🦋 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.
00:00 reportable6 left 00:02 reportable6 joined 00:11 vrurg_ joined 00:15 vrurg left 00:23 jjido joined 01:06 razetime joined
SmokeMachine antoncube: I was more reacting to the backlog when lizmat suggested the trait “is shortcircuit”, I needed overriding && and || for Red, but had to do that on a different way… 01:16
guifa SmokeMachine what are you trying to do?
01:16 JonathanTart joined 01:18 JonathanTart left
SmokeMachine guifa: I’m not trying that anymore… I was trying doing that for this: github.com/FCO/Red/blob/ac69941c40...s.pm6#L832 01:19
but I’ve done that in a different way… 01:20
guifa ah okay
when I finish it you might be interested in my new and improved regex module 01:21
SmokeMachine And now, with out my computer, on vacation, I realised an easy way of doing what I was expecting custom compiler passes to…
guifa i do like you and have moved to a ECMA262-AST class model
SmokeMachine Interesting 01:22
guifa but I still compile ultimately to RakuAST, via a GENERATE method
SmokeMachine After seeing the BF module I’m really wanting to play again with EEL (github.com/FCO/EventExpressionLanguage) 01:23
After seeing the BF module I’m really wanting to play again with EEL (github.com/FCO/EventExpressionLanguage) 01:24
antononcube @SmokeMachine See the section "Infix operators" for the alternatives of && and || I came up with: raku.land/zef:antononcube/FunctionalParsers .
SmokeMachine guifa: any snippet I could have to have a taste?
guifa actually, lemme through it up on a branch
throw* 01:25
antononcube @SmokeMachine Yes, event processing is very important! 🙂 01:26
SmokeMachine antononcube: that looks good, but for Red the intent is to make it looks like normal Raku code… 01:27
antononcube @SmokeMachine I know what you are saying, but using these operators does look like normal Raku code. 01:28
my &p = (&p1 (|) &p2 (|) &p3 (|) &p4) (&) (&pM (^) {10**6}) (&) &pTh; is completely normal!!! 🙂 01:29
Agh... bad example above. This one is better : (&p1 «|» &p2 «|» &p3 «|» &p4) «&» &pM «o {10**6} «&» &pTh;
SmokeMachine Now my intention is, when I get my computer back, is to create an role to mix with raku’s action class and for (lexically) every map/grep/first get the ast for the block, translate it to red-ast and pass it to the method call… because those are methods, it will just ignore the unknown named parameters… but the methods from Red will use that to create the sql… 01:31
01:31 jjido left
SmokeMachine antononcube: but you with Red, we need to be able to do something like this: github.com/FCO/Red/blob/ac69941c40...-map.t#L25 currently I have to run that code many times to try to understand what’s inside that block… if what I’m thinking work… that will not be needed anymore! And everything will be done before runtime! :) 01:37
01:56 Maylay joined 01:58 guifa_ joined
[Coke] github.com/Raku/Blin/issues/31 - blin doesn't even compile here. 01:58
guifa_ github.com/alabamenhu/PolyglotRege...ee/rakuast
to see how I'm doing the intermediate AST. Actions make the nodes defined in Classes, and those have a method called GENERATE that create the equivalent in RakuAST::Regex nodes (unfortunately I jumped the gun, somehow the Frugal modifier is causing compilation problems because its not deparseable ATM) 01:59
SmokeMachine ^^ 02:04
There's a LOT of holdover in the actions nodes from my old code and it'll be cleaned up a lot, but Classes is much nicer looking and readable\ 02:05
guifa_ is afk 02:06
02:12 guifa_ left
SmokeMachine guifa: that’s looking good! :) 02:14
03:11 rf left 03:56 rf joined 04:04 razetime left 04:16 askmeaboutloom left 04:42 teatwo joined 04:45 razetime joined, tea3po left 04:59 rf left 05:59 evalable6_ left, coverable6 left, bloatable6 left, squashable6 left, committable6 left, tellable6 left, nativecallable6 left, benchable6 left, unicodable6 left, bisectable6 left, shareable6 left, quotable6 left, statisfiable6 left, greppable6 left, sourceable6 left, reportable6 left, linkable6 left, releasable6 left, notable6 left 06:00 quotable6 joined, benchable6 joined, squashable6 joined, linkable6 joined, sourceable6 joined, committable6 joined, releasable6 joined, coverable6 joined 06:01 nativecallable6 joined, bloatable6 joined, bisectable6 joined, tellable6 joined, reportable6 joined, statisfiable6 joined, shareable6 joined, evalable6 joined 06:02 greppable6 joined, unicodable6 joined, notable6 joined 06:17 jpn joined 06:42 jpn left 06:45 jpn joined 07:38 sena_kun joined 07:53 sena_kun left 08:17 jpn left 08:29 jpn joined 08:30 Sgeo left 08:39 razetime left 08:57 razetime joined 09:57 jpn left 10:04 jpn joined
antononcube @guifa Can you make the sub ecma2raku which you use in the tests of "PolyglotRexen" to be a module function. I want to opportunistically extract regular expression from LLM outputs. Using "PolyglotRexen" seems to be a good fit. (The use case is literally in the name.) 10:24
@guifa Hmm... I still cannot install "PolyglotRegexen". I think we discussed that before... 10:32
