»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.
00:04 n1lp7r joined 00:05 kyclark joined 00:12 n1lp7r left 00:20 RabidGravy left
timotimo you know how there's now no longer a list of subs you have to remember "which work on $_ if no argument given and which don't" 00:20
we usually say "well, now you can .say"
but now there's the list of methods of Any that might be interesting to you that you'd have to remember
so you can decide whether you can .subname or have to .&subname
00:25 n1lp7r joined 00:26 pierre_ joined 00:29 jcallen_ joined, Jonis_ left, Jonis joined 00:30 jcallen left 00:31 espadrine left 00:36 mohae_ left, n1lp7r left
SmokeMachine____ hi! 00:37
this should work?
m: use experimental :macros; sub a(:$b!){}; macro m($b){quasi{ a(:{{{$b}}}) }}; m("bla")
camelia rakudo-moar f2df2c: OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in sub a at <tmp> line 1␤ in any at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
SmokeMachine____ m: use experimental :macros; sub a(:$b!){}; macro m($b){quasi{ a(:b({{{$b}}})) }}; m("bla")
camelia rakudo-moar f2df2c: OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in sub a at <tmp> line 1␤ in any at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
00:38 mohae joined
SmokeMachine____ should one of those work? 00:38
should any of those work? (sorry) 00:39
00:39 r_i_d left, r_i_d joined
kyclark lpaste.net/181419 00:41
I’m working on a script to find overlapping genes. I’m not getting expected output. Perl is reporting overlap when there is none. 00:42
Sample input data given at the top.
Comments/suggestions welcome! 00:43
Maybe it’s too late and I should try again tomorrow. 00:46
AlexDaniel SmokeMachine____: well, macros do not work at the moment 00:53
SmokeMachine____: so yes, it should work in the future… hopefully :)
SmokeMachine____ AlexDaniel: does not work at all? 00:58
AlexDaniel kyclark: and the expected output is?
SmokeMachine____: yeah, I think so
SmokeMachine____ 😞
AlexDaniel kyclark: ah, I see 00:59
00:59 BenGoldberg joined, kaare_ joined
kyclark Expected is that the last two overlaps should not appear 01:03
r_i_d ls 01:04
In writing tests, how do you "use" the file you want to test? 01:05
AlexDaniel kyclark: a, pft 01:06
kyclark: they actually overlap
kyclark: this part: %gene1<pos>[0,*-1].join('..')
kyclark: assumes that the list is ordered
but it is not
01:06 adu left
AlexDaniel kyclark: change it to this and you'll see what is going on: %gene1<pos>.sort[0,*-1].join('..') 01:07
01:08 cdg joined
r_i_d I have two files, a program.pm6 and a program-test.pm6. How do I make the test aware of the program file? 01:10
AlexDaniel r_i_d: not sure, but maybe you want to set PERL6LIB env variable to . or use ｢use lib '.';｣ in one of the files 01:11
kyclark: the problem itself, however, is slightly different 01:14
m: .say for ‘7235’..‘9016’
camelia rakudo-moar f2df2c: OUTPUT«7235␤7236␤7225␤7226␤7215␤7216␤7135␤7136␤7125␤7126␤7115␤7116␤7035␤7036␤7025␤7026␤7015␤7016␤8235␤8236␤8225␤8226␤8215␤8216␤8135␤8136␤8125␤8126␤8115␤8116␤8035␤8036␤8025␤8026␤8015␤8016␤9235…»
r_i_d i just tried use lib '.'; and it failed with "Undeclared name:".
AlexDaniel kyclark: so it looks like you have to do this: pos => [+%data<start> .. +%data<end>], 01:15
kyclark: … honestly ranges with numeric Strs hardly make any sense… and that's what you have encountered
r_i_d: what's the full error? Perhaps something is indeed undeclared? :) 01:16
r_i_d the program declares an object -- the test works fine inside the program, but fails when in it's own .t file. 01:18
the failure is that the object is undeclared, so I'm not linking it right. 01:19
AlexDaniel r_i_d: well, the dot means that it should be in the current directory. Maybe that's not the case? If so, provide another path? 01:20
r_i_d It's just the two files in the same directory file.pm6 and test-file. Do I have to declare it in the test file like I would in C? 01:23
01:26 cdg left 01:28 labster left 01:29 _slade_ left
r_i_d I don't have panda (I'm on a raspberry pi) so I cant build a module with a DATA6.json file. I'm just trying to test an object to learn how testing works in perl6. I can't imagine it being complicated. 01:29
01:31 user9 left
grondilu r_i_d: I have a raspberry pi and I have pand installed. 01:31
*panda
(I'm not using it much lately, though)
(ever since I've realized even a VM is much faster than a Pi) 01:32
r_i_d i just want to 1: declare a class in a .pm6 file and 2: declare an object of that class in a .t file. Nothing fancy 01:33
grondilu it's very simple
gfldex r_i_d: did you export the class?
grondilu $ echo -e "unit class Foo;" > Foo.pm6; echo "use Foo;" > test.t; perl6 -I. test.t 01:34
(also use Test; of course) 01:35
r_i_d oh. 01:36
grondilu did that help?
01:37 gfldex left
grondilu (the suspense is killing me) 01:37
r_i_d oops 01:40
I succeeded in overwriting my document with "unit class Foo;" 01:41
but I think I'm on the right track.
my file is gone, but perl doesn't complain about an undeclared object. 01:43
01:48 MasterDuke joined 01:49 bob777 joined
grondilu you may have a ~ backup 01:52
kyclark AlexDaniel, thanks! This fixes it:
pos => [+%data<start> .. +%data<end>],
You were right on. 01:53
r_i_d I don't need it, it was just a dummy object.
AlexDaniel kyclark: docs.perl6.org/language/traps#Stri.../Sequences 01:55
kycl
kyclark Thanks very much. It makes sense now. 01:56
AlexDaniel kyclark: basically I'd just wipe stringy ranges from the language… this is such a fail… 01:57
one of the greatest ideas behind perl 6 is different operators for numeric stuff
and yet we managed to step into the same turd with .. op 01:58
kyclark I would agree with that. It’s something I dig about Haskell. I know the operator has my back, will throw up on bad types.
r_i_d :grondilu Thank YOU! This works perfectly.
tbrowder ugexe: changes to zef look good! thanks! 02:11
ugexe: i should say they work great! 02:13
dalek ateverable: dc40d25 | MasterDuke17++ | Whateverable.pm6:
Be less restrictive about mime types

Allow any mime type that contains 'text/plain' or 'perl' in it. Also, attempt to decode the response based on the given charset, so ISO-8859-1 should be supported in addition to UTF-8. Closes #17.		 02:14
02:28 pierre_ left 02:31 r_i_d left
grondilu took the liberty to create an issue about the enumeration constants discussed earlier: 02:32
github.com/perl6/roast/issues/147
02:34 wamba joined 02:35 pierre_ joined 02:39 pierre_ left 02:42 noganex joined 02:46 noganex_ left, zacts joined
pmichaud ...can we or should we figure out a way to separate rakudobug tickets from language change request tickets? 02:53
yoleaux 22 Aug 2016 13:59Z <[Coke]> pmichaud: if I can have access to edit web pages on rakudo.org
02:53 pierre_ joined
pmichaud for example: RT #129131 is a language RFC, not a rakudo bug. 02:53
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=129131
pmichaud yoleaux: tell [Coke] You're currently listed as an editor on rakudo.org... shall I increase that to "administrator"? 02:55
.tell [Coke] You're currently listed as an editor on rakudo.org... shall I increase that to "administrator"?
yoleaux pmichaud: I'll pass your message to [Coke].
BenGoldberg m: my $foo = '7236'; say ++$foo; 03:04
camelia rakudo-moar f2df2c: OUTPUT«7237␤»
BenGoldberg m: my $foo = '7236'; say ++$foo.WHAT;
camelia rakudo-moar f2df2c: OUTPUT«Cannot resolve caller prefix:<++>(Str); none of these signatures match:␤ (Mu:D $a is rw)␤ (Mu:U $a is rw)␤ (Int:D $a is rw)␤ (int $a is rw)␤ (Bool $a is rw)␤ (Num:D $a is rw)␤ (Num:U $a is rw)␤ (num $a is rw)␤ in …»
BenGoldberg m: my $foo = '7236'; (++$foo).WHAT.say;
camelia rakudo-moar f2df2c: OUTPUT«(Str)␤»
pmichaud m: .say for 'file08.jpg'..'file15.jpg' 03:05
camelia rakudo-moar f2df2c: OUTPUT«file08.jpg␤file07.jpg␤file06.jpg␤file05.jpg␤file18.jpg␤file17.jpg␤file16.jpg␤file15.jpg␤»
BenGoldberg m: ‘a00.b’..‘a16.b’
camelia rakudo-moar f2df2c: OUTPUT«WARNINGS for <tmp>:␤Useless use of ".." in expression "‘a00.b’..‘a16.b’" in sink context (line 1)␤»
BenGoldberg m: .say for ‘a00.b’..‘a16.b’ 03:06
camelia rakudo-moar f2df2c: OUTPUT«a00.b␤a01.b␤a02.b␤a03.b␤a04.b␤a05.b␤a06.b␤a10.b␤a11.b␤a12.b␤a13.b␤a14.b␤a15.b␤a16.b␤»
BenGoldberg m: .say for ‘a00.b’..‘a10.b’
camelia rakudo-moar f2df2c: OUTPUT«a00.b␤a10.b␤»
BenGoldberg m: ‘a00.b’.SUCC.say; 03:07
camelia rakudo-moar f2df2c: OUTPUT«Method 'SUCC' not found for invocant of class 'Str'␤ in block <unit> at <tmp> line 1␤␤»
BenGoldberg m: ‘a00.b’.succ.say;
camelia rakudo-moar f2df2c: OUTPUT«a01.b␤»
pmichaud that last one looks like a rakudobug to me.
the 'a00.b'..'a10.b' one.
m: .say for 'a00.bb'..'a10.bb' 03:08
camelia rakudo-moar f2df2c: OUTPUT«a00.bb␤a10.bb␤»
BenGoldberg Pretty strange :)
pmichaud m: .say for 'a00.txt'..'a10.txt'
camelia rakudo-moar f2df2c: OUTPUT«a00.txt␤a10.txt␤»
pmichaud m: .say for 'a00.txt'..'a11.txt'
camelia rakudo-moar f2df2c: OUTPUT«a00.txt␤a01.txt␤a10.txt␤a11.txt␤»
pmichaud m: .say for 'a01.txt'..'a11.txt'
camelia rakudo-moar f2df2c: OUTPUT«a01.txt␤a11.txt␤»
pmichaud m: .say for 'a01.txt'..'a12.txt'
camelia rakudo-moar f2df2c: OUTPUT«a01.txt␤a02.txt␤a11.txt␤a12.txt␤»
BenGoldberg m: .say for '0b000' .. '0b111'
camelia rakudo-moar f2df2c: OUTPUT«0b000␤0b001␤0b010␤0b011␤0b100␤0b101␤0b110␤0b111␤»
BenGoldberg m: .EVAL.say for '0b000' .. '0b111' 03:09
camelia rakudo-moar f2df2c: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤»
BenGoldberg m: .EVAL.say for '0o00' .. '0o77' 03:10
camelia rakudo-moar f2df2c: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤10␤11␤12␤13␤14␤15␤16␤17␤18␤19␤20␤21␤22␤23␤24␤25␤26␤27␤28␤29␤30␤31␤32␤33␤34␤35␤36␤37␤38␤39␤40␤41␤42␤43␤44␤45␤46␤47␤48␤49␤50␤51␤52␤5…»
BenGoldberg m: .EVAL.say for '0x00' .. '0xff'
camelia rakudo-moar f2df2c: OUTPUT«0␤1␤2␤3␤4␤5␤6␤7␤8␤9␤5===SORRY!5=== Error while compiling /home/camelia/EVAL_10␤Confused␤at /home/camelia/EVAL_10:1␤------> 030x0:7⏏5<EOL>␤ expecting any of:␤ colon pair␤»
BenGoldberg m: .say for '0x00' .. '0xff' 03:11
camelia rakudo-moar f2df2c: OUTPUT«0x00␤0x01␤0x02␤0x03␤0x04␤0x05␤0x06␤0x07␤0x08␤0x09␤0x0:␤0x0;␤0x0<␤0x0=␤0x0>␤0x0?␤0x0@␤0x0A␤0x0B␤0x0C␤0x0D␤0x0E␤0x0F␤0x0G␤0x0H␤0x0I␤0x0J␤0x0K␤0x0L␤0x0M␤0x0N␤0x0O␤0x0P␤0x0Q␤0x0R␤0x0S␤0x0T…»
03:15 zacts left
pmichaud that also looks like a bug to me. 03:16
timotimo is it?
