»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by masak on 28 November 2015.
Zoffix m: class Foo { has @!d = 'tmp', 'foo'; method m { EVAL qq{use lib @!d}; say $*REPO.repo-chain; } }; Foo.new.m 00:01
camelia rakudo-moar 5e4927: OUTPUT«5===SORRY!5=== Error while compiling /home/camelia/EVAL_0␤Variable @!d used where no 'self' is available␤at /home/camelia/EVAL_0:1␤------> 3use lib @!d7⏏5<EOL>␤»
Zoffix hm? How do I make self available in that eval?
Or do I haveta use a temp var?
m: class Foo { has @!d = 'tmp', 'foo'; method m { my @d = @!d; EVAL qq{use lib @d}; say $*REPO.repo-chain; } }; Foo.new.m 00:02
camelia rakudo-moar 5e4927: OUTPUT«(file#/home/camelia/foo file#/home/camelia/tmp inst#/home/camelia/.perl6/2015.11-359-g5e49279 inst#/home/camelia/rakudo-m-inst-1/share/perl6/site inst#/home/camelia/rakudo-m-inst-1/share/perl6/vendor inst#/home/camelia/rakudo-m-inst-1/share/perl6)␤»
00:03 roguelazer left 00:04 pierre-vigier left
lucasb m: my @a = <a b c>; EVAL 'use lib @a'; say $*REPO.repo-chain[^3] 00:04
camelia rakudo-moar 5e4927: OUTPUT«(file#/home/camelia/c file#/home/camelia/b file#/home/camelia/a)␤»
Zoffix If anyone got a better solution, please comment on this PR: github.com/tokuhirom/p6-Crust/pull/66 00:05
lucasb interesting, they get pushed in the reverse order
00:05 roguelazer joined
Zoffix 46 dists to go... this is taking longer than I thought... 00:05
hoelzro MadcapJake: did you end up getting that Perl6::Compiler change to work? 00:08
00:09 BenGoldberg joined, cygx left, rurban left 00:10 csd_ joined
jdv79 Zoffix: you're doing PRs for all repos on gh for the @*INC thing? 00:10
00:11 skids joined, lucasb left
Skarsnik zoffix is fixing everything! 00:12
00:12 lizmat joined
Zoffix jdv79, for now, I'm doing all the @*INC things that can simply be replaced by "use lib". In total there are 65 dists that have @*INC stuff and I'm moving stuff that isn't "use lib" into a special folder that I hope to learn how to fix and send PRs for those later... Stuff like "for @*INC -> $f is copy {"... IIRC that's not yet implemented. 00:12
And nine++ is working on it (making module like Pluggable working, etc)
Skarsnik Leave if to the module author I say x) 00:13
jdv79 yeah
Zoffix Module author might not always be present, and with influx of users we'll get soon, it's be helpful to at least have a fixed forked version they could clone and install 00:14
leedo could @*INC be made some sort of proxy that does the right thing with the new method of doing things?
Zoffix leedo, don't think it's worth it TBH.
leedo ah k, haven't noticed much fallout myself from that 00:15
Zoffix I mean, after these 65 modules are fixed, we won't need to deal with @*INC
jdv79 you would think authors wold notice in 20 days or so
Zoffix ¯\_(ツ)_/¯ 00:16
masak ah; I see perl6advent.wordpress.com/2015/12/...s-in-them/ is published now :)
'night, #perl6
Zoffix \o/
night
.tell masak masak++ I've learned something: you can't have $foo-2 :) 00:19
yoleaux Zoffix: I'll pass your message to masak.
Zoffix (as a variable name)
Skarsnik hm 00:20
jdv79 that's why underscores are still better;) 00:21
Zoffix And I totally love that you can have hyphens in identifiers... Perl 5 made me hate how other langs use regexes... Perl 6 will make me hate how other langs don't allow hyphens :)
*shrug* my pinky appreciates not having to mash SHIFT as often :)
flussence being able to use “” (altgr+vb for me) instead of "" is a great improvement too 00:22
Skarsnik err wtf http::useragent lost a part of the cookies between two requests 00:23
AlexDaniel Zoffix: surprisingly, when I was designing my keyboard layout I decided that ‘-’ is a rare character, so now it's on altgr+shift level…
Zoffix heh
AlexDaniel Zoffix: while both _ and – are easier to press
00:23 Ben_Goldberg joined
Zoffix .u – 00:23
yoleaux U+2013 EN DASH [Pd] (–)
Zoffix AlexDaniel, yeah, if you use proper en/em dashes, hyphen *is* a rare char :) 00:24
timotimo .u –
yoleaux U+2013 EN DASH [Pd] (–)
timotimo .u —
yoleaux U+2014 EM DASH [Pd] (—)
AlexDaniel Zoffix: yup
Zoffix: Perl 6 can possibly make me rethink that…
timotimo oh my 00:25
in just 26 days we're going from HORIZONTAL BAR to DOUBLE VERTICAL LINE
Zoffix timotimo, huh?
jdv79 Skarsnik: there's bugz
flussence
.oO( these mojolicious release codenames are getting weird )
jdv79 i've been too lazy to fix them 00:26
00:26 BenGoldberg left
timotimo m: say uniname 0x2015; say uniname 0x2016; 00:26
camelia rakudo-moar 5e4927: OUTPUT«HORIZONTAL BAR␤DOUBLE VERTICAL LINE␤»
hoelzro MadcapJake: it looks like you got it!
jdv79 should probaly report them at least
Zoffix timotimo, what's that about?
flussence m: say uniname 2015;
camelia rakudo-moar 5e4927: OUTPUT«NKO LETTER LA␤»
timotimo nothing much
Zoffix oh
timotimo m: say uniname 0o2015; say uniname 0o2016
camelia rakudo-moar 5e4927: OUTPUT«CYRILLIC CAPITAL LETTER I WITH GRAVE␤CYRILLIC CAPITAL LETTER SHORT U␤»
Zoffix timotimo, i didn't bother actually READING those UTF codes :P
Skarsnik jdv79, they get cut gist.github.com/Skarsnik/c345897985eab0edeb40 00:27
timotimo :P
flussence m: say uniname (2015 + 4714); # julian date
camelia rakudo-moar 5e4927: OUTPUT«TAI THAM LETTER HIGH HA␤»
flussence guess the hex one is more interesting after all
jdv79 Skarsnik: yes
Skarsnik let me grab some tea and have a look xD 00:28
jdv79 i have a partial patch but this damned hotel wifi is broken of course
AlexDaniel m: my \o-o'o-o'o-o = 5; say o-o'o-o'o-o 00:29
camelia rakudo-moar 5e4927: OUTPUT«5␤»
00:30 geraud left, RabidGravy left
jdv79 Skarsnik: gist.github.com/anonymous/ae41051c8c0ffb1cc45c 00:30
maybe that's relevant
dalek kudo/nom: bb59896 | TimToady++ | src/Perl6/Grammar.nqp:
import should reuse precedence of exported ops

Fixes #126761.
AlexDaniel m: my \o-o = 10; say o-o- -o-o 00:31
camelia rakudo-moar 5e4927: OUTPUT«20␤»
AlexDaniel m: my \o-o = 10; say -o-o- -o-o 00:33
camelia rakudo-moar 5e4927: OUTPUT«0␤»
AlexDaniel m: my \o-o = 10; say -o-o- - -o-o- - -o-o
camelia rakudo-moar 5e4927: OUTPUT«-30␤»
jdv79 Skarsnik: not sure (not that i was able to load your gist) 00:34
Skarsnik the cookie returned by website is Set-Cookie: __cfduid=d264c3aa05f93f44b672cc132da7c55811449275073; expires=Sun, 04-Dec-16 00:24:33 GMT; (and other stuff) and it get cut to "Set-Cookie: __cfduid=d264c3aa05f93f44b672cc132da7c55811449275073; expires=Sun, 04" 00:36
jdv79 should get cut on comma from what i was seeing 00:37
weird
AlexDaniel m: my \o-o = {10}; say o-o o o-o
camelia rakudo-moar 5e4927: OUTPUT«-> |args is raw { #`(Block|86615248) ... }␤»
dalek ast: fae390f | TimToady++ | S06-operator-overloading/imported-subs.t:
test for RT #126761, import precedence correctly
 00:38
Skarsnik let's see 00:41
now it get cut in 3 pieces and some are lost 00:42
00:42 leont left
Skarsnik but yes , seens to be an issue 00:42
TimToady jdv79: fixed your import bug 00:43
00:43 raiph left
jdv79 TimToady: thank you sir! i'll check it out once i have enough internets to get it:) 00:44
hoelzro around, Linenoise should work again 00:45
Skarsnik how I put like a say in the field-value methid? before make it work?
hoelzro *alright
jdv79 Skarsnik: maybe the value token in the cookies grammar is missing -? seems weird that would be the case
i don't know why i'm always surprised when tests are not covering something "obvious" 00:46
Skarsnik why there is a grammar in the test file Oo 00:48
AlexDaniel m: my \O'O-O = 'O'.ord; say '0'-O'O-O-'0'-O'O-O-0-'0' 00:49
camelia rakudo-moar 5e4927: OUTPUT«-158␤»
jdv79 yeah that test file is funny 00:50
because "DATE" is the same thing as the actual thing:)
good luck
Skarsnik why the start of the grammar is <cookie>, ? 00:51
like why the coma
jdv79 no idea 00:52
Skarsnik www.ietf.org/rfc/rfc2109.txt no coma in this x) 00:54
MadcapJake hoelzro: you tried it and it works? o_O 01:01
jdv79 we don't have enough devs to fix all the bugs:(
Skarsnik lol wtf date converted to a DATE string 01:02
+get
I want to patch this to have AT-POS on HTTP::Cookies and AT-KEY on HTTP::Cookie x) 01:05
this grammar does not take the - 01:08
Zoffix w00t! I'm done (for today) 46 PRs sent! :) I wish GitHub were doing their t-shirt context this month :) 01:11
dalek kudo/nom: 56836b6 | TimToady++ | src/core/Rat.pm:
last patch had copy/paste error
Skarsnik T-shirt contest? 01:12
Zoffix Skarsnik, yeah, they've done Hacktober in October: send 4 PRs and get a T-Shirt: github.com/blog/2067-hacktoberfest...in-october 01:13
... which reminds me... I still haven't gotten mine :/
01:15 pierre-vigier joined
Skarsnik hm <-;> is not the syntax to everything exepct ;? 01:16
Zoffix No
<-[;]>
Skarsnik thx, that fixed this dumb grammar xD
Zoffix \o/
Skarsnik not sure it's really exact, but all test still pass 01:17
01:18 BenGoldberg joined 01:19 Ben_Goldberg left
Skarsnik look better: gist.github.com/Skarsnik/c345897985eab0edeb40 01:20
but it still does not work. I am not logged x)
why the httponly disapear lol 01:22
Zoffix "X-Powered-By: PHP/5.5.9-1ubuntu4.14"... well, there's your problem X) 01:23
01:24 Zoffix left 01:26 pierre-vigier left 01:27 pierre-vigier joined
Skarsnik hm, how I say to a token to ingore the case? 01:29
01:29 Ben_Goldberg joined
PerlJam Skarsnik: token foo { :i foo } 01:29
yoleaux 1 Dec 2015 22:23Z <Zoffix> PerlJam: I've seen several times today people who [I assume want to write] disappointed that Advent list is full. Maybe there's some sort of "official" way we can have a free-for-all and launch a bunch of articles per day, if people wanna write, since this Christmas is THE Christmas?
Skarsnik Oh it's in the token
thx!
PerlJam Hmm. 01:30
01:30 BenGoldberg left, moei joined
Skarsnik Good night here! 01:33
01:35 Skarsnik left
timotimo i'm up way later than i wanted 01:35
PerlJam .tell Zoffix Sorry I'm late but ... Sure, if people want to do multiple articles per day, I don't see why not. Everyone may need to increase the communication factor a little bit though. And I'm not sure if later-in-the-day articles will hide earlier ones or not, but either way, perhaps for those days that have multiple articles, the main one for the day could just be pointers to the others. 01:38
yoleaux PerlJam: I'll pass your message to Zoffix.
gtodd that would be cool :-D
01:40 cdg left
PerlJam And the schedule is not completely full. I see that days 6 and 13 are still open (at least according to what's in mu) 01:40
PerlJam is happy to see autarch in the list of contributors this season :) 01:42
01:57 pierre-vigier left 02:03 llfourn joined 02:05 kjs_ joined
bzipitidoo In doc.perl6.org/language/subscripts, what does the guillemet do in this code? $element = do given $value».split(/',' \s+/).flat { 02:08
02:09 keix left 02:11 keix joined
geekosaur bzipitidoo, design.perl6.org/S03.html#Hyper_operators 02:13
PerlJam bzipitidoo: it hyperizes. If $value has more than one thing in it, they all will have .split called on them. 02:14
MadcapJake what would be the equivalent way of using `call` in javascript with a thisArg (lexical context) 02:17
02:18 yqt left 02:24 FROGGS_ joined 02:27 raiph joined, FROGGS left 02:28 flussence left, flussence joined, adhoc left
PerlJam MadcapJake: you mean you want to call a method with a different self? 02:39
MadcapJake: or that you just want to do some-method($this, ...); instead of $this.some-method(...); 02:41
02:43 pierre-vigier joined 02:46 adhoc joined 02:54 flussence left, flussence joined 03:06 noganex_ joined
MadcapJake i ended up just opting for a first argument as a context 03:07
MadcapJake is off to bed!
03:07 MadcapJake left 03:08 pierre-vigier left
lucs Can I have sub MAIN (Bool $foo) ... , and if so, how do I pass the boolean on the command line? 03:08
(tried 1 and False, no luck)
ugexe yes, and --foo
lucs Ah, right 03:09
Thanks.
03:09 noganex left, pierre-vigier joined
ugexe sub MAIN (Bool :$foo) tho (named :$foo) 03:09
PerlJam I think that changes the nature of lucs' original question substatially ;) 03:10
lucs Bah, good enough :) 03:11
(better even)
ugexe i was just hoping he wasnt crazy enough to want positional bool command line arguments 03:12
PerlJam lucs: sub MAIN(Bool(Cool) $foo) { ... } would allow you to run it as ./foo 1 or ./foo 0
lucs PerlJam: Aha. Thanks, noted. 03:13
PerlJam probably over-uses "foo"
lucs Different namespaces, no problem :)
03:17 xtreak left
raiph m: EVAL(now.perl) 03:17
camelia rakudo-moar 56836b: OUTPUT«===SORRY!===␤Cannot find method 'compile_time_value'␤»
raiph sorry, ww
03:20 pierre-vigier left 03:21 pierre-vigier joined
PerlJam lucs: it occurs to me that I should probably mention that my version will generate true values for practically anything you pass to the program, except for things that evaluate to 0 (that's the price for being a Cool Bool :) 03:21
i.e. ./foo goober -> True ./foo 000 -> False 03:22
lucs PerlJam: Makes sense, and perfectly perlish. 03:25
gtodd can dd show me the body of a sub in the REPL ? dd &some_sub shows the signature and then { #`(Sub|34524460272) ... } 03:26
ugexe whats the best way to pass False to Bool :$foo from the command line?
gtodd jjust wondered if I could expand that and see what is in "#`(Sub|34524460272) "
PerlJam ugexe: ./foo --/foo 03:28
(I'm not sure why / was chosen as the character for "not") 03:30
ugexe PerlJam: nice. dunno how ive never seen anything on it
PerlJam ugexe: it's in S19 somewhere 03:31
gtodd I'm thinking of the way "dd" (Data::Dumper) in perl5 can set $Data::Dumper::Deparse = "1"; to deparse code refs
lucs It's a pretty good choice, as it should't interfere with shell expansion.
n
PerlJam aye 03:32
03:36 pierre-vigier left 03:43 pierre-vigier joined 03:45 keix left, keix joined
flussence gtodd: well, there's &?ROUTINE.file and &?ROUTINE.line, and we have CUR stuff now, so it may not be completely impossible to look up the original source... won't be easy though 03:46
03:49 japhb left
TEttinger PerlJam: the not equal sign in mathematics is a slashed equals sign, maybe some reason like that? 03:49
03:49 japhb joined 03:52 uruwi_ joined 03:53 skids left, uruwi left
gtodd flussence: wow cool 03:54
flussence: I just meant something I had written way back up in the repl and didn't want to scroll back and find ... but finding source of functions from loaded modules would be a great feature ... 03:55
for the repl
03:56 vendethiel left 03:58 webstrand joined 04:03 kid51 joined
webstrand I'm having trouble getting panda to work. I've tried v2015.10 but it fails due to some nil error, and the latest fails with dpaste.com/3ZPERBB . I'm using rakudo and moarvm 2015.11, is there some version incompatibility? 04:07
awwaiid I'm seeing some holes in the docs that can be filled with things from like S02. Two questions (1) why weren't the Synopsis used as the seed for the docs? (2) can I take entire swaths of these, such as design.perl6.org/S02.html#Names, and bring them somewhere into docs? 04:19
04:19 vendethiel joined 04:21 xtreak joined 04:23 kid51 left 04:27 Zoffix joined 04:28 xtreak left 04:31 abaugher_ left, webstrand left 04:32 Ben_Goldberg left, abaugher_ joined, Ben_Goldberg joined 04:33 Ben_Goldberg left, Ben_Goldberg joined
AlexDaniel .tell Skarsnik Whoops! “Failed to copy '/home/alex/.panda-work/1449289967_1/lib/mysql/SQLite/StatementHandle.pm6” 04:35
yoleaux AlexDaniel: I'll pass your message to Skarsnik.
04:43 vendethiel left
AlexDaniel Hm! I've never tried editing a file right on GitHub. That was pretty quick. Anyway: github.com/perl6/DBIish/pull/38 04:44
04:51 webstrand joined
dalek c/qualified-names: e2054c9 | (Brock Wilcox)++ | doc/Language/ (2 files):
Add basic info on qualified names, taken from S02
 04:54
04:57 Calibellus left 05:00 pierre-vigier left 05:06 kjs_ left 05:12 khw left 05:13 Actualeyes joined
chansen_ gist.github.com/chansen/6076c2c2c7236dda5fe1 05:19
^^^ Named parameters multiple dispatch bug?
05:20 zacts joined 05:21 webstrand left
ugexe order by constraint count desc 05:21
chansen_ Ahh, works 05:24
Why is that necessary? 05:25
05:25 Ben_Goldberg left 05:28 yeahnoob joined
ugexe because the original first new matched all your signatures. new doesn't imply *only* the arguments passed in (maybe inheritence will add them later for instance) 05:30
chansen_ ok, thanks 05:31
05:31 AlexDaniel left
ugexe you might want to look at submethod BUILD 05:33
grondilu m: module tau { constant tau is export = 2*pi }; import tau; say tau 05:43
camelia rakudo-moar 56836b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/emrpvFheer␤Cannot import symbol tau from tau, because it already exists in this lexical scope␤at /tmp/emrpvFheer:1␤------> 3stant tau is export = 2*pi }; import tau7⏏5; say tau␤»
grondilu ^this does work if I put the module in a file. 05:44
05:44 pierre-vigier joined 05:45 pierre-vigier left 05:46 kaare_ joined
grondilu github.com/grondilu/tau/blob/master/tau.pm6 05:47
05:53 mohae left
ChoHag But isn't the whole point of tau that pi is half it, not that it's twice pi? 05:56
And why do you hate unicode?
grondilu in Perl 6, pi is defined as 'pi', not 'π'. I think a tau module should have the same restreint 05:58
m: sat π
camelia rakudo-moar 56836b: OUTPUT«5===SORRY!5=== Error while compiling /tmp/D1PZCqt4TY␤Undeclared routine:␤ sat used at line 1. Did you mean 'set', 'say'?␤␤»
grondilu m: say π
camelia rakudo-moar 56836b: OUTPUT«3.14159265358979␤»
grondilu oh
sorry
I thought it was not defined. 05:59
ChoHag I had no idea anyway.
06:17 n0tjack joined 06:19 raiph left 06:22 n0tjack left
ChoHag Has anyone written (or adapted cperl into) a perl 6 emacs mode? 06:27
06:30 Actualeyes left 06:36 Actualeyes joined 06:58 km3 joined
uruwi_ m: $_ % 65536 for ^1000000; say now - INIT now; 07:02
camelia rakudo-moar 56836b: OUTPUT«2.36885743␤»
uruwi_ m: $_ +& 65535 for ^1000000; say now - INIT now;
camelia rakudo-moar 56836b: OUTPUT«0.3290170␤»
07:15 redhands joined 07:24 rurban joined 07:32 molecules joined, molecules is now known as Guest59403, Guest59403 left 07:45 yeahnoob left
uruwi_ m: my \a = 5; a for ^1000000; say now - INIT now; 07:47
camelia rakudo-moar 56836b: OUTPUT«0.24396763␤»
uruwi_ m: my $a = 5; $a for ^1000000; say now - INIT now;
camelia rakudo-moar 56836b: OUTPUT«0.3849377␤»
uruwi_ m: my \a = 5; a for ^1000000; say now - INIT now;
camelia rakudo-moar 56836b: OUTPUT«0.2504394␤»
uruwi_ m: my $a = 5; $a for ^1000000; say now - INIT now;
camelia rakudo-moar 56836b: OUTPUT«0.24502646␤»
uruwi_ m: my \a = 5; a for ^1000000; say now - INIT now;
camelia rakudo-moar 56836b: OUTPUT«0.24859563␤»
uruwi_ m: my $a = 5; $a for ^1000000; say now - INIT now;
camelia rakudo-moar 56836b: OUTPUT«0.2445257␤»
07:50 CIAvash joined
uruwi_ Anyone know where I can get the source for gen/moar/m-BOOTSTRAP.nqp? 07:53
07:53 domidumont joined 07:55 RabidGravy joined 07:58 domidumont left 07:59 domidumont joined
moritz uruwi_: it's src/Perl6/BOOTSTRAP.nqp 08:01
er, no
src/Perl6/Metamodel/BOOTSTRAP.nqp 08:02
08:02 rurban left
uruwi_ Apparently this part is slow? github.com/rakudo/rakudo/blob/nom/....nqp#L2000 08:06
There's supposed to be an anonymous routine there, isn't it? 08:07
moritz uruwi_: what makes you think it's slow?
looks like something that usually runs at compile time 08:08
08:08 redhands left
uruwi_ Ah, then. 08:08
moritz uruwi_: I'm still curious, why do you think it's slow? 08:12
did it show up in a profile or something?
uruwi_ Yes
I was profiling Digest::MD5, to be exact.
The question is: Why did it show in the profile? 08:13
08:13 uruwi_ is now known as uruwi
moritz uruwi: does Digest::MD5 do some fancy multi dispatch? 08:13
uruwi: and second question, how long is the run time of the code that you profiled? 08:14
uruwi I typically manage only 30 hashes per second 08:15
moritz uruwi: that doesn't quite answer my question
uruwi Over 3 seconds to calculate 100 hashes
moritz ok
so it should be long enough to drown out compile time 08:16
uruwi High-performance Perl 6 is a nightmare. 08:19
08:26 darutoko joined 08:28 xpen joined
RabidGravy is Digest::MD5 doing it all in Perl? 08:29
uruwi Yes
And not very performant Perl, apparently, either.
RabidGravy looking at the code it could probably do with a bit of love 08:30
08:30 xpen left
uruwi Hold on 08:31
I'm working on optimizing it
github.com/bluebear94/perl6-digest-md5 08:32
moritz I'm sure that works with arrays of integers somehow, right?
if you manage to make them native and shaped ("my int @a[64]") or so, that should give a huge boost
dalek Iish: 4a26f60 | (Aleks-Daniel Jakimenko)++ | META.info:
Typo in META.info

