»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
gfldex m: role R {}; sub f(R $r){ dd $r}; my Int $v = 1 but R; f($v); 00:11
camelia rakudo-moar fa84f1: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Calling f(Int) will never work with declared signature (R $r)␤at <tmp>:1␤------> 3b f(R $r){ dd $r}; my Int $v = 1 but R; 7⏏5f($v);␤»
gfldex should that work?
m: role R {}; sub f($r){ dd $r}; my Int $v = 1 but R; f($v); dd $v, $v ~~ R; 00:12
camelia rakudo-moar fa84f1: OUTPUT«Int+{R} $r = 1␤Int+{R} $v = 1␤Bool::True␤»
BenGoldberg bisectable, m: my (::N); N::x 00:16
bisectable BenGoldberg: On both starting points (good=2015.12 bad=fa84f13) the exit code is 1 and the output is identical as well
BenGoldberg: Output on both points: Cannot call method 'AT-KEY' on a null object␤ in block <unit> at /tmp/sPmv84pXBd line 1␤
BenGoldberg bisectable, m: role Foo[$x] { }; say Foo[42].WHO.^name 00:21
bisectable BenGoldberg: On both starting points (good=2015.12 bad=fa84f13) the exit code is 0 and the output is identical as well
BenGoldberg: Output on both points: VMNull
BenGoldberg m: say chr 2**63 00:23
camelia rakudo-moar fa84f1: OUTPUT«␀␤»
BenGoldberg m: say chr 2**63-2
camelia rakudo-moar fa84f1: OUTPUT«chr codepoint cannot be negative␤ in block <unit> at <tmp> line 1␤␤»
BenGoldberg m: multi sub cross() { } 00:24
camelia ( no output )
BenGoldberg bisectable, multi sub cross() { }
bisectable BenGoldberg: On both starting points (good=2015.12 bad=fa84f13) the exit code is 0 and the output is identical as well
BenGoldberg: Output on both points:
BenGoldberg bisectable, m: for 0..0xffff -> $i { say $i if $i %% 100; my $c = try { :16(uniprop($i.chr, "Bidi_Mirroring_Glyph")).chr }; say "{$i.fmt("%04x")}: {$i.chr} ~ $c" if $c.defined} 00:25
bisectable BenGoldberg: On both starting points (good=2015.12 bad=fa84f13) the exit code is 0 and the output is identical as well 00:26
BenGoldberg: gist.github.com/e95d4d857b72a32284...9d74d3a598
BenGoldberg m: for 0..0xffff -> $i { say $i if $i %% 100; my $c = try { :16(uniprop($i.chr, "Bidi_Mirroring_Glyph")).chr }; say "{$i.fmt("%04x")}: {$i.chr} ~ $c" if $c.defined}
camelia rakudo-moar fa84f1: OUTPUT«(signal SEGV)0␤100␤200␤»
AlexDaniel commit releases for 0..0xffff -> $i { say $i if $i %% 100; my $c = try { :16(uniprop($i.chr, "Bidi_Mirroring_Glyph")).chr }; say "{$i.fmt("%04x")}: {$i.chr} ~ $c" if $c.defined} 00:29
ah right 00:30
commit: releases for 0..0xffff -> $i { say $i if $i %% 100; my $c = try { :16(uniprop($i.chr, "Bidi_Mirroring_Glyph")).chr }; say "{$i.fmt("%04x")}: {$i.chr} ~ $c" if $c.defined}
committable AlexDaniel: gist.github.com/539428256414be39c0...16bfa35e69
AlexDaniel BenGoldberg: bisectable does not catch SEGVs properly yet :( 00:31
… especially when they are fixed? 00:32
BenGoldberg But not fixed for camelia?
AlexDaniel m: for 0..0xffff -> $i { say $i if $i %% 100; my $c = try { :16(uniprop($i.chr, "Bidi_Mirroring_Glyph")).chr }; say "{$i.fmt("%04x")}: {$i.chr} ~ $c" if $c.defined}
camelia rakudo-moar fa84f1: OUTPUT«(signal SEGV)0␤100␤200␤»
AlexDaniel commit: HEAD for 0..0xffff -> $i { say $i if $i %% 100; my $c = try { :16(uniprop($i.chr, "Bidi_Mirroring_Glyph")).chr }; say "{$i.fmt("%04x")}: {$i.chr} ~ $c" if $c.defined}
committable AlexDaniel: ¦«HEAD»: 0␤100␤200 «exit signal = SEGV (11)»
AlexDaniel ah 00:33
AlexDaniel MasterDuke: am I reading that incorrectly or does the one above say that it is OK on HEAD? 00:34
MasterDuke: I mean this one: gist.github.com/Whateverable/53942...16bfa35e69
MasterDuke: I mean, if it does, then it's a bug :)
konobi .tell pmurias well, es6 has support on all modern browsers... though a few still have aspects that aren't implementent yet. Thankfully we can use babel to transpile from pure es6 to es5, or combo of es5/es6 for different browsers support (or use polyfills) 02:12
yoleaux konobi: I'll pass your message to pmurias.
31 Jul 2016 19:31Z <pmurias> konobi: I'm applying your refactoring piece by piece as there was a bit of things that where breaking stuff and also I prefer run the tests after relatively small changes rather then debug complex interactions
konobi .tell pmurias yup... it had all been working up until a small change around code-ref/core/sixmodel =0( 02:13
yoleaux konobi: I'll pass your message to pmurias.
konobi .tell pmurias i have a feeling it might be due to a missunderstanding somewhere of how the 'this' keyword works 02:30
yoleaux konobi: I'll pass your message to pmurias.