pmichaud m: say '0x00'.succ
camelia rakudo-moar f2df2c: OUTPUT«0x01␤»
pmichaud m: say '0x09'.succ
camelia rakudo-moar f2df2c: OUTPUT«0x10␤»
timotimo don't we only do the magic succ-ing when the ends are in a magic sequence?
pmichaud possibly. 03:17
timotimo like 0 to 9 or a to z or something
pmichaud According to S05 (which I accept may be out of date), ranges for non-numeric types are formed by repeatedly doing .succ on the values 03:18
03:19 brrt` joined
pmichaud so, if that's the case, it seems to me that after '0x09' I should get '0x10' in the range sequence. 03:19
not '0x0:'
timotimo right. well, i *think* the current behaviour for things that are not recognized to be of the same sequence is to go through unicode codepoints until you wrap 03:21
pmichaud okay, so that's a change from S05.
03:22 brrt left 03:24 kyclark left
BenGoldberg According to docs.perl6.org/language/traps#Stri.../Sequences rakudo is actually doing what it's supposed to. 03:25
Which is to say, it's not supposed to use .succ on the whole string, supposedly since if that was what you'd wanted, you could easily write: $first, *.succ .. $last 03:26
03:26 Xliff joined
timotimo you mean ..., yes? 03:27
BenGoldberg Yes.
03:27 Actualeyes left 03:40 wamba left 03:41 Actualeyes joined
shantanu Is there a way to implement observable streams in Perl 6 like RxJs or RxPy? I love promises but it looks like observables could provide a little more useful functionality for event based programming. 03:49
AlexDaniel pmichaud: we don't have any other place to submit language change tickets, so it seems like [RFC] tickets are exactly for that purpose now
timotimo shantanu: Supply is exactly what observable streams are 03:50
and they are already implemented for you
shantanu ohh thanks for that!! :) 03:51
timotimo they are really cool, and we have a bunch of combinators, too. like you'd expect from the other implementations of Rx*
a big part of the concurrency primitives are modelled with a big influence from C#
shantanu C#? I am not fluent in that. 03:52
timotimo me neither
but our master of technical design is fluent in it, among many other things
shantanu It does look really cool!
timotimo the "supply"/"react" and "whenever" syntax is *really* nice 03:53
Xliff Why am I getting this error? "You cannot create an instance of this type (Color)" 03:56
timotimo check what it's HOW is 03:57
it's potentially a PackageHOW
Xliff Yeah. PackageHOW. 03:58
shantanu HOW?
timotimo yeah, HOW
a bit different from WHAT
shantanu ohh ok
timotimo m: say Int.WHAT;
camelia rakudo-moar f2df2c: OUTPUT«(Int)␤»
timotimo m: say Int.HOW
camelia rakudo-moar f2df2c: OUTPUT«Perl6::Metamodel::ClassHOW.new␤»
timotimo m: subset Foo of Int where 10 < * < 99; say Foo.HOW 03:59
camelia rakudo-moar f2df2c: OUTPUT«Perl6::Metamodel::SubsetHOW.new␤»
timotimo package test { }; say test.HOW
m: package test { }; say test.HOW
camelia rakudo-moar f2df2c: OUTPUT«Perl6::Metamodel::PackageHOW.new␤»
shantanu m: use Supply; say Supply.HOW; 04:00
camelia rakudo-moar f2df2c: OUTPUT«===SORRY!===␤Supply is a builtin type, not an external module␤»
shantanu m: say Supply.HOW;
camelia rakudo-moar f2df2c: OUTPUT«Perl6::Metamodel::ClassHOW.new␤»
shantanu m: say Supply.WHAT;
camelia rakudo-moar f2df2c: OUTPUT«(Supply)␤»
Xliff So how do I change a PackageHOW to a ClassHOW? 04:01
timotimo switching something's HOW from one thing to another is not possible
when you have a PackageHOW, you have a package of things
shantanu my $a = Supplier.new; say $a.HOW;
m: my $a = Supplier.new; say $a.HOW;
camelia rakudo-moar f2df2c: OUTPUT«Perl6::Metamodel::ClassHOW.new␤»
timotimo i expect the ClassHOW-ed thing you want is inside that package somewhere
shantanu m: my $a = Supplier.new; say $a.WHAT; 04:02
camelia rakudo-moar f2df2c: OUTPUT«(Supplier)␤»
Xliff perl6 -e 'require ::("Color"); say ::("Color").HOW.^name' --> Perl6::Metamodel::ClassHOW
timotimo and you can't .new that? 04:03
Xliff However when I do that in a script, it becomes "Perl6::Metamodel::PackageHOW"
I can new the former, but not the latter.
timotimo well, something else is off, then
Xliff And I don't know why there's a difference.
timotimo is it the difference between grey and gray again :P 04:04
do you perhaps have -I in one and not the other?
04:07 skids left
Xliff Yes 04:09
No change using -I
And no, I think this is more subtle than "gr(e||a)y" 04:10
Or at the very least more annoying.
timotimo i can imagine
sorry about it :(
Xliff :/
I'm wondering if it's something to do with the "unit module" or "unit package" behavior. 04:12
04:12 MasterDuke left, khw left 04:13 skids joined, harmil left
Xliff Because things I am doing in "perl6 -I..." shouldn't be so significantly different when those same things are done in a file based compunit. 04:14
(I almost said "script", but this isn't a script)
And that should really read "perl6 -e..."
04:21 kyclark joined
grondilu trying to make Perl 6 shine on HN: news.ycombinator.com/item?id=12387492 04:21
04:24 kyclark left
Xliff timotimo: Golfed. 04:24
I will RT
Actually, before I do, I will gist and let ppl check it out.
04:30 Actualeyes1 joined, Actualeyes left 04:31 skids left
Xliff gist.github.com/Xliff/6fa8e4e6283e...0c708948aa 04:31
And yes... at this point I wll except ¯\_(ツ)_/¯, but I just wanted to get that piece of weirdness out there. 04:33
Wow! All it took was a ::! 04:35
perl6 -Ilib -e 'unit package Color::Test; INIT { require ::("Color"); }; sub test { ::("Color").HOW.^name.say }; test'
Perl6::Metamodel::PackageHOW
04:43 pierre_ left 04:45 pierre_ joined 04:49 pierre_ left 04:50 notostraca joined 04:52 TEttinger left, notostraca is now known as TEttinger 04:53 MilkmanDan left 04:57 labster joined 05:11 MilkmanDan joined 05:28 MilkmanDan left 05:54 CIAvash joined 05:55 pierre_ joined, brrt joined 06:01 pierre_ left 06:06 pierre_ joined 06:10 Xliff left 06:11 Xliff joined 06:22 brrt left 06:25 darutoko joined, firstdayonthejob joined
domm what's the correct syntax to import Perl5 functions using Inline::Perl5? 06:26
I need 'use Data::Random qw(:all)' in a p6 script
hm, $p5.call('Data::Random::rand_words'); 06:31
06:31 margeas joined
domm works, but is not too nice 06:31
ok, back onto the bike...
moritz ... to bikeshed :-) 06:32
domm :-)
06:33 BenGoldberg left
Xliff domm: 'use Inline::Perl5; my $varable; my $p5 = Inline::Perl5.new; p5.run('use Data::Random qw(:all); $variable = rand_words...'); 06:35
With correct quoting this time...
domm: "use Inline::Perl5; my $varable; my $p5 = Inline::Perl5.new; p5.run('use Data::Random qw(:all); $variable = rand_words...');"
nine domm: use Data::Random:from<Perl5> <all>; 06:36
Xliff Or that. nine++ (and more succinct)
nine: If you have time, can you look at this gist and tell me if it's a bug?
gist.github.com/Xliff/6fa8e4e6283e...0c708948aa
nine Maybe it _is_ a good idea to give my talk again :)
domm: use Data::Random:from<Perl5> <:all>; # correct version with the colon before all 06:38
06:46 firstdayonthejob left
nine Xliff: easier way to reproduce: unit package Color::Test; use Color; ::("Color").HOW.^name.say 06:49
Xliff Huh
Any ideas as to solution? 06:50
I'm trying to write a module where the use of the Color class is optional.
06:52 bbkr left
nine Actually, this could just be a plain merge_globals bug 06:53
06:56 brrt` left
geekosaur isn't this the thing where it creates packages for intermediate levels (so Color::Test implicitly creates package Color containing package Test)? 06:56
(in terms of namespacing, at least) 06:57
nine yes
Xliff O_o
I think I actually grokked that! 06:58
\o/
nine But it should replace the package with the class. Otherwise Color.new doesn't even work
Xliff Exactly my problem.
geekosaur right; possibly these should be different namespaces somehow. I think there may be some RTs about related issues already
(noodling) so maybe you get a Color that has a PackageHOW only if the name is not already in use, otherwise it is accessible via a PACKAGE:: namespace. in a case like this where it' 07:00
s created first, the compiler should see that and arrange to not make the package one directly accessible
...or maybe packages just shouldn't be in the default namespace as such but the PACKAGE:: namespace is checked by things that care, and if you want to access it directly you must specify 07:01
Xliff Would that be $?PACKAGE... or something else? 07:04
nine The odd thing is that merge_globals should already DTRT. When it detects that the target is a stub (PackageHOW) and the source is a e.g. a class, it takes all symbols from the package, adds them to the class and replaces the package with the class. 07:05
geekosaur I'm thinking of "pseudo-packages" docs.perl6.org/language/packages#P...o-packages
nine github.com/rakudo/rakudo/blob/nom/...r.nqp#L132 07:06
geekosaur see, I always considered that a hack... mostly because there kept being weird bugs
like, oh, this one
nine But the weird bug in this case is just that it doesn't seem to do this at all.
If the plan was followed through, it would work just fine. 07:07
07:09 pierre_ left 07:11 nadim_ joined, pierre_ joined 07:18 leont left 07:22 zakharyas joined 07:40 aindilis left, jonas2 joined 07:47 domidumont joined 07:51 domidumont left 07:52 smls joined, domidumont joined 07:55 g4 joined
Xliff Here's an odd one: "no EXPORT sub, but you provided positional argument in the 'use' statement" but there is an export sub in the module being imported. O_o 08:01
08:01 leont joined
Xliff github.com/Xliff/p6-color-names/bl...mes.pm#L38 <- Proof 08:02
Generates error: perl6 -Ilib -e 'use Color::Names <:b>; dd color("gray42")'
smls Xliff: That's the same problem I had yesterday. Moving the `sub EXPORT` above the `unit package Color::Names;` should make it work.
no idea *why* that's required though 08:03
Xliff smls: That's going to be hell on scoped sub names, though. 08:04
smls ¯\_(ツ)_/¯
Xliff I mean, if I am going to export routines outside of the EXPORT scope, how would I do it? 08:05
::() doesn't seem to work.
