00:00:14 which inner [] bind by reference 00:00:15 I can always pass the vector -say- to itself as a field? 00:00:24 yes 00:00:36 but the vector isn't part of th DAG 00:00:39 it's separat 00:00:52 you just have a reference to vector that is a primitive type like a number 00:00:53 Ah. Hmm. 00:01:04 that's what lisp does 00:01:28 and python and ruby etc 00:01:47 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77106&oldid=77105 * SoundOfScripting * (+468) /* Language specifications */ 00:01:55 All those languages allow cycles. Only early Lisp rewrote DAGs. 00:02:10 I don't think even early lisps did 00:02:24 Yah, sure. The first one did. I read the report. 00:02:25 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77107&oldid=77106 * SoundOfScripting * (-1) 00:03:28 I think all serious lisps have set-car and set-cdr, even the oldest ones (possibly with a different name), but these days we want programs to not use it anywhere, so that conses are known to be globally immutable and the interpreter can optimize using that 00:03:55 and yes, some toy lisps might not have it, like my toy lisp for example 00:04:13 Nono. The actual first Lisp was reference counted. 00:04:33 By -whatshisname- McArthy. In the report. 00:04:35 yes 00:04:46 that doesn't stop this 00:05:21 you can just end up in uncollectable cycles if your program is not careful 00:05:55 but you can free a cycle either by breaking it, or using some newer data structure that allows weak referencing 00:06:31 Right. My hope was that just being pure would cut it. But it's all way to slow. If it wasn't as slow as it is now, I wouldn't add vectors. 00:06:48 and it's John McCarthy (1927..2011) 00:06:58 Ah right. 00:07:03 McCarthy. 00:07:37 if you don't want mutability, you can consider lazy promises that you have to explicitly evaluae too 00:07:53 that solves a few of the problems, though not most 00:08:18 but if you want to keep immutability completely, you can also do that 00:08:27 I would like to avoid mutability but it doesn't seem worth it. 00:08:50 -!- adu has joined. 00:09:00 In the sense that, I should give programmers mutable fast containers where they can't use lists. 00:09:11 But that's assuming 'programmers'. 00:09:13 Tall order. 00:09:15 I think it's worth in a non-toy language, but this is a toy language 00:09:41 -!- Arcorann__ has joined. 00:09:47 Yah. But sometimes I flip opinion and hope it could be something like a bash or python. 00:09:50 sorry, that's conflicting 00:09:58 And then I think: nah. 00:10:02 I think it's worth *to adds mutable structures* in a non-toy language, but this is a toy language 00:10:12 And then I flip opinion again. 00:10:24 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77108&oldid=77107 * SoundOfScripting * (+341) /* Language specifications */ More instructions :) 00:11:44 I admit that olvashato doesn't have mutability, because I didn't want them for the programs I wrote, but you could add them in a library 00:11:59 it's a bit messy because of the prolog side, but possible 00:12:26 Today I wanted to compare my abysmal performance to Python. 00:12:43 my toy lisp does have mutability 00:12:48 Which gave an out of stack after going 900 deep into recursion. 00:13:01 heh 00:13:03 So then I thought: maybe there's a use case after all. 00:13:14 And now I think: nah. 00:13:37 ¯\_(ツ)_/¯ 00:13:48 what if esolang where computations have to be done with constantly changing operator precedence, and said operators modify precedence? 00:27:44 I could add explicitly managed vectors. Where you indirectly reference a vector pool with an index. And you're responsible for managing memory. 00:28:45 But if you're managing memory already you might as well make sure that your vector doesn't contain cycles. 00:29:11 Ah well. Gonna hack on the double semicolon, I think. 