00:17 SrainUser joined
SrainUser I passed an array of strings to a field of an object, then I printed that array from a method within that class, but the array printed is empty. Is there a certain syntax for accessing array attributes within classes? 00:26
guifa 5rainUser: not really, it sounds like for some reason it's not binding during creation 00:29
can you post your code?
m: class Foo { has @.text; method print { say @!text.join } }; my $foo = Foo.new(text => <a b c d>); $foo.print 00:30
camelia abcd
SrainUser class Class { 00:31
has @!arr;
method yo() {
say @!arr;
}
}
my @array = ["hello", "goodbye", "what"];
my $example = Class.new(
arr => @array
);
$example.yo();
ab5tract @!arr is defined as a private instance variable 00:32
You can set it manually via a method new / submethod BUILD / submethod TWEAK
guifa just change it to has @.arr (will make it by default settable via .new) or you can also use `has @!arr is built`, or you can do the adjustments as ab5tract notes in BUILD or TWEAK 00:33
ab5tract but if you define it as @.arr it becomes public and the default new / BUILD implementations will make it work as you currently are creating the $example object
SrainUser Thanks! 00:34
00:43 tejr left
SrainUser I this a bug?: 01:40
my @array1; 01:41
my @array2 = [["hello", "goodbye"], True];
@array1 = @array2[0];
say @array1;
Prints a list inside a list.
01:44 Aedil joined
SrainUser A temporary solution is: 01:59
my @array1;
my @array2 = [["hello", "goodbye"], True];
for @array2[0][0..*] -> $item {
@array1.push($item);
}
say @array1;
phogg I was porting a Perl module today and found myself missing wantarray. What do people do instead? 02:06
02:06 guifa left 02:09 guifa joined
SmokeMachine m: my @array1; my @array2 [[“hello”, “goodbye”], True]; @array1 =@array2[0]; say @array1 02:21
camelia ===SORRY!===
Signatures as constraints on variables not yet implemented. Sorry.
at <tmp>:1
------> my @array2 [[“hello”, “goodbye”], True]<HERE>; @array1 =@array2[0]; say @array1
Other potential difficulties:
Literal values in s…
SmokeMachine m: my @array1; my @array2 = [[“hello”, “goodbye”], True]; @array1 =@array2[0]; say @array1 02:22
camelia [[hello goodbye]]
SmokeMachine That looks correct to me…
m: my @array1; my @array2 = [[“hello”, “goodbye”], True]; @array1 := @array2[0]; say @array1 # SrainUser is this what you want? 02:24
camelia [hello goodbye]
SmokeMachine m: my @array1; my @array2 = [[“hello”, “goodbye”], True]; @array1 = |@array2[0]; say @array1 # Or maybe this? 02:25
camelia [hello goodbye]
SmokeMachine m: my @array1; my @array2 = [[“hello”, “goodbye”], True]; @array1 = @array2[0][]; say @array1 # not sure about this one…
camelia [hello goodbye]
SrainUser Yes, although why would it not work with the assignment operator? 02:27
SmokeMachine SrainUser: @array2[0] returns an item that in this case is an array, so it’s put as an item inside the other array 02:28
m: my @array = [[“bla”, “ble”], True]; dd @array[0]; dd @array[0][] 02:29
camelia @array = $["bla", "ble"]
["bla", "ble"]
SmokeMachine SrainUser: compare the difference: ^^ 02:30
ab5tract In other words, @foo[] is a slice 02:31
m: my @a = [1,2], True; dd @a[0], @a[0][], @a[0][*] 02:33
camelia @a = $[1, 2]
[1, 2]
(1, 2)
SmokeMachine phogg: what was the case you needed the wantarray?
SrainUser That is rediculous. I don't want a slice. I want the first element of the array. 02:40
phogg SmokeMachine: a method returning a list of words or the same list joined on a string (in a scalar context)
it would be easy if return type was used as part of method dispatch 02:41
02:46 ming joined 02:47 hulk joined 02:48 kylese left
SmokeMachine What about something like `return <bla ble ble> but role { method Str { $.join } }`? 02:59
m: sub a { <bla ble ble> but role { method Str { $.join } } }; say a; say “{ a }” 03:00
camelia (bla ble ble)
blableble
SmokeMachine phogg: ^^ 03:01
timo SrainUser: when stuff is put into an array, every entry gets a scalar container, so when you take the first element out of @array2 you're getting back a single item that you're assigning to @array1. you'll want to .list, or <> to decont, or [], or something else to get rid of the container that protects it from being iterated while assigning to @array1. another thing you can do is bind instead of 03:08
assign to @array1 with the := operator
03:10 ming left
phogg SmokeMachine: yeah that was what I was thinking, but it seems kind of hacky. Looking for best practice advice. 03:11
I will have a handful of similar methods and I don't like the look of all of that inline. When does it become reasonable to make a custom type instead? 03:13
03:15 hulk left
SrainUser So it doesn't copy the value of the item, it points to the item in the container (or array), or it creates a new container for the value of the item. 03:15
03:15 kylese joined 03:17 yewscion joined
timo when you bind, yes you'll end up with the same item living inside @array2 that now has the "extra name" @array1 03:20
when you assign an array to an array variable, you'll get fresh scalar containers for the elements in it, which is important to know
i guess it would be quite surprising if it were different 03:21
03:23 stackspace left
ab5tract phogg: my recommendation would be to write two multi candidates and use a named argument to distinguish between the two 03:27
phogg if I could I would 03:28
ab5tract What do you mean?