08:06 dakkar joined
Xliff Returning this from sub EXPORT: 08:06
# cw: Really want a SELECTIVE way to load these, instead of doing
# them all at compile time.
Gar...
Returning this: 08:07
{
'&color' => ::("&Color::Names::color"),
'&hex' => ::("&Color::Names::hex"),
'&rgb' => ::("&Color::Names::rgb")
}
And trying the script above, I get: "Too many positionals passed; expected 0 arguments but got 1"
So do I need to export signatures, too?
moritz you don't need to export signatures 08:08
Xliff wishes you could grep repositories on GitHub 08:09
moritz Xliff: are those our-subs?
Xliff Um. No.
moritz then the namespaced lookup won't work
Xliff Seems like they work because I'm not getting a "not found" error. I am getting a "I can't find a signature match" error. 08:10
However I will try.
moritz and any reason you're not just using export tags?
Xliff Same ish.
Can you use EXPORT tags with a custom sub EXPORT? 08:11
I thought that rendered the tags useless.
Again... will try.
Oh. LOL. Already had 'em in.
So export tags AREN'T working.
Again, on suggestion from smls++, sub EXPORT is moved ouside of unit package scope. 08:12
08:12 MilkmanDan joined
smls Xliff: It's just what cygx++ suggested to me... :) 08:13
Xliff Again I need a custom sub EXPORT.
So based on current rakudo behavior, export tags are superfluous.
At least with current implementation, which I am stubborly still attached to.
Give me a few more days with this roadblock. 08:14
凸ಠ益ಠ)凸
08:15 vytas joined
Xliff Emoticon was for roadblock... not for #channel. 08:15
08:18 cgfbee left
TimToady sitting in Munich waiting for MUC --> DEV --> SJC 08:25
08:25 cgfbee joined
moritz waves from just 200km away 08:32
nine waves from 183km away 08:34
lizmat waves from 576 km away :-) 08:39
08:41 pierre_ left
leont waves from a good 100km further than lizmat :-p 08:43
08:50 pierre_ joined 08:55 pierre_ left 08:57 pierre_ joined 09:07 leont left
TimToady boarding & 09:08
09:09 diego_k joined 09:10 diegok left
Xliff Now getting this in parse stage: "Cannot find method 'merge-symbols': no method cache and no .^find_method" 09:11
09:11 lizmat left
El_Che I had a look yesterday at the packaging options (as a consequence about what rakudo a user must install). Pretty hellish 09:11
Xliff Error goes away when I remove sub EXPORT; 09:12
Just for the halibut, I will rakudobrew.
09:13 go|dfish left 09:15 RabidGravy joined
Ulti anyone reported the broken rakduobrew builds? 09:17
nine El_Che: what do you mean? 09:20
09:22 mls left
El_Che nine: I was looking for a fast way to create debs/rpms/what-have-you for rakudo. 09:22
nine: so users can install an up to date pkg instead of rakudobrew 09:23
moritz El_Che: if you provide some debian/ dirs for building the .deb packages, I'm willing to build a pipeline/process that automtaically runs the builds and uploads them into a repo 09:27
El_Che moritz: I was looking at that specifically. What do we want exactly. Something less integrated in the OS, but quickly built: github.com/jordansissel/fpm 09:29
Xliff Ulti: Just rebuild rakudo with no problems.
El_Che "Proper" packages will be supplied by the OS. I was thinking in putting everything in /opt/rakudo or something like that
nine El_Che: creating rpm packages is actually quite simple: build.opensuse.org/package/view_fi...c?expand=1 09:31
ambs does this ring any bell? --- paste.perldancer.org/2pO6zrOjkNpPz
El_Che moritz: maybe an anti-pattern, but docker killed my packaging efforts for $work (I used to produce rpms for our apps)
nine build.opensuse.org/package/view_fi...c?expand=1
build.opensuse.org/package/view_fi...c?expand=0
El_Che: with a bit of luck those even build on other distributions or may be easily adaptable.
El_Che nine: you're extraordinary. I got an headache looking at the opensuse build system docs (and the zillion dead links)
nine ambs: I suspect that you are running a very old perl6 version? 09:32
moritz El_Che: I'm fine with fpm
melezhik Hi all!
ambs nine: ok, it might be
nine El_Che: yes, took me quite a while to figure it out. But once you got through that the build service is just awesome
melezhik Anybody suggest a simple logger for perl6?
moritz note(), say(), github.com/moznion/p6-Log-Minimal 09:33
nine loves note
melezhik mortiz: thanks, Log::Minimal does not allow to write into file? 09:35
El_Che moritz, nine: there is also this: github.com/alanfranz/fpm-within-docker . But I don't think running someone elses containers --never mind how easy they make the process-- to produce packages is a good idea
smls Did Rakudo's start-up time improve recently? 09:37
`perl6 -e ''` takes 0.14s now on my PC, I thought it had been 0.18s before but I could be mistaken. 09:38
ambs nine: exactly that. thanks 09:39
smls Or could switching from rakudobrew to building from git perl nine's instructions, have affected that?
*per
nine smls: I can't imagine rakudobrew having an influence on that. I claim it's innocent for once :) 09:40
moritz iirc rakudobrew adds a layer of shell scripts around executables, but that shouldn't make 0.04s difference 09:42
Xliff LOL, nine++ # /me hugs his rakudobrew
09:42 bjz joined
Xliff Is there a mechanism for supplying dummy export tags? 09:43
09:43 pmurias joined
Xliff I can get "sub EXPORT(*%a)" to work, but not "sub EXPORT (*@a)" without a parse error. However "sub EXPORT (*%a)" will bail at parse without a supplied export tag matching argument to the use statement. 09:44
09:46 mls joined
pmurias are there any bad consequences of me publishing a rough 0.0.1 nqp-js version on npm? (it's not really meant for direct use just for easier building of rakudo.js) 09:46
Xliff tries to imagine shoe-horning rakudo.js into a browser. 09:47
pmurias: Any ideas as to the size of rakudo.js?
pmurias Xliff: I don't have a currently build one as I was mostly working on nqp-js recent, and I haven't tried minifing it 09:52
Xliff: rakudo.js will need to be included on REPLs 09:53
Xliff Aaieee... minifyinh. 09:54
...minifying, even.
Ok, I can just feel RTs future pain from the minifying bugs. 09:55
(depending on what you use to minify)
It's been a while since I've had to worry about minifying Javascript.
Fortunately, I have never had to write JS where it was necessary to Hide What Was Done for $work. 09:56
09:56 Actualeyes1 left
pmurias nqp-js is currently 9.5M unminified with currently no effort spend in reducing the size of the emitted code 09:56
nqp-js-running-on-top-of-js 09:57
09:58 go|dfish joined
pmurias Xliff: re minification bugs, I think the current minifiers have different settings where you can set how much potential breakage you want 09:58
10:00 bjz left
Xliff Ah! That's nice. 10:00
I've never bought the minifying-for-size argument, but then again, I've not written a suite of JS over 250k...ever. 10:01
OOB - HAH! Oh Charlie Stross. You usually get the future right, but the references about MIPS is just.... cute. 10:02
10:07 pierre_ left
konobi oh wow... just looked at arrayref declarations of attributes or modifiers in Moose for the first time in a long time 10:10
10:10 cgfbee left
konobi i remember adding it for practicality reasons (and somewhat for compositionality) but now that I look back on it, I can't decide if it's the best representation 10:11
10:12 Actualeyes joined 10:13 pierre_ joined
timotimo anyone want to write a .ASSIGN for CArray? :) 10:16
tadzik oh, Moose. I actually haven't heard that name in a long time :o
it's Moo, Moo, Moo for everything I do
10:18 cgfbee joined
Xliff timotimo: What would be involved for that? 10:18
timotimo Xliff: take a list and fill the CArray with it; look at the native array roles' implementation of ASSIGN for inspiration 10:19
10:19 bjz joined
Xliff Is that MoarVM or Rakudo? 10:20
pmurias konobi: if I want to release a rough version of nqp-js on npm (which will be mostly intendend for building rakudo.js) how should I mark that people should beware before using it
timotimo that's rakudo
Xliff timotimo: ./nqp/MoarVM/src/6model/reprs/MVMArray.c 10:24
?
timotimo no, in rakudo 10:26
under src/core/
well, the CArray thing is in lib/NativeCall somewhere
10:28 aindilis joined
Xliff moar-nom/lib/NativeCall/Types.pm6 10:29
However I see references to ASSIGN-POS, but no .ASSIGN anywhere in rakudo source.
timotimo oh
Xliff Lemme modify grep.
timotimo i think i meant to say STORE
Xliff Heh. 10:30
timotimo that's the method that should make "my @foo := CArray[int].new; @foo = Bool.roll(100_000)" work 10:32
rather than having to "for Bool.roll(100_000) { @foo.push($_) }" 10:33
Xliff timotimo: I'll try looking into this. I will let you know if I have to throw my hands up in defeat. 10:34
timotimo thanks for taking the time!
timotimo is AFK for a bit
Xliff np
timotimo i hope others can assist :)
Xliff Yeah. There's a whole lot of nqp here to unravel. I can understand most of Array's STORE, but mapping that into the proper assignment methods may tricky for me. 10:42
Will look more into it, tomorrow (ie Tuesday, EDT, pm) 10:43
I am up way past my bedtime... again.
(will someone please mute that massive fusion generator in the sky, please?)
<-- vampire 10:44
10:45 dLeCamarae joined
Xliff And it looks like CArray is missing some attributes in Array and List. 10:46
10:46 pierre_ left
Xliff goes *poof* 10:46
10:46 Xliff is now known as Xliff_zzzzz
mst timotimo++ # Oh. My. Gods. 10:59
timotimo mst: what did i do wrong? :( 11:01
mst < timotimo> mst: Opan GangnaMSTyle!
timotimo oh, hehe.
mst that was *AWFUL*
hence the ++
Xliff_zzzzz HAHAHAHA!
timotimo i was pretty proud of that one
nine Well spotted :)
11:02 labster left
timotimo shit. with this tweet i've doomed myself to write a Shakespeare-based regex slang :( 11:08
twitter.com/loltimo/status/770578530879012865
smls m: say (reverse 1 .. 100_000_000_000_000_000_000_000_000)[^2] 11:10
camelia rakudo-moar f2df2c: OUTPUT«(100000000000000000000000000 99999999999999999999999999)␤»
smls Nice, Range.reverse returns a Seq that iterates the range backwards efficiently
timotimo everything else would be terrible :) 11:11
smls This is why I love the GLR, jnthn++
timotimo but how do we actually implement that ... :P
nine reverse(@a) calls @a.reverse which in the case of a Range returns a reverse-iterator 11:12
timotimo oooh 11:13
dLeCamarae Guys, I have this Perl Web app. When I run it at home, a call to Web::App::Dispatch.new(SCGI.new) works; at work it fails with an argument error. Weirdest thing ever.
nine So it's actually more an artifact of using OO
moritz dLeCamarae: unless it's written in Perl 6, I recommend asking in #perl
nine moritz: the .new suggests it's indeed Perl 6 11:14
dLeCamarae It is in Perl 6, of course. What do you take me for, a slug? :-D
Only Perl 6 and Haskell for the foreseeable future. Life is too short.
moritz dLeCamarae: no offense meant, we just get a high number of 5ers asking here by mistake
11:15 ilmari_ is now known as ilmari
timotimo can you give us your perl6 --version for both at home and at work? 11:15
moritz dLeCamarae: so, have you compared rakudo versions on boht machines?
dLeCamarae (And, if God wills it, another, this-time-complete Pugs; Perl 6 in Haq! Yay!)
