00:20 rypervenche left 00:27 rypervenche joined, dogbert17 left 00:40 leont left
[Coke] .tell patrickb I'm not enough of a windows expert to know if that makes sense. 00:45
tellable6 [Coke], I'll pass your message to patrickb
[Coke] .tell patrickb ... and I see it's already hit main. 00:46
tellable6 [Coke], I'll pass your message to patrickb
00:53 maggotbrain left 01:07 maggotbrain joined 01:18 guifa left 01:35 Kaiepi left 01:36 Kaiepi joined 02:06 lucasb left 03:19 Kaiepi left 03:20 Kaiepi joined 03:32 Kaiepi left 03:33 Kaiepi joined 04:09 Kaiepi left 04:10 Kaiepi joined 04:18 Kaiepi left, Kaiepi joined 04:43 Kaiepi left, Kaiepi joined 05:08 MasterDuke left 06:28 sena_kun joined
sena_kun releasable6, status 06:43
releasable6 sena_kun, Next release in ≈8 days and ≈12 hours. There are no known blockers. 0 out of 48 commits logged
sena_kun, Details: gist.github.com/4cb54db3549671d9a3...87114d236e
07:20 patrickb joined 07:55 domidumont joined 08:14 leont joined 09:05 leont left, leont joined
lizmat Files=1336, Tests=113630, 219 wallclock secs (29.34 usr 8.24 sys + 3056.47 cusr 296.12 csys = 3390.17 CPU) 09:09
also from the nqpif2ternary branch
09:14 Ven`` joined 09:15 Ven`` left
Geth_ rakudo/rakuast: da66583ab7 | (Brad Gilbert)++ (committed using GitHub Web editor) | src/Raku/ast/pair.rakumod
(:!foo).IMPL-INTERPRET should have a False value

It seems like the code for `RakuAST::ColonPair::True.IMPL-INTERPRET()` was copied to `RakuAST::ColonPair::False` but without the proper alteration.
rakudo/rakuast: 46e2ab613d | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/Raku/ast/pair.rakumod
Merge pull request #3960 from b2gills/patch-6

  [RakuAST] (:!foo).IMPL-INTERPRET should have a False value
09:20 Altai-man joined 09:23 sena_kun left 09:46 domidumont left 09:48 domidumont joined
Geth_ rakudo: mid1221213++ created pull request #3962:
Fix NativeCall/C++ to allow double instantiation
09:54 Kaeipi joined 09:55 Kaiepi left
Geth_ rakudo: ecaf334c85 | (Alexandre Jousset)++ | 2 files
Fix NativeCall/C++ to allow double instantiation

The test of the arguments concreteness using the Op
  <isconcrete> instead of <isconcrete_nd> was preventing
the instantiation more than once of the same class using the following code:
... (9 more lines)
rakudo: 0f1952792d | niner++ (committed using GitHub Web editor) | 2 files
Merge pull request #3962 from mid1221213/fix-nativecall-1