03:28 Aedil left
phogg Doing that would change the API. I'm trying not to. 03:28
if I'm changing the API I can just use different method names 03:29
ab5tract Sure, doing that versus adding an adverb is a matter of taste 03:30
I guess I don’t know enough about your use case to provide valuable input 03:31
That said, if I were stuck with an API I couldn’t change, I would probably just wrap it with a local sub. 03:33
Either way, if the call site knows what it wants, it can call a specialized version of the code to get what it expects. wantarray has been an anti pattern for a 03:34
03:45 Aedil joined
ab5tract * few decades already 03:52
04:40 yewscion left 04:42 yewscion joined 04:45 yewscion left, yewscion_ joined 04:48 yewscion_ left 04:49 yewscion joined 05:22 SrainUser left 06:15 yewscion left 06:16 guifa left 07:16 ulipink joined 07:17 ulipink left 07:18 ulipink joined, ulipink left, ulipink joined, Sgeo left 08:22 sorenson_ joined 08:23 sorenson left, sorenson_ is now known as sorenson 09:11 dakkar joined 09:54 sena_kun joined
patrickb Is there a module to set up an ad hoc ecosystem from a local folder with a few module repos checked out? 11:48
I have a few interdependent modules I'd like to test installation of. So only installing one of them won't work as it depends on the others. 11:49
lizmat I usually bump the versions of the modules locally, and just install them locally when testing that one module 11:52
one of the advent posts described a content management system 11:53
raku-advent.blog/2024/12/13/day-13...ributions/ 11:54
patrickb But I'm concerned about the dependency specifications those I want to test. So I want zef to be able to resolve and install the dependencies. 11:57
It should be pretty trivial to write a ecosystem generator, right? I basically only have to copy together all the META6.json files, right? 11:59
antononcube Is using Docker an alternative ? Generating a dockerfile with the required packages and versions should be easy. 12:01
lizmat hmmm.. create a CURFS with a standard prefix, install the dependencies to that with zef --to (I believe), and then use that standard prefix with -I always ?
12:03 ulipink left
patrickb Nah, I can just install to the default locations. That's fine. I'm only interested in the installation process itself, not the resulting installation. 12:04
I think I'll just hack something up. Thanks everyone for chiming in.
12:52 ulipink joined, ulipink left, ulipink joined 12:57 ulipink left 13:22 guifa joined
[Coke] docs.raku.org/syntax/wantarray%20-%20perlfunc 13:26
13:59 ulipink joined, guifa left 14:01 guifa joined 14:04 ulipink left 14:07 hudo_ left 14:08 hudo joined, hudo_ joined 14:33 ulipink joined, ulipink left, ulipink joined 14:38 ulipink left 14:53 ulipink joined, ulipink left, ulipink joined 14:57 ulipink left 15:28 ulipink joined 15:33 ulipink left 16:07 ulipink joined 16:12 ulipink left 16:15 Manifest0 joined 16:41 ulipink joined 16:46 ulipink left 16:47 SrainUser joined 16:48 SrainUser left 16:52 Simba111 joined 17:28 yewscion joined 17:29 ulipink joined, ulipink left, ulipink joined 17:33 ulipink left 17:37 dakkar left
ugexe patrickb: I would look at these 17:44
github.com/ugexe/zef/blob/main/xt/...y.rakutest
github.com/ugexe/zef/blob/main/lib...od#L16-L34
The slightly harder part is getting `zef` to see those type of custom repositories, but if the above links make sense it should be trivial for you to modify the synopsis code here to use instead `bin/zef`: github.com/ugexe/zef/blob/080304e1...od#L26-L49
alternatively you can add a Zef::Repository::Ecosystems entry in the config file similar to the fez one, and point the mirror at a file (instead of a url) that is an array of hash (an array of meta6.json data) 17:45
there is also github.com/tony-o/raku-zeco being worked on (or rather more like waiting for me to review it more) which spins up a docker container with fez ecosystem software which could then be used as the mirror/url for an ecosystem in the zef config 17:49
lizmat and there's raku.land/zef:lizmat/Zef::Configuration :-) 17:58
18:03 ulipink joined 18:09 ulipink left 18:24 ulipink joined 18:28 ulipink left 18:43 ulipink joined, ulipink left, ulipink joined
patrickb ugexe: Thanks for the hints! I plan to create a tiny script that copies together a valid ecosystem.json and then add an entry to the zef config. 18:45
18:47 ulipink left 18:50 msiism joined
msiism Is there a way to customize the Rakudo prompt? 18:51
jdv the repl one? 18:53
msiism Yes. 18:54
I'd like to disable the line count.