geekosaur dalek pinged out about 8 hours ago... 03:03
MasterDuke BenGoldberg: FYI, you were calling bisectable with an 'm: ' in front of your code. it should actually be valid Perl 6 (the 'm: ' will just be a label), but it's probably not what you intended 03:09
AlexDaniel: yeah, something seems off there 03:10
harmil Think I found my problem with an attribute becoming the type, Callable, but not sure why... 03:27
m: class NotComposed { has &.func = ~*; }; class IsRat does Rational[Int,Int] { has &.func = ~* }; class IsReal does Real { has &.func = ~* }; say "In NotComposed, func={NotComposed.new.func.perl}"; say "In IsRat, func={IsRat.new.func.perl}"; say "In IsReal, func={IsReal.new.func.perl}"
camelia rakudo-moar fa84f1: OUTPUT«In NotComposed, func={ ... }␤In IsRat, func=Callable␤In IsReal, func={ ... }␤»
harmil It seems that any class that does Rational has its callable attributes munged. 03:28
geekosaur thinks that looks like a rakudobug... 03:29
harmil Yeah, I was just checking to see if I was doing something stupid before sending email 03:30
andrzejku hi :) 05:06
holyghost hello 05:10
andrzejku I am looking for Perl friend 05:12
holyghost try #perl 05:14
lots of good mainframe basic dude around here :-) 05:15
andrzejku ok
holyghost s/dude/dudes
andrzejku thank you
andrzejku holyghost, you are very good 05:18
moritz \o 06:59
Zoffix o/
ufobat hello :-) 07:06
Woodi hi today :) 07:07
I wonder what unmapped memory is... FreeBSD got patch that improves mem performance 35% - 135% reviews.freebsd.org/D7195 do they just removed we-have-infinity-lot-of-memory layer ? 07:10
moritz "this can improve the performance 1.35X ~ 2.35x for FIO on hyperv." 07:12
what is FIO?
Woodi file IO ? 07:22
probably this wiki.mikejung.biz/Benchmarking 07:24
Zoffix New blog post "Hacking on The Rakudo Perl 6 Compiler: Mix Your Fix": perl6.party/post/Hacking-on-Rakudo-...x-Your-Fix 07:36
moritz Zoffix: you seem to be unaware of local ettiquette. It's called "lol I blug" here :-) 08:00
ab6tract good *, #perl6 09:04
i ran into something interesting while working on the Sereal port last Friday
an array of Nums failed to smartmatch against an array of equivalent Rats 09:05
calling .perl.EVAL on the array of Nums would make the smart match work
i'm not sure how to craft a minimal example for this, because every approach i have used to reproduce would Just Work, except for the original test case based on a real Sereal blob 09:06
eg: " not ok 3 - ARRAY -- got: $[1.23e0, 4.56e0, 7.89e0]expected: $[1.23, 4.56, 7.89]" 09:10
m: say $[1.23e0, 4.56e0, 7.89e0] ~~ $[1.23, 4.56, 7.89]
camelia rakudo-moar 4b934d: OUTPUT«True␤»
ab6tract so, wtf?
ab6tract too early for the heavy questions, maybe? 09:21
moritz ab6tract: what types do you get there? 09:25
ab6tract: are these "vanilla" Array objects, or parameterized? (Array[Num] or so)?
ab6tract moritz: they are vanilla. the data gets encoded in Sereal as a double and gets read back out through a C function that casts to double (NativeCall return type specified as num64) 09:26
ab6tract JSON::Tiny and JSON::Fast don't seem to support Nums, and only pass back Rats (the expected result is saved into a json file when generating the corpus of Sereal files (p5 script)) 09:27
moritz m: my num64 $n = 1e1; say [$n] ~~ [1e1]
camelia rakudo-moar 4b934d: OUTPUT«True␤»
moritz m: my num64 $n = 1e1; say [$n] ~~ [10.0] 09:28
camelia rakudo-moar 4b934d: OUTPUT«True␤»
ab6tract moritz: indeed. exactly as happens to me: any attempt to get a non-ok result from $[1.23e0, 4.56e0, 7.89e0] ~~ $[1.23, 4.56, 7.89] fails
and even calling .perl.EVAL on the object (should be a no-op, correct) makes it pass
ab6tract so it's unclear how to even create a portable test case 09:33
i can work around it by calling .Str.Rat on the result from the read_double C call 09:34
(custom function we wrote for getting doubles out of a blob)
ab6tract but that feels very much like a hack, because the expression works fine when i type it in manually! 09:35
moritz ab6tract: can you get a minimal example that mimics the double extraction?
s/get/produce/
ab6tract without the underlying C code? 09:39
was just about to try it with pack/unpack, but it doesn't seem to support nums yet 09:40
psch m: use NativeCall; sub fabs(num64 $) is native {*}; my $x = fabs(1e1); say [$x] ~~ [10.0] 09:54
camelia rakudo-moar 4b934d: OUTPUT«False␤»
psch m: use NativeCall; sub fabs(num64 $) is native {*}; my $x = fabs(1e1); say [$x] ~~ [10.0]; say $x 09:54
camelia rakudo-moar 4b934d: OUTPUT«False␤(Mu)␤»
ab6tract psch: lol, why didn't i think of that
psch m: use NativeCall; sub fabs(num64 $) is native returns num64 {*}; my $x = fabs(1e1); say [$x] ~~ [10.0]; say $x
camelia rakudo-moar 4b934d: OUTPUT«True␤10␤»
psch ab6tract: no idea :)
ab6tract damnit, same issue
fuck 09:55
so, that doesn't actually help in the end
psch well, it does help insofar the its probably not NC directly
s/the/that/
ab6tract the C code just does a memcpy from the relevant buf and it's return values are consistent with the inputs 09:56
what really gets my goat is the .perl.EVAL part... 10:00
psch m: use NativeCall; sub fabs(num32 $) is native returns num32 {*}; my $x = fabs(1e1); say [$x] ~~ [10.0]; say $x 10:05
camelia rakudo-moar 4b934d: OUTPUT«True␤10␤»
psch well, definitely weird :)
ab6tract i've noticed that other serialization libs just do .Str.Rat 10:11
i was wondering if that is based on consensus for a best approach or..
