🦋 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, reportable6 joined 00:28 deferen joined 01:28 evalable6 left, linkable6 left 01:30 m_athias left, nine left, evalable6 joined 01:31 linkable6 joined, nine joined 01:32 m_athias joined 01:36 derpydoo left 01:45 rf joined 01:48 MasterDuke left 02:01 rf left 02:06 RonaldR34g4m left 02:10 Vyrus joined 02:26 teatwo joined 02:27 teatwo left 02:29 teatime joined 02:30 tea3po left 02:36 teatime left, teatime joined 03:17 Maylay left 03:27 vrurg left 03:42 rf joined 03:58 vrurg joined 04:04 rf left
andinus lizmat 04:30
ah sorry, meant to /query
05:04 linkable6 left, evalable6 left 05:06 linkable6 joined, evalable6 joined 05:36 jpn joined 06:00 reportable6 left 06:01 reportable6 joined 06:08 jpn left 06:09 jpn left 06:13 jpn joined 07:00 elcaro left, elcaro joined 07:03 codesections1 joined 07:04 codesections left, codesections1 is now known as codesections 07:06 jpn left 07:09 codesections1 joined 07:10 jpn joined, codesections left, codesections1 is now known as codesections 07:13 abraxxa joined 07:16 jpn left 07:21 jpn joined 07:26 jpn left 07:27 jpn left 07:30 jpn joined 07:42 jpn left 07:46 jpn joined 07:50 jpn left 07:54 ab5tract joined 08:00 jpn joined 08:05 jpn left 08:12 dakkar joined 08:34 sena_kun joined 08:54 jpn joined 08:59 jpn left, Sgeo left 09:32 sena_kun left
andinus looks like rakudo on rockylinux repos is broken 09:41
rakudo is currently at v2023.02, moarvm and nqp are at v2023.05, (seems like moarvm, nqp package size has reduced quite a bit too) 09:42
lizmat recently, like on 2023.06? or longer alreay?
andinus iirc until last month or so it was working, i upgraded a system today and rakudo fails 09:43
with Unhandled exception: Missing or wrong version of dependency 'gen/moar/stage2/MASTNodes.nqp' (from 'gen/moar/Pod.nqp')
09:48 jpn joined 09:53 jpn left 10:04 crimsonred joined 10:06 crimsonred left 11:06 bloatable6 left, committable6 left, shareable6 left, linkable6 left, unicodable6 left, coverable6 left, evalable6 left, benchable6 left, notable6 left, releasable6 left, reportable6 left, bisectable6 left, greppable6 left, statisfiable6 left, squashable6 left 11:07 unicodable6 joined, coverable6 joined, bisectable6 joined, committable6 joined, statisfiable6 joined, squashable6 joined 11:08 shareable6 joined, reportable6 joined, bloatable6 joined, linkable6 joined, evalable6 joined 11:09 benchable6 joined, notable6 joined, greppable6 joined, releasable6 joined 11:18 merp left 11:23 merp joined 11:51 ab5tract left, ab5tract joined 12:00 reportable6 left 12:02 reportable6 joined 12:06 jpn joined 12:13 jpn left 12:15 Some-body_ joined, DarthGandalf left 12:18 Some-body_ is now known as DarthGandalf 13:01 jpn joined 13:06 jpn left 13:11 vrurg left 13:13 vrurg joined
tbrowder__ hi, looking for an opinion: given creation of a LOT of class objects, space and time wise, is it better to minimize attributes in favor of methods or vice versa? 13:21
that is, for example, i can have either attribute z or method z that calculates z if needed. 13:24
i think i'll favor fewer attrs and more methods. 13:30
lizmat YMMV really 13:48
how expensive is the calculation?
if you don't always need z, then calculating it (maybe only once and then store it in an attribute) would be more lazy :-) 13:49
13:54 Xliff joined, Xliff left
vrurg And the last lizmat statement is exactly where AttrX::Mooish shines. :p 14:01
lizmat tbrowder_ ^^ 14:02
14:17 rf joined
rf Morning folks 14:24
tellable6 2023-06-27T16:52:04Z #raku <tonyo> rf: did that work for you?
rf tonyo: lgtm 14:25
14:25 xinming left 14:26 teatime left, teatime joined 14:28 xinming joined 14:38 dakkar left 14:39 dakkar joined 14:49 jpn joined 14:54 jpn left, Voldenet_ joined 14:55 Voldenet left, Voldenet_ is now known as Voldenet 14:58 Maximus joined, Sevalecan left, Maximus is now known as Sevalecan 15:37 Sgeo joined 15:59 [Coke]_ joined
tonyo tak 16:00
16:01 [Coke] left 16:31 jpn joined 16:38 dakkar left 16:41 jpn left 16:45 abraxxa left 17:00 bigdata joined 17:21 amenonsen left, amenonsen joined 17:42 sena_kun joined 17:49 Sgeo left, jpn joined 17:51 Sgeo joined 18:00 reportable6 left 18:01 reportable6 joined 18:51 jpn left 18:57 [Coke]_ is now known as [Coke] 19:46 MitarashiDango[m joined
MitarashiDango[m Is the directory specified in the checksum for Rakudo 2023.06 there intentionally? Its not consistent with previous releases. If it is, I'll need to update the asdf plugin for raku. 19:46
19:49 bigdata left 20:14 oodani left 20:15 oodani joined, jpn joined 20:20 jpn left
Voldenet tbrowder__: contrary to what people say (ymmv) I believe in never storing state that's not needed – computations scale well on 128-core systems, but no kind of storage scales well 20:58
storing state in such cases is an optimization technique for specific use case only, but it shouldn't be go-to strategy by default ever 20:59
Even if returning a value allocates and JIT is not optimizing it anyhow at the moment, it might in the future and state will be a lot harder to optimize 21:02
not that I say to never use state, just use it only when it speeds up a piece of code by 50% 21:07
or something equally impressive
21:20 evalable6 left, linkable6 left 21:22 evalable6 joined, linkable6 joined 21:34 sena_kun left
librasteve tbrowder: I have found that lazy generation of classes is best - for example in Physics::Unit generating all 270 classes on start up was very slow ... so now I just make them when they are needed 21:37
its here github.com/librasteve/raku-Physics...it.rakumod at about line 320 21:40
21:42 jpn joined
lizmat jdv MitarashiDango[m> has a question re checksum 21:42
librasteve PS. I tried leveraging precomp but that route was a dead end - here is a blog post that hopefully illuminates the progress rakujourney.wordpress.com/2021/01/...ysicsunit/ 21:45
21:49 jpn left
Voldenet if I remember correctly using postfix operators reloads the grammar on every new op :/ 21:59
that's why I'm using methods instead of symbols even if syntax isn't as nice 22:01
lizmat Voldenet: that's only true with compilation though, so anything precompiled doesn't have to do that 22:02
Voldenet but then you end up needing to put everything in the module
it works, but it's not as low-maintenance as simply having one file 22:03
lizmat afaik scripts are precompiled as well nowadays?
22:49 linkable6 left, evalable6 left, linkable6 joined 22:50 evalable6 joined 22:51 ab5tract left 22:52 ab5tract joined 23:14 shmup left, shmup joined 23:38 squashable6 left 23:40 squashable6 joined 23:54 codesections left 23:58 codesections joined 23:59 xinming left