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.
01:09 razetime joined 03:32 camelia left 03:37 camelia joined 03:48 MasterDuke left 04:11 razetime left 04:32 razetime joined 04:42 teatime left, teatime joined 04:44 teatime left, teatime joined 05:06 teatwo joined 05:08 teatwo left, teatwo joined 05:09 teatime left 05:40 siavash joined 07:42 habere-et-disper joined
habere-et-disper I grok : 07:49
m : elems gather .take if Date.new( year => 2013, month => $_, day => 13 ).day-of-week == 5 for 1..12
But I really wanted to write :
m : elems gather .take if Date.new( year => 2023, month => (1..12).any, day => 13 ).day-of-week == 5
m: elems gather .take if Date.new( year => 2013, month => $_, day => 13 ).day-of-week == 5 for 1..12
camelia ( no output ) 07:50
habere-et-disper m: elems gather .take if Date.new( year => 2023, month => (1..12).any, day => 13 ).day-of-week == 5
camelia Cannot call Date.new with these named parameters: month year day
in block <unit> at <tmp> line 1
lizmat (1..12).map( { my $dt := Date.new( year => 2013, month => $_, day => 13 ); $dt if .day-of-week == 5 } ) 07:51
habere-et-disper Is the (1..12).any my misunderstanding of Junctions or only that Date does not support them ? 07:54
lizmat 1..12 is a Range, not a Junction 07:59
ah, .any should read better :-)
not sure what is going on there, but the Positional interface does work 08:01
m: dd Date.new(2023, (1..12).any, 13)
camelia any(Date.new(2023,1,13), Date.new(2023,2,13), Date.new(2023,3,13), Date.new(2023,4,13), Date.new(2023,5,13), Date.new(2023,6,13), Date.new(2023,7,13), Date.new(2023,8,13), Date.new(2023,9,13), Date.new(2023,10,13), Date.new(2023,11,13), Date.new(2023,…
habere-et-disper Thank you lizmat  That is exactly how I wanted to write it ! 08:09
m: say elems gather .take if .day-of-week == 5 for Date.new( 2023, (1..12).any, 13 )
camelia 2
habere-et-disper 😀
08:09 dakkar joined
lizmat :-) 08:11
although the use of "for" there is a bit iffy :-) 08:12
nemokosch I didn't even know junctions are iterable, is this a public interface? 08:18
lizmat they aren't iterable, that's why I said the "for" was "iffy"
m: .say if $_ %% 2 for 42 08:19
camelia 42
lizmat m: .say if $_ %% 2 for 43
camelia ( no output )
lizmat m: .say if $_ %% 2 for 42 |43 08:20
camelia any(42, 43)
lizmat m: .say if $_ %% 2 for (42 |43)
camelia any(42, 43)
nemokosch Heh 08:21
lizmat m: .say gather .take if $_ %% 2 for 42 |43
camelia ===SORRY!=== Error while compiling <tmp>
Two terms in a row
at <tmp>:1
------> .say⏏ gather .take if $_ %% 2 for 42 |43
expecting any of:
infix
infix stopper
statement end
statement mo…
lizmat m: say gather .take if $_ %% 2 for 42 |43
camelia (42)
lizmat m: say gather .take if $_ %% 2 for 42 | 43 | 44 08:22
camelia (42 44)
nemokosch So where did the junction ultimately distribute into its eigenvalues? 08:23
lizmat the conditional .take
nemokosch m: say gather (1|2|3).take 08:25
Raku eval (any(1, 2, 3))
nemokosch Welp, i tried
08:26 razetime left
habere-et-disper m: say ( Date.new( 2023, (1..12).any, 13 ).day-of-week == 5 ).iterator 08:26
