gcd No one has to like the paperclip factory, but pretending the world is going back to loose leaf isn't a strategy. 00:01
00:04 Sgeo left
timo i'm not going to re-tread all the arguments that i've already spent way too many words on in that moarvm issue 00:13
00:24 Sgeo joined
timo meanwhile i'm going to keep pretending that the entire global banking system hasn't been annihilated and replaced with cryptocurrencies 00:27
00:46 [Coke] left 00:54 Sgeo left 01:10 Sgeo joined 01:50 kylese left 01:51 kylese joined 02:15 kylese left, kylese joined 02:42 [Coke] joined
samebchase Thanks for all your efforts timo 🙏🏽 03:12
05:40 abraxxa joined 06:12 Sgeo left 06:52 gfldex left 07:19 gfldex joined 07:48 dakkar joined 10:29 lizmat left, lizmat joined
disbot <aruniecrisps> @timo is there some way to contact you? I have an idea 10:38
lizmat aruniecrisps if you hop on IRC, you can privmsg with timo 10:39
disbot <melezhik.> So I am using my free groq credits to generate dsci pipeline for typical Raku layout ( including test coverage) - dsci.sparrowhub.io:8080/report/ui2/job.run/2078 @antononcube you may take interest in this / or anyony willing to play with dsci . openai/gpt-oss-120b is pretty accurate , besides some minor glitches , and it takes only 7 secs to generate ! 10:57
<melezhik.> And this may become a party of dsci offering out of the box 10:58
<melezhik.> Here is what I am asking for : “You are an expert DevOps assistant specialized in DSCI (Dead Simple CI). DSCI replaces complex CI/CD YAML with a tiny, high-level structural YAML file (jobs.yaml) and executes all actual logic using standard Bash or Raku in jobs (job.raku/job.bash) and tasks (task.bash/task.raku)” 10:59
<melezhik.> Generate a DSCI pipeline configuration for: project written on Rakulang - The source code of the project will be commited later. - Pipeline should test, build and push code as image to container registry - Test stage should run unit tests using prove6 cli - Prove6 should be installed before using zef install App::Prove6 command 11:00
<aruniecrisps> Thanks @lizmat 11:06
tbrowder timo, esp. thanks for fixing JSON::Fast, aa 12:12
as well as all the other things. mine sharing why you're leaving and where you're going? 12:13
[Coke] tbrowder: I believe it is primarily around github.com/MoarVM/MoarVM/issues/2015 12:18
tbrowder ah, yes, thnx 12:37
12:45 atcroft left
[Coke] folks - thoughts on how to do a soft pre-release appreciated. As we get closer to flipping the switch for RakuAST to be the default, we'll need a bunch of folks testing that aren't the core developers+Blin 12:49
If we can figure out a clean way to put a -rc-ast-1 version in the pipeline and get native builds for it, that'd be nice. (but we'd have to be careful it didn't get picked up as the latest version by folks) 12:51
13:01 Pixi` joined 13:04 Pixi left 13:08 atweedie left, eax_ left, pierrot left, atweedie joined 13:09 eax_ joined 13:14 pierrot joined
disbot <antononcube> Here is a brainstorming question: > What would be the rakubrew command for a "soft pre-release"? 13:41
<antononcube> To put it another way, can I specify to rakubrew do I want to see/download RakuAST distributions or not? 13:43
ugexe so long, and thanks for all the fish 14:34
[Coke] for rakubrew, we could have people do a build of a specific commit.
tbrowder woul 14:52
erg, i will try to help on my old laptop 14:53
but i'm not using rakubrew, just the old fasioned way 14:54
disbot <melezhik.> @antononcube now markdown generated by LLM is rendered - dsci.sparrowhub.io:8080/file_view/d.../answer.md for readability 15:26
ugexe no offense but it seems a bit tone deaf to mention all this LLM stuff today given that someone not even 12 hours ago posted here that they are leaving the raku project because of genai 15:31
timo don't worry, I'm adopting a "not my circus, not my monkeys" attitude and trying to just laugh it off 15:36
disbot <melezhik.> ugexe: dsci itself is written by me , without LLM. LLM is just an excellent tool to attract new users to any ( Raku ) related projects , whats wrong with that ? People could use LLM baked feature to bootstrap pipelines …
<melezhik.> I myself never consider LLM is evil if one uses it properly 15:37
15:39 [Coke] left
tbrowder amen 15:39
15:40 [Coke] joined, [Coke] left
tbrowder and test well 15:40
ugexe because no one really cares about the llm part
disbot <melezhik.> Btw if you please point me the ref to this 12 hours ago post ? Ooc
<melezhik.> Well that’s the point - people cares about results which LLM gives them 15:41
ugexe yes, so why lead with how much you spent on grok or whatever
disbot <melezhik.> Cause I use LLM to generate ( of good quality in my case ) code and probably going to make this LLM backed feature part of dsci , and hopefully this will attract more users and eventually new users to Raku , cause dsci ci engine is written on it 15:45
ugexe why do you think people care about the llm part of what you just said?
