»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_log/perl6 | UTF-8 is our friend! 🦋
Set by Zoffix on 25 July 2018.
00:06 Kaiepi joined 00:07 Manifest0 left 00:09 eseyman left 00:11 xiaoyafeng left 00:12 Manifest0 joined 00:18 eseyman joined 00:20 Sgeo left 00:22 zacts joined 00:28 zacts left 00:30 zacts joined 00:34 sivoais left, sivoais joined 00:35 Manifest0 left 00:39 zacts left 00:40 Manifest0 joined 00:45 Sgeo joined 00:49 Manifest0 left 00:55 Manifest0 joined 01:00 zacts joined 01:02 Manifest0 left 01:04 zacts left 01:07 Manifest0 joined 01:26 aborazmeh left 01:27 Manifest0 left 01:33 Manifest0 joined 01:40 qiqi joined 01:48 Manifest0 left 01:52 molaf left 01:53 Manifest0 joined 01:54 noisegul_ joined 01:57 noisegul left 02:03 Manifest0 left 02:04 molaf joined 02:05 Sgeo_ joined 02:07 Manifest0 joined 02:08 Sgeo left 02:19 Manifest0 left 02:24 Manifest0 joined 02:26 squashable6 left 02:29 kst joined 02:30 squashable6 joined, ChanServ sets mode: +v squashable6 02:31 Manifest0 left 02:38 Manifest0 joined 02:53 Manifest0 left 02:58 Manifest0 joined 03:04 Manifest0 left 03:09 epony left, Manifest0 joined
AndroidKitKat Is there a more graceful way to "pop" an item from the front of an array other than to delete[0] 03:16
I'm used to Python lists where a .pop() removes an item from the front so if you had list = [1, 2, 3] and did num = list.pop() 03:17
you would get 1
MasterDuke AndroidKitKat: that's `shift` in perl (5|6) 03:19
AndroidKitKat Thanks!
MasterDuke np
AndroidKitKat I was looking through the docs for that, appreciate it
I must've failed reading comprehension lmao 03:20
MasterDuke hm, wonder if the docs for `pop` should mention `shift`... 03:21
03:23 xiaoyafeng joined 03:25 Manifest0 left 03:30 Manifest0 joined 03:34 Manifest0 left
cpan-p6 New module released to CPAN! FileSystem::Parent (0.2.1) by 03LEMBARK 03:35
03:37 xiaoyafeng left 03:40 Manifest0 joined 03:43 epony joined 03:44 mowcat left 03:50 Manifest0 left 03:54 Manifest0 joined 03:59 Manifest0 left 04:05 Manifest0 joined 04:06 grayrider joined 04:18 cpan-p6 left 04:20 cpan-p6 joined, cpan-p6 left, cpan-p6 joined 04:24 Actualey` left 04:29 Manifest0 left 04:37 Manifest0 joined 04:49 ayerhart joined 04:50 curan joined 04:59 qiqi left
holyghost I'm off to my kids 05:09
05:09 holyghost left
Geth doc: Elronnd++ created pull request #2795:
Replace "two's complement" with "logical negation" to better reflect the meaning and usage of the operator		 05:12
Elronnd Arghh 05:15
why does it say logical negation
05:17 sauvin joined 05:22 ufobat__ left 05:23 feldspath joined
Voldenet bitwise negation may be technically incorrect when the two's complement doesn't match bitwise negation of the number 05:28
05:28 qiqi joined, Manifest0 left
Elronnd is what it's doing not bitwise negation? 05:30
Voldenet I'm not sure how is "163**251" stored or if it's even specified anywhere 05:31
bitwise negation implies that it flips the bits in given bytes 05:32
Elronnd ohh because width is ambiguous 05:33
Voldenet indeed, and length is certainly not flipped 05:34
Elronnd hah
hmmm
I just think that _the_ two's complement is an odd phrasing because there are any number of those 05:35
Voldenet docs.perl6.org/language/5to6-perlop does this right
>use prefix +^ for bitwise integer negation. Assumes two's complement. 05:36
05:36 Manifest0 joined 05:38 feldspath left 05:40 Woodi left 05:41 domidumont joined 05:44 molaf left 05:49 jmerelo joined
jmerelo squashable6: status 05:49
squashable6 jmerelo, Next SQUASHathon in 13 days and ≈22 hours (2019-06-01 UTC-14⌁UTC+20). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
05:51 Manifest0 left
Geth doc: a640eda155 | Elronnd++ | doc/Language/js-nutshell.pod6
Replace "two's complement" with "bitwise negation" to better reflect the meaning and usage of the operator
 05:54
doc: a69bccb37d | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | doc/Language/js-nutshell.pod6
Merge pull request #2795 from Elronnd/master

Replace "two's complement" with "bitwise negation" to better reflect the meaning and usage of the operator
synopsebot Link: doc.perl6.org/language/js-nutshell
05:56 Manifest0 joined 05:59 squashable6 left 06:00 Manifest0 left
Elronnd Has anyone used Terminal::Print? How can I move the cursor to a specific spot? 06:02
06:02 squashable6 joined, ChanServ sets mode: +v squashable6
jmerelo Elronnd: not me, sorry... 06:03
Elronnd it looks really nice but...no docs, compared with ncurses
jmerelo .seen japhb 06:04
yoleaux I saw japhb 16 May 2019 17:58Z in #moarvm: <japhb> I love that algorithm.
jmerelo Elronnd: you can try and talk to him in that channel ^^^ 06:05
Elronnd alright, I will take a look there
thanks! 06:06
jmerelo Elronnd: sure. Good luck!
Elronnd err, here is here too :P
06:06 Manifest0 joined
Elronnd japhb: do you know how I can move the cursor to a specific place with Terminal::Print? 06:07
06:11 feldspath joined 06:19 Manifest0 left 06:22 Woodi joined 06:24 ChoHag joined, Manifest0 joined 06:31 feldspath left 06:35 feldspath joined 06:39 Manifest0 left 06:40 antoniogamiz joined 06:45 Manifest0 joined 06:48 jmerelo left 06:50 Manifest0 left 06:53 netrino joined, feldspath left 06:55 Manifest0 joined 06:56 abraxxa joined 06:59 feldspath joined 07:07 ChoHag left 07:08 Manifest0 left, robertle joined 07:10 lizmat joined 07:14 Manifest0 joined, feldspath left, feldspath joined 07:20 lizmat left, Manifest0 left 07:21 ChoHag joined 07:23 feldspath_ joined, woolfy left, feldspath left 07:25 Manifest0 joined
masak morning, #perl6 07:28
07:29 Actualeyes joined 07:30 Manifest0 left
antoniogamiz masak: o/ 07:33
mm is there an operator to compare lists despite the order of the elements?