It seems like the typo was made during copy-pasting.
  (“DBDish/SQLite” was renamed to “mysql/SQLite” instead of “DBDish/mysql”)		 08:34
Iish: f1be21e | moritz++ | META.info:
Merge pull request #38 from AlexDaniel/patch-1

Typo in META.info
uruwi moritz: Sadly I can't test that because I only have 2015.09 installed. 08:35
moritz uruwi: time to upgrade :-)
uruwi: but even 2015.09 shoulld support arrays of native ints, even if not shaped 08:36
uruwi Heh, I was just waiting for MSI's to be released.
bartolin_ there is a failing test in S32-hash/adverbs.t which looks very strange. the code similiar to the following does not die as expected when I run the test file: 08:37
m: my %h = b => 2, c => 3; say %h<b>:zorp
camelia rakudo-moar 56836b: OUTPUT«Unexpected named parameter 'zorp' passed␤ in block <unit> at /tmp/m8_t6hyU7j:1␤␤»
bartolin_ maybe it is somehow related to precompilation of lib/Test.pm? If I add a line to that file, the first execution of the command with 'use Test' works 08:39
you can see the behaviour in that gist: gist.github.com/usev6/13268ed977a0b1ff69a6
is that a known issue?
moritz bartolin_: not as far as I know 08:40
grondilu there's a RC entry for MD5, btw. rosettacode.org/wiki/MD5/Implementation#Perl_6 Feel free to propose perf improvements there. 08:41
08:41 nadim left
bartolin_ moritz: thanks. I'll rakudo bug it ... 08:42
08:42 nadim joined 08:51 n0tjack joined
RabidGravy in markdown, how does one stop eg $*REPO being rendered as $<i>REPO</i> ? 08:54
moritz RabidGravy: `$*REPO` 08:55
RabidGravy: backticks are for code
08:55 n0tjack left
RabidGravy ah ha! 08:55
cheers 08:56
Woodi gfldex++ # ~~ Proxy; # True 09:11
moritz masak++ # advent post 09:14
09:15 muLinuxer joined 09:17 rindolf joined 09:18 domidumont left 09:30 muLinuxer left 09:35 Alina-malina joined 09:39 RabidGravy left
nine Oh how I love not having to re-install modules after building rakudo :) 09:47
lizmat good *, #perl6!
yes!
Although, after a recompile, the following still errors: 09:48
$ perl6 -MText::CSV -e ''
===SORRY!===
Cannot find method 'run_alt'
doing it a second time, works fine
[Tux] probably also the cause for Travis to report Failure
lizmat so I assume the precomp is messing up something, causing the loading to ultimately fail
the first time 09:49
[Tux] (no timings today: I'm working on a better IO::String, but that proves harder than I imagined)
09:51 Alina-malina left
[Tux] IO::Blob is too overengineered for my purpose and IO::String doesn't offer enough (yet) 09:52
bartolin_ lizmat: getting a different result at the first run of a command while using a module sound similiar to the spectest failures mentioned here: gist.github.com/usev6/13268ed977a0b1ff69a6 09:53
lizmat: also irclog.perlgeek.de/perl6/2015-12-05#i_11661151
lizmat bartolin_: ah, interesting! 09:54
I hadn't realised thattest-files were being precomped ? 09:55
bartolin_ but lib/Test.pm get's precompiled, maybe?
lizmat BTW, is it correct that Inline::Perl5 still doesn't install because of Pod::Coverage hanging ? 09:56
nine Inline::Perl5 doesn't use Pod::Coverage
But it needs to locate it's p5helper.so which used to crawl @*INC 09:57
lizmat nine: so Inline::Perl5 still doesn't install, right ? Just a reality check 10:03
nine correct
lizmat :-(
nine YES YES YES YES YES YES YES!
lizmat sorry, I guess you've been asked that too many times the past week... I haven't been online much :-( 10:04
nine It works!
lizmat you mean you fixed it just now ?
10:04 nadim left
dalek kudo/dist_resources: 995375e | (Stefan Seifert)++ | src/ (7 files):
Implement %?RESOURCES

Allows for e.g. Inline::Perl5 to find its p5helper.so both when loaded from a file system directory and when loading an installed version by just accessing
  %?RESOURCES<p5helper.so>
Needs to be cleaned up a little before merging.		 10:06
nine lizmat: ^^^
rakudo's dist_resources branch and panda's resources branch make it happen
lizmat way cool, nine++
10:06 Begi joined
nine Needs just a little more work before merging, but we're almost there :) 10:07
Now I have to go buy a christmas tree and many other errands, so I won't be online till probably in the evening.
lizmat nine: happy shopping! 10:08
nine .tell FROGGS I worked around the problem by not referencing the CompUnit::Repository directly in Distribution::Resources. I only store its path-spec which I can use to get the object later on (at runtime instead of compile time)
yoleaux nine: I'll pass your message to FROGGS.
10:09 Begi2 joined 10:12 nadim joined 10:13 Begi left 10:18 Begi2 is now known as Begi 10:21 spider-mario joined 10:42 iH2O joined 10:48 _nadim joined 10:52 Peter_R joined, nadim left 11:03 virtualsue joined 11:04 csd_ left, virtualsue left 11:05 virtualsue joined, xinming_ left 11:07 vendethiel joined 11:08 Ven joined 11:10 xinming joined 11:12 Ven left
dalek kudo/nom: 0236c43 | lizmat++ | src/core/CurrentThreadScheduler.pm:
Make sure we provide a scope for the CATCH block
 11:13
ast: 377935b | lizmat++ | S17-scheduler/at.t:
Unfudge now passing test
stmuk blogs.msdn.com/b/vcblog/archive/201...ate-1.aspx 11:14
!
jdv79 do ww have auth and ver support yet? 11:15
we
11:17 n0tjack joined, Skarsnik joined 11:21 n0tjack left
_nadim Good morning, two questions, #1 given a little perl code, can I get a dump of how it is parsed in some way? 11:21
gfldex _nadim: perl6 --target=ast -e 'say "foo"' 11:24
_nadim #2 I have a sub S in module M, S uses multis defined in M, in the context in which I call S, I have a bunch of multis in another context;so those are not visible to S, what's the clean way to get my multis in M? 11:25
gfldex: thanks
11:25 iH2O left
Woodi _nadim: objects and methods ? 11:26
11:27 rurban joined
_nadim the multis are filters used by S to display objects. Classes have the possibility to define methods if they want to but the usage I have shall not modify the classes thus the use of multis 11:29
Woodi _nadim: overriding multis when loading M is not cleanest way probably... 11:31
maybe lexically...
or some plugin with specific api? 11:33
_nadim adding multis is what is done, overriding possible but will probably not be used a lot. having a multi per type is the cleanest I believe. and this is needed only when a specific object type should be handled specially, in most cases the ulti for Any just works.
I'd do the API thingy, I believe that all multies should be kept in the dumpers namespace, the question I have is how does one do that. 11:34
Woodi so you have multi sub filter( Int i ) ... and same thing in M ? 11:35
_nadim short of defining them in the M namespace but that look less good than an API call. And I'd rather have the multis exist only during the call to S
Woodi: I have a fixed set of multis in M only. for a new class, I control, I can add a method, but for classes outside my control I want to add a multi for that class, one that would match instead for filter(Any), which is one of M's pre-declared filters 11:37
Woodi _nadim: there are things like .nextsame but I have no idea how they work. you want to override standard control multis or extend them somehow ? 11:39
11:43 _nadim left, _nadim joined 11:49 Myron joined 11:50 virtualsue left
llfourn anyway to disable REPL behaviour in p6 so it works like p5 (read program from STDIN)? 11:52
lizmat perl6 <foo ?? 11:54
llfourn: where foo is the filename ? 11:55
llfourn lizmat: but I don't have a file I have stdin
Woodi echo "say 2" | perl6 works but gives '>'
llfourn perl6 -- # seems to be a thing in --help
Woodi: but that will only work for one line
Woodi hmm
llfourn try printf '{\n}\n'|perl6 11:56
I think -- switch might be broken
try printf '{\n}\n'|perl6 -- # should work
lizmat hmmm... now that we have IO::Handle.t on moar, we should be able to get that to work
Woodi cat scr.pl | perl6 gives: > 1\n> 2\n> *
llfourn Woodi: yep because it's treating it like REPL 11:57
Woodi yea... 11:58
perl6 -- scr.pl works :)
llfourn lizmat: I'll try and see what's going on with -- implementation atm
(assuming it's what I want)
12:00 _mg_ joined 12:02 virtualsue joined
_nadim I can declare two modules in the same file, how do I need/use/require the first one in the second one? if I do nothing then subs in the first module are not visible. if I need it or I use it, I get an error as the modules' file does not exist. if I require it, the subs in the first module are not visible. 12:02
llfourn _nadim: need use and require are for compilation units not modules aka module foo { } 12:04
_nadim: you can try import
_nadim nopaste.linux-dev.org/?876289 12:05
import worked perfectly, thank you
llfourn _nadim: no worries. Actually we need to document import :) 12:06
_nadim it "is" documented here, under the 'use' entry doc.perl6.org/language/modules
but it would be nice with an own entry and an example
llfourn yes I wrote that and atm that is actually a lie 12:07
_nadim hehe
llfourn use is not the same as need + import
use loads and imports from the compilation unit (aka file)
_nadim then we all want to know all about it :)
llfourn import imports from a package
but according to design docs they are sort of meant to be the same
so it's a case of tell the truth about rakudo and lie about perl6 12:08
_nadim Now I'll try my hand at squrelling multis in a package that is not the one I am working with
llfourn: maybe just telling the truth about both is best, we'll all need to live with Rakudo too, not just P6
llfourn _nadim: I agree :) 12:10
12:12 Myron left
dalek ast: ce5872f | lizmat++ | S17-scheduler/basic.t:
Unfudge now passing tests
 12:18
kudo/nom: 519a081 | lizmat++ | src/core/ (2 files):
Move substr internals to Rakudo::Internals
llfourn lizmat,Woodi It's perl6 - # read program from stdin 12:22
( of course ) 12:23
lizmat :-)
still, now that we have IO::Handle.t, the - should probably not be necessary
llfourn++ # for finding out and telling us :-)
llfourn lizmat: oh you mean autodetect? 12:24
lizmat yeah 12:25
llfourn not bad idea unless pipeing a REPL session is a thing 12:26
Woodi llfourn: that realy works :)
llfourn +1 from me :)
probably should have been the first thing I tried 12:27
so what does -- mean then
Woodi lizmat: what you mean ? .t changes something ? ;)
lizmat leave the rest of the parameters for the script? 12:28
llfourn ah
lizmat no, the "t" method on an IO::Handle is True if there is a terminal attached to it
in other words: there is most likely an actual person sitting at the other end
Woodi lizmat: ah, oki :)
12:29 nadim joined 12:32 leont joined 12:33 _nadim left 12:39 risou is now known as risou_awy, risou_awy is now known as risou
llfourn m: say { ... } # is there any explanation as to why ;; is there in an anon block 12:40
camelia rakudo-moar 519a08: OUTPUT«-> ;; $_? is raw { #`(Block|85255656) ... }␤»
lizmat llfourn: that's to make sure the parameters following it are not part of multi-dispatch matching 12:51
llfourn lizmat: how is multi dispatch used with anon blocks?
incase I decide to make it a candidate later?
lizmat well, it's more about a Signature, really
the Block has a signature, and what you see there is the .perl of a Signature 12:52
afaik :-)
llfourn mm I'm just wondering why by default blocks have a ;; as part of their sig when it's unlikely they are going to be used in multi dispatch 12:53
m: say -> $a,$b { }
camelia rakudo-moar 519a08: OUTPUT«-> $a, $b { #`(Block|64754752) ... }␤»
llfourn I'm writing a presentation for Sydney.pm with demoing the fact that blocks are first level objects 12:54
wondering how to answer if someone asks
lizmat m: sub a($a,$b;; $_) { }; say &a.perl 12:55
camelia rakudo-moar 519a08: OUTPUT«sub a ($a, $b;; $_) { #`(Sub|70411880) ... }␤»
lizmat Block isa Code, Code hasa Signature, Block.perl does $!signature.perl 12:56
that's why it shows up like that: Signatures are generic 12:57
dalek kudo/nom: 3cb6ae9 | lizmat++ | src/core/ (4 files):
Introduce Rakudo::Internals.error-rcgye

