»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
psch eh, still behaves wierd with .list tbh 00:00
m: say (m/@<a>=A+/)<a>.list for <A AAAA AAAAAAA>
camelia rakudo-moar c880f1: OUTPUT«()␤()␤()␤»
psch not sure what's up there, but then i don't know if i remember @<a>= correctly
pippo I think a single match has to have the .list method working as for several matchs.
m: say 1.list;
camelia rakudo-moar c880f1: OUTPUT«(1)␤»
psch m: $_ = "abc"; say /(.)+/; say $/.list 00:01
camelia rakudo-moar c880f1: OUTPUT«/(.)+/␤(Nil)␤»
psch m: $_ = "abc"; m/(.)+/; say $/.list
camelia rakudo-moar c880f1: OUTPUT«([｢a｣ ｢b｣ ｢c｣])␤»
psch m: $_ = "abc"; m/(.+)/; say $/.list
camelia rakudo-moar c880f1: OUTPUT«(｢abc｣)␤»
psch pippo: it does, if you don't give it a name
m: $_ = "abc"; m/$<a>=(.+)/; say $/.list
camelia rakudo-moar c880f1: OUTPUT«()␤»
pippo psch: and why with a name it should not work? 00:02
psch pippo: because a name means "this gets saved as if $/ was a hash"
m: $_ = "abc"; m/$<a>=(.+)/; say $<a>.list
camelia rakudo-moar c880f1: OUTPUT«()␤»
psch m: $_ = "abc"; m/$<a>=(.+)/; say $<a>
camelia rakudo-moar c880f1: OUTPUT«｢abc｣␤»
psch m: $_ = "abc"; m/[$<a>=(.)]+/; say $<a> 00:03
camelia rakudo-moar c880f1: OUTPUT«[｢a｣ ｢b｣ ｢c｣]␤»
psch m: $_ = "abc"; m/[$<a>=(.)]+/; say $/.list
camelia rakudo-moar c880f1: OUTPUT«()␤»
psch m: .say for %(a => 1, b => 2).list
camelia rakudo-moar c880f1: OUTPUT«a => 1␤b => 2␤»
psch i admit that's inconsistent
zengargoyle m: $_ = "abc"; m/(.)+/; say $/.list
camelia rakudo-moar c880f1: OUTPUT«([｢a｣ ｢b｣ ｢c｣])␤»
psch as in, %h.list gives pairs as a list
00:04 advwp joined
psch i don't know if we want that for $/, though, and i'm not willing to try to decide if we could use it :) 00:04
zengargoyle well yeah, %h.pairs gets pairs
psch m: $_ = "abc"; m/[$<a>=(.)]+/; say $/.pairs
camelia rakudo-moar c880f1: OUTPUT«(a => [｢a｣ ｢b｣ ｢c｣])␤»
psch match objects are weird :)
00:04 firstdayonthejob left
pippo I agree!!! :-)) 00:05
psch pippo: feel free to RT this, from what i can see there's something to think about wrt the different contexts for $/ and how other things behave in those contexts
i mean, on one hand i can see wanting .hash and .list on $/ different 00:07
'cause if i'm doing a for @$/ { } i might not want to suddenly deal with a pair that has as submatch as value
on the other hand it seems inconsistent with how hashes work in the same context... vOv 00:08
so it's probably a design decision, which i wash my hands of :)
00:09 lsm-desktop joined
zengargoyle i'm sorta resigned to six months of messing up grammars before getting the hang of just how everything works. 00:10
pippo psch: filed bug. 00:13
Hotkeys I think the rakudo repl is broken on windows 00:14
pippo good night psch and zengargoyle o/ 00:15
00:15 pippo left
Hotkeys i.imgur.com/latkrqh.png 00:17
basically the repl gets hung up sometimes, most repeatably after the second expression inputted 00:18
I've tried two windows pcs, one on 7 and one on 10 both do it
but my ubuntu vps does not
00:18 lichtkind left
Hotkeys i.imgur.com/1vJbO5y.png 00:21
psch Hotkeys: yes, the repl is more likely to not work than work. i tried working on it and might have left bits in there that cause the current problems
Hotkeys ah 00:22
alright
I just wanted to make sure it wasn't some weird coincidence
psch Hotkeys: i'm rather occupied with rakudo.jvm, though, and unfortunately there's no one else who has the time or energy to try and make it work
Hotkeys I'd try and help but I doubt I have the ability 00:23
psch Hotkeys: well, it's mostly understanding src/Perl6/Grammar.pm 00:24
Hotkeys: the issue is that we'd want a repl that's smart with newline, to an extend
i'll try and restore the pre-moreinput state in a bit though, 'cause that really needn't be in there anymore 00:25
better to have something clean whenever someone wants to try and implement moreinput...
00:26 softmoth joined, chenryn joined
psch Hotkeys: basically, the repl we had before was "one line or bust", i.e. you'd to make sure to have a complete and valid statement (minus the terminator if that's the only thing missing) 00:26
Hotkeys right
psch Hotkeys: what i tried (and thought i'd removed) was the ability to split a statement over multiple lines, but something in there is hard wrt preserving parser state 00:27
like, <.say^Mfor ^5> would show the * prompt after the ^M, and print 0..4 on one lines each 00:28
00:28 cygx left
psch eh, actually <.say^Mfor ^5^M^M> 00:28
or <.say^Mfor ^5;^M> as equivalent
but yeah, it's a bit hard :P 00:29
00:29 leont left
Hotkeys sounds like it :p 00:30
it's alright I'll just do my fooling around onmy vps
00:30 softmoth left
Hotkeys I have two linguistics classes which by the end will have taught me most of the same stuff 00:31
but the teacher in my night ling class is mcun more engaging 00:32
sprocket Not sure if anyone’s interested in using perl6 and the Raspberry Pi, but I’ve just pushed the first cut of an RPi library to github: github.com/cspencer/perl6-raspberry-pi
Documentation is still lacking, but there’s an example in the bin/ dir giving the basic usage.