07:36 rindolf joined 07:37 Manifest0 joined
masak antoniogamiz: if multiplicity also doesn't matter, I'd throw the lists into Sets, and compare 07:43
antoniogamiz: otherwise, call .unique on both and compare
07:43 Manifest0 left
masak hm, I wonder if one could write an `infix:<eqv>` multi with an `:under(&transform)` adverb... 07:44
07:45 feldspath_ left
masak m: multi infix:<eqv>($l, $r, :under(&transform)!) { transform($l) eqv transform($r) }; say([1, 2, 3] eqv [1, 3, 2] :under(*.unique)) 07:46
camelia 5===SORRY!5=== Error while compiling <tmp>
You can't adverb &infix:<eqv>
at <tmp>:1
------> 3[1, 2, 3] eqv [1, 3, 2] :under(*.unique)7⏏5)
antoniogamiz mmm ty masak :)
masak aww
why can't I adverb &infix:<eqv>? it'd be so much fun! :)
07:46 feldspath_ joined
masak I think I've hit this limitation before. can't adverb built-in operators 07:46
m: sub infix:<EQV>($l, $r, :under(&transform)!) { transform($l) eqv transform($r) }; say([1, 2, 3] EQV [1, 3, 2] :under(*.unique)) 07:48
camelia False
masak dunno why that one doesn't say `True`...
(halp)
having an operator like that would be genuinely useful -- not arguing for dumping it in CORE right now, but I'd sure like to be able to define it in my code 07:49
07:49 Manifest0 joined 07:50 netrino left 07:51 dolmen joined 07:55 qiqi left 07:56 feldspath_ left, Manifest0 left 08:01 Manifest0 joined
tobs m: sub infix:<EQV>($l, $r, :under(&transform)!) { transform($l) eqv transform($r) }; say([1, 2, 3] EQV [1, 3, 2] :under(*.sort)) 08:02
camelia True
08:02 antoniogamiz left
tobs masak: unique didn't do anything on those operands, you wanted to sort them 08:02
08:05 netrino joined
masak oh, right *facepalm* 08:05
masak can't even blame lack of coffee 08:06
08:07 netrino_ joined
masak anyway, adverbs are underused (and mostly for good reason; they're dangerous/weird), but this is exactly what they're good for 08:07
08:09 feldspath_ joined 08:10 netrino left 08:12 nadim_ joined 08:16 reach_satori joined 08:18 ChoHag left 08:19 ChoHag joined 08:24 Manifest0 left 08:25 Ekho left, zakharyas joined 08:26 qiqi joined, pecastro joined 08:31 Manifest0 joined 08:37 Ekho joined 08:38 Manifest0 left 08:43 Manifest0 joined 08:56 Manifest0 left 08:59 aindilis left, qiqi left 09:01 aindilis joined 09:02 Manifest0 joined 09:04 qiqi joined
Geth doc: lukasvalle++ created pull request #2796:
document antipairs for Setty		 09:13
09:13 SergiusUA joined 09:19 Manifest0 left 09:23 domidumont1 joined 09:24 Manifest0 joined 09:26 domidumont left, aindilis left 09:27 aindilis joined 09:31 qiqi left 09:34 Cabanossi left
Geth doc: 5ddf5a7585 | lukasvalle++ | doc/Type/Setty.pod6
document antipairs for Setty
 09:35
doc: af5b559f6a | lukasvalle++ | doc/Type/Setty.pod6
document antipairs for Setty,reorder in page
doc: a7c427895d | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | doc/Type/Setty.pod6
Merge pull request #2796 from lukasvalle/master

document antipairs for Setty
synopsebot Link: doc.perl6.org/type/Setty
09:36 rgrau left 09:38 qiqi joined 09:45 Manifest0 left 09:47 Cabanossi joined 09:50 Manifest0 joined 09:59 Manifest0 left 10:04 Manifest0 joined 10:17 sena_kun joined 10:19 Sgeo_ left 10:20 Sgeo_ joined 10:26 klapperl left
sena_kun 6c: sub a { fail "One" }; try a.new; say $!.message 10:28
committable6 sena_kun, ¦6c (37 commits): «One␤»
10:33 klapperl joined, Manifest0 left 10:38 Manifest0 joined 10:43 dolmen left, Manifest0 left 10:49 Manifest0 joined 10:58 Manifest0 left 11:01 zakharyas left 11:03 Manifest0 joined 11:05 filthyjvmuser joined
filthyjvmuser hello 11:05
moritz hi there. Nice nickname :D 11:06
filthyjvmuser thanks, it's very related to the questions i'm going to ask really quick
11:06 dolmen joined
filthyjvmuser can you use Java annotations on a Perl 6 function? and are there any documented examples of using Perl 6 to integrate within a more complicated Java framework? 11:08
specifically, i'd like to write plugins for this Minecraft server: www.spongepowered.org/
but i don't expect there to be any examples for that 11:09
but maybe there's something for Spring or another more common Java framework
timotimo oh, interesting. i don't think we have something for that yet
i.e. getting a function be annotated some specific way
filthyjvmuser got it 11:10
timotimo would be a fascinating puzzle to figure out i'm sure
you should be able to write a thin wrapper in java that just has the functions annotated with the annotations you need and they just call into perl6 code, though 11:11
11:12 qiqi left
filthyjvmuser that's still way better than writing Java directly :-) 11:12
thanks for the advice
timotimo i hope it works! :)
filthyjvmuser me too! thanks again :-) 11:13
11:13 filthyjvmuser left 11:19 Manifest0 left 11:24 Manifest0 joined 11:29 SergiusUA left 11:30 SergiusUA joined 11:31 Manifest0 left 11:34 Cabanossi left 11:36 Manifest0 joined 11:47 Manifest0 left 11:48 Cabanossi joined 11:52 Manifest0 joined
hahainternet i see there's a couple of language server implementations on github, is there one that people prefer? 12:14
12:19 Manifest0 left
Kaiepi m: my Instant $b = now; (my @ = (1..100)) for 1..10_000; say now - $b 12:20
camelia 0.118631
Kaiepi m: my Instant $b = now; (my @foo; @foo.STORE: (1..100)) for 1..10_000; say now - $b
camelia 0.12148629
timotimo m: my Instant $b = now; (my @foo; @foo.STORE: (1..100)) for 1..10_000; say now - $b
camelia 0.1338259
timotimo oops
m: my Instant $b = now; ((my @).STORE: (1..100)) for 1..10_000; say now - $b
camelia 0.116912
timotimo watch out for noisy timings on camelia 12:21
12:22 Guest5089 joined
Kaiepi .STORE gets called on assignment to @ sigilled variables right? 12:23
timotimo i believe so, yeah
it might be .STORE directly or it might be nqp::p6store
Kaiepi nqp::p6store?
