🦋 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 00:06 Manifest0 left
el m: perl my @a = 0..*; say @a 00:16
Raku eval [...]
el lazy arrays are a thing
leont When using Getopt::Long, if you do something like «sub MAIN(DateTime :$date is options(&datetime-interpret)) { ... }» that will DWIM :-) 00:18
el i smh can't wrap my head around that 00:19
Nemokosch what is &datetime-interpret? 00:46
what is options in the first place
01:03 xinming left 01:05 xinming joined
guifa_ options I think is something that will modify the $date argument, and &datetime-interpret from DateTime::Grammar parses the date 01:10
01:13 kybr left 01:16 perlmaros left, perlmaros joined
Anton Antonov @Nemokosch 1. The function datetime-parse is a "synonym" of DateTime::Grammar.parse. The function datetime-interpret is DateTime::Grammar.parse with actions -- by default those are Raku-actions. (I.e. DateTime objects are generated.) 2. I use the word "options" as a short cut for "named arguments". (That is my Mathematica background showing up...) 01:23
01:26 kybr joined
@Nemokosch Hmm... I am not sure how leont uses options. "Options" is also a concept used in describing signatures of CLI scripts. 01:27
Nemokosch yeah that name just gives me no clue what its purpose is.. 01:31
Anton Antonov Which name?
leont The option (I typod an extra s) trait in Getopt::Long sets up how named CLI options are turned into named arguments, it can be used if you want something other than the default 01:32
Nemokosch oh this is pretty useful 01:35
funnily enough, Getopt::Long even fixes some issues with the default CLI, like enum lookup 01:36
leont Yeah, enum lookup in the default parser is really broken
01:36 frost joined
leont Try typing something like «zef install True» and see what happens… 01:36
Nemokosch 🤣 01:37
www.youtube.com/watch?v=AR8D2yqgQ1U
Anton Antonov @leont Would you use sequences of date specs in a CLI script? For example: (1) new-temporal-data --dates=yesterday,today , (2) new-temporal-data --dates=2022-02-12,2023-01-12, (3) new-temporal-data --dates=monday,wednesday,thursday. 01:40
[Coke] jdv: fixed it on HEAD, I think 01:49
huh. I have a local script that works on 2022.07 but emits no output on 2022.12 01:51
er, on HEAD
Nemokosch tbh there were regressions in 2022.12 alread 01:55
y
[Coke] I'll see if I can golf it
guifa_ sweet I just got percent and permille formats added to Intl::Format::Numbers 02:02
Anton Antonov @guifa Pictures or never happened! 02:03
guifa_ say format-number .987654, :type<percent>; # 98.7654% 02:05
say format-number .987654, :type<permille>; # 987.654‰
I'll be uploading it shortly
next step is scientific 02:06
Anton Antonov Cool!
02:09 ispyhuma1fly left
Nemokosch Tbh Raku could really be more known outside of, well, us, given the amount of work that is going into it (mainly thinking of Rakudo), basically on a daily basis 02:15
Have you heard of Self? Historically an interesting language, basically all dead
and that's just one random (extreme) example
el it is even more underrated than smalltalk 02:16
Nemokosch Self is at least prototypal while Smalltalk is just your usual class-based OO language without power
My point is also that Raku could be seen as an interesting experiment as well, if we decide to ignore the real-world motives 02:18
say you want to learn about language design, parsers, compilation, etc. 02:19
and all this in a fairly coherent context, in the footsteps of gifted people 02:20
el smalltalk is radical compared to other oo langs like java and c#
Nemokosch sure but I didn't buy into the vision 02:22
leont Self may be dead, but Javascript's semantics are highly based on it. It lives on in a way
Nemokosch and the language is not that powerful unless you are really into this "all interaction happens via message passing" stuff
for Raku, well, I could imagine "The language with ambitions" or something similar as a motto 02:25
the only language I can quickly think of that has this crippling feature mania must be C++ 02:26
leont The main problem with C++ is that it doesn't have the syntax to support all those features, so everything ends up cramped
el scala?
