[00:04] *** sjn left [00:07] leont: actually, rust could be awesome for jiting - its macros system could add a layer of type constraints for emitting jit code, since macros can be recursive and also typed [00:08] *** Manifest0 left [00:09] Yes, in the end you'd use some unsafe to cast allocated page into code, but before that, you could get a lot of checks before even emitting code [00:10] you could get them in C but definitely not with nice parsed-macro syntax e.g. `jit! { tagged_add r3, r1, r2; guard_type, r2, Int }` etc. [00:10] (without , after guard_type perhaps ;p) [00:12] not that I've tried to write it in rust, I theoretically know how it could work [00:12] and compile times could probably become real pain [01:02] *** sibl joined [01:17] *** topnep left [01:19] *** topnep joined [01:24] *** johnjay left [01:25] *** johnjay joined [01:26] *** johnjay left [01:28] *** johnjay joined [01:28] *** johnjay left [01:29] *** hulk joined [01:29] *** johnjay joined [01:29] *** kylese left [01:33] *** johnjay left [01:35] *** [Coke] left [01:35] *** johnjay joined [01:37] *** [Coke] joined [02:15] *** hulk left [02:15] *** kylese joined [02:46] *** sibl left [02:47] *** sibl joined [02:57] *** sibl left [02:58] *** sibl joined [03:23] *** topnep left [03:24] *** topnep joined [03:34] *** human-blip left [03:35] *** human-blip joined [03:40] *** sibl left [03:41] *** sibl joined [03:47] *** lichtkind_ joined [03:49] *** lichtkind__ left [04:24] *** johnjay left [04:28] *** johnjay joined [04:36] *** sibl left [04:57] *** johnjay left [04:58] *** johnjay joined [05:12] *** johnjay left [05:13] *** johnjay joined [05:17] *** johnjay left [05:21] *** johnjay joined [05:28] *** topnep left [05:30] *** topnep joined [05:37] *** wikipedia joined [05:38] raku is an improvement on perl [05:38] vlang is an improvement on golang [05:38] I wonder if raku has ideas that vlang can use [05:43] *** hurufu joined [06:16] *** belluzj joined [06:24] *** belluzj left [06:54] well Raku was “designed” by an experienced language author with a feel for natural language and gathered 361 RFPs [06:54] so, please dig in and tell us the answer 🙃 [06:57] I don't know raku [06:58] I asked so you might give ideas [07:17] *** abraxxa joined [07:20] *** abraxxa left [07:21] *** Sgeo left [07:21] *** abraxxa joined [07:34] *** topnep left [07:35] *** topnep joined [08:01] *** dakkar joined [08:23] raku is just so very different from vlang that it's hard to even begin [08:25] maybe stuff that would be useful in parsing stuff or something [08:26] probably grammars [08:26] is that some sort of syntax? [08:26] https://slangify.org/examples [08:27] there's whole page about it, but it's kinda like ebnf [08:28] other hugely useful things are concurrency using `react { whenever … }` [08:29] could it have advantages over goroutines? [08:32] yes, it's very different mental model [08:34] raku handles channels too and you can do `start react whenever $channel { .say }` [08:34] m: my $ch = Channel.new; my $task = start react whenever $ch { .say }; for ^5 { $ch.send($_); }; $ch.close; await $task [08:34] rakudo-moar f628473b8: OUTPUT: «0␤1␤2␤3␤4␤» [08:34] it's kinda goroutine-ish approach [08:34] m: react { whenever Supply.interval(.3) { say 1 }; whenever Supply.interval(.5) { 2.say }; whenever Promise.in(2) { done } }; say "react block has ended"; [08:34] rakudo-moar f628473b8: OUTPUT: «2␤1␤1␤2␤1␤1␤2␤1␤1␤2␤1␤2␤react block has ended␤» [08:35] I like unique designs based on actual use experience [08:35] Yes, the above is doable using channels, but it's just easier to understand [08:36] ah, vlang has select that does similar thing… [08:40] there are things that raku has in signatures and I'm not sure if vlang has - signature coercions, signature parameter constraints (and subsets) [08:40] m: sub Log(Num() $x where * > 0) { }; Log(1); Log(-1) # Log coerces whatever into floating point number and checks if it's greater than 0 [08:40] rakudo-moar f628473b8: OUTPUT: «Constraint type check failed in binding to parameter '$x'; expected anonymous constraint to be met but got Num (-1e0)␤ in sub Log at line 1␤ in block at line 1␤␤» [08:43] Voldenet: what do the things in signatures do? [08:44] add additional constraints that die when given thing does not satisfy the value [08:44] erm, the cosntraint [08:46] the above example has type coercions `Num()`, constraints `where * > 0` and WhateverCode syntax (instead of `$_ > 0` you can write `* > 0` in a lot of contexts) [08:46] e.g. in map: [08:46] m: say (^5).