[00:21] *** jjido left [01:08] *** wayland joined [01:08] *** wayland76 left [01:37] *** hulk joined [01:37] *** kylese left [01:51] *** hvxgr left [02:15] *** hulk left [02:15] *** kylese joined [02:30] *** nine left [02:35] *** nine joined [03:06] *** sibl joined [04:50] multi MAIN ($x where 'a' = 'a') { $x.say; }; multi MAIN ($x where 'b' = 'b') { $x.say; } [04:50] xinming, rakudo-moar 5789d349e: OUTPUT: «a␤» [04:51] with this example, Is there shorter way to remove not duplicate the $x where 'a' = 'a' part? [04:52] I mean the $x where 'a' = 'a', remove the duplicated 'a' here. [04:56] if 1 { say $x }; [04:56] or just say $x [04:56] vlad__: I mean the signature part [04:56] no clue [04:56] but someone will wake up who does have a clue [04:57] hang about [04:59] * vlad__ suddenly notices he's in #raku and not #perl [05:01] *** sibl left [05:21] *** sibl joined [05:35] *** hurufu joined [05:37] *** hurufu left [05:38] *** hurufu joined [05:44] is it still briged with Discord ? [05:44] 2026-07-20T03:51:02Z #raku sibl That has been mentioned as a problem. [06:02] *** sibl left [06:02] *** sibl joined [06:35] *** hwj joined [06:44] *** sibl left [06:44] *** sibl joined [06:55] m: multi MAIN ($x where 'a') { $x.say; }; multi MAIN ($x where 'b' ) { $x.say; } [06:55] rakudo-moar 6c5175469: OUTPUT: «Usage:␤ ␤ ␤» [06:57] xinming: Interesting! I'm not sure I understand what you're trying to achieve though. If we knew that, we might be able to help better. [07:06] *** Sgeo left [07:26] *** kylese left [07:31] *** kylese joined [07:59] *** dakkar joined [08:57] *** wayland left [09:28] *** eseyman left [10:06] *** sergot joined [10:12] *** sibl left [10:25] *** hurufu left [10:25] *** hurufu joined [10:32] *** sp1983 joined [10:32] https://news.ycombinator.com/item?id=49004510 [10:35] *** sp1983 left [10:58] *** manu_ joined [10:58] *** manu_ is now known as eseyman [12:40] *** hurufu left [13:06] *** sftp left [13:29] *** hurufu joined [14:09] Where clauses should fail loudly when you are using assign instead of equality [14:10] Though actually I’ve enjoyed using assignment in clever ways with where clauses. So maybe a worry is enough, as those are straightforward to suppress [14:23] i think that's not an assignment but the default value [14:33] *** acidsys left [14:34] m: multi MAIN ($x where "a"|"b" = "a") { $x.say; }; # xinming [14:34] rakudo-moar 6c5175469: OUTPUT: «a␤» [14:35] if you mean that you want "a" to be the default and also constraint, then you can't avoid duplication I think [14:36] m: constant a = "a"; multi MAIN ($x where a = a) { $x.say; }; # maybe this [14:36] rakudo-moar 6c5175469: OUTPUT: «a␤» [14:38] it doesn't entirely remove the duplication, but you can change the value in one spot [14:42] a subset would read better [14:42] where clauses on optional parameters have this problem of having to supply a default that matches the where clause :( [14:52] yeah, subset would definitely read better otherwise it's confusing syntax [14:54] usage generated is quite useless anyway [14:55] *** acidsys joined [14:59] *** Aedil left [15:01] otoh subset can be named and may show up in some usage contexts, e.g. `multi sub MAIN(SomeSubset :$x)` will show the subset name, and adding `#= ` comment will throw in default value [15:05] I wonder if usage could use some description for the subset instead of the type, in the case above it's always SomeSubset [15:05] *** Aedil joined [15:10] *** human-blip left [15:12] *** human-blip joined [15:34] *** hwj left [15:35] *** hwj joined [16:02] we can set a .WHY on the subset type can't we? that'd be sensible to output in the generated help / usage [16:31] *** dakkar left [16:46] *** Geth joined [16:53] possibly, I think currently it's always ^name: https://github.com/rakudo/rakudo/blob/main/src/core.c/Main.rakumod#L318 [17:01] that does seem to couple generated help output to .WHY in a way that wouldn't always make sense. for example the way i might describe a subset for a generated help may not be the same as i'd otherwise write to describe 'why'. describing the constraint in english feels more akin to describing 'what' [17:02] time to introduce .HUH [17:02] although being able to do like `#| A string with 'a'` inside of the main signature itself on the parameter might be fine [17:04] doing that doesn't already have an effect? [17:05] this is what I do and while it kind of