and the disclaimer is that it’s VERY slow. :)
psch sprocket: you might wanna PR that into the ecosystem :)
sprocket psch: will do - i’ll flesh out the docs and tests a bit more before doing so. :) 00:33
00:33 grumble joined
sprocket the example I’d commited in the bin/ directory was used to turn on and off and mechanical relay, that will eventually become an irrigation controller for our greenhouse. i’d written the initial version in python, but perl 6 sounded like a more-fun project :) 00:35
00:37 BenGoldberg joined 00:38 aborazmeh joined, aborazmeh left, aborazmeh joined 00:45 adu joined
psch hm, dalek's gone 00:50
Hotkeys: i've removed the last remains of my moreinput attempt, the repl now again tries to parse completely as soon as the user hits enter
Hotkeys is the change live? 00:51
psch github.com/rakudo/rakudo/commit/b9...3d4f545827 00:52
'cause dalek's gone :)
Hotkeys cool 00:53
00:53 grumble left
Hotkeys time to rebuild 00:53
I'm glad I figured out building on windows
'cause the rakudo msis are still on 2015.09 00:54
01:02 chenryn left 01:11 yeahnoob joined
Hotkeys hmm 01:11
it's still doing the thing
psch Hotkeys: uh..? what exactly are you trying? 01:13
Hotkeys i.imgur.com/Mh4zfjA.png 01:14
psch Hotkeys: i don't see that here, but i'm on debian. i'd (with low confidence) suspect something about the \n changes, but... vOv 01:16
Hotkeys if there's anything I can try to help debug let me know
It works fine on my ubuntu vps
psch $ ./perl6-m --version
This is perl6 version 2015.10-193-gc880f16 built on MoarVM version 2015.10-56-g9fd3005
Hotkeys just on my two windows machines
psch is that cmd.exe or powershell?
maybe that matters..?
Hotkeys c:\Program Files (x86)\Microsoft Visual Studio 10.0\VC>perl6-m --version 01:17
This is perl6 version 2015.10-194-gb9383b3 built on MoarVM version 2015.10-56-g9fd3005
cmd.exe
but it does it on powershell too
psch hrmm
Hotkeys I also tried restarting the machine when it was happening on my desktop
psch well, \n kind of could make sense, but i really don't know... 01:18
shows you how neglected the repl is... :)
Hotkeys i.imgur.com/xqhcUAA.png 01:19
psch sprocket: i'd second Zoffix++ comment on POSIX for the RPi module, fwiw, but you probably already expected that :)
sprocket psch: i think i missed the comment - where was it?
psch sprocket: github.com/perl6/ecosystem/pull/86 the POSIX ecosystem PR 01:20
it has a cat!
sprocket psch: oh - hah! yes, that was expected, and that’s my current weekend task :)
psch Hotkeys: well, as mentioned, semicolon or no, debian doesn't mind vOv 01:21
Hotkeys yeah it seems to be windows only
anyone else running windows that can confirm?
psch Hotkeys: i don't know if cmd.exe and powershell send \xa\xd or \xd, but maybe the grammar can't deal with whatever it is they send...
Hotkeys: best bet is probably bugging jnthn about it whenever he's around and not digging too hard into all those xmas tickets... 01:22
Hotkeys: *or* digging really deep yourself, starting with a bisect (with the moreinput removal backported, probably...)
Hotkeys: i wouldn't recommend that, though, 'cause NFG seems a bit spooky :P
Hotkeys I was going to build a few different commits to see where it starts breaking 01:23
psch ya, you probably want to start a bisect at 385850bf, after rebasing my latest commit before that 01:24
and as long as you stay on moar, everything should build and only break in that spot, i guess...
Hotkeys should I build jvm and see if that breaks too 01:25
or is this moar only
psch jvm is already broken in quite a few ways, the repl is the least of our worries there...
Hotkeys lol
alright
01:28 aborazmeh left 01:29 zacts left 01:32 aborazmeh joined, aborazmeh left, aborazmeh joined, zacts joined 01:33 z8 joined 01:34 Actualeyes joined 01:35 lue is now known as ShimmerFairy
Hotkeys I'm getting close 01:47
except this version breaks on the first expression rather than the second
I assume it's the same issue though
02:02 dalek joined, ChanServ sets mode: +v dalek 02:03 solarbunny joined 02:11 AlexDaniel left 02:14 softmoth joined 02:19 softmoth left
Hotkeys looks like the bug was introduced in 75b2451 02:24
going to try and see where it started breaking on the second expression instead 02:28
02:31 chenryn joined 02:38 xpen joined 02:43 kid51 left, pippo joined, stanrifkin joined
pippo m: $_ = "abc"; m/$<a>=(.+)/; say $<a>».Str; 02:43
camelia rakudo-moar b9383b: OUTPUT«()␤»
02:43 yqt left, vendethiel joined
pippo m: $_ = "abc"; m/@<a>=(.+)/; say @<a>».Str; 02:44
camelia rakudo-moar b9383b: OUTPUT«[abc]␤»
pippo Hmm… 02:45
02:45 pippo left, ilbot3 left 02:47 ilbot3 joined 02:51 softmoth joined 02:55 softmoth left
gfldex m: my $a = 42; my $b = 666; sub foo(\p1, \p2 is copy){ say p1.VAR.name; say \p2.VAR.name }; foo($a, $b); 02:57
camelia rakudo-moar b9383b: OUTPUT«$a␤\(666)␤»
02:57 kaare_ joined
gfldex m: my $a = 42; my $b = 666; sub foo(\p1, \p2){ say p1.VAR.name; say p2.VAR.name }; foo($a, $b); 02:57
camelia rakudo-moar b9383b: OUTPUT«$a␤$b␤»
ugexe m: my $a = 42; my $b = 666; sub foo(\p1, \p2 is copy){ say p1.VAR.name; say \p2.VAR.name }; foo($a, $b);
camelia rakudo-moar b9383b: OUTPUT«$a␤\(666)␤»
ugexe oops i was going for the same thing 02:58
gfldex m: my $a = 42; my $b = 666; sub foo(\p1, \p2 is copy){ say p1.VAR.name; say p2.VAR.name }; foo($a, $b);
camelia rakudo-moar b9383b: OUTPUT«$a␤$b␤»
gfldex that leaves the question what '\p2.VAR.name' actually does 03:00
03:00 lizmat_ joined, lizmat left 03:02 cxreg2 is now known as cxreg
zengargoyle pippo: When used as a hash, a Match object pretends to be a hash of all its named captures. The keys do not include any sigils, so if you capture to variable @<foo> its real name is $/{'foo'} or $/<foo>. However, you may still refer to it as @<foo> 03:03
looks like @<a> just forces it to be [] even if only one match. 03:06
03:09 vendethiel left 03:12 yeahnoob left 03:16 Praise joined, Praise left, Praise joined
zengargoyle m: $_ = "abc"; m/$<a>=(.+)/; say @($<a>,)>>.Str 03:19
camelia rakudo-moar b9383b: OUTPUT«(abc)␤»
zengargoyle m: $_ = "abc"; m/$<a>=(.+)/; say [$<a>]>>.Str
camelia rakudo-moar b9383b: OUTPUT«[abc]␤»
03:19 adu left 03:29 mr-foobar left 03:34 aborazmeh left 03:37 ShimmerFairy left 03:38 BenGoldberg left 03:43 kaare_ left 03:49 ShimmerFairy joined
stanrifkin is there a book about perl6? I think we need a printed book. 03:51
Hotkeys books are for nerds 03:53
03:53 nys left
Hotkeys if we get a book it should wait until 6.christmas probably 03:53
stanrifkin 6.christmas? 03:56
timotimo currently we're at 6.Birthday, aka 6.Beta 03:57
6.christmas will be "the" release
04:01 chenryn left 04:02 chenryn joined 04:03 chenryn left
skids m: my @a <== 4,5; @a.perl.say; @a.join("").say 04:04
camelia rakudo-moar b9383b: OUTPUT«[(4, 5),]␤4 5␤»
timotimo oh, that's item-y now? 04:05
skids I don't know what went on there. I think feeds kinda got left by the GLR wayside.