Fix NativeCall/C++ to allow double instantiation
nine (my $a, my $b) = 42, 666, 314 should most definitely not warn 10:03
That's a pretty common and useful construct, just usually with a function on the right hand side and some return values just not being interesting 10:04
lizmat yeah, I see that
the () = 42 case feels... different ?
I ran into this when writing tests for %h{a;b;c} 10:05
m: my %h = a => { b => { c => 42 } }; dd %h{"a";*;"x"} 10:06
camelia (Any,)
lizmat there are some cases where an empty list is returned with certain adverbs 10:08
any adverbed slice is a right value
however, the empty list silently absorbs any assignment
Geth_ rakudo/fix-pod-to-text-warn: dbc7c1e5d7 | Altai-man++ (committed using GitHub Web editor) | lib/Pod/To/Text.rakumod
Use a more efficient code, lizmat++ for a tip
lizmat m: my %h = a => { b => { c => 42 } }; dd %h{*;"b";"x"}:kv: 10:11
camelia Unexpected named argument 'kv' passed
in block <unit> at <tmp> line 1
lizmat m: use v6.e.PREVIEW; my %h = a => { b => { c => 42 } }; dd %h{*;"b";"x"}:kv:
camelia Don't know setting RESTRICTED
at gen/moar/ModuleLoader.nqp:249 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/Perl6/ModuleLoader.moarvm:previous_setting_name)
from gen/moar/ModuleLoader.nqp:254 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/…
lizmat ?? 10:12
m: use v6.e.PREVIEW
camelia Don't know setting RESTRICTED
at gen/moar/ModuleLoader.nqp:249 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/Perl6/ModuleLoader.moarvm:previous_setting_name)
from gen/moar/ModuleLoader.nqp:254 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/…
lizmat that is... weird ?
timotimo most probably a camelia artifact from how exactly it invokes rakudo to get the restricted setting? 10:13
lizmat no idea 10:14
moritz possible
lizmat m: use v6.d
camelia Don't know setting RESTRICTED
at gen/moar/ModuleLoader.nqp:249 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/Perl6/ModuleLoader.moarvm:previous_setting_name)
from gen/moar/ModuleLoader.nqp:254 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/…
lizmat m: use v6.c
camelia Don't know setting RESTRICTED
at gen/moar/ModuleLoader.nqp:249 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/Perl6/ModuleLoader.moarvm:previous_setting_name)
from gen/moar/ModuleLoader.nqp:254 (/home/camelia/rakudo-m-inst-2/share/perl6/lib/…
lizmat I guess we haz a bug
nine I don't think it's camelia's fault. Sounds more like a wrong assumption in previous_setting_name 10:19
the assumption being that settings are always about language versions and there are not other settings than the one corresponding to language versions
lizmat yeah, that would be my assessment as well 10:20
in any case: %h{"a";"b";"x"}:kv gives () 10:21
because there is no key "x" in the deepest hash
and then %h{"a";"b";"x"}:kv = 42 then silently does not die 10:22
it's a corner case, yes, but still :-)
actually, the simpler case already works in the same way: 10:23
m: my %h; %h<a b>:kv = 42
camelia ( no output )
lizmat m: my %h; %h<a>:kv = 42 # simpler still 10:25
camelia ( no output )
AlexDaniel` restricted setting has been broken for a long time 10:26
moreover it doesn't really work anyway, so just remove the restricted setting
and by “doesn't work” I mean that blacklists don't work for that purpose 10:28
nine m: open "testfile", :w 10:51
camelia Failed to open file /home/camelia/testfile: Permission denied
in block <unit> at <tmp> line 1
nine yeah...no restrictions at all
11:02 squashable6 left 11:03 squashable6 joined
Geth_ rakudo: 18f3808e19 | (Elizabeth Mattijsen)++ | 6 files
Rename multislice related files

  - for consistency and thus easier maintenance
  - does *not* change *any* functionality whatsoever
  - yet preserves git blame history
rakudo/master: 5 commits pushed by (Elizabeth Mattijsen)++ 11:18
AlexDaniel` nine: not sure what you mean exactly, so just in case: that error message isn't related to the restricted setting, and I think it stopped working since v6.d was released 11:51
but, it being a blacklist, it never really worked anyway
nine AlexDaniel`: there was no sarcasm there. I really meant that the restricted setting doesn't restrict anything anymore at all. And even if it worked as designed, it still wouldn't cover simple cases like "somefile".IO.open IIUC 12:03
AlexDaniel` ok :) 12:04
Geth_ rakudo: 93503c24b4 | (Elizabeth Mattijsen)++ | src/core.e/hash_multislice.pm6
Make %h{;} slices with * non-assignable

Because a slice with * at any level, is non-deterministic. Making it die with an appropriate error message like:
   $ raku -e 'my %h; %h{*} = 42'
   Cannot assign to *, as the order of keys is non-deterministic
... (5 more lines)
roast: 675e0ab065 | (Elizabeth Mattijsen)++ | S32-hash/multislice-6e.t
Adapt test to non-assignability of * slices
lizmat afk for a few hours& 13:04
13:21 sena_kun joined 13:23 Altai-man left
Geth_ nqp: 6049b3806b | Coke++ | docs/ops.markdown
Document isconcrete_nd

Pulled from IRC conversation of lizmat++ nine++
14:37 domidumont left 14:45 squashable6 left 14:46 squashable6 joined 14:51 [TuxCM] left
[Tux] Rakudo version 2020.09-57-g93503c24b - MoarVM version 2020.09-10-g8d7a80d55
csv-ip5xs0.815 - 0.864
csv-ip5xs-207.979 - 8.272
csv-parser27.862 - 30.659
csv-test-xs-200.387 - 0.391
test7.502 - 7.814
test-t1.866 - 1.896
test-t --race0.860 - 0.902
test-t-2031.589 - 32.380
test-t-20 --race8.763 - 9.295
14:56 [TuxCM] joined, domidumont joined
lizmat [Coke]++ 15:15
15:31 jmerelo joined 15:58 patrickb left 16:17 MasterDuke joined 16:18 domidumont left 16:28 domidumont joined 16:37 domidumont left 16:50 finsternis joined 16:51 domidumont joined 17:09 domidumont left
Geth_ rakudo: b60772fa70 | (Elizabeth Mattijsen)++ | src/core.e/hash_multislice.pm6
Don't pass current dim around anymore with %h{;}

