🦋 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.
melezhik o/ 07:42
just applied this Minecraft theme to my WP site, looks cool, huh? sparrowdo.wordpress.com/2025/02/03...enerators/ )) 07:43
timo "Discover the ultimate Minetest resource – your go-to guide for expert tutorials, stunning mods, and exclusive stories. Elevate your game with insider knowledge and tips from seasoned Minetest enthusiasts." - i'm not sure that's the right text for the sparrowdo site :) 08:53
"I’m Bedrock. Discover the ultimate Minetest resource – your go-to guide for expert tutorials, stunning mods, and exclusive stories. Elevate your game with insider knowledge and tips from seasoned Minetest enthusiasts." - this neither 08:54
melezhik . 09:17
weekly: sparrowdo.wordpress.com/2025/02/04...ow-checks/
notable6 melezhik, Noted! (weekly)
melezhik post thing on linux channel programming.dev - programming.dev/post/24947338?scro...ments=true , not sure if Linux people are familiar with Raku, please upvote if you like ... thanks 09:34
timo did you see my messages from ~45min ago melezhik? 09:39
melezhik timo: I can check by history logs 09:40
or feel free to repost it
^^ @timo
timo it was just that there's some text refering to minetest specifically that's now in your website 09:41
melezhik ohhh - the title 09:42
it must have crept in when I changed the theme
thanks timo
fixed now, thanks a lot @timo 09:44
anyways this was fun - my 2 boys used to play Minecraft a lot )) 09:45
timo i still see the two blocks of text 09:52
in the footer of the page and in the top block of the right-side bar next to an article 09:53
and the twitch, twitter, tiktok, facebook, instagram buttons all just link to the front page of that service, so can probably just be removed 09:54
melezhik oh 09:57
I guess I am going revet this theme back, too annoying to change all the bits, anyway - thanks! 09:58
melezhik timo: finally got back to bedrock , just have removed all irrelevant traces 10:21
lizmat clickbaits rakudoweekly.blog/2025/02/03/2025-...ie-awaits/ 13:14
[Coke] is mi6 standardizing on http:// instead of git for github URLs? 13:26
lizmat yeah, it's a bit annoying
fwiw, I use ssh:/ for mine 13:27
takes a little tweaking in .git/config and only once
[Coke] I understand it, was just surprised.
lizmat BTW, I think I will have the basics for ecosystem source searching done later today 13:45
for *all* the current modules :-)
timo it would be kinda neat if we had like a search index based on AST instead of just input text 13:54
though tbh i don't have a use case for that yet
lizmat I have thoughts about integrating that into rak, fwiw
timo oh, parsing raku source to ast, then searching through the structure? 13:55
lizmat yeah... problem is that too much code out there doesn't parse to RakuAST yet :-( 13:56
timo right. plus side-effects from parsing code, of course
lizmat and there's that, indeed :-)
timo who wants to try if we can modify the rakuast of a program and use DEPARSE to generate a diff of the source code that represents the changes of the ast? 13:57
kind of like structured editing / structured patches / semantic patches
something for the very future maybe 13:58
lizmat who wants to try get RakuAST finished
feels like it should be done first
:-)
timo phew. 13:59
lizmat afk for a few hours& 14:10
[Coke] lizmat: cut a release of date-slang with your update. sorry that fell through the cracks. 14:20
librasteve I have raised an issue on mi6 to request using yaml (or something else) for config rather than dist.ini since it's kinda weird seeing that on a mac ... got no joy from the author ... but I suspect that skaji may have been more amenable had I provided in the form of a PR 14:52
perhaps that (a PR) will work, maybe as a setting for git:// instead of http:// ... maybe too late to change 14:53
[Coke] FWIW, I'm fine with dist.ini for something that I setup once and almost never have to edit again (I also release all my dists on a mac.) 15:23
I also wouldn't necessarily push people to git or ssh over http to github urls. http is fine for most of what users of our modules are going to do. 15:24
(but if authors want to skip that override, they should be able to) 15:26
El_Che There is big group of people that dislike ini (or TOML). :) 15:34
[Coke] fair 15:39
librasteve yeah well it works fine with dist.ini on a mac … just makes me retch a little 15:49
[Coke] De gustibus non est disputandum 15:53
El_Che [Coke]: Etiam de eo disputandum est! 16:11
:)
antononcube @El_Che Good to know. (About TOML.) I am following B. Duggan's "lead" for many of infrastructure packages and he uses TOML. (E.g. "LLM::DWIM".) 16:25
For now, I mostly stick to JSON for special initializations.
El_Che I suspect that TOML was a joke on .ini and people ran with it 16:27
antononcube: JSON is awful for config. No YAML because of the superset thingy? 16:28
but as [Coke] states De gustibus non est disputandum
antononcube Yeah, I tend to dislike YAML. 🙂 TOML seemed a good alternative. 16:29
El_Che I can't easily parse the flatness in my mind 16:31
tonyo you guys know ingy is in this channel, right? say nothing bad of YAML ! 16:35
(haha)
jdv yaml is not a superset of json:) 16:36
El_Che are you going there, jdv? :) 16:37
jdv nah, i don't care
tonyo i ran to metacpan to find that doc battle 16:38
ugexe git:// is deprecated 17:05
antononcube YAML is used too much in R. But, I -- and others -- tend to distrust languages with significant white space. So, I avoid using YAML. (Except when working with R.)