timotimo could very well be 04:06
ugexe i believe its supposed to flatten but doesnt 04:07
timotimo interesting
you can look at the --target=ast -e 'my @a <== 1, 2, 3'
i think all it needs to do is replace .push with .append 04:08
skids Are loop/while is still planned to eventually return a list of block results like for does? 04:09
timotimo good question; i remember it was supposed to be one of the simplest ways to create a list programmaticaly, eh? 04:10
ugexe: i'm a bit surprised (but then again ... not that surprised) that spec tests didn't show that my @a <== 1, 2, 3 thing 04:12
skids I'm wondering if efficiency gains acheived by while-izing for-like things will remain if it does (or why for is generally slower for that matter) 04:13
04:13 chenryn joined
timotimo good question 04:13
i'm going to spectest changing the .push to .append 04:14
because that's rather obviously what it meant to do
ugexe i ran into it awhile ago, and i only mention the flattening because someone else asked in irc and toady commented on it (which is how i learned it was incorrect in the first place)
timotimo mhm 04:15
ugexe but it was around glr i believe
timotimo i used feeds SO rarely in the past
ugexe why? always use the coolest looking operator 04:16
timotimo that's how that piece of code cygx wrote got so popular :D
04:21 Actualeyes left
perlawhirl hi hall 04:24
er.. i mean... hi all
i'm building perl 6 on my aws instance... i've done it before... but this time it seems to be getting hung at "Stage mast : " 04:25
i know this step takes a long time... but it's been hours.... in fact my instance appears to no longer be responding (it's a free tier, maybe out of mem) 04:26
timotimo quite likely that it could be out of memory
perlawhirl anyways. i'm certain perl 6 will run fine, how do i skip these test and just install it?
timotimo "Stage mast :" isn't a test
perlawhirl oh
so... what are my options
timotimo but you can really just build it locally and copy it over, as long as your paths match up
perlawhirl ok... and same arch too? like 64bit? 04:27
or does that not matter
timotimo 64bit is very recommended, because that'll give you a JIT compiler
for quite some performance improvements
perlawhirl yep, ok... i'm sure i built an older release on free aws without issues... oh well... will spin up a local system and build it locally 04:28
should it also be the same distro?
timotimo probably not necessary. it could help
perlawhirl k... will see how i go... i generally --prefix=/opt/perl6 so matching paths is of minimal hassle 04:29
timotimo if you're spinning up a vm locally to build it anyway, no reason to choose something different, i'd say
good
perlawhirl yep, i just already have a 64-but centos system running here, and my aws us ubuntu
eh... will copy it up and see how it goes.. any issues i'll spin up a ubuntu vm to build it 04:30
timotimo i see
i mean, i believe you can build moarvm with a static libmoar.so
i've never tried that, so no clue if it actually works
but that could make your build much more independent from versions on your system and such
perlawhirl yeh... will stick to what i know 04:31
ok thanks
timotimo good luck! i should probably really go to bed
hm. my change to feeds caused deprecations to break ... ?!? 04:32
perlawhirl it's only 1530pm here in australia... so you must be operating on australian time
timotimo what the ... 04:34
04:36 Actualeyes joined
timotimo derp 04:36
i had a dirty core setting
that changes things about exceptions, or rather backtraces
04:36 sprocket left 04:37 eliasr left 04:40 softmoth joined 04:44 softmoth left 04:49 mr_ron joined
dalek kudo/nom: a2a34c2 | timotimo++ | src/Perl6/Actions.nqp:
feeds aren't supposed to itemize things

there's no difference in the spec tests with or without this change, so we'll really want to up the coverage of feeds		 04:50
mr_ron m: use newline :crlf; "abc.txt".IO.spurt: "abc123\ndef456\n"; my $f = slurp "abc.txt"; dd $f 04:51
camelia rakudo-moar b9383b: OUTPUT«Str $var = "abc123\r\ndef456\r\n"␤»
mr_ron methinks slurp is buggy here 04:52
timotimo isn't camelia supposed to prevent .IO from being used?
mr_ron don't know? 04:53
m: unlink 'abc.txt'
camelia rakudo-moar b9383b: OUTPUT«unlink is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub unlink at src/RESTRICTED.setting:19␤ in block <unit> at /tmp/L_OQ0AZoQA:1␤␤»
04:55 mr_ron left 05:05 khw left 05:10 nightfrog left 05:12 yeahnoob joined 05:13 nightfrog joined 05:22 skids left 05:41 tardisx joined 06:02 telex left 06:04 telex joined 06:05 cognominal left 06:10 laouji joined 06:16 apejens_ is now known as apejens 06:27 laouji left 06:29 softmoth joined 06:33 softmoth left 06:35 quester joined 06:36 domidumont joined 06:40 domidumont left 06:42 sftp joined 06:43 tardisx left, tardisx joined, tardisx left 06:57 geraud left
dalek kudo/nom: 1458f0e | TimToady++ | src/Perl6/ (2 files):
must canonicalize name adverb declarations too!