And make sure all copy-n-paste of RAKUDO_ERROR_COLOR handling is using that.		 12:58
12:59 boegel left 13:00 telex left 13:01 telex joined 13:03 boegel joined
dalek kudo/nom: 79e3e9d | lizmat++ | src/core/Label.pm:
Oops, it's a method, not an our sub
 13:05
llfourn lizmat: just looked at Signature. You are right it is more of a byproduct of block's default signature not having .multi-invocant set for the $_ arg. So nothing too special. Thanks :). 13:08
chansen_ ugexe: is there a more idiomatic way using submethod BUILD rather than using multi method new? (<gist.github.com/chansen/6076c2c2c7...a5fe1>) 13:16
13:17 raiph joined
mspo m: {}.WHAT 13:20
camelia ( no output )
timotimo huh, we still have proto.perl6.org and it forwards to modules.perl6.org
grondilu anybody doing adventofcode.com/ ?
timotimo well, not quite "forwards"
but it gives you the same thing as modules.perl6.org does
chansen_: candidates with their only difference in named parameters are tried in order of appearance in the code file 13:21
mspo timotimo: www.perl6.org seems to handle the rest. I think someone should setup a few redirects :)
timotimo chansen_: so because the "year, day" candidate comes first and matches in every case, it's always the one being chosen 13:22
13:23 abaugher_ left
chansen_ timotimo: Thanks! ugexe pointed that out yesterday. He said I should take a look at submethod BUILD 13:23
mspo github.com/perl6/perl6-examples/bl...xamples.pm ? 13:25
timotimo hm, not sure what he may have meant
mspo put your multimethod inside of BUILD? 13:26
timotimo surely you mean "the decision logic"? 13:28
mspo anyway don't listen to me, I'm struggling to write my first p6 module ;) 13:29
grondilu m: say i.narrow
camelia rakudo-moar 79e3e9: OUTPUT«Attempt to divide 1 by zero using /␤ in block <unit> at /tmp/NtZv0lmiS_:1␤␤Actually thrown at:␤ in block <unit> at /tmp/NtZv0lmiS_:1␤␤»
13:29 kid51 joined
dalek kudo/nom: 65de9c3 | lizmat++ | src/core/Block.pm:
Don't use a var for Block.perl
 13:31
grondilu division by zero error in i.narrow is easy to fix. I'll submit a PR 13:33
lizmat ++grondilu 13:34
13:38 lucasb joined
dalek ast: 47943ae | lizmat++ | S06-other/main-usage.t:
Fix another print streamlining casualty
 13:38
13:40 pecastro left 13:42 pippo joined
pippo o/ #perl6 13:42
mspo is there an IO::Socket for udp?
and unix sockets
lizmat grondilu: wouldn't something like $!re != 0e0 && $!im / $!re ≅ 0e0 ?? $!re.narrow !! self make more sense ? 13:43
grondilu it's the same isn't it? 13:45
timotimo mspo: no :(
moar doesn't have udp sockets yet
grondilu m: say .re !== 0e0 && .im / .re =~= 0e0 ?? .re.narrow !! $_ given i 13:46
camelia rakudo-moar 79e3e9: OUTPUT«0+1i␤»
grondilu m: say .re == 0e0 ?? $_ !! .im / .re =~= 0e0 ?? .re.narrow !! $_ given i
camelia rakudo-moar 79e3e9: OUTPUT«0+1i␤»
lizmat hmmm... interesting data point: 13:47
RAKUDO_MODULE_DEBUG=1 make t/spec/S32-hash/adverbs.t # succeeds, no errors
13:47 _mg_ left
lizmat without the RAKUDO_MODULE_DEBUG it fails 13:47
13:47 virtualsue left 13:49 WizJin left 13:52 WizJin joined
grondilu m: for ^4 { say "hi" } 13:53
camelia rakudo-moar 65de9c: OUTPUT«hi␤hi␤hi␤hi␤»
lizmat m: my $a = 42; my $b ::= $a; $b = 666; say $a # if I understand the spec correct, the assignment to $b should fail
camelia rakudo-moar 65de9c: OUTPUT«666␤»
grondilu m: for ^4 { when 3 { say "three!" }; say "hi" }
camelia rakudo-moar 65de9c: OUTPUT«hi␤hi␤hi␤three!␤»
grondilu m: for ^4 { when 1 { say "three!" }; say "hi" }
camelia rakudo-moar 65de9c: OUTPUT«hi␤three!␤hi␤hi␤»
lizmat perhaps we should make ::= a NYI error, to prevent code in the wild that has wrong assumptions about ::= ? 13:54
grondilu sorry, I thought that would reproduce an unexpected behavior I got
13:54 hankache joined
mspo timotimo: weird 13:57
Zoffix What does "getlex: outer index out of range" error mean?
yoleaux 01:38Z <PerlJam> Zoffix: Sorry I'm late but ... Sure, if people want to do multiple articles per day, I don't see why not. Everyone may need to increase the communication factor a little bit though. And I'm not sure if later-in-the-day articles will hide earlier ones or not, but either way, perhaps for those days that have multiple articles, the main one for the day could just be pointers to the others.
13:58 pecastro joined
grondilu wonders if Complex.narrow should use atan2 13:58
mspo there's no lower level Socket either, it seems
14:02 WizJin left
mspo christmas is soon ;) 14:03
Zoffix Hm, caused by "is cached"
hankache hello #perl6
Zoffix \o
Skarsnik ua is messy >< 14:04
yoleaux 04:35Z <AlexDaniel> Skarsnik: Whoops! “Failed to copy '/home/alex/.panda-work/1449289967_1/lib/mysql/SQLite/StatementHandle.pm6”
14:04 WizJin joined
Skarsnik hm, typo x) 14:04
Zoffix m: sub chars64with (@*chars) is cached { 42 }; say [chars64with('+', '/'), chars64with('-', '_')] 14:05
camelia rakudo-moar 65de9c: OUTPUT«Unhandled exception: Method 'rcgye' not found for invocant of class 'Rakudo::Internals'␤ at <unknown>:1 (/home/camelia/rakudo-m-inst-2/share/perl6/runtime/CORE.setting.moarvm:print_exception:4294967295)␤ from gen/moar/m-CORE.setting:19749 (/home/ca…»
Zoffix m: sub foo (*@chars) is cached { 42 }; say [foo('+', '/'), foo('-', '_')] 14:06
camelia rakudo-moar 65de9c: OUTPUT«[42 42]␤»
Skarsnik how to show the caller in a sub/method?
Zoffix so "is cached" is an issue with the new precomp stuff? 14:09
Skarsnik what cached do? x) 14:10
Zoffix Skarsnik, doc.perl6.org/routine/is%20cached
Skarsnik how it's neat 14:11
*ho
Zoffix yeah
Skarsnik Why do I have to patch everything just to scrap webpages xD
Zoffix :) 14:12
hm "getlex: outer index out of range in any enter at gen/moar/m-Metamodel.nqp:3763" 14:14
lizmat Zoffix: fixing
Zoffix lizmat, is it related to precomp? 14:15
lucasb m: sub f(@*a) {}; f(1); f(2) # Zoffix++
camelia rakudo-moar 65de9c: OUTPUT«Unhandled exception: Method 'rcgye' not found for invocant of class 'Rakudo::Internals'␤ at <unknown>:1 (/home/camelia/rakudo-m-inst-1/share/perl6/runtime/CORE.setting.moarvm:print_exception:4294967295)␤ from gen/moar/m-CORE.setting:19749 (/home/ca…»
lizmat no, something stupid I just did
Skarsnik m: class A { method foo(say "hello"); my $a = A.new; $a.$("foo");
camelia rakudo-moar 65de9c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/SixX6briLB␤Invalid typename 'say' in parameter declaration.␤at /tmp/SixX6briLB:1␤------> 3class A { method foo(say7⏏5 "hello"); my $a = A.new; $a.$("foo");␤»
Zoffix lucasb, it's @*a vs *@a
Skarsnik m: class A { method foo {say "hello"}}; my $a = A.new; $a.$("foo");
camelia rakudo-moar 65de9c: OUTPUT«Cannot find method 'CALL-ME'␤ in block <unit> at /tmp/bKKbq7e6PP:1␤␤»
Skarsnik hm
Zoffix lizmat, aha! :) Someone dissed me on my bug report, saying this was precomp stuff
Skarsnik It's a new trend this days "Does not work -> blame precomp stuff" 14:16
lucasb idk about *@a, but I think that sub f(@*a) is bug, isn't it? I never saw the name 'rcgye' before... dunno what it is 14:17
Zoffix Blame is one thing, but closing a real Issue and saying I should be "more restrictive" is another.
Anyway. I guess that's getting addressed now, so *me goes down through other 30+ github notifications*
14:17 skaji left
dalek kudo/nom: b8122e9 | lizmat++ | src/core/ (3 files):
It's error-rcgye, not just rcgye!
 14:18
14:19 skids joined
Zoffix oh 14:19
Skarsnik how do I call a method by its name?
Zoffix lizmat, i thought you meant the "getlex: outer index out of range in any enter at gen/moar/m-Metamodel.nqp:3763" issue
lizmat ."$foo"
moritz ."$foo"()
Skarsnik Ooh it's so simple?
lizmat moritz: are the parens needed ?
moritz lizmat: I think so
m: say 1."cos" 14:20
camelia rakudo-moar 65de9c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/SbtPMNyPS2␤Quoted method name requires parenthesized arguments. If you meant to concatenate two strings, use '~'.␤at /tmp/SbtPMNyPS2:1␤------> 3say 1."cos"7⏏5<EOL>␤»
moritz m: say 1."cos"()
camelia rakudo-moar 65de9c: OUTPUT«0.54030230586814␤»
lizmat moritz++
14:20 skaji joined
lucasb ah, lizmat++; I didn't saw your previous commit :) 14:21
rcgye == red clear green yellow eject 14:22
14:22 bjz_ left 14:23 bjz joined
[Tux] test 50000 24.180 24.069 14:24
test-t 50000 20.364 20.252
csv-parser 50000 25.847 25.736
lizmat yeah, wanted to implicitely document the return values
14:24 virtualsue joined
hankache is panda working? 14:25
[Tux] yes 14:26
hankache i just updated rakudo and panda using rakudobrew and panda stopped working
Zoffix Mine just lounges on the couch all day... lazy critter..
hankache ===SORRY!=== Could not find Shell::Command:ver<True>:auth<True>:api<True>
Zoffix hankache, rebuild it 14:27
hankache Zoffix ok using rakudob rew?
Zoffix hankache, yeah, rakudobrew build-panda 14:28
hankache Zoffix: tried again, no luck :(
===SORRY!=== 14:29
Could not find Shell::Command:ver<True>:auth<True>:api<True> in:
/home/kmel/.perl6/2015.11-366-g65de9c3
/home/kmel/.rakudobrew/moar-nom/install/share/perl6/site
/home/kmel/.rakudobrew/moar-nom/install/share/perl6/vendor
/home/kmel/.rakudobrew/moar-nom/install/share/perl6
14:29 WizJin left
Zoffix hankache, did it actually build or did it say "Run bootstrap instead of rebootstrap"? 14:29
lizmat hankache: I've heard good stories about nuking the entire installation first
Zoffix Yeah, that's what I've done to fix that: nuke ~/.rakudobrew and ~/.perl6, then install everything from scratch 14:30
hankache Zoffix: first time it said Run bootstrap instead of rebootstrap then i updated rakudobrew and rebuilt panda 14:31
lizmat Zoffix i'll nuke it and see what happens
Skarsnik hm why $h."{$a<name>.lc}"() = ~$a<value>; broke me the whole object. $a<name> is an attribute name. 14:33
stmuk ==> Installing Term::termios
Method 'key' not found for invocant of class 'Any'
arnsholt Is there a way to get the PID of the interpretr running my program? 14:34
lizmat m: say $*PID
camelia rakudo-moar b8122e: OUTPUT«8233␤»
lizmat arnsholt: ^^^ ??
arnsholt Oh, derp. Of course 14:35
I grepped for pid, but didn't think to do grep -i
Cheers!
lizmat you're welcome!
moritz m: use NativeCall; sub getpid() return int64 {*}; say getpid 14:36
camelia rakudo-moar b8122e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Udza2qOJYI␤Missing block␤at /tmp/Udza2qOJYI:1␤------> 3use NativeCall; sub getpid()7⏏5 return int64 {*}; say getpid␤ expecting any of:␤ new name to be defined␤»
moritz m: use NativeCall; sub getpid() returns int64 is native(Str) {*}; say getpid
camelia rakudo-moar b8122e: OUTPUT«8316␤»
arnsholt Heh. That'd work too =D
dalek kudo/nom: 15ea690 | lizmat++ | src/Perl6/Grammar.nqp:
Make ::= a NYI, in light of 6.c

