This channel is intended for people just starting with the Raku Programming Language (raku.org). Logs are available at irclogs.raku.org/raku-beginner/live.html
Set by lizmat on 8 June 2022.
00:16 discord-raku-bot left 00:17 discord-raku-bot joined 01:16 discord-raku-bot left 01:17 discord-raku-bot joined
jaguart irclogs.raku.org/raku/2023-01-04.html#01:34 02:15
At the moment I limit Github tests on Windows to 2022.07 and they work ok 02:16
02:17 discord-raku-bot left, discord-raku-bot joined 03:03 rf left 03:17 discord-raku-bot left, discord-raku-bot joined 03:48 rf joined 03:55 rf left, rf joined 03:59 rf left
stevied Ran into something I couldn’t figure out: how do you get a method in a child class to call the parent class method? Tried various things like Parent.blah but nothing worked. 04:02
Couldn’t find any documentation for this from within the child class. Only from outside it. 04:03
jack9 `self.Parent`? 05:13
jaguart Docs - docs.raku.org/language/objects#Inheritance which you probably already found, but redispatch is covered by docs.raku.org/language/functions#sub_nextsame and friends - in other words, up-the-inheritance chain might not mean what you think, so you can `nextsame` or `nextwith` and trust Raku to find the next candidate 05:26
actually - start reading the block on redispatch here: docs.raku.org/language/functions#Re-dispatching 05:28
jack9 what is the equivalent of `if __name__ == '__main__'` in raku 05:31
like this?``` 05:33
if ($*PROGRAM-NAME eq $?FILE) {
main()
}
```
07:14 destroycomputers left 07:15 destroycomputers joined 07:17 elcaro joined 07:20 Heptite left 08:17 discord-raku-bot left, discord-raku-bot joined
gfldex <@842225380051124254> see: docs.raku.org/routine/MAIN 08:35
jaguart @jack9 - maybe you are looking for docs.raku.org/language/variables#&?ROUTINE e.g. 09:32
jaguart m: say &?ROUTINE.name
camelia ===SORRY!=== Error while compiling <tmp>
Undeclared name:
?ROUTINE used at line 1. Did you mean 'Routine'?
jaguart m: sub x { say &?ROUTINE.name }; x 09:34
camelia x
Nahita when you `import` a .py file, \_\_name\_\_ is set to the name of the file. when you run a .py file, it is set to "\_\_main\_\_" which allows to distinguish between running directly or importing 09:37
in Raku there are .raku files and .rakumod files so no need for this distinction perhaps 09:38
10:23 Heptite joined
Nemokosch I also couldn't figure that out 10:41
Lately I've seen a peculiar method call syntax that might help: self.ParentName::methodName(arguments) 10:59
jaguart isn't assuming an inheritance hierarchy kind of an anti-pattern? vs nextwith() using the .^mro? 11:17
Nemokosch I don't know if it is but it's better than nothing and that's how you would do it in C++ as well 11:18
jaguart but nextwith() isn't nothing? 11:19
Nemokosch it's definitely a smaller anti-pattern than tight coupling of inheritance and multiple dispatch which have very different goals and manifestations
it's not even guaranteed to do the right thing if you have multiple candidates
or better said, it probably won't do the right thing
jaguart yeah - which is why composition and roles are probably a better pattern 11:20
multi doesn't always work as I expect it - still having to train my brain
Nemokosch I consider this another design fault that might be too costly to do anything about it at this point
and it's not particularly well documented, the least to say 11:21
jaguart this -> being inheritance or composition?
Nemokosch the fact that method hierarchy has no proper mechanism for itself
it's just hacked into multiple dispatch
jaguart on the multi - is there a syntax for several types? e.g. `multi sub ( (Str | RatSr) $x )` or do I have to use a `where {}` 11:22
Nemokosch I wouldn't know for one. Leaning more towards "no". If there is anyone who has a good grasp of syntax at the moment it's probably lizmat, with all the RakuAST work she's been doing recently 11:23
jaguart oh interesting - I assmued that multis just sorted from most-to-least specific at each point in the ^mro...
