00:07 sortiz left 00:21 JRaspass left 00:25 frost-lab joined 00:44 frost-lab left 00:56 frost-lab joined 02:04 Kaeipi joined 04:11 MasterDuke left 07:29 squashable6 left 07:30 squashable6 joined 07:54 domidumont joined 08:02 Altai-man joined 08:49 JRaspass joined 09:42 sena_kun joined 09:44 Altai-man left 10:08 squashable6 left 10:09 squashable6 joined
Geth ¦ rakudo: lizmat self-assigned REPL repeating statements github.com/rakudo/rakudo/issues/4160 11:48
12:07 jmerelo joined
jmerelo #FOSSAsia call for speakers is open eventyay.com/e/fa96ae2c/cfs 12:07
15 minute talks, + 10 minute interaction, in any language you want. Great chance to introduce Raku!
timotimo sounds like i need to brush up on my lojban 12:08
lo gerku ku vofli ue sai
nwc10 yes, first thing I thought was "Esparanto?" 12:26
jmerelo In the Perl/Raku world, I would say Latin is the most widespread. 12:30
So I'd go for that... 12:31
13:41 Altai-man joined 13:44 sena_kun left 14:08 jmerelo left
Geth roast: 5f67d2a474 | (Elizabeth Mattijsen)++ | 2 files
Tweak checks on error messages to match array slice re-imagination
roast: f708802a2d | (Elizabeth Mattijsen)++ | S32-array/exists-adverb.t
Fix improper test on :exists:!v validity for array slice re-imagination
roast: b901130854 | (Elizabeth Mattijsen)++ | S32-list/seq.t
Array slice assignment is no longer lazy after array slice re-imagination
rakudo: ca7bc91e71 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 22 files
Re-imagine array slicing, fixing issues, between 2x and 10x faster (#4131)

This commit makes array slice accesses and assignment between 2x and 10x as fast *and* fixes several issues found in the previous implementation. It also creates a framework for handling adverbs that are allowed with postcircumfix adverbs (in a separate source-file Rakudo/Internals/PostcircumfixAdverbs).
... (16 more lines)
linkable6 RAKUDO#4131 [closed]: github.com/rakudo/rakudo/pull/4131 Re-imagine array slicing, fixing issues, between 2x and 10x faster
14:42 frost-lab left 14:55 domidumont left, domidumont joined 15:09 b2gills joined 15:21 patrickb joined 17:42 sena_kun joined 17:43 Altai-man left 17:44 domidumont left 17:46 lucasb joined 18:01 JRaspass left
vrurg . 18:41
tellable6 2021-01-08T16:59:16Z #raku <[Coke]> vrurg: my method private_method_table(Metamodel::PrivateMethodContainer: $obj) # this is now a compile time error in rakudo, is this change a bug?
vrurg [Coke]: with regard to the changes in docs, I think it's the right thing to do. 18:43
[Coke] +1 18:47
(already done)
vrurg [Coke]: BTW, I'm not sure it ever compiled under rakudo. 18:48
[Coke] we've had that code in the docs for years, with an xt/ test that was passing 18:52
gist.github.com/Whateverable/68c5c...a6dbfd36bb complains about the "missing block", and in the test, we add enough around the method (including a class) that the block is there. 18:53
the line as shown throws a compile error before we even get as far as noting the block. 18:54
vrurg [Coke]: actually, the compilation error is a bug. I'd like to ask you to open an issue on it. I see what the problem is, but not sure what'd be the best way to fix it. 19:13
dogbert17 m: say (1,2,3; 4,5,6; 7,8,9)[*;{0,1}] 19:20
camelia (3 6 9)
19:35 MasterDuke joined 19:46 JRaspass joined
lizmat dogbert17: looks like the return value is not interpreted as a list, when it should 19:56
a 2 elem list is returned, and it handles it as [*;2] 19:57
Geth rakudo: c039b28fc8 | (Elizabeth Mattijsen)++ | src/core.c/Rakudo/Iterator.pm6
Streamline some iterator setups

You can assign attributes in a signature like:
   method foo($!foo,$!bar)
but this costs 2 Scalar allocations for each invocation. In normal ... (5 more lines)
lizmat m: my %h = "a" .. "j" Z=> 1..*; dd %h{"a","b",("c","d")} 20:30
camelia (1, 2, Any)
lizmat yuck
timotimo but how would you access entries that are keyed on a list? 20:31
lizmat m: my @a = <a b c d e f g h i j>; dd @a[0,1,(2,3)]
camelia ("a", "b", ("c", "d"))
lizmat it should iterate the list, like array slices do ?