camelia List::Todo.new
lizmat I guess this is the reason: 08:28
multi method iterator(Junction:D:) {
# If we're asked for an iterator, we should really give one rather than
# auto-thread over the `iterator` method. Otherwise we get decidedly
# confusing outcomes from things that do `.iterator` and then expect it
# to follow the iterator API.
list(self).iterator
}
interesting... :-)
nemokosch m: (1|2|3).list.say 08:29
Raku eval any((1), (2), (3))
nemokosch Lmao
lizmat github.com/rakudo/rakudo/commit/bceaeb1a9a 08:30
nemokosch m: list(1|2|3).say
Raku eval (any(1, 2, 3))
nemokosch Okay but how does this explain the countability in that snippet? 08:31
Okay, the if succeeds, the junction passes on, right? What do I miss? 08:32
lizmat m: say gather .take if .day-of-week == 5 for Date.new( 2023, (1..12).any, 13 )
camelia (2023-01-13 2023-10-13)
lizmat the gather returns a Seq, and elems just counts the number of elems in it
nemokosch But why isn't that Seq only the one junction? 08:33
lizmat so I guess the "for" wasn't that "iffy" after all
because the "for" calls .iterator on the Junction under the hood, and thus iterates
nemokosch Didn't that commit just say that it will iterate as one? 08:34
lizmat no, it said it will iterate, because making .iterator return a Junction would seriously be confusing to a lot of the internals 08:35
nemokosch Not a Junction but an iterator that produces a Junction, like with scalars 08:36
But either way, NOT the eigenvalues
That's what the commit message states
at least this dragged me out of bed 😂 08:38
m: .say for (1|2|3)
Raku eval any(1, 2, 3)
nemokosch yeah
lizmat hmmmm so the "for" *is* iffy after all 08:39
hmmm
nemokosch m: say gather .take if .day-of-week == 5 given Date.new( 2023, (1..12).any, 13 )
Raku eval (any(2023-01-13, 2023-02-13, 2023-03-13, 2023-04-13, 2023-05-13, 2023-06-13, 2023-07-13, 2023-08-13, 2023-09-13, 2023-10-13, 2023-11-13, 2023-12-13))
nemokosch m: say gather .take if .day-of-week == 5 for Date.new( 2023, (1..12).any, 13 ) 08:40
Raku eval (2023-01-13 2023-10-13)
nemokosch but it still isn't for some reason xd
lizmat ah.. the list(self) is a bit of a red-herring
nemokosch what does it produce? 08:41
lizmat multi sub list(+l) { l }
using the single arg rule, it returns the arg 08:42
hmmm...
ah, but that autothreads
08:44 razetime joined
lizmat m: dd list(1|2).iterator.pull-one 08:44
camelia any(1, 2)
nemokosch m: dd gather list(1|2).iterator.pull-one.take
Raku eval (any(1, 2),).Seq
nemokosch what really bothers me is that I can't reproduce or even imagine what happened with that for 08:45
lizmat m: my $i := list(1|2).iterator; dd $i.pull-one for ^2
camelia any(1, 2)
IterationEnd
lizmat so it produces an iterator that produces 1 element, itself
so the use of "for" *is* really "iffy" in the original case 08:46
nemokosch that may be true but then wouldn't for be equivalent with given? 08:48
lizmat I guess the for deconts 08:49
nemokosch which part takes that into account, of the if, the take and the gather? 08:50
m: my \test = (1|2|3); say gather test.take if test %% 3; 08:52
Raku eval (any(1, 2, 3))
lizmat hmmm... given will just topicalize, whereas the for will return an iterator
meh.. this is tricky :-)
nemokosch m: $_ := (1|2|3); say gather .take if $_ %% 3; 08:53
Raku eval (any(1, 2, 3))
nemokosch m: say gather .take for (1|2|3)
Raku eval (1 2 3)
nemokosch m: say .say for (1|2|3) 08:54
Raku eval any(1, 2, 3) True
nemokosch is "gather for" something special?
