01:41 hulk joined 01:42 kylese left 02:15 hulk left, kylese joined 03:00 Aedil joined
guifa timo: because it's actually 2024b :) 03:07
oh yeah also the word is, but that's actually present in the readme. Not worth an update though to fix those typos lol
but antononcube will be happy to know I'm venturing into the LLM (and other ML) territory for this upcoming project 03:10
03:43 floyza left 03:44 floyza joined
librasteve i’m a physicist, so there is only zero, one and many 04:32
nijmegenzuigt you're making me very phycisist 😠 04:47
05:49 tous joined, tous is now known as zapmoth 06:21 zapmoth left 06:32 Sgeo left 06:37 andinus2 left 06:38 andinus joined 07:37 sorenson left 07:45 floyza left 07:59 sorenson joined 08:17 ptc joined, ptc left 09:11 sena_kun joined 09:27 sena_kun left 09:28 sena_kun joined 09:30 xinming left 09:31 xinming joined 09:33 sena_kun left
antononcube Here is the “CopyEdit” prompt page : resources.wolframcloud.com/PromptR.../CopyEdit/ 10:03
That prompt is also in “LLM::Prompts”. 10:04
10:20 andinus left
ab5tract New comma beta released this morning. It removes the old SDK integration entirely, so you may be once again prompted for an SDK — but this should really be the last time! 10:55
Another highlight is that ctrl/ cmd clicking on an external package name should seamlessly load that package in a file — assuming that the distribution that provides it is installed. 10:57
Accessing the same via the External Libraries tree should also Just Work.
I’d appreciate some feedback on the external libraries mechanism. Right now you get a flat list of all provided packages from the distribution. Is this fine? Or should I recreate the directory structure of the package instead? 10:59
Or should we go with “why not both?” :)
12:28 sena_kun joined 12:33 wayland joined, wayland76 left 12:39 andinus joined
lizmat weekly: github.com/Raku/problem-solving/issues/445 13:01
notable6__ lizmat, Noted! (weekly)
13:08 Voldenet_ joined 13:09 Voldenet left, Voldenet_ is now known as Voldenet 13:33 dutchie left
antononcube FYI, Raku-Random-Spirographs-demo-1: 14:01
cdn.discordapp.com/attachments/633...72f26&
timo that's pretty 14:03
weekly: <antononcube> cdn.discordapp.com/attachments/633...72f26&
notable6__ timo, Noted! (weekly)
antononcube @timo Thanks! I was looking for a way to demo Raku's complex number calculations. Spirographs are a very fruitful domain. 14:11
timo love me some circles 14:13
antononcube @timo How about mandalas? 14:20
timo colored in lots of those on paper when i was a wee lad 14:22
antononcube I think this is how my mandalas work started -- someone asked in StackExchange for generation of mandala drawings for coloring. 14:28
14:35 gabiruh left 14:36 gabiruh joined 14:37 teatime left
timo they are also fun to make by hand with a compass 14:39
does anyone else find it annoying that compass is the english word for both a compass and a compass?
german has "zirkel" which is rather descriptive, but distinct from the word for circle which is "kreis"
antononcube Yeah, English is very polysemic. 14:40
timo also, "recorder" 14:41
antononcube > openai-playground --max-tokens=4096 list the top 20 polysemic words in English with brief meanings descriptions 1. Bank - can refer to a financial institution or the side of a river 2. Bat - can refer to a flying mammal or a piece of sports equipment 3. Bear - can refer to a large mammal or to tolerate or endure 4. Bill - can refer to a statement of money owed or a specific piece of legislation 5. Book - can 14:46
refer to a written work or to reserve a spot 6. Box - can refer to a container or to engage in a sport 7. Case - can refer to a legal proceeding or a container for holding items 8. Date - can refer to a day on the calendar or a romantic outing 9. Face - can refer to the front of a person's head or to confront a situation 10. Line - can refer to a row of text or a boundary 11. Match - can refer to a sports competition or a
corresponding item 12. Note - can refer to a written message or a musical tone 13. Plane - can refer to an aircraft or a flat surface 14. Ring - can refer to a piece of jewelry or a sound 15. Rock - can refer to a type of mineral or to move back and forth 16. Scale - can refer to a measuring device or to climb 17. Seal - can refer to a marine mammal or to close off 18. Spring - can refer to a season or a coiled piece of metal
19. Table - can refer to a piece of furniture or to postpone something 20. Watch - can refer to a timepiece or to observe something closely.