timotimo that works because indices can only be numbers
hm. 20:32
you're right, you can already %foo<a b c> and get a, b, c
this is tricky 20:33
lizmat tricky how?
for the core developer?
timotimo to decide how it should be handled
lizmat same as with array slices? 20:34
m: my %h = "a" .. "j" Z=> 1..*; dd %h{"a","b",("c","d")} # should be (1,2,(3,4))
camelia (1, 2, Any)
timotimo m: my %a; %a{(1,2),} = "hi"; %a{(2,3),} = "bye"; dd %a; %a{(1,2),(2,3)} = 99, 100; dd %a
camelia Hash %a = {"1 2" => "hi", "2 3" => "bye"}
Hash %a = {"1 2" => 99, "2 3" => 100}
timotimo m: my %a{Any}; %a{(1,2),} = "hi"; %a{(2,3),} = "bye"; dd %a; %a{(1,2),(2,3)} = 99, 100; dd %a
camelia Hash[Any,Any] %a = (my Any %{Any} = (1, 2) => "hi", (2, 3) => "bye")
Hash[Any,Any] %a = (my Any %{Any} = (1, 2) => 99, (1, 2) => "hi", (2, 3) => "bye", (2, 3) => 100)
timotimo ok they aren't value types 20:35
lizmat nope
timotimo but there are value types that are iterable
lizmat not in core, I don't think 20:36
timotimo i don't feel very strongly about this, since i don't know if anybody uses this 20:37
m: my %a{Any}; %a{set(1,2),} = "hi"; %a{set(2,3),} = "bye"; dd %a; %a{set(1,2),set(2,3)} = 99, 100; dd %a
camelia Hash[Any,Any] %a = (my Any %{Any} = Set.new(1,2) => "hi", Set.new(3,2) => "bye")
Hash[Any,Any] %a = (my Any %{Any} = Set.new(1,2) => 99, Set.new(3,2) => 100)
timotimo here's an iterable value type, right?
lizmat ah, like that... I guess you're right then 20:39
timotimo i'm not happy about it :) 20:40
lizmat perhaps it should be different for object hashes
timotimo i was thinking to differentiate on whether the type is iterable or not, but i think you could just have a subtype that is iterable and things won't work any more like i imagined 20:41
lizmat m: my %h = a => 42, "b c" => 666; dd %h{"a",("b","c")} # that doesn't feel right either 20:42
camelia (42, 666)
lizmat afk& 20:55
20:58 Xliff joined
Xliff What does "Failed to set temporary DH" mean from IO::Async::SSL? 21:09
jnthn: ^^
tellable6 Xliff, I'll pass your message to jnthn
timotimo hmm, related to diffie hellman key exchange? 21:10
Xliff Could be... but why would it fail? I would expect the key stuff to be done in memory.
Anway there is an upgrade, so I hope that will fix it. 21:11
How are you doing timo^2?
timotimo i'm okay
Xliff That's good!
timotimo my mind boggles from the new details i'm seeing from the USA thing
but i did some good work today
how are you? 21:12
Xliff Hah! I'm living it.
4 years. I'm so numb.
I'm doing OK. Got another project I need to solve a segfault for.
Almost 480,000 lines now. 21:13
timotimo wow
Xliff Also working on my first Raku project at $dayJob
Hence the IO::Socket::Async::SSL thing
Trying to run Cro as HTTPS and things are vexing me.
timotimo nice! but then also: oh no 21:14
worst case, you can use a web server in front of the cro instance to terminate TLS connections for you 21:15
since it's very common already to run any kind of web app behind a reverse proxy
21:36 maggotbrain left 21:41 Altai-man joined 21:43 sena_kun left
dogbert17 lizmat: I took the example from R#2488 21:44
linkable6 R#2488 [open]: github.com/rakudo/rakudo/issues/2488 Multi-dimensional slice with Callable returning Iterable
dogbert17 btw, are you still interested in github.com/Raku/roast/pull/712 21:45
22:00 patrickb left 22:03 Altai-man left
[Coke] vrurg: github.com/rakudo/rakudo/issues/4162 22:34
vrurg [Coke]: Just've got the email. :)
[Coke] np, this was the easy part.
Geth ¦ rakudo: vrurg self-assigned metamodel role invocant compilation bug. github.com/rakudo/rakudo/issues/4162 22:37
23:08 maggotbrain joined