jdv idk. override interactive_prompt? 18:58
github.com/rakudo/rakudo/blob/6363...kumod#L411
msiism Hm… I was more looking for run-time configurability. 19:01
It not all that tragic to have the line numbers in there. It's just that I often copy examples from the REPL into my notes and then have to filter them out. 19:03
19:03 ulipink joined
antononcube Hmmm... this might mean, that you have to use different REPL. 19:03
For example, notebooks. 19:04
jdv you could submit a feature req or a pr but idk if itd get in
[Coke] lizmat did some work to help customize the REPL previously - I think having something pull from an env var there might be worth considering.
msiism Yeah, that's what I was thinking. 19:05
lizmat is thinking waterbed 19:06
[Coke] msiism: github.com/Raku/problem-solving ?
lizmat the line numbers (actually result numbers) were added to be able to refer to previous values by means of $*0, $*1, etc 19:07
antononcube For example, "Jupyter::Chatbook" notebooks take (i) a Raku file with initialization code, and (ii) a JSON file with LLM personas.
19:07 ulipink left
msiism [Coke]: Okay, I'll see what I can do. 19:07
antononcube And notebooks support In[1] and Out[12] type of references for previous inputs and outputs. 19:08
msiism lizmat: I see.
[Coke] so that might correspond to %*ENV{RAKU_REPL_PROMPT} = '[{$N}] > ' as a default?
lizmat [Coke]: possibly yes, issue / PR welcome
[Coke] msiism: if you open that problem solving ticket, I can see about putting together a PR 19:09
msiism Okay, nice. I won't be able to get to that today, but withing the next days should be possible. 19:10
lizmat ++msiism
antononcube Using Raku file with initialization code is not in "Jupyter::Kernel" yet -- I have to file an issues so bduggan would implement it. 19:11
19:37 Aedil left 19:39 ulipink joined 19:44 ulipink left 20:26 yewscion left 20:30 ulipink joined 20:34 ulipink left 20:41 yewscion joined
msiism If `||` "returns the first argument that evaluates to True in Boolean context" and "otherwise returns the last argument", why does `'Alpha' || 'Beta'` return `Alpha`? 20:49
ugexe because 'Alpha' evaluates to True in Boolean context
m: say so 'Alpha'
camelia True
msiism Okay, but why? 20:50
lizmat because that was so decided 20:51
ugexe docs.raku.org/syntax/Boolean%20context
"In general, non-zero, non-empty will be converted to True; zero or empty will be equivalent to False. But .so can be defined to return any Boolean value we want, so this is just a rule of thumb."
msiism I see, thanks.
21:04 ulipink joined
Geth ecosystem/main: b3f73ae412 | (Elizabeth Mattijsen)++ | META.list
Remove two of Kamil Kułaga's distributions

They have bitrotted, and one of them interferes with the new Test::Coverage distribution		 21:07
21:08 ulipink left 21:39 ulipink joined 21:43 ulipink left 21:57 Xliff joined
Xliff m: my \💩 = 69; say 💩 21:57
camelia ===SORRY!=== Error while compiling <tmp>
Malformed my
at <tmp>:1
------> my<HERE> \💩 = 69; say 💩
Xliff O_o .... NOoooooooo! Whyyyy! My hopes for this years April Fool's is dashed! 21:58
No. Seriously. Why is that symbol not allowed?
I mean, I know it loosely translates to "My crap is pr0n" but c'mon!!! 22:00
guifa m: my term:<💩> = 69; say 💩 22:02
camelia ===SORRY!===
Type 'term:<💩>' is not declared
at <tmp>:1
------> my term:<💩><HERE> = 69; say 💩
Malformed my
at <tmp>:1
------> my term:<<HERE>💩> = 69; say 💩
guifa my sub term:<💩> = {69}; say 💩
m: my sub term:<💩> = {69}; say 💩
camelia ===SORRY!=== Error while compiling <tmp>
Missing block
at <tmp>:1
------> my sub term:<💩><HERE> = {69}; say 💩
expecting any of:
new name to be defined
guifa m: my sub term:<💩> {69}; say 💩
camelia 69
guifa there you go
Xliff \o 22:05
guifa: \o/ You've saved my 4/1 # I muzt troll!
guifa I guess you could just do sub term without the my too 22:06
depends on the scope :)
Xliff Yah. Thanks!
22:28 msiism left 22:34 kjp left 22:36 kjp joined 22:38 kjp left, kjp joined 22:47 ulipink joined 22:51 ulipink left 23:19 sena_kun left 23:22 ulipink joined, ulipink left, ulipink joined 23:26 ulipink left 23:41 Sgeo joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!