 06:59
07:02 diana_olhovik_ joined
dalek ast: 6afbefb | TimToady++ | S17-scheduler/in.t:
detach adverb from name it glommed onto
 07:02
ast: 70293c9 | TimToady++ | S02-names-vars/varnames.t:
Make sure name adverb declarations canonicalize
[Tux] test 50000 35.146 35.026
test-t 50000 17.972 17.852
07:07 rindolf joined
bartolin_ good morning, #perl6 07:18
if someone wants to look at the JVM breakage: With the following bandaid, I'm able to build a current rakudo.jvm: gist.github.com/usev6/f7eaf093a38fb893e76b 07:20
... maybe that could give a hint for a cleaner patch 07:21
07:21 domidumont joined 07:29 firstdayonthejob joined 07:30 brrt joined 07:32 _mg_ joined 07:34 jkva joined 07:39 _mg_ left 07:42 diana_olhovik_ left 07:43 diana_olhovik_ joined 07:50 _mg_ joined 07:54 hlafarge joined 07:58 jkva left, jkva joined 08:15 quester left 08:16 darutoko joined 08:18 softmoth joined 08:23 softmoth left, firstdayonthejob left, hlafarge left 08:28 shmibs left 08:29 shmibs joined 08:35 andreoss joined
brrt good netsplit #perl6 08:35
or wasn't that a netsplit 08:36
nine Good Maybe[netsplit] brrt!
brrt ohai nine! 08:37
08:39 zakharyas joined
mrf good * #perl6 08:40
08:44 pierrot_ joined, pierrot left 08:46 abraxxa joined
andreoss why does smoke.perl6.org show GTK::Simple tests failed? where i can see logs? 08:46
08:48 hlafarge joined 08:54 ely-se joined, virtualsue joined
stmuk andreoss: it's probably because it needs a x server 09:02
maybe it's possible to setup a headless one 09:05
lizmat_ good *, #perl6! 09:09
09:09 lizmat_ is now known as lizmat
lizmat .botsnack 09:09
yoleaux :D
dalek kudo/nom: dd1d7fe | lizmat++ | src/core/Str.pm:
Implement :k/:v/:kv/:p for Str.split(@needles)

Also now make a limit functional without needing to resort to regexen.		 09:10
k-simple: 1c1a8cc | andreoss++ | t/01_sanity.t:
Run tests only if $*ENV<DISPLAY> is set
09:19 andreoss left, g4 joined, g4 left, g4 joined 09:26 arunes007 joined 09:27 arunes007 left 09:31 Psyche^_ joined 09:33 xpen left 09:35 Psyche^ left 09:36 cognominal joined, xpen joined 09:37 yeahnoob left 09:46 xfix joined 09:49 RabidGravy joined
lizmat wendyga.wordpress.com/2015/10/08/f...-speakers/ # FOSDEM 2016 09:49
dalek kudo/nom: 2017471 | lizmat++ | src/core/Str.pm:
If limit, only look for that many in each needle
 09:52
brrt lizmat; good point 10:03
10:05 espadrine joined 10:07 softmoth joined, domm joined 10:09 tokuhir__ joined 10:10 eliasr joined 10:11 tokuhiro_ left, softmoth left 10:14 hlafarge left
dalek pan style="color: #395be5">perl6-examples: b981db8 | (Konrad Borowski)++ | / (2 files):
[euler] Add solution for 4
 10:14
10:15 shamu left, shamu joined 10:26 eternaleye joined 10:29 stanrifkin left 10:31 _mg_ left
jnthn morning, #perl6 10:38
yoleaux 5 Nov 2015 18:35Z <nine> jnthn: fun starts when people do use Test; use lib 't/spec/packages'; use Test::Util; because then we load a precomped Test from one repo and a precomped Test::Util from another and the latter is based on a different precomp version of Test.
dalek pan style="color: #395be5">perl6-examples: bc64a7e | (Konrad Borowski)++ | categories/ (11 files):
Remove executable bit from non-executable files

Those files aren't executable files, as in a compiled executable, but rather Perl 6 source files. Without a shebang, executable bit is useless, and potentially even confusing.		 10:39
10:44 z8 left
lizmat jnthn o/ 10:44
jnthn o/ lizmat 10:46
timotimo o/
|Tux| lizmat, :v, :skip-empty => 10:48
All tests successful.
Files=25, Tests=19732, 58 wallclock secs ( 3.05 usr 0.17 sys + 203.51 cusr 2.14 csys = 208.87 CPU)
xfix Odd, perl6-examples tests get stuck at testing LWP::Simple...
travis-ci.org/perl6/perl6-examples...s/89617797
lizmat |Tux| :-)
|Tux| Nice to note: Inline::Perl5+XS 16.022, plain perl6: 17.578 10:51
SO CLOSE!
jnthn .tell mr_ron slurp is fine, we don't transform \r\n to \n on I/O, the newline pragma changes the meaning of \n in strings
yoleaux jnthn: I'll pass your message to mr_ron.
timotimo oh? i thought "make \n magic" would make it actually magical 10:52
jnthn timotimo: "Means what $?NL says it should" is already quite magical for an escape sequence ;) 10:54
timotimo oooh 10:55
sorry, i misread!
10:57 jkva left
xfix I'm thinking about replacing the third example on perl6.org/ 10:58
nine |Tux|: looks like I have to speed up Inline::Perl5 soon ;)
xfix That is, instead of `fail "Not a Natural Number in Factorial" if $n < 0;`, it would have `Int $n where $n >= 0`
Does that look fine? 10:59
timotimo xfix: that'll throw an exception rather than a Failure, though, right?
jnthn |Tux|: Looks like I need to find more time for optimizing ;-)
xfix Eh, I guess.
11:00 woolfy joined
xfix But at the same time, negative numbers in factorial are exceptional condition, aren't they :-)? 11:00
are an exceptional condition*
jnthn nine: Saw your comment; brane is still not firing on all cylinders yet so will get back to you on it later today :)
woolfy London Perl Workshop is really looking for more presentations about Perl 6.
nine jnthn: thanks!
woolfy Mark Keating writes: "Heya Perl6 folksies. LPW CforP officially closes today but the site will actually accept submissions all the way until Monday - don't forget we may be closing Lightning Talk submissions early as well - we need more Perl6 if we can get it." 11:01
xfix Hm, there is a test for giving a Failure...
woolfy So, if you want to show the awesomeness of Perl 6 to an audience of at least 150 people, go here and submit a talk: act.yapc.eu/lpw2015/newtalk 11:02
|Tux| ENOTIME
xfix dies-ok { (-1)! }, "Factorial for -1 throws an exception"
11:05 sitaram joined 11:09 AlexDaniel joined
dalek ast: 858b2e0 | lizmat++ | S32-str/split.t:
Add tests for Str.split(@needles)
 11:10
xfix That depends on how bad is passing negative value to factorial is, I guess. 11:11
Division by zero is a failure, so I guess that makes sense. 11:12
RabidGravy weird that on travis sometimes access to the github api directly results in a 403 and sometimes it doesn't 11:25
11:32 z8 joined
moritz RabidGravy: rate limiting? 11:33
11:36 xfix left 11:37 Sqirrel left 11:39 sitaram left
dalek kudo/nom: 6b4845d | lizmat++ | src/core/Str.pm:
Better error on mixed nameds exception
 11:39
kudo/nom: e3713a0 | lizmat++ | src/core/Str.pm:
Handle an empty needle by ignoring it

The idea behind this is that if a list of needles is generated from some external input, that you can group types of needles resulting in fixed index numbers (even if they are not all specified).