At work:
This is Rakudo version 2016.06-83-gb93043a built on MoarVM version 2016.06-9-g8fc21d5 implementing Perl 6.c. 11:16
This is the one that fails. It seems that, with it, Web::App::Dispatch.new sends a second argument to Web::App (which it inherits from), failing in the process, while on the home setup it does not. Let me try to log into home and get the version. 11:18
Hmm ... my Tor access to home is failing. :-(
github.com/supernovus/perl6-web/bl...pp.pm6#L11
That is the line that complains, but only in this setup. In every other it works. It says here: 11:19
Too many positionals passed; expected 1 argument but got 2
in method new at /home/revence/.rakudobrew/moar-nom/install/share/perl6/site/sources/336A8A6602DE7CDDA67F44458EC605AD94184E17 (Web::App) line 11 11:20
in method set-up at ./egrapha.pl6 line 799
And that line 799 in egrapha.pl6 is:
my $wapp = Web::App::Dispatch.new(SCGI.new(:$port));
nine dLeCamarae: you linked to Web::App.new but showed code for Web::App::Dispatch.new 11:21
dLeCamarae So, this follows the spec pretty closely, and it works every time at home. A pity I cannot get the version on that side; I doubt, though, that it would be the one that has to chang.
nine ah is Web::App; sorry for the noise 11:22
dLeCamarae nine Okay, hold on, let me get the other one.
Yes, they are the same. It inherits pretty directly, especially the new. Alas, it is weird here.
nine self.bless(*, :$engine) looks suspicious?
I think the * there should have been gone for a long time.
dLeCamarae I think it does. 11:23
Okay ... pull requests revving up. But why does it work on my other Rakudo Star?
nine Because most probably the one at home is older
dLeCamarae Hmm.
nine Don't remember exactly when but lizmat fixed that a couple months ago
11:24 TEttinger left
dLeCamarae I have been leaving behing Rakudo installations for some 10 years now, so I would not be surprised. But I always say use v6.c in all my code now; should it not catch such an inconsistency? 11:24
timotimo we've deprecated putting a * as the first argument to bless a long time ago, and completely removed it at some point, too 11:27
are you able to get the version you're using at home? 11:28
nine More or less. The spec test suite doesn't even cover all the features implemented in rakudo. There's no hope at all for it covering 100 % of the non-features, e.g. being able to pass invalid parameters. The * was such an oversight.
11:31 cyphase left 11:32 cyphase joined
pmurias npm install -g nqp-js-on-js; nqp-js-on-js -e 'say("Hello World")' works 11:34
11:36 akosgheri joined
tbrowder help: trying to execute a prog and get error about no such file in /path2rakudosharedist*hash*/ no such file; i get no error when i comment out a module i use, but the module has been uninstalled. any way to fix a corrupted compunit repository? 11:44
nine tbrowder: can you paste more details like the actual paths somewhere?
tree output of the repository in question could also be helpful 11:45
tbrowder i can do that; give me a few minutes... 11:46
masak today's fun puzzle: how can you demonstrate whether a prefix op that you declared is tighter than another prefix op that you declared? 11:47
(hi #perl6)
mst blinks
I'm not sure how precedence even works for prefix ops 11:48
I mean, like, in general
timotimo yeah, since prefixes hug the thing they prefix so tightly anyway
masak mst: that's the reaction I was expecting
timotimo i could imagine asking whether a prefix is tighter than a postfix
masak it does make sense. let me explain.
yes, what timotimo said. 11:49
consider pre- and postfixes to share a precedence space
mst ooooh, right
masak then they can compete for eval-first
...and that's also the answer to the riddle
moritz well, my approach would be create a postfix operator with the same precedence as one of the prefix operators
timotimo aye
masak moritz: correct.
or a nearby precedence, is enough
moritz and then one prefix against the artificial postfix
timotimo ++@foo[1]
masak for actual code that demonstrates this (currently a bug in 007), see github.com/masak/007/issues/189 11:50
mst right, sorry, you said 'one prefix op precedence higher than another prefix op'
that's the one I can't understand
masak mst: yes, I said that.
mst if we're allowing postfixes as well, the question makes more sense
masak mst: you test them *through* a postfix.
mst: that's the only way you can get an observable.
mst aha
timotimo right, the postfix is just the piece of litmus paper
masak right 11:51
mst yes, ok, I now follow
neat :D
jnthn m: say infix:<cmp> (&prefix:<|>, &prefix:<++>)>>.prec>>.<prec>
camelia rakudo-moar f2df2c: OUTPUT«Too few positionals passed; expected 2 arguments but got 1␤ in block <unit> at <tmp> line 1␤␤»
jnthn m: say infix:<cmp> |((&prefix:<|>, &prefix:<++>)>>.prec>>.<prec>)
camelia rakudo-moar f2df2c: OUTPUT«Less␤»
masak not all of us have your fancy introspection, sir :P
moritz introspection is cheating!
:-)
masak introspection is technique! 11:52
moritz (and cheating is technique... )
masak: hah, we're in resonance again!
masak was gonna say
it's the famous .se-.de entanglement
11:54 obfusk left, obfusk joined 11:55 smls_ joined, smls left
tbrowder nine: the repository file tree is here: "gist.github.com/tbrowder/3880cbf51...61094b6c7" 11:56
11:57 ilbot3 left, canopus left
Xliff_zzzzz Help: Is it not possible to grab the arguments to a custom "sub EXPORT()" before it goes looking for export tags. Or any way I can force dummy export tags. Since my error occurs in "stage parse" I am a little worried this is not possible. 11:58
nine tbrowder: how exactly did you uninstall a dist? 11:59
Xliff_zzzzz Currently, I can only do "sub EXPORT(*%l)". rakudo bails on "sub EXPORT(*@l)" if positional is given.
timotimo well, yeah, :foo and :bar are named arguments, of course
tbrowder nine: i was in the directory containing the module, under construction, and executed: "zef uninstall ." 12:00
Roamer` *head scratch* am I misunderstanding something about what Hash[type] $var is supposed to mean? I mean...
m: use v6.c; my Hash[Int:D] $wtf = Hash[Int:D].new(); $wtf<a> = 7.62; dd $wtf; dd $wtf<a>; my Int:D %fine; %fine<a> = 7.62;
camelia rakudo-moar f2df2c: OUTPUT«Hash[Int:D] $wtf = (my Int:D % = :a(7.62))␤Rat <element> = 7.62␤Type check failed in assignment to %fine; expected Int:D but got Rat (7.62)␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz Yeah, well I don't give a toss about export tags, really. I just want to pass custom arguments to EXPORT so I can do some magic. 12:01
nine tbrowder: please try removing /usr/local/rakudo-git.d/share/perl6/site/short/8144BFF032FBA84D9128DCD60D2A111A91A29BC8
Roamer` ...how exactly did it manage to put a Rat into a Hash[Int:D]?
12:01 gabiruh joined
tbrowder nine: WILCO 12:01
nine Roamer`: Hash[Int:D] means the keys are defined Ints
moritz nine: but that didn't validate either, because there's a Str key in there 12:02
Xliff_zzzzz m: use v6.c; my Hash[Int:D] $wtf = Hash[Int:D].new(); $wtf<2> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<2> = 7.62;
nine So...wait a second. Then you shouldn't be able to assign to <a>
camelia rakudo-moar f2df2c: OUTPUT«Hash[Int:D] $wtf = (my Int:D % = "2" => 7.62)␤Rat <element> = 7.62␤Type check failed in assignment to %fine; expected Int:D but got Rat (7.62)␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz m: use v6.c; my Hash[Int:D] $wtf = Hash[Int:D].new(); $wtf<2> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<a> = 7.62;
camelia rakudo-moar f2df2c: OUTPUT«Hash[Int:D] $wtf = (my Int:D % = "2" => 7.62)␤Rat <element> = 7.62␤Type check failed in assignment to %fine; expected Int:D but got Rat (7.62)␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz m: use v6.c; my Hash[Int:D] $wtf = Hash[Int:D].new(); $wtf<2> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<a> = 700
camelia rakudo-moar f2df2c: OUTPUT«Hash[Int:D] $wtf = (my Int:D % = "2" => 7.62)␤Rat <element> = 7.62␤»
Roamer` oooookay, so I had indeed grossly misunderstood things, but there's still some good about to come out of it? :) 12:03
Xliff_zzzzz m: use v6.c; my Hash[Int:D] $wtf = Hash[Int:D].new(); $wtf<2> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<a> = 700; dd %fine
camelia rakudo-moar f2df2c: OUTPUT«Hash[Int:D] $wtf = (my Int:D % = "2" => 7.62)␤Rat <element> = 7.62␤Hash[Int:D] %fine = (my Int:D % = :a(700))␤»
12:03 ilbot3 joined
tbrowder nine: success!! thanks!! how did you know what to remove? 12:03
Xliff_zzzzz First, requires int keys. Last requires int Values.
12:04 pierre_ joined
nine tbrowder: the directories in short/ are just the SHA-1 hashes of the short-name of modules. The files contained in such directories are named after the dist that contains such a module. You seem to have removed the dist but not the short-name lookup directory. 12:04
12:04 canopus joined
nine tbrowder: helps a lot being the one who implemented all of that :) 12:04
Roamer` Xliff_zzzzz, ahh, I get it now... almost 12:05
12:05 user9 joined
Roamer` Xliff_zzzzz, is there then a syntax that would allow me to specify a *type* (like e.g. a function return type) of a hash containing arrays of strings? 12:05
I thought Hash[Array[Str:D]] was it, but apparently not
Xliff_zzzzz Aiee!
Roamer`, That specifies key types. 12:06
So all keys must be Array[Str:D]
tbrowder nine: thanks a heap--i'm putting that in my notes. is that documented some where? sounds like a good candidate for the docs...
Roamer` Xliff_zzzzz, yep, I almost got that part, although I still don't get what moritz and nine are wondering - how does $wtf<a> work then :)
nine ugexe: see tbrowder ^^^. zef uninstall . left a short-name lookup file. I assume zef just uses CURI.uninstall and I guess the local version of the dist does no longer contain the module the short-name lookup belonged to.
moritz Roamer`: it's most likely a bug somewhere 12:07
nine tbrowder: sounds like a better candidate for a fix. There's even a comment about this in the uninstall method.
Roamer` moritz, yeah, I thought so, I'll report it. still, any ideas about a function returning a hash of arrays of strings? :)
(and, yes, of course I can encapsulate everything into objects or even subsets of classes, but still...) 12:08
Xliff_zzzzz m: use v6.c; my Hash[Int:D] $wtf; $wtf<a> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<2> = 7.62;
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in assignment to $wtf; expected Hash[Int:D] but got Hash (${})␤ in block <unit> at <tmp> line 1␤␤»
moritz Roamer`: have you tried it with a %-variable?
Xliff_zzzzz m: use v6.c; my Hash[Int:D] $wtf; $wtf<2> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<2> = 7.62;
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in assignment to $wtf; expected Hash[Int:D] but got Hash (${})␤ in block <unit> at <tmp> line 1␤␤»
Roamer` Xliff_zzzzz, you need to initialize the variable when you declare it (another thing that kind of threw me, maybe another bug?) 12:09
Xliff_zzzzz m: use v6.c; my Hash[Int:D] %wtf; %wtf<2> = 7.62; dd $wtf; dd $wtf<2>; my Int:D %fine; %fine<2> = 7.62;
camelia rakudo-moar f2df2c: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '$wtf' is not declared. Did you mean '%wtf'?␤at <tmp>:1␤------> 3my Hash[Int:D] %wtf; %wtf<2> = 7.62; dd 7⏏5$wtf; dd $wtf<2>; my Int:D %fine; %fine<␤»
Xliff_zzzzz m: use v6.c; my Hash[Int:D] %wtf; %wtf<2> = 7.62; dd %wtf;
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in assignment to %wtf; expected Hash[Int:D] but got Rat (7.62)␤ in block <unit> at <tmp> line 1␤␤»
Roamer` Xliff_zzzzz: try my Hash[Int:D] $wtf .= new();
Xliff_zzzzz m: use v6.c; my Hash[Int:D] %wtf; %wtf<a> = 7.62; dd %wtf;
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in assignment to %wtf; expected Hash[Int:D] but got Rat (7.62)␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz Roamer: I am trying without the .new for a reason.