In the hope that forgiveness > permission :-)
14:36 csd_ joined
stmuk ah that's just a missing provides 14:36
lizmat m: use nqp; say nqp::getpid # arnsholt, yet another way 14:37
camelia rakudo-moar b8122e: OUTPUT«8332␤»
nadim is it possible to fiddle in a multi sub in another module? 14:38
14:40 bjz_ joined 14:41 bjz left
lizmat afk for a few hours& 14:47
stmuk github.com/tadzik/panda/pull/261
14:48 rindolf left
Skarsnik hm has $.foo is rw; create a setFoo accessor? 14:49
14:51 FROGGS_ is now known as FROGGS
FROGGS o/ 14:52
yoleaux 4 Dec 2015 22:54Z <nine> FROGGS: I could fix the install-core-dist.pl issue. Sadly it's really the disabling of %.loaded and %.seen that makes my test work, confirming our working theory.
4 Dec 2015 23:05Z <nine> FROGGS: I got a reasonable workaround! Since we may not keep the lexpad either, we just don't. After merging the globals I clean the CompUnit::Handle. We have no further use for it anyway. At least none that I'm aware of.
4 Dec 2015 23:09Z <nine> FROGGS: ok, my simple test works, but I still get the same failure in Inline::Perl5 :(
10:08Z <nine> FROGGS: I worked around the problem by not referencing the CompUnit::Repository directly in Distribution::Resources. I only store its path-spec which I can use to get the object later on (at runtime instead of compile time)
Skarsnik Hi FROGGS x)
moritz Skarsnik: no, it means you can say $obj.foo = 'newvalue'; 14:53
Skarsnik m: class A { has $.a is rw}; my $p = A.new; $p."a"() = 42; say $p.perl;
camelia rakudo-moar 15ea69: OUTPUT«A.new(a => 42)␤»
14:53 xpen joined
FROGGS hi Skarsnik 14:53
Skarsnik I don't get it a : $h."{$a<name>.lc}"() = ~$a<value>; destroy my $h object 14:54
14:54 rindolf joined
Skarsnik and $a<name> only contain attribute name 14:55
moritz what do you mean by "destroy"? 14:56
masak g'day, #perl6
yoleaux 00:19Z <Zoffix> masak: masak++ I've learned something: you can't have $foo-2 :)
masak :)
Skarsnik nvm it work but now http::ua stop adding cookies Oo 14:58
cognominal what is the official status of HLL::Grammar, it contains all sort of useful methods that are not documented anywhere in the specs. 15:00
15:00 dbrock joined, mbrock joined, mbrock left
moritz cognominal: HLL::Grammar is an implementation detail, IMHO 15:02
zengargoyle i was thinking HLL::Grammar is a NQP thing vs a Perl thing.
15:03 xpen left 15:04 xpen joined
zengargoyle in the NQP/Rakudo workstop stuff it's sort of a base used to implement whatever language you wish. like Ruby-ish or PHP-ish in the examples. 15:04
Skarsnik Yay I fixed cookies I think? 15:05
FROGGS cognominal: it is just a helper exposed by NQP to implement high level languages 15:07
virtualsue blows a raspberry at travis 15:08
FROGGS cognominal: so, it does not belong to the design docs or spec tests
stmuk virtualsue: wrong channel :P
virtualsue could be 15:09
nadim moritz: anywhere I could find docus about putting a sub in another module? 15:10
hankache lizmat Zoffix i nuked .perl6 and built moar again. when i try to build panda it gets stuck at ==>Testing JSON::Fast 15:11
any idea why? 15:12
or what i should do?
moritz nadim: no idea; what's your goal? 15:17
zengargoyle hankache: have you done a git-pull on rakudobrew itself recently? 15:20
masak nadim: still having trouble with exporting multis?
jnthn waves from a train
moritz \me waves a train 15:23
hankache zengargoyle i just nuked rakudobrew and cloning it again, maybe that would work 15:24
moritz also confuses solidus and reverse solidus
stmuk ==> Successfully installed JSON::Fast
DrForr hankache: As an experiment I deleted ~/.rakudobrew, that seems to be working better than my installation was last night.
Though Inline::Perl5 still has the same issues. 15:25
.seen nine
yoleaux I saw nine 10:08Z in #perl6: <nine> .tell FROGGS I worked around the problem by not referencing the CompUnit::Repository directly in Distribution::Resources. I only store its path-spec which I can use to get the object later on (at runtime instead of compile time)
zengargoyle i'm starting a rebuild... just think that 'hey, rakudobrew itself needs updating sometimes' isn't particularly well pointed out in the whole use rakudobrew to build thing.
stmuk Task::Star installs now apart from DBIish and JSON::RPC (latter fixed via PR) 15:26
zengargoyle hankache: yah, make it through the build moar, build panda, without any issue on my machine... 15:27
nadim masak: yes, still the same problem :)
hankache DrForr zengargoyle i just deleted ~/.rakudobrew and rebuilding. Lets see what happens 15:28
stmuk is there anyway to introspect adverts like ^methods? 15:31
errr adverbs
zengargoyle docs probably need a: in dire cases after great changes... $ rakudobrew nuke moar; cd ~/.rakudobrew; git pull; rakudobrew build moar; rakudobrew build panda; 15:32
moritz m: class A:ver<42> { }; say A.^ver
camelia rakudo-moar 15ea69: OUTPUT«Method 'gist' not found for invocant of class 'NQPMu'␤ in block <unit> at /tmp/4W2KNITxuy:1␤␤»
jnthn We don't pass on ver and auth to the meat-object yet
zengargoyle and maybe a: rm -rf ~/.perl6 # for good measure.
moritz any volunteers for tomorrow's Perl 6 advent? 15:33
masak nadim: I might have misunderstood your problem. I just tried the simplest thing possible, and it worked: gist.github.com/masak/3b2e4b16482781431b3b 15:35
15:37 mohae joined
zengargoyle hrm, Task::Star bailed on Installing Grammar::Debugger: Cannot unbox a type object 15:37
stmuk zengargoyle: try again it usually works second time
zengargoyle worked second time
:)
stmuk :)
15:38 khw joined 15:39 cygx joined
zengargoyle tho ??? not a good thing to randomly fail then work. 15:39
Bailador: Cannot find method 'run_alt' 15:40
cygx moritz: I probably could write something for the calendar, though my more interesting modules are too much in flux for me to be comfortable writing about them
15:41 rurban left
cygx perhaps something about the as of today still undocument extended open modes 15:41
masak yes, please
cygx (I'm working on the docs: github.com/cygx/doc/commit/d1d5a17...bffa2a735d )
moritz cygx: that would be awesome
15:41 rurban joined, rurban left
moritz cygx: do you already have a login for perl6advent.wordpress.com/wp-admin/ ? 15:42
15:43 kjs_ joined
dalek : 0fe59dd | moritz++ | misc/perl6advent-2015/schedule:
Advent 2015: update schedule

  * cygx++ does the 6th (tomorrow)
  * add topic for yesterday		 15:48
masak thank you.
...for today, actually :)
15:48 AlexDaniel joined
moritz erm, yes 15:48
:-)
masak: in my defense: you wrote it yesterday :-) 15:49
masak ;) 15:50
what's an off-by-one between friends, eh? :P 15:51
moritz masak: I could also use your "I'm a (new) parent, time is meaningless to me" defense :-)
masak: about which I pondered yesterday, fwiw
jnthn If time is now meaningless, will you always be a new parent, and so always be able to use this defense? :) 15:52
moritz masak: because it's quite a bit of a paradox. You see the child(ren) grow up, and they evolve so fast that it should be good clock to adjust by
masak jnthn: I... I don't know! O.O
moritz masak: and yet I have the same feeling of time not meaning so much (on topics not related to children)
masak moritz: I've always hated the phrase "time goes by so quickly", on purely esthetic grounds. 15:53
moritz: but I think what people *mean* is that they didn't notice time pass by. they were distracted by something.
moritz: and that definitely happens when you become a parent.
moritz masak: I guess it's related to $work + @children consuming so much time, that there is little time for other stuff
and thus things happening on the outside in this "other stuff" section seem to be quite fast 15:54
aye
masak right
seems we're vibrating on the same frequency on this one
cygx is back 16:00
moritz: no, not yet
moritz: do I need to setup a wordpress account first? 16:03
16:03 xpen left
moritz cygx: it's enough if /msg me your email address, I believe 16:05
cygx: maybe that'll prompt you to set up a wordress account or so, dunno
AlexDaniel OK, so even though DBIish is now installable with --notests and --force, still it does not work. This error: travis-ci.org/perl6/DBIish#L962 .
stmuk does panda not handle github.com type urls? 16:06
AlexDaniel So it seems like the rumours that you can skip tests on this one are wrong
any idea what could this find_symbol thing be about? 16:07
arnsholt That's a very internal error 16:09
jnthn Well, find_symbol is the method the compiler internals use to find symbols...dunno if it's the same one
jnthn blogged on what he's been working on: 6guts.wordpress.com/2015/12/05/get...christmas/ 16:10
arnsholt From the stack trace it says it's find_symbol in CORE.setting.moarvm, so looks like it
masak jnthn++
jnthn That's an odd plae to do a find__symbol
You may not be in the context of the compiler
hankache jnthn++ 16:11
stmuk my $suplier = Supplier.new; needs an extra p on the LHS
16:12 virtualsue left
jnthn People will figure it out :) 16:13
jnthn had enough of the wordpress editor for one day
16:17 atta left
moritz jnthn: rt.perl.org/Ticket/Display.html?id=126816 I can reproduce the find_symbol error with a single require ::('DBDish::mysql'), whereas "use DBDish::mysql;" works fine 16:18
nadim moritz: I am writing a dumper, it uses multis to dump objects. I want to implement clss specific sumpers. the problem is that the dumper is in a cls (with its own filters), and it is used from another name space; filter (multi sub) that are defined in my script are not visible in the dumper module. I want to make those filter art of the dumper namspace. 16:19
cygx jnthn: I did something about open(:bin) - github.com/cygx/rakudo/commit/3290...795e1479d4 16:20
moritz nadim: one option is to use methods, of course
cygx jnthn: not quite sure if it's agood idea, though ;)
moritz nadim: the other is to use the CLIENT:: pseudo namespace to look for the multis
jnthn As long as you encode it using utf8-c8 16:21
yes, you do
It'll do
nadim moritz: I am using methods of course. people ho want to add the specific method are welcome, and I can do that on my own classes. But the whole idea is that one can decide at run time that class XX should be dumped in a specific way without having to fiddle with that class.
cygx jnthn: it's part of github.com/rakudo/rakudo/pull/619 16:22
nadim best would be to be able to scope all that of course so that the filters are not part of the dumper after the call to the dumper
moritz: where can I find documentation, alt examples, for the CLIENT:: pseudo namespace? 16:23
moritz jnthn++ # blog post, xmas fixes
16:24 atta joined
AlexDaniel jnthn: “GitHub doesn't let us configure per-branch permissions, so we can't prevent pushes to master that way.” – are you sure? I thought that it was a new shiny feature 16:25
jnthn Maybe it is, but I couldn't find anything about it 16:27
I don't really feel a need to do it anyway though
AlexDaniel jnthn: hm, indeed. I can only find this: help.github.com/articles/configuri...-branches/ 16:32
cygx afk for a bit, then I'll get started on that advent posting thing
AlexDaniel and this: github.com/blog/2051-protected-bra...tus-checks
ah 16:36
there is a comment about it, whoops
nadim Anyone else knows about CLIENT:: and can point me somewhere? 16:41
16:42 khisanth_ joined 16:44 Khisanth left, Fleurety joined
JimmyZ jnthn: $supply.emit('oh$supply.emit('hai # I can't understatnd it ... 16:46
16:46 dolmen joined
jnthn wtf 16:47
flussence looks fine to me...
dolmen The DateTime API for Perl 6 is definitely *NOT READY* for a 1.0 release: its handling of timezone is just naive
dalek kudo/nom: 494e456 | jnthn++ | src/core/Channel.pm:
Give Channel its own Supply impelmentation.

Which does not block. We can optimize this some, but for now get it to work. This means you can use react/supply blocks and whenever more than one Channel to multiplex them without the busy-waiting that can happen with earliest in such situations.		 16:48
dolmen doc.perl6.org/type/DateTime#method_timezone
jnthn Looks fine to me too...
dolmen A TimeZone class is missing in the API
16:49 webstrand joined
jnthn We're not going to do timezones in the cor language 16:49
And should not.
flussence dolmen: is your specific complaint that we have full ISO8601 support in the core, or that DateTime::Timezone isn't? :)
jnthn They change too fast.
JimmyZ jnthn: both windows and ubuntu the same to me :(
jnthn You should be able to upgrade it at a different rate than your Perl
JimmyZ firefox 16:50
and ie ...
dolmen jnthn: Having only UTC and Local timezones in the core language would be ok for me. My point is that I don't see a provision for extension for support of the whole IANA timezone 16:52
16:52 bzipitidoo left 16:53 domidumont joined
dolmen jnthn: in Perl 5, ZEFRAM has made modules that can load timezones from the database managed by the operating system. This is the approach that workaround the main issue with timezone: keeping the database up to date. metacpan.org/pod/DateTime::TimeZone::Tzfile 16:57
flussence so, the same as what we're already doing? github.com/supernovus/perl6-timezone
pippo m: my $supplier = Supplier.new; my $supply = $supplier.Supply; $supply.tap({ .say }); $supply.emit('oh'); 16:58
camelia rakudo-moar 15ea69: OUTPUT«Method 'emit' not found for invocant of class 'Supply'␤ in block <unit> at /tmp/62pGCAoMva:1␤␤»
hankache deleted ~/.rakudobrew and ~/.perl6 and started fresh. While building panda it still gets stuck at ==> Testing JSON::Fast 16:59
jnthn oh those second two should be supplier 17:00
*sigh*
hankache any ideas what i should do?
pippo jnthn: :-)
jnthn Fixed it
Did the typo too 17:01
lucasb hankache: there is a single test in JSON::Fast that is hanging. You can comment the line and it'll install fine
gfldex El_Che: IIRC you asked for strictness of named parameters for new. This is how it's done: gist.github.com/gfldex/c6c81d00a06d19e95651
webstrand I've been trying to package rakudo-star for Archlinux, but from what I can tell, it's Configure.pl uses the destination prefix both as a writable destination to place binaries and to lookup those same binaries to use. If I choose the prefix to be some writable directory (not /usr) and then move the files into /usr, the binaries still attempt to search for libraries in the original prefix. I've only a limited knowledge of perl6, am I missing 17:02
something that would fix this issue?
dolmen flussence: no, lib/DateTime/TimeZone/Zone/*.pm are generated from the IANA db. Which implies you have to upgrade the module every time the IANA db is updated, while your operating system also updates it. Two copies of the IANA db.
hankache lucasb where should i find my local copy of the test file? 17:03
ugexe 3rd copy of iana db is a charm
flussence webstrand: try --sysroot
lucasb panda/ext/JSON__Fast/t/01-parse.t: line 153: # '{ "a : false }', # missing quote
hankache: ^^ comment that line 17:04
dalek kudo/precomp-singleprocess: 299adc5 | arnsholt++ | src/core/CompUnit (2 files):
No need to use binding; use assignment instead.

This fixes the circular loading error of the previous commit version, but uncovers a new error.
kudo/precomp-singleprocess: 5252cd8 | arnsholt++ | src/core/CompUnit/PrecompilationRepository.pm:
Use spurt instead, to ensure the .deps file exists.