11:42 liufeng joined 11:47 xpen left 11:50 xpen joined 11:51 Skarsnik joined 11:53 jkva joined 11:54 perlIRC joined 11:55 xpen left 11:56 softmoth joined
RabidGravy moritz, ah yes - anyhow I'll leave the test TODO for the time being ;-) 11:56
perlIRC For coding with autocompletion which IDE is best for perl 6?????? 11:57
11:58 jkva left
DrForr I think Microsoft's new Visual Studio supports Perl 6, not sure how far autocomplete goes. 11:58
tadzik wasn't that Visual Studio Code? 11:59
brrt yeah, that's VS code 12:00
not 100% sure what the relation between these are
RabidGravy many a long year that I've used Visual Studio
12:00 softmoth left
RabidGravy ^ since 12:00
tadzik Code is just a fancy, trendy webby editor, a bit like github's Atom 12:01
iv it's very well integrated. more cohesive than atom's "a big pile of plugins" approach 12:02
perlIRC Code?? send me the link plz 12:07
timotimo "visual studio code" 12:08
12:08 brrt left
perlIRC tadzik "Code" editor?? 12:09
12:10 azawawi joined
timotimo it's called "visual studio code" 12:10
azawawi hi
RabidGravy erp
DrForr perlIRC: Maybe read what people are saying? "visual studio code" as timotimo said :)
azawawi since i was looking at the logs :)
perlIRC: your best bet for windows and *inux is atom.io 12:11
perlIRC: visual studio code is just github's atom editor with faster integration on windows
timotimo "faster"?
iv code.visualstudio.com 12:12
azawawi timotimo: yup on windows visual studio code is faster a bit
timotimo interesting
azawawi timotimo: they preload dlls... etc
timotimo: it is their OS lol... they can optimize for it :)
timotimo well, sure
12:12 jkva joined
DrForr Works for Volkswagen. 12:12
perlIRC perl6 script ends with what?? Is it .p6?? 12:13
timotimo .p6 for scripts, .pm6 for modules; that is quite common 12:14
12:14 azawawi_v2 joined
azawawi_v2 perlirc: if you use .p6 and .pm6, syntax highlighting work even with github 12:15
perlIRC: s/work/works/
RabidGravy a *script* doesn't have to have an extension on most platforms, however on windows the association depends on the extension
flussence
.oO( until they break everything again )
timotimo that's true, too
flussence: well, we have very talented people who can usually work around the issue given enough time
azawawi_v2 also atom editor identifies shebangs #/usr/bin/env perl6 12:16
12:16 azawawi left
timotimo that's worth a lot. windows doesn't though. except perl, i've heard 12:16
or was that just prove that does it? no, i think it's perl.
12:17 Sqirrel joined
flussence timotimo: yeah, the people writing those syntax highlighters (usually in a horrible language :) deserve a *lot* of credit. 12:17
timotimo i fully agree
azawawi_v2 i was just looking at atom's syntax highlighter for perl6... github.com/atom/language-perl/blob...pec.coffee 12:18
12:18 azawawi_v2 is now known as azawawi
timotimo that's just a test file 12:18
github.com/atom/language-perl/blob...l%206.cson - this is the highlighter grammar
azawawi i hate atom's not identifying perl6-identifiers github.com/atom/language-perl/blob...l%206.cson :0 12:19
github.com/atom/language-perl/blob....cson#L256 # for identifiers 12:20
RabidGravy gently nudges tadzik about the github.com/tadzik/JSON-Unmarshal/pull/10
:) 12:21
azawawi looks at the channel graphs at irclog.perlgeek.de/
timotimo azawawi: not bad, eh? 12:23
tadzik oh, oh :)
azawawi perl 6 graph is interesting
tadzik will look
azawawi moritz++
RabidGravy talkative bunch aren't you 12:24
12:24 andreoss joined
RabidGravy anyway farm shop - need more vegetables 12:24
see y'all later
azawawi we have also a channel called #masakism :)
timotimo oh yes
that was the name of a little workshop masak did 12:25
tadzik that brings fun memories
timotimo i'd call them "fund memories"
azawawi www.theregister.co.uk/2015/10/31/br...r_fortran/ # imagine that for Perl in twenty years :) 12:27
12:29 _mg_ joined 12:30 RabidGravy left
azawawi i think p6 syntax highlighting in general is an interesting topic that i tackled earlier when i first joined #perl6 12:30
moritz++ was focused on correctness ofcourse 12:31
PerlJam if i remember correctly was focused on speed
most syntax highlighter cover 90% of the most common cases 12:32
s/highlighter/highlighters/ 12:33
i wonder if our earlier approach to the project was right... i.e. we should have built tests first for what a perl 6 syntax highlighter should do 12:34
and then focused on speed for implementations that pass common test cases 12:35
.tell moritz please see irclog.perlgeek.de/perl6/2015-11-06#i_11494509 12:36
yoleaux azawawi: I'll pass your message to moritz.
azawawi going out for a walk & 12:37
12:37 azawawi left, kaare_ joined
moritz sees 12:37
yoleaux 12:36Z <azawawi> moritz: please see irclog.perlgeek.de/perl6/2015-11-06#i_11494509
12:38 weihan joined, sufrostico joined
dalek ast: fd2cdc2 | jnthn++ | S16-io/handles-between-threads.t:
Tests for reading from $*IN from threads.

Covers RT #124005.		 12:39
12:42 kid51 joined
dalek ast: f98782d | jnthn++ | S16-filehandles/unlink.t:
Remove utterly bogus tests.

