonton!~ubolonton@171.233.75.238] has quit (Quit: ubolonton).
[05:08:40] * Sgeo_ [Sgeo_!~Sgeo@ool-18e4354b.dyn.optonline.net] has joined the channel.
[05:10:49] * Sgeo [Sgeo!~Sgeo@ool-18e4354b.dyn.optonline.net] has quit (Ping timeout: 240 seconds).
[05:31:16] * jtimon [jtimon!~quassel@117.29.134.37.dynamic.jazztel.es] has quit (Ping timeout: 246 seconds).
[08:34:09] <kernelj> decided to have stack effects like [ => ] or [ =>$ ] for ones that have side effects
[08:34:26] <kernelj> so if you have the first form you know it's pure
[08:36:32] <kernelj> actually nah, it might signify that it *might* be pure
[08:37:08] <kernelj> really I need a way to show that the purity is dependent on some inputs
[08:44:24] <kernelj> *might* be pure isn't very useful, *definitely* pure is
[09:01:24] <kernelj> so then =>$ is more general than =>
[09:37:04] <kernelj> @multi?[`@zss?{{`@*}+} head(zss).. ~v ~fs\({drop(zss) tail(zss)}zipwith:#xs#ys [xs.. v =>$ ys..]) =>$[fs..] top(zss)..]: #v#l l each[#f v f#]
[09:37:34] <kernelj> just gets more and more fun these types
[09:51:05] <kernelj> (realises that to_both is the same as bi)
[09:51:18] <kernelj> (both_to rather)
[09:52:42] <kernelj> meh I'm just gonna call to_both both again it's not like the proper both you can't just write dip[f#] g#
[09:56:17] <kernelj> actually I'll rename it twice
[09:56:27] <kernelj> then I can have thrice as well :D
[09:57:45] <kernelj> #f dip(twice(f)) f#
[09:59:26] <kernelj> fail I of course mean @thrice:#f dip[twice(f)] f#
[10:23:59] <kernelj> now I'm thinking I should rename twice and thrice to just apply2 and apply3
[10:29:12] <kernelj> interesting
[10:29:14] <kernelj> @curry?[`@xs `@ys ~t ~f\[xs.. t =>$ ys..] => [xs.. =>$[f] ys
[10:29:33] <kernelj> ..]]:              #x#f [x f#]
[10:30:20] <kernelj> so curry itself is pure but it creates a quotation that is unpure if the input one is impure
[10:41:24] <kernelj> weird... in kitten if you do [1 2] you get 1 [2]
[10:41:59] <kernelj> [1 2,3 4] gives 1 2 [3,4]
[10:44:20] <kernelj> 1 2 3 [(+) 4] gives 1 5 [4]
[10:45:34] <kernelj> not sure I like this way of doing things
[10:45:49] <kernelj> it requires your lists to be fixed arity
[10:47:44] <kernelj> Starpial gives you a new stack when you open with {
[10:48:13] <kernelj> with stack underflow, and everything you put in goes into the list
[10:48:50] <kernelj> Kitten doesn't even allow you to have trailing commas to control the arity :(
[10:50:40] <kernelj> define nop (->) { }
[10:51:36] <kernelj> [1 2 3 4 5, nop, nop, nop, nop]
[10:51:46] <kernelj> laaaaaaaame
[11:05:36] <kernelj> @compose?[`@xs `@ys `@zs ~f\[xs.. =>$ ys..] ~g\[ys.. =>$ zs..] => [xs.. =>$[f g] zs..]]: #f#g [f# g#]
[11:15:34] <kernelj> where does 'draw' in kitten draw to...?
[11:18:41] <kernelj> off topic: Exploding Kittens sounds like a fun card game
[11:32:26] <kernelj> ok I figured it out from looking at the code
[11:32:28] <kernelj> https://iterm2.com/documentation-images.html
[11:36:51] <kernelj> *sigh* mac only
[12:31:08] <kernelj> man I just thought of a epic major change to Starpial
[12:31:23] <kernelj> thinnking of getting rid of the whole {>} {<} pattern thing
[12:32:19] <kernelj> and just going with variables combined with .. (spill) * (kleene star with projections) + (like regex) and ? (like regex)
[12:33:20] <kernelj> came across this when in type land I end up building [=>] as a data structure and {} is also a data structure but I'm treating them differently
[12:33:36] <kernelj> they should all use the same syntax
[12:35:05] <kernelj> man this is such a huge change, good job I haven't written any interpreter yet
[13:07:34] * jtimon [jtimon!~quassel@117.29.134.37.dynamic.jazztel.es] has joined the channel.
[13:30:10] <kernelj> not sure the projections part makes sense
[13:37:05] <kernelj> would help if I'd actually used projections anywhere
[13:37:57] <kernelj> otherwise zipwith and map do the job
[13:42:55] <kernelj> but yeah I'm getting rid of all those awful `@ with this simple matching trick
[14:26:15] <hackerfoo> I've updated the asm.js version of the PoprC. It should be more robust now: http://hackerfoo.com/eval.html
[14:47:09] <kernelj> cool
[14:47:56] * lonjil [lonjil!~quassel@2a02:418:6050:ed15:ed15:ed15:e741:32d6] has quit (Quit: No Ping reply in 180 seconds.).
[14:48:40] * lonjil [lonjil!~quassel@2a02:418:6050:ed15:ed15:ed15:e741:32d6] has joined the channel.
