[00:31] *** lizmat_ joined [00:33] *** lizmat left [01:40] *** hulk joined [01:42] *** kylese left [01:55] *** librasteve_ left [02:05] *** topnep_ left [02:07] *** topnep joined [02:15] *** hulk left [02:17] *** kylese joined [09:15] *** Geth left [09:16] *** Geth joined [09:16] melezhik up again [09:16] 2026-07-09T07:01:53Z #raku lizmat: looks like irclog.raku is out of sync [09:30] *** rmv joined [09:30] *** rmv left [09:30] *** rmv joined [09:35] *** rmv left [10:06] *** abraxxa1 left [10:09] *** abraxxa joined [10:19] *** abraxxa left [10:19] *** abraxxa1 joined [10:24] *** acidsys left [10:28] *** acidsys joined [10:52] *** topnep left [10:52] *** topnep joined [10:55] *** lizmat left [10:55] *** lizmat_ left [10:56] *** lizmat joined [11:01] *** rmv joined [11:06] *** rmv left [11:21] *** guifa_ joined [11:58] hi, i thought my old debian bullseye wasn't going to get rakudo-pkg again but it was there this moring and installed fine. thanks to patrickb and crew [12:45] *** oodani left [12:46] *** oodani joined [12:53] *** rmv joined [12:53] *** rmv left [12:53] *** rmv joined [12:58] *** rmv left [13:06] *** wayland76 joined [13:07] *** wayland left [13:13] tbrowder: All thanks for rakudo-pkg go to El-Che. [13:19] URL? [13:47] <[Coke]> ah, bah, working on a refactor and ending up with circular modular dependencies. :| [14:02] *** nine left [14:02] *** nine joined [14:39] *** japhb left [14:42] *** japhb joined [14:55] melezhik.: `sudo cat /etc/sudoers` tells me `@includedir /etc/sudoers.d`, which is suboptimal :> [14:55] 2026-07-09T06:50:32Z #raku Voldenet: "and sudo has no `show me current full config`", you can just use `sudo cat /etc/sudoers` [15:02] *** topnep left [15:03] *** topnep joined [15:26] *** rmv joined [15:28] *** atweedie left [15:28] *** thatonelutenist left [15:28] *** atweedie joined [15:29] *** thatonelutenist_ joined [15:29] *** thatonelutenist_ is now known as thatonelutenist [15:31] *** rmv left [15:36] *** Eric22 joined [15:38] <[Coke]> Is there no way to use JSON::Fast to jsonify a custom object? [15:38] *** Eric22 left [15:38] is there any fool proof way to serialize raku objects? [15:38] <[Coke]> be nice if there was a method I could define that it would use. [15:39] <[Coke]> I don't expect it to do it without me writing any code. But it would be nice if I could put that code in the object and not have to call it explicitly. [15:39] you could use like JSON::Class but even something like that isn't going to be fool proof [15:57] hm, ideally a lib would separate serializing/deserializing from objects for separation of concerns [16:02] e.g. json.register-serializer(:type(X), :write(-> \j, X $x { j.start-object; j.property('x', $x.x); json.end-object; }, :read(-> \j { j.start-object; my $x = X.new; if(j.property('x')) { $x.x = j.int; }; json.end-object; }); [16:02] erm, should be `json.register-serializer(:type(X), :write(-> \j, X $x { j.start-object; j.property('x', $x.x); j.end-object; }, :read(-> \j { j.start-object; my $x = X.new; if(j.property('x')) { $x.x = j.int; }; j.end-object; });` [16:03] or [16:04] > json.register-serializer(:for(* ~~ X), :write(-> $_, X $x { .start-object; .property('x', $x.x); .end-object; }, :read(-> $_ { .start-object; my $x = X.new; if .property('x') { $x.x = .int; }; .end-object; }); [16:05] and afterwards, json.write(X.new(x => 42)); my $x = json.read(X, $str) [16:06] OR to-json and from-json using dynamic $*JSON variable for customization [16:10] it could be even very similar to JSON::Fast, but entry point would first attempt to find a sub to read/write and have tiny wrapper for interacting with json (without knowing if it's pretty