timotimo find it in src/vm/moar/Perl6/Ops.nqp 12:24
it's a bit more than just calling .STORE
12:24 Manifest0 joined
timotimo first it checks if there is a container 12:24
if it's a container, it will nqp::assign instead 12:25
Kaiepi m: use nqp; my @foo; nqp::p6store(@foo, [1,2,3]); @foo
camelia WARNINGS for <tmp>:
Useless use of @foo in sink context (line 1)
Kaiepi m: use nqp; my @foo; nqp::p6store(@foo, [1,2,3]); say @foo
camelia [1 2 3]
Kaiepi m: my Instant $b = now; nqp::p6store(my @, (1...*)) for 1..10_000; say now - $b 12:26
camelia 5===SORRY!5=== Error while compiling <tmp>
Could not find nqp::p6store, did you forget 'use nqp;' ?
at <tmp>:1
------> 3nt $b = now; nqp::p6store(my @, (1...*))7⏏5 for 1..10_000; say now - $b
Kaiepi m: use nqp; my Instant $b = now; nqp::p6store(my @, (1...*)) for 1..10_000; say now - $b
camelia 0.0973934
Kaiepi m: use nqp; my Instant $b = now; ((my @).STORE: (1...*)) for 1..10_000; say now - $b
camelia 0.0988337
Kaiepi m: use nqp; my Instant $b = now; nqp::p6store(my @, (1...*)) for 1..100_000; say now - $b 12:27
camelia 0.6454862
Kaiepi m: use nqp; my Instant $b = now; ((my @).STORE: (1...*)) for 1..100_000; say now - $b
camelia 0.6283517
Kaiepi should bench my feed operator parallelization pullreq with nqp::p6store vs STORE method calls
even though they look to be the same speed 12:28
timotimo yeah, i wouldn't expect a speed difference
12:35 Manifest0 left 12:37 zakharyas joined 12:41 Manifest0 joined 12:47 Manifest0 left
Kaiepi timotimo, p6store is actually faster than the method call to STORE 12:51
timotimo that's a little odd 12:52
12:54 Manifest0 joined
Kaiepi this is what i used to bench hastebin.com/azijorohis.pl 12:54
nqp::p6store's average is 3 seconds faster than the STORE method call
's
timotimo oh
try switching the two run-with-blah around 12:55
also, you can drop the "gather" and "take" by just putting a "do" before the "for"
12:58 abraxxa left
timotimo not very important, just an optimization opportunity 13:01
13:02 Manifest0 left 13:05 MasterDuke left
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.('a') 13:06
camelia No such method 'CALL-ME' for invocant of type 'A'
in block <unit> at <tmp> line 1
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.'a'
camelia 5===SORRY!5=== Error while compiling <tmp>
Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.
at <tmp>:1
------> 3Int) { 1 } }; my A $a .= new; say $a.'a'7⏏5<EOL>
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.('A::a') 13:07
camelia No such method 'CALL-ME' for invocant of type 'A'
in block <unit> at <tmp> line 1
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.A::{'a'}
camelia No such method 'A' for invocant of type 'A'. Did you mean 'a'?
in block <unit> at <tmp> line 1
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.{'a'}
camelia Type A does not support associative indexing.
in block <unit> at <tmp> line 1
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.::{'a'}
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed class-qualified postfix call
at <tmp>:1
------> 3-> Int) { 1 } }; my A $a .= new; say $a.7⏏5::{'a'}
expecting any of:
dotty method or postfix
postfix
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.A::<a> 13:08
camelia No such method 'A' for invocant of type 'A'. Did you mean 'a'?
in block <unit> at <tmp> line 1
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.<a>
camelia Type A does not support associative indexing.
