🦋 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.
SmokeMachine If any one is interested: I changed some things (like it have it's own iterator), but I think it is starting to look good: usercontent.irccloud-cdn.com/file/.../image.png 01:27
there, I'm saving the op (a group that groups all operators: infix, prefix, postfix, etc) node, child of apply-op (that groups the apply of all operators) where the `.right` node is 1 using the name `op` and getting it after the query. 01:31
SmokeMachine That's cool that I didn't realised that RakuAST::Call::Method was an operator... 01:35
ab5tract Yeah, the one thing that makes me nervous about creating categories like 'op' over the AST classes is that it's maybe not a 100% mapping, either conceptually or in the RakuAST type hierarchy 07:29
also, not sure how to parse ' ... op$op|).<op>' ... but I also haven't had coffee yet 07:30
Manifest0 Hi! Yesterday i got a core dump are you guys interested in it? 08:54
lizmat is NativeCall involved ? 09:00
Manifest0 probably. I'm not using directly, but maybe libxml or dbiish might use it 09:02
i just checked and both libxml and dbiish use nativecall. So most probably nativecall is involved 09:06
Manifest0 here is the bt output paste.opensuse.org/pastes/c8454c67446e 09:37
SmokeMachine ab5tract: on the categries, we can have it defined class by class... avoiding parents if needed... 09:52
ab5tract maybe one way to go would be to have pluggable categories 09:53
lizmat Manifest0: sorry, I think that tells us something went wrong in garbage collecting 09:54
ab5tract so you can create a specific set (or maybe even chain) of types to match
lizmat and not a lot more: if you have a reproducable case, then it would be interesting 09:55
SmokeMachine about the ' ... op$op|).<op>', ast-query now returns a ASTQuery::Match (that is Positional and Associative) and when you define a `$name` the match will be assigned to $match<name>... (kinda like $/)
ab5tract ah, got it
SmokeMachine ab5tract: yes, I plan to allow the user to customise the categories... 09:56
ab5tract `.op$op` still feels a bit squished to me
SmokeMachine: awesome!
Manifest0 lizmat, it was the first time i got this error. I'm not able to reproduce it :-(
ab5tract Manifest0: :(
SmokeMachine ab5tract: still about the categories, you can, instead of using `.group`, use just the class name like: `RakuAST::Call`, or `RakuAST::Call#say` 09:59
or `RakuAST::Call#say$call-say` 10:00
SmokeMachine ab5tract: for example: usercontent.irccloud-cdn.com/file/.../image.png 10:06
ab5tract ohhhh 10:17
SmokeMachine: that clears it up nicely, thanks :)
SmokeMachine ab5tract: just to point out that the positional part of the ASTQuery::Match will store a list of the complete matches: usercontent.irccloud-cdn.com/file/.../image.png 10:19
(I mean, not only the named ones) 10:20
melezhik. Hi! I’ve run my r3 test suite against Rakuast enabled Rakudo ( head ) and got some failures , @liz are you interested in the report ? 11:00
sparky.sparrowhub.io/report/gh-mel...tool/22855 11:06
lizmat melezhik tbh, there are still so many roast files with failures, so it's a bit overwhelming 11:18
if you could golf some of these cases, and create an issue with them and mark them with the RAKUDO_RAKUAST=1 label, that would be much appreciated 11:19
Xliff \o 12:09
I am getting some weird behavior. I am populating a hash at INIT time, but it appears to be serialized after that. If I make changes to the hash file, those changes aren't reflected unless the module is itself recompiled. 12:10
Is this a bug?
Note all changes to the hash are made at INIT, not BEGIN
timo you can try outputting the hash's WHICH to verify you're looking at the same object at different times (though WHICH does not stick around across serialization) 12:13
Xliff timo: Is the INIT closure built at compile time? 12:13
timo that's possible, but it should be closing over the lexical, rather than the literal object, right? 12:14
Xliff Yes, but the object is global. 12:15
timo is this from bug hunting that report from today or yesterday or so, or something for your own projects?
if loading the module causes the mainline to be run, I would assume a "my %foo" would be fresh each run, but an "our %foo" probably wouldn't be 12:16
Xliff Yeah. It's our 12:17
And no, this is from my mass of growing code. 12:18
timo unfortunately, i'm not too familiar with the bits involved here
Xliff ROFLMAOACK 12:19
Actually, that's KTC not CK
I added $hash = DateTime.now to the population routine and now it works.
timo what's AKTC expand to? 12:20
Xliff 🤣
"And Killing the Cat"
Which is mean. I lub mah kittehs!
timo yeah i much prefer none(&KTC) 12:21
Xliff em-content.zobj.net/source/animate..._1f923.gif 12:22
If only we could animate unicode.
timo: no pets?
timo no, just no killing of the cats
Xliff Yeah. This is from way back in college. We abused "ROFL" for fun and profit back then. 12:27
I'll have to be better about that, but "ROFLMAO" just didn't match the level of amusement.