Nemokosch I basically only know the name
el "the language with ambitions and weak typing (in 2015 even)" 02:27
Nemokosch leont: if C++ doesn't have the syntax for it, some of its meta-languages will! xD
and it has quite a couple of meta-languages 02:28
C preprocessor, templates, constexpr stuff... 02:29
02:30 Xliff joined
el raku be becoming the next C++ 02:30
Nemokosch to be honest, Typescript was mentioned... that can also be a scary language
exactly because it's kept from turning into a proper preprocessor sort of language, while it also has no runtime 02:31
02:32 ispyhumanfly joined
so basically you can do all the crazy shit you could do with C++ templates and a bit more - however, you need to write essays for it 02:32
basically Typescript allows literal values to be types 02:33
and on this basis, you can start hacking the type system as compile-time values - as long as you turn the obtained values into types again 😄 02:34
02:36 ispyhumanfly left
weak typing... I think Raku has a reasonable system, no need to jump up to the bandwagon 02:36
ngl I even feel safer when I know what my values are treated as (as opposed to what they really are) 02:37
guifa_ Anton: I just got it to respect language tag numbering systems too! 02:38
format-number 0.987654, :language<en-u-nu-limb>, :type<permille>; # ᥏᥎᥍.᥌᥋᥊‰ 02:39
guifa_ doesn't expect anyone to actually use that, but it's technically part of the standard (probably more for, e.g., Arabic script users to give preference to Eastern or Western style numerals) 02:40
02:41 rf joined
Anton Antonov @guifa Looks interesting. I have to say, I have not heard that before. (Should know about it for my numeric word forms parsers.) 02:42
guifa_ the language codes can give quite a bit of information. When I get back to date formatting, the language code can even encode that! 02:44
for instance, the language coed on my system is
en-US-u-ca-gregory-fw-sun-ms-ussystem
(gregorian calendar, first day of the week is Sunday, and measurements prefer US customary) 02:45
rf .tell tonyo I found an interesting bug in Fez, submitted an issue 02:47
tellable6 rf, I'll pass your message to tonyo
02:50 ispyhumanfly joined
Anton Antonov @guifa Is that language information accessible in Raku? Using your packages? 02:58
guifa_ yes. Take a look at Intl::LanguageTag 03:02
there's a whole international ecosystem
03:02 squashable6 left
guifa_ Intl::CLDR is the gigantic compendium of information 03:03
and right now I'm slowly rewriting all of the Intl::Format::* modules
03:03 ispyhumanfly left, rf left 03:04 squashable6 joined
guifa_ is afk 03:07
03:08 razetime joined 03:14 ispyhumanfly joined
Anton Antonov @guifa Thank you! Will do. 03:39
03:53 razetime left 04:02 pingu joined 04:12 freeside left 04:21 vrurg_ joined, vrurg left 04:26 derpydoo left, freeside joined 04:38 razetime joined 05:11 [Coke] left 05:19 [Coke] joined 05:36 jmcgnh left 05:47 jmcgnh joined 05:55 freeside left 06:00 reportable6 left, reportable6 joined 06:53 Xliff_ joined 06:55 Xliff left 07:11 Sgeo_ left 07:54 freeside joined 08:04 razetime left 08:16 razetime joined 08:25 freeside left 08:49 razetime left 08:53 jmcgnh left 09:00 jmcgnh joined 09:03 Manifest0 joined 09:07 jpn joined 09:10 dakkar joined 09:42 freeside joined 09:45 pingu left, sena_kun joined
Geth ecosystem/main: 8cafd30d90 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | META.list
Remove Sustenance (#614)

Because it is providing almost daily updates without ever increasing the version number, causing false positives in updates and etc.
