00:09 deoac joined 00:38 deoac left 00:53 deoac joined 03:04 frost joined 03:16 frost left 03:35 frost joined 04:02 saint- joined 06:38 Oshawott left 06:39 Oshawott joined 06:51 SmokeMachine joined 06:55 crystalfrost[m] joined 07:49 dakkar joined
summer how do you test that a value is in a list? 09:20
I went with @.first(thing,:k).defined;
(Nil,).first(Nil,:k).defined; # True 09:22
but is there something nicer?
Nemokosch well, using (elem) (also known as ∈) is nicer, no? 09:26
m: my @values = (1 .. 10).roll: 5; dd @values; dd 1 (elem) @values; 09:28
let's try this a couple of times
dakkar or `$value ~~ @list`
(which may be faster if ∈ converts the list to a set internally…) 09:29
Nemokosch that doesn't seem alright 09:30
m: 1 ~~ (1, 2, 3) andthen .say
anyway, that doesn't work, and that's according to the specs
dakkar ooh! 09:31
works for ranges, not lists…
Nemokosch so yeah, what solutions do we have 09:32
we have .first
we have (elem)
dakkar docs.raku.org/routine/ACCEPTS#(Lis...od_ACCEPTS I was confused, thank you for checking
Nemokosch m: for ^10 { my @values = (1 .. 10).roll: 5; dd @values; dd @values ~~ (*, 1, *); }
meh, too bad 09:33
dakkar I'd say ∈ is the clearest option
summer hyperwhatevers for that approach
Nemokosch oh okay 09:34
m: for ^10 { my @values = (1 .. 10).roll: 5; dd @values; dd @values ~~ (**, 1, **); }
nice 🙂
so yeah, this is one with smartmatching
I want one with junctions as well
summer @.contains: 1; 😎 09:35
Nemokosch m: for ^10 { my @values = (1 .. 10).roll: 5; dd @values; dd so @values.any eqv 1; } 09:36
Well, then contains is the clearest option
Nahita nope, `.contains` doesn't work 09:37
summer @ ~~ /\b 1 \b/
Nahita redirects to (elem) :p 09:38
Nemokosch List inherits from Cool
so it's textual containment apparently
then yeah, backing off from contains
Then yeah, there is `first`, the junction way, `(elem)` and the smartmatch way 09:39
Nahita there's also `.grep(...).so`
junction one doesn't yet shourtcircuit, `grep` does, so i use that sometimes... 09:40
junction one doesn't yet shortcircuit, `grep` does, so i use that sometimes...
Nemokosch interesting
09:40 discord-raku-bot left, discord-raku-bot joined
Summer oh grep is lazy? 09:41
09:41 summer left
Nemokosch almost everything is lazy, or better said, is potentially lazy 09:42
grep, map, Z, X
the hyper meta-operator generates lists, on the other hand
Summer wait, is `so` la boolifying a lazy list lazy? 09:44
Nemokosch I think `so` returns True if there is at least one element 09:45
that seem pretty easy 09:46
Nahita yeah `so` makes it give a boolean result as soon as possible
the `any` equivalent is therefore directly transfarable to there; but for `all`, i visit De Morgan :\| 09:47
negate the predicate and put `.not` instead of so
lizmat all shortcuts as soon as it sees a False value
Nemokosch but then doesn't any do the same with True? 09:49
Nahita lizmat: but when i do `(0, |(1 xx 100_000_000)).all.so` it hangs 09:51
lizmat that's probably in the *creation* of the Junction 09:52
m: (0, |(1 xx 100_000_000)).all
camelia (timeout) 09:53
Nahita `(0, |(1 xx 100_000_000)).grep(* != 0).not` is instant on the other hand.
2022.04 i tried; is the behaviour on some newer versions?
yes that's what i meant by not shortcircuiting...
