guifa MasterDuke indeed. I wonder if a grammar by default should avoid the more Raku-y ones like <identifier>. <digit> makes sense (and can even be usefully overridden with <[0..9]> ) 03:28
actually, nm. They aren't included (I think I've been in slang world too long). There shouldn't be a problem with identifier or token, etc, in a normal grammar. 05:32
08:50 TempIRCLogger left 08:51 lizmat left, RakuIRCLogger left 08:59 dakkar joined 09:00 lizmat_ left, lizmat joined 11:39 frost left
Nemokosch Hello 12:35
yes, I'm thinking about 154-2...
is there a nice way to remove duplicates from a Seq? 12:36
MasterDuke .unique 13:17
Nemokosch wow thanks, this is good to know 13:24
what is the runtime cost of using it? I guess it needs to store the elements of the sequence?
MasterDuke github.com/rakudo/rakudo/blob/mast...1846-L1882 13:32
so yeah, just checked against a hash it creates 13:34
15:07 frost joined 15:36 frost left 17:40 dakkar left 18:07 Hello16 joined 18:08 Hello16 left
gfldex <@297037173541175296> .squish might work too 19:26
Nemokosch Oh? 19:28
Nahita if sequential
m: [12, 12, 4, -1, 12].squish.say
Nemokosch Yeah, that is for something else, right?
Nahita yeah, they did say \*might\*
Nemokosch It's more effective but tbh I wouldn't want to depend on a "trivial detail" like this 19:32
gfldex @Nemokosch#9980 do you care to share your 154-2 solution? 20:41
Nemokosch <github.com/manwar/perlweeklychalle...2.raku> 20:45
gfldex here is mine: gfldex.wordpress.com/2022/03/01/pushing/ 20:46
Nemokosch Nice :) 20:52
I see you love containerless stuff 20:53
Nahita `×` instead of 1 more `*` could be better to avoid confusion IMHO, while you're using ∞ 20:55
so many *s
gfldex m:``` 20:59
use experimental :cached;
proto P(Int $n) is cached { * }
multi sub P ($n where * == 0|1|2) { 1 }
multi sub P ($n where * ≥ 3) { P($n - 2) + P($n - 3) }
(0..*).map(&P).grep(&is-prime).unique.head(10).say;
```
@Nemokosch#9980 is cached works for multis too
Nemokosch I know but I preferred given when here 21:01
gfldex Sadly, we can't use junctions as literals in signatures.
Nemokosch I suspected so 21:05
I thought if I couldn't use the constant value elegantly, why even bother with the multi 21:07
gfldex I like to use containerless symbols for lazy lists because they act rather like functions. 21:15
22:48 codesections5 joined 22:50 codesections left, codesections5 is now known as codesections