The REA contains an older version of Sustenance, so people will still be able to install it.		 09:50
10:09 ab5tract joined 10:27 sena_kun left 10:43 Oshawott joined 10:44 archenoth left 11:01 peder left, freeside left 11:02 peder joined 11:03 frost left 11:12 sena_kun joined 11:15 jpn left 11:16 frost joined, ab5tract left 11:29 ab5tract joined 11:42 ab5tract left 11:53 ab5tract joined 12:00 reportable6 left 12:01 reportable6 joined 12:05 TieUpYourCamel joined 12:07 thundergnat joined
thundergnat >[Coke]: "What sort of Unicode properties would folks want 'uni -v <char>' to report out on?" It may be useful to also include the "long name" of the Unicode property. I know them pretty well now, but when I first started I constantly needed to look up what the abbreviations stood for. 12:09
tellable6 2023-02-14T14:30:30Z #raku <[Coke]> thundergnat - Yup, that's very close to what I need.
thundergnat m: gist.github.com/thundergnat/da95af...6e4854c8a7
camelia (Lu Ll Ll Ll Po Pi So Zs Po Pd Pe No)
((Letter, uppercase) (Letter, lowercase) (Letter, lowercase) (Letter, lowercase) (Punctuation, other) (Punctuation, initial quote) (Symbol, other) (Separator, space) (Punctuation, other) (Punctuation, dash) (Pun…
12:10 thundergnat left 12:13 frost left 13:13 linkable6 left, evalable6 left 13:14 linkable6 joined 13:15 evalable6 joined 13:22 merpaderp left
tbrowder ref CucumisSextus: should it not be renamed and made a group project? my initial thought on the subject was to make the Cucumber ppl aware of Raku interest in participating 13:24
lizmat all for making it a group projecty 13:25
13:28 saint- left
Anton Antonov @tbrowder I think there should be a more general project named, say, “Test::Workflows” or “DSL::English::TestWorkflows” that facilitates the creation and utilization of natural language test specifications. Then the functionalities of “CucumisSextus” can be “just” actions under the “Gherkin” sub-namespace. (With further concrete language localizations, like, Raku, Ruby, Java, etc.) 13:37
13:58 jpn joined 14:09 QhpAptyj9hj0RQwM joined 14:18 merp joined 14:33 jgaz joined
tonyo ingy: didn't you have a project in the same vein ^? 14:52
tellable6 2023-02-15T02:47:52Z #raku <rf> tonyo I found an interesting bug in Fez, submitted an issue
tonyo .tell rf let's keep the bugs uninteresting and easy to fix 14:53
tellable6 tonyo, I'll pass your message to rf
tonyo .tell rf also, i will take a look shortly (currently vacationing and bike packing until end of feb)
tellable6 tonyo, I'll pass your message to rf
[Coke] Not a fan of the name of the cucumber projectin light of the big rename. 14:55
Nemokosch what does that even mean? 14:59
Anton Antonov @Coke Duly noted.
[Coke] Looks like it's maybe a latin reference to Cucumber6 - which made more sense when we were Perl 6. Now it's just esoteric. 15:00
It's already a raku project, we can just call it Cucumber 15:01
Anton Antonov @Coke Sounds good. I still consider having a module "Gherkin" that is used in combination with a module "Cucumber". 15:02
Nemokosch well makes sense 15:04
there is this love-hate situation with the number 6 at the moment
Anton Antonov LOL 15:09
15:14 Xliff_ left 15:15 razetime joined 15:31 freeside joined 15:36 freeside left
[Coke] .tell thundergnat - I've been printing [Lu], e.g. in the output of 'uni -s A' - I can print out the expanded text you provide there with 'uni -v A' 15:40
tellable6 [Coke], I'll pass your message to thundergnat
[Coke] IWBNI if I could get raku to give me that info instead of managing my own table, but that's easy enough to include. 15:41
15:43 freeside joined 15:44 Sgeo joined
sjn who's taking care of vim-raku these days? 15:47
zostay? ^
15:47 freeside left
sjn hm. I might have write access.... 15:50
tonyo sjn: updates coming? 15:54
15:54 rf joined
rf tonyo: Does fez bundle the .git directory when uploading? 15:55
tellable6 2023-02-14T13:25:56Z #raku <tonyo> rf: are you parsing in C ?