The previous variant didn't close properly, so use spurt() instead.
webstrand flussence: Thanks. no idea how I missed that. 17:05
flussence webstrand: I did all that blind stumbling around while writing an ebuild the other day, so you don't have to! :D 17:06
webstrand flussence: I searched for an up-to-date ebuild to cannibalize, where can I find yours?
flussence it's not ready for real world use yet, things compile but they explode at runtime 17:07
(afaik, all of the official ebuilds have similar problems) 17:08
ugexe if a module gets installed, but that module loads/depends on a module that is not installed (but available via -I../OtherModule/lib) will the .precomp of the installed module be tied to the uninstalled OtherModule? 17:09
17:10 CIAvash left
cognominal thx FROGGS for you answer about HLL::Parser 17:11
Skarsnik hm http::ua seem to not work with the new rakudo
colomon ==> Installing Path::Iterator
X::IO::Copy.new(from => "/home/colomon/emmentaler/.panda-work/1449335442_1/lib/Path/Iterator.pm6", to => q|/home/colomon/.rakudobrew/moar-nom/install/share/perl6/site/sources/9AB877291EBF0EEC80ABC4B1A9605CBC0EF63801|.IO(:SPEC(IO::Spec::Unix)), os-error => "Failed to copy file: no such file or directory")
Skarsnik t/070-ua-simple.t .......... ===SORRY!===
Cannot find method 'run_alt'
17:12 dolmen left
arnsholt So those commits there should give us precompilation in a single process 17:13
Not tested beyond installing the core dist, though
hankache lucasb thanks for the hint. panda built successfully. shouldn't we fix the test file?
ugexe youve saved countless man hours on r-j win32 im imagine
may even make smoke testing on r-j rpi bearable again 17:15
lucasb hankache: cool :) I think timotimo++ will fix his module later 17:16
hankache lucasb thanks again :)
flussence ...r-j on a pi‽ I barely have the patience for it on a desktop!
ugexe if you're lucky... getting it to install without running out of memory is seemingly random 17:17
17:18 Begi left, khisanth_ is now known as Khisanth 17:19 km3 left 17:24 znpy left
AlexDaniel jnthn: what kinda makes me sad is that it looks like qx// thing is going into the release as is. And it's ok if some helper subs will be added later (“shellout” and friends, once appropriate names are found), but the idea is that qx// should probably use ｢run｣ underneath (well, as suggested by Larry), which definitely is not going to be backward compatible… Sooo, some amount of pain is expected aftewards 17:29
speaking of which, by the way, I've looked into the panda sources and found things like this one everywhere: github.com/tadzik/panda/blob/maste...her.pm#L46 17:31
17:31 kjs_ left
AlexDaniel I'm not going to raise remote code execution flag, since panda runs a lot of stuff anyway… but it might be worth fixing 17:31
17:38 rurban joined 17:40 new joined
jnthn AlexDaniel: afaik, Larry still is going to decide on that 17:46
hankache Perl 6 for the web. What do you recommend Bailador or Web::App ? 17:49
17:50 ZoffixW joined
AlexDaniel hankache: well, I'm using Bailador and it works 17:51
ZoffixW Bailador was broken for me ~1-2 days ago (it was looking for psgi.input, but it was actually p6sgi.input) 17:52
17:52 thejawn joined
Skarsnik AlexDaniel, can you make a small tutorial or something about it? I am confused on how that work x) 17:52
ZoffixW Anyone wanna help me debug my grammar? gist.github.com/zoffixznet/4cff770d3b81980c236f As you can see from the trace output, the string seems to match completely, but the TOP fails for some reason. Any ideas?
AlexDaniel Skarsnik: take a look at examples: github.com/tadzik/Bailador/tree/master/examples 17:53
ZoffixW jnthn++ Grammar::Debugger is awesome
skids ZoffixW: you sure \r\n is working like you think? 17:54
17:54 RabidGravy joined
ZoffixW skids, no idea. How can I test that? 17:54
darutoko is there an editors list that support syntax highlight?
Skarsnik AlexDaniel, yes but how you setup thix?
ZoffixW aha
Skarsnik darutoko, vim/atom
ZoffixW skids, if I remove \r\n from both the regex and the string, then match is successfull :/ 17:55
AlexDaniel darutoko: emacs with perl6-mode works too
Skarsnik kate has a kind of working p6 syntax mode, but it's pretty imcomplete
skids ZoffixW: Yeah so \n matches the system newline now I think.
17:55 robnox joined
darutoko Skarsnik, AlexDaniel, I gues its NO :) 17:55
skids ZoffixW: use \x0a if you mean that. 17:56
Skarsnik Ah you mean a list
AlexDaniel Skarsnik: well, just run any of these examples
Skarsnik I don't know if the doc has a list somewhere
AlexDaniel Skarsnik: it will start and the link will be printed to stdout. Just open it and you will see something
Skarsnik hm, you set nothing with your http server? 17:57
AlexDaniel yes
17:57 new left
jdv79 i think i have internet again 17:57
ZoffixW skids++ thanks. Chaning \r\n in the regex to just plain \n seems to have done the trick.
AlexDaniel Skarsnik: getting this thing to work on a real server is going to be a bit more complicated… as any other stuff is, like ruby on rails… pain! 17:58
ZoffixW Skarsnik, you could reverse proxy, if you wanted to
Skarsnik jdv79, I fixed these damn cookie (I think)
jdv79 nice
so anyone want to point me in the right direction of building a cpan installer?
ZoffixW likes P5's App::cpanminus 17:59
Skarsnik AlexDaniel, well that the intersting part x)
jdv79 i guess i'll just can panda til i can't take it anymore
s/can/hack/
Skarsnik Does zef still work?
jdv79 ugexe: what say yee on the topic of cpan and zef? 18:00
iirc there was a refactor or something affot
*afoo
t
robnox I tried ruby on rails once, I really didn't like it -- I don't see what all the hype is about 18:01
ZoffixW Is it still hyped up though?
I thought phoenix/elixir is where the cool kids are at 18:02
robnox yeah, was looking for local contract work earlier and at least half the jobs wanted ruby on rails.. lol
18:02 ZoffixW left
Skarsnik I choose to stay on october rakudo/panda for now x) 18:03
jdv79 i'm building latest now
darutoko Skarsnik, do I need to get some thing for atom or it can from the box? 18:04
Skarsnik perl6advent.wordpress.com/2015/12/...r-support/
jdv79 56s parse stage - not bad
ugexe jdv79: it will be possible in a couple of days. i'm working on abstracting the behavior between having the entire package list vs querying something like metacpan for an identity
jdv79 this is fantastic
darutoko Skarsnik, thank you 18:05
jdv79 oh right you have a service thingee
well, we have somewhat lame indices at cpan now
that metacpan6 is trying to use
didn't quite finish that part yet
ugexe i tried looking at the examples from your blog, but i think i was confused at what was showing up at the time 18:06
AlexDaniel robnox: what I still don't understand is how to make it work on a real server *correctly*. Everyone suggests Phusion Passenger but I refuse to use crippled down software. So right now using reverse proxy thing, but it does not sound like the best solution… 18:07
jnthn .tell lizmat Think the ::= patch has some spectest fallout to fix
yoleaux jnthn: I'll pass your message to lizmat.
jdv79 ugexe: what exactly?
there are no real examples 18:08
i think
ugexe it was json of like 20 items
none of which i recognized
lizmat computer, messages 18:09
yoleaux 18:07Z <jnthn> lizmat: Think the ::= patch has some spectest fallout to fix
jdv79 huh. i just looked and don't see anyting liek that
lizmat jnthn: yes, it does, they're basically all faulty tests
that were actually passing for the wrong reason 18:10
jdv79 since there's maybe 2 valid dists on cpan it makes sense though:)
robnox AlexDaniel: best solution is to uninstall ruby and load up perl ;)
lizmat m: my $a ::= 42; $a++ # fails because of MMD failure of finding a candidate that dpesn't have "is rw"
camelia rakudo-moar 494e45: OUTPUT«5===SORRY!5=== Error while compiling /tmp/szpk4i9ZM8␤"::=" not yet implemented. Sorry. ␤at /tmp/szpk4i9ZM8:1␤------> 3my $a ::= 427⏏5; $a++ # fails because of MMD failure ␤»
lizmat *failed
jdv79 ugexe: www.cpan.org/authors/p6*
lizmat jnthn: but are you otherwise ok with the patch ? 18:11
ugexe jdv79: i dont remember exactly but it looks like im recalling the 2 links to hack.p6c.org:500\d
jdv79 tomorrow i plan to dig into panda and rakduo though for the first time
18:11 rurban left
jdv79 i think hack rebooted since then and they aren't setup again yet 18:11
ugexe do you have an example of how to make a query or possibly get all perl6 package listings?
jdv79 that's just a metacpan instance over faux-dists generated from the ecosystem 18:12
Skarsnik Zoffix, did you reach http::ua yet in your quest to install everything?
18:12 rurban joined
jnthn lizmat: I think so 18:12
jdv79 yeah, there are 2 ok indices on cpan. p6dists and p6provides but they probably need help
ugexe so i take it there will be a perl6 flag or something? 18:13
jdv79 the metapcan instance will be lying right now cause its faked out for prototyping
not tracking cpan
metacpan6 will be a seperate instance from metacpan
jnthn lizmat: I'm just spectesting some patches that fix our ability to use whenever properly with channels and eliminating the last uses of earliest in core setting. 18:14
jdv79 and at cpan its just that dir structure
lizmat jnthn++ :-)
ugexe a question: if i fetch metadata from metacpan but its source-url points to github, which is the correct default selection to download from?
jnthn wonders whether we want a few days deprecation grace or just to rip it out :)
jdv79 source-url is not to spec
i tjhink its support.source im META6.json 18:15
lizmat jnthn: at this stage, I would say: rip it out
jdv79 source-url is a panda/META.info/eco sthing
*thing
ugexe its documented there is a source-type and source-url im pretty sure 18:16
skids jnthn: oh that reminds me, do you want an RT for .head not filing its done, or will that just likely fall by the wayside in other fixes?
jdv79 plus you won't download using that. you download the dist file.
jnthn skids: Please RT
It's a separate issue to anything I've been doing now
jdv79 design.perl6.org/S22.html#support
Skarsnik m: sprintf("%s"); 18:17
camelia rakudo-moar 494e45: OUTPUT«Directives specify 1 argument, but no argument was supplied␤␤»
Skarsnik Should I make a tiket for this?
dalek kudo/nom: 3865bdd | jnthn++ | src/core/Channel.pm:
Fix a race in Channel.Supply; update .list.
kudo/nom: b45ef16 | jnthn++ | src/core/Supply.pm:
Re-implement Supply.list without Channel.
kudo/nom: 87fed0e | jnthn++ | src/core/Promise.pm:
Promise.Supply is an on-demand Supply.

Otherwise it will be racey.
Skarsnik m: sprintf("%s%"); 18:18
camelia rakudo-moar 494e45: OUTPUT«'s%' is not valid in sprintf format sequence '%s%'␤␤»
ugexe jdv79: ah, indeed you are correct. i was imagining ecosystem/spec.pod was a rakudo doc (which is where source-type and source-url are mentioned) 18:19
lizmat jnthn: I'm pretty sure the ripping out will be mentioned in the P6W :-)
dalek ast: 5eb2477 | lizmat++ | S0 (3 files):
Fix fallout of making ::= NYI
jdv79 looks like metacpan kicks back a download_url field that cpanm uses 18:20
i imagine that's what you may be after in te end
not ready yet:)
soon
ugexe im actually glad... it was a pain to implement sane defaults when i considered it spec
dalek ast: 0b9c6c6 | jnthn++ | S17-supply/syntax.t:
Test using whenever with channels.
 18:22
Skarsnik thx jdv79 for merging x) 18:24
ugexe you mean RabidGravy ... ? 18:25
RabidGravy I just merged something in HTTP::UserAgent from you :) 18:26
Skarsnik Oh sorry for the mistake x)
RabidGravy but Skarsnik++ # had been meaning to put the debug in myself 18:27
Skarsnik I forget to document the debug thing ~~
RabidGravy oh well
dalek ast: 0003213 | jnthn++ | S17-supply/syntax.t:
Test multiple whenevers on channels too.

Since that is what we advise for multiplexing channels now, instead of earliest.
jdv79 sweet 18:28
18:29 ZoffixW joined
jdv79 Skarsnik: you fixed h:ua to do a loging? 18:29
i've been meaning to do that since ovid squawked about it 18:30
ZoffixW Writing character classes like this is not correct, is it? My regex seems to be matching against literal "n" <-[\ \a\0\r\n,]>
Or rather, my grammar (my token)
Skarsnik Not really, it display on stdout what is send/recv, if you do UA.new(:debug)
could probably be on stderr, I don't know 18:31
jdv79 m: say so /<[\n]>/ for "n", "\n"
camelia rakudo-moar 494e45: OUTPUT«Resource temporarily unavailable»
jdv79 huh?
WUT
that worked fine locally
is camelia ill? 18:32
18:32 adhoc left
ZoffixW m: say so /<[\n]>/ for "n", "\n" 18:32
camelia rakudo-moar 87fed0: OUTPUT«False␤True␤»
ZoffixW weird
jdv79 meh
[Tux] rakudobrew:
lizmat jnthn: any reason why Supply.list doesn't use an Iterator ?
jnthn Well, did say temporarily :)
[Tux] Permission denied (publickey).
fatal: Could not read from remote repository.
jnthn lizmat: Harder to implement correctly 18:33
lizmat jnthn: also, is it correct that each queue gets its own ConcQueue ?
ZoffixW m: gist.github.com/zoffixznet/9f64995f16faa6fc7925 18:34
camelia rakudo-moar 87fed0: OUTPUT«False␤»
jnthn lizmat: And it's an unlikely thing to be on a hot path...
ZoffixW ^^ how come that code says "True" if I uncomment Grammar::Tracer?
jnthn sync/async transitions are by definition bad for performance; the gather/take is the least of the worries :)
lizmat: "gets its own"?
lizmat each gather does a ConcQueue.CREATE ? 18:35
18:35 corbyhaas left
jnthn doesn't quite get what's being asked 18:35
lizmat aaahhhh
I read it wrong
the type is used as the sentinel
jnthn Yes, it's correct
Ah, ok :)
(Correct because we tap the Supply each time, so it needs its own queue instance too)
[Tux] note to self: do not run two make again commands on the same folder simultaneous
lizmat jnthn: what if the Supply is live and queue gets empty before the supply is done? 18:37
jnthn: shouldn't it block on an empty queue ?
18:38 corbyhaas joined
jnthn lizmat: It does block on an empty queue 18:39
nqp::shift on the queue blocks
lizmat ah, so that is hidden in the repr(ConcBlockingQueue)
jnthn omg I'm in Norway 18:40
:)
lizmat is it Tuesday ?
bartolin_ fwiw, I created ticket RT #126823 for the failing tests in S32-array/adverbs.t and S32-hash/adverbs.t (and I'm going to fudge both tests)
jnthn lizmat: Yes
18:40 hankache left 18:41 herby_ joined
herby_ Hello, everyone! 18:41
[Tux] o/
lizmat herby_ o/
FROGGS good evening 18:42
18:43 rurban left
dalek kudo/nom: 733529c | jnthn++ | t/spectest.data:
Don't run earliest.t; it's going away.
 18:43
ZoffixW Here's a shorter version. If I comment out "use Grammar::Debugger", the match fails... and I've no idea why :/ 18:44
gist.github.com/zoffixznet/629e13e0214cc6e3e2e2
lizmat FROGGS o/ 18:45
dalek ast: eddca86 | jnthn++ | S17-channel/earliest.t:
Remove earliest.t; this feature is going away.
herby_ Trying to learn a little Perl 6 and embrace the holiday cheer at the same time with Advent of Code :) 18:46
adventofcode.com/day/5
18:46 hankache joined, lsm-desktop left
herby_ Trying to stumble my way through that problem, which is in Regex land 18:46
FROGGS lizmat / nine: btw, I want to extend the CU repositories by a bin-dir attribute, so you can create a CUR with a bin-dir other than <prefix>/bin 18:47
lizmat FROGGS: and that would be shared by all versions of a distro ? 18:48
FROGGS lizmat: what do you mean by 'distro' here?
[Tux] test 50000 23.534 23.417
test-t 50000 16.555 16.438
csv-parser 50000 26.023 25.906
lizmat [Tux]++
[Tux] posting now as I note that the diff between test and test-t increases 18:49
and because 16.555 is a really good figure
lizmat FROGGS: never mind, I guess I must put on my thinking cap before saying anytihng 18:50
[Tux] hands lizmat a bivakmuts
lizmat distro is not applicable in this case anyway, right ?
RabidGravy it's saturday why think more than necessary
lizmat
.oO( I think I get that :-)
 18:51
FROGGS lizmat: I just know that I need *one* bin dir for star, and that the repo where the star dists are installed in a unfortunate bin directory (i.e. not install/bin)
is a*
lizmat FROGGS: no objections from me :-) 18:52
FROGGS k :o)
[Tux] lizmat, care to have a second pair of eyes at gist.github.com/Tux/afec03490da258d48671 to see where I go wrong in extending hoelzro's IO::String with «IO::String.new ("foo");» ? 18:55
ZoffixW (╯°□°）╯︵ ┻━┻
I give up.
lizmat Tux: isn't nl-in an IO::Handle attribute? 18:56
[Tux]: also, method print-nl is referencing .nl, not .nl-in 18:57
dalek ast: 9828048 | jnthn++ | integration/advent2013-day14.t:
Re-work test to use react/whenever, not earliest.
lizmat jnthn: re ::= , couldn't we implement this for now by codegenning the LHS of ::= with a Proxy ? 19:00
dalek kudo/nom: 539b5f4 | jnthn++ | src/ (3 files):
Remvoe earliest syntax.

