🦋 Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 6 September 2022.
00:43 guifa joined 00:47 guifa left 00:50 oodani left 00:51 oodani joined 01:37 kylese left, kylese joined 02:15 kylese left, kylese joined 04:04 _________ left 04:05 Aedil joined 04:12 _________ joined 05:30 Sgeo left 06:08 thaewrapt left 06:12 thaewrapt joined 06:17 thaewrapt left 06:28 thaewrapt joined 06:51 thaewrapt left 07:05 thaewrapt joined 07:11 thaewrapt left 07:22 thaewrapt joined 07:26 sena_kun joined 07:30 thaewrapt left
andinus mail 07:30
oops wrong pane
07:38 sena_kun left 07:41 thaewrapt joined 08:27 thaewrapt left 08:38 thaewrapt joined 08:53 thaewrapt left 08:56 MasterDuke left 09:04 thaewrapt joined
patrickb I have a role that needs some init logic. I can't put it in a `submethod TWEAK` as that collides in the does-ing class. Currently I rely on a `method TWEAK_SpanHandling()` that needs to be called in the mixing in classes TWEAK. Is there a way to do this that does not require cooperation of the does-ing class? 09:36
ab5tract patrickb: docs.raku.org/language/phasers#COM...plemented) 09:41
lizmat m: role A { sub TWEAK($) { dd; nextsame }; ::?CLASS.^find_method("TWEAK").wrap(&TWEAK) }; class B does A { method TWEAK() { dd } }; B.new 09:42
camelia sub TWEAK($)
method TWEAK(B $:: *%_)
ab5tract ah, nice
patrickb ab5tract: But I need this at object creation time, not compile time.
lizmat patrickb: I think my solution is what you're looking for ? 09:43
patrickb lizmat: That works. Thanks!
But wow, this is hacky..
ab5tract I imagine COMPOSE will be used often to mix in a run time candidate as above, but you raise a good point 09:44
This comes up fairly often, it would be great to explore some solutions
lizmat FWIW, I don't think a COMPOSE phaser is that necessary, as the mainline of the role is basically the COMPOSE phaser 09:47
ab5tract Fair, I think it was proposed as a way to create hooks like the ones that have been asked for twice in 48 hours alone. But just a hook or two in themselves would likely be plenty 09:56
librasteve just playing with the role / TWEAK conundrum - you could do this: 10:56
m: role Thingy { submethod TWEAK { say 'yo' } } class A is Thingy { submethod TWEAK { say 'ho' } } my $a = A.new; 10:57
Raku eval Exit code: 1 ===SORRY!=== Error while compiling /home/glot/main.raku Strange text after block (missing semicolon or comma?) at /home/glot/main.raku:1 ------> Thingy { submethod TWEAK { say 'yo' } }⏏ class A is Thingy { submethod TWEAK { s expecting any of: infix infix stopper statement end statement modifier statement modifier loop
evalable6__ (exit code 1) 4===SORRY!4=== Er…
librasteve, Full output: gist.github.com/2850c869c3272537a4...6ca051efb3
librasteve m: role Thingy { submethod TWEAK { say 'yo' } }; class A is Thingy { submethod TWEAK { say 'ho' } }; my $a = A.new;
Raku eval yo ho
evalable6__ yo
ho
10:58 guest555 joined
librasteve that is to go is rather than does in the user class and force the role to be punned...? 10:58
11:18 guest555 left 12:22 Chanakan left 13:26 jgaz joined
tbrowder .ask patrickb have you tried installing Raku on a Chromebook? 13:38
tellable6 tbrowder, I'll pass your message to patrickb
tbrowder or anyone else... 13:39
14:13 coleman left
patrickb no I haven't 14:15
tellable6 2024-08-19T13:38:46Z #raku <tbrowder> patrickb have you tried installing Raku on a Chromebook?
14:21 coleman joined 14:28 Chanakan joined 14:37 derpydoo joined
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2024/08/19/2024-...ing-ahead/ 14:51
15:10 derpydoo left
jgaz When will Raku Land stop listing up non-Zef modules? 16:32
16:40 sena_kun joined 16:50 lizmat left 16:51 lizmat joined 17:19 Xliff left
patrickb jgaz: Probably nevercor very far in the future. Since p6c is effectively frozen, there is no harm coming from it anymore. 18:00
18:19 Aedil left 18:24 Aedil joined
antononcube Just finished watching the presentation and Q&A session of "Large Language Models and The End of Programming - CS50 Tech Talk with Dr. Matt Welsh" (www.youtube.com/watch?v=JhCl-GeT4jw) My TL;DR review is: > Yes, yes, yes, and blah, blah, blah. 19:30
The presenter either lacks imagination or really likes portraying himself as such.
As for programming with LLMs -- I challenge him to show me how with LLMs he is going to produce working code for "multiscale Truchet patterns" -- say in Python, R, or JavaScript -- via cursory, human understanding of the this post: community.wolfram.com/groups/-/m/t/3247480
See these LLM-derived breakdowns and summaries of that presentation: github.com/antononcube/RakuForPred...t-Welsh.md
19:31 dawids_ joined 19:32 Aedil left
jaguart.v1 One thing I note on LLM generated code, with prompts are for Perl - the code is legacy Perl style - no modern idioms e.g. Corinna. Might differ for other languages I guess. 20:57
20:59 guifa joined 21:15 jgaz left 21:41 dawids_ left
antononcube Understanble -- LLMs are statistical in nature, hence larger-footprint outcomes are more probable. That is why people make complicated prompts, etc. (And make different claims about those... ) 21:46
21:47 sena_kun left
_grenzo So, you're saying, to protect our jobs we need to put more honey pots on github with code that suffers from the Halting problem, or other hard to spot bugs? 22:07
And use misleading comments...
Rule 1: We don't talk about AI fight club. 22:11
22:27 Sgeo joined
antononcube Or just invent new programming languages that immediately have huge mind-share. 22:46
_grenzo Have an LLM do that
23:01 MasterDuke joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!