in block <unit> at <tmp> line 1
Kaiepi m: use nqp; my class A { method a(--> Int) { 1 } }; my A $a .= new; say nqp::callmethod(nqp::decont($a), 'a')
camelia 1
13:09 Manifest0 joined
Kaiepi m: my class A { method a(--> Int) { 1 } }; my A $a .= new; say $a.^lookup('a')($a) 13:15
camelia 1
Kaiepi surely there's a better way to call a method given a method name as a string 13:16
jnthn $a."$the-string"()
13:17 Manifest0 left 13:25 Manifest0 joined 13:29 curan left, jeromelanteri joined 13:30 kurahaupo_ joined 13:32 SergiusUA left 13:33 kurahaupo left 13:34 skids joined 13:43 SergiusUA joined 13:45 SergiusUA left, SergiusUA joined 13:57 Manifest0 left 14:02 Manifest0 joined 14:04 jelkner joined 14:11 Manifest0 left, aborazmeh joined, aborazmeh left, aborazmeh joined 14:14 nadim_ left
Geth doc: lukasvalle++ created pull request #2797:
add .map, .flat, .concise, and .summary for Backtrace		 14:15
14:16 Manifest0 joined 14:21 zacts joined 14:23 Manifest0 left 14:28 Manifest0 joined 14:30 domidumont joined
Geth doc: 3b6c405aa7 | Coke++ | doc/Language/control.pod6
fix typo
 14:32
synopsebot Link: doc.perl6.org/language/control
14:32 zacts left 14:33 domidumont1 left, pmurias joined
pmurias . 14:33
14:36 zacts joined 14:37 mornfall joined, lizmat joined 14:39 dolmen left 14:40 lizmat_ joined 14:43 powerbit left, dolmen joined, lizmat left 14:47 lizmat_ left 14:48 quipa joined 14:49 robertle left 14:55 Guest1336 left 14:56 aborazmeh left 15:06 feldspath_ left 15:11 Manifest0 left 15:15 sauvin left 15:16 Manifest0 joined 15:18 ravenousmoose_ joined 15:19 MilkmanDan left 15:20 MilkmanDan joined 15:23 domidumont left 15:27 sauvin joined 15:31 sauvin left 15:32 sauvin joined 15:35 sauvin left 15:40 zacts left 15:45 reach_satori left 15:47 jmerelo joined
jmerelo releasable6: status 15:48
releasable6 jmerelo, Next release in ≈34 days and ≈3 hours. 1 blocker. 140 out of 405 commits logged (⚠ 2 warnings)
jmerelo, Details: gist.github.com/8b1c52402183b1adfe...a8d05a09d3
15:53 sauvin joined 15:55 ravenousmoose_ left, zakharyas left 16:02 Manifest0 left 16:07 Manifest0 joined
El_Che we had a release? 16:12
Geth doc: f1e7907e66 | (JJ Merelo)++ | type-graph.txt
VM not correctly included in type-graph.txt
 16:13
16:14 ravenousmoose_ joined
sena_kun El_Che, we didn't, just the date was delayed to June 16:15
16:16 kurahaupo_ left, kurahaupo joined 16:21 Manifest0 left
Geth doc: 8c2c9828a0 | (JJ Merelo)++ | doc/Type/Iterable.pod6
Reflow and check definition
 16:24
doc: 7b9506cc08 | (JJ Merelo)++ | doc/Type/Test.pod6
Eliminates wrong preamble (for a method), closes #2799
synopsebot Link: doc.perl6.org/type/Iterable
Link: doc.perl6.org/type/Test
AndroidKitKat I know this is a noob question, but I'm not sure why I can't do what I'm doing on line 21 gist.github.com/c5ce0d0e8d16994e5b...e177eee30a
are members of an array not mutable? 16:25
jnthn No, scalar parameters to a block are immutable by default
jmerelo AndroidKitKat: that might be a list, not an array. lists are immutable
jnthn Use <-> instead of ->
jmerelo m: my @numbers; say @numbers.^name
camelia Array
jnthn Which will default to rw
jmerelo jnthn++ 16:26
AndroidKitKat Thanks
I'll probably have more questions later, I appreciate it
jmerelo AndroidKitKat: use this or StackOverflow (or both). You'll be very welcome.
sena_kun: what was the problem with this release?
sena_kun: there was only one blocker...
sena_kun jmerelo, no idea, ping kawaii 16:27
probably personal reasons
16:27 lizmat joined
jmerelo sena_kun: OK. 16:27
El_Che sena_kun: ah good, I thought I was out of the loop :)
16:29 lizmat left 16:30 lizmat joined 16:31 Manifest0 joined
namibj Fuck, I just dropped my phone screen-first from like 120~130cm onto an i think basalt paving stone. 16:32
Screen works, touch works, glass is pretty shattered. 16:33
Oh sorry, wrong chat ;)
16:35 dolmen left
AndroidKitKat Sorry to hear about your phone though 16:36
jmerelo AndroidKitKat: :-) 16:37
16:40 jeromelanteri left
kurahaupo jmerelo: have a break, have a … ? 16:53
16:53 pmurias left 16:55 robertle joined, zacts joined 16:56 domidumont joined
jmerelo kurahaupo: :-) 16:59
17:00 feldspath_ joined 17:01 natrys joined, reach_satori joined 17:04 feldspath joined 17:05 feldspath_ left 17:07 ravenousmoose joined 17:10 dolmen joined, ravenousmoose_ left, dolmen left, feldspath left
HarmtH Is there a way to set a default value for a multi-dimensional hash? 17:11
jmerelo HarmtH: as in, all values being the same? I don't think so. A multidimensional hash is a hash of hashes... So I don't see how you could do that. But I might be wrong. 17:12
HarmtH: why do you want to do that? 17:13
17:13 domidumont left, reach_satori left, reach_satori joined
HarmtH In my particular case? A programming puzzle. Have a 2-dimensional grid for which I would like a default value. 17:13
17:13 ravenousmoose left
jmerelo HarmtH: wouldn't a multidimensional array be better? 17:14
17:14 ravenousmoose joined 17:15 zacts left, reach_satori left, reach_satori joined
HarmtH Sometimes 17:15
But sometimes the array is sparse, or you'd like to use negative indices 17:17
17:17 reach_satori_ joined, a3f left 17:18 spaceotter joined, reach_satori left 17:19 satori__ joined
jmerelo lizmat has published this: modules.perl6.org/dist/Array::Spar...:ELIZABETH Maybe it can be useful. 17:21
spaceotter Hey guys I have a small issuse with the repl. If I type characters like üöä§ and want to delete them i get a wired questionmark :o
17:22 reach_satori_ left, satori__ left 17:23 satori__ joined
SyrupThinker HarmtH: Is the // operator viable for your usecase? 17:24
HarmtH SyrupThinker: Mostly, yes
SyrupThinker: But a default value is more elegant imho
jmerelo spaceotter: yep, that will happen in most consoles. I use the emacs console for that reason. 17:25
17:25 a3f joined 17:26 satori__ left
b2gills spaceotter: You could just keep hitting backspace until the question marks disappear. Or you could use `rlwrap` or I think the readline binding fixes that. (I use `$ rlwrap perl6` when I use the REPL.) 17:29
17:29 reach_satori joined
spaceotter Oh, good to know. Thank you guys :) 17:30
17:30 molaf joined, reach_satori_ joined 17:33 rindolf left, reach_satori left, molaf left 17:34 molaf joined 17:35 rindolf joined
Geth doc: 7bc42b8329 | (JJ Merelo)++ | doc/Language/exceptions.pod6
Reflow and some clarifications
 17:40
synopsebot Link: doc.perl6.org/language/exceptions
jmerelo Anyone can help with this? stackoverflow.com/questions/561906...ng-or-file
17:41 reach_satori joined, lala_ joined, reach_satori_ left, lala_ left
timotimo jmerelo: i think you have to set the :target to "syntaxcheck" in EVAL 17:44
what do you mean "safe", though?