Roamer` ah, right
Xliff_zzzzz Looks like there is type checking weirdness when attempting to attach a type check to keys. 12:10
But that's just my Wild Aassed Guess based on 2 minutes with camelia.
m: use v6.c; my Hash[Int:D, Int:D] $wtf; $wtf = Hash[Int:D, Int:D].new; $wtf<a> = 2; 12:11
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in binding to key; expected Int:D but got Str ("a")␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz m: use v6.c; my Hash[Int:D, Int:D] $wtf; $wtf = Hash[Int:D, Int:D].new; $wtf<2> = 2;
camelia ( no output )
smls_ m: my Array[Str] %hash; %hash<foo> = Array[Str].new("a"); %hash<foo> = ["b"];
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in assignment to %hash; expected Array[Str] but got Array ($["b"])␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz And there you go.
m: use v6.c; my Hash[Any, Int:D] $wtf; $wtf = Hash[Any, Int:D].new; $wtf<a> = 2; 12:12
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in binding to key; expected Int:D but got Str ("a")␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz m: use v6.c; my Hash[Int:D, Any] $wtf; $wtf = Hash[Int:D, Any].new; $wtf<a> = 2;
camelia ( no output )
Roamer` Xliff_zzzzz, hmm, I *thought* I'd seen the Hash[something, something] syntax, I should've tried it
Xliff_zzzzz, thanks
Xliff_zzzzz Wow. So it's Hash[ValueType, KeyType].... 12:13
Feels a little backward byt.
s/byt/but/
jnthn Just consistent parameter order
12:13 kyclark joined
jnthn Hash[ValueType] is a hash with the specified value but keys coerced to strings 12:13
smls_ Yeah, it would be weird if the optional one came first.
Roamer` btw, here's another one, is it a bug of dd? Part of the reason I thought it was Hash[ValueType]...
m: use v6.c; my Array[Str:D] %foo; %foo<a> = Array[Str:D].new(<a b c>); dd %foo;
camelia rakudo-moar f2df2c: OUTPUT«Hash[Array[Str:D]] %foo = (my Array[Str:D] % = :a(Array[Str:D].new("a", "b", "c")))␤»
Xliff_zzzzz Roamer`, why do you think that's a bug? 12:14
Roamer` Xliff_zzzzz, because the way the dd output starts ("Hash[Array[Str:D]]") suggests to me that this is how I should declare a hash with arrays as values 12:15
brb 12:16
Xliff_zzzzz Hummm.....
tbrowder nine: can you check my issue #100 in zef and ensure i said the right thing" thanks VERY much!
12:16 smls_ is now known as smls 12:17 gfldex joined 12:19 domidumont left 12:21 domidumont joined
nine tbrowder: added my findings 12:23
tbrowder thanks
12:26 kyclark left
nine Sometimes programming really is like a game of chess. You stare at a piece of code for half an hour, occasionally stroking your beard. And then change a single line to fix the bug. 12:28
masak nine: "single line" tends to correlate with "good architecture", IME
arnsholt And exponentially long time-to-bug-found, sometimes 12:31
12:32 wamba joined
masak dunno if I see any correlation between time-to-bug-found and architecture quality... there probably is one, but it's not entirely immediate to me. 12:33
in both cases you need to go and look; you need to get in a certain state of mind, etc
Xliff_zzzzz Can someone tell me if something like "sub EXPORT(*%l)" can parse "use MyModule <:a :b :c>" without the need of having those export tags present in the code? 12:35
Hmmm.... Or can I cheat and use dummy tags, but that would require adding dynamic subs to a package.
Roamer` jnthn, did you just suggest that Hash[ValueType] $foo should constrain $foo to only accepts values of this type? I guess you must have missed my original golf snippet then, so here it is again... 12:37
m: use v6.c; my Hash[Int:D] $wtf = Hash[Int:D].new(); $wtf<a> = 7.62; dd $wtf; dd $wtf<a>; my Int:D %fine; %fine<a> = 7.62;
camelia rakudo-moar f2df2c: OUTPUT«Hash[Int:D] $wtf = (my Int:D % = :a(7.62))␤Rat <element> = 7.62␤Type check failed in assignment to %fine; expected Int:D but got Rat (7.62)␤ in block <unit> at <tmp> line 1␤␤»
Roamer` jnthn, how did I put a Rat into a Hash[Int:D]?
12:38 adu joined
jnthn Roamer`: No idea. o.O Probably by constructing it directly, and uncovering a bug (because all the tests do it the `my Int:D %fine` way, I guess). Please RT. 12:39
Roamer` jnthn, ok, I will, thanks
CIAvash m: Hash[Str, Int].new: 1,2 12:44
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in binding to x; expected Str but got Int (2)␤ in block <unit> at <tmp> line 1␤␤»
CIAvash what is x?
m: Hash[Str, Str].new: 1,2
camelia rakudo-moar f2df2c: OUTPUT«Type check failed in binding to key; expected Str but got Int (1)␤ in block <unit> at <tmp> line 1␤␤»
12:47 kyclark joined 12:50 MasterDuke joined 12:54 MasterDuke left
timotimo internal name, apparently 12:54
12:55 rindolf joined 13:01 patrickz joined
patrickz pmurias++ #nqp-js on npm, yay! 13:03
13:04 domidumont left 13:05 domidumont joined
timotimo jnthn wrote he "Elimianted" something ... i wonder how that works 13:05
13:05 sufrostico joined
dalek c: 02f4ff8 | gfldex++ | doc/Language/typesystem.pod6:
show how to test for classness and enumness
 13:08
Xliff_zzzzz OK. 13:09
13:09 M-Illandan left
Xliff_zzzzz Another question. How can I access PACKAGE::EXPORT from a BEGIN phaser? 13:09
13:09 Sgeo left, ilmari[m] left
Xliff_zzzzz %PACKAGE::EXPORT doesn't work. Neither does ::("PACKAGE::EXPORT"). 13:09
13:09 ilmari[m] joined
Xliff_zzzzz ::("%PACKAGE::EXPORT") isn't the same thing. 13:10
13:15 M-Illandan joined
melezhik Hi! One question concerning perl6 gramma 13:19
timotimo grammophone :D
melezhik well, I mean grammar
let's say I have two tokens foo and bar 13:20
13:20 skids joined
perlpilot yes, perl 6 grammars hold the key to world peace 13:20
masak .oO( grammarophobe )
perlpilot (unfortunately, the world doesn't have a lock that will accept that key)
mst ugexe: why did your one liner wrt Build.pm need ::("Build") even though it had -MBuild on it btw? 13:21
melezhik I want to define that both token could be in input data
13:21 adu left
melezhik obviously token TOP { <foo> | <bar> } does not work 13:21
nine melezhik: how is that obvious? 13:22
perlpilot melezhik: did you mean both tokens could be simultaneously in the input? 13:23
melezhik ok, let me reshape my statement. I mean if input data has entries matched <foo> token and if the come first then only action binded to <foo> will be trigered
despite the fact input data may have entires matched to <bar> token 13:24
yes, exactly
I want to fire actions on both <foo> and <bar> sort of entries 13:25
moritz you could have <foo> & <bar> to require both to match
melezhik but every time it will be only <foo> or onle <bar>
moritz so you could do something like [<foo> & <bar>] || [<foo> | <bar> ]
though let me add that this is a pretty atypical use case for regexes or grammers 13:26
arnsholt melezhik: Are the entries matched *either* a foo *or* a bar, or can they be both at once?
melezhik mortiz: well, sounds good, but what if I have many of such types, like <foo>, <bar>, <baz> , then it's getting complicated
arnsholt It's not quite clear to me what you mean
moritz melezhik: I'm not called moritz.
arnsholt moritz: You're Spartacus? =) 13:27
moritz melezhik: then you instead find a common way to tokenize your input, and do the rest in user-level code
melezhik sorry for typo moritz:
moritz arnsholt: moricus!
masak .oO( Ave Imperator, moritzuri te salutant ) 13:28
melezhik moritz: not sure if I understand what you mean
arnsholt melezhik: Anyways, are you quite sure what you want isn't just /[<foo> | <bar>]+/? 13:29
moritz melezhik: I mean that regexes and grammars aren't a good tool to find all possible ways to parse a string
melezhik yes I am sure
13:29 MilkmanDan left
moritz melezhik: so instead you use regexes what they are good for, and parse in just one, more primitive way 13:29
melezhik: and write user-space code to generate all the possible ways to combine the primitive matches into more advanced matches 13:30
perlpilot notes that P6 does have an :exhaustive modifier ;)
melezhik moritz: yes, I know how to do what I need in primitive way, using plain regexp, I wanted to have it via grammars
moritz wants a pony in his grammar 13:31
melezhik sadly if I can't have grammars which works like I say ...
13:31 MilkmanDan joined
perlpilot m: "foo" ~~ m/(foo)/; say @(); say "@()"; 13:36
camelia rakudo-moar 68443f: OUTPUT«(｢foo｣)␤@()␤»
perlpilot Should @() interpolate withing double quoted strings? I think it should.
Xliff_zzzzz masak: "We who are about to die, salute you!" ??? 13:38
m: "foo" ~~ m/(foo)/; say @(); say "{@()}"; 13:39
camelia rakudo-moar 68443f: OUTPUT«(｢foo｣)␤foo␤»
CIAvash m: "foo" ~~ m/(foo)/; say @(); say "@()[]";
camelia rakudo-moar 68443f: OUTPUT«(｢foo｣)␤foo␤»
masak Xliff_zzzzz: yes, but note the extra "z"
Xliff_zzzzz: (it's a famous quote. even has its own Wikipedia page. gladiators used to said it to the emperor.) 13:40
Xliff_zzzzz masak: I didn't grok that so it fell into my ignore-and-hope-for-the-best filter
13:40 lambd0x joined
masak Xliff_zzzzz: sometimes my jokes are too clever for my own good. I subject #perl6 to them anyway (sometimes) because in my mind, this channel is associated with geeky in-jokes 13:41
13:44 cdg joined
arnsholt FWIW, I thought it was pretty amusing =) 13:44
masak arnsholt++ is on my mental checklist of people who would get the reference at-sight 13:45
perlpilot I dunno ... I thought it should have been "mortizuri" to keep things consistent ;)
masak perlpilot: but the actual word is "morituri"
I did consider both, but the way I did it fits better with the original word
perlpilot sure, but it all started from the "mortiz" typo
masak that spelling always reminds me of The Adams Family, for some reason 13:46
perlpilot ITYM "Addams Family" :)
masak oh, I most likely do
masak hangs his shed in hame 13:47
lambd0x Hello everybody!
timotimo greetings lambd0x
masak helloooo λ0x
ugexe mst: is just an artifact from when it used to be `require "Build.pm"` (before that stopped working for some Build.pm) 13:48
[ptc] moritz: it's Perl6, so I'm sure you *can* have a pony in your grammar ;-) 13:49
mst ugexe: ah. how come that stopped working but 'use' still does?