@timo You are making me think including js-d3-spirograph in "JavaScripdt::D3". I might make to do curve fill-ins too. 15:30
16:53 [Coke] joined
[Coke] . 16:53
nijmegenzuigt just a tiny one? 16:54
[Coke] trims that sentence down: "finds English annoying" 16:56
[Coke] hurls www.funtrivia.com/askft/Question41640.html 16:57
nijmegenzuigt go get a nosebeer 🙂
[Coke] I prefer to ignore the first sentence in the quote, but the second one seems pretty spot on 16:58
antononcube Wow! Very interesting propaganda message masking as linguistic quote. 🙂 17:01
nijmegenzuigt chatgpt 😠
antononcube Basically, when British Isles get invaded, or conquered, or whatever, then English gets "enriched." 17:04
nijmegenzuigt well from what gets said, the Brits do like the stuff 17:12
😛
antononcube They absolutely do like the "enrichment", there some relavant pie charts proving it. 17:14
nijmegenzuigt my friends who go there sometimes just hide the fact that theyre Dutch at all
:kek:
antononcube I know of several related jokes. 🙂 17:21
@nijmegenzuigt Are you, by any chance, using TOML configurations? 17:22
nijmegenzuigt never 17:24
prefer gura if I do have to make config files
antononcube Is there a Raku package handling Gura specs? 17:25
Meaning, I do not see such package in raku.land, but otherwise one might exist. (Or three.) 17:26
Gura seems to compare itself with TOML. 17:27
nijmegenzuigt Gura is good because it gets me out of the JSON/YAML/TOML discussion 🙂 17:30
but Gura itself obv has some funny stuff
antononcube I am looking for making a new grammar "from scratch" for a conference talk. I was thinking to use TOML, but I might use Gura instead. 17:31
nijmegenzuigt used it once with ZMQ and felt way better than JSON
tbh it's pretty clean 17:32
Gura is "potentially" slower but look ath ow much garbage JSON is being passed around either way
antononcube Well, it has an officially published BNF grammar, so that is a good reason to try it out.
The TOML parser is Raku seems to be using NQP. (I.e. is fast.) 17:33
17:34 Aedil left
@ab5tract Trying the new beta of Comma right now. So far so good. 17:36
But it is over a small project. 17:37
ab5tract nice! I hope it performs the same for a large one
nijmegenzuigt who (except ROckstar) parses that much for it to matter
ab5tract nijmegenzuigt: theoretically there should be no difference 17:38
antononcube What is really cool about having an up to date Comma plugin, is that I can use that plugin in CLion and get seamless Raku-and-C work environment for Raku packages that use NativeCall / C-implementations. 17:39
lizmat antononcube maybe you should blog about that :-) 17:40
antononcube @nijmegenzuigt Well, "LLM::DWIM" uses TOML, so, I decided to look closer into it.
@lizmat Yeah. Upcoming blog posts for sparse matrices. 17:41
nijmegenzuigt I went from Object Pascal to a Objective-C job and then C & C++ got banned from use for future projects
17:41 kaholaz joined
so D is kinda my go-to 😛 17:41
antononcube Right, use Swift, then.
nijmegenzuigt back in 2010
awwyeh
well Nim as ORC! 🙂 17:42
antononcube @lizmat I tried to come up with a few data structure designs that are "sparse arrays", not "just" sparse matrices. But that is a little too much thinking for a functionality that I have used no more than 3 times in the last 15-20 years. 17:45
I mean using multidimensional sparse arrays with dimensions 3 or (much) more.