2023-02-15T14:53:12Z #raku <tonyo> rf let's keep the bugs uninteresting and easy to fix
2023-02-15T14:53:56Z #raku <tonyo> rf also, i will take a look shortly (currently vacationing and bike packing until end of feb)
tonyo .tell rf the test fixes you mentioned are up
tellable6 tonyo, I'll pass your message to rf
15:55 QhpAptyj9hj0RQwM left
tonyo it shouldn't but it might 15:55
rf Hmm. I need to source a submodule in my project, maybe I'll just have to do some wonky stuff 15:56
tonyo do you mean include it in the dist? 15:57
rf Yeah
tonyo can probably do that with a command
is it in your gitignore or anything? 15:58
rf Yeah I'm gonna see if I can force CMake to do it
tonyo rf: it's possible to provide your own tar.gz to the fez upload with the `--file` flag 16:03
jdv [Coke]: thanks! 16:05
rf tonyo: I found a really nice naive way to do it, no worries :D 16:06
tonyo just cloning them into the repo? 16:09
zostay I'm acting as project manager on vim-raku, but I don't do much raku anymore and I am probably not doing a very good job. If someone wants to take it over, I'd be willing to pass control of it over. 16:21
16:21 freeside joined
patrickb zostay: Addopting into community-modules might be a good fit. 16:22
16:26 freeside left 16:46 xinming left 16:48 xinming joined
rf tonyo: Basically. 16:48
It's only one sub-module and CMake supports it so w/e
16:52 freeside joined 16:56 freeside left 16:59 freeside joined, xinming left 17:00 xinming joined 17:03 freeside left 17:06 freeside joined 17:10 freeside left 17:15 freeside joined 17:20 freeside left 17:25 razetime left 17:37 freeside joined, dakkar left 17:43 freeside left
sjn tonyo: I've just mulled about applying some of the pull-requests, that's all 17:48
18:00 reportable6 left 18:01 freeside joined, reportable6 joined 18:02 sena_kun left 18:05 freeside left 18:16 ab5tract left 18:30 ab5tract joined 18:37 sena_kun joined 18:38 derpydoo joined 18:50 freeside joined 18:57 freeside left 19:18 MoC joined 19:26 jpn left 19:30 ab5tract left, jpn joined 19:35 jpn left
Nahita if I have a .rakumod file for common use, what can I do so that use Module; from anywhere uses it 19:35
tried moving the file to some places where it reports it couldn't find it, it still didn't find it 19:36
tonyo you can add `RAKULIB=<path to directory containing lib/Module.rakumod>` or, better, turn it into a full dist and install it to the system with `zef install .` 19:38
Nahita thanks why didn't moving the file to one of the places it reports it couldn't find it in work 19:41
and the minimum it takes to make it a "full dist" is an accompanying META6.json file? 19:42
[Coke] I think it's RAKULIB=<path to directory containing Module.rakumod> 19:43
that is, you don't get a free 'lib' in their.
gah. "in there"
Nahita what would "zef install ." do that i can do manually
[Coke] If you don't want to use zef, then it's not worth duplicating what it's doing, and I'd recommend picking a common install path and then always using RAKULIB or -I on the command line 19:48
if you can show an example of what didn't work, that'd be helpful. (the command line, a directory listing of the path in question...) 19:50
Anton Antonov Ha! I never asked myself (and others) that question!
[Coke] sorry, not "common install" path since it's not really installing at that point, but just a common folder. 19:51
Nahita re: If you don't want to use zef, then it's not worth duplicating what it's doing i was trying to understand what it was doing
rf zef does a lot, what specifically? 19:52
Nahita re: example of what didn't work i have a single Module.rakumod file, then i do use Module in an arbitrary place. then it reports it couldn't find it, and says where it looked or something, then i move this file to one of those places, use again, it doesn't work still
19:53 ab5tract joined
i do this in Perl, it reports where it looked (@INC), then i move one of such places, it works. i do this in Python, i put it in a place in sys.path, it works. 19:53
[Coke] Yes, that should work in raku as well, I do it at $dayjob all the time. 19:54
If you can provide a gist showing what you're running, and a directory listing, that will help us figure out why it's not working for you
if you're trying to just move it into one of the system folders, that isn't going to work. 19:56
if you have a "raku -I/home/coke/modules -e 'use myModule'", and a /home/coke/modules/myModules.rakumod, that should work. 19:57
"isn't going to" -> "probably isn't going go"
Nahita yes that's what i am doing, moving into those system folders; why doesn't it work, was one of my questions 19:58
Anton Antonov @Coke Does your $day-job involve Raku explicitly, or you are sneaking it in? 19:59
Nemokosch I also thought it would work but it doesn't
[Coke] more the latter; I use it where I can, make sure legal team's on board, etc. 20:00
I believe there's a compunit registry that is updated when zef does the install
Nemokosch Actually, the content of named folders is some absolute obfuscated mess, lol
20:00 freeside joined
[Coke] so even though you've put a file out there, the index doesn't have it 20:00
Anton Antonov @Coke Good to know!