There is no implicit file close at scope exit in Perl 6, and the tests only appeared to work in any sense on non-Windows because you can unlink an open file there (and on Windows not).		 12:43
sunnavy_ sunnavy
12:43 sunnavy_ is now known as sunnavy 12:51 kid51 left, weihan left 12:58 chenryn left, perlIRC left
tadzik Rab<TAB><TAB> 13:00
dang
.tell RabidGravy pure gold, merged it :) Thanks!
yoleaux tadzik: I'll pass your message to RabidGravy.
nine Rab tab tab dang.....sounds like a song ;)
tadzik or a music genre :P 13:02
masak evening, #perl6 13:03
timotimo chinevening, masak
... chinesening?
masak UTC+08:00-vening
ely-se should run automatically append .exe, .bat and .cmd like some other functions in other libraries of other programming languages do on Windows? 13:04
masak last evening in .cn, then I'll be downgrading to European time again.
ely-se: sounds... plausible.
13:08 aborazmeh joined, aborazmeh left, aborazmeh joined 13:12 joydo joined 13:13 yqt joined
|Tux| lizmat, gist.github.com/Tux/0bf175a73e713b2e43e9 13:13
lizmat oki, looking 13:14
ely-se masak: I don't actually know which Windows API functions do and do not do this 13:18
|Tux| was preparing a :k branch to see if there is any more performance to gain
lizmat |Tux|++ 13:19
|Tux| damn you perl6 folk! I don't have time for this, but this speed gain motivates me to do things I shouldn't do :)
lizmat :-) 13:21
|Tux| lizmat, is it defined/documented which index is returned when there are multiple equal entries in @needle?
13:21 joydo left
lizmat no, but it will be the first 13:21
|Tux| expected the first, but it is quite often the last
it isn't 13:22
lizmat hmmm.... actuallu
atm it's undefined
not sure how stable p6's sort is
|Tux| I can live with that, but then again, my expectation was different
lizmat I will make it so it is the first 13:23
that will match the expectation of LTM, I think
|Tux| lizmat++
lizmat of course, it's better to not do this :-)
|Tux| @chx.push: $sep, $quo, $noe ?? Str !! $esc;
=> @chx.push: $sep, $quo, $esc; 13:24
lizmat but I will make it so it will always do the first
|Tux| if it were defined to be the first
13:25 virtualsue left
dalek kudo-star-daily: f06b4cb | coke++ | log/ (7 files):
today (automated commit)
 13:29
13:31 duncan_dmg joined 13:32 cognominal left
dalek kudo/nom: 365449e | lizmat++ | src/core/Rakudo/Internals.pm:
Add NOT_ALL_TYPE internal method
 13:35
kudo/nom: 4c4b4c1 | lizmat++ | src/core/Str.pm:
Identical needles always index to the first
13:41 n0tjack joined 13:42 rurban joined, yakudza left 13:43 n0tjack_ joined 13:45 softmoth joined 13:46 n0tjack left 13:48 n0tjack_ left
[Coke] I'm going to merge bartolin's modification of my patch to get the jvm unbusted. anyone wanting a cleaner fix can hunt for all the #?if jvm markers... 13:53
13:53 aborazmeh left 13:56 yakudza joined
dalek rl6-most-wanted: 35dd0dd | (Zoffix Znet)++ | most-wanted/modules.md:
Mark Data::GUID as stalled / List UUID

Doing GUID right is harder than I expected. Marked it as stalled, in case someone smarter wants to have a crack at it, before I figure it out.		 13:56
13:59 zakharyas left 14:01 Actualeyes left 14:02 RabidGravy joined 14:11 brrt joined 14:15 Actualeyes joined 14:17 zengargoyle left 14:19 zengargoyle joined 14:25 cfloare left 14:26 cgfbee left 14:28 muraiki joined
RabidGravy the weather is yuck out there 14:30
yoleaux 13:00Z <tadzik> RabidGravy: pure gold, merged it :) Thanks!
14:30 skids joined
RabidGravy tadzik++ # superstar 14:30
14:35 virtualsue joined
dalek kudo/nom: ccefe10 | coke++ | src/core/Failure.pm:
Revert "Fix some strange failures", at least on the JVM

This broke rakudo-jvm
This reverts commit 5349d05c87387e1c4f70e4e7d1591ed17114cc35.		 14:37
kudo/nom: 771995d | coke++ | src/core/Failure.pm:
partially revert "Failure should also not return useful integer"

This broke rakudo-jvm
This reverts commit ad08a1d71935a12e6e3e0814d14ae357b2064c62.
[Coke] ^^ bartolin++
kudo/nom: 8b60779 | coke++ | src/core/IO/Handle.pm:
partially revert "Don't forget to set seps on special handles."

This broke rakudo-jvm
This reverts commit cc4d72c4abb243f44500a7c435e2e9946325f3bd.
14:41 sprocket joined, cgfbee joined, mr_ron joined 14:42 cfloare joined 14:48 xpen joined, sprocket left, sprocket joined 14:53 xpen left 14:54 tokuhir__ left 14:55 rurban left
sprocket hello all 14:56
RabidGravy yo 14:57
moritz ilmari++ # pull requests for moar, nqp, rakudo, roast 15:00
15:02 pippo joined
ilmari moritz: well, I needed to test that it worked, so I had to wire up all the bits between moar and roast 15:02
15:02 abraxxa left
pippo m: my regex first { ^\<a }; my regex second { ^\<b }; my @a = /<first>/, /<second>/; my $t = now; grep /<first>|<second>/, "aaaaaaaaaaa bb" xx 45_000; say now - 15:02
camelia rakudo-moar 8b6077: OUTPUT«5===SORRY!5=== Error while compiling /tmp/bP0JY19eJe␤Confused␤at /tmp/bP0JY19eJe:1␤------> 3d>/, "aaaaaaaaaaa bb" xx 45_000; say now7⏏5 -␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement …»
pippo m: my regex first { ^\<a }; my regex second { ^\<b }; my @a = /<first>/, /<second>/; my $t = now; grep /<first>|<second>/, "aaaaaaaaaaa bb" xx 1_000; say now - $t; 15:03
camelia rakudo-moar 8b6077: OUTPUT«0.1188711␤»
15:03 duncan_dmg left, rurban joined
pippo m: my regex first { ^\<a }; my regex second { ^\<b }; my @a = /<first>/, /<second>/; my $t = now; grep /@a/, "aaaaaaaaaaa bb" xx 1_000; say now - $t; 15:03
camelia rakudo-moar 8b6077: OUTPUT«2.3813503␤»
pippo ^^can someone explain to me why that huge time difference?