moritz speaking of nerdy (and slightly off-topic) things, I recommend the "Talk Nerdy" podcast: carasantamaria.com/podcast/
about hour-long interviews with scientists, science communicators, authors and other funny folks
kyclark How can I unpack a list of pairs in a for loop? 13:50
m: my @pairs = foo => 11, bar => 12; for @pairs -> [$key, $val] { put "$key = $val" }
camelia rakudo-moar 68443f: OUTPUT«Type check failed in binding to <anon>; expected Positional but got Pair (:foo(11))␤ in block <unit> at <tmp> line 1␤␤» 13:51
moritz m: for a => 1, b => 2 -> (:$key, :$value) { say $key }
camelia rakudo-moar 68443f: OUTPUT«a␤b␤»
ugexe mst: specifically I dont know. -M happens to take a different module loading path which somehow worked around the problem (which was precomp related)
moritz kyclark: ^^
mst weird 13:52
kyclark moritz, how/why does that work? I see I can’t use :$val but have to say :$value. Is it calling a method on the Pair? 13:53
moritz kyclark: yes 13:54
kyclark: it works for all objects, not just pairs
timotimo melezhik: i think you can go with / [ .*? "foo" .*? ] & [ .*? "bar" .*? ] /, but that could have some catastrophic backtracking, so may have to fiddle around a bit more
moritz kyclark: (and the [] unpacking in a signature only works for list-like things)
timotimo melezhik: personally, i woudln't do it with regex, but with a contains-substring-check :) 13:55
actually, if you put ^ at the beginning and end inside the brackets could work well
moritz then you could just as well match two separate regexes
13:55 kurahaupo joined
kyclark OK, then for my case a List will be a better structure. Thanks. 13:55
ugexe mst: it might have been fixed recently. the problem seemed related to precompiling the modules used by the Build.pm and nine did some commits that looked related 13:56
melezhik as It told, it is going to be very complicated if one have many data tokens ( not only <foo> and <bar>)
and in some case there are only a <foo> entires, and probably this solution won't work ... 13:57
timotimo well, the thing is, you're going against what regex wants. you're not caring about "match this first, then this"
melezhik or only <bar> entries in input data 13:58
timotimo you want "does this match anywhere? does this other thing match anywhere? how about this thing?"
personally, i'd split it into one .match or ~~ per token you want to find
melezhik I want 2 things 1) match all possible entries, if any (foo,bar, etc) 2) once entries matched - fire an actions for them ( this point is result of previous one ) 13:59
and third thing - I want to do it with a single perl6 grammar 14:00
moritz but why? 14:01
timotimo you can have "method TOP" which calls a match for each token
and runs the action methods manually perhaps?
melezhik moritz: why what? why I want it via perl6 grammar? 14:02
timotimo yes, why indeed 14:03
moritz melezhik: yes. And why do you want to do it in the first place? 14:04
melezhik ahh, I already have a perl5/simple regexps implementation for my parser, I wanted to port it to perl6, then I noticed a perl6 grammar which I found quite interesting and potentially minimizing a code to express such a things 14:05
arnsholt Do you have an example of the kind of text you're trying to parse?
melezhik sure
arnsholt This feels like an X Y problem
melezhik this is dsl specification - github.com/melezhik/outthentic-dsl...ode-syntax , I want to port it to perl6 using perl6 grammars, looks like I stick 14:07
14:07 harmil joined
melezhik let's have a simple case, I want to express generators: expression and regexp: expression in a single perl6 grammar rules 14:08
I know how to do this for every type of entry (regexp: and generator:) but when I need to have may grammar to match any of them I fail
like <generator> | <regexp> does not work 14:09
timotimo well, with a grammar, it's anchored at start and end
melezhik as if generator entires comes first in input stream then grammar matches them and stop
timotimo so in order for that to work it has to find generator at the beginning and it has to reach the end. alternatively, regexp has to come at the beginning and go to the end
melezhik and other vise if regexp: entries comes first then grammar find them and stop ( ignoring generator entries ) 14:10
timotimo entries? 14:11
oh, sorry
it's confusing that the parts of the regexp are named "regexp" and something else :)
so, did you try my solution with the .*? and the & in-between?
harmil melezhik: I think it could perhaps be clearer if you didn't specify the implementation via grammars that you think works, but rather an example text you wish to parse and the ultimate goal you wish to accomplish with it e.g. "1 + 2" and "I want to end up with a parse tree that looks like OP('+", VALUE("1"), VALUE("2"))". 14:13
melezhik yes, your solution works
then only concern is the number of entires types 14:14
harmil I say this coming in late, and only having skimmed the history, but I'm not clear on what you actually want to accomplish, here.
melezhik I have many like - generators, code, regexp, plain strings, asserts, text blocks ( you may take a look at the doc link I shared )
harmil : an ultimate goal is to port github.com/melezhik/outthentic-dsl...ode-syntax to perl6, and I am going to use grammars if it possible 14:16
timotimo what does "plain regular text within asserts range" mean?
arnsholt melezhik: gist.github.com/arnsholt/4ee0fa479...2beb922b7d 14:17
That's a sketch of how I'd start
The different declarations in your DSL are distinct, so it's a question of either mathing one kind of declaration or another
Then, if some of them share functionality, that's a question of how you structure the underlying logic classes, not the AST building
timotimo it already says your DSL is line-based 14:18
14:19 khw joined
timotimo so i wonder why you would want to use a full-text-at-once based thing, which is pretty much what grammars are 14:28
you'd have to re-create a line-by-line thing in grammars
it's not hard, but you could leave it out completely if you wanted
mst if you can't do line-by-line how do you have a repl? 14:29
14:34 kyclark left
timotimo the repl actually evaluates each line separately 14:36
which is something we'd like to change in the future, too
we've wanted to change that for a long time, actually
smls m: my $str = "a"; dd (($str ~~ /b/) // "").Str 14:38
camelia rakudo-moar 68443f: OUTPUT«""␤»
smls ^^ is there a neater way to stringify a regex match, with 'no match = empty string'? 14:39
my $str = "a"; say ~($str ~~ /b/) // ""
m: my $str = "a"; say ~($str ~~ /b/) // ""
14:39 nbg23 left
camelia rakudo-moar 68443f: OUTPUT«Use of Nil in string context in block <unit> at <tmp> line 1␤␤» 14:39
smls ^^ Would be cool if this worked, but I suppose there's no Failure-like system for warning? 14:40
*warnings
14:45 acrussell joined
timotimo you can "quietly { ... }" 14:46
m: say quietly ~Nil
camelia rakudo-moar 68443f: OUTPUT«␤»
timotimo m: say ~Nil
camelia rakudo-moar 68443f: OUTPUT«Use of Nil in string context in block <unit> at <tmp> line 1␤␤»
moritz m: say quietly ~('a' ~~ /x/)
camelia rakudo-moar 68443f: OUTPUT«␤»
moritz m: say quietly ~('x' ~~ /x/)
camelia rakudo-moar 68443f: OUTPUT«x␤»
14:51 kyclark joined
melezhik arnsholt: thanks for giving a hint, I will examine your draft soon more precisely , basically my code looks very familiar to what you suggest, I will let you know if run into troubles, 14:51
kyclark Are there profiling tools for Perl 6? 14:52
14:53 lostinfog joined
timotimo kyclark: we have "perl6 --profile ..." 14:53
kyclark Oh, cool. Sorry, should have been obvious.
timotimo hardly :) 14:54
14:55 pierre_ left 14:57 g4 left 14:58 cdg left
gregf_ llegedly 14:58
oops :/ sorry
15:00 cdg joined 15:10 sakuya joined 15:11 pmurias left, cdg left
perlpilot It would be neat if we had something like alpha.trycarbide.com/ for Perl 6 15:15
15:22 rmmm left, lambd0x left
Xliff_zzzzz Carbide would be a good thing for nqp.js, though.... right? 15:23
I would shudder to think of the work necessary to get that working for MoarVM. 15:24
But stilll...nifty.
sakuya Is IO::Socket::INET support bind a local port like perl5 does . 15:25
15:25 jonas2 left
sakuya perl5 code : IO::Socket::INET->new( PeerAddr => '127.0.0.1', PeerPort => 3333, LocalPort => 23356,Proto => 'tcp'); works fine 15:25
perl6 : IO::Socket::INET.new(:host<127.0.0.1>, :port(3333), :localhost<127.0.0.1>, :localport(23356); not working 15:26
15:27 zacts joined, Khisanth left 15:28 rmmm joined
[Coke] timotimo: (angular 2) nope, sorry. Did find docs on how to make a single-page pre-built app, though. 15:28
yoleaux 02:55Z <pmichaud> [Coke]: You're currently listed as an editor on rakudo.org... shall I increase that to "administrator"?
15:33 cognominal left
[Coke] .tell pmichaud that question predates you fixing the thing I couldn't fix 15:40
yoleaux [Coke]: I'll pass your message to pmichaud.
15:40 Khisanth joined
[Coke] .tell pmichaud regarding segregating roast queue vs. rakudobug queue - no, I don't think it's worth our time necessarily to force the tickets into the right queues. We can probably do just as well by having some metadata about tickets in RT 15:41
yoleaux [Coke]: I'll pass your message to pmichaud.
[Coke] .tell pmichaud for some tickets, it's easy or obvious to move them; but anyone working on anything in the toolchain or lang spec is going to deal with RT. 15:42
yoleaux [Coke]: I'll pass your message to pmichaud.
15:45 MetaZoffix joined
geekosaur sakuya, docs.perl6.org/type/IO$COLON$COLON...method_new 15:46
MetaZoffix 🎺🎺🎺 There are currently 55 RFC tickets: perl6.fail/t/RFC I invite users to review and comment on those RFCs, so the core devs would have some feedback about the proposed changes. 15:48
15:52 ufobat joined 15:56 patrickz left
sakuya geekosaur, the page has not what i want 15:57
dalek line-Perl5: dd99758 | niner++ | configure.pl6:
Fix configure.pl6 on rakudo-j

Sadly $*VM.config keys are not standardized among backends.		 15:58
16:00 cognominal joined 16:02 sakuya left 16:03 cdg joined 16:05 edehont joined 16:07 canopus left 16:10 domidumont left 16:14 itaipu joined 16:16 canopus joined
nine moritz++ # thanks for perl6-all-modules making my life easier 16:16
16:16 zacts left 16:21 rudolfochrist joined 16:26 MetaZoffix left 16:32 canopus left 16:39 canopus joined
moritz nine: you're welcome. Nice that others find it useful too 16:42
timotimo i just cloned it again the other day, but i wasn't good enough to find the right grep incantation for what i wanted to know 16:43
moritz timotimo: what did you want to know? 16:45
16:46 lizmat joined
timotimo wanted to know more about export/import tags 16:46
16:46 rudolfochrist left
moritz if/how they are used? 16:46
16:47 ChoHag left 16:49 dLeCamarae left
kyclark m: my @y = (1..10), (5..11), (12..15); my $x = (2..7); for @l -> $y { say so $x (&) $y } 16:49
camelia rakudo-moar 68443f: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Variable '@l' is not declared␤at <tmp>:1␤------> 3 (5..11), (12..15); my $x = (2..7); for 7⏏5@l -> $y { say so $x (&) $y }␤»
kyclark Sorry 16:50
m: my @y = (1..10), (5..11), (12..15); my $x = (2..7); for @y -> $y { say so $x (&) $y }
camelia rakudo-moar 68443f: OUTPUT«True␤True␤False␤»
kyclark Anyway to do that with hyper operator (>>)?
timotimo i was trying to help out xliff with the color names module
16:51 woolfy joined
nine moritz: in case you wonder. That's what I just used it for: github.com/perl6/toolchain-bikeshe...a2faa31593 16:55
16:56 pierre_ joined 17:00 pierre_ left 17:15 leont joined, zakharyas left 17:17 MilkmanDan left, mephinet joined 17:22 mcmillhj joined 17:26 MilkmanDan joined 17:40 firstdayonthejob joined 17:41 leont left, rudolfochrist joined 17:42 dakkar left, margeas left 17:45 LeCamarade joined
LeCamarade This is Rakudo version 2016.04 built on MoarVM version 2016.04 17:45
implementing Perl 6.c. 17:46
Hmm.