(* ** 2) [08:46] rakudo-moar f628473b8: OUTPUT: «No such method 'CALL-ME' for invocant of type 'Range'␤ in block at line 1␤␤» [08:46] m: say (^5).map(* ** 2) [08:46] rakudo-moar f628473b8: OUTPUT: «(0 1 4 9 16)␤» [08:53] also, hyperops (`say ^5 >>**>> 2` does the same as map example), metaops (`[+] ^5` means to apply sum on all elements of ^5), postfix function calling [08:54] so given `sub a($x) { };` `a(2)` and `2.&a` are equivalent [08:54] you can make anonymous blocks for postfix function call too [08:54] m: 2.&{ $_ + 3 }.say [08:54] rakudo-moar f628473b8: OUTPUT: «5␤» [08:55] that's the most of things I know that are useful to me [09:01] and would be probably implementable too - things like multiple dispatch or user-defined operators are not easy to have [09:01] or junctions [09:38] *** topnep left [09:40] *** topnep joined [09:48] *** hurufu left [11:01] *** sjn joined [11:29] *** wikipedia left [12:55] are we in a spot already where we can give source code instead of "expected anonymous constraint to be met"? [12:55] m: sub Log(Num() $x where 0 ^.. *) { }; Log(1); Log(-1) # could be easier to generate a good error for [12:55] rakudo-moar f628473b8: OUTPUT: «Constraint type check failed in binding to parameter '$x'; expected anonymous constraint to be met but got Num (-1e0)␤ in sub Log at line 1␤ in block at line 1␤␤» [13:04] don't we have "will complain" for that ? [13:07] where does "will complain" go, on subsets? [13:09] hhmmm... good question [13:09] oh am i missing a "use v" [13:09] use experimental :will-complain I think [13:10] ah, yes there it is [13:11] RAKUDO_RAKUAST=0 doesn't give me "not rakudo rakuast" btw, it's the same as RAKUDO_RAKUAST=1 [13:21] yeah... :-( [13:22] if nqp::getenvhash() [13:22] ah yeah that gives strings [13:22] if we try to intify / numify there we'd throw an exception if it's something like "yes" [13:26] could mix in a role that can give the source of a constraint by storing the beginning and end character index and then using $?SOURCE in the right spot [13:26] since we now store $?SOURCE in many cases [14:26] *** librasteve_ joined [15:21] *** abraxxa left [15:22] *** abraxxa joined [16:05] notable6: weekly [16:05] librasteve_, 3 notes: https://gist.github.com/bae426247368849096a89032952cbd6b [16:19] notable6: weekly reset [16:19] librasteve_, Moved existing notes to “weekly_2026-06-01T16:19:32Z” [16:31] *** dakkar left [16:51] *** wikipedia joined [17:19] *** johnjay left [17:20] *** johnjay joined [17:24] *** wikipedia left [17:25] https://rakudoweekly.blog/2026/06/01/2026-22-some-ast-fruit/ [17:26] *** johnjay left [17:27] *** johnjay joined [17:27] *** johnjay left [17:27] *** wikipedia joined [17:28] *** johnjay joined [17:28] *** johnjay left [17:29] *** johnjay joined [17:29] *** johnjay left [17:30] *** johnjay joined [17:50] *** belluzj joined [18:00] *** wikipedia left [18:19] *** wikipedia joined [18:51] *** belluzj left [19:11] *** belluzj joined [19:13] *** belluzj left [19:18] *** johnjay left [19:50] *** abraxxa left [20:05] weekly: https://github.com/Raku/problem-solving/issues/520 [20:05] lizmat, Noted! (weekly) [20:09] *** topnep left [20:12] *** topnep joined [21:02] *** acidsys left [21:28] think i finally figured out what is going on with the delete option [21:29] .tell guifa what timezone are you in? [21:29] tonyo, I'll pass your message to guifa [21:35] *** wikipedia left [21:40] *** acidsys joined [22:00] *** wikipedia joined [22:15] *** topnep left [22:17] *** topnep joined [22:18] *** Sgeo joined [22:44] *** wikipedia left [22:59] *** kylese left [23:24] *** Nobody joined