no other idea left 08:55
m: my @vals = gather { .take for (1|2|3) }; dd @vals 08:56
Raku eval Array @vals = [any(1, 2, 3)]
nemokosch m: my @vals = gather .take for (1|2|3); dd @vals
Raku eval Array @vals = [1, 2, 3]
lizmat interesting 08:57
nemokosch Well, it must be a bug, I see no other way out at this point 🙃
lizmat hehe... or a feature :-)
m: say gather .take for (1|2|3) 08:58
camelia (1 2 3)
lizmat m: .say for gather .take for (1|2|3)
camelia 1
2
3
09:55 habere-et-disper left
lizmat github.com/rakudo/rakudo/issues/5320 10:07
.tell habere-et-disper your version may actually depend on a bug in Rakudo: github.com/rakudo/rakudo/issues/5320
meh
nemokosch ^^ 10:08
nemokosch 🧐 thanks 10:14
lizmat it appears to be a codegen issue with "for" inside a "gather" 10:17
librasteve m: say Date.new( 2023, (1..12).any, 13 ).iterator 10:33
Raku eval List::Todo.new
librasteve ^^ guess that means something is a bug - (and there was me thinking it was the single argument rule) 10:34
lizmat m: my $i := Date.new( 2023, (1..12).any, 13 ).iterator; dd $i.pull-one for ^2
camelia any(Date.new(2023,1,13), Date.new(2023,2,13), Date.new(2023,3,13), Date.new(2023,4,13), Date.new(2023,5,13), Date.new(2023,6,13), Date.new(2023,7,13), Date.new(2023,8,13), Date.new(2023,9,13), Date.new(2023,10,13), Date.new(2023,11,13), Date.new(2023,…
lizmat m: my $i := (1|2|3).iterator; dd $i.pull-one for ^2 10:35
camelia any(1, 2, 3)
IterationEnd
lizmat nothing wrong there librasteve
librasteve ok - so docs.raku.org/type/Junction#method_iterator 10:36
Returns an iterator over the Junction converted to a List. is wrong then?
in which case, what is the best way to "unpack" a junction ...? 10:37
(or is that idea forbidden in principal?)
lizmat the idea is that the eigenstates of a Junction are opaque 10:38
librasteve btw -- really enjoying the entertainment!
lizmat that's why I made raku.land/zef:lizmat/eigenstates :-)
librasteve ok - so its forbidden - but you have provided a workaround ? 10:41
lizmat that's the gist of it
librasteve ^^ that's a doc bug then? (I can file an issue) 10:42
lizmat m: my @a; sub a($a) { @a.push($a) }; a 1|2|3; dd @a # another way to unpack a junction 10:43
camelia Array @a = [1, 2, 3]
lizmat m: sub unpack(Mu $a) { my @a; sub a($a) { @a.push($a) }; a $a; @a.List }; dd unpack 1|2|3 # another way to unpack junctions 10:44
camelia (1, 2, 3)
librasteve use eigenstates; my $out = Date.new( 2023, (1..12).any, 13 ).day-of-week == 5; $out.&eigenstates.grep(*.so).elems.say; 10:48
nemokosch "unpacking a junction" is kind of a confusing concept anyway. The eigenvalues are in a certain relation to each other. 10:56
librasteve yeah - I get why it is forbidden (sorry I was using the word unpack loosley)
nemokosch oh I didn't mean to pick on the terminology, we could call it listing or iterating or whatever 10:57
by the way
10:58 habere-et-disper joined
m: my %multikey = (1|2) => True, (3&4) => False; dd %multikey 10:58
Raku eval Hash %multikey = {"1" => Bool::True, "2" => Bool::True, "3" => Bool::False, "4" => Bool::False}
nemokosch quite funny isn't it
when this came up roughly a year ago, the & version actually benchmarked significantly faster for some reason 10:59
lizmat the power of junctions :-)
nemokosch m: my %multikey = (1^2) => True; dd %multikey
Raku eval Hash %multikey = {"1" => Bool::True, "2" => Bool::True}
nemokosch anyway, they all work but there is no meaningful semantics behind it
lizmat true 11:00
DIHWIDT
nemokosch like if I were to read it "either 1 or 2 is the key of this hash, with True as the value" 11:01
that would actually be a junction of two hashes