ugexe github.blog/security/application-s...ty-github/ 17:06
i guess if you aren't using github that may not be relevant though
[Coke] finds blog.readthedocs.com/github-git-pr...precation/
jdv wut? git:// isn't deprecaated. old key algos are. right? 17:57
melezhik. My 2 cents to yaml “holy war” - yaml as configuration form is fine as long it’s simple and not too nested . Coding in yaml - imho not the best idea. 17:58
jdv i stumbled on a circleci conf once that took me a while to figure out. think it was using multiple refs or something. too much fun for me. 18:00
ugexe git:// is deprecated on github yes 18:30
git@ is still fine 18:31
jdv ah, gotcha 18:48
coleman Toml is great IF you don't have deep nesting. 19:46
El_Che toml is great if you don't use it. :) 20:06
(just kidding, use what you like) 20:07
melezhik. Well the same ( my statement for yaml ) works for toml - when nesting is not too deep )) , it’s just ( when one has too many nested layers ) in yaml you struggle with bunch of tabs, subtle errors due to that and with toml - I am not sure - but I don’t think it is designed for complex structures ad well 20:08
El_Che as long it's just configuration, you'll be fine. It's only painful when you try to be smart 20:09
melezhik. Yeah . Or when you try to make your configuration smart 😀
El_Che json could be ok, besides commas and comments 20:11
so, no
melezhik. JSON is ok, but I would say for machine interactions, for people it’s a bit verbose 20:16
timo what's the name of that ini-like format again ... 20:17
TOML?
El_Che yes
timo so far that seems like a good format to me 20:20
El_Che it's annoying when you need lists of dictionaries 20:21
s/of/or/ 20:22
I don't need to look up stuff up when using whatever config format
I need to look everything up in toml 20:23
and I can't understand it later
the perl is write-only joke does apply to toml
(the python whitespace thing applies to yaml)
timo i'm thinking if you have configuration files in whatever format, you should also offer a template file with examples and commented default values and all that. in that case it wouldn't be write-only for people making changes, at least i think so 20:26
ugexe definitely a big drawback to json 20:27
coleman I am a big fan when software is packaged with a 99% commented-out and documented config file that shows the defaults 20:29
Or when cli tools have a generate-config subcommand that spits an example file to stdout.
timo that too, yeah
and if there's something like "includes" or anything involving conditionals and maybe overrides from environment or commandline arguments, it's great to have a "write out the effective configuration" flag as well 20:31
and a flag to check a given config for correctness without doing anything that has side-effects ... if possible 20:32
coleman yes! that is such a nice feature
every incident I've ever caused was config-related 20:33
librasteve on this general topic, I have a couple of modules that write an example YAML to eg .raws-config/aws-ec2-launch.yaml during Build.pm ... but then I also have to cp out any existing file to a backup location to avoid overwriting any changes made by the user - is there any best practice out there? - is there any standard for where to stick my yaml configs? 21:06
timo ah, that's tricky. dpkg and rpm have configuration file handling built in, so they check if you have changes and then offer you a choice how to deal with it. but i'm not sure that's "in scope" for zef 21:09
librasteve wordpress provides wp-config-sample.php and a configurator to make the final wp-config.php ... 21:12
timo for finding where configuration files should go, the XDG specification can be good. XDG_CONFIG_HOME is where you would probably put thees yaml files, though i'm not sure what they do exactly. could also be for XDG_DATA_HOME instead. and of course a fallback for systems that don't have XDG 21:18
tonyo well, yamlscript is aiming to take care of templating 21:19
librasteve timo: I recall some discussion here about XDG and the general sense was that it is not widely enough supported to be a decent standard iirc (sorry if I have poor memory) .... what I would like is for there to be something like a consensus that the path HOME/.raku-config/CLI/AWS/EC2-Simple/config/aws-ec2-launch.yaml is good where the filename.yaml are up to the discretion of the module writer - and an easy way for 21:34
the user to get the path and to go in and edit the yamls
timo well, if the auth isn't part of the path, it's no good :) 21:38
librasteve I guess this could be somewhere like ~/.rakubrew/versions/moar-2024.10/share/perl6/site/config/CLI/AWS/EC2-Simple/aws-ec2-launch.yaml but the site dir these days is quite obfuscated in terms of the module name / lib path and afaik there may be a good reason that the module paths are not generally used and also this is pretty deep for someone to go in and edit a config vs eg. ~/.zshrc
timo it depends very very much on how you want to handle the configurations 21:39
librasteve ah yes probably you will want the auth and the ver
timo if you put it below a specific version of rakudo, then a rakudo update will make your configs disappear
i mean your user's custom changes wouldn't be found any more
librasteve good point
but the config file format may change from time to time so maybe some notion of migration 21:40
timo yes indeed 21:41
it's a topic that requires a lot of careful thought, and there is probably no one-size-fits-all solution at all?
librasteve yeah - I do see that 21:42
coleman XDG is widely used 21:43
Caveat: for software post ~2015-ish 21:44
There's a lot of magical systemd integrations with XDG, but even if you don't use or pay attention to that stuff, it's a very well thought-out system. 21:49
librasteve irclogs.raku.org/search.html?query...2025-02-04 21:53
yeah - looks like my recollection is unfairly reflecting on the popularity of XDG - having found and scanned the irclogs ^^ 21:54
ugexe i dislike that xdg puts stuff in multiple locations. it is quite convenient to remove a single directory to clean up that stuff. it is also nice to look in that directory to see all the things you could possibly want to edit 22:11
in other words i dislike xdg because im lazy 22:13
librasteve laziness is a virtue 22:15