I also assumed that the ^mro was going to be definable / changeable given it's current implementation 11:24
I will have to read up on Signature matching again 11:25
Nemokosch welp... literally, does anyone actually know the multiple dispatch resolution order?
jaguart it shows up in my grok tool - I just assumed the ordering from the MOP was meaningful 🤷‍♂️ 11:26
Nemokosch it probably does have a logic - but not a logic that would be properly written down anywhere, it seems 11:27
and since it needs to handle the self and the parameters, the question of specificity is quite severe 11:28
by the way. Sorry for hijacking the topic but I'm trying to estimate who would be willing to participate if we were to set up a "squashathon"
jaguart I'd give it a go - maybe if you gave me an example to do a couple of weeks ahead 11:29
is it like a nerdy zoomfest? or more tinezone distributed? 11:30
Nemokosch github.com/Raku/ecosystem-unbitrot this was the original version and this is my vision github.com/2colours/Raku-ideas/blo...0reborn.md
ideas are also welcome about the format
I'd be happy if we could get something real-time but yes, there needs to be some consideration of those who cannot be accessible in the same timeframe 11:31
I think it's more a prerequisite to even know if there are people willing to participate and under what terms 11:32
jaguart Yeah - I'd like to participate
Nemokosch a format idea I have is: let's collect modules with accessible authors that can "use some sprucing up", as the fez tool would say 11:34
then give it some time while people can choose and do some of the possible tasks
and there could be a kind of meetup where we can share results and partially also look for more stuff that might have come up 11:35
jaguart you could see about starting a new channel and use that to invite / exhort the troops - longer term, a place for newbies to start with small contributions and then sustain a fix a week etc. would be cook
if you have a feel for peeps level you could pumpmaster them towards ones that you think would suit them and that would be useful 11:36
Nemokosch yes, I think this is something that could even just go to a news feed, similarly to the Rakudo weeklies
jaguart I think you might be better with the social aspects of a channel 11:37
Nemokosch at this stage, my primary concern is to even get enough interest, to gather and estimate those people who are potential participants
jaguart maybe start your channel and send out a few invites... discord or irc work for me
Nemokosch depends on what channel. Discord is handy but I think it faces some (partially legitimate) rejection from the broader community, and so far I have bad experience with IRC - people just don't keep up with the different channels 11:38
regarding Discord - it occured to me yesterday that now that we have 3-4 regular learners showing up, even some casual voice sessions with whoever wants to discuss something about Raku, could increase cohesion 11:40
jaguart My feeling though is that you need the social interactive if you want steam. newsfeeds are low on my attention list, likewise reddit etc. 11:41
You could always use #raku I guess but the noise might annoy stalwarts 11:42
or spin something up on Matrix / FossToDon
or even twitter groups - my thought is it needs to be social casual, a place where it's ok to look a bit stupid 😶 11:44
in any case - post what you think, I'll sign up 11:45
Nemokosch Thank you for widening the perspective
Not gonna lie, I don't like Reddit, and especially Twitter and its "spin-off" Mastodon, and this does have political reasons. These seem to exist for a certain type of people that definitely isn't me. 11:46
tbh not just "isn't me", more like, we seem to be mutually incompatible 11:47
jaguart how about Matrix then?
you can host your own
11:49 discord-raku-bot left, discord-raku-bot joined
jaguart on branding - I like #raku-rejuve - kind of like a school for delinquent newbies, on the path to rakuzen. 11:50
unbitrot looks like something that relay's irc chat ;)
Nemokosch rakuzen haha 😄 11:51
yeah well, unbitrot existed in some form, for a couple of months or a year? I think there's stuff that can be reused but my vision is probably somewhat different
jaguart the Raku branding is way cool :) I have a soon to be released dist called Kaolin - clay for Raku (reusable components).
Nemokosch and perhaps the target itself shifted in the meantime 11:52
jaguart raku-juve - pronounced joo-vee - it's your penitence for asking too much in the beginner channel - before you attain your raku zen. 11:53
oh - no objection to unbitrot - pronounced oon-bye-trot? and it does have a lot of dust to tidy 11:54
Nemokosch un+bit+rot 11:55
jaguart off to bet at 11pm Melb
let me know and I'll sign up
Nemokosch okay, nighties
jaguart blushes and covers the cam...
