🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
00:00 reportable6 left, reportable6 joined 00:17 NemokoschKiwi left 01:43 frost47 joined 01:44 frost47 left 01:46 frost69 joined 01:48 frost69 left 01:50 frost31 joined 01:51 frost31 left 01:53 frost20 joined, frost20 is now known as frost 02:34 frost left 02:54 melezhik joined 02:57 melezhik left 02:59 frost joined 04:03 squashable6 left, squashable6 joined 06:00 reportable6 left 06:02 reportable6 joined 07:35 squashable6 left 07:36 squashable6 joined
[Tux] Rakudo v2022.12-1-gd52342eb0 (v6.d) on MoarVM 2022.12-19-g92aac35c6
csv-ip5xs0.854 - 0.966
csv-ip5xs-205.454 - 5.669
csv-parser3.724 - 3.851
csv-test-xs-200.402 - 0.410
test6.561 - 6.568
test-t1.492 - 1.567
test-t --race1.004 - 1.263
test-t-2020.497 - 21.945
test-t-20 --race6.486 - 6.532
08:52 sena_kun joined 09:01 sena_kun left, shareable6 left, tellable6 left, unicodable6 left, [Coke] left, tbrowder left, [Coke] joined, tellable6 joined, tbrowder joined, shareable6 joined, sena_kun joined
Geth rakudo/main: 4 commits pushed by (Stefan Seifert)++ 09:05
nine lizmat: that typo in the type didn't cause anything to blow up because we don't actually check those types 09:08
lizmat: why did you remove Attaching from subset? It was good as it was and in fact that was something I explicitly pointed out reviewing. Now it's wrong again. 09:10
lizmat I thought Attaching was not needed because there was nothing to attach anymore ? 09:21
it was only used to attach the current package, but that was only needed in the BEGIN handling, and could be looked up there ? 09:22
perhaps I'm misunderstanding the use / need for Attaching ? 09:24
nine You never know when BEGIN is actually run and what the state of the resolver is at that point. It could be in a completely different package.
attach is the only time where you can really be sure where the resolver is at that point 09:25
lizmat but the "attach" wasn't attaching anything... so it seems to me it was being used because of the timing, not because of its intended use 09:26
Nemokosch Is there anything about architecture besides those two markdown files in the rakudo repo? 09:49
Geth rakudo/main: 5c66515d46 | (Elizabeth Mattijsen)++ | src/core.c/List.pm6
Make List.sum adhere to container semantics

Fixes github.com/rakudo/rakudo/issues/5205
09:56 unicodable6 joined
ab5tract lizmat: I'm worried about the changes you made to $!of in RakuAST::Type::Subset 10:02
lizmat ab5tract: explain :-)
ab5tract why did you stop the type checking?
lizmat I stopped the type checking ? 10:03
ab5tract '+ $visitor($!of) if $!of; # && !nqp::istype($!of, RakuAST::Declaration::External::Constant);'
lizmat hmmm.... I did do that, didn't I 10:04
ab5tract partly I'm worried because the roast tests for subsets cover too much at once.
I will have to get my tests into roast sooner rather than later
lizmat can you tell me why you put that test there? for that particular class / case ? 10:05
ab5tract it's explained in the coment above
I've got to run, will be back with more later
lizmat ok... will re-instate the check 10:06
ab5tract but basically, without this, `my subset Foo of Int; my subset Bar of Int` will fail
in fact, as it is, my subset Foo of Int should fail already with an IMPL-QAST-DECL missing
bbl 10:07
lizmat m: Q|my subset Foo of Int; my Foo $a = "bar"|.AST(:run)
camelia Type check failed in assignment to $a; expected Cool::Foo but got Str ("bar")
in method AST at SETTING::src/core.c/core_epilogue.pm6 line 1
in block <unit> at <tmp> line 1
lizmat m: Q|my subset Foo of Int; my Foo $a = 42|.AST(:run)
camelia ( no output )
lizmat looks to me it works ?
hmmm.. but it did get attached to the wrong package it seems :-( 10:09
Geth rakudo/main: 4663edbfe4 | (Elizabeth Mattijsen)++ | src/Raku/ast/type.rakumod
RakuAST: re-instate proper order for resolving