15:05 duncan_dmg joined 15:06 raiph joined 15:07 gtodd left
moritz m: my $t = now; grep rx { ^\<a }; grep /@a/, "aaaaaaaaaaa bb" xx 1_000; say now - $t 15:07
camelia rakudo-moar 8b6077: OUTPUT«5===SORRY!5=== Error while compiling /tmp/V5TNhkHNzF␤Variable '@a' is not declared␤at /tmp/V5TNhkHNzF:1␤------> 3my $t = now; grep rx { ^\<a }; grep /7⏏5@a/, "aaaaaaaaaaa bb" xx 1_000; say now ␤»
moritz m: my $t = now; grep rx { ^\<a }, "aaaaaaaaaaa bb" xx 1_000; say now - $t
camelia rakudo-moar 8b6077: OUTPUT«0.07905799␤»
moritz m: my $t = now; grep rx { ^\<b }, "aaaaaaaaaaa bb" xx 1_000; say now - $t
camelia rakudo-moar 8b6077: OUTPUT«0.07801403␤»
moritz pippo: I'd say "because interpolation of arrays into regexes is slow". Why though I don't know. 15:08
jnthn Probably because it can't rely on the array's contents not changing over time, and because | has LTM semantics it has to re-compute the NFA every time
m: my regex first { ^\<a }; my regex second { ^\<b }; my @a = /<first>/, /<second>/; my $t = now; grep /||@a/, "aaaaaaaaaaa bb" xx 1_000; say now - $t; 15:09
camelia rakudo-moar 8b6077: OUTPUT«2.7571054␤»
jnthn hm, that's not much of a speedup
15:09 domidumont left
jnthn In fact, it's slower 15:09
15:09 cfloare left, pmurias joined
jnthn m: my regex first { ^\<a }; my regex second { ^\<b }; my @a = &first, &second; my $t = now; grep /@a/, "aaaaaaaaaaa bb" xx 1_000; say now - $t; 15:10
camelia rakudo-moar 8b6077: OUTPUT«1.94181371␤»
pippo jnthn: moritz apparently it dipends on the lenght of the string to be matched.
jnthn That removes the unrequired level of indirection
pippo: Sure, 'cus it'll be scanning through the string
15:10 cgfbee left
pippo jnthn: but there is ^ anchor 15:10
moritz grepping for one regex is 80ms; grepping for two is 2000ms 15:11
so I'm pretty sure it's the array that's causing the overhead
jnthn pippo: The top-level regex interpolating them can't know that 15:12
moritz pippo: I don't think we propagate the information upwards that each alternative is anchored
jnthn m: my regex first { ^\<a }; my regex second { ^\<b }; my @a = &first, &second; my $t = now; grep /^@a/, "aaaaaaaaaaa bb" xx 1_000; say now - $t;
camelia rakudo-moar 8b6077: OUTPUT«0.2299179␤»
jnthn Move the anchor there and it's a lot faster
pippo Ha. OK. TY!
moritz m: my $t = now; grep / ^\<a | ^\<b /, "aaaaaaaaaaa bb" xx 1_000; say now - $t;
camelia rakudo-moar 8b6077: OUTPUT«0.0987924␤»
moritz still much faster with a direct alternation 15:13
m: my $t = now; grep / ^\<a || ^\<b /, "aaaaaaaaaaa bb" xx 1_000; say now - $t;
camelia rakudo-moar 8b6077: OUTPUT«0.0846258␤»
moritz ... even when it doesn't LTM
so there's still a huge overhead from the array interpolation
jnthn Aye 15:14
pippo moritz: indeed. 15:15
15:16 cfloare joined 15:17 cgfbee joined 15:18 yqt left
mr_ron m: use newline :crlf; "abc.txt".IO.spurt: "abc123\ndef456\n"; my $f = slurp "abc.txt"; dd $f 15:19
yoleaux 10:51Z <jnthn> mr_ron: slurp is fine, we don't transform \r\n to \n on I/O, the newline pragma changes the meaning of \n in strings
camelia rakudo-moar 8b6077: OUTPUT«Str $var = "abc123\r\ndef456\r\n"␤»
mr_ron it looks like it translated \n to \r\n on the way out but wondering why it didn't do the reverse when it read 15:20
jnthn Read what I wrote.
zengargoyle m: my $s = 'abc'; $s .= flip; say $s;
camelia rakudo-moar 8b6077: OUTPUT«cba␤»
mr_ron m: use newline :crlf; dd "abc123\ndef456\n"; 15:21
camelia rakudo-moar 8b6077: OUTPUT«Str $var = "abc123\r\ndef456\r\n"␤»
zengargoyle is space before '.' allowed everywhere or just somewheres? some docs still say no space allowed before '.'
ugexe id think [.*?] wouuld work 15:22
er sorry im stuck on regex mode 15:24
pippo zengargoyle: p6weekly.wordpress.com/2015/09/28/...-birthday/ 15:26
mr_ron Stuck on a problem with a stray \r\n installing panda on win32 with an up to date Perl ... still working on it ...
pippo See "Larry’s Birthday Present To Developers" 15:27
zengargoyle: Hope this is related to your question.
15:28 zakharyas joined, pippo left, sprocket left 15:31 sprocket joined
zengargoyle pippo: that looks like the answer 15:31
dalek p: d323210 | jnthn++ | tools/build/MOAR_REVISION:
Bump MOAR_REVISION for encoding improvements.
 15:32
p: 6261d63 | (Dagfinn Ilmari Mannsåker)++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
Wire up encoderep MoarVM op
 15:33
p: 40cba3f | jnthn++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
Merge pull request #251 from ilmari/encode-replacement

Wire up encoderep MoarVM op
zengargoyle my $s = 'abc'; say [.] $s flip
ilmari jnthn++ # mergage
jnthn Rakudo one coming up soon 15:34
zengargoyle m: my $s = 'abc'; say [.] $s flip 15:35
15:35 sprocket left
camelia rakudo-moar 8b6077: OUTPUT«5===SORRY!5=== Error while compiling /tmp/gheMGbIWP0␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/gheMGbIWP0:1␤------> 3my $s = 'abc'; say [.]7⏏5 $s flip␤» 15:35
RabidGravy ilmari++ # dodgy scandiwegian 15:37
dalek kudo/nom: a87e868 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for encoding improvements.
ilmari someone else can fix it for jvm and js :-P 15:38
dalek kudo/nom: e7bc028 | (Dagfinn Ilmari Mannsåker)++ | src/core/Str.pm:
Add :replacement(Str) adverb to Str.encode

Without an argument it uses a per-encoding default.
kudo/nom: 76ee2d8 | jnthn++ | src/core/Str.pm:
Merge pull request #576 from ilmari/encode-replacement

Add :replacement(Str) adverb to Str.encode
ilmari (the rakudo bits are #?if moar-ed out)
jnthn Yes, noticed that. Thanks :)
dalek ast: d6712a9 | (Dagfinn Ilmari Mannsåker)++ | S32-str/encode.t:
Test non-encodable characters
 15:39
ast: c9eab1a | jnthn++ | S32-str/encode.t:
Merge pull request #76 from ilmari/encode-replacement

Test non-encodable characters
ilmari couldn't test that that actually worked, as rakudo-jvm wouldn't build for me
jnthn All merged, I think :) 15:40
BooK so the other day in a bar, after quite a few beers, I made the following observation to ab5tract_: 15:41
Only perl can parse Perl, but of course Perl6 can parse Perl6.
does that make sense?