[14:54:46] <kernelj> ok I think I'm going to keep projections
[14:55:25] <kernelj> when you have a normal binder enclosed in a * + or ? it turns into a list, then list of lists etc.
[14:55:57] <kernelj> can be used with lists {a b c}* or stack effects [xs.. => ys..]*
[14:58:01] <kernelj> if you don't want a projection you enclose in () so {a (b) (c)}* only a is projected
[14:58:21] <kernelj> b and c are fixed values that are repeated and bound elsewhere
[15:01:37] <kernelj> interestingly the new syntax was already valid according to one of my notes
[15:05:19] <kernelj> now I'm rewriting if/elif/else/endif so the combinator object doesn't use mutation and it just returns a new value of itself
[15:05:37] <kernelj> the way it was working it changed type which is kinda wrong
[15:06:12] <kernelj> the type is parameterised by an input vector and a list of possible stack effects from the various branches
[15:07:15] <kernelj> and then non-determinism is used in the type system to execute the branch
[15:07:41] <kernelj> it then unifies the possible branches with the actual branch that is taken
[15:11:30] * Sgeo_ [Sgeo_!~Sgeo@ool-18e4354b.dyn.optonline.net] has quit (Ping timeout: 240 seconds).
[15:22:39] * Sgeo [Sgeo!~Sgeo@ool-18e4354b.dyn.optonline.net] has joined the channel.
[15:36:19] <kernelj> I've still got the ifchain changing type..
[15:38:05] <kernelj> because when the branch is already set it returns itself
[15:42:23] <kernelj> I guess just because it's one type it can't be another more general one
[15:42:56] <kernelj> seems like a weird sort of subtyping
[15:50:44] <kernelj> anyway that looks done now, I shall continue on with the prelude
[16:26:40] <kernelj> @fix?[xs.. ~f\[xs.. [xs.. =>$[f] ys..] =>$ ys..] =>$[f] ys..]: #f f[f fix]#
[16:27:33] <kernelj> whether the argument to f is allowed to be pure depends on whether f is pure
[16:28:00] <kernelj> things getting more complicated and interesting...
[17:46:00] <kernelj> I think I'll rename a few shuffle words to be more like Factor
[17:47:09] <kernelj> dip -> over -> pick -> ... getn
[17:47:21] <kernelj> for the dependent variable arity one
[17:47:36] <kernelj> s/dip/dup
[17:49:18] <kernelj> currently I don't allow my words to start with digits so things are a bit backwards
[17:49:57] <kernelj> dup -> over -> pick -> nget
[17:50:08] <kernelj> dup2 -> over2 -> pick2 -> nget2
[17:50:14] <kernelj> dup3 -> over3 -> pick3 -> nget3
[17:50:46] <kernelj> 4 dupn -> 4 overn -> 4 pickn -> 4 ngetn
[17:51:09] <kernelj> I guess... that works...!!1!
[18:28:27] <kernelj> yeah the Factor names are a bit weird
[18:28:40] <kernelj> bi bi* bi@
[18:29:10] <kernelj> Starpial: bi both apply2
[20:02:13] <kernelj> mmmph I can't get the option syntax to work in Kitten
[20:08:49] <kernelj> I'm totally lost
[20:09:00] <kernelj> match doesn't seem to work even
[20:12:05] <kernelj> ok I'm using it wrong for it's new syntax
[20:29:35] <kernelj> don't see much reason for not keeping the old syntax for option and choice
[20:42:30] * tgunr [tgunr!~davec@cpe-76-173-83-160.hawaii.res.rr.com] has quit (Ping timeout: 240 seconds).
[20:56:21] * tgunr [tgunr!~davec@cpe-76-173-83-160.hawaii.res.rr.com] has joined the channel.
[21:04:54] <kernelj> ugh kitten doesn't have File I/O
[21:19:40] <tgunr> could anyone explain how to use popen in factor? A quick search thru the Factor sources reveals only the FFI declaration. I tried a "pwd" "r" popen and get back an alien, then I tried to read from the  FH but little success
[21:20:26] <tgunr> or is there a better way to interface to the shell? I want to make a program that controls my nftables
[21:45:15] <kernelj> now I don't know how to make a UInt32 in Kitten
[21:49:30] <kernelj> ah 0u32
[22:25:10] <kernelj> and yet the get/set list functions use Int32 and so I have to abandon the idea of using UInt32
[22:37:11] * flogbot [flogbot!~flogbot@2001:4800:7814:0:2804:b05a:ff04:4ba7] has quit (Ping timeout: 245 seconds).
[22:37:19] * flogbot [flogbot!~flogbot@2001:4800:7814:0:2804:b05a:ff04:4ba7] has joined the channel.
[22:37:20] :leguin.freenode.net 353 flogbot = #concatenative :flogbot tgunr Sgeo lonjil jtimon groovy2shoes FreeFull mollerse dustinm` MDude kernelj rotty rgrinberg flout carvite earl_ shachaf erg koisoke merry rjungemann jeaye ephe_meral doublec diginet m_hackerfoo hackerfoo strmpnk kanzure jeremyheiler bmp PiDelport shmibs otoburb puckipedia
[22:37:20] :leguin.freenode.net 366 flogbot #concatenative :End of /NAMES list.
[23:02:49] <kernelj> eeeeeeehhhhh I can't do match on numbers, how lame is this
[23:53:12] * evincar [evincar!~jon@173-228-12-107.dsl.dynamic.fusionbroadband.com] has joined the channel.