psch m: use nqp; use NativeCall; sub fabs(num32 $) is native returns num32 {*}; my $x = fabs(1e1); say nqp::what($x) 10:12
camelia rakudo-moar 4b934d: OUTPUT«(Num)␤»
ab6tract m: use nqp; use NativeCall; sub fabs(num32 $) is native returns num32 {*}; say nqp::what( fabs(1e1) ) 10:13
camelia rakudo-moar 4b934d: OUTPUT«(Num)␤»
psch i guess that also boxes on moar..? 10:14
ab6tract psch: sorry, how do you mean? 10:16
psch ab6tract: well, the native sub returns a native type 10:16
ab6tract: but Num isn't
fwiw, i'm not sure r-j does it differently - we can't really have a WHAT for native types i think 10:17
ab6tract right, yeah
psch j: use nqp; say nqp::what(my num32 $ = 1e1)
camelia rakudo-jvm cd19db: OUTPUT«(Num)␤»
ab6tract this actually makes sense to me 10:18
as otherwise you would never have a solid way to get the value into anything else
since you can't call a coercion method on an unboxed
timotimo i inputted my password wrong 2 times and now it seems like i'm getting to wait for an hour before i'm allowed to try again? 10:19
timotimo oh ... 10:20
dalek died again
no, i mean
hack.p6c.org died again
psch oh
i logged in like 20 minutes ago
$ uptime 10:20:38 up 25 days, 17:15, 4 users, load average: 6409.08, 6379.62, 6306.41
timotimo yeah 10:21
psch that i think is a record..?
timotimo *shrug*
i'm going to force-kill hack
moritz timotimo: thank you 10:22
timotimo no need for thanks 10:23
timotimo ... if only we could figure out what makes the virtual discs freak out like that? 10:23
TEttinger maybe thanking people fixes it but only temporarily
timotimo makes sense 10:24
moritz TEttinger: more permanent solutions would be very welcome, yes
TEttinger and saying "thank you" actually reduces the reliability of the discs. clearly we need to switch to "cheerio, guvnah!"
ab6tract psch: here's the C code -- github.com/ab5tract/Sereal/blob/pe...code.c#L86 10:25
and here's the sereal decoder -- github.com/ab5tract/Sereal/blob/pe...er.pm6#L64
just in case you get curious :) 10:26
moritz TEttinger: why do I get the distinct feeling you're trolling? 10:26
timotimo not trolling, just making light of the situation 10:28
TEttinger if I was trolling it would be a very different sense 10:33
joking is more transparent
trolling might be,
TEttinger hey can perl6 help me find this one prime number, it's like my favorite but I forgot it and it's 4242 decimal places 10:34
which would be followed by attempts at guessing primes and returned with "no, not it" 10:35
moritz I guess I just don't appreciate your sense of humor, then 10:38
TEttinger that's fine. I don't have a great one 10:39
I'm also very sleepy
good night
moritz sleep well, and dream of an improved sense of humor
TEttinger oh, long running process
this thing will stay online, watching, judging, in silence
ab6tract the intrigue continues... if I call .Str.Num on the num64 it *also* passes the smartmatch 11:45
ab6tract of course, smart match is not all that intelligent about arrays, so what one really wants to use is is-deeply 11:47
which fails when comparing Nums to Rats
m: use Test; is-deeply [1.1e0], [1.1], "the same?"; 11:48
camelia rakudo-moar 4b934d: OUTPUT«not ok 1 - the same?␤␤# Failed test 'the same?'␤# at <tmp> line 1␤# expected: $[1.1]␤# got: $[1.1e0]␤»
ab6tract m: say [ [], [], [] ] ~~ [ [], [ [] ], [] ] 11:48
camelia rakudo-moar 4b934d: OUTPUT«True␤»
ab6tract m: [ [], ['a'], [] ] ~~ [ [], [ ['a'] ], [] ] 11:49
camelia ( no output )
ab6tract m: say [ [], ['a'], [] ] ~~ [ [], [ ['a'] ], [] ]
camelia rakudo-moar 4b934d: OUTPUT«True␤»
ab6tract wow... is-deeply agrees with smartmatch here 11:50
great
masak m: say @?INC
camelia rakudo-moar 4b934d: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '@?INC' is not declared␤at <tmp>:1␤------> 3say 7⏏5@?INC␤»
masak no @?INC ?
do S11 and S28 need an update? or will @?INC come back? 11:51
ilmari m: say $*REPO.repo-chain 11:51
camelia rakudo-moar 4b934d: OUTPUT«(inst#/home/camelia/.perl6 inst#/home/camelia/rakudo-m-inst-2/share/perl6/site inst#/home/camelia/rakudo-m-inst-2/share/perl6/vendor inst#/home/camelia/rakudo-m-inst-2/share/perl6 CompUnit::Repository::AbsolutePath.new(next-repo => CompUnit::Repository::NQ…»
masak I just wanted to know whether '.' was in @?INC (which I guess it's not) 11:52
ilmari there's no '.' my $*REPO.repo-chain 11:53
ilmari AFAICT 11:53
jnthn No, @?INC won't come back :) 11:56
You'd have to use lib '.' to get the cwd into the repo chain 11:57
masak I just wanted to check that it wasn't there by default
ilmari and I see 'use lib' resolves it to an absoulte path at use time, unlike perl5 11:58
ab6tract m: my @a = [ [], ['a'], [] ]; my @b = [ [], [ ['a'] ], [] ]; say @a ~~ @b; use Test; is-deeply @a, @b, "how are these the same?" 12:01
camelia rakudo-moar 4b934d: OUTPUT«True␤ok 1 - how are these the same?␤»
ab6tract why is it that i always manage to use p6 in a way that shows weird and disturbing seams? 12:02
masak m: say [ ['a'] ].perl 12:04
camelia rakudo-moar 4b934d: OUTPUT«["a"]␤»
masak ab6tract: it's the one-element rule you're encountering
m: my @a = [ [], ['a'], [] ]; my @b = [ [], [ ['a'], ], [] ]; say @a ~~ @b; use Test; is-deeply @a, @b, "how are these the same?"
camelia rakudo-moar 4b934d: OUTPUT«False␤not ok 1 - how are these the same?␤␤# Failed test 'how are these the same?'␤# at <tmp> line 1␤# expected: $[[], [["a"],], []]␤# got: $[[], ["a"], []]␤»
ab6tract fair enough, i thought that might be the case 12:05
masak I don't know if there's a good doc page somewhere explaining that...