but if somebody adds multiple junction keys, this interpretation either collapses or at least the number of cases will multiply and the whole thing becomes bloated in the blink of an eye 11:02
lizmat m: my %h{Mu} = (1|2) => 42; dd %h 11:04
camelia Hash[Any,Mu %h = (my Any %{Mu} = any(1, 2) => 42)
lizmat if you want Junctions to be keys in an object hash
nemokosch yes, I anticipated this would work 11:05
anyway... It's not fun to be the party pooper but I think the way junctions are, the idea to count friday 13 through junctions is unsound
lizmat I would agree 11:06
nemokosch in an alternative world there could be additive junctions that wouldn't collapse into a boolean value but a sum of the individual values 11:07
without that, the "most algebraic" solution seems to be to start from 1..12 and apply a "map-reduce" logic, transforming the value into 0's and 1's and summing them 11:09
m: (1..12).map({ Date.new( 2023, $_, 13 ).day-of-week == 5 }).sum.say 11:10
Raku eval 2
librasteve ^^ ++ 11:11
nemokosch I'm personally biased towards this approach in general, it can be applied to a wide range of problems, the data flows in one direction and it's easy to intercept at any point (especially now with .snitch) 11:12
habere-et-disper Looks like I stood on a junction in raku-beginner -- thank you everyone ! 11:15
'=D
nemokosch btw I could imagine a module that alters Date.new(2023, *, 13) in a way that it produces all the possible dates for this pattern
In general I think the WhateverStar could be used in clever ways but it's fairly dangerous to try to guess a general case for its meaning 11:16
it's more of a slangy thing 11:17
there is another manifesto-like statement I wanted to make 😅 11:25
it's true that a simple and straightforward tool can be easier to understand and master
lizmat I can see * for months and day to be okish, but for years, that would be interesting ? 11:26
nemokosch but it's also a virtue if one doesn't have to understand something in order to use it
and this is what I do believe that Raku is not a "write-only language" 11:27
not everyone reading code even reads that code with the intention to change it
let alone radically change it 11:28
on the other hand, this approach is drastically based on trust towards whoever gave you the high-level tool, and I am often concerned about that trust 11:30
lizmat: um, yeah... part of the reason I can only imagine something like that in user space 11:31
11:52 razetime left 12:09 razetime joined 12:24 siavash left 12:33 habere-et-disper left
antononcube A certain LLM prompt can be made for making "manifesto statements." 14:09
15:32 razetime left 15:50 Phytolizer joined 16:35 dakkar left 16:42 tea3po joined 16:45 teatwo left 17:24 Phytolizer left 18:44 tea3po left 18:46 teatime joined 19:01 Phytolizer joined 19:07 Phytolizer left, Phytolizer joined
SmokeMachine m: say Date.new(2023, 1, 13), Date.new(2023, 2, 13) … * 20:36
camelia (...)
SmokeMachine m: .say for (Date.new(2023, 1, 13), Date.new(2023, 2, 13) … *).head: 5 20:37
camelia 2023-01-13
2023-02-13
2023-02-14
2023-02-15
2023-02-16
nemokosch it just uses succ probably 20:40
Phytolizer m: .say for (Date.new(2023, 1, 13), *.later(:1month) ... *).head: 5 20:47
camelia 2023-01-13
2023-02-13
2023-03-13
2023-04-13
2023-05-13
librasteve .say for ( Date.new(2023, 1, 13), *.later(:1month) ... *).head: 5 21:22
seems that Discord swallowed a couple of stars... good way to get sequence of dates 21:23
nemokosch yeah, it was markdown for italics 21:24
Phytolizer oh, ha. i guess that wouldn't show up on irc 21:32
21:41 Phytolizer 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!