00:29:55 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77109&oldid=77108 * SoundOfScripting * (+78) 00:30:07 [[1CP=1ICL]] https://esolangs.org/w/index.php?diff=77110&oldid=77109 * IFcoltransG * (+9) /* Language specifications */ Commented out unnecessary section 00:32:21 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77111&oldid=77110 * SoundOfScripting * (+3) 00:33:22 spruit11: by "explicitly managed", do you mean that the user has to free them explicitly, they're not refcounted or tracked by the refcounter or garbage collector that handles your DAG nodes? 00:33:47 to be clear, all that mutability stuff is less important than what I mentioned earlier 00:34:11 Right. You reference a vector indirectly in a pool. That way you break the cycle. But the cost for that is also that you would need to free it. 00:34:35 hmm, that would be much more inconvenient 00:34:52 I was thinking it would at least be refcounted 00:35:12 Yah. Doesn't seem worth it. 00:35:31 Oh. 00:35:35 Wait. 00:35:56 I could free it if the reference is destroyed? 00:36:02 What voodoo is this? 00:36:12 that is the voodoo of recounting 00:36:36 I mean don't you already have something like that for the DAG nodes? 00:37:23 Uh. I abuse C++'s native refcounting. 00:38:36 C++'s "native refcounting"? what do you mean? do you mean std::shared_ptr, which is a reference-counted smart pointer type? 00:38:42 Yah. 00:38:49 why is that an abuse? 00:39:01 you can use that for the vectors too 00:39:35 Uh, yes and no. 00:40:05 I need a form of indirection over a vector pool. But that's all yes. 00:40:40 I don't see why it's an abuse, and I'm not sure in what sense it counts as "native", but then that's because I remember the past when that wasn't in the standard library yet, and also I know it's implemented as a pure library feature 00:41:02 um... vector pool? 00:41:48 what why? 00:42:04 Well. The thing is that in the 'graph' you construct you don't want cycles. But if a 'vector' object only stores an index into a global vector pool, you can't create cycles. 00:42:49 sure you can, if it's reference-counted both ways, as in the vector keeps alive the dag node it points to and the dag node keeps alive the vector it points to, then it can form cycles 00:42:56 it doesn't matter if you allocate them from a pool or not 00:43:23 Damn. 00:43:35 Right. 00:43:49 No vector pool than but unsafeness. 00:44:00 why? what's unsafe in this? 00:44:02 Thanks for that. For a moment I thought I found a way out. 00:44:28 make it reference-counted both ways, and document that it's only reference counted so if you leave a cycle it won't be collected and that that's the user's responsibility 00:44:32 The unsafe part is that I would allow the programmer to form cycles in what should be a DAG. 00:44:52 Right. We agree. 00:46:04 but it's not a cycle that will bother the pattern matcher or the pretty-printer or anything, because the pattern matcher doesn't go arbitrarily deep through vectors because there's no such pattern, and you can just make the pretty-printer either not print the contents of vectors at all, only a reference, like it doesn't print the content of a function now, or make it track which vector it met and not 00:46:10 print repeated vectors 00:46:25 so I don't see why the cycles are a problem 00:47:09 unless this is like your PhD thesis and you have "DAG" prominently in the plan you submitted and now you have to resubmit the plan and wait for permission from a committee to be allowed to write non-DAGs 00:47:12 I think I fully agree with you. 00:47:41 Nono. If I would be doing a PhD I would already been kicked out for not having a type system. 