@nijmegenzuigt I have heard but forgotten about Nim. 17:47
Thanks for mentioning it! (Not dropping Raku anytime soon for Nim, though.) 17:48
17:48 kaskal left 17:49 kaskal joined
nijmegenzuigt well Nim has a lot of interopt, can make Python modules or addons for blender and such 17:50
but Nim is in flux so who knows what it is like next year 17:51
antononcube Can it use Raku modules?
nijmegenzuigt granted as a staunch Wirthian I cant help but love it ofc
antononcube PASCAL programmers tend to know the value of everything and the Wirth of nothing. 17:52
Or similar... 17:53
nijmegenzuigt he's a nickles worth or smth
plenty of jokes along those lines 😛
antononcube There is should Nim localizations.
nijmegenzuigt but eh, 50 years later wheres modules for C++ eh? Pascal can compile its whole IDE, compiler and backend for a dozend architectures before LLVM is done with the first template while still being single-threaded 😎 17:55
antononcube Mm... right.
Nim makes me think of Eiffel, though. (Then again, Eiffel is inspired a lot from Ada.) 17:56
nijmegenzuigt ppl can talk about Raku's sigils,twigils, trigils and all that 17:57
but the sequence "unit module" is still the weirdest to me
antononcube Can be seen as a shortcut. 18:00
scullucs From "tasty_freeze" at Hacker news: Besides his contribution to language design, he authored one of the best puns ever. His last name is properly pronounced something like "Virt" but in the US everyone calls him by "Worth". That led him to quip, "In Europe I'm called by name, but in the US I'm called by value." 18:02
lizmat antononcube yeah, they'd have to be pretty sparse to make it worthwhile to handle them differently from ordinary arrays / matrices 18:09
[Coke] (virt) *groan* 18:13
nijmegenzuigt maybe some more ... 18:16
antononcube @lizmat I use very sparse matrices often (e.g. with fill-in 0.002%.) I am saying that have not used multidimensional sparse arrays or tensors that often. 20:07
20:54 dutchie joined
patrickb Hey! I need a hint for a problem where I don't know how to do best in Raku. I have an object with some private state. There is a role with some unimplemented methods I need to implement (used by some separate module). The methods will want to modify my private state. But I don't want to expose the roles methods publicly. 20:56
My first idea was to have an inner class implement the role. But inner classes can't access private attributes of their surrounding class.
timo can you "trusts" the inner class? 20:57
patrickb Yeah, I can do that. But trusts only allows accessing private methods, not attributes. It get's awkward quickly.
timo OK. you could have a private accessor method at least :D 20:58
patrickb Passing references to all the state I want to modify into my inner class is another approach, but I have to keep the references up to date. That also turns messy fast.
timo hmm. what happens if you have a sub inside the outer class that is called by the inner class, passing the instance. does that have access to private attributes?
patrickb I can pass the instance while constructing the inner object. But that object is still sealed. 20:59
timo right 21:00
antononcube @patrickb Can you put your state to be a Mao/Hash? Then you can transfer it however you want.
patrickb Mao?
antononcube No, Xi 21:01
patrickb Xi?
antononcube I meant: "Map/Hash" .
patrickb Ah :-P
That's an idea. But it feels really strange that moving over to using a Hash as data structure is the best solution. 21:02
I could have a State class with only public rw attributes and pass that to the inner class.
It's a bit sad that all modifications from then on are method calls, but maybe that's the price to pay... 21:03
antononcube You can have a special state class to transform into and from a Hash.
timo time to build Raku::InsideOut 21:04
patrickb Or I go with "**** it!", and just implement the role directly exposing the interface publicly.
antononcube There is a recent package for object (de-)serialiation to Hash. 21:05
See "Hash2Class" : raku.land/zef:lizmat/Hash2Class 21:06
nijmegenzuigt hmm virtualmin is Perl 21:07
is it a PITA to make plugins for with Raku? 😛 21:08
antononcube And concretemax is Django
nijmegenzuigt it's for a gateway/managament system for the prox mox ones 21:09
antononcube @nijmegenzuigt Haven't heard of "virtualmin" before: > The strength of Virtualmin lies in its minimalistic approach, utilizing the OS's provided software to keep your systems lean. Its non-intrusive nature and respect for your server configurations, along with the ease of migration from cPanel/Plesk and straightforward backup solutions, makes it indispensable.