17:45 reach_satori left, pmurias joined
timotimo because syntaxcheck will also cause some code to run 17:49
17:50 zakharyas joined
ugexe gist.github.com/ugexe/b3de0a3a78290061f116 :facemelt: 17:50
ah that doesn't take a string though 17:51
jmerelo timotimo: mainly not gobbling up a lot of memory like it happened here github.com/perl6/doc/issues/2764# 17:55
ugexe: right. It will maybe be a bit faster than perl6 -c, but I think the slowest part is creating the file and calling that on it. 17:56
timotimo: and I'm not sure that's documented... 17:58
timotimo huh, really?
18:00 quipa left
jmerelo timotimo: where's that :target defined? github.com/rakudo/rakudo/blob/1ec4...gnCode.pm6 18:01
tobs m: use nqp; say so try { nqp::getcomp('perl6').compile($_) } for 'my $s = π', 'my $s = π-'
camelia True
False
jmerelo tobs: hey, that's cool...
tobs that's what I managed to read, but if there's an option to EVAL that'd be much less nqp-y
timotimo it's part of %n
but apparently it calls compiler.compile without passing that 18:02
jmerelo timotimo: so you say it's not really working?
timotimo so actually there doesn't appear to be a way to reach it with the official EVAL sub
jmerelo timotimo: %n is not used actually there...
timotimo: that a bug? 18:03
timotimo not sure
ugexe I'm not sure if you'd want to reuse the same compiler object
tobs good point 18:04
timotimo yaeh, with nqp::getcomp you can get at everything the compiler offers
i think the compiler objects we currently have don't store any state that changes just because of compilation
so re-using it should be fine
on the other hand, the objects are so lightweight that making a new one should have barely any cost to it 18:05
ugexe i was thinking if you have 2 unrelated files, that both do something like change the grammar, which are both compiled with the same object
timotimo ah, no i believe the grammar, actions, and world object are built inside the compile method and not shared 18:06
and changes to the grammar like with slangs and such are lexically scoped, which means we make copies and revert later as well 18:07
jmerelo oh wow
that's fast
tobs++
tobs glad I could contribute a bit to my own issue :p 18:08
timotimo \o/
18:12 vrurg left
jmerelo tobs: :-) 18:12
18:13 vrurg joined, dotdotdot joined
ugexe fwiw you dont need to call .compile if you are just interested if it will .parse 18:16
jmerelo tobs: it's not failing for errors, however, I don't know why...
18:16 elmer667 joined 18:18 espadrine_ joined
jmerelo ugexe: thanks. But it's not failing, for some reason 18:18
20:01:11 tobs | m: use nqp; say so try { nqp::getcomp('perl6').compile($_) } for 'my $s = π', 'my $s = π-' │+squashable6 18:19
m: use nqp; say so try { nqp::getcomp('perl6').parse($_) } for 'bad code', 'my $s = π-'
camelia False
False
timotimo m: say bad code 18:22
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routines:
bad used at line 1. Did you mean 'bag'?
code used at line 1
ugexe fwiw probably need a `my $*LINEPOSCACHE;`
jmerelo m: use nqp; say so try { nqp::getcomp('perl6').parse($_) } for '2+3', 'my $s = π-' 18:23
camelia 1
False
jmerelo m: use nqp; say so try { nqp::getcomp('perl6').parse($_) } for 'bad code', 'my $s = π-'
camelia False
False
tobs that seems to work, or am I missing something?
18:24 sauvin left
ugexe m: use nqp; say nqp::getcomp("perl6").parse($_).so for q|bad code|, q|my $s = π| 18:24
camelia Dynamic variable '$*LINEPOSCACHE' not found
ugexe m: use nqp; my $*LINEPOSCACHE; say nqp::getcomp("perl6").parse($_).so for q|bad code|, q|my $s = π| 18:25
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routines:
bad used at line 1. Did you mean 'bag'?
code used at line 1
tobs ahh, I think .compile took care of setting that up, which is why it worked initially 18:26
ugexe well you used try, which i removed
tobs m: use nqp; say so nqp::getcomp('perl6').compile($_) for 'my $s = π'
camelia True
tobs m: use nqp; say so nqp::getcomp('perl6').parse($_) for 'my $s = π' 18:27
camelia 1
ugexe the LINEPOSCACHE is probably for reporting the error
tobs I see, that makes sense
18:28 CepriN joined
jmerelo ugexe: parse is faster and does not produce warnings, which were not caught by the redirected $*ERR since they were produced in NQP 18:29
ugexe: it's producing an error in "<anon|4047>::Metamodel::MethodContainer is not composable, so <anon|4071>::Metamodel::ParametricRoleHOW cannot compose it 18:30
I remember I fixed that somehow... That error was produced before with EVAL too.
18:31 SergiusUA left, espadrine_ left
jmerelo m: use nqp; say so nqp::getcomp('perl6').parse($_) for 'my $s = πp' 18:32
camelia Dynamic variable '$*LINEPOSCACHE' not found
jmerelo m: use nqp; my $*LINEPOSCACHE; say so nqp::getcomp('perl6').parse($_) for 'my $s = πp' 18:33
camelia 5===SORRY!5=== Error while compiling <tmp>
Undeclared routine:
πp used at line 1
jmerelo right
we need the try anyway to avoid that.
18:38 Manifest0 left
Geth doc: d51503be0d | (JJ Merelo)++ | 2 files
Using low-level parse to speed up things

Thanks to @ugexe and @taboege for hints on how to make this. I've checked memory, and it's actually taking quite a bit of memory, but not as much as EVAL and it's actually faster than that.