So why do you think "our %config" doesn't pick up the changes unless "%config<random-key> = DateTime.now" is present? 12:28
timo i wonder if writing to a key in it triggers repossession that otherwise doesn't happen? 12:36
tbrowder good morning fellow raku ppl. can zef search cpan for old raku modules? 13:05
i didn't see anything in its help about cpan
lizmat perhaps raku.land/zef:lizmat/Zef::Configuration is what you're looking for? 13:07
zef-configure enable cpan
is what should do the trick I think
antononcube @Xliff Why are lubing your cats? (Maybe it is an Australian thing...) 13:24
@Xliff "If only we could animate unicode." -- I have been trying to figure out how to do animations in Jupyter notebooks, with a Raku kernel, for awhile. I strongly suspect a different solution that Jupyter is needed. 13:26
Xliff antoncube: I am in the US, not Australia. Although I would love to visit there! 13:28
antononcube Agh, dully noted. (And, I was not exercising Cunningham's Law...) 13:30
As for animantions -- I have been curious how easy it is to re-implement this in Raku: nbedos.github.io/termtosvg/ 13:36
timo well, jupyter allows you to embed arbitrary(?) HTML widgets with javascript, doesn't it? and i think there's some protocol to stream data back and forth between kernel and frontend as well that you can use together with that 13:49
antononcube Yeah, the exchange is via ZeroMQ. 13:54
But I would rather do the animations on "JavaScript level." 13:55
timo when you use \r or "move cursor up" ansi escapes, and then print, the output cell on the jupyter notebook will also update, correct? that should be usable for simple things at least
antononcube Hmm... it would be interesting to try that out. 13:57
Unfortunately, the main JavaScript library I choose to invest for Raku graphics -- D3.js -- is not that good for dealing with images or interactive content. 13:58
Animations by themselves are useful only to a point. A much more useful functionality is to have interactive interfaces. (JavaScript code generated using Raku code.) 14:01
I am not sure that latter problem/task can be done using Jupyter. And, to some extend, I do not want to learn the Jupyter framework that much. 14:02
Ideally, Raku I would have a notebook solution that is tailor-made from scratch using JavaScript (or TypeScript.) 14:05
nijmegenzuigt a wild TiddlyWiki appears 😛
antononcube Ha! Have not heard of it! 14:07
I was thinking hijacking WLJS in some way: jerryi.github.io/wljs-docs/
See the dynamics section: jerryi.github.io/wljs-docs/frontend/Dynamics 14:08
nijmegenzuigt js.wiki/ maybe?
timo building something compatible with ipython's widgets package would be a good thing in general, since things like euporia specifically support that 14:09
antononcube This is what "Jupyter::Kernel" is for.
I consider Python is "too stupid" and worth investing time to know it better. 14:11
timo this is less about python and more about the ecosystem of tools that work with the jupyter ecosystem
antononcube If really want interactive functionalities for Raku, I will "finish" my Mathematica re-implementation in Raku and use the "RakuMode" connection in Mathematica notebooks. 14:12
@timo I get -- I work(ed) with that ecosystem. I find it good, but not great. 14:13
nijmegenzuigt guessing things like loqseq also arent quite there? 14:24
antononcube Is that an Obsidian alternative? 14:26
nijmegenzuigt I never used Obsidian
antononcube I have not eigther, but "Markdown::Grammar" supports it. One can create Jupyter notebooks from Obsidian-flavored Markdown files. 14:30
melezhik. @lizmat - you don’t need to golf it neither create issues - r3 tests are for GitHub issues in Rakudo repository , so that means in test named 1234 fails - there is issue number 1234 in Rakudo GitHub repo that describes and golfs that case , and now it fails under rakuast , 99% golf code is the what original issue author presented when created an use 14:39
an use -> an issue 14:40
lizmat melezhik understood, thanks! 15:59
tbrowder lizmat: thanks! 18:33
jdv anyone else see this when trying to install Whateverable from a checkout?: Attribute $!max-age not declared in class Whateverable::Webhooks::StrictTransportSecurity 18:52
at
[Coke] github.com/Raku/whateverable/issues/390 19:09
yes. :)
jdv oh, why isn't that commited then?... 19:13
who can do that?
[Coke] I didn't commit it because I have no idea if it's the right fix. 19:16
... but I'll do it and then you can report any further issues you find?
jdv that'd be cool, thanks. 19:17
[Coke] there, now you can just press a button: github.com/Raku/whateverable/pull/393 19:19
:) 19:20
jdv thanks, i'll look later, maybe after i figure out how this rando errata test is failing 19:21
ab5tract done clicked 20:37
[Coke] thanks 20:42
ab5tract Comma 2.0 Beta 21 is hot off the press. Semi-automatic dependency management! 20:52
tonyo \o/ 20:54
ab5tract Which is to say, you can install any missing dependency as a context action.
We are getting close to gold :)
I've made my life easier by not worrying about Pakku at the moment though. So if we have any Pakku users here who want/need it supported in Comma, please speak up. 20:58
jdv wonders if anyone uses it
ab5tract It makes my life easier to assume that nobody does, but you know, torture the implementor and all that 21:05
ab5tract But it seems like I'll have to climb the hill of :auth<>:ver<>:api<> before I can call dependency management done for proper 21:26
jdv who can make a new release of Whateverable to fez or wherever? 21:33
it appears there's an unreleased fix for the SSL symbol error 21:36