printed etc) [16:11] just trivial `write property name, write str/bool/int value, write array, write object` [16:12] and symmetric api for `if property name, read str/bool/int value, expect array, expect object` [16:12] maybe with another `.read(X)` and `.write($x)` inside serializer [16:15] *** abbe_ joined [16:18] in fact, this could be built based on quite basic json interface – `my $j = json-write; $j.int(42); my $str = $j.end; my $j = json-read($str); my $i = $j.int; $j.end` (in write case, end/start are just writing tokens, in read case they expect that the token is there) [16:19] that probably wouldn't be as fast, but could be used as building block for anything more advanced [16:23] *** abbe left [16:23] *** abbe_ is now known as abbe [16:23] ah, right, reader could also have `is-array, is-int, is-string, is-bool, is-null, is-object` for probing next token type [16:24] in case someone would want to handle `["42"], [42], [{x:42}]` the same [16:26] *** vrurg_ joined [16:26] even better, `is-start-array, is-end-array, is-property, is-start-object, is-end-object` etc., of course [16:29] *** vrurg left [16:31] which would basically wrap `nqp::iseq_i($token,91)` etc. [16:44] El_Che: thnx for the rakudo-pkg for old debian bullseye! (and i'm finally in the process of upgrading to bookworm) [16:52] * Voldenet is watching [DKB] Katainaka no Ossan, Kensei ni Naru - S02E01 [------0-1--2--3-------0-1-] [ 10s / 23:33 ] [16:53] *** dakkar left [16:54] yes, totally bad channel, sorry [17:06] <[Coke]> for my particular small use case, will implement a 'method get-hash' which returns a Hash which is easily json-ifiable. [17:07] *** topnep left [17:08] I vaguely remember using that trick once or twice [17:08] <[Coke]> and then @array-of-stuff.map(*.get-hash). good enough for today. [17:09] *** topnep joined [17:13] weekly: https://dev.to/lizmat/governance-and-the-raku-foundation-1b6b [17:13] lizmat, Noted! (weekly) [17:17] *** abraxxa1 left [17:19] in case of JSON::Fast you can implement Associative [17:19] m: use JSON::Fast; class X is Associative { has $.x; method list { "x" => $!x } }; to-json(X.new(x => 42)).say [17:19] rakudo-moar f59a02a58: OUTPUT: «===SORRY!=== Error while compiling ␤Could not find JSON::Fast in:␤ /home/camelia/.raku␤ /home/camelia/rakudo-m-inst-1/share/perl6/site␤ /home/camelia/rakudo-m-inst-1/share/perl6/vendor␤ /home/camelia/rakudo-m-inst-1/share/pe…» [17:21] that's totally not what Associative should be implementing, but it works in this case [17:29] *** Sgeo joined [18:24] *** tadzik joined [19:01] *** rmv joined [19:01] *** rmv left [19:01] *** rmv joined [19:13] *** topnep left [19:15] *** topnep joined [20:03] *** librasteve_ joined [20:07] i'm a little late to the party - but I had to swap replace JSON::Class with JSON::Fast recently and ended up with a method get-hash {...} of sorts [20:07] (digs around) [20:08] https://www.irccloud.com/pastebin/bwr4Gr07 [20:09] I called it "action-hash" but ymmv [20:09] and then [20:10] https://www.irccloud.com/pastebin/KraCgR6S [20:10] *** leppard joined [20:34] *** vrurg_ left [20:49] yeah, that works but I wouldn't want to benchmark it [20:50] Voldenet: I've been very much enjoying the registration pattern tbh. [20:51] *** vrurg joined [22:32] Isn't the .raku method supposed to be object serialisation? [22:34] Or is the problem that that would introduce security concerns, and serialising as JSON is safer? [22:42] *** librasteve_ left [22:54] *** dmvrtx left [23:00] *** sjn left [23:01] *** sjn joined [23:05] *** dmvrtx joined [23:06] I don't think there's security concerns, just interoperability for serialization more than anything I'd guess