no one is seeing llm generated code and saying "oh it is interesting that llm generated that"
people are interested in results sure, but that is different than glorifying the method used to get there 15:46
disbot <melezhik.> Cause they just do 😊 cause LLM is the future whether we want it or not. But I am against bad use or f LLM like vibe coding and stuff
<melezhik.> And I am fire good use of LLM
<melezhik.> Fire -> for 15:47
ugexe i think you're missing my point
disbot <melezhik.> What’s wrong with method ? What’s wrong with saying - the method is good is the one provides the good ? 15:48
<vendethiel> sad day for the Raku project
ugexe do you think people would find it interesting if you mentioned "i used vscode to write this code"? 15:49
disbot <melezhik.> No. But LLM is not vscode
ugexe for the sake of the argument im making the comparison is apt. yes obviously thing A is not thing B 15:52
disbot <melezhik.> Ugexe: I would rather avoid the discussion , I can try to avoid LLM related messages here if this bothers you but I just don’t get why. There is tons of code I believe generated by LLM which is really bad and I am realistic about that. But also there is good ( at least for my purposes ) code which is produced by LLM as well
<melezhik.> And if it’s good I am just sharing . That’s it ) 15:54
15:56 [Coke] joined
tadzik I am now curious if there are cases where (advertised) LLM use attracts someone to a project. I've seen plenty of data points for the opposite, but it may be just my bubble 16:24
16:37 dakkar left
disbot <melezhik.> Oh, I just read about timo decision . Very sad 😔, by no means I want trigger any sort of disturbance here by referencing LLM in this chat . But fundamentally I personally do consider LLM is evil , it just need proper usage and wisdom 16:45
<antononcube> Thanks for bringing this up -- I missed it. 16:58
<melezhik.> Do consider -> do not consider. Sorry for the typo 17:01
<antononcube> Facilitating creation of LLM pipelines and making LLM / AI agents. Having LLMs at your fingertip in interactive sessions. 17:02
17:04 human-blip left 17:06 human-blip joined
disbot <antononcube> To repeat and exaggerate a point made by others -- to ignore LLMs and the use of LLMs in programming is simply stupid. 17:07
<antononcube> I also agree that the use LLMs can be turn off -- there are (at least) a dozen of Raku packages that are obviously written by AI-agents, and, because of that, I completely not interested trying them out or using them 17:09
<antononcube> Mostly, because I have doubts they are well designed. I.e. why spent time and efforts adopting some AI slop... 17:11
tadzik "simply stupid?" 17:20
aside of the tongue-in-cheek "stupid because if you ignore them then you won't notice that you're adopting slop", I'm curious what you mean by that 17:21
disbot <antononcube> I will give examples. 17:33
<melezhik.> ugexe: would I have known the news I’d be more careful in sharing this LLM stuff , sorry for that . Did not want to upset you 17:48
18:48 abraxxa left, abraxxa joined 18:52 abraxxa left, abraxxa joined
tbrowder antononcube i welcome your analysis of my latest module to see if i’m wasting my time: App::RakuDocToPDG 19:17
or your time 19:18
er PDF 19:28
disbot <antononcube> 🤔 19:38
20:54 abraxxa left 21:15 rantanplan left 21:17 snonux joined
tbrowder module is App::RakuDocToPDF (pardon my bad typing) 22:42
23:11 Sgeo joined
lucs I think I don't understand the ratchet attribute. 23:33
m: '4a2ab' ~~ /:r (\d) 'a' 'b'/; say ~$0; # Shouldn't this fail? 23:34
camelia 2
23:49 kjp left 23:56 vasko45355861 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!