00:49:09 [[1CP=1ICL]] https://esolangs.org/w/index.php?diff=77112&oldid=77111 * IFcoltransG * (+875) /* Language specifications */ Add oculi for converting between bits and bytes 00:49:52 well, technically there's another case when you don't want this: if you want to make a sandbox version of egel to run untrusted code by multiple users (like a javascript interpreter in an old browser that runs everything in one procesS), and want to be able to free every object created by a user by just freeing a few top-level references, in which case you need either a garbage collector, or allocate 00:49:58 everything into arenas, different users never share an arena, and free each arena owned by a user when you're done with them 00:50:10 but you would need a lot of other changes too if you wanted that 00:50:53 Life is never easy. 00:51:34 I already regret compiling to bytecode. And I also regret having a global pool for recursive structures. 00:51:53 Life would be so much nicer without those two. 00:52:10 spruit11: ah yes, that's when you make an incompatible egel2 00:52:25 where you keep the ideas you like and redesigne and rewrite everything 00:53:36 [[1CP=1ICL]] https://esolangs.org/w/index.php?diff=77113&oldid=77112 * IFcoltransG * (+291) Clarified Tokigun's half cell movements 00:53:44 Hmyah. I'll end up with a toy of a toy. 00:53:49 A noble cause! 00:54:27 Anyway, you got me fired up. I am hacking on the double semicolon now. 00:54:37 nice 00:59:46 Isn't a double semicolon just one whole colon? 01:00:37 fizzie: no, because English sucks 01:00:48 but good point 01:02:11 spruit11: also, if you do a major rewrite where you keep some code and rewrite some of it, you will need to learn branching in version control 01:06:17 [[1CP=1ICL]] https://esolangs.org/w/index.php?diff=77114&oldid=77113 * IFcoltransG * (+273) /* Language specifications */ Added equality check command 01:19:02 I have not used branched version control so far, but that would probably be one case where it could be used. 01:23:52 -!- egelbot has joined. 01:24:14 egelbot: using System;; 1+1;; 2+2 01:24:14 2 01:24:14 4 01:30:22 egelbot: using String;; strlen "hello" 01:30:22 5 01:31:06 Well, that takes care of that. 01:32:52 -!- adu has quit (Quit: adu). 01:38:44 egelbot: getenv "EGEL_PS0" 01:39:21 egelbot: getenv "PATH" 01:39:21 "/home/marco/.local/bin:/home/marco/bin:/usr/share/Modules/bin:/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin" 01:39:51 Weird path. 01:39:58 Ah well. 01:44:52 spruit11: nice, that was quick 01:45:16 Yah, it was painless. 01:47:42 egelbot: 3+3 01:47:42 6 01:47:54 egelbot: getenv "HOME" 01:47:54 "/home/marco" 01:48:00 egelbot: getenv "DISPLAY" 01:48:00 ":0" 01:48:09 Hey! 01:48:09 egelbot: getenv "LANG" 01:48:09 "en_US.UTF-8" 01:48:18 egelbot: getenv "LC_ALL" 01:48:45 egelbot: {getenv "USER", getenv "TERM", getenv "LC_ALL"} 01:48:45 (System:cons "marco" (System:cons "xterm-256color" (System:cons System:nop System:nil))) 01:49:36 egelbot: (getenv "_",) 01:49:36 internal:1:14:syntactical:primary expression expected 01:49:43 egelbot: tuple (getenv "_") 01:49:43 (System:tuple "./egel-bot") 01:50:43 egelbot: using Math;; sin 1.0 01:50:43 0.8414709848078965 01:51:55 egelbot: data foo;; let bar = [X -> 2];; foo;; bar foo 01:51:55 internal:1:31:syntactical:in expected 01:52:02 def 01:52:10 egelbot: data foo;; var bar = [X -> 2];; foo;; bar foo 01:52:10 internal:1:21:syntactical:= unexpected 01:52:16 egelbot: data foo;; def bar = [X -> 2];; foo;; bar foo 01:52:17 foo 01:52:17 2 01:52:37 egelbot: data foo;; def bar = [X -> 2];; bar foo;; quux 01:52:37 2 01:52:37 internal:1:44:semantical:undeclared quux 01:52:52 egelbot: data foo;; def bar = [X -> 2];; quux;; bar 01:52:52 internal:1:34:semantical:undeclared quux 01:53:01 egelbot: bar 01:53:01 bar 01:53:04 egelbot: bar foo 01:53:04 2 01:53:13 right 01:53:49 egelbot: val id = [ X -> X ];; val one = 1;; id one 01:53:49 1 01:54:00 Okay, that works too. 01:54:28 egelbot: data foo;; def foo0 = foo;; data foo;; def isfoo = [foo -> true | _ -> false];; (foo0, foo, isfoo foo0, isfoo foo) 01:54:28 (System:tuple foo foo System:true System:true) 01:54:36 hmm 01:54:44 egelbot: data foo;; def foo0 = foo;; 01:54:44 internal:1:30:syntactical:primary expression expected 01:54:49 egelbot: data foo;; def foo0 = foo;; 1 01:54:49 1 01:55:04 egelbot: data foo;; 2 01:55:04 2 01:55:10 egelbot: def isfoo = [foo -> true | _ -> false];; (foo0, foo, isfoo foo0, isfoo foo) 01:55:11 (System:tuple foo foo System:true System:true) 01:55:14 still true 01:55:26 egelbot: def foo = 5; data foo;; 2 01:55:26 internal:1:15:syntactical:primary expression expected 01:55:31 egelbot: def foo = 5;; data foo;; 2 01:55:31 2 01:55:34 egelbot: def isfoo = [foo -> true | _ -> false];; (foo0, foo, isfoo foo0, isfoo foo) 01:55:34 (System:tuple foo foo System:true System:true) 01:55:38 egelbot: def foo = 5;; 01:55:38 internal:1:16:syntactical:primary expression expected 01:55:42 egelbot: def foo = 5;; foo 01:55:42 5 01:55:45 egelbot: def isfoo = [foo -> true | _ -> false];; (foo0, foo, isfoo foo0, isfoo foo) 01:55:45 (System:tuple 5 5 System:false System:false) 01:55:48 what 01:56:12 foo = 5 01:56:18 yes 01:56:38 So isfoo reduces to false. 01:56:40 but I thought foo0 still has the binding to the old foo, which is a different foo even if its name is the same 01:56:48 No. 01:57:11 oh well 01:57:16 One global table for definitions you can interactively modify. 01:57:30 and anything lowercase refers to that? 01:57:31 ok 01:57:45 I mean they're resolved late, even when they're already in a definition? 01:57:53 egelbot: def isfoo = [foo -> true | _ -> false];; 01:57:53 internal:1:43:syntactical:primary expression expected 01:57:56 Yah. 01:58:00 egelbot: data foo 01:58:07 egelbot: def isfoo = [foo -> true | _ -> false] 01:58:11 egelbot: isfoo 01:58:11 isfoo 01:58:16 egelbot: def Isfoo = [foo -> true | _ -> false] 01:58:16 internal:1:2:syntactical:combinator or operator expected 01:58:30 egelbot: data foo 01:58:34 egelbot: isfoo foo 01:58:34 System:true 01:58:54 so isfoo has a definition that refers to foo by name and dynamically resolves it 01:58:57 ok 01:59:09 Combinators lowercase, variables uppercase. It was the only thing which I thought made sense in an untyped environment. 01:59:47 yeah, it's like prolog I guess 01:59:56 The problem is [ x -> x ], I need/want to know whether x is a variable of combinator. 02:00:05 except for, you know, the implicit unevaluated function thing 02:00:16 Yah, Prolog chose the solution. 02:00:46 Well, Prolog is another operational model. This is just a broken down FP. 02:01:09 this is not the syntax I like too much, but it works and is consistent 02:01:09 *the same solution. 02:02:00 Yah. But `[x -> x]` right? In a typed environment you can check whether it binds correctly to a combinator or otherwise should bind to a variable. 02:03:29 I would suggest no case difference, in a pattern if a name stands alone rather than as the head of an application then it's a freshly bound variable, there's some way to disambiguate it to a constructor (with an empty parenthesis when you use parenthesis function calls, or with a namespace prefix), and a few important nullary constructors (eg. true, false, just, unit) are keywords 02:03:37 Making the distinction up front means you can visually see the difference, and simplifies the interpreter (which otherwise would need heuristics). 02:04:09 but it might not work well in egel because application doesn't use parens, so you can't use an empty paren 02:04:21 No. But Egel is untyped. You can have a variable in the head. 02:04:44 egelbot: [ (X Y Z) -> Z Y X ] {1} 02:04:44 (System:nil 1 System:cons) 02:05:24 And for orthogonality I want to keep that. 02:05:37 and constructors are not namespaced under enum typenames, so you don't have a convenient namespace syntax that you can use to name any constructor already 02:05:52 spruit11: yeah, egel does that unevaluated thing, so this wouldn't work 02:07:13 No. It's weird stemming from Haskell/OCaml but if you see Egel as an LC+constants it's the only thing which makes sense. 