It's also producing a false positive with `class ... (6 more lines)		 18:38
jmerelo ugexe , taboege: you might want to answer it in StackOverflow. Thanks a lot. 18:39
tobs m: use nqp; say so try nqp::getcomp('perl6').parse('class Metamodel::ParametricRoleHOW {}') 18:41
camelia False
tobs m: use nqp; my $*LINEPOSCACHE; say so try nqp::getcomp('perl6').parse('class Metamodel::ParametricRoleHOW') 18:42
camelia False
tobs m: use nqp; my $*LINEPOSCACHE; say so try nqp::getcomp('perl6').parse('class Metamodel::ParametricRoleHOW {}')
camelia 1
tobs I don't know why it's 1 under `so` and why it again needs LINEPOSCACHE but there's a difference at least. 18:43
18:43 Manifest0 joined
tobs jmerelo: I might be able to fix that, having touched the heuristic that adds empty blocks after things to make them compile last squasthon 18:45
jmerelo tobs: that would be nice. I don't know why it happens, and precisely in that class. There are others like it 18:53
Geth doc: taboege++ created pull request #2803:
Point out difference between finite and infinite Range subscripts		 18:55
doc: 68fbdb7d40 | (Tobias Boege)++ | doc/Language/list.pod6
Point out difference between finite and infinite Range subscripts
 18:56
doc: d8b0b50613 | (Juan Julián Merelo Guervós)++ (committed using GitHub Web editor) | doc/Language/list.pod6
Merge pull request #2803 from taboege/subscript-range-whatever

Point out difference between finite and infinite Range subscripts
synopsebot Link: doc.perl6.org/language/list
jmerelo tobs: but you have privs, right?
tobs oh, then I'm not sure I understand what's going on, but will investigate…
18:56 zakharyas left
tobs jmerelo: do I? I'm "in the perl6 organization", but I'm not sure what it entails to be honest 18:57
(maybe whoever invited me placed too much trust in me)
jmerelo tobs: it entails a great honor and a great responsability :-)
tobs: but that includes perl6/doc, too. I'm happy to review your PRs if you're not sure, of course. 18:58
(but I'll also review most commits anyhow, so... )
18:58 jmerelo left
tobs .tell jmerelo good, will read up and keep that in mind for easy changes 18:59
yoleaux tobs: I'll pass your message to jmerelo.
19:04 CepriN left 19:10 Manifest0 left 19:12 spaceotter left 19:18 jelkner left, Manifest0 joined
tobs .tell jmerelo huh, the only mention of that class is in Type/Metamodel/ParametricRoleHOW.pod6 and that tested successfully 10 times in a row for me just now. Is it maybe flapping for you? 19:19
yoleaux tobs: I'll pass your message to jmerelo.
19:22 molaf left 19:29 Manifest0 left 19:31 lizmat left 19:32 lizmat joined 19:35 Manifest0 joined 19:36 lizmat left 19:38 reach_satori joined
ugexe timotimo: wouldn't augments live across compiler objects? 19:41
er, across compiles from the same compiler object 19:42
class Foo { }; EVAL(q|use MONKEY-TYPING; augment class Foo { method bar { say 42 } }|); Foo.bar 19:43
evalable6 42
ugexe that also shows a problem with using EVAL
so if you have two unrelated modules and one augments a core class, the second module will not have the expected class 19:44
m: use nqp; my $*LINEPOSCACHE; nqp::getcomp("perl6").compile(q|use MONKEY-TYPING; augment class Int { method bar { die 42 } }|); nqp::getcomp("perl6").compile(q|BEGIN say Int.bar|); 19:47
camelia 5===SORRY!5=== Error while compiling <tmp>
An exception occurred while evaluating a BEGIN
at <tmp>:1
Exception details:
42
in method bar at <tmp> line 1
in code at <tmp> line 1
ugexe ^
similar if calling .parse 19:49
when instead i think what is expected is an error about method bar not existing on Int
well, what is expected of a solution to the earlier problem rather 19:50
19:53 cpan-p6 left 19:54 elmer667 left
ugexe creating a new compiler object doesn't change the outcome though... 19:58
m: use nqp; use Perl6::Grammar:from<NQP>; use Perl6::Actions:from<NQP>; use Perl6::Compiler:from<NQP>; my $*LINEPOSCACHE;
camelia ( no output )
ugexe sub compiler() { my $c = Perl6::Compiler.new; $c.parseactions(Perl6::Actions.new); $c.parsegrammar(Perl6::Grammar.new); $c }; my $c1 = compiler(); my $c2 = compiler(); $c1.compile(q|use MONKEY-TYPING; augment class Int { method bar { die 42 } }|); $c2.compile(q|BEGIN say Int.bar|);
timotimo ugexe: ouch :) 20:02
20:02 HoloIRCUser joined 20:03 HoloIRCUser is now known as zzzzzzzzz, cpan-p6 joined, cpan-p6 left, cpan-p6 joined, zzzzzzzzz is now known as Guest62255
tobs hmm, could I change &ok to always succeed that way? 20:05
timotimo you could .wrap it 20:06
ugexe and if it was a class it would require supercedes instead of augments
20:06 Guest62255 left 20:09 zacts joined 20:15 Manifest0 left 20:17 xinming_ joined 20:19 xinming left 20:22 Manifest0 joined 20:23 espadrine_ joined 20:25 zacts left 20:27 pmurias left 20:37 reach_satori_ joined
Geth doc/master: 5 commits pushed by lukasvalle++, Altai-man++ 20:39
20:40 reach_satori left 20:56 dolmen joined 20:57 ravenousmoose left
Geth doc/note-start-special-vars: 9025ba0559 | Altai-man++ | doc/Language/control.pod6
Explain $! and $/ in start block
 21:02
doc: Altai-man++ created pull request #2804:
Explain $! and $/ in start block
21:03 skids left
sena_kun 6c: my $a = anon sub foo {}; my $b = anon sub bar {} 21:04
committable6 sena_kun, ¦6c (37 commits): «»
sena_kun 6c: my $a = anon sub foo {}; my $b = anon sub foo {}
committable6 sena_kun, ¦6c (37 commits): «»
21:06 zacts joined 21:07 Sgeo_ left, Sgeo_ joined 21:16 natrys left
sena_kun 6c: say (1, 2, 3).Seq eqv (1, 2, 3) 21:18
committable6 sena_kun, gist.github.com/2e8260c96c6cd73380...98a81d0bf3 21:19
21:20 molaf joined