And other repairs to fix the damage done by yours truly in b915dd3a. The subset now gets attached to the proper package again.
  nine++ ab5tract++
10:32 sena_kun left
lizmat anyways, now at 682 / 1355 10:38
t/spec/S12-class/instantiate.t now passing 10:39
nine lizmat: "attach" is just a callback for when a node gets attached to the AST. Lots of nodes already just use it to take not of where they are attached to. 11:00
11:08 sena_kun joined 11:26 lizmat_ joined 11:30 samebchase5 joined 11:31 bartolin_ joined, sjn_ joined 11:34 lizmat_ left, lizmat_ joined 11:36 Geth left, lizmat left, bartolin left, samebchase left, sjn left, Geth joined, samebchase5 is now known as samebchase, lizmat_ left 11:37 lizmat joined, RakuIRCLogger left 11:39 RakuIRCLogger joined 11:40 kawaii_ is now known as kawaii 12:00 reportable6 left 12:01 reportable6 joined
lizmat notable6: weekly 12:19
notable6 lizmat, 1 note: 2023-02-12T21:03:08Z <vrurg>: weekly youtu.be/sJs6jNroRXU
12:21 sena_kun left, sena_kun joined
lizmat notable6: weekly reset 12:24
notable6 lizmat, Moved existing notes to “weekly_2023-02-13T12:24:43Z”
nine Attribute initializers are on the one hand relatively simple to support. It's really just a paragraph of code and a few incidental minor fixes. On the other hand they may become a well that shakes the very foundation of the republic^W^WBEGIN time evaluation. Yes, my old friend again. 12:49
lizmat is glad she didn't get into that well 12:54
nine In case anyone's curious. My problem is that if IMPL-BEGIN-TIME-EVALUATE is given a Code object, it will run it. But now there's a case where we need a Code object to just evaluate to its meta-object instead. 12:57
ab5tract oof 12:58
Nemokosch why does the same thing try to solve two problems? 13:00
ab5tract okay, I feel like I'm going crazy.. This token definition *should* match the string 'enum', correct? gist.github.com/ab5tract/a894cb156...b9b5d197ba 13:07
as in 'enum', nothing but 'enum' 13:08
I don't remember the specifics of the journey to get subset parsing, but I don't remember it being so difficult 13:11
nine It usually isn't. Yes, that token looks fine. What's happening? 13:12
ab5tract Thanks to your recent fix, I get: 'Undeclared routine'
emum used at line -1. Did you mean ''? 13:13
I need to increase the font size in my terminal, clearly
still happens with ./rakudo-m -Ilib -e 'enum' 13:14
enum used at line -1. Did you mean ''?
nine LOL emum :) 13:22
ab5tract :)
I wish fixing the typo had solved the issue 13:23
nine Try it again with more of that token put back in. It might be that your token simply loses to name parsing because it doesn't contain anything more than that
ab5tract fair enough!
nine Remember: we do longest token matching. Your token should match more than others :)
ab5tract excellent point 13:25
damn you, "emum"! 13:26
I lost a surprising amount of time to that typo
nine That's why it's always good to not just post the problem, but also what exactly you did to reproduce it here. 13:27
ab5tract sometimes I feel like I'm chatting in the channel enough already but your point makes a lot of sense and I will take it to heart 13:30
Geth rakudo/main: b010e70b50 | (Elizabeth Mattijsen)++ | 2 files
Add first subset tests