Nemokosch timeout 😬
lizmat because then you don't have a Junction
the .grep treats it as a lazy list 09:54
m: say (0, |(1 xx 100_000_000)).first(0)
camelia 0
lizmat m: say (0, |(1 xx 100_000_000)).first(1)
camelia 1
lizmat m: say (0, |(1 xx 100_000_000)).first(1, :k)
camelia 1
lizmat m: say (0,0, |(1 xx 100_000_000)).first(1, :k)
camelia 2
Nahita yes; my desire is Junction seeing the `.so` or `.not` upon it and exhibit shortcircuit behaviour 09:55
Nemokosch It's obvious why `(0, |(1 xx 100_000_000)).grep(* != 0).not` is fast
Nahita like `grep` does
Nemokosch it's also "obvious" why it's not obvious to do the same with junctions
but in order to make them worth considering, they still should somehow mimic the lazy behavior
lizmat well, at this point in time, Junction creation is *not* lazy
Nemokosch it's too much of a drawback that they basically build a list
lizmat feels a problem-solving issue coming up :-) 09:56
Nemokosch not quite list but an in-memory eager data structure 09:57
like yes, the whole language is built around making everything as lazy as possible
Nahita an observation: e.g., Python persons are bragging (rightly) about their `all` and `any` combined with generator expressions to do lazy and readable things super fast 09:58
i matched that behaviour with `.grep(...).so` or DeMorganized version of this in Raku 09:59
Summer ∈ is p fast
Nemokosch okay but all and any are essentially like grep in Python
Kaiepi `all`, `any`, etc. are lazy in haskell
Nemokosch anyone who knows Haskell
ACTION looks around
Kaiepi sometimes there's someone 😉
Nahita but that's a specific predicate, isn't it... 10:00
Nemokosch anyway, the finish the thought - if we want junctions to be anything but a cute syntax sugar, they should also be lazy as much as possible
there are no ordering constraints, like with hyper 10:01
Nahita not quite?
Nemokosch yep, pretty much
Nahita what's "grep in Python"?
Nemokosch filter
or Array#some and Array#every in JS
Nahita whose "all and any" are you talking about? Raku's?
Nemokosch except for iterables, not arrays
Summer 💪 first, first 10:02
Nemokosch all in Python is like map, filter and all these 10:03
just an iterable management function
in Raku, it's completely different
Nahita i see now
Kaiepi try `[&&]`
Nemokosch one could create a function or method like that in Raku as well and then it would be fast by the same means
Nahita all i'm saying is, it'd be cool junctions shortcircuited; that's all
Nemokosch one could perhaps give them similar names
which is fair enough 10:04
I'm just saying that this is indeed a nontrivial thing to achieve, as desirable as it might be 10:05
lizmat well... Junctions go deep into the guts 10:06
it's basically the only core class that doesn't inherit from Any
which is how they can work transparently at all levels of the code 10:07
Summer `cat reduce`
END say now - INIT now;
sink for ^10_000 { [&&] (^100).roll(50).map(*==0) }```
`$ raku reduce`
I should be using a benchmarker really
Nemokosch well, both things are true at the same time. Junctions are a special design that need special care when implementing - and regardless, it's a major drawback if they have to eagerly fetch all data in order to be constructed and start any sort of evaluation 10:09
Summer if I want to have a stable random, its just srand(0) right? 10:17
~~if I want to have a stable random, its just srand(0) right?~~ yep 10:18
Nemokosch what makes it stable? 10:22
Summer sets it to a given seed, I'm going with the assumption that anything using random is deterministic after that, if not then I'm going to cry 10:24
Nemokosch 😄
well it should
Summer a `$*RANDOM` would be neat to be able to change the rng 10:26
10:30 gfldex left, discord-raku-bot left 12:20 jgaz joined 13:09 frost left
lizmat and yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2022/08/01/2022-...merelease/ 13:44
14:23 human-blip left 15:05 human-blip joined 15:30 jgaz left 15:31 jgaz joined 16:12 guifa joined 16:15 guifa_ left 16:38 dakkar left 16:41 discord-raku-bot joined 16:54 deoac left 17:02 sivoais left 17:15 sivoais joined 17:31 sivoais left 17:39 sivoais joined
gfldex `Num.rand` calls `nqp::rand_n`. If you want to change the RNG, wrapping `Num.rand` should suffice. 17:43
21:40 mjgardner_ joined 21:43 lucs_ joined 21:48 lucs left, mjgardner left, mjgardner_ is now known as mjgardner
Jodi hello, im super new to raku 22:40
was wondering how I can remove a range from a string, i.e
in python I can just do `new = old[min : max]`
hello, im super new to raku 22:42
was wondering how I can remove a range from a string, i.e
in python I can just do something like `new = old[0: start:] + old[stop + 1::]`
nvm figured it out 22:55
its just
my $new = $old.substr(0 .. $START) ~ $old.substr($STOP .. *)
guifa Hi jodi, that works, but you can also do 23:39
$old.substr-rw($START, $STOP - $START) = ''; 23:41
although that mutates the original
Also, with your solution (which is the one I'd use), you can simplify the code a bit 23:42
my $new = $old.substr(^$START) ~ $old.substr($STOP);
eh that might be an off by one on start, too lazy to double check lol 23:43
Jodi thanks 23:50
what does the `^` mean though? is that a regex thing