nijmegenzuigt Django is waht Allura uses now but Im switching that
ah ye this is for dedicated servers to set up VMs and partially cloud 21:11
you know all that serverless blabla stuff ;p
antononcube @nijmegenzuigt I was joking, BTW, making a semantic inverse of "virtual-min" into "concrete-max."
nijmegenzuigt nice try chatgpt
antononcube Hmmm... there is a prompt for that, sure. 21:13
It is probably in the "Black Friday Prompts" repository.
21:44 sena_kun left
[Coke] stares blearily with dilated eyes at his high contrast IRC session 21:53
kaholaz Hello! Is it hard to get into Raku as a Perl developer? 22:00
22:01 wayland left
_grenzo @kaholaz No not really. Most concepts carry over. See: docs.raku.org/language/5to6-perlfunc 22:02
Then have a look at the Raku features that perl doesn't have 22:03
(docs.raku.org/introduction) 22:04
22:11 Sgeo joined
guifa How does the < twigil work? 22:18
timo you mean $<foo>? 22:22
[Coke] that really seems like it's not a twigil, not sure why it's listed there. I think it's referring to using $<submatch> and referring to a part of $/
timo it's just short for $/<foo>
[Coke] docs.raku.org/language/variables#Twigils does list it as a twigil. I think the <> is indexing there, not a twigil.
timo it's made to mirror the $<blop>=[ hello ] in regex syntax 22:23
[Coke] guifa - you seeing it anywhere other than that link I just posted? Might be a doc bug
guifa Coke: that's where I'm seeing it
I've never interpreted $<foo> as a twigil, interesting 22:24
timo the headings are all missing text
no, not all of them, but a bunch of them 22:25
[Coke] guifa: if you can, please open a bug at github.com/raku/doc pointing that out.
timo i think it's fine to show the < in that list, it's a spot where a user is likely to look when they see $<foo bar baz> in code and don't know what it means 22:26
[Coke] timo: if you could file a bug on that at github.com/raku/doc-website for that.
I think the "not really a variable" should maybe be "not really a twigil"
timo yes, a little bit more context wouldn't be bad 22:27
[Coke] (website for the non-rendered headings)
PRs welcome!
timo hm, where is issuable?
[Coke] ?
oh, a bot to open a bug? :) 22:28
timo something like that yeah 22:30
issuable, bug-a-boo, something like that. no, please not buggable.
bugbearer 22:31
m: no scrubs;
camelia ===SORRY!===
Don't know how to 'no scrubs' just yet
[Coke] "just yet". :| 22:52
23:04 Guest7 joined
timo we're working hard! 23:11
23:27 Guest454 joined
Guest7 I see a command on github ("nqp") to run nqp opcodes from a file, but what do I use to run qast, and what file extension do I use? 23:32
timo you might have misunderstood. the nqp command is for compiling nqp code which is still high level compared to just opcodes 23:36
Guest7 My bad. How do I run qast nodes from moarvm then? 23:38
timo you use the MASTCompiler for that purpose 23:42
we might have tests that exercise this, by which i mean starting with a qast tree and running the code it encodes 23:43
may i ask what you want to use this for? RakuAST may be a better fit
or you might desire to go the opposite direction and actually write moar bytecode by hand? or maybe start with MAST actually 23:44
23:45 teatime joined
timo parts of what you need do live in the nqp repo, fwiw 23:52
rakudo uses the HLL::Compiler and HLL::Backend bits that are inside nqp
Guest7 I don't know Raku, but I'd like to target moarvm from another language. I don't know RakuAST, but generating qast nodes seems like an option. 23:55
timo ah, the nqp repo has an example language called rubyish in it 23:56
i'm not sure if anyone is preventing it from bit-rotting

Powered by Cro and the Raku® Programming Language.

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

Thank you!