🦋 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.
00:24 Summer left, itaipu left, Summer joined 00:37 itaipu joined 00:50 Manifest0 left 00:55 Summer left, Summer joined 01:25 Summer left 01:26 Summer joined 01:28 bdju joined 01:54 oodani left 01:55 oodani joined 01:56 Summer left 01:57 Summer joined 02:20 MasterDuke joined 02:32 MasterDuke left 02:33 hulk joined 02:34 kylese left 03:15 hulk left, kylese joined 03:21 Summer left, Summer joined 03:31 andinus left, andinus joined 03:52 Summer left, Summer joined 04:22 Summer left 04:23 Summer joined 04:41 vrurg left 04:53 Summer left, Summer joined 05:24 Summer left, Summer joined 06:34 constxqt_ left 07:11 verzit_ joined 08:01 samebchase left 08:03 verzit_ left 09:02 snonux left 09:04 Manifest0 joined 09:26 samebchase joined 09:35 snonux joined 10:37 sena_kun joined 11:13 verzit_ joined 11:14 derpydoo joined 11:39 constxqt_ joined 11:46 derpydoo left 11:51 derpydoo joined 11:59 Sgeo left 12:30 verzit_ left 13:44 epony left 13:49 verzit_ joined 14:19 vrurg joined, verzit_ left 14:54 izel joined
izel hi isn't the following contradictory 14:56
m: my %h{List} = (0, 0), 3; %h{%h.keys.head}.say
camelia Type check failed in binding to parameter 'key'; expected List but got Int (0)
in block <unit> at <tmp> line 1
izel may i never reach values of %h by manual key 14:57
ugexe something seems off about that. it works when using List.new: but not with .List 15:08
m: my %h{List} = (0, 0), 3; %h{%h.keys.head.List}.say
camelia Type check failed in binding to parameter 'key'; expected List but got Int (0)
in block <unit> at <tmp> line 1
ugexe m: my %h{List} = (0, 0), 3; %h{List.new: %h.keys.head}.say
camelia (3)
lizmat probably better with my %h{Positional} ? 15:12
izel Positional gives the same error 15:19
well except expected Positional instead of List
List.new version gives back a List as a value not single value but thanks 15:20
i'm using Tuple it seems to work
except i cannot `say %h`
it errors with "This representation (VMArray) does not support attribute storage (for type Tuple)"
nemokosch what even is "type Tuple"? 15:29
izel google it 15:30
15:34 vrurg_ joined
nemokosch there is no such thing in Raku 15:36
15:36 vrurg left 15:37 vrurg_ is now known as vrurg 15:53 MasterDuke joined
lizmat nemokosch probably raku.land/zef:lizmat/Tuple 16:17
roguerakudev Is it possible to hook into object construction before accessors for attributes are generated? I thought BUILD would do it, but doing .set_rw on an attribute in BUILD doesn't make the generated accessor return-rw 16:24
For context, I want to reuse the same event class for both sync and async events, and for the former case I want to make any publicly-readable attributes also publicly-writeable by mixing in a role 16:26
16:26 izel left
might need to do traits on the members instead 16:30
trait would be better as it would give more fine-grained control, but I'm not sure how that would look, because you don't know until a specific instance is created whether it's for sync or async, and thus whether an rw accessor should be generated 16:33
actually that's not even good enough, because the object has to be constructed before the mixin is applied anyway 16:36
nvm you can mix in before calling new 16:37
16:37 derpydoo left
lizmat BUILD and TWEAK are runtime, you're really talking about compile (composition) time 16:38
roguerakudev so X but Y is compile-time? 16:44
or for this purpose, I guess the question is will it be if X and Y are known at compile time 16:45
or maybe I don't need to worry about that, if I just add_method in the trait_mod:<is> for my custom trait, then have the method itself look at whether the role in question got mixed in 16:46
I may have rubber-ducked into a solution :P 16:47
16:49 epony joined
lizmat roguerakudev please post a link to your solution :-) more eyes on code help! 16:50
16:57 Summer left
roguerakudev Not quite there yet... need to figure out how to get the ClassHOW for the class the Attribute is part of, so I can .add_method 17:05
lizmat call .^add_method ? 17:06
roguerakudev well, on what is the question... in multi trait_mod:<is>(Attribute $attr, :$sync-rw!) all I have access to is the Attribute itself 17:09
17:24 dpb left
I could pass $?CLASS in when using the trait, but that feels icky 17:41
17:48 derpydoo joined
guifa Attribute has a .package() method 18:08
I use that in Intl::CLDR for adding aliases via traits: github.com/alabamenhu/Intl-CLDR/bl...akumod#L38 18:09
roguerakudev Aha, finally got it: gist.github.com/landyacht/1b35ed25...41cff69046 18:41
It took some digging through Rakudo source and many rounds of trial and error, but the result is actually not all that messy 18:43
quite happy with this
paging lizmat re: "more eyes on code help" 18:45
guifa so just a heads up: wrapping can be a bit slow. You might consider actually generating the method yourself. I don't do a whole lot with that but lizmat I know has some better experience 18:47
roguerakudev Hm, good point - I didn't go that route b/c the code for generating the method inside Attribute::compose is rather long and uses nqp 18:49
guifa that's fair
the nice thing is unless it becomes a bottleneck for whatever you're doing, what you have works great
roguerakudev But I do want this software I'm writing to be pretty fast, so wrap being slow is a real concern
I wonder if it's any faster to straight up remove the generated method from the table and add back in one that calls that method in a wrap-y way without actually using .wrap 18:50
lizmat roguerakudev Am in the middle of 3 things at the moment, and 4 would be at least 1 too many... will look in an hour or so
roguerakudev I don't know how much hackage the "remove the method" part entails though :P 18:51
guifa is going to try something quickly 18:53
antononcube @lizmat I wrote the Jupyter notebook with article. Since the images are two big I have to some image export and resizing. I hope to have the Markdwon version within few hours.
lizmat antononcube WP is pretty good at resizing 18:54
guifa roguerakudev: how about something like this? 19:00
tio.run/##XY9NbsJADIX3c4qHlAVIkAWq...A9K3sq77AA
19:00 verzit_ joined
roguerakudev Oh right, Attribute's composition skips generating if there's already a method there 19:02
so mine is overcomplicated haha 19:03
nice, thanks
I also didn't realize get_value would return something writeable 19:04
Is there a reason you're doing set_rw?
guifa it might not be necessary, I just noticed you had it. I'd assume there's some internal machinations to set up the writability part 19:06
but it actually seems to work okay without it so
roguerakudev I set it because Attribute's compose method looks at it, and I wanted the method it generated to return-rw 19:07
since I figured trying to return-rw something that was originally returned as a readonly container wouldn't work
Awesome, your solution structure works for my use case as well 19:10
19:12 epony left 19:14 epony joined
lizmat it appears roguerakudev no longer needs my eyes :-) 19:27
Geth advent/main: 4736296ce5 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | raku-advent-2023/authors.md
move vushu's post to the 19th
 19:44
20:16 Sgeo joined
librasteve my 3rd is ready for Scheduling btw 20:56
lizmat librasteve cool, will attend to it nnow 20:58
librasteve ok if you want to shift the Day 21:10
lizmat ok, will make it the 20th then :-)
El_Che hi
librasteve k
yo 21:11
Geth advent/main: 90e529f95a | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | raku-advent-2023/authors.md
Move craggy to the 20th
 21:19
22:30 MasterDuke left 23:10 verzit_ left 23:18 sena_kun left

Powered by Cro and the Raku® Programming Language.

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

Thank you!