00:00 Xliff left 00:13 oddp left 00:14 sena_kun joined 00:17 Altai-man left 02:14 Altai-man joined 02:16 sena_kun left 02:29 chansen_ left 02:35 chansen_ joined 03:48 lucasb left 04:14 sena_kun joined 04:16 Altai-man left 06:14 Altai-man joined 06:16 sena_kun left 07:08 JJMerelo joined
JJMerelo m: subset RealT of ::T where ::T.WHAT == ::T 07:08
camelia 5===SORRY!5=== Error while compiling <tmp>
No such symbol 'T'
at <tmp>:1
------> 3subset RealT of ::T where ::T7⏏5.WHAT == ::T
tellable6 2020-08-08T05:43:09Z #raku <ab5tract> JJMerelo yeah that's the one! the Hilbert matrix does cute things to floats when it encounters it's inverse.
07:40 leont joined 07:58 oddp joined 08:14 sena_kun joined 08:16 Altai-man left 09:06 JJMerelo left 09:36 JJMerelo joined
JJMerelo m: say v.0.0.1 < v0.0.1 09:36
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call
at <tmp>:1
------> 3say v.7⏏050.0.1 < v0.0.1
JJMerelo m: say Version.new("v.0.0.1") cmp Version.new("0.0.1") 09:42
camelia Less
JJMerelo OK, I see the problem now github.com/softmoth/perl6-pod-to-m.../issues/22 09:43
10:14 Altai-man joined 10:16 sena_kun left 10:53 JJMerelo left 11:19 Kaiepi left, Kaiepi joined 12:14 sena_kun joined 12:16 Altai-man left 12:57 domidumont joined 13:14 oddp left
dogbert17 am I the only one getting a failure in t/spec/S14-traits/package.t 13:43
Type check failed in binding to parameter '<anon>'; expected Description but got Perl6::Metamodel::ClassHOW+{Description[Str]} (Perl6::Metamodel::Cl...) 13:46
AlexDaniel > Junctions are, despite their internal complexity and rarity in programming languages are something that are so well thought out and integrated into common Raku coding styles that most use them without any thought. 14:11
oh no :)
timotimo danger, danger! 14:13
14:14 Altai-man joined 14:17 sena_kun left
AlexDaniel the blog posts about RFCs are really good, even if you take unintentional irony away :) 14:21
m: sub foo($x where { 2 < $_ } where { $_ < 4 }) { say $x }; foo(3) 14:26
camelia 3
AlexDaniel heh!
or without blocks: 14:27
m: sub foo($x where 2 < * where * < 4) { say $x }; foo(3)
camelia 3
rypervenche Or without the extra where :) 14:41
m: sub foo($x where 2 < * < 4) { say $x }; foo(3)
camelia 3
AlexDaniel rypervenche: the point is that you can use multiple where's 14:42
which I didn't know about
rypervenche Ahh ok. 14:43
AlexDaniel if I really wanted it short I would've written something like
sub foo($x where 2^..^4) { say $x }; foo(3)
evalable6 3
AlexDaniel in hindsight, supporting junctions as objects is a mistake 14:48
timotimo you would have preferred only syntax?