The react/whenever syntax should be used instead; it now works well on channels, and doesn't have the busy waiting problem when multiplexing two channels.		 19:01
jnthn lizmat: I don't think so; I think it's meant to mean "binding a la signatures"
Rather than simply "readonly"
lizmat yeah, but if we bind it *and* put the result of that in a Proxy ? 19:02
m: my $a = 42; my $b := proxy-ro($a); sub proxy-ro(\x) { Proxy.new( FETCH => { x }, STORE => -> $a, $b { die() }) }; $b = 42
camelia rakudo-moar 733529: OUTPUT«Died␤ in block <unit> at /tmp/v6jylhqU9l:1␤␤»
lizmat m: my $a = 42; my $b := proxy-ro($a); sub proxy-ro(\x) { Proxy.new( FETCH => { x }, STORE => -> $a, $b { die() }) }; $a = 666; say $b
camelia rakudo-moar 733529: OUTPUT«666␤»
jnthn But tht isn't really the semantics we want 19:03
lizmat ok, then we leave it for post 6.c 19:04
jnthn my @a ::= ...something that gives a Seq... # should do what a signature bind does also, and .cache it
lizmat ah, of course, ::= is not just working on scalars
duh
ok
19:04 ChoHag left 19:05 ChoHag joined
jnthn hopes the earliest => react thing won't be too bad on the ecosystem 19:07
herby_ regex question... how would I check if a string had any letter appearing twice in a row? 19:09
jaala would be true cause "aa" 19:10
ZoffixW m: say so "azfdfsd" ~~ /(.)$0/
camelia rakudo-moar 733529: OUTPUT«False␤»
ZoffixW m: say so "aazfdfsd" ~~ /(.)$0/
camelia rakudo-moar 733529: OUTPUT«True␤»
ugexe wonders if there is a way around precompile happening twice when first testing Module A followed by installing Module A
herby_ hmmm
zoffix, so that is say capture any letter, then see if that capture follows that letter? 19:11
regex gives me a headache :)
ZoffixW herby_, yeah
herby_ great, thanks!
jnthn ugexe: I suspect not since we test in the filesystem, but install to an installation
herby_ m: say so "xxxaxxxexxxi" ~~ / <[aeiou]> ** 3 / 19:14
camelia rakudo-moar 539b5f: OUTPUT«False␤»
dalek kudo/precomp-singleprocess: eb3f580 | arnsholt++ | src/core/CompUnit/PrecompilationRepository.pm:
HLL::Compiler.evalfiles is more appropriate than command_eval.

The latter does exception handling and calls exit(1) if an exception is caught. Using evalfiles lets the exception propagate into the module loader, which lets us handle exceptions during precomp.
Currently it'll just propagate into the same error handler a level up, but this avoids weird exception behaviour in the future, should we want to handle some exceptions.
herby_ hmmm
herby_ m: say so "xxxaxxxexxxi" ~~ / <[aeiou]> /
camelia rakudo-moar 539b5f: OUTPUT«True␤»
herby_ m: say so "xxxaxxxexxxi" ~~ / <[aeiou]> + /
camelia rakudo-moar 539b5f: OUTPUT«True␤»
jnthn ** 3 implies "3 in a row" 19:15
herby_ hmmm. how would i check for a total of 3 vowels?
or, at least 3 19:16
lizmat jnthn: getting a SIGABRT on syntax.t now
ok 50 - react/whenever with supply that immediately emits values works
Abort trap: 6
ZoffixW m: say so "xxxaxxxexxxi" ~~ / [<[aeiou]>.*?]**3 /
camelia rakudo-moar 539b5f: OUTPUT«True␤»
ZoffixW m: say so "xxxaxxxexxx" ~~ / [<[aeiou]>.*?]**3 /
camelia rakudo-moar 539b5f: OUTPUT«False␤»
jnthn m: say so ("xxxaxxxexxxi" ~~ m:g/ <[aeiou]> /).elems >= 3
camelia rakudo-moar 539b5f: OUTPUT«True␤»
herby_ you smart people blow my mind :)
jnthn lizmat: OSX-specific bug then.
lizmat afraid so 19:17
jnthn If nobody else gets to it in the next days, I'll ask for a shell account :) 19:18
lizmat ok... 19:19
jnthn (Will be doing the years last bits of teaching/travel)
ZoffixW m: grammar Foo { token TOP { <foo> }; token foo { .+ } }; class Actions { method TOP ($/) { $/.make: foo => $<foo>}}; say Foo.parse('foobar', :actions(Actions)) 19:21
camelia rakudo-moar 539b5f: OUTPUT«Too few positionals passed; expected 2 arguments but got 1␤ in method TOP at /tmp/MWQBlthA0a:1␤ in regex TOP at /tmp/MWQBlthA0a:1␤ in block <unit> at /tmp/MWQBlthA0a:1␤␤»
ZoffixW Why is it telling me this?
jnthn foo => is a named arg
19:21 skids left
ZoffixW m: grammar Foo { token TOP { <foo> }; token foo { .+ } }; class Actions { method TOP ($/) { $/.make: {foo => 'bar'}}}; say Foo.parse('foobar', :actions(Actions)) 19:21
camelia rakudo-moar 539b5f: OUTPUT«｢foobar｣␤ foo => ｢foobar｣␤»
ZoffixW jnthn++ thanks
bartolin_ fwiw I also see failures on S17-supply/syntax.t on FreeBSD 10.2: "Abort trap (core dumped)" after 'ok 29 - Multiple whenevers run concurrently' 19:22
"This is rakudo version 2015.11-359-g5e49279 built on MoarVM version 2015.11-34-gc3eea17"
ZoffixW jnthn, is there a way to escape from whenever {} ?
jnthn, like react { whenever { next if $not_thing; do_this_when_thing; } } 19:23
jnthn ZoffixW: done or last, depending on whether you want to escape from the whenever or the entire react (or supply) block
next shoud work to...though maybe NYI
ZoffixW Ah
Thanks, jnthn++
jnthn yeah, NYI... 19:24
Can fix it soon, but bit tired now and train should be getting to Oslo soon anyways...
bartolin_ (S17-supply/syntax.t) there are aborted tests on linux too: github.com/coke/perl6-roast-data/b....out#L4900
jnthn Ah 19:25
So it's really "not Windows" that hsa issues
19:26 quester joined
lizmat looks like it then 19:26
jnthn Well, that's easier; I'm more familiar with the Linux debugging toolchain than the OSX one :)
lizmat and then there's hack, right ?
jnthn Yes, though I suspect I can reproduce it on my own Linux box 19:27
RabidGravy I think we need another 32 modules in the ecosystem before xmas
500 would be a nice round number 19:28
Skarsnik I already on one!
jnthn If we reject the tau PR, that'll be an easy extra one :P
cygx when are the advent postings supposed to go live?
RabidGravy I've got one maybe two I should finish
ZoffixW I'm definitely finishing 1 more
RabidGravy jnthn++ # lateral thinking
Skarsnik at midnight for your country time I think?
ZoffixW Not "your country time". Mine went at like 8pm my time 19:29
arnsholt jnthn: How long are you in Oslo?
ZoffixW cygx, if you go to "Blogs" and look at your post, it should say "Posting in X time" (if you've scheduled it)
And I scheduled mine at 00:00 19:30
cygx ZoffixW: Haven't even finished writing it ;)
jnthn arnsholt: Approx 14 hours :)
cygx the rambling part is done, the meat is still missing
jnthn arnsholt: Flying up to Trondheim tomorrow midday or so
ZoffixW cygx, you can still schedule it :) 19:31
jnthn Not until he's written the meat... :) 19:32
herby_ m: say so "xax" ~~ / (.)<-[$0]>$0 /
camelia rakudo-moar 539b5f: OUTPUT«True␤»
herby_ m: say so "xafx" ~~ / (.)<-[$0]>$0 / 19:33
camelia rakudo-moar 539b5f: OUTPUT«False␤»
arnsholt jnthn: Oh, right. No need to ask about $beer, in that case =)
jnthn arnsholt: Yeah, 'fraid not
jnthn woulda said if he was going to be around for a bit
arnsholt jnthn: Oh, and the OS X debugging toolchain (at least for me when I'm working on OS X) is basically the same as on Linux: printf and gdb 19:34
jnthn :D
arnsholt Yeah, I figured. But I've been a bit absent lately, so figured I'd check in case I'd missed something
Better safe than sorry, and all that
=)
jnthn :)
Indeed
colomon lldb instead of gdb on os x these days
at least for me.
and lots and lots of printfs. ;) 19:35
robnox I just started using llvm recently, seems legit
jnthn I probably want helgrind for this one, tbh :)
ugexe jnthn: darn, i was hoping something like an external compunit that did FileSystem but staged in ::Installation without updating the sha1 -> path lookup table (and deleting the files if install fails or updating the table if install succeed)
colomon oh, latest valgrind seems to work well on 10.10.5, it’s a nice improvement over the situation a couple of years ago. 19:36
19:36 dbrock left
jnthn ugexe: Maybe some day, but let's make it work before being clever :) 19:36
19:36 domidumont left
herby_ thanks for the help, I'll be back! 19:36
19:36 herby_ left
jnthn colomon: Ah, nice. Last time I wanted it on OSX then the box I had acces to had latest OSX and valgrind wasn't updated for it yet. 19:37
so I had to use malloc_history and friends :) 19:38
colomon yeah, I did a happy dance when I realized valgrind worked properly on my OS X box again. 19:40
19:40 yqt joined 19:41 Begi joined
RabidGravy phew I put Supplies before Channels in the concurrency document, must have been a magical foresight so I didn't have to explain react twice ;-) 19:44
jnthn :)
ZoffixW Well... I spent ~3 hours on this IRC parsing grammar and I'm still getting bugs I can't reproduce in a simple test case -_- 19:45
ZoffixW lights the whole thing on fire and starts again from scratch
moritz btw I'm doing the first few tasks of adventofcode.com/ in Perl 6 right now
ZoffixW I may have been over-excited with how easy it is to convert BNF to grammars :P
DrForr ZoffixW: Yeah, there are some quirks to it. 19:46
(he says, going back to trying to create a Prolog grammar) 19:47
There's an ABNF grammar for ANTLR, have you tried converting that? (he says, knowing there's a tool to help) 19:48
moritz (actually I did the first one in shell, just because I could :-))
ZoffixW :) 19:49
DrForr, not idea what neither ABNF nor ANTLR are, so I think I'll try starting from scratch and including only the RFC components I care about instead of blindly copy-pasting BNF :) 19:50
DrForr ANTLR is effectively the new lex/yacc, if you know what those tools are.
ZoffixW shakes head 19:51
negatory
DrForr Parser generation tools. 19:52
Skarsnik IRC grammar is not stupidly easy? like :sender cmd ...stuff... 19:53
ZoffixW Skarsnik, essentially, yes: twitter.com/zoffix/status/673183128756854784 19:54
geekosaur at a gross level
ugexe then its well now i need a grammar for ipv4, and ipv6, and probably something needs to parse time stamps, and... 19:56
19:57 eyck joined
ugexe no longer so simple 19:58
bartolin_ the test file integration/advent2014-day05.t (run as stresstest) aborts, because it uses the old Supply API (see perl6advent.wordpress.com/2014/12/05/) 20:02
should the test file be removed from spectest.data?
20:06 lsm-desktop joined
nadim masak: I answered in gist.github.com/masak/3b2e4b16482781431b3b (#comment formated as FAT!) 20:07
I can hack it in 5 lines with a filter list that I can match with the bject type but it is ugly and I would rather see the addition of the filter like a trait added at run time to an object. 20:10
20:14 FROGGS left
ZoffixW laughs hysterically 20:14
20:14 adhoc joined
ZoffixW So, I remade everything, and back against this stupid bug. Does NOT parse, unless I "use Grammar::Tracer" -_- 20:14
m: gist.github.com/zoffixznet/306e95627aae4fa4487e
camelia rakudo-moar 539b5f: OUTPUT«Nil␤»
ZoffixW (╯°□°）╯︵ ┻━┻
DrForr ZoffixW: FWIW I'm rerunning my test suite under the latest rakudo and finding out it's broken. 20:15
20:15 nadim left
[Tux] lizmat, If I don't include nl-in as attribute, it barfs when setting it later 20:15
20:15 nadim joined
[Tux] .nl is "stolen" from hoelzro's IO::String 20:16
[Tux] doesn't use it (I think)
ZoffixW I just don't get how using a module affects whether my grammar succeeeds
ugexe ZoffixW: tools that evaluate code can change the behavior sometimes... 20:17
m: my $x = { :a(1) }; say $x.perl
camelia rakudo-moar 539b5f: OUTPUT«${:a(1)}␤»
ugexe m: use trace; my $x = { :a(1) }; say $x.perl
camelia rakudo-moar 539b5f: OUTPUT«2 (/tmp/t7tsKPGQW8:1)␤my $x = { :a(1) }␤7 (/tmp/t7tsKPGQW8:1)␤say $x.perl␤-> ;; $_? is raw { #`(Block|65770936) ... }␤»
dalek c: 04779c1 | RabidGravy++ | doc/Language/concurrency.pod:
Remove the earliest examples and replace with react
ZoffixW I notice that if I remove the portion after <servername> here gist.github.com/zoffixznet/306e956...rser-p6-L7 or remove both "!" and "@" from here gist.github.com/zoffixznet/306e956...ser-p6-L11 THEN it works :/ 20:21
ugexe what if you change the |s to ||s? 20:22
20:23 nadim left
ZoffixW ugexe, it works... What does || do? 20:23
20:23 nadim joined
ugexe its the same as the perl6 operator 20:24
| will attempt all the submatches at once and take the longest one or something
ZoffixW ugexe++ thanks.
( docs.perl6.org/language/regexes#Alternation )
AlexDaniel m: say (‘°□°）’ ~ “┻━┻”)
camelia rakudo-moar 539b5f: OUTPUT«°□°）┻━┻␤»
20:25 ollej joined 20:26 darutoko left
AlexDaniel m: sub prefix:<(╯°□°）╯︵>($e) { die $e }; (╯°□°）╯︵ “oops, something went wrong” 20:29
camelia rakudo-moar 539b5f: OUTPUT«oops, something went wrong␤ in sub prefix:<(╯°□°）╯︵> at /tmp/ModkpjVz4Z:1␤ in block <unit> at /tmp/ModkpjVz4Z:1␤␤»
geekosaur now you just need something to render the lettering upside down :p 20:30
ZoffixW :D 20:32
AlexDaniel geekosaur: so that you can write ｢(╯°□°）╯︵ ‘ƃuoɹʍ ʇuǝʍ ƃuıɥʇǝɯos 'sdoo’｣? Yeah, that would be cool 20:33
however, I find it weird that to make it work as you expect you have to create an operator
I wish it was a sub 20:34
but this means that you have to use term:<…> thingy, which changes the grammar or something
which adds up to the startup time… hm, but new operator does the same thing, isn't it? 20:35
ZoffixW Yeah
AlexDaniel yeah, right. Ok then it could as well be a sub 20:36
20:40 hankache left
ZoffixW Oh, man 20:40
m: sub prefix:<(╯°□°）╯︵ ┻━┻> {}
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/PBrZQj4ZLI␤Too many symbols provided for categorical of type prefix; needs only 1␤at /tmp/PBrZQj4ZLI:1␤------> 3sub prefix:<(╯°□°）╯︵ ┻━┻>7⏏5 {}␤»
ZoffixW There's a limit on the number of chars? :(
And that error is weird too 20:41
"needs only 1" part
lizmat ZoffixW: I think you forgot a > for the prefic
*prefix
m: sub prefix:<(╯°□°）>╯︵ ┻━┻> {}
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/rWXe5eID3v␤Missing block␤at /tmp/rWXe5eID3v:1␤------> 3sub prefix:<(╯°□°）>7⏏5╯︵ ┻━┻> {}␤ expecting any of:␤ new name to be defined␤»
lizmat m: sub prefix:<(╯°□°）> ╯︵ ┻━┻> {}
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/PeGmbhRVe7␤Missing block␤at /tmp/PeGmbhRVe7:1␤------> 3sub prefix:<(╯°□°）>7⏏5 ╯︵ ┻━┻> {}␤ expecting any of:␤ new name to be defined␤»
lizmat ZoffixW: hmmm... perhaps not 20:42
m: sub prefix:<(╯°□° ╯︵ ┻━┻）> {}
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ZQ3q9AGc7g␤Too many symbols provided for categorical of type prefix; needs only 1␤at /tmp/ZQ3q9AGc7g:1␤------> 3sub prefix:<(╯°□° ╯︵ ┻━┻）>7⏏5 {}␤»
ZoffixW The cause seems to be the space before the table 20:43
m: sub prefix:<(╯°□°）╯︵┻━┻> {}
camelia ( no output )
ZoffixW m: sub prefix:<(╯°□°）╯︵┻━┻xxxxxxxxxxxxxxxxxx> {}
camelia ( no output )
ZoffixW .u
yoleaux Search for a Unicode character by codepoint, name, or raw character
ZoffixW :(
m: say " ".uniname
camelia rakudo-moar 539b5f: OUTPUT«SPACE␤»
ZoffixW Oh well, right, why would there be a space in an operator! :) 20:44
AlexDaniel ZoffixW: sounds LTA anyway
I mean the error message
ZoffixW Yeah, the error is weird
m: sub prefix:<[ ]> {}
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VUI4GciK7Z␤Too many symbols provided for categorical of type prefix; needs only 1␤at /tmp/VUI4GciK7Z:1␤------> 3sub prefix:<[ ]>7⏏5 {}␤»
ZoffixW m: sub prefix:< > {} 20:45
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/pCLWGPBpZb␤Null operator is not allowed␤at /tmp/pCLWGPBpZb:1␤------> 3sub prefix:< >7⏏5 {}␤»
ZoffixW m: sub prefix:<. > {}
camelia ( no output )
ZoffixW Ah, I see, the <> are a white-space-separated-list
dalek c: effb68d | lizmat++ | doc/Language/faq.pod:
Answer 'Why should I learn Perl 6?' better