11:57 discord-raku-bot left 11:58 discord-raku-bot joined
lizmat there is no such syntax atm, afaik 12:40
14:17 discord-raku-bot left 14:18 discord-raku-bot joined
jack9 what are the difference between blocks and functions 14:42
besides that blocks can be used in control flow statements
Nemokosch there is no return in block, that's for sure 14:58
I'm not 100% sure but you might be unable to use blocks for recursion
well, to be more accurate about "return" 15:02
if you do "return" inside a block, obviously it won't only terminate the block but the enclosing routine itself
15:06 discord-raku-bot left, discord-raku-bot joined 15:16 NemokoschKiwi joined 15:44 guifa_ joined 15:46 guifa left
jack9 m: 16:00
$fact = -> $n {
n == 0 ?? 1 !! $n * $fact($n - 1)
};
say $fact(10);
m:
my $fact = -> $n {
n == 0 ?? 1 !! $n * $fact($n - 1)
};
say $fact(10);
m:
my &fact = -> $n {
$n == 0 ?? 1 !! $n * fact($n - 1)
};
say $fact(10);
m: 16:01
my &fact = -> $n {
$n == 0 ?? 1 !! $n * fact($n - 1)
};
say fact(10);
you can do recursion in blocks smh
blocks are functions that don't return 16:03
like procs in ruby
Nemokosch that's a very interesting order of binding tbh 16:04
jack9 can i monkey patch a class like this?``` 16:11
class Str {
method charcount() {
self.comb.Bag
}
}
```
Nemokosch you might need some monkey pragma and the augment keyword 16:12
github.com/2colours/Int-polydiv/bl...iv.rakumod this is a module I created 16:16
not sure if it's best practice, perhaps not
but it does work
and is an example as such
it brings a polydiv method to Int
(a counterpart of polymod)
jack9 not even this?``` 16:19
method charcount(Str:D:) {}
```
Nemokosch well, do try it 😄 16:20
this is something I did once out of curiosity and pretty much that was it
jack9 m: 16:21
method charcount(Str:D:) {
self.comb.Bag
}
say 'aaabb'.charcount
m:
my method charcount(Str:D:) {
self.comb.Bag
}
say 'aaabb'.charcount
what
Nemokosch well yeah, why would a plain dumb 'aaabb' string know about your method 😅 16:22
it needs to be attached some way, that much is for sure
probably with the metamodel
jack9 because i defined a `charcount` method to its type
Nemokosch well, not to its type, it's not living in the Str class 16:23
you made a constraint for a method floating in the sky
jack9 anyways what does `:U` and `:D` stands for 16:24
up and down? 16:26
Nemokosch U is, hm unfortunate xd 16:31
"Undefined", except it's rather "indefinite" which is a distinct concept
And then D would be "definite" 16:32
D is, like, give me a valid instance of this type
U is like, give me a "hollow" instance - a type object - of this type 16:33
The default value is, unless specified, the type object
m: my Int $foo; say $foo; 16:34
This is the same thing that 5.WHAT would return, actually. A type object fits the type constraint 16:35
But it's not Int:D
And conversely, 5 is not an Int:U 16:36
If you subtyped Int, the type object of your subtype could be an Int:U
Does this make sense?
jack9 hmm 16:38
16:40 discord-raku-bot left, discord-raku-bot joined
Nemokosch 🙊 16:42
By the way, there can be times when we can discuss stuff like this in the voice chat, if you want. I think some things could be easier. 16:43
17:17 discord-raku-bot left, discord-raku-bot joined
stevied alright, my latest creation is relatively mature: github.com/sdondley/Directory 17:19
feedback welcome
I may add an iterator to it at some point in the future. I really need practice with iterators. they still kind of confuse me with that lazy loading stuff 17:20
p6steve <@563790557831495691> ... I came very close to using your %?RESOURCES module earlier (and then gave up on stupid resources anyway) 18:01
stevied which one was that, the one that lets you use a menu to navigate it? 18:02
18:02 raku-discord joined
raku-discord <4stevied> which one was that, the one that lets you use a menu to navigate it? 18:02
<4raku-bot> <raku-discord> <4stevied> which one was that, the one that lets you use a menu to navigate it?
<4raku-bot> <raku-discord> <4raku-bot> <raku-discord> <4stevied> which one was that, the one that lets you use a menu to navigate it?
<4raku-bot> <raku-discord> <4raku-bot> <raku-discord> <4raku-bot> <raku-discord> <4stevied> which one was that, the one that lets you use a menu to navigate it?
18:02 raku-discord left
stevied bot looks recursive 18:02
p6steve <@563790557831495691> ... out of curiousity, when would I reach for Directory (instead of plain dir
)?
yes that one 18:03
Zephyr apologies for that, it's been fixed
a new bridge was being tested and both bridges collided 18:04
stevied I created Directory mostly because I wanted to create a class with a `has Directory` attribute
and it would do some basic checking for me to make sure it wasn't already a file, etc.
p6steve oh - OK so it's for OO style stuff - cool!
stevied then I decided to throw in some extra goodies like tying it in to IO::Dir and File::Directory::Tree so it could do some useful stuff too out of the box. 18:05
so it's just a little more convenient that IO::Path, especially for newbs. 18:06
so it's just a little more convenient than IO::Path, especially for newbs.
p6steve nice!