nemokosch what happens, good sir?
antononcube @nemokosch Lots of CompUnit and test errors. The first one is Could not find Polyglot::Regex::ECMA262::Grammar. 10:33
nemokosch what did you run? Let me try as well 10:34
antononcube I tried both zef install Polyglot::Regexen and zef install github.com/alabamenhu/PolyglotRegexen.git 10:37
10:38 razetime left
nemokosch confirmed... 10:41
antononcube @nemokosch Thanks!
Here is my use case: my &fr = llm-function({"Give regular expression for parsing $_"}); say &fr('emails'); # The following regular expression can be used for parsing emails: # # \b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,}\b 10:43
The LLM output is text. but with &fr('emails', form => Regex) I would get an array of strings and regex(es). 10:44
nemokosch the META6 exposes the modules with different names than what the tool uses internally, I think it's a bug
Polyglot::Regex::ECMA::Actions when it really should be Polyglot::Regex::ECMA262::Actions 10:45
guifa I'll try to fix that today, about to go to work 10:48
10:56 MoC joined
nemokosch thanks 10:57
11:12 Xliff joined
Xliff \o 11:12
nemokosch what's up: 11:13
11:50 jgaz left, jgaz joined 12:00 reportable6 left 12:01 reportable6 joined 12:08 MoC left
Xliff My blood pressure? ;) 12:09
I'm working on GIMP (pre)3 bindins for Raku
With the eventual hope that I can make Raku into a target for GIMP Plugin Writers. 12:10
lizmat sounds cool! 12:12
Xliff Working around the whole issue of plugin registration for now, though. 12:15
There are several UI elements available in libgimpui that don
There are several UI elements available in libgimpui that don't need GIMP at all! :)
I am currently in the proccess of liberating those widgets! ;) 12:16
12:21 jjido joined
antononcube @Xliff Ditto, sounds cool. Can Raku be used a scripting language in Gimp? (I assume you are working on Gimp-in-Raku.) 12:52
Xliff antoncube: That's my hope is that it soon can be. yes. 13:18
Raku-Fu
I just need to figure out how to register with Gimp. 13:19
antononcube Haha - so cool!
Gimp in Raku would be great too -- I can then hook it to LLMs (ChatGPT, DALL-E, etc ...) 13:20
Xliff And the snowball grows so well and rolls down the hill of cool things we can make it do. 13:21
13:21 linkable6 left, evalable6 left 13:22 evalable6 joined 13:23 linkable6 joined
antononcube Because of all the gimps who rolled boulders high? 13:42
13:43 jjido left
Xliff antoncube: I would resemble said remark. 13:45
1,040,000 lines of Raku as of today.
antononcube Sisyphus work! (Sisyphusian?) 13:46
SmokeMachine guifa: fosstodon.org/@smokemachine/110792173543505423 14:33
15:06 Sgeo joined
Xliff SmokeMachine: Many of Raku's Grammars go above and beyond the functions of ECMA Regex. You might not be able to convert everything. 15:20
15:35 Guest16 joined 15:47 Guest16 left 15:49 razetime joined
tbrowder__ leont: ref TAP windows failures, have you tried, as an interim step for now, commenting out any parallel testing? 15:56
ugexe I thought we knew the TAP issues were related to long paths 15:58
TAP issues on windows rather 15:59
[Coke] is there a bug for this issue? 16:01
(my long path issues on win were not test related)
ugexe I'm guessing its github.com/Raku/tap-harness6/issues/61 16:02
snonux SmokeMachine, oh, you are here, too (seen you on Mastodon reg. that Regexen) :-)
[Coke] added a comment about long path, maybe they can repro it 16:16
16:27 squashable6 left, squashable6 joined
tbrowder__ disregard the above 16:28
looking at github debug logs i don't see path lengths obviously too long (<255) 16:38
but log does show some kind of path error. i'll see if i can shown it in a gist... 16:40
the path error is on the line showin #site sources of App::Prove6 16:44
16:44 tea3po joined 16:46 teatime joined 16:47 teatwo left 16:48 teatwo left, teatwo joined 16:49 teatwo left, tea3po left, teatwo joined 16:52 teatime left, tea3po joined 16:55 teatwo left 17:00 rf joined 17:36 jjido joined 17:43 sena_kun joined 17:45 jjido left 17:53 rf left
tbrowder__ i’ve been interrupted will post logs in a while 17:56
18:00 reportable6 left 18:02 reportable6 joined 18:21 razetime left
tbrowder__ here's main part of log: 18:45
gist.github.com/tbrowder/ed2f2ff9f...b93fdabbef
19:21 evalable6 left, linkable6 left 19:23 linkable6 joined, evalable6 joined 19:42 rf joined 19:45 kst left 20:06 raiph joined 20:11 tea3po left 20:12 teatime joined
ugexe yeah that doesnt look like a path length issue 20:50
its failing when rakudo precompiles it during install
the path is weird though - D:\a\path-finder\path-finder/site#sources\5D2A0C938B6800BF1A3F66830C6AD233F1FC6C01 20:51
where is the config for your GH action? 20:55
SmokeMachine snonux: yes, I’m always here… usually a bit slow, but always here :)
21:12 linkable6 left 21:13 linkable6 joined 21:21 rf left 21:41 raiph left 22:20 sena_kun left 22:47 Xliff left 23:35 rf joined 23:37 jpn left

Powered by Cro and the Raku® Programming Language.

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

Thank you!