[Coke] and what nemokosch said - if you had installed anything with zef, you'd see that (I think) it's not a direct correlation of pre-installed file names to the installed area. 20:01
Nemokosch I mean it's okay that zef doesn't do something obvious - but the obvious thing could work regardless. 20:03
In fact, it does work if you do it the other way around, i.e add your path to the lookup
20:04 freeside left 20:05 xinming left
tonyo nahita: moving it to those sites is probably skipping precomp 20:06
tellable6 tonyo, I'll pass your message to Nahita
20:09 xinming joined
Nemokosch I'd like to think that I've gotten pretty good at explaining how a whole string of Raku behaviors make sense in some way but the situation around modules seems to be a bottomless pit 20:14
Nahita thanks tonyo and [Coke] 20:23
tonyo there's a lot of intricacy in making all of the precomp/modules work together so that when you `use Something` it just kind of works 20:24
Nemokosch there's just always more work than time and actual motivation but this field might need the most help 20:26
20:31 ispyhumanfly left 20:32 ispyhumanfly joined
tonyo in which way? 20:34
20:36 ispyhumanfly left 20:38 codesections joined, ispyhumanfly joined
ugexe just copying the files somewhere and linking RAKULIB (i.e. not installing) means the repository chain works different (since its highest version from first repo - installable repos contain multiple modules but non-installed modules use a single distribution per repo). it also means it might not work the same on e.g. case sensitive file system vs case insensitive file system. it also means things like 20:38
unicode in the names might not work on various systems
20:38 swaggboi left 20:39 freeside joined
ugexe its inherently complex. make easy things easy and hard things possible doesn't imply making hard things easy 20:39
github.com/ugexe/Raku-CompUnit--Repository--Lib -- heres an approximate implementation of something closer to perl where you could just copy the files over, so its not like the possibility hasn't been considered 20:41
20:42 ispyhumanfly left 20:43 freeside left, ispyhumanfly joined
ugexe docs.raku.org/language/compilation#Why_change? 20:47
20:49 ispyhumanfly left 20:50 ispyhumanfly joined 20:53 pingu joined
Nemokosch where does creating something that represents a module and a compunit (i.e a .rakumod file with one sole module in it), and making it available to Raku scripts, fall on the easy-hard scale? 20:54
20:56 ispyhumanfly left 20:57 ispyhumanfly joined
tonyo more difficult than the way that is phrased makes it sound 20:57
ugexe $ raku -e 'require "{$*CWD}/lib/Zef/Utils/FileSystem.rakumod" qw<&list-paths>; say list-paths($*CWD).head' 20:59
"/Users/nlogan/repos/zef/META6.json".IO
Nemokosch and the fact that it's banally simply achieved in the languages in the comparison
ugexe $ raku -Ilib -e 'use Zef::Utils::FileSystem; say list-paths($*CWD).head'
"/Users/nlogan/repos/zef/META6.json".IO
so what exactly is difficult about those?