...but that's clearly what's happening
ab6tract masak: unfortunately this doesn't explain my "array of Nums fails to smartmatch with array of Rats unless that array of Nums has .perl.EVAL called on itself" 12:06
masak m: say [1e0, 2e0] ~~ [1.0, 2.0]
camelia rakudo-moar 4b934d: OUTPUT«True␤»
masak worksforme :)
ab6tract that is exactly my issue 12:07
masak you sure you're not experiencing floating-point fuzziness or something? 12:07
ab6tract m: not ok 3 - ARRAY -- got: $[1.23e0, 4.56e0, 7.89e0]expected: $[1.23, 4.56, 7.89]
camelia rakudo-moar 4b934d: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Two terms in a row␤at <tmp>:1␤------> 3 ARRAY -- got: $[1.23e0, 4.56e0, 7.89e0]7⏏5 expected: $[1.23, 4.56, 7.89]␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤…»
ab6tract oops
" not ok 3 - ARRAY -- got: $[1.23e0, 4.56e0, 7.89e0]expected: $[1.23, 4.56, 7.89]"
now, try to craft a case where those two don't pass ok
masak: and if I call .Str.Num on those Nums it will pass the smartmatch 12:08
masak m: say [1.23e0, 4.56e0, 7.89e0] ~~ [1.23, 4.56, 7.89] 12:09
camelia rakudo-moar 4b934d: OUTPUT«True␤»
masak dunno what you're doing, but it works here
ab6tract that's my point, masak. i am running into an error case
you cannot get those to fail smartmatch
masak are you sure your test is using smartmatching?
ab6tract yes
masak wait, you want them to *fail* smartmatching?
m: say 1.23e0 ~~ 1.23 12:10
camelia rakudo-moar 4b934d: OUTPUT«True␤»
ab6tract no, i want them to pass. and they pass when inputting manually
they only fail in the code that i am testing
but even the got and expected values will pass smartmatch when put in manually
and the code i am testing will pass smartmatch if i do $a_of_nums.perl.EVAL ~~ $b 12:11
masak what's the name of the assertion sub you're calling?
ab6tract but not if i do $a_of_nums ~~ $b
'ok'
masak aha, and the expression you pass contains a smartmatch?
ab6tract yup 12:12
masak how have the Nums been produced?
ab6tract from a C routine: github.com/ab5tract/Sereal/blob/pe...code.c#L86
masak the plotchki thickens
ab6tract here is the nativecall definition: github.com/ab5tract/Sereal/blob/pe...ve.pm6#L21 12:13
and here is the line of code that pulls a double from the buf
github.com/ab5tract/Sereal/blob/pe...er.pm6#L39
masak ok, I now urge you to golf this
like, down to something ridiculously small, like three lines of code
ab6tract really not sure how to golf something that needs an object file
masak you'll think of something, I'm sure :D 12:14
ab6tract doing it just with NativeCall doesn't trigger the same issue
masak right now there is too much code not contributing to the bug
ab6tract right, but the smallest possible test case still requires a C compilation phase 12:16
i'll see what i can do
masak ++ab6tract 12:16
sometimes during especially intense golfing sessions, the C compilation phase (or equivalent) goes away 12:17
ab6tract m: use nqp; use NativeCall; sub fabs(num32 $) is native returns num32 {*}; say nqp::what( fabs(1e1) ); say [ fabs(1e1) ] ~~ [ 10.0 ] 12:20
camelia rakudo-moar 4ee104: OUTPUT«(Num)␤True␤»
ab6tract m: use nqp; use NativeCall; sub fabs(num32 $) is native returns num64 {*}; say nqp::what( fabs(1e1) ); say [ fabs(1e1) ] ~~ [ 10.0 ] 12:21
camelia rakudo-moar 4ee104: OUTPUT«(Num)␤False␤»
ab6tract m: use nqp; use NativeCall; sub fabs(num64 $) is native returns num64 {*}; say nqp::what( fabs(1e1) ); say [ fabs(1e1) ] ~~ [ 10.0 ]
camelia rakudo-moar 4ee104: OUTPUT«(Num)␤True␤»
ab6tract whoa
should i expect the second one to work?
ab6tract answer: no, because num32 <> num64 and thus it gets mangled on way through the fabs routine 12:25
naser hello 12:31
what are some good books on perl6, i found some on amazon but they are 10+ years old..
tadzik there aren't any 12:32
tadzik whatever's in perl6.org/documentation/ is your best bet at learning it :) 12:32
kalkin- Is it true that TT is working on a Perl6 book?
naser I am feeling confused about this language, looks like an old yet new language, no books except few old ones.. 12:33
perlpilot naser: if the books you have found on amazon are "Perl 6 and Parrot Essentials" and "Perl 6 Now", the first is no longer relevant and the second is more of a Perl 5 book :)
kalkin- naser: What helped me a lot is reading learnXinYminutes and the docs. Anything beyond that, wouldn't be in a Beginners Book anyway. 12:34
This channels is also very helpful and filled with kind people 12:35
What I personally missing the most a Good Practice guide, to learn what is the Perl6 way to do this or that, but there can be a Good Practices guide, because now one really knows how Good Practices should be. 12:36
naser i wish i could find something i can view offline on an e-reader.. 12:38
probably i could scrap the content of perl6 site and turn it into an epub document...
jeek www.facebook.com/DaTweekaz/videos/...034206794/ 12:39
asciidoc converts to epub pretty easily
kalkin- pod2mobi would be nice, to read the docs on the kindle 12:40
I like the way you can read The Art of Unix Programming or Ruby on Rails docs on a Kindle. It is very comfortable 12:41
apejens kalkin-: it looks like metacpan.org/pod/EBook::MOBI kan do pod -> mobi? It has a plugin for it, and a small example, but might need to write the actual script `pod2mobi` 12:50
kalkin- apejens: My impression was that Pod5 and 6 differ, am I wrong? 12:51
apejens ahh, lol, I didn't see I was in #perl6, sorry!
kalkin-: but perhaps a pod6 input plugin wouldn't be too hard to write 12:52
kalkin- apejens: i will look in to it, when i have time. Thanks for the hint
[Coke] .tell pmurias I'm ok with using a cutting edge JS. There are compilers to get us to an older version if needed, no? (I think getting something working all the way through the stack is more important at this point, so if the newer JS makes your job easier...) 13:25
yoleaux [Coke]: I'll pass your message to pmurias.