AlexDaniel most of the examples I see is people wanting syntactic sugar to compare against multiple values
that can be done during compilation by just expanding the expression for the user 14:49
then people also want to check all/any/one values of an iterable
ok, easy: all(@a X== 42)
timotimo oh no, a slippery sloap :) 14:50
sippery soap
AlexDaniel or something like that, basically a sub to check that all values are truthy
no, a slippery sloap is introducing junctions as objects that you can pass around
people get away with it because they limit themselves to the two edge cases I mentioned above 14:51
but *any* code you write is probably buggy if you suddenly get junctions as input
for example: 14:52
m: my @a; @a.push(42|43); @a.push(4|8); say @a; say @a.sort
camelia [any(42, 43) any(4, 8)]
This type cannot unbox to a native integer: P6opaque, Junction
in block <unit> at <tmp> line 1
AlexDaniel none of this really has any meaning 14:53
although, `sum` “works”:
m: my @a; @a.push(42|43); @a.push(4|8); @a.push(2|3); say @a; say @a.sum
camelia [any(42, 43) any(4, 8) any(2, 3)]
any(any(any(48, 49), any(52, 53)), any(any(49, 50), any(53, 54)))
AlexDaniel and it's not like the most basic use of junctions is straightforward 15:03
people always get confused by the way it negates
but the situation highlights the additive nature of Raku – if something like that was proposed today I'm sure it wouldn't fly at all, but if it slipped in earlier then we praise the feature insead of thinking how to reimplement it with a proper solution :) 15:08
timotimo i'm not that happy with the negation, honestly. that's a "make it how humans would intuitively understand it" thing 15:14
16:13 oddp joined 16:14 sena_kun joined 16:16 Altai-man left
lizmat . 16:28
18:14 Altai-man joined 18:16 sena_kun left 18:20 Altai-man left 18:27 Altai-man joined 18:43 domidumont left, domidumont joined 18:53 domidumont left 18:54 domidumont joined
AlexDaniel timotimo: I'm planning to remove lrzip stuff from whateverable. I hope zstd can work just as good in its long-range mode 19:04
fingers crossed :)
hopefully will remove all the extra code paths too
timotimo i hope it'll work out :) 19:12
AlexDaniel benchmarking… 19:16
surprisingly, it's not really better when it comes to size 19:25
timotimo hm
AlexDaniel it's much faster I think, but that needs more benchmarking
9.5M all.tar.lrz # -c 19 --long
8.3M all.tar.lrz
I tried --long=31 which seems to be the max value, that resulted in 9.4M 19:26
not a real win I think
--ultra -22 --long=31 # 9.2M 19:28
that's as far as you can push it, hmm…
--ultra -22 --long=31 --format=lzma # 7.3M 19:30
sorry that paste above the first line is .tar.zstd 19:39
timotimo: actually, that slow command that results in 7.3M files… 19:59
timotimo: it takes more than a minute to compress vs 20s for lrzip 20:00
but, unsurprisingly
it's faster at decompression
1s vs 1.5s
so, like usual, I win *both* space and decompression speed by just switching to zstd :) 20:01
this time, unfortunately, at the cost of compression speed :( 20:03
timotimo whoops, huh 20:10
hey, i wonder
actually ... no, forget about it
AlexDaniel what? :) 20:12
20:15 sena_kun joined 20:16 Altai-man left
[Tux] Rakudo version 2020.07-56-g798614cb7 - MoarVM version 2020.07-14-g656142e21
csv-ip5xs0.826 - 0.828
csv-ip5xs-207.913 - 7.980
csv-parser24.479 - 25.756
csv-test-xs-200.380 - 0.381
test7.503 - 7.607
test-t1.860 - 1.890
test-t --race0.830 - 0.874
test-t-2032.050 - 32.824
test-t-20 --race8.720 - 8.910
AlexDaniel timotimo: github.com/Raku/whateverable/issue...-670971601 20:40
not entirely sure what to do actually
basically, switching straight to zstd will make bisectable much much faster 20:41
bisect: old=2020.02.1 new=2020.05.1 say Date.today.last-date-in-month 20:42
bisectable6 AlexDaniel, Bisecting by exit code (old=2020.02.1 new=2020.05.1). Old exit code: 1
AlexDaniel, bisect log: gist.github.com/d039ce57f35c60a776...b00d50ed2d
AlexDaniel, (2020-04-17) github.com/rakudo/rakudo/commit/2f...78951717fe
AlexDaniel hmm
the server is a bit busy right now :) 20:43
but it'd be a few seconds faster, big difference
MasterDuke i think faster is better 20:50
AlexDaniel faster is better yeah, unless it is running out of space… x) 20:51
I actually deleted all moarvm builds because I need space to clean things up… 20:52
it's not significantly faster on larger archives :( 20:55
ah wait…
yeah it is more than three times faster on decompression x) 20:58
22:14 Altai-man joined 22:16 sena_kun left 23:14 oddp left