sena_kun 6c: Num.new(5).WHAT.say 21:23
committable6 sena_kun, ¦6c (37 commits): «(Num)␤»
sena_kun 6c: say Nil.chrs; 21:24
committable6 sena_kun, gist.github.com/0e9e83fbbd5f32d98e...eb68ea2875
sena_kun 6c: say Hash<foo>:delete; 21:26
committable6 sena_kun, gist.github.com/faf21ad658eb570275...4ad153146b
21:30 zacts left
Geth doc/add-cando-example: b25e802cfe | Altai-man++ | doc/Type/Code.pod6
Add cando example with Block
 21:38
cpan-p6 New module released to CPAN! Font::FreeType (0.1.9) by 03WARRINGD 21:40
Geth doc: Altai-man++ created pull request #2805:
Add cando example with Block
21:43 dolmen left
Geth doc/pair-pair-method: 1719332eec | Altai-man++ | doc/Type/Pair.pod6
Document Pair.Pair method
 21:45
doc: Altai-man++ created pull request #2806:
Document Pair.Pair method
21:48 |oLa| joined 21:49 |oLa| left
Geth doc/Enumeration-Int-method: 619fa6a1d3 | Altai-man++ | doc/Type/Enumeration.pod6
Document Enumeration.Int
 21:59
doc: Altai-man++ created pull request #2807:
Document Enumeration.Int		 22:00
22:05 adu joined
sena_kun just loves it when theere are docs available 22:09
22:12 rindolf left 22:16 zacts joined 22:17 netrino_ left
Geth doc/note-start-special-vars: 1124099187 | Altai-man++ | doc/Language/control.pod6
Clarify a sentence; jnthn++
 22:17
jnthn sena_kun++ # working on docs 22:18
sena_kun slightly easier than fixing rakudo bugs. :)
sena_kun is still a bit sad about & && issue though
22:20 MasterDuke joined, MasterDuke left, MasterDuke joined
jnthn sena_kun: haha... github.com/rakudo/rakudo/blob/mast....nqp#L5585 22:20
22:20 rindolf joined
jnthn That's...suspect :) 22:21
A space matches \W
That's why 'a' & 'a' is OK but a & a is not :P
I bet if you tighten up that lookahead it'll be OK
sena_kun >a totally different bug location 22:22
www.youtube.com/watch?v=tXKTkZ8gyUA
jnthn Maybe try <+[\W]-[\s]> instead of the \W
sena_kun jnthn, thanks!
jnthn Though maybe <:punct> is better
sena_kun I'll try it out now...
22:25 woolfy joined
sena_kun yay, the first fix works. running spectest... 22:31
jnthn Cool; it might just have an impact on what's considered the declarative prefix, but I hope not an adverse one.
(In theory it shouldn't, but iirc there's some small gremlins there...) 22:32
sena_kun do you recommend <:punct> over it? 22:33
22:36 rindolf left 22:37 zacts left
jnthn sena_kun: Yes, but I might have totally made it up :) 22:40
m: say '!.?:*' ~~ /<:punct>+/ 22:42
camelia ｢!.?:*｣
jnthn Hm, seems workable
May be a bit cheaper, and should be free of LTM concerns 22:43
sena_kun is sad to kill spectest, so will keep it until the end to show respect
jnthn Hm, not sure we complete those <:uniprop> things in Comma yet. Another once for the wish list. :)
*one
sena_kun Result: FAIL 22:45
oops
t/spec/S05-mass/rx.rakudo.moar, no TAP plan found... trying out with <:punct> 22:46
jnthn Did it fix the bug despite the regression? :) 22:51
'cus if we so at least we know we're fixing the right place :)
sena_kun it did
jnthn I did a bigint opt today and failed precisely one spectest. It was an off-by-one. :) 22:52
(In my code)
sena_kun I also see tons of warnings like `Space is not significant here`, so I am rebuilding it with \W to check if they were there before...
yes, warnings were there from the start 22:53
jnthn btw, for those here who don't follow MoarVM/Rakudo guts: factor of > 2 speedup on Rat math is coming. :) 22:54
sena_kun jnthn, it means an opportunity to do some math grinding and not being too shy when folks ask e.g. "Do you have numpy here?"? 22:55
(yes, I know that numpy is based a lot on arrays, not just numbers, but it has plenty of math things included too) 22:56
jnthn Well, for that you probably want natives anyway 22:57
I did some opt on those recently, though there's certainly more to come :)
timotimo at some point i hope we can get 128bit floating points into moarvm :)
sena_kun isn't PEA based on lowering objects on heap into natives in registers? 22:58
(if you super-duper simplify it, just so I could get it a tiny bit)
timotimo that's the "scalar replacement" bit 22:59
PEA is the analysis that we do to figure out when that's possible, and when it has to be undone in the code for things to stay safe
23:00 zacts joined
timotimo EA would let us do scalar replacement when objects that are created will never be seen again, but PEA will allow us to do scalar replacement in cases where the object has some secape routes, but they aren't too commonly used 23:00
sena_kun well, I mean the "action" part, not the "deciding" part.
jnthn sena_kun: What timotimo said, but: yes, it's true that for Num it is a pretty much "straight" unbox, and once we get good at that then for scalars `num` vs `Num` will matter a bit less. BUT, with arrays, there's still quite a significant difference in storage size, and thus in cache efficiency. 23:01
sena_kun so the test case is fixed, but gist.github.com/Altai-man/0ced0c83...8e78bdd04e
jnthn Int vs. int is far less clear-cut because that's not a straight unboxing
sena_kun are .rakudo.moar files generated from .t ones or not? 23:02
jnthn Since Int is auto-upgrade to big integer
sena_kun seems like the answer is "yes"
jnthn And so we need upgrade checks in the JIT output
ugexe yes
jnthn sena_kun: make t/spec/foo/bar.t will do the fudge and then run the .rakudo.moar if one is produced; spectest does that at the start too; only ever edit the .t file 23:03
(the the .rakudo.moar is good to read) 23:04
('cus the line numbers match up)
sena_kun so it seems that `/a |& b/` regex cannot be parsed with <:punct> instead of \W.
jnthn What on earth...