avar Is Inline::Perl6 unmaintained? Trying to compile it on Rakudo star 2016.07 yields an error about a missing 'frames' struct member. 14:03
Looking at moarvm.git jnthn removed that back in march in 824748d
[Coke] buggable, source? 14:04
avar;it's not tested as part of the rakudo release process. 14:05
avar Which is apparently an internal API now, so my attempt to patch this to use the new API failed: metacpan.org/source/NINE/Inline-Pe...rl6.xs#L91 14:05
avar Right, but it's also looking like I'm the only one who's tried to install this on a star release in ~6 months :){ 14:06
ab6tract :S 14:07
ugexe ahem i think you mean 4 months and 28 days 14:08
avar Anyway, a monkeypatch of getting rid of everytihng after ? seems to "work": metacpan.org/source/NINE/Inline-Pe...rl6.xs#L91
avar nine: *ping* 14:20
[Coke] avar; can you open an issue for this here: github.com/niner/Inline-Perl6 so nine sees it?
heh. didn't see your ping yet. :)
avar will open an issue anyway..
avar github.com/niner/Inline-Perl6/issues/1 :) 14:22
[Coke] avar++ 14:23
nine: might want to update links in the p5 distro of I:6 so that it points back to github. 14:24
(metacpan wants me to open a ticket in cpan's RT. 14:25
avar [Coke], nine : I also opened this issue asking "why the multiplicity?" github.com/niner/Inline-Perl6/issues/2 14:29
kid51 First time I've ever seen Perl 6 in a job posting: jobs.perl.org/job/20342 14:30
[Coke] kid51: woohoo
kid51: I'm actually using Perl 6 in a project at work myself, very happy. 14:31
ugexe wanted: senior perl6 web developer
TimToady maybe someone just fatfingered it :) 14:33
moritz do I qualify if I'm a senior web developer, and a senior Perl 6 developer? :-)
[Coke] TimToady: killjoy. :P 14:34
stmuk_ it says PHP 7 as well .. just looks like a BS jobpost mangled by an agent or manager 14:36
nine avar: thanks for the reports! I'm a bit ill right now but will have a look once I feel better. Inline::Perl6 hasn't seen much action as you may actually be the very first user ;) 14:45
avar nine: Yeah no rush at all, I'm just playing with this stuff for a hobby project 14:53
ab6tract well, some of us colleagues of avar have evil-er plans in mind ;) 15:04
but it isn't a rush in our case, either 15:05
pmurias maybe that job post is a bait and switch trap, they will promise you Perl 6 and make you write PHP? ;) 15:10
yoleaux 02:12Z <konobi> pmurias: well, es6 has support on all modern browsers... though a few still have aspects that aren't implementent yet. Thankfully we can use babel to transpile from pure es6 to es5, or combo of es5/es6 for different browsers support (or use polyfills)
02:13Z <konobi> pmurias: yup... it had all been working up until a small change around code-ref/core/sixmodel =0(
02:30Z <konobi> pmurias: i have a feeling it might be due to a missunderstanding somewhere of how the 'this' keyword works
13:25Z <[Coke]> pmurias: I'm ok with using a cutting edge JS. There are compilers to get us to an older version if needed, no? (I think getting something working all the way through the stack is more important at this point, so if the newer JS makes your job easier...)
pmurias .tell konobi arrow function having different this semantics was one of the problems 15:15
yoleaux pmurias: I'll pass your message to konobi.
dalek c: 245236e | (Tom Browder)++ | doc/Language/tables.pod6:
correct example table
 15:21
c: 7632895 | (Tom Browder)++ | doc/Language/tables.pod6:
correct spelling
 15:22
tbrowder don't see pragma when searching docsl zoffix's last blog post is great but first time i've seen "use monkey-guts" mentioned to enable nqp in p6 15:43
tailgate say I have an array @r = <(a b) (c d) (e f)>. Can I name the tuple values in a lambda, like @r.map({->($x, $y) $x ~ $y})? 15:45
jnthn Yes, pretty much like you did it, just need the block around the lambda body 15:47
Also, <(a b) (c d) (e f)> looks a bit confused :)
tailgate jnthn: lol I just figured that out
thnaks 15:48
jnthn m: my @r = <a b>, <c d>, <e f>; say @r.map(-> [$x, $y] { $x ~ $y })
camelia rakudo-moar 4ee104: OUTPUT«(ab cd ef)␤»
tbrowder ref rakudo dev: i've installed panda now system-wide (using its bootstrap instructions) along with rakudo head; however, panda can't find the source of p6 modules. do i have to download my own p6 module repo so "panda install" will work? 15:48
gtodd1 any heavy git users have an opinion on the best/simplest/safest way to maintain source control for an application developing and using/working with at the same time? 16:02
[Coke] tbrowder: what are you doing that can't find the source of p6 modules? 16:02
also, are you behind a firewall?
gtodd by that I mean say I'm writing a module and using it in an app ... eventually I will have a repo for my application based on Module.p6m and a repo for Module.p6m 16:03
[Coke] I would just start out with repos for both things 16:04
gtodd I was thinking of instead of pushing an pulling a ton up and down from bitbucket or github I could just have a branch "myapp" I switch into to add my own configuration etc. and then the main branch for the module itself ... 16:05
I guess if I'm going to go that way eventually I might as well do it now
and have two repos
jnthn Provided the module is in a folder and isolated, it's possible to extract its history separately using git filter-branch later 16:06
If you go the two repo route from the start, git subtree may help make it less annoying 16:07
gtodd ok ... I thought I could have a branch specific remote so when I pushed it would be automagically pushing the right stuff to the right place
it just doesn't seem big enough yet
jnthn Remtoes are always branch specific anyway :)
Well, uh
gtodd Oh good :) ... 16:08
jnthn Branches each have their own remote branch, and those might have different remotes
pmurias [Coke]: it should be possible to compile the runtime parts using the new javascript into the old version
jnthn I'd probably use subtree or something though...