Time for an upgrade.
mst LeCamarade: do you happen to have a local::lib setup for perl5 and feel like testing something? 17:47
17:50 domidumont joined
LeCamarade I do not really understand Perl 5 stuff anymore (if ever), but I guess I can run whatever you have. 17:50
I mean, I do not know what local::lib would even be. I guess a library of some sort. 17:51
Xliff_zzzzz Moritz: How large is p6-all-modules 17:52
I may need to do some grep diving myself.
17:54 skrshn joined
skrshn Newbie question 17:54
m: sub avg(Int @n) { ([+] @n) / @n.elems }; say avg(1..10);
camelia rakudo-moar 68443f: OUTPUT«Type check failed in binding to @n; expected Positional[Int] but got Range (1..10)␤ in sub avg at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
Xliff_zzzzz Well... 187M; (Answering my own question)++ 17:55
That's actually not bad.
skrshn m: sub avg(Int @n) { ([+] @n) / @n.elems }; say avg([1..10]); 17:56
camelia rakudo-moar 68443f: OUTPUT«Type check failed in binding to @n; expected Positional[Int] but got Array ($[1, 2, 3, 4, 5, 6, 7...)␤ in sub avg at <tmp> line 1␤ in block <unit> at <tmp> line 1␤␤»
kyclark m: sub avg(*@n) { ([+] @n) / @n.elems }; say avg(1..10); 17:57
camelia rakudo-moar 68443f: OUTPUT«5.5␤»
kyclark You can't put constraints on slurpy args. I've learned this the hard way.
17:59 MasterDuke joined
MasterDuke m: sub avg(Int @n) { @n.sum / @n.elems }; say avg(Array[Int].new(1..10)) 17:59
camelia rakudo-moar 68443f: OUTPUT«5.5␤»
18:00 rudolfochrist left
tailgate How do I get File::Spec::Functions? 18:00
MasterDuke tailgate: you could clone and install it manually, or use panda or zef 18:02
tailgate Ah. I tried variations of `panda install File::Spec` and it wasn't in the ecosystem
and it's a dependency for perl6/doc 18:03
Xliff_zzzzz Much of the nitty gritty EXPORT examples seem to be in rakudo-p5
MasterDuke you may need to do a 'panda update' first
kyclark tailgate, I've been using the $*SPEC object for those functions, but I confess I hold my nose. I do not like that black magic and would prefer to explicitly bring in the module/functions I want. 18:04
Tell me if there is a better way.
Xliff_zzzzz Some of this looks promising... 18:05
18:06 bob777 left
MasterDuke tailgate: ahh, that's a Perl 5 module 18:07
18:07 itaipu left
MasterDuke kyclark: you might be interested in SPEC::Func 18:07
18:08 leont joined, itaipu joined
kyclark MasterDuke, how do I pull up docs on things like that? "perl6 --doc SPEC::Func" doesn't work. 18:08
MasterDuke kyclark: i'm not actually all that sure. i usually just go directly to the repo, github.com/zoffixznet/perl6-SPEC-Func in this case 18:09
kyclark OK, that's useful. Thanks! 18:11
18:22 itaipu left 18:23 domidumont left 18:27 itaipu joined
Xliff_zzzzz \o/ 18:30
\o\
/o/
Almost working.
I need to do "use Color::Names <a>.list" for some reason.
Otherwise I get a weird error.
If I "use Color::Names <a b>", it's a proper list and I don't need to .list 18:31
Doing "multi sub EXPORT($a)" doesn't seem to work. 18:32
m: dd <a>
camelia rakudo-moar 68443f: OUTPUT«"a"␤»
Xliff_zzzzz m: dd <a>.WHAT
camelia rakudo-moar 68443f: OUTPUT«Str␤»
18:34 cdg left, MetaZoffix joined
MetaZoffix Xliff_zzzzz: what's your export sub is like? Sounds like you just need to make it slurp 18:35
sub EXPORT (*@names) { ... }
kyclark: there's no magic. Perl 6 has dynamic variables. $*SPEC is a dynamic variable that has one of the IO::Spec::* objects in it (which, depends on the OS), and that object privides the methods you seek. 18:36
m: my $obj = $*SPEC; say $obj.splitpath: "a/b/c"; 18:37
camelia rakudo-moar 68443f: OUTPUT«( a/b/ c)␤»
MetaZoffix m: with $*SPEC { say join .dir-sep, .splitpath: "a/b/c";
camelia rakudo-moar 68443f: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Missing block␤at <tmp>:1␤------> 3 say join .dir-sep, .splitpath: "a/b/c";7⏏5<EOL>␤»
MetaZoffix m: with $*SPEC { say join .dir-sep, .splitpath: "a/b/c"; }
Xliff_zzzzz MetaZoffix, has always slurped.
camelia ( no output )
Xliff_zzzzz However if I don't specify argument to "use", I get "Cannot find method 'merge-symbols': no method cache and no .^find_method 18:38
"
MetaZoffix Got code you can pastebin?
Xliff_zzzzz MetaZoffix: Already pushed to GitHub --> github.com/Xliff/p6-color-names/bl...ames.pm#L7 18:39
18:41 girafe joined
MasterDuke Xliff_zzzzz: btw, [email@hidden.address] $_ eq @a.any })' could be [email@hidden.address] 18:43
MetaZoffix Xliff_zzzzz: is "Missing serialize REPR function for REPR VMException" your "weird error"?
MasterDuke m: my @a = <a b c d e f>; my @b = <c d z x f>; say @a.grep(@b.any) 18:44
camelia rakudo-moar 68443f: OUTPUT«(c d f)␤»
Xliff_zzzzz MetaZoffix, one of them.... yeah 18:45
MasterDuke++ # Conciseness!
MetaZoffix No idea. It may something to do with doing require in "mainline" or something. I had to avoid that in some of my modules. 18:47
nadim_ Hi, do we have something that can give us an object size, in bytes, like Devel::Size? 18:50
18:50 gabiruh left
MetaZoffix m: my @a = <a b c d e f>; my @b = <c d z x f>; say @a.Set{@b}:k 18:50
18:50 DANtheBEASTman left 18:51 kaare__ joined, camelia joined, M-Illandan left, sufrostico left, kaare_ left, krunen_ left, DrForr left, sjn left, japhb left, pmichaud left
skrshn MasterDuke: thanks 18:51
18:51 sufrostico joined 18:52 rindolf left, krunen left, Ulti left, masak left
MetaZoffix benchable6: HEAD m: for ^10000 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.Set{@b}:k } 18:52
18:52 literal left, avar left
benchable6 MetaZoffix, starting to benchmark the 1 given commits 18:52
MetaZoffix, ¦«HEAD»:3.6870
MetaZoffix benchable6: HEAD m: for ^10000 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.grep(@b.any) }
benchable6 MetaZoffix, starting to benchmark the 1 given commits
18:52 sjn joined, benchable6 left, hanekomu joined, masak joined, bitmap left, japhb joined, pmichaud joined, krunen_ joined 18:53 BooK joined
Xliff_zzzzz Well this is good progress for me! 95% complete. 18:53
18:53 ChanServ sets mode: +v camelia, Ulti joined, avar joined, avar left, avar joined
Xliff_zzzzz MetaZoffix++; MasterDuke++, timotimo++ 18:53
18:54 benchable6 joined
MasterDuke bench: compare HEAD for ^10000 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.Set{@b}:k } ||| for ^10000 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.grep(@b.any) } 18:54
benchable6 MasterDuke, starting to benchmark the 1 given commits
MetaZoffix benchable6: so... any word on that bench I asked you to do?
ooh, neat
MasterDuke it died on the second one
18:55 Matthew[m] left, sufrostico left, benchable6 left
MetaZoffix bench: compare HEAD for ^100 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.Set{@b}:k } ||| for ^100 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.grep(@b.any) } 18:55
MetaZoffix chuckles 18:56
MasterDuke the server is in the middle of building old commits
18:56 bitmap joined, DANtheBEASTman joined, chee joined, gabiruh joined, benchable6 joined
MetaZoffix bench: compare HEAD for ^100 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.Set{@b}:k } ||| for ^100 { my @a = <a b c d e f>; my @b = <c d z x f>; say @a.grep(@b.any) } 18:56
benchable6 MetaZoffix, starting to benchmark the 1 given commits
18:57 M-Illandan joined, Bucciarati joined 18:58 DrForr joined
benchable6 MetaZoffix, gist.github.com/9d2ec8216117a1b2fc...22f0b3876e 18:58
18:59 ugexe left 19:00 literal joined
harmil_wk MetaZoffix: what's ||| 19:01
I don't see it in the Routines doc
MetaZoffix harmil_wk: new operator we ivented. Makes stuff run about 1000x faster :P
harmil_wk: it's a special syntax for benchable6 robot to specify to different pieces of code to bench
*two
MasterDuke fyi, you can do more than two 19:02
harmil_wk Ah, darn, I was about to go retrofit my modules for turbo mode. :-)
MetaZoffix bench: compare HEAD my @r; for ^100 { my @a = 1e5..1e6; my @b = 1.1e5..1.1e6; @r.push: @a.Set{@b}:k } ||| my @r; for ^100 { my @a = 1e5..1e6; my @b = 1.1e5..1.1e6; @r.push: @a.grep(@b.any) } 19:03
benchable6 MetaZoffix, starting to benchmark the 1 given commits
19:04 krunen joined
El_Che If you build rakudo from source, it clones nqp from github and builds it first. If you build nqp it needs moar first. Circular dependecies much? (~_~) 19:04
19:04 rindolf joined 19:05 frew joined
MasterDuke MetaZoffix: btw, the compare option is a little fragile. because of the way it runs the code, it will complain about some things (like declaring a sub in each option with the same name) 19:05
El_Che (If you try to build a pkg, you don't want it to dynamically clone a master branch remotely)
19:07 benchable6 left, nebuchadnezzar left
MasterDuke huh, it's still running your benchmark, don't know why it quit 19:08
Xliff_zzzzz Weird... now I am getting this "Cannot invoke this object (REPR: Uninstantiable; EXPORT)" if I define multi EXPORTs
MetaZoffix MasterDuke: a wild guess is it's not parallelizing the requests, so it wasn't able to respond to a ping request 'cause it's busy doing the bench 19:09
Also... why is it taking so long to run it? There are just 100 iterations?
19:09 nebuchadnezzar joined 19:10 darutoko left, Matthew[m] joined
MetaZoffix Oh, I guess I over-did it with the array sizes 19:10
MasterDuke it runs each option 10 times 19:11
MetaZoffix oh
Well, even 1 time is taking forever on my box so.. If it's still running, it may be useful to kill it :P
MasterDuke and the server is pretty busy right now building+compressing commits 19:12
19:12 lostinfog left
MasterDuke ha, will do 19:12
19:13 benchable6 joined
MetaZoffix m: say "It would've taken more than {(9.930 + 127) * 2000 / 60} minutes" 19:15
camelia rakudo-moar 68443f: OUTPUT«It would've taken more than 4564.333333 minutes␤»
MetaZoffix Had to kill @a.Set{@b}:k version after 2 minutes, for a single iteration. 19:16
19:16 MetaZoffix left
nine El_Che: it only clones nqp if you tell it to with --gen-nqp 19:21
skids m: enum FF <zero one two three>; FF(2).perl.say; FF(two).perl.say 19:25
camelia rakudo-moar 68443f: OUTPUT«FF::two␤FF::two␤»
skids m: enum FF <zero one two three>; FF(FF(2)).perl.say;
camelia rakudo-moar 68443f: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤Unable to parse expression in typename; couldn't find final ')' ␤at <tmp>:1␤------> 3enum FF <zero one two three>; FF(FF(7⏏052)).perl.say;␤»
19:27 ugexe joined 19:32 wamba left 19:35 Wiertek joined 19:40 setty2 joined 19:41 rindolf left 19:42 Actualeyes left 19:46 rindolf joined 19:48 MasterDuke left 19:54 MasterDuke joined 19:55 eliasr joined 19:56 CIAvash left, Wiertek93 joined 19:57 smls left 19:58 cdg joined 20:00 Wiertek left, Wiertek joined 20:04 Wiertek93 left, cpage_ left, cpage_ joined
El_Che nine: yes, I get the --gen thing. But when you try to build nqp (to suppy to the rakudo build) it needs moar. I'll have a look later 20:05
[Coke] --gen-nqp implies --gen-moar, IIRC.