There's no need to pass it around, we just need a outside lexical to remember where we are: saves 1 arg and 1 native int allocation per recursion, makes it about 10% faster.
17:20 Altai-man joined 17:23 sena_kun left
[Coke] I agree with zoffix on github.com/Raku/nqp/issues/389 - I recommend we close this as a wontfix. 17:24
lizmat: github.com/Raku/nqp/issues/101 - ISTR you (et al.?) have done work to speed up JSON since 2013. 17:28
lizmat [Coke]: I've worked on JSON::Fast a lot 17:29
MasterDuke looks like method Numeric isn't docced for Match. the page has a method from Capture, but it does something different 17:31
m: say \(1,2,3, apples => 2).Numeric; my $a = "abc123def456" ~~ /\d+/; dd $a.Numeric 17:32
camelia 3
jmerelo MasterDuke: can you please raise the issue? 17:33
tellable6 2020-10-12T20:16:24Z #raku <melezhik> jmerelo: - see github.com/melezhik/RakuDist/blob/...modules.sh
2020-10-12T20:18:18Z #raku <melezhik> JJMerelo: here is full CICD example with pulling RakuDist build status and waiting it finishes - github.com/melezhik/RakuDist/blob/...on-example
2020-10-12T20:19:21Z #raku <melezhik> JJMerelo: you basically give it a list of modules and run them all, RakuDist will place request in queue and start executing them one by one ...
jmerelo .tell melezhik thanks!
tellable6 jmerelo, I'll pass your message to melezhik
MasterDuke sure 17:35
17:38 Ven`` joined
jmerelo MasterDuke: thanks! 17:51
17:51 jmerelo left
Geth_ rakudo: 198f603043 | (Elizabeth Mattijsen)++ | src/core.e/hash_multislice.pm6
Optimize key creation in %h{;}:adverbs slices

  - make the key list a visible lexical IterationBuffer
  - don't need to pass current keys as a parameter anymore
  - makes these slices about 25% faster
20:02 dogbert17 joined 20:19 Altai-man left 20:27 ggoebel left 20:29 ggoebel joined
[Coke] m: use nqp: dd nqp::indexingoptimized 22:09
camelia 5===SORRY!5=== Error while compiling <tmp>
at <tmp>:1
------> 3use nqp7⏏5: dd nqp::indexingoptimized
[Coke] m: use nqp: dd nqp::indexingoptimized();
camelia 5===SORRY!5=== Error while compiling <tmp>
at <tmp>:1
------> 3use nqp7⏏5: dd nqp::indexingoptimized();
[Coke] nqp: nqp::say(nqp::indexingoptimized) 22:10
camelia Arg count 0 doesn't equal required operand count 2 for op 'indexingoptimized'
at gen/moar/stage2/QAST.nqp:1568 (/home/camelia/rakudo-m-inst-2/share/nqp/lib/QAST.moarvm:compile_mastop)
from gen/moar/stage2/QAST.nqp:1781 (/home/camelia/rakudo-m-inst-2/sh…
[Coke] nqp: nqp::say(nqp::indexingoptimized("what"))
camelia what
[Coke] nqp: nqp::say(nqp::indexingoptimized("what","now"))
camelia now
[Coke] nqp: nqp::say(nqp::indexingoptimized("what","now","brown","cow"))
camelia Arg count 4 doesn't equal required operand count 2 for op 'indexingoptimized'
at gen/moar/stage2/QAST.nqp:1568 (/home/camelia/rakudo-m-inst-2/share/nqp/lib/QAST.moarvm:compile_mastop)
from gen/moar/stage2/QAST.nqp:1781 (/home/camelia/rakudo-m-inst-2/sh…
[Coke] ... Could use some help documenting this op. Looks like at some point it's collapsing the strands in a Moar string object, but <shrugs vaguely> 22:11
ah. I can probably just lift the comment from the MoarVM source. 22:12
nqp: nqp::say(nqp::indexingoptimized("what","now","brown"))
camelia Arg count 3 doesn't equal required operand count 2 for op 'indexingoptimized'
at gen/moar/stage2/QAST.nqp:1568 (/home/camelia/rakudo-m-inst-2/share/nqp/lib/QAST.moarvm:compile_mastop)
from gen/moar/stage2/QAST.nqp:1781 (/home/camelia/rakudo-m-inst-2/sh…
[Coke] unsure why it's allowing two args.
timotimo [Coke]: indexingoptimized only takes one arg, but it also returns something, i believe? 22:13
i wonder if it just takes the string register that holds the first string and overwrites it with the result
22:23 patrickb joined
timotimo [Coke]: collapsing the strands is supposed to make running regexes over strings faster, since access to the string will then be fewer data lookups 22:33
22:41 Ven`` left 23:06 patrickz joined 23:10 patrickb left 23:36 patrickz left