Or just fix it later with filter-branch
dalek c: f39e6d4 | Altai-man++ | doc/Type/A (3 files):
Now examples from pages of Any/Array/Attribute are compilable. See github.com/perl6/doc/issues/776
 16:09
gtodd jnthn: OK thanks the other bit of the puzzle is I have a number of toy apps (in perl6 perl5 Catalyst etc.) that are set up specifically to deploy or run on heroku or aws and where I want to abstract out the the non-specifc to me bits in order to have a public branch I can push to git-hub say 16:11
gtodd jnthn: filter-branch seems interesting 16:11
I think I will do one that way and one with 2 repos and see which is more "natural" 16:12
for each application :)
jnthn :)
dalek c: 4258331 | (Wenzel P. P. Peppmeyer)++ | util/extract-examples.p6:
extract example based on .modified unless --force
 16:14
jdv79 unmatched}: your blogging pace is impressive 16:31
moritz not just pace, also quality 16:34
jdv79 i already commended him on that point months ago:) 16:35
but yes, of course
and very much an inspiration to try to do more 16:36
tbrowder [Coke}: I am no longer using rakudobrew. I am using the head of rakudo (git clone) and installing it in a system-wide location (with suitable path in my .bash_aliases). I installed panda from a git clone using its bootstrap and it also is installed system-wide using its default path (which I also have in my bash_aliases). Yhus panda and perl6 (and nqp, 17:05
etc.) are available from any dir. The problem is panda can't find modules when I command installation, so I need help to fix my dev environment so all play nicely together.
s/Yhus/Thus
tbrowder [Coke]: see finger fumble above 17:06
[Code]: no firewall 17:07
dalek c: 5ae4076 | (Wenzel P. P. Peppmeyer)++ | doc/Language/operators.pod6:
make operators.pod6 compile
 17:44
[Coke] tbrowder: I still don't understand what you mean. 17:51
can't find a module to -install-? or cannot find a module to use?
and rakudobrew or not, if you installed panda with the perl6 you already isntalled, you should be ok.
if it's somehow tied to the old rakudobrew you mentioned, I wouldn 17:52
t expecet that to work
konobi moo 18:56
yoleaux 15:15Z <pmurias> konobi: arrow function having different this semantics was one of the problems
dogbert17 o/ #perl6 19:08
huggable buggable 19:09
huggable: buggable
dogbert17 have written up some docs for Pair.invert. Comments welcome : gist.github.com/dogbert17/211c3769...a826ecc16c 19:11
perlpilot dogbert17: what about an example with ('Perl' => %hash).invert ? 19:15
or even ('Perl' => { foo => "bar", baz => "blat" }).invert # I'm not sure your description is quite right either. I think it's not about Positional, but rather Iterable 19:18
dogbert17 perlpilot: that might force me to make some changes :) 19:19
perlpilot: I believe that you're right, back to the drawing board then 19:20
m: say ('Perl' => (1..5)).invert 19:21
camelia rakudo-moar 5c5ef1: OUTPUT«(1 => Perl 2 => Perl 3 => Perl 4 => Perl 5 => Perl)␤»
dogbert17 yeah, the positional rubbish must go 19:22
hankache good evening #perl6 19:24
perlpilot hello hankache 19:25
hankache hi perlpilot
lizmat starts working on the P6W 19:30
hankache lizmat++
masak ++lizmat 19:32
hankache [+] (lizmat) 19:33
lizmat feels identified
AlexDaniel m: say [+] ‘lizmat’ 19:34
camelia rakudo-moar 5c5ef1: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5lizmat' (indicated by ⏏)␤ in block <unit> at <tmp> line 1␤␤Actually thrown at:␤ in block <unit> at <tmp> line 1␤␤»
AlexDaniel m: say [-] ‘lizmat’
camelia rakudo-moar 5c5ef1: OUTPUT«lizmat␤»
hankache how come? ^^^ 19:35
AlexDaniel hankache: #128757 #128758 19:36
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128757
Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128758
bdmatatu Hi all, is it possible to inherit from Date but override the formatter?
TEttinger lizmat is a number in radix 36 or higher 19:37
jdv79 is react a "supply" block? 19:38
AlexDaniel m: say :36<lizmat>
camelia rakudo-moar 5c5ef1: OUTPUT«1301684645␤»
zostay jdv79: yes, it just doesn't return a Supply
gfldex bdmatatu: see github.com/rakudo/rakudo/blob/nom/...ate.pm#L24 19:39
timotimo yup, it immediately noms all the values that you'd normally get from the supply block
AlexDaniel m: say :35<lizmat>
camelia rakudo-moar 5c5ef1: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Couldn't process entire number: 2/6 int chars, 2/-1 fractional chars␤at <tmp>:1␤------> 3say :35<lizmat>7⏏5<EOL>␤»
AlexDaniel ↑ what?
2/-1 fractional chars? What is it all about?
lizmat AlexDaniel: perhaps because that excludes "z" Or do you mean the LTAness ? 19:40
dogbert17 perlpilot: have updated the gist, i.e. gist.github.com/dogbert17/211c3769...a826ecc16c (feels like I have entered deep waters ...)
AlexDaniel lizmat: LTAness
TEttinger yeah, radix 35 would be 0-9a-y, but agreed on the unusual error
AlexDaniel I feel like I already rakudobugged this…
TEttinger it does say the general point though, that it can't process the number
bdmatatu Thanks gfldex, looks like inheriting from that is a bit tricky 19:41
gfldex bdmatatu: all you need to do is provide your own .new 19:41
lizmat bdmatatu: looks to me that you don't need to inherit from Date, but just specify a :formatter parameter ?