works, it requires some typing [17:06] yeah i guess it does work [17:08] especially if you repeat have a parameter inside multiple MAINs [17:08] s/have/; [17:08] I had this idea to support proto MAIN to make this slightly easier [17:12] i tried to prototype a shared reusable signature thing that could flatten into another signature definition ala my $shared-sigs = |(...); MAIN(||$my-shared-sigs) { ... } but i gave up after hitting some compile time vs runtime impedance [17:28] <[Coke]> Users: how do you use install/use rakudo? How often do you upgrade? (if you commit to any of moarvm/nqp/rakudo, your vote doesn't count for these purposes :) [17:28] m: my constant XY = :(Int $x, Str $y); BEGIN say XY # probably something like this [17:28] rakudo-moar 6c5175469: OUTPUT: «(Int $x, Str $y)␤===SORRY!===␤QAST::Block with cuid 1 has not appeared␤» [17:28] in fact signature can't become constant too [17:30] [Coke]: builds georg who sits in a datacenter and builds every single commit, usually multiple times in a row, is an outlier adn shouldn't have been counted [17:31] <[Coke]> c: abcd 3.say [17:31] [Coke], ¦abcd: «Cannot find this revision (did you mean “v6.c”?)» [17:31] <[Coke]> c: 3.say [17:31] [Coke], I cannot recognize this command. See wiki for some examples: https://github.com/Raku/whateverable/wiki/Committable [17:31] <[Coke]> c: 467694501 3.say [17:31] [Coke], ¦4676945: «3␤» [17:31] <[Coke]> ^^ georg [17:33] yup, that and our CI [17:35] on the VMs where I use rakudo I only update when I need a new feature or bugfix [17:35] I'm pretty sure that's how it goes for anything [17:36] <[Coke]> I ask with the thought of reducing our release schedule from monthly. [17:36] <[Coke]> to 4 to 6 times a year, or ad hoc, or.. [17:44] the model of monthly releasing is not that popular, iirc java, golang and node release versions every 6 months [17:45] there are things like .net that update monthly, but there's been exceptions to this rule as well [17:45] so some patch releases were skipped [18:21] who is georg [18:21] Spiders Georg [18:22] https://en.wikipedia.org/wiki/Spiders_Georg actually has a wikipedia article! [18:23] okay I was very online in 2013 but this one missed me [18:23] it's a delightful little joke [19:08] *** elcaro left [19:27] *** hurufu left [20:10] *** hwj left [20:12] [Coke]: Could you tell me if any of these work for you? https://fco.github.io/Raku-Playground/webkit-3d-test.html [20:22] <[Coke]> SmokeMachine: b-d all render [20:22] <[Coke]> a does not (this on DDG on mac) [20:22] <[Coke]> in chrome a-e all work. [20:23] <[Coke]> firefox a-e work. [20:23] <[Coke]> oh, wait. one sec. [20:23] <[Coke]> yup. what I said [20:29] [Coke]: thanks! [20:29] *** jjido joined [20:41] [Coke]: Can you see if it's working now? https://fco.github.io/Raku-Playground [20:42] <[Coke]> /win 12 [21:12] <[Coke]> verified jdv will eventually come back, but that's a few months out. [21:13] <[Coke]> ww [21:16] [Coke]: please, let me know how it goes when you have time to test it... [21:54] <[Coke]> that URL works the same as before on #6 [21:54] <[Coke]> the rock tile is colored by no rock. [21:55] <[Coke]> (on DDG) [21:55] <[Coke]> and it shows up if I find just the right viewing angle [22:07] *** Sgeo joined [22:34] *** jjido left [23:01] *** human-blip left [23:05] *** wayland76 joined [23:10] Just a heads-up that one of my next projects (ie. something I'm probably going to release in the next month or two) is an addon called CommandObject that takes a different approach to self-documenting MAIN functions. [23:15] .tell [Coke]: As a user, I'm currently using rakubrew because my current projects require RakuAST and I need the bleeding edge. When I move past that stage (hopefully after the 6.e release, though possibly after macros, depending on where things go), my preferred options would be a) OS package, and b) as a secondary option, a docker container that has raku in it (would be handy for installing different versions of Raku without needing rakubrew). [23:15] wayland76, I'll pass your message to [Coke] [23:21] .tell [Coke]: And, while I tend to only update every, say, 1-3 months, I'm currently on latest, because there was a serious bug that was blocking me. Given the current rate of development (RakuAST fixes) I'd encourage us to stick to monthly releases until at least 6.e. [23:21] wayland76, I'll pass your message to [Coke]