Nemokosch knowing the path to the file 21:00
ugexe and what languages solve all the issues listed in the Why change? link
21:01 swaggboi joined
ugexe how is knowing the path to the file an issue? you have to know where you want to copy it to begin with 21:01
tonyo this is ignoring how awful other ecosystems are and package managers.. it's very easy in go to add a module but it's not easy to do this. python? same headache but different 21:02
ugexe if you know a language where you can copy an arbitrary file to an arbitrary location and import it using some name i'd be interested to know about it
but afaik you would need to know at least one location (probably a library directory)
21:03 ispyhumanfly left
ugexe and you have to know the name 21:03
so, as in my second example, you combine the library path (which must be known) and the name (which must be known) and you know the location
Nemokosch the funny thing was rather that you are given paths at every failed module lookup - but that paths don't seem to mean a thing in practice, they aren't looked up in the sense a RAKULIB entry is 21:04
21:04 ispyhumanfly joined
ugexe RAKULIB=site 21:04
what do you think that does?
RAKULIB doesn't imply the repository type
you seem to want the lowest common denominator of behavior 21:05
21:05 freeside joined
ugexe at the expense of everything else that has been listed 21:05
Nemokosch well, one clear behavior wouldn't hurt
ugexe how would you know? 21:06
Nemokosch exactly
definitely not by using it
ugexe i mean how would you actually know. you personally 21:07
you claim it wouldn't hurt
but you also claim to be naive to most of it
Nemokosch If there is something I claim, is that the current situation is a mixture of highly incompatible behaviors, and one might conclude that sticking to one - even at the expense of "easy things easy", in some arbitrary interpretation - would do, well, more harm than good, most probably 21:10
more good than harm*, lol
the Freudian slip...
tonyo no one said anything about making `use Something` an easy task 21:11
Nemokosch well, an arbitrary subset of it (use lib and then use) is made seem simple 21:12
ugexe I mean I can claim the moon is made of green cheese, and i'd be wrong 21:13
"I dont know about this subsystem, but I can confidently say it would be better if it accommodated my mental model"
21:14 freeside left
ugexe the world is flat is pretty simple too 21:14
Nemokosch No. I can confidently say that from the user point of view, it stacks surprise on top of surprise, by the rules it gives away about itself via the interface
gfldex lolibloggedalittle: gfldex.wordpress.com/2023/02/15/yes-but-dont/ 21:15
ugexe Well I guess its a shame you weren't around to design everything 21:16
Nemokosch Or at least to witness if it was a definite design paradigm to do everything differently from established practices /s 21:17
ugexe I mean we need language designers with confidence
and yet you haven't address all the stuff it solves
a common theme i've noticed
21:17 freeside joined
ugexe its like you can't comprehend that there are problems you aren't facing 21:18
Nemokosch I'm literally saying that the whole "use lib" hack could go for all I care
that doesn't account for the problems "it solves", does it
21:18 jpn joined
ugexe I'm so confused at what you think established practices are now that I'm not sure what you actually believe 21:19
Nemokosch perhaps because you are mashing two things together in your head 21:20
ugexe I don't think so 21:21
Nemokosch yes, and you could claim the moon is made of green cheese
21:22 freeside left
ugexe ok, the SME is the confused one here not you 21:22
tonyo nemo, i think you're missing the point ugexe is making. making an arbitrary just dump files into a folder that can be `use Something` from anywhere is not as easy as it sounds. there's other things not considered like how would `supercedes` work in that context? 21:23
21:23 jpn left
Nemokosch That's a fair point actually 21:24
Does supercedes actually work?