This is basically a copy of the list of reasons as posted on the Perl 6 FaceBook group		 20:46
ZoffixW m: sub prefix:("(╯°□°）╯︵ ┻━┻") ($e) {}
camelia ( no output )
ZoffixW m: sub prefix:("(╯°□°）╯︵ ┻━┻") ($e) { say $e }; (╯°□°）╯︵ ┻━┻ "I give up"
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ia_QyX6JqM␤Bogus statement␤at /tmp/ia_QyX6JqM:1␤------> 3x:("(╯°□°）╯︵ ┻━┻") ($e) { say $e }; (7⏏5╯°□°）╯︵ ┻━┻ "I give up"␤ expecting any of:␤ pref…»
geekosaur I think you can't use "(" as part of an operator... 20:47
ZoffixW Yeah, you can 20:48
m: sub prefix:<(╯°□°）╯︵┻━┻> ($e) { say $e }; (╯°□°）╯︵┻━┻ "I give up"
camelia rakudo-moar 539b5f: OUTPUT«I give up␤»
ZoffixW Ehehe
m: sub term:<(╯°□°）╯︵┻━┻> { run 'shutdown -h NOW' }; (╯°□°）╯︵┻━┻
camelia rakudo-moar 539b5f: OUTPUT«run is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub run at src/RESTRICTED.setting:14␤ in sub term:<(╯°□°）╯︵┻━┻> at /tmp/_0kSEXYE7u:1␤ in block <unit> at /tmp/_0kSEXYE7u:1␤␤»
ZoffixW ^^ something like that could be useful :P 20:49
(FWIW, I knew run was disallowed and I wasn't attempting to kill camelia :P)
This deserves a module.... 20:52
AlexDaniel yeah, your run arg is wrong too
｢run «shutdown -h now»｣ should work though 20:53
jnthn -h gets help, right? :)
20:53 rurban joined
AlexDaniel only if you think that halting your system is helpful 20:54
jnthn ;)
ZoffixW :D
jnthn It's a wonderful bit of command design :)
lucasb ZoffixW: I don't fully understand grammars, but I tried your snippet and like ugexe said, changing | to || only in the prefix token, it works: token prefix { <servername> || ... } 20:58
flussence ZoffixW: I see you've joined the "why the hell does Grammar::Debugger change this parse outcome" club :D 20:59
ZoffixW flussence, hah :D
lucasb, yeah, it did. And after a few more tweaks I managed to get it to parse all of my sample input correctly: gist.github.com/zoffixznet/cbf1d50b472ef998e6b4
21:04 thundergnat joined 21:05 quester left
thundergnat m: say (5.123456789+3.987654321i).round(.01); # Hmmm why yes, yes they do... 21:06
camelia rakudo-moar 539b5f: OUTPUT«Ambiguous call to 'round'; these signatures all match:␤:(Complex:D $: Real $scale, *%_)␤:(Complex:D $: Real $scale, *%_)␤ in block <unit> at /tmp/EeLadBVvYw:1␤␤»
thundergnat ^ Bug? or am I missing something?
moritz m: say (5.123456789+3.987654321i).round 21:07
camelia rakudo-moar 539b5f: OUTPUT«5+4i␤»
moritz thundergnat: bug
thundergnat 'k
robnox hmm.. are you guys really typing all those \u's or do I need to recompile my irc client with unicode support.. lol 21:08
TimToady the latter 21:09
thundergnat m: say (5+5i).rand;
camelia rakudo-moar 539b5f: OUTPUT«Can not convert 5+5i to Num: imaginary part not zero␤ in block <unit> at /tmp/aFs6AcsJFq:1␤␤Actually thrown at:␤ in block <unit> at /tmp/aFs6AcsJFq:1␤␤»
thundergnat :(
m: use MONKEY-TYPING; augment class Complex { method rand () { [+] self.reals».rand Z* 1,i } } say (5+5i).rand;
camelia rakudo-moar 539b5f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/O5_NHQhia5␤Strange text after block (missing semicolon or comma?)␤at /tmp/O5_NHQhia5:1␤------> 3and () { [+] self.reals».rand Z* 1,i } }7⏏5 say (5+5i).rand;␤ expecting any of:␤ infix…»
thundergnat m: use MONKEY-TYPING; augment class Complex { method rand () { [+] self.reals».rand Z* 1,i } }; say (5+5i).rand;
camelia rakudo-moar 539b5f: OUTPUT«3.60462065430806+1.82499575714017i␤»
[Coke] will deprecated handle v6.d style versions instead of dates going forward? 21:10
21:11 z8 joined 21:13 thejawn left 21:16 robnox left, lucasb left 21:17 raiph left
dalek osystem: 6564197 | (Zoffix Znet)++ | META.list:
Add Acme::GiveUp to eco

  (╯°□°）╯︵┻━┻ when you just want to give up for the day: github.com/zoffixznet/perl6-Acme-GiveUp		 21:18
ZoffixW Ahh... now *that's* the way to spend the day :)
21:19 z8 left
masak :D 21:20
dalek kudo/nom: 75ca9dc | (Stefan Seifert)++ | src/ (2 files):
Avoid CompUnits carrying permantent references to VM Contexts

VM Contexts are non-serializable by design, so they may not be referenced from compile time data lest they break on precompilation. So we get rid of them as soon as we don't need them anymore.
kudo/nom: 6c0f93c | (Stefan Seifert)++ | src/ (5 files):
Implement %?RESOURCES

Allows for e.g. Inline::Perl5 to find its p5helper.so both when loaded from a file system directory and when loading an installed version by just accessing
  %?RESOURCES<p5helper.so>
Since library loaders and other operating system facilities may need file extensions to accept those files, we keep the extensions while using a SHA hash as file name to avoid collisions.
21:21 rindolf left
flussence nine++ 21:21
timotimo ^- i like this 21:22
ZoffixW nine++
RabidGravy, ^ seems there's now a way to fix your @*INC-using modules
jnthn nine++ 21:24
Skarsnik I love this module 21:26
timotimo found that the terminal he uses wasn't buggy, they just put "mark text to put it into cut buffer" as an optional feature that's off by default
maximum facepalm
flussence whaaaaaaaaaaaat
who makes that *optional*?!
masak Windows. 21:27
flussence well I can understand it there...
21:27 ZoffixW left
timotimo MATE Terminal. 21:27
flussence but *there*!
timotimo i'm using it instead of gnome terminal which had been good enough before that
until they got rid of something ... i think see-through background?
AlexDaniel Zoffix: “You can use and distribute this module under the same terms as Perl 6 itself.”
Zoffix: what does it mean?
timotimo AlexDaniel: artistic license 2 21:28
geekosaur ??? ctrl-shift-c works fine for me...
timotimo oh ... hmm. should probabl yread "rakudo itself" rather than "perl 6 itself"
AlexDaniel Zoffix: sure, but it says “Perl 6 itself”. What is it?
geekosaur auto to clipboard I dislike, random stuff ends up wiping out something I had selected deliberately
flussence geekosaur: that's why there's more than one clipboard :)
timotimo i use the cut buffer for things i want to transfer from one second to another
Skarsnik Zoffix, make me want to figure is there is a way to shutdown (save) a vm from a guest with virtualbox xD 21:29
geekosaur people still use cut buffers.
...
"I still use DOS"
AlexDaniel timotimo: if it starts saying “rakudo” then why is it limited to just one implementation 21:30
nadim if roles are applied on "proto objects", does that mean that roles can also be applied to "normal" objects thus augmenting the type of an object at compile/run time? in that case how does one do that?
timotimo AlexDaniel: because rakudo has a license and this sentence is only related to a license 21:31
tbh i don't even know where that was quoted from
masak nadim: do you mean "type objects"?
timotimo geekosaur: my opinion is actually that every time i have to use windows i get utterly frustrated that middle click doesn't paste what i selected elsewhere
masak nadim: they haven't been called "proto objects" for something like five years... 21:32
21:32 kaare_ left
masak I only remember that term because I'm an old geezer, in Internet years 21:32
AlexDaniel timotimo: it just makes me wonder why not use artistic license 2 explicitly
flussence to me, a UI that requires ^C-^V is like using a car that requires double-declutch 21:33
Skarsnik because you have to include it I think
AlexDaniel timotimo: e.g. what happens if 10 years from now rakudo dies and another implementation takes over
timotimo flussence: what is "double-declutch"?
AlexDaniel timotimo: not a big issue though
nadim masak: yes I meant type object when I said "proto object"
timotimo AlexDaniel: yeah, true. why not put artistic license 2? i don't know. i'd do that.
AlexDaniel timotimo: it was from here: github.com/zoffixznet/perl6-Acme-GiveUp 21:34
timotimo nadim: you apply a role to an instance with "does" or "but". the first one will mutate an object in-place, the other will give you a copy with the role mixed in
nadim timotimo: great, any place where it is documented?
flussence timotimo: like normal manual transmission, but you have to let off the clutch pedal in neutral, rev-match, then shift to the next gear. 21:35
nadim masak: the problem is that googling for P6 brings back docs from the past ;)
timotimo nadim: "infix does" and "infix but" on doc.perl6.org is only a little snippet of text 21:36
nadim masak: did you look at my answer to you?
timotimo doc.perl6.org/language/objects#Roles
doc.perl6.org/routine/... <- 404 :( 21:37
cygx and done with the advent article
flussence timotimo: doc.perl6.org/routine/.. is worse
cygx next on the schedule: eat something, re-read the post and get it into wrodpress
timotimo hahahaha
flussence it should probably be written routine/&infix:<..>, it's never going to work the normal way 21:39
AlexDaniel Same with doc.perl6.org/routine/.
timotimo yeah 21:40
AlexDaniel and a bunch of other stuff like .::
or even .=
I thought that it was a known issue
timotimo is .:: a thing?
AlexDaniel well, it is listed: doc.perl6.org/language/operators#postfix_.%3A%3A 21:41
nadim I can try but it is faster to ask, can a role contain a multi sub? and in that case, am I right when I think that I can add a multi sub to an object by using 'does' on it? 21:42
[Coke] I would definitely be explicit about the license you mean instead of mentioning a different sw project as holding your license. 21:43
21:48 xfix joined
masak nine: yes, old docs are a problem 21:48
(not just for Perl 6)
timotimo where did that bite us this time?
masak nadim: found your answer in the backlog. looking now. 21:49
timotimo: nadim found an old do that said "proto object" instead of "type object"
timotimo wha ... seriously?
where the hell?
nadim timotimo: it was in old presentation not in the docs
timotimo hm, okay
nadim :) 21:50
21:50 kjs_ joined
masak nadim: this doesn't immediately sound to me like a problem multis ought to be a solution to 21:50
nadim: I think you want the module to have some kind of hooking/registration mechanism
nadim: where you can give it a type and a callback 21:51
nadim: I don't know offhand if you can do that with multis somehow... maybe. but I almost wouldn't expect it to work. 21:52
multis add behavior in *your* scope, not in someone else's.
timotimo i expect that'd require augment rather than does/but 21:53
masak right.
and I wouldn't recommend `augment` as a solution :) it's brittle
timotimo yes
DrForr Hrm. Can someone recommend a clean way to write $/<foo>>>.ast, since it seems to no longer return a list of matched trees? 21:58
Or maybe I've uncovered a bug.
timotimo @<foo>>>.ast?
DrForr That's almost what I'm after... 22:00
22:04 leat joined
DrForr Seems there's just one layer of dereferencing I need to get rid of somewhere... 22:07
dalek c: 96be575 | RabidGravy++ | doc/Language/concurrency.pod:
spaces not tabs
 22:08
c: b1f31c2 | RabidGravy++ | doc/Language/faq.pod:
Merge branch 'master' of github.com:perl6/doc
c/qualified-names: ea23fcd | (Brock Wilcox)++ | doc/Language/ (3 files):
Move into a dedicated Packages doc
nadim masak: I disagree about the design of the interface, multis do just what a lookup does, re-implementing it sounds illogical. But this is still not fixed, who knows, maybe I completely agree with you in a few days. right now I like the idea of using 'does' or maybe inheritance. 22:11
masak :)
nadim: I'm also not 100% set in my opinion. that was just my initial reaction at seeing the problem you were solving. 22:12
nadim masak, on the other hand, and although it is bad design, I would like to know how to inject suff in other classes/namespaces. I also noted that method 'can' just takes a name, it would be nice to know how to give it a name and types to see if it can that too.
masak: as they say in France, only idiots never change their opinion, being a bit open to other solutions is always a sign of intelligence :) 22:13
masak :)
the notion of injecting stuff in other namespaces is... what's the term? "risky"? "tightly coupled"? 22:14
of course, it all depends.
lizmat nadim: do you realize you can add methods to object at runtime ?
nine Oh my, anyone else seeing a massive spec test regression?
lizmat m: my $a = 42 but role { method zippo { "zippo: {self}" } }; say $a; say $a.zippo # nadim 22:15
camelia rakudo-moar 6c0f93: OUTPUT«42␤zippo: 42␤»
lizmat massive as in how many tests ? 22:16
lizmat builds and spectests
22:16 espadrine joined
nadim lizmat: Yes, got the info 20 mn before your post :) I am trying that right now. 22:16
flussence nine: I see a few S02's crashing...
nine lizmat: as in 65 files
lizmat no, haven't seen that yet 22:17
22:17 brabo left
jnthn o.O 22:17
lizmat but I don't think I tested after your commits
22:17 brabo joined
nine Undeclared symbols diag ok skip 22:17
nadim although I am stubborn and trying to use multi sub but since it makes no difference, I'll use the methods
flussence 25% done with spectest, seen S02 and a few S17 explode so far but nothing surprising
jnthn nine: Sounds like an import oops... 22:18
nine May be 75ca9dcbb495763a2fb9e9f4829b9ba885db2a8e.
flussence I'm using TEST_JOBS=8, would that affect it?
22:18 rob joined, rob is now known as robnox
lizmat flussence: no 22:19
yup, looks like import is borked
flussence oh, quite a few explosions in S06
robnox ok someone please post some unicode.. I want to see if I fixed it lol
flussence ‽
jnthn RabidGravy: The third code example in doc.perl6.org/language/concurrency#Channels seems mis-formatted (ended up in many code boxes) 22:20
robnox darn, still seeing \u.. prob gotta tweak a setting now brb
lizmat ę
22:20 robnox left, Skarsnik left
lizmat nine: looks like import fails in precomped files? 22:21
===SORRY!=== Error while compiling /Users/liz/Github/rakudo.moar/t/spec/packages/Test/Util.pm
Undeclared routines:
diag used at lines 59, 67, 70
masak that "ę" is the codfish of vowels.
RabidGravy jnthn, I saw that pushed one with spaces instead of tabs
jnthn ah, cool, then guess it's just behind
RabidGravy++
m: say uniname "ę" 22:22
camelia rakudo-moar 6c0f93: OUTPUT«LATIN SMALL LETTER E WITH OGONEK␤»
22:22 rurban1 joined
flussence these failure patterns are weird; "test returned 1" in short bursts, grouped close together 22:22
lizmat flussence: they're the test files with "use Test::Util" I will bet 22:23
dalek kudo/nom: 88d93f8 | (Stefan Seifert)++ | src/ (2 files):
Revert "Avoid CompUnits carrying permantent references to VM Contexts"