More to come later!
nine We're all very experienced rubber ducks :) 13:31
lizmat :-) 13:32
nine: re 035544f28dfb570cb97 wouldn't it make more sense to create a ApplyPostfix.as-method-call(ArgList), which would take the first element of the given ArgList as the invocant 13:48
and not change ArgList at all ?
13:48 linkable6 left 13:49 linkable6 joined
lizmat ah, but then it wouldn't know how to codegen sym<name> 13:49
hmmm.. perhaps a subclass of ArgList: InvocantArgList ? 13:50
nine You don't like the current implementation? 14:16
[Coke] the nice thing about rubber ducking is that for something like that, the person doesn't even have to know how to fix the thing you were working on! 14:18
nine Sometimes the person doesn't even have to actually be there :)
lizmat: but yes, InvocantArgList would have a tiny performance advantage. 14:19
[Coke] I am very happy and very frustrated when that works. :)
lizmat nine: not talking about performance, but developer interface really 14:20
14:28 codesections joined
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2023/02/13/2023-...ore-class/ 14:42
vrurg lizmat: correction for the Weekly: it's only the second part of the class. The first one is lost due to poor video quality. 14:43
Ah, "second"... I'm still sleeping... :) 14:44
Geth rakudo/main: 615b9bf929 | (Elizabeth Mattijsen)++ | src/core.c/List.pm6
List.head should not decontainerize

Fixed ValueMap from the latest Blin run
15:27 sjn_ is now known as sjn 15:29 Xliff joined
Xliff P6-GLib suite timing statistics for Raku 2022.12.1192.g.2.b.6.b.4019.d 15:50
Total number of projects: 33 (502818 loc)
Total non-parallel compile times: 8869.070 (268.760 avg)
Total parallel compile times: 1435.714 (43.506 avg) 6.177x speedup
Note: Great compile times! Much better than the last 2 weeks! ;)
lizmat FWIW, I don't think anybody's done anything in that direction, so *I* have no explanation for what you're seeing / have seen 15:51
Xliff lizmat: I can understand your not understanding.
lizmat but I would *like* to understand ! 15:52
Xliff I'm almost certain it's environment.
I'm having open file issues and I can't seem to nail them down.
I will let you know if I ever lock it down. 15:53
lizmat please!
bisectable6 old=2022.12 say Date.today.Int 16:50
bisectable6: old=2022.12 say Date.today.Int
bisectable6 lizmat, Bisecting by exit code (old=2022.12 new=615b9bf). Old exit code: 1
lizmat, (2023-01-09) github.com/rakudo/rakudo/commit/72...46c378d01e 16:51
lizmat, bisect log: gist.github.com/6f04016638c8c9134e...c1016e2bc3
Xliff lismat: Are you still working on the weekly? 16:52
s/lismat/lizmat/ 16:53
lizmat nope, published already
2.5 hours ago or so
Geth Proxee/main: 06a6d325cf | (Elizabeth Mattijsen)++ | 6 files
Xliff lizmat=+ 16:59
What's wrong with my keyboard skills today?!? o_0 17:00
lizmat mp ofrs
Nemokosch the problem is that = and + are close to each other in the first place 17:11
Geth Proxee/main: fc7226bc5a | (Elizabeth Mattijsen)++ | 4 files
Fix tests
Proxee/main: 4f1c5f23c1 | (Elizabeth Mattijsen)++ | Changes
17:56 ab5tract left 18:00 reportable6 left 18:03 reportable6 joined 18:09 sena_kun left 18:43 sena_kun joined 19:13 melezhik joined 19:23 melezhik left
Geth Proxee: coke++ created pull request #2:
[Coke] I had to fork that, couldn't do it in place. 20:06
Geth Proxee/main: a823858b61 | (Will Coleda)++ | README.md
Proxee/main: e744dd2295 | (Márton Polgár)++ (committed using GitHub Web editor) | README.md
Merge pull request #2 from coke/main

Nemokosch good catch
22:00 frost left 22:08 ab5tract joined 22:59 sena_kun left 23:38 Xliff left