AlexDaniel hmm maybe not 19:42
bdmatatu I was going to use the same formatter in a few places, so thought a class would be better.
perlpilot dogbert17++ Looks good to me. I don't have any other commentary. :)
tx0h hi
dogbert17 perlpilot: :) thx
lizmat class MyDate is Date { method new(|c) { self.Date.new(|c, :formatter( { ... } ) } } 19:43
self.Date::new
tx0h i'm not sure if i have overseen something, but is there no way in bailador to use a scalar as a Template::Mojo ? 19:44
jdv79 are supply and react doc'd?
search seems to fail me
dalek c: 522a231 | (Jan-Olof Hendig)++ | doc/Type/Pair.pod6:
Added docs for Pair.invert. perlpilot++
 19:46
bdmatatu lizmat: thanks, works great! 19:47
dalek c: 5f18bbf | (Wenzel P. P. Peppmeyer)++ | doc/Language/operators.pod6:
better example
c: ca3eef2 | (Wenzel P. P. Peppmeyer)++ | Makefile:
add make examples-clean
dalek c: 71caed5 | (Wenzel P. P. Peppmeyer)++ | doc/Language/concurrency.pod6:
add whenever and react to index
 19:50
gfldex jdv79: ^^^
jdv79 thanks 19:51
pochi m: my $a; $a = $a+1 19:52
camelia rakudo-moar 5c5ef1: OUTPUT«Use of uninitialized value of type Any in numeric context in block <unit> at <tmp> line 1␤»
pochi perl5 tells me which variable it thinks is uninitialized: "Use of uninitialized value $a in addition (+) at -e line 1." 19:53
hoelzro perl 6 does, sometimes
pochi would that be possible in perl6 as well?
hoelzro m: my $a; say "$a"
camelia rakudo-moar 5c5ef1: OUTPUT«Use of uninitialized value $a of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at <tmp> line 1␤␤»
hoelzro it is possible, we just need to find the right code to change =) 19:54
pochi :-)
travis-ci Doc build errored. Wenzel P. P. Peppmeyer 'add whenever and react to index' 19:59
travis-ci.org/perl6/doc/builds/148988171 github.com/perl6/doc/compare/ca3ee...caed56380e
dogbert17 is "string".comb the fastest way to split a string into a list of characters or might there be an alternative hiding in the docs somewhere? 20:02
hoelzro huh, turns out that's super easy to add (the uninitialized value variable name thing) 20:03
pochi: want to make a contribution to Perl 6? =)
pochi I was looking in Mu.pm :-)
hoelzro that's the place to look! 20:04
pochi compiling ...
hoelzro want more hints, or would you prefer to figure it out on your own?
pochi I just tried to copy-paste the code from the string case 20:05
sena_kun I don't see anywhere in the docs(grep too) usage of "=begin output". Why nobody used it over years? It seems to be a valid pod code. 20:05
hoelzro pochi: did it work?
sena_kun: that's the POD code for "example output of a chunk of code", right? 20:06
sena_kun hoelzro, sure.
pochi hoelzro: still compiling
hoelzro sena_kun: yeah, it seems like it would be very useful
gfldex sena_kun: did you check if the renderers support it too?
sena_kun gfldex, nope. That's why I'm asking. (: 20:07
gfldex sena_kun: i believe they don't. It wouldn't be hard to change that tho. 20:07
sena_kun Let's try it locally... 20:07
gfldex, it's just we often use `code` block to show output. The easy solution is to just do a `:skip-test` and to worry, but I'm interested. 20:09
pmurias hi 20:10
sena_kun s/to/don't
Ah.
pochi hoelzro: no go :-( Seems $name just is the empty string 20:12
hoelzro pochi: could you show me the diff of the change you made? 20:18
AlexDaniel m: sub hello { say 42 }; say &hello 20:20
camelia rakudo-moar 5c5ef1: OUTPUT«sub hello () { #`(Sub|54567296) ... }␤»
AlexDaniel how can I do the same thing with a method?
and is it even possible? 20:21
m: class Foo { method hello { say 42 } }; my $x = Foo.new; say $x.&hello
camelia rakudo-moar 5c5ef1: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Undeclared routine:␤ hello used at line 1. Did you mean 'shell'?␤␤»
gfldex AlexDaniel: you have to use the MOP
jnthn Foo.^lookup('hello') 20:22
AlexDaniel m: class Foo { method hello { say 42 } }; my $x = Foo.new; $x.^lookup(‘hello’)($x) 20:23
camelia rakudo-moar 5c5ef1: OUTPUT«42␤»
AlexDaniel so does not work as nicely on an instance 20:24
pochi hoelzro: pastee.org/3ncbm
jnthn "does not work as nicely"?
Works just the same
AlexDaniel jnthn: well, I have to pass the object itself if I want to call it 20:25
tbrowder [Coke]: Sorry, when I first installed panda I tried to install a module (don't remember which) and panda couldn't find it. I just tried "panda install Linenoise" and it recognized it being installed. Sorry for the wasted bandwidth--looks like a non-problem now.
jnthn Sure, methods need an object to work on.
AlexDaniel jnthn: sure, I just hoped that it will be done for me
jnthn Ah. That's now how Perl 6 OO works :) 20:26
And if you do want that, closures are really compact in Perl 6.
hoelzro pochi: that's weird - a similiar patch worked for me
AlexDaniel yeah
jnthn {$x.hello} is quite a bit shorter than $x.^lookup('hello') :)
pochi hoelzro: You are right, it works on the 'my $a ...' example :-) 20:27
gfldex jnthn: .assuming is defined for Routine, that includes Method, but there is no easy way to call that method.
zostay tx0h, i'm not sure i understand the question, the call to template() in bailador takes a string (scalar) containing the template, doesn't it?
AlexDaniel jnthn: except that .hello is a multimethod in my case, and it seems like it will be a bit longer to preserve that
pochi hoelzro: but not on my actual usecase :-)
jnthn m: say '-> |c { $x.foo(|c) }'.chars 20:29
camelia rakudo-moar 5c5ef1: OUTPUT«20␤»
pochi m: class Foo { has Numeric $.a; method foo() { $.a = $.a + 1 } }; Foo.new.foo
camelia rakudo-moar 5c5ef1: OUTPUT«Use of uninitialized value of type Numeric in numeric context in method foo at <tmp> line 1␤Cannot modify an immutable Numeric␤ in method foo at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
jnthn m: say '$x.^lookup("foo")'.chars
camelia rakudo-moar 5c5ef1: OUTPUT«17␤»
jnthn Only just and I was generous with whitespace :)
sena_kun It seems current renderers treat `output` as plain paragraph. 20:30
pochi hoelzro: seems the warning about uninitialized variables can't handle attributes, because the string example doesn't work if it's a string attribute 20:32
hoelzro pochi: ahhh...what's your exact usecase?