ugexe i mean i posted an entire repo of code that, if copied into rakudo, would do exactly what you presumably keep talking about
you've just asserted that i'm confused 21:27
Nemokosch by the way, how many times does something need to happen to become a "common theme"? 21:31
phogg Five percent of the time. 21:32
ugexe do i need a license to use the term common theme from you? 21:33
tonyo this issue does come up a lot 21:34
usually stemming from not knowing how a module is actually loading or the complexity involved with dependency resolution etc that _needs_ to happen in order for things to just work 21:35
and by a lot i mean more than 5% 21:38
Nemokosch On one hand, it doesn't hurt knowing what somebody's words mean, on the other hand, I wanted to see if I can use this phrase for the mindset I've seen with library-related design 21:44
tonyo it's a common theme with an underlying problem, if you wanted to tackle documenting more explicitly the complexities so it can just be linked to; no one would complain 21:48
Nemokosch Personally I don't find this module resolution situation "issue worthy", it's just unfortunate that the same tools set up an expectation with the -I flag and use lib and all, and then it gets ripped apart
tonyo there's no issue with using the -I flag globally if you don't mind writing `raku -I$HOME/dev/my-own-module-store/lib ...` every time you want to run something and then shoving everything in `$HOME/dev/my-own-module-store/lib/....` 21:49
Nemokosch However, coming across this shortly after the dependency resolution situation, which I'm again either completely misunderstanding or it really is grotesque... well, that starts to set up a "common theme" 21:50
I'd hope that if the former is the case, I could have understood it a long time ago, had I gotten any sort of guidance but bantering 21:51
tonyo not sure that's helpful 21:54
dependency resolution is very difficult and a well written about subject. most ecosystems do it in a very dumb way, raku doesn't. grab "XYZ", "XYZ" needs "ABC", "GHI", grab "ABC", "GHI", repeat and 20 minutes later you find out that there's no module to resolve or the version conflicts between two modules and the process fails 21:56
raku allows you to put 1.0.0 of ABC for XYZ and version 0.1.0 for something else 30 modules down in the resolver 21:57
and then raku internals also need to figure that out, it's really not cut and dry
rf Yeah I find that the Raku dependecy resolving is top tier
tonyo the same is true of the other side of that, package managers have a lot of issues around it. even ones that have been around for a long time (apt?) 21:59
haskell has the best i've seen but it's so pedantic as to be annoyingly useless 22:00
moritz if it's useless, it can't be the best, no? :-) 22:01
Nemokosch Tonyo: yes, this makes sense, and it's clearly a good thing
What I cannot wrap my head around, though, is that units of dependecies are modules, even though they cannot even provide all the metadata needed to pin them; not at module resolution time at least 22:03
Because that metadata comes on a per-dist basis
*dependency resolution time, sorry
gfldex moritz: uless it's art. 22:05
tonyo moritz: technically correct is the best kind : ) 22:10
they can, they just don't yet and there is a PR that ugexe and (i forget who) are reviewing 22:11
Nemokosch That's definitely a resolution (of the issue, that is) if modules learn to "stand for themselves" 22:14
That would also mean that they can have their own version within a dist, though, wouldn't it?
22:18 squashable6 left
[Coke] hurls github.com/coke/raku-uni#-v-verbos...-character 22:18
22:19 MoC left 22:20 squashable6 joined 22:23 freeside joined 22:29 freeside left
Nemokosch meaning, v2023.02 seems good to go? 22:35
22:46 jpn joined
[Coke] .tell thundergnat github.com/coke/raku-uni#-v-verbos...-character 22:46
tellable6 [Coke], I'll pass your message to thundergnat
[Coke] (includes the expanded items he mentioned)
22:51 derpydoo left 22:56 thundergnat joined
thundergnat [Coke] 👍 22:56
tellable6 2023-02-15T15:40:56Z #raku <[Coke]> thundergnat - I've been printing [Lu], e.g. in the output of 'uni -s A' - I can print out the expanded text you provide there with 'uni -v A' 22:57
2023-02-15T22:46:37Z #raku <[Coke]> thundergnat github.com/coke/raku-uni#-v-verbos...-character
[Coke] uploaded a new version of "uni" that avoids the test failure on recent raku and adds -v
thundergnat While you are editing things, may want to change the repository description Perl 6 -> Raku. 22:58
[Coke] thundergnat: done, thanks 23:00
fez++ 23:01
guifa_ weekly: arstechnica.com/information-techno...nts-page=1 <-- speaking of modules ha 23:23
notable6 guifa_, Noted! (weekly)
23:33 codesections left 23:43 thundergnat left 23:47 freeside joined 23:51 freeside left
tbrowder my 2 cents but if ppl would just create modules for personal use as if they would be published (and use App::Mi6 to create them), then use zef to install them locally, a lot of the "problems" using modules would disappear. 23:53
ugexe and tonyo and others have put in immense energy, talent, and time to give us a great module system that just "works" if you play by the rules. 23:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!