20:06 espadrine joined
[Coke] too much backlog: what's the question? 20:06
moritz [Coke]: it's the other way 'round 20:07
[Coke]: --gen-moar implies --gen-nqp 20:08
El_Che 1:04 < El_Che> If you build rakudo from source, it clones nqp from github and builds it first. If you build nqp it needs moar first. Circular dependecies much? (~_~)
21:05 < El_Che> (If you try to build a pkg, you don't want it to dynamically clone a master branch remotely)
[Coke]: I was looking at packaging, i.e. without the "generate from github part"
[Coke] it's not circular; rakudo needs nqp needs moar. 20:09
El_Che but moar is not release as a tar?
(or part of the rakudo dist?)
moritz sure there are more tarballs
20:10 pecastro left
moritz El_Che: take a look at a the rakudo-star makefile; it works with tarballs 20:10
[Coke] El_Che: moarvm.com/releases.html
El_Che moritz: great tip. Thanks
20:11 rudolfochrist joined, pmurias joined
[Coke] if you're packaging, I would recommend doing as the macports port does and have 3 packages. 20:11
pmurias perlpilot: is that carbide thing usable for big programs or just for short snippets
El_Che [Coke]: I am looking if fpm and docker could be a base for simple packages (I was thinking ubuntu and centos) 20:12
20:12 labster joined
[Coke] I know most of those words. 20:12
pmurias perlpilot: there are a lot of things that are ultra awesome for short things but don't scale at all 20:13
[Coke] on docker, there's already a prebuilt thing that includes docker/nqp/moarvm
El_Che [Coke]: I have my own rakudo docker image (using rakudobrew) for my apps
[Coke] sorry includes rakudo/nqp/moarvm
El_Che [Coke]: this docker is to build packages
[Coke] ... why? 20:14
... nevermind.
The time it would take you to explain it to me isn't going to help you. :)
El_Che so we get rpms & debs for poeple to install that want to try it
the docker part is that everyone here can run it, and create the packages when needed 20:15
20:15 rudolfochrist left 20:16 calopter joined
El_Che asking users to compile it themselves may be a deterrent, and several people here consider rakudobrew a rakudo dev tool not meant for end users 20:17
[Coke] "the docker part is" ahhhh, thanks
20:17 TEttinger joined
El_Che [Coke]: the easy way it to create full vm's, but that's not portable 20:18
[Coke] yah, that would have been a lovely option to have for the macports so everyone had the tools to build a port. Great idea. 20:19
20:19 skrshn left
calopter Hi! Is there a perl6 equivalent to haskell's iterate function? I'm looking for a way to generate a lazy list from a seed value and a function that produces the next element 20:23
geekosaur docker's kinda not really a thing on OS X (you can build and use it, but it's controlling virtualbox. there is zero chance cgroups get added to the OS X kernel)
moritz calopter: yes, the sequence operator can do that
m: say (1, { $_ * 2 + 5} ... *)[^10] 20:24
camelia rakudo-moar 68443f: OUTPUT«(1 7 19 43 91 187 379 763 1531 3067)␤»
El_Che geekosaur: yeah, I had specifically Ubuntu 16.04 and Centos7 in mind
calopter moritz: Thanks!
geekosaur that was aimed more at [Coke]'s comment re macports
[Coke] geekosaur: I'm using the native Docker.app now
geekosaur sure, and it's using a vbox under the covers 20:25
El_Che I had a look at Snap (the multi distro pkg by Ubuntu) but I haven't seen it in the wild yet
geekosaur docker needs cgroups, otherwise it can only control a hypervisor
[Coke] geekosaur: I'm not sure if by vbox you mean something as heavyweight as the oracle VirtualBox app which it was using before.
it's not using that now
geekosaur pretty sure it's using the virtualizaton engine under the covers. there's plenty of references to docker on os x and how it has no choice but to control a hypervisor 20:26
*maybe* they have some cut-down version that can partially --- but only partially --- simulate it using the sandbox framework.
but that will have lots of limitations 20:27
(and for some uses it's even slower than a hypervisor)
El_Che [Coke]: fpm does also create osx pkg files, but I'm not sure how the crosscompiling would work in that case
(you can run it natively of course)
20:27 itaipu left, pecastro joined 20:28 itaipu joined
calopter moritz: would this approach work with a stateful object like Inline::Perl5 WWW::Mechanize? I'm looking to build a lazy list of webpages, crawling from one to the next with follow_link() 20:28
geekosaur [Coke], also confusing the issue is 10.11 has a framework for wrapping hypervisors, which means it can be using one without you seeing it... 20:30
(there might even be a minimal hypervisor in there, which might be used by the sandbox on 10.11)
El_Che geekosaur: what you're describing is the exact way that it will work on windows (through hyper-v onder the covers) 20:31
geekosaur yeh
geekosaur already knew windows uses hyper-v for WoW (32 bit windows apps running in 64-bit windows)
El_Che github.com/mist64/xhyve <-- probably the one on osx 20:32
hharnisc.github.io/2016/06/16/deve...-2016.html 20:34
geekosaur indeed " The Docker engine is running in an Alpine Linux distribution on top of an xhyve Virtual Machine on Mac OS X "
there you go
moritz calopter: sure, why not?
geekosaur vaguely familiar with xhyve, knows it's modeled after bhyve (freebsd's answer to linux-kvm) 20:35
20:35 cpage_ left
calopter moritz: Cool. Coming from haskell so still wrapping my head around interacting with stateful stuff :) Thanks! 20:35
moritz calopter: though that sounds like something you might want to write as gather while condition { fetch; parse; take $result }
calopter moritz: that does sound good! 20:36
geekosaur the early betas were using virtualbox; I didn't know they'd switched to xhyve for the release
20:40 pullphinger joined
El_Che [Coke]: mac ports is pretty neat 20:43
looking at the ports now
20:45 jjido joined 20:52 zengargoyle left, jjido left 20:53 jjido joined, espadrine left 20:54 zengargoyle joined 20:59 zengargoyle left 21:00 skids left, pmurias left, zengargoyle joined 21:01 Wiertek left 21:04 cpage_ joined 21:13 mcmillhj left
El_Che ok, the container is building some stuff. I am curious :) 21:19
21:19 calopter left 21:22 nadim_ left 21:23 edehont left, nadim_ joined, pullphinger left
El_Che root@09a2f80615fe:~# /opt/rakudo/bin/perl6 -v 21:26
This is Rakudo version 2016.08.1 built on MoarVM version 2016.08
implementing Perl 6.c.
this could work \o/ :)
21:27 rindolf left 21:29 acrussell left 21:30 kyclark left
harmil_wk If I have two Promises, a long-running task and a timeout (ala Promise.in($seconds)) what should I do when the timeout happens to terminate the other Promise? 21:31
timotimo harmil_wk: check for the timeout regularly inside your work code
harmil_wk Unfortunately, that work code is calling into a library that might take too long. This is testing code. What I really want to do is shoot it in the head... 21:32
21:32 jjido left 21:33 mcmillhj joined
timotimo you can use NativeCall's CPointer to write to a memory location that's shared that you know will crash the thread without crashing the whole application 21:34
harmil_wk That seems ... graceless, but I guess it would work. There's really no ability to terminate a thread?
21:34 jjido joined
harmil_wk But thank you for answering, I don't want you to think I'm not grateful, just grumbly 21:35
21:38 mcmillhj left, AndyBotwin joined
timotimo i think at some point we may get a "cause that thread to have an exception thrown" feature 21:40
but terminating threads is pretty problematic
when you have the possibility to throw an exception, you can get things like "proper clean-up of shared resources and locks" 21:42
21:43 edehont joined
leont terminating threads cleanly is very non-portable, AFAIK 21:44
21:45 edehont left
timotimo that's what i've heard, yes 21:46
21:46 mcmillhj joined, margeas joined
timotimo i must say python's "KeyboardException" or "KeyboardInterrupt" or something seemed really cool until i tried to use it with a Real Program in the wild 21:46
leont I mainly know it because of some completion script that doesn't like being killed with control-C 21:47
timotimo heads towards bed
leont It annoys me every time
timotimo lik dis if you cri everitim 21:48
21:51 mcmillhj left
El_Che I am under the impression that panda and zef do not have something to specify a --prefix to be installed, am I right? 21:51
ugexe zef has -to / --install-to
El_Che ugexe: also for zef itself? 21:52
ugexe sure
El_Che ok, trying, thx
21:54 nadim_ left 21:56 woolfy left
El_Che ugexe: probably extremely silly, but zef complains about the dir: paste.ubuntu.com/23113650/ 21:59
ugexe: the idea is to add zef to a fresh compiled rakudo in order to pkg it
ugexe did you read the README? 22:00
El_Che yes, sorry if I missed something
ugexe github.com/ugexe/zef#custom-instal...-locations
El_Che thx
it works 22:03
22:03 grondilu left, mcmillhj joined, bbkr joined 22:04 cpage_ left
leont lizmat: have an issue for that test counter issue 22:04
Will try to polish tomorrow 22:05
22:07 kyclark joined 22:08 mcmillhj left, Sgeo joined
TimToady waves from DEN 22:09
lizmat waves from LJU (well, actually a few clicks away from that) 22:11
leont I challenge you to pronounce it properly! :-p 22:14
22:14 jjido left 22:15 mcmillhj joined
El_Che ugexe: I wanted to provide a generic zef on a system localtion so the user could install what I needed. I didn't take into account that when I do that, zef will try to install modules in this path (no writable to the user). 22:15
leont ugexe: I'd really like a zef on TAP::Harness
El_Che ugexe: so it's probably better the user installs his own zef 22:16
22:18 itaipu left, firstdayonthejob left
El_Che ugexe: maybe the config.json could help me here 22:18
good night, going to bed 22:19
22:19 itaipu joined, mcmillhj left 22:20 maybekoo2 joined
dalek c: 8cfffc2 | coke++ | doc/Type/Range.pod6:
remove trailing whitespace
 22:22
22:24 maybekoo2 left 22:25 maybekoo2 joined 22:27 skids joined 22:33 Wiertek joined, mcmillhj joined 22:35 cpage_ joined 22:36 kyclark left
ugexe leont: it already exists, you just have to disable the 2 default testing adapters that come first in the config: `perl6 -Ilib bin/zef --/prove --/default-tester test .` 22:38
22:38 mcmillhj left 22:40 RabidGravy left
ugexe github.com/ugexe/zef/blob/master/l...ce/TAP.pm6 # just basics 22:41
the indirect names are used so that older rakudos don't complain 22:42
El_Che: you can also install your modules invoking zef directly; it never needs to be installed at all 22:48
22:48 Wiertek left
leont Ah, cool 22:48
22:51 kurahaupo left 22:52 mcmillhj joined 22:56 harmil_wk left 22:57 mcmillhj left 23:05 leont left, Cabanossi joined 23:06 akosgheri left 23:21 mcmillhj joined 23:26 mcmillhj left
skids could someone with the appropriate bits flag RT#124251 as testsneeded? 23:32
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=124251
23:52 mcmillhj joined 23:53 MasterDuke left 23:56 mcmillhj left 23:58 pierre_ joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!