stevied hides a layer of abstraction from them 18:08
p6steve I need a way to get a dir sample test_data files (csvs) in GH /bin/test_data through zef install and accessible as an abs or rel filepath string by both bundled scripts and newly written scripts after the install ... the docs on Module resources are less than awesome ... can you maybe point me to a minimal module that can do that? 18:11
right now I just do a git clone after the zef install and use the files from there 18:12
stevied I don't know about a module but I recently wrote a script that can be adapted to search all the tarballs in the entire REA repository 18:13
It just gets the listings inside the tarball but with some tweaks it can read the files, too
p6steve ok - my guess is that all I need is in Distribution::Resources::Menu so I will pick that apart next time 18:14
stevied ok, I misread your question
p6steve well I don't need to do the menu picker step 18:15
stevied if I understand you, what I would do is put that data in /resources directory. then install it over to the user's home directory 18:16
I don't think you need my module for that
you just have to know the path to the data 18:17
p6steve yeah - that's what I would like to do ... but I tried every combination of the docs and came up short ... do you know of a working example I could look at? 18:18
stevied you only need to copy to the home directory if you are modifying them
well, resources works just like a hash: off the top of my head you do something like `%*RESOURCES<path/to/file>` 18:19
my memory is a little fuzzy
docs.raku.org/language/variables#%?RESOURCES 18:20
have you seen that?
p6steve yes ... that's what I couldn't get to work for me ... but I see that Distribution::Resources::Menu has test resources and so I will be able to plunder that for a working example!! perfect... 18:22
stevied ok, cool. yeah, as far as I recally, you just throw your files and direcgtories into the /resources directory in your module 18:23
18:33 NemokoschKiwi left 18:48 josh24 joined 18:51 josh24 left 19:47 discord-raku-bot left 19:53 raku-discord joined
raku-discord <4stevied> I've got use Drectory in my script. Is there a way to determine which version of Directory the script is using? 19:57
<2Nahita> Directory.^ver maybe? idk 20:00
<4stevied> my script insists on using a stale version of the module. no idea where it's getting it from 20:01
<4stevied> oh, nvm. i'm using a different raku version in this shell. ugh 20:04
<4stevied> just lost about 2 friggin hours trying to figure out why tests were acting so weird. 20:06
20:13 raku-discord left, discord-raku-bot joined
gfldex This is a bridge test. 20:14
stevied I want to create a signature for the following scenario: if new is passed (), default $dir to '.'. If new is passed 'some string, I want $dir to be 'some string'. And if new is passed dir => '/some/dir', I want $dir to be 'some/dir' 21:00
so far, I have this: multi submethod BUILD(Str :$dir? = '')
I guess I just need multi submehotd BUILD(Str:D $dir) 21:02
gfldex @stevied Str:D(Numerical) is a better constraing in your case. 21:06
stevied ok, so I have this now: submethod BUILD(Str $dir-arg?, Str :$dir? = '', *%_ ()) and my $obj = Directory.new($dir-arg || $dir); just feels real awkward
hmm, what does that do? 21:07
gfldex It wont fail on literal numbers.
gfldex m: sub foo(Str:D $s) { dd $s }; foo(42); 21:08
camelia Type check failed in binding to parameter '$s'; expected Str but got Int (42)
in sub foo at <tmp> line 1
in block <unit> at <tmp> line 1
Nemokosch Str:D is a bit broken with coercions, for that matter
gfldex m: sub foo(Str:D(Numerical) $s) { dd $s }; foo(42);
camelia ===SORRY!=== Error while compiling <tmp>
Coercer is too complex. Only type objects, with optional type smileys, or empty parentheses, implying 'Any', are supported.
at <tmp>:1
------> sub foo(Str:D(Numerical)⏏ $s) { dd $s }; foo(42…
stevied huh
gfldex m: sub foo(Str(Numerical) $s) { dd $s }; foo(42);
camelia ===SORRY!=== Error while compiling <tmp>
Unable to parse expression in typename; couldn't find final ')' (corresponding starter was at line 1)
at <tmp>:1
------> sub foo(Str(⏏Numerical) $s) { dd $s }; foo(42);
Nahita Numerical doesn't exist though, does it
gfldex m: sub foo(Str:D(Numeric) $s) { dd $s }; foo(42); 21:09
camelia "42"
gfldex Raku's type system is so advanced, it is really easy to get wrong. :-> 21:10
21:11 rf joined
stevied all of it is tricky. a lot to keep track of 21:16
rf Anyone know how to tackle this, I reinstalled Linux yesterday and I'm seeing: Aborting due to fetch failure: Test::META:ver<0.0.19>:auth<zef:jonathanstowe>:api<1.0> when trying to install Test::META 21:24
lizmat feels like a connection issue, not a Raku issue ?
rf It's possible, but my internet is fine (at the moment), maybe my DNS is broken or something 21:25
lizmat can you install other modules ?
rf Yeah, I can pull everything but this one module 21:26
Hmm, it's failing on a different module now, maybe I am having internet problems. 21:28
lizmat afk& 21:29
rf Looks like Starlink (my ISP), is having issues. Well Humming-Birds new HTTPServer is delayed until that's fixed :L 21:36