This reverts commit 75ca9dcbb495763a2fb9e9f4829b9ba885db2a8e which seems to break import in precompiled modules.
nine Down to 16 failing files.
Lots of S17 failures in there
22:24 robnox joined
lizmat is building and spectesting again 22:24
nine What is "cas" supposed to be?
22:24 Begi2 joined
flussence and t/spec/integration/advent2013-day14.t hung... 22:24
masak .oO( min commit e cas )
flussence updates and tests also
22:25 rurban left
lizmat nine: en.wikipedia.org/wiki/Compare-and-swap 22:25
22:26 Begi left
nine Appears in the failing S17 tests but nowhere else? 22:26
nadim m: my $a = 42; $a does role { method zippo { "zippo: {self}" } }; say $a; say $a.zippo 22:27
camelia rakudo-moar 6c0f93: OUTPUT«42␤zippo: 42␤»
22:28 Begi2 left
nine ack '\bcas\b' in my rakudo directory yields only test S17 tests and MoarVM .h files 22:29
flussence panda bootstrap.pl sounds unhappy right now: «Use of uninitialized value %ENV...» but it seems to have installed anyway. 22:30
jnthn 'night, #perl6
masak 'night, jnthn 22:31
nine I know I shouldn't push stuff when actually spending an evening with my girlfriend :/ But people have been wasting time with workarounds all week...
cygx It Is Done: perl6advent.wordpress.com/?p=3176
flussence nine: spectest looks a lot happier now 22:32
nine Ah, seems like jnthn++ removed cas and some spec tests need updating.
flussence: btw. I'm really not sure if we can centralize precomp directories
22:32 pecastro_ joined 22:33 pecastro left
flussence I'm not sure either. asking people to install perl6 to user-writable directories isn't gonna fly though. 22:33
dalek c: 383c032 | RabidGravy++ | doc/Language/concurrency.pod:
Really rid the tabs this time
 22:34
nine flussence: we definitely need some patches to make failures to write non-fatal. If we cannot precomp, we should still continue, just slower.
dalek kudo/nom: 8b77688 | lizmat++ | src/core/asyncops.pm:
Restore cas()
 22:35
masak cygx++ # nice! I enjoyed it :) 22:36
lizmat nine flussence : I'm not sure jnthn removed cas() intentionally, so I restored it
at least for now 22:37
PerlJam cygx++ I'm only here for a few minutes and your post is very nice
Zoffix AlexDaniel, it's an artifact of me copying over my P5 module boilerplate and changing P5 to P6 in the license field. Thanks for spotting it. I'll change it to the proper name of the license. 22:38
AlexDaniel :)
flussence I'm half-wondering if it'd be saner, at this point, to just store precomp data in sqlite and let that worry about the concurrent access stuff.
cygx masak, PerlJam: anything you'd like me to change? 22:39
there's still ime, but not a lot ;)
*time
lizmat flussence: that is an option, that's why it's just modules
and in some installations, you could only have precomped files, *without* source 22:40
or load the precomp on the fly from the cloud :-)
PerlJam cygx: There's one sentence that didn't make sense to me: Note that providing :r,:w is semantically different from the combination of :r and :w as existing files will not be truncated.
AlexDaniel OK I have some progress with DBIish
PerlJam cygx: was one of those supposed to be :rw ?
cygx PerlJam: :r,:w behaves identically to :rw 22:41
but in contrast to :w, it doesn't truncate
so :r,:w != :r+:w, so to speak
PerlJam ah
btw, it's too bad you didn't use :e for exclusive, then we could have an open mode of :ew :) 22:42
AlexDaniel “Cannot find method 'find_symbol'” this error does not mean anything, but if we look at when it happens: ｢require ::($module);｣. This basically means that it will be expanded to something like ｢require DBDish::Pg;｣. Now, it is pretty simple, if you try to *require* it then it fails, but if you try to *use* it then there is no problem
timotimo i was about to point that same thing out PerlJam just put here
lizmat nine: confirm spectest is as clean as it was before your revert and the cas() revert 22:43
AlexDaniel in fact, I've changed it to “use DBDish::Pg;” and it seems like the rest of the DBIish is working
Zoffix cygx++ nice article. This line reads weird tho: "and binary files do currently not return buffers instead". should it be "do not currently"?
jdv79 whaat article? 22:44
AlexDaniel So if I got it right, then the question is “why you can ｢use｣ it but can't ｢require｣ it?”
jdv79 oh, there's new one by jnthn...
s//a/
Zoffix jdv79, the scheduled Advent post 22:45
nine lizmat: great :) 22:46
lizmat: 7 test files failing now which is about average 22:47
PerlJam Is there some known weirdness with panda (or rakudobrew)? I just tried "rakudobrew build-panda" and it seems to be hanging on "==> Testing JSON::Fast"
timotimo yes, there is a weirdness
colomon I’ve been getting hangs in panda for a month or two now?
timotimo JSON::Fast uses nqp::eqat to find the closing " for a quoted string
colomon seem kind of random, and usually if I crtl-C all the work has been done.
timotimo recent moarvm doesn't throw an exception when eqat goes beyond the end of string 22:48
Zoffix nine, BTW, someone mentioned this bug was due to precomp stuff. Not sure if you know of it/care: rt.perl.org/Ticket/Display.html?id=126818
timotimo i'll have to bite the bullet and put bounds checks in everywhere :(
nqp-m: say nqp::graphs("hi") 22:50
camelia nqp-moarvm: OUTPUT«Confused at line 2, near "say nqp::g"␤ at gen/moar/stage2/NQPHLL.nqp:521 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/NQPHLL.moarvm:panic:105)␤ from gen/moar/stage2/NQP.nqp:921 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/nqp.moarvm:comp_unit:872)␤ from …»
timotimo nqp-m: say(nqp::graphs("hi"))
camelia nqp-moarvm: OUTPUT«No registered operation handler for 'graphs'␤ at gen/moar/stage2/QAST.nqp:1576 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/QAST.moarvm:compile_op:64)␤ from gen/moar/stage2/QAST.nqp:5488 (/home/camelia/rakudo-m-inst-1/share/nqp/lib/QAST.moarvm:compile_node…»
dalek kudo/nom: fdec627 | lizmat++ | src/core/Complex.pm:
Remove bogus coercion, fixes RT #126825
timotimo which one do i use in userland nqp?
nine Zoffix: I seem to remember jnthn mentioning "is cached" being an ongoing source of problems?
timotimo nqp-m: say(nqp::graphs_s("hi")) 22:51
camelia nqp-moarvm: OUTPUT«No registered operation handler for 'graphs_s'␤ at gen/moar/stage2/QAST.nqp:1576 (/home/camelia/rakudo-m-inst-2/share/nqp/lib/QAST.moarvm:compile_op:64)␤ from gen/moar/stage2/QAST.nqp:5488 (/home/camelia/rakudo-m-inst-2/share/nqp/lib/QAST.moarvm:compile_no…»
timotimo hum.
cygx I've tweaked some sentences to address the raised issues 22:52
PerlJam cygx++
timotimo i suppose nqp::chars
PerlJam cygx: Maybe you could mention slurp and spurt at the end somewhere too? (I mean, that's why many people open files in the first place) 22:54
AlexDaniel surprisingly the error only happens with DBIish files, you can ｢require｣ any other module just fine 22:56
AlexDaniel is talking to himself 22:57
RabidGravy nine, that %?RESOURCES thing does it still rely on changes to the installer?
nine RabidGravy: what do you mean?
RabidGravy e.g. panda or what ever? to read the resources from the META.info and so on 22:58
nine I already pushed a change to panda
RabidGravy ah okay, you rock I'm being slow 22:59
nine "resources" : [ "p5helper.so" ],
timotimo panda shouldn't hang any more when running JSON::Fast tests.
please update your panda to the latest revision
lizmat good night, #perl6! 23:00
nine Good night! 23:01
RabidGravy toodlepip
PerlJam g'night lizmat; sleep well!
timotimo gnite lizmat
23:02 molecules joined, molecules is now known as Guest99713
AlexDaniel if I'm getting a bunch of “at …/runtime/CORE.setting.moarvm:1” errors, how can I make it meaningful? 23:04
nadim private multi methods are not supported, by rakudo or perl6? 23:06
23:06 xfix left
dalek line-Perl5: 03e8ac1 | (Stefan Seifert)++ | / (4 files):
Use new %?RESOURCES to locate p5helper.so

Fixes finding p5helper.so when installed into a CompUnit::Repository::Installation and brings us much closer to an installable version again.		 23:08
nine RabidGravy: ^^^ should give you a hint on how to utilize %?RESOURCES
RabidGravy cool
nine Off to bed now. Good night!
PerlJam good night nine! 23:09
masak good nine, night 23:10
AlexDaniel in other words, is there any way to get the line number if the error happens in nqp or CORE.setting?
TimToady it would be nice if we had the option to install a cockpit recorder that would minimally record line numbers as they are executed 23:16
RabidGravy nadim, correct 23:17
jdv79 TimToady: i was actually looking at that spot in the code (add_categorical) when i attempted to gain some understanding
at least i wasn't off in left field 23:18
masak nadim: there's no theoretical reason they couldn't exist, so they're simply not-yet-implemented in Rakudo.
nadim: possible workaround is to delegate in a single-entry-point private method to multi subs.
23:20 Zoffix left
AlexDaniel it seems like I can skip ｢require｣ and use ｢use｣ instead, though that would be more code 23:23
TimToady there's no particular reason you can't just use subs instead of private methods 23:24
23:24 TEttinger left
TimToady since private methods are really just subs with a different call syntax 23:24
well, as long asyou aren't planning to trust someone else 23:25
23:27 herby_ joined
herby_ Good evening, everyone! 23:27
PerlJam hello herby
23:28 bpmedley joined
dalek ast: c02bf4b | thundergnat++ | S32-num/rounders.t:
Add tests for RT 126825
 23:29
masak herby_: ahoy! 23:35
herby_ o/ 23:37
23:37 espadrine left
herby_ m: say "herby".comb(/\w/); 23:39
camelia rakudo-moar fdec62: OUTPUT«(h e r b y)␤»
herby_ say "h2erb3y".comb(/\w/); 23:40
m: say "h2erb3y".comb(/\w/);
camelia rakudo-moar fdec62: OUTPUT«(h 2 e r b 3 y)␤»
herby_ m: say "h2erb3y".comb(/\S/);
camelia rakudo-moar fdec62: OUTPUT«(h 2 e r b 3 y)␤»
herby_ m: say "h2erb3y".comb(/\s/); 23:42
camelia rakudo-moar fdec62: OUTPUT«()␤»
herby_ m: say "h2erb3y".comb(/\d/);
camelia rakudo-moar fdec62: OUTPUT«(2 3)␤»
herby_ I really like this irc interpreter, if thats what its called :)
m: say "h2erb3y".comb(/\w**2/); 23:43
camelia rakudo-moar fdec62: OUTPUT«(h2 er b3)␤»
timotimo m: say "h2erb3y".comb(2)
camelia rakudo-moar fdec62: OUTPUT«(h2 er b3 y)␤»
herby_ ahhh 23:44
timotimo that version is 1) much faster, 2) gives you a rest at the end if there is one
herby_ m: say"aaabbbccc".comb(3);
camelia rakudo-moar fdec62: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/Pt7_G4Bvjf:1␤------> 3say7⏏5"aaabbbccc".comb(3);␤Two terms in a row␤at /tmp/Pt7_G4Bvjf:1␤------> 3say7⏏5"aaabbbccc".comb(3);␤ expecting any of:␤ infix…»
herby_ m: say "aaabbbccc".comb(3);
camelia rakudo-moar fdec62: OUTPUT«(aaa bbb ccc)␤»
herby_ sweet, learn something new every day
23:45 spider-mario left
timotimo :) 23:45
herby_ m: split("", "herby").say 23:46
camelia rakudo-moar fdec62: OUTPUT«( h e r b y )␤»
herby_ m: split(2, "herby").say
camelia rakudo-moar fdec62: OUTPUT«(herby)␤»
timotimo there's no 2 in herb@
herby_ i was fishing a little there :) 23:47
timotimo m: say split(2, "her2by2").perl
camelia rakudo-moar fdec62: OUTPUT«("her", "by", "")␤»
jdv79 what is the precision of numeric ops?
timotimo numeric as in working on Num type?
23:47 kjs_ left
jdv79 m: say 0.99999999999999999999999 < 1 # why this? 23:48
camelia rakudo-moar fdec62: OUTPUT«False␤»
timotimo m: say 0.99999999999999999999999.perl
camelia rakudo-moar fdec62: OUTPUT«<99999999999999999999999/99999999999999991611392>␤»
timotimo because our rats are broken?
jdv79 so it is supposed to be actually arbitary and that's a bug?
*arbitrary 23:49
timotimo it must be we're using a wrong type somewhere in parsing that number
Guest99713 m: <0.99999999999999999999999999999999999999999999999999999999999> < 1 23:50
camelia rakudo-moar fdec62: OUTPUT«WARNINGS:␤Useless use of "<" in expression "<0.99999999999999999999999999999999999999999999999999999999999> < 1" in sink context (line 1)␤»
Guest99713 say <0.99999999999999999999999999999999999999999999999999999999999> < 1
herby_ m: say so 42.list.elems
camelia rakudo-moar fdec62: OUTPUT«True␤»
AlexDaniel m: say <0.99999999999999999999999999999999999999999999999999999999999> < 1 23:51
camelia rakudo-moar fdec62: OUTPUT«True␤»
23:52 bjz joined
herby_ the perl 6 docs say that "say so 42.list.elems" should ouput 1 23:53
does camelia translate that to True?
m: say so 42.list.elems
camelia rakudo-moar fdec62: OUTPUT«True␤»
23:54 bjz_ left
AlexDaniel herby_: why would it output 1 if there's “so”? 23:56
jdv79 what is 0.A? 23:57
arnsholt herby_: Where in the docs does it say that? That's almost certainly outdated
jdv79 m: say 0.9999999999999999999999.perl 23:58
camelia rakudo-moar fdec62: OUTPUT«0.A␤»
Guest99713 arnsholt: I found it. Forking it to fix it.
jdv79 m: say 0.9999999999999999999999.WHAT
camelia rakudo-moar fdec62: OUTPUT«(Rat)␤»
AlexDaniel doc.perl6.org/type/Any#method_list
yeah 23:59
herby_ Thanks Guest, I was looking for it
AlexDaniel just remove “so”, I guess
cygx T-2m
herby_ I didn't really understand the example so I was going to run it myself and test it out
Guest99713 Done: Pull request is 'Removed erroneous use of "so" #225'

Powered by Cro and the Raku® Programming Language.

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

Thank you!