02:07:22 (lambda calculus) 02:07:34 [[User:Hakerh400]] https://esolangs.org/w/index.php?diff=77115&oldid=76891 * Hakerh400 * (+24) 02:08:43 [[User:Hakerh400]] https://esolangs.org/w/index.php?diff=77116&oldid=77115 * Hakerh400 * (-20) 02:09:44 Right, in an untyped LC+constants `1 2` is a term and you want to be able to pattern match on that. 02:10:29 spruit11: no, only if you allow unevaluted applications 02:10:45 I don't see that? 02:10:48 normally `1 2` would try to call the function `1` which is an error because it's not a function 02:11:11 but in spruit you allow any application and they just default to unevaluated 02:11:16 It's normally allowed by the grammar... 02:11:25 it's not the grammar 02:11:45 you could write (1 2) in the code, and it wouldn't even give an error until you try to evaluate it 02:11:48 T := V | C | T T | \V. T 02:12:05 but if you want to pattern match to something, then it has to try to evaluate it, fully in an eager language, or to top level constructor in a lazy language 02:12:08 like haskell 02:12:22 Okay. Then I just difer on what I consider an untyped LC+constants. 02:12:28 yeah 02:12:29 *differ 02:17:35 [[Special:Log/newusers]] create * R123124 * New user account 02:17:44 -!- Heavpoot has quit (Remote host closed the connection). 02:21:06 [[Esolang:Introduce yourself]] https://esolangs.org/w/index.php?diff=77117&oldid=76916 * R123124 * (+108) /* Introductions */ 03:32:30 [[Special:Log/upload]] upload * RocketRace * uploaded "[[File:Linear transformation latex.gif]]" 03:34:18 [[1CP=1ICL]] https://esolangs.org/w/index.php?diff=77119&oldid=77114 * RocketRace * (+321) \ 03:36:39 -!- adu has joined. 03:37:01 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77120&oldid=77119 * RocketRace * (+26) /* Language specifications */ 03:51:46 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77121&oldid=77120 * RocketRace * (+222) 03:52:37 [[1CP=1ICL]] M https://esolangs.org/w/index.php?diff=77122&oldid=77121 * RocketRace * (+12) /* Language specifications */ 03:57:33 [[Babalang]] https://esolangs.org/w/index.php?diff=77123&oldid=76378 * RocketRace * (+0) Remove from 2020 category and put into 2019 category 03:58:00 [[Babalang]] M https://esolangs.org/w/index.php?diff=77124&oldid=77123 * RocketRace * (+0) Undo revision 77123 by [[Special:Contributions/RocketRace|RocketRace]] ([[User talk:RocketRace|talk]]) 04:06:46 [[Consequential]] https://esolangs.org/w/index.php?diff=77125&oldid=74207 * TheCoderPro * (+5) /* Instructions */ 04:24:05 -!- adu has quit (Quit: adu). 04:24:42 -!- adu has joined. 04:45:09 [[,,,]] https://esolangs.org/w/index.php?diff=77126&oldid=77071 * TwilightSparkle * (+27) /* Syntax */ 04:46:01 -!- Lord_of_Life has quit (Ping timeout: 258 seconds). 04:46:49 -!- Lord_of_Life has joined. 04:48:11 -!- Frater_EST has joined. 04:55:46 -!- Matthias29 has joined. 04:56:00 hey 04:56:37 I'm writing the first implementations of .5d5dbfwmttwmtt and then compiling using .5dbfwmtt 05:01:57 man, it takes a lot of hackery to write a reasonably fast interpreter. 05:04:21 -!- Frater_EST has left. 06:05:26 -!- adu has quit (Ping timeout: 240 seconds). 06:09:22 -!- adu has joined. 07:04:44 -!- adu has quit (Quit: adu). 08:01:14 -!- kritixilithos has joined. 08:10:12 -!- hendursa1 has joined. 08:10:20 [[,,,]] M https://esolangs.org/w/index.php?diff=77127&oldid=77126 * SunnyMoon * (-22) Grammar fix! 08:12:23 -!- hendursaga has quit (Ping timeout: 240 seconds). 