AlexDaniel jnthn: okay, that might work
jnthn: thanks
pochi hoelzro: see a couple of lines above, I asked camelia :-)
hoelzro ahhh, that makes sense 20:35
hmm
I really want to make that work now.
tx0h zostay: no, it's just a path. that means that every time you request a page the file must get opened, read and closed. 20:45
zostay ah, i see, i misread the code... skipped the template method in Bailador::App 20:52
app.renderer.render('text', ...) ought to do what you want
tx0h oh, i didn't saw that? let me try. ty. 20:54
zostay np, i haven't used bailador in a while myself, i just read the code on github, but something like what i wrote is possible 20:56
tx0h nods 20:59
tx0h it works, 2nd argument (3rd) must be an array with the vars for the template 21:00
jdv79 was there any word on using private class attrs from a role? 21:05
jdv79 at present, and from memory, it seems there is no solution 21:09
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2016/08/01/...of-an-era/ 21:11
dalek c: 5bb9af9 | Altai-man++ | doc/Type/ (2 files):
AttributeContainer and Proc/Async now compile
 21:12
travis-ci Doc build passed. Altai-man 'AttributeContainer and Proc/Async now compile' 21:34
travis-ci.org/perl6/doc/builds/149013626 github.com/perl6/doc/compare/71cae...b9af92f822
jdv79 m: my @a = 1,2,3;my $h = {k => {k2 => @a}}; for $h.values>><k2>.list {.say} 21:46
camelia rakudo-moar 67b654: OUTPUT«[1 2 3]␤»
jdv79 shouldn't that list flatten out there?
gfldex sena_kun: i tried to implement =begin output and found that the parser eats indentation and newlines 22:11
sena_kun gfldex, you investigated Pod::To::HTML? 22:16
gfldex sena_kun: i'm using Pod::To::BigPage, easier to change because it's basicly just a flat list of multies 22:17
gfldex sena_kun: gist.github.com/gfldex/5f99bea644d...42e0d5c028 22:20
sena_kun gfldex, I see. Maybe I should to magically know it, but still. It looks like every Pod::To-like module has it's own way to parse Pod content, am I wrong? Or parsing already built-in into compiler as a part of language? 22:25
sena_kun Ah, I'm wrong, actually. 22:26
gfldex sena_kun: rakudo does all the parsing of the actual .pod6 file. It outputs the data structure in the gist. There are a few shortcomings that different renderers deal with in different ways. 22:27
sena_kun gfldex, ah, now I understand it. Thanks. Let's see... 22:28
gfldex sena_kun: how nodes are rendered depends on their parent nodes, at times generations deep. util/extract-examples.p6 is a simple example how to deal with that. 22:29
gfldex i tried to build a round-tripping Pod::To::Perl6 for easy testing both of a renderer and the pod6 part of the grammer and got stuck. 22:30
sena_kun gfldex, I'm looking at Grammar.nqp now and it seems we have to extend it a little bit if we want it to treat `=begin output` in a special way. 22:32
Since internal parser treats `=begin something` as just plain paragraph, where all spaces and newlines are eaten. 22:33
gfldex sena_kun: if you want to dig into it, Pod::Block could do with a parent attribute and if filename/linenumber could be preserved it would make it much easier for renderers to complain about dupe index entries and other problems. 22:34
sena_kun gfldex, I'm not sure my attempts can give some result, but I'll try to investigate it further. Also, I thought about the examples compilation: we can fix up at least two common-case problems in automatical way using grep. I'll prepare PR a bit leter, I suppose. 22:39
*later
It'll take me some time to do a proper review of the diff. 22:40
sena_kun s@grep@grep/sed@ 22:43
lambd0x Hi everyone! 23:28
jdv79 europe might be sleeping
plus its monday 23:29
sena_kun lambd0x, 0/
lambd0x sena_kun: \o 23:29
lambd0x People, how can I declare a continuous extending bidimensional array in perl 6. 23:31
dalek c: 057226c | dmaestro++ | doc/Type/Int.pod6:
Clarify description of lsb, msb

Description of msb was slightly misleading, as the indexing referred to is always from the right. Clarified both lsb and msb.
c: 8405f0b | (Wenzel P. P. Peppmeyer)++ | doc/Type/Int.pod6:
Merge pull request #779 from dmaestro/clarify_Int_msb_description

Clarify description of lsb, msb
c: d0dfc5a | Altai-man++ | html/js/main.js:
The TOC animation was removed
 23:32
c: ce02d33 | (Wenzel P. P. Peppmeyer)++ | html/js/main.js:
Merge pull request #783 from perl6/animation-has-gone

The TOC animation was removed
lambd0x Is there a way to do it? E.g. my @array; is a single-dimensional array that's continuous extending..
gfldex m: my @a[*;*] 23:34
camelia rakudo-moar e411e5: OUTPUT«Jagged array shapes not yet implemented. Sorry. ␤ in block <unit> at <tmp> line 1␤␤»
gfldex lambd0x: ^^^ 23:35
m: my @a; @a[1;1] = 10; dd @a; 23:36
camelia rakudo-moar e411e5: OUTPUT«Array @a = [Any, [Any, 10]]␤»
gfldex that means you can use multi dim arrays but there is no constraint on the dimensions. If you need that you will have to implement that by hand. 23:37
m: my @a; @a[1;1] = 10; dd @a.pairs;
camelia rakudo-moar e411e5: OUTPUT«(0 => Any, 1 => $[Any, 10]).Seq␤»
gfldex also, .pairs and a few other methods don't do what they should. 23:38
m: my @a; @a[1;1] = 10; dd @a.elems;
camelia rakudo-moar e411e5: OUTPUT«2␤»
lambd0x gfldex: so this is really a feature that still needs to be implemented.. wow 23:44
Well, at least I can use a statically defined multidimensional array at will. That might be sufficient for the time being. 23:45
gfldex: thanks for answering 23:46

Powered by Cro and the Raku® Programming Language.

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

Thank you!