What does the diff you have look like? 23:05
sena_kun or I might got the wrong line...
diff of the "fix"?
jnthn Yeah
But...is that even valid syntax?
sena_kun one second
jnthn The leading & is only allowed at the start of a regex for a [...] group, ain't it?
sena_kun gist.github.com/Altai-man/baf7fa4d...9f2c049284 23:06
^ diff
jnthn m: use Test; ok 'a' ~~ /a |⏏& b/, 'alternation and conjunction (|&)
camelia 5===SORRY!5===
Unrecognized regex metacharacter ⏏ (must be quoted to match literally)
at <tmp>:1
------> 3use Test; ok 'a' ~~ /a |7⏏5⏏& b/, 'alternation and conjunction (|&)
Unable to parse expression in single quotes; couldn't f…
jnthn m: use Test; ok 'a' ~~ /a |& b/, 'alternation and conjunction (|&)
camelia 5===SORRY!5=== Error while compiling <tmp>
Unable to parse expression in single quotes; couldn't find final "'" (corresponding starter was at line 1)
at <tmp>:1
------> 3|& b/, 'alternation and conjunction (|&)7⏏5<EOL>
expecting …
jnthn m: use Test; ok 'a' ~~ /a |& b/, 'alternation and conjunction (|&)'
camelia ok 1 - alternation and conjunction (|&)
sena_kun github.com/perl6/roast/blob/master.../rx.t#L828 <- from roast 23:07
jnthn I'm not sure I trust this test at all
Moment, let me check something
Yes, the "allow a leading alternation/conjunction operator" is a nibbler level thing, like I thought 23:08
23:09 woolfy left
jnthn m: use Test; ok 'a' ~~ /a | [& b]/, 'alternation and conjunction (|&)' 23:09
camelia ok 1 - alternation and conjunction (|&)
23:10 zacts left
jnthn Seems to me the test was passing for the completely wrong reason before now, and I'm not entirely sure its expectation is legitimate. 23:12
Though I wouldn't mind a second opinion...
sena_kun well, I see a commit...
github.com/perl6/roast/commit/ee1b...f26fd21c0b <- this one, by TimToady 23:13
jnthn ooh, that's an answer then
sena_kun it says that the precedence differs
jnthn Yeah, it's right on that
OK, so if we want to fix the immediate bug, we need to fix something else first :) 23:14
The something else being src/QRegex/P6Regex/Grammar.nqp in the NQP repo
sena_kun looked there from the start 23:15
hmm
jnthn Yeah, in nibbler at the moment we have an alternation of those things to just silently "eat" at the start
But I guess we need to move those 23:16
sena_kun jnthn, if you are fancy to bring the fight to the floor^W^W^W^W^W^W hunt it down, I'll present results nicely, otherwise I'll at least update the ticket description with findings
sena_kun .oO ( www.youtube.com/watch?v=NM7wpkL-Amw ) 23:17
jnthn sena_kun: If you fancy trying it: probably toss the group starting line 135 in nibbler. Then in termaltseq put [ <!rxstopper> '||' { $*SEQ := 1; } <.ws> ]? at the start. And in termalt put at the start [ <!rxstopper> '|' <.ws> ]?, etc. 23:18
sena_kun let's see... 23:19
jnthn So we just move the eating of the initial throw-away thingy into the rule that parses things at that level
Rather than doing them all at the top nibbler level
It's only termaltseq that is "weird" and sets a $*SEQ, the rest are just simple matchings.
sena_kun let's see if I can work as a simple diff applier here... 23:20
jnthn :) 23:21
sena_kun jnthn, is `etc.` and the end implies something after `termalt`? :) 23:22
jnthn I figure you've written enough grammars it makes some kind of sense too. Admittedly writing the grammar for grammars is a tiny bit meta, but that's compiler life...
sena_kun well, to be fully honest I cannot say I have a lot of concentration right now
jnthn sena_kun: Yes, the leading && handling goes in termconjseq, and the leading & in termconj :) 23:23
haha, me either :D
sena_kun ok, building nqp-rakudo and then trying out the test...
jnthn Spent all my brane on graph theory and the horrors of the x64 instruction set today :P 23:24
sena_kun I just a bit non-chalantly downloaded newest Idea... Just to see that we have a lot of exceptions on mere attempt to run Comma from run configuration 23:25
and they say "You need at least 183 to work with 191", but the truth it - it _doesn't_ work at all, you still need 191. 23:26
jnthn Oh heck, yes
sena_kun apparently failed as patch, because there is still a null 23:27
let me re-read it again and try to grasp...
jnthn And you can't use 191 to develop something built on an earlier enough platform version either. 23:28
Or at least, not without more fun than I was willing to have :)
sena_kun ah, I see what I did wrong, likely...
jnthn Feel free to send a diff if you want me to chck
*czech
oh goodness, *check 23:29
23:29 aborazmeh joined, aborazmeh left, aborazmeh joined
sena_kun gist.github.com/Altai-man/588562b1...c06cec79a4 23:30
23:30 pecastro left
jnthn Yeah, that's what I had in mind 23:30
sena_kun does a build/test 23:31
jnthn make test in NQP should give some clue of if it's right-ish :)
sena_kun nqp has a test suite itself?
I see, that's admirable
rx.t passes, regex.t passes. \o/ 23:33
running spectest again...
jnthn :) 23:41
sena_kun Unrecognized regex metacharacter @ (must be quoted to match literally) 23:48
sigh
gist.github.com/Altai-man/806a226e...e877455c9e <- output 23:49
23:50 Elronnd left
sena_kun `my @var = <a b ab c>; my \$aref = @var; 'a0' ~~ m/@\$aref[0]/` fails 23:51
m: my @var = <a b ab c>; my $aref = @var; 'a0' ~~ m/@$aref[0]/;
camelia Potential difficulties:
Apparent subscript will be treated as regex
at <tmp>:1
------> 3ab c>; my $aref = @var; 'a0' ~~ m/@$aref7⏏5[0]/;
sena_kun m: my @var = <a b ab c>; my $aref = @var; say 'a0' ~~ m/@$aref[0]/;
camelia Potential difficulties:
Apparent subscript will be treated as regex
at <tmp>:1
------> 3>; my $aref = @var; say 'a0' ~~ m/@$aref7⏏5[0]/;
｢a0｣
sena_kun github.com/perl6/roast/blob/master....t#L34-L35 <- those two are not too ok 23:54
m: '@' ~~ /<:punct>/ 23:58
camelia ( no output )
sena_kun m: say '@' ~~ /<:punct>/
camelia ｢@｣
sena_kun m: say '@' ~~ /\W/
camelia ｢@｣

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!