08:15:38 [[,,,]] M https://esolangs.org/w/index.php?diff=77128&oldid=77127 * SunnyMoon * (-1) Oh, the one is not needed. 08:20:22 [[,,,]] M https://esolangs.org/w/index.php?diff=77129&oldid=77128 * SunnyMoon * (-10) There is hence twice on a sentence. 08:49:29 -!- Sgeo has quit (Read error: Connection reset by peer). 08:52:56 -!- imode has quit (Ping timeout: 240 seconds). 09:45:59 -!- Matthias29 has quit (Remote host closed the connection). 10:05:58 -!- iovoid has quit (Quit: iovoid has quit!). 10:06:31 -!- FraterEST has joined. 10:07:03 -!- iovoid has joined. 10:07:41 -!- FraterEST has quit (Read error: Connection reset by peer). 10:09:38 -!- clog has quit (Ping timeout: 260 seconds). 10:12:00 -!- glowcoil has quit (Ping timeout: 244 seconds). 10:12:31 -!- ^[_ has quit (Ping timeout: 244 seconds). 10:12:53 -!- dnm has quit (Read error: Connection reset by peer). 10:13:53 -!- dnm has joined. 10:13:55 -!- ^[_ has joined. 10:22:43 -!- haavard has quit (Ping timeout: 260 seconds). 10:26:07 -!- haavard has joined. 10:44:41 -!- glowcoil has joined. 11:02:56 -!- hendursa1 has quit (Remote host closed the connection). 11:03:50 -!- hendursa1 has joined. 11:07:12 -!- glowcoil has quit (Ping timeout: 244 seconds). 11:08:42 -!- glowcoil has joined. 11:30:21 egelbot: 1;; raise "oops";; 2 11:30:21 1 11:30:21 internal:1:6:semantical:undeclared raise 11:30:38 egelbot: def raisepls = raise "oops" 11:30:38 internal:1:17:semantical:undeclared raise 11:30:48 egelbot: def raisepls = (raise "oops") 11:30:48 internal:1:18:semantical:undeclared raise 11:30:56 egelbot: def raisepls = (0; raise "oops"; 0) 11:30:56 internal:1:21:semantical:undeclared raise 11:30:59 hmm 11:31:15 egelbot: (print 3; raise "oops"; print 4) 11:31:15 internal:1:3:semantical:undeclared print 11:53:54 throw 11:54:12 egelbot: (print 3; throw "oops"; print 4) 11:54:13 internal:1:3:semantical:undeclared print 11:54:28 egelbot: (say 3; throw "oops"; say 4) 11:54:28 3 11:54:29 4 11:54:29 exception(System:nop) 11:54:36 egelbot: Io.print "hello, world" 11:54:36 internal:1:2:semantical:undeclared Io 11:54:55 IO prints to stdout and isn't included in the bot. 11:55:16 The bot uses say. I guess I could change that? 11:55:32 ah 11:55:37 egelbot: say "hello, world" 11:55:38 hello, world 11:55:48 wasn't there a declaration to load any module though? 11:56:17 Yah. If you know the path on my system you can make the bot print to my stdout. 11:56:26 egelbot: (say "hello, world"; raise "oops"; say "hello, world") 11:56:26 internal:1:23:semantical:undeclared raise 11:56:38 import "path_local/io.ego" 11:56:51 egelbot: (env "PWD", env "_") /* those paths? it might be able to guess. */ 11:56:51 internal:1:37:syntactical:? unexpected 11:57:01 egelbot: (getenv "PWD", getenv "_") /* those paths? it might be able to guess. */ 11:57:01 internal:1:43:syntactical:? unexpected 11:57:04 hmm 11:57:11 Use # for comments 11:57:33 I had C style syntax but I changed. Made it all much better. 11:57:52 egelbot: (getenv "PWD", getenv "HOME", getenv "_") 11:57:52 (System:tuple "/home/marco/Programming/egel-bot/src" "/home/marco" "./egel-bot") 11:57:54 Also #!/usr/bin/env egel now works for scripting. 11:57:57 right, so those 11:58:02 yay! 11:58:15 does that include mixing declarations and expressions? 11:58:38 egelbot: [_ -> say "hello, world"; raise "oops"; say "hello, world"] 0 11:58:38 internal:1:28:semantical:undeclared raise 11:58:42 what... still? 11:58:45 throw 11:58:49 oh 11:58:50 egelbot: throw 1 11:58:50 exception(1) 11:59:02 egelbot: 3; throw 4; 5 11:59:02 exception(5) 11:59:11 egelbot: (13; throw 14; 15) 11:59:11 exception(15) 11:59:12 Oops. 11:59:22 You found a bug! 11:59:26 egelbot: 3;; throw 4;; 5 11:59:26 3 11:59:26 exception(4) 11:59:26 5 11:59:32 that's not a bug, it's just creative precedence 11:59:50 it interprets that as throw(4;5) 11:59:58 egelbot: 3; (throw 4); 5 11:59:58 exception(4) 12:00:04 Oh. Right. Phew. 12:00:11 egelbot: ((say 3); (throw 4); (say 5)) 12:00:11 3 12:00:12 exception(4) 12:00:16 see 12:00:29 So I guess I should take a look at the syntax. 12:00:31 egelbot: def raisepls: throw "oops" 12:00:31 internal:1:14:syntactical:= expected 12:00:48 egel had strange precedence already when I last looked 12:00:49 IIRC 12:00:56 Moving say to print makes sense. There's no reason it should be say in the bot. 12:01:17 spruit11: you may want both print and println though 12:01:21 for convenience 12:01:32 Yah, the operator table doesn't allow for equal precedencies. 12:01:41 But that's for way later. 12:02:03 For now, I'll change say to print and do your pretty printing hack. 12:02:04 egelbot: def raisepls = (throw "oops") 12:02:20 egelbot: ((say 3); raisepls; (say 5)) 12:02:20 3 12:02:21 exception("oops") 12:02:41 egelbot: let Raisepls = (throw "Oops") in ((say 3); Raisepls; (say 5)) 12:02:41 exception("Oops") 12:03:02 egelbot: def saypls = "please" 12:03:09 egelbot: def raise = [ X -> throw X ] 12:03:15 egelbot: raise 6 12:03:15 exception(6) 12:03:31 egelbot: (say 3; 4) + (saypls; 5) 12:03:31 3 12:03:31 9 12:03:41 egelbot: def saypls = say "please" 12:03:45 egelbot: (say 3; 4) + (saypls; 6) 12:03:45 please 12:03:46 3 12:03:46 10 12:03:50 hmm 12:03:54 creative evaluation order too 12:04:11 I know raise is popular in academia but I rather stick to the now more popular throw. 12:04:22 Yah, it's left to right. 12:04:24 throw is fine, I just programmed too much python 12:04:31 Uh, right to left. 12:05:28 It's a bit weird but also more orthogonal I feel. `f x y` goes from right to left. 12:06:07 anyway, this means you can put just one word instead of two in the [ ... | _ _ -> raisebadargs] trailer in function definitions 12:06:53 well, the evaluation order is noticable only when you use impure functions like say, and there are ways to force it in that case, so it doesn't matter too much 12:07:10 egelbot: (say 3; say 4) 12:07:10 3 12:07:11 4 12:07:23 I was thinking on throwing exceptions now on everything, including unevaluated combinators. But I still think I'll leave it up to the programmer. 12:07:58 There has to be some weirdity! 12:08:00 ok. not great for your bot that doesn't tolerate infinite loops, but doesn't matter as much for HackEso which will just time out in like a minute 12:08:53 (it's not really a minute, but something on that order of magnitude, don't depend on the exact timeout value) 12:09:14 Great. I'll hack on pretty printing today, I think. 12:09:40 Does the build script work for you? 12:10:59 And install script? 12:11:18 I guess I should add a 'local' option to install once. 12:15:09 -!- egelbot has quit (Remote host closed the connection). 12:15:11 spruit11: dunno, I think I wrote a simple custom install script for HackEso, because that was simpler than modifying your install script, but I used the Makefile with some arguments to set pathnames 12:15:58 -!- egelbot has joined. 12:16:00 but basically I just had to invoke a make target for each library and one for the interpreter, then copy files and install a wrapper script 12:16:39 egelbot: print 5 12:16:39 5 12:16:43 egelbot: agitprop 12:16:43 visit https://egel-lang.github.io/ 12:17:18 HackEso has somewhat unusual filesystem hierarchy for legacy reasons, so I needed custom paths anyway 12:18:34 It would be nicer if people could get just a local install by running `install local`, right? 12:18:53 admittedly those legacy reasons include my changes... 12:19:21 spruit11: sure, that's basically what I'm doing, but you need to be able to specify make parameters for the paths, like six of them 12:19:28 -!- hendursa1 has quit (Quit: hendursa1). 12:19:32 make that eight 12:19:40 I suck? 12:19:45 -!- hendursaga has joined. 12:19:46