15:42 _mg_ left
BooK or am I misunderstanding the power of Perl6? 15:42
RabidGravy well in the sense that the Perl6 grammar engine does parse Perl6 yes
BooK RabidGravy: yes, that's what I meant
RabidGravy Yay! 15:43
BooK writing PPI in Perl6, or rather anything that depends on PPI is much simpler in Perl6
RabidGravy BooK++
JimmyZ m: use nqp; nqp::say(unbox_s(Str));
camelia rakudo-moar 8b6077: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Y5thaxkCrA␤Undeclared routine:␤ unbox_s used at line 1␤␤»
JimmyZ m: use nqp; nqp::say(nqp::unbox_s(Str)); 15:44
camelia rakudo-moar 8b6077: OUTPUT«Cannot unbox a type object␤ in block <unit> at /tmp/aQHYe2f8e8:1␤␤»
BooK RabidGravy: that's a very private joke, since you not only have to know the difference betweem PERL and Perl, but also between perl and Perl :-)
JimmyZ ilmari: I saw you code trying to unboxs Str
jnthn ilmari: I took care of RT #123673
ilmari JimmyZ: no, note the "if $replacement.defined"
15:44 g4 left
ilmari m: say Str.defined 15:45
camelia rakudo-moar 8b6077: OUTPUT«False␤»
JimmyZ oh, missed it ..
RabidGravy what about "PeRl" eh? eh?
jnthn There's one more off the xmas list :)
RabidGravy everyone++ # "they're getting there"
JimmyZ ilmari++ # 1 off the xmas list 15:46
mrf #perl6++ # You are all awesome people 15:47
ilmari JimmyZ: and I haven't even written my first perl6 program yet :-P
hacking on the implementation is more fun... or something
jnthn ilmari: Though since the implementation is partly in Perl 6, you have at this point written some Perl 6 ;) 15:48
ilmari jnthn: true, but nothing that qualifies as "a program"
JimmyZ the core setting is your first Perl 6 program :P
jnthn ilmari: True also :) 15:50
RabidGravy ilmari, I've got a list if you want to play a long at home ;-)
mrf has apparently now written more NQP than Perl6 :D
15:51 xpen joined, khw joined
arnsholt mrf: The biggest part of my contributions to the Perl 6 project have involved writing C code =) 15:53
--
mrf arnsholt: :D
JimmyZ andreoss: me too ..
arnsholt Oh, and some Java code
JimmyZ arnsholt: ^^
jnthn isn't sure what language he's written most of for Perl 6 :) 15:54
15:55 tokuhiro_ joined
jnthn I'm pretty sure it's not Java, even if it felt like it at the time :P 15:55
arnsholt =D
Yeah, I'm pretty sure you've written more C. At least if we include MoarVM (which we should) 15:56
15:56 xpen left
arnsholt And large chunks of my C code wouldn't have happened without your C code as a model, so there's that too =) 15:56
JimmyZ repr in parrot and moarvm !
jnthn yeah, MoarVM may push it towards "more C" 15:57
arnsholt Although if we ask GitHub, it'll probably reply Parrot bytecode =) 15:58
nine jnthn: I've had a nice discussion with lizmat++ about my use lib issue yesterday. The only way forward that I see is not precompiling modules loaded after a use lib. Or to be more precise: after a second PrecompilationStore object is created. 15:59
15:59 tokuhiro_ left
zengargoyle hrm, Linenoise failed during the panda auto update step thing from `rakudobrew build moar` 15:59
jnthn nine: Yes, that seems reasonable; I was mostly wondering "is that sufficient"
nine jnthn: as it is now, even a use lib done after a use lib could be a problem since lib is just a module loaded from a repository. 16:00
zengargoyle but installed fine with `panda install Linenoise`
raiph BooK: stackoverflow.com/a/5923181/1077672
nine jnthn: just to be sure, you did intend for us to automatically precompile on module load?
jnthn nine: For the FileSystem repo, yes
nine :) 16:01
BooK raiph: I so wanted to be original, and I'm four years behind :-)
masak BooK: many years ago, Adam Kennedy showed up on #perl6 and said that PPI would be lots *harder* in Perl 6 than in Perl. bordering on impossible.
jnthn nine: So that at development time you can just make test and know that at most you'll only pre-comp your module once for the first test.
masak BooK: it would be interesting to find that backlog and talk about it.
16:01 advwp left
jnthn nine: Until you change things and do another test run, of course :) 16:02
nine Do we have any file locking capabilities yet?
BooK masak: is there a search tool for the irc logs?
16:02 ilbelkyr_ is now known as ilbelkyr
jnthn nine: They're certainly impl'd in Moar, not sure if exposed. Can do so if you need 'em. 16:03
16:03 ely-se left
BooK masak: I don't know Perl6 enough to have an informed opinion 16:03
16:03 advwp joined
pmurias masak: Perl 6 is really oriented towards using the Perl 6 compiler for what PPI does 16:03
masak pmurias: agreed. 16:04
zengargoyle panda issue? rakudobrew issue? gist.github.com/zengargoyle/2a1ca0...e1ed4b3547 16:05
masak pmurias: I'm also running into some interesting issues (in 007) where sometimes you want macro-expanded Qtrees (like when you're compiling/manipulating the program before runtime), and sometimes macro-unexpanded ones (like when you're linting/refactoring it)
colomon Is qq:to/END/ supposed to convert tabs to spaces? It’s very inconvenient for my $work script…. 16:06
ugexe zengargoyle: i think that only happens when its reinstalls line noise after updating rakudo
pmurias masak: for some things you might even want some even more "expanded" version like and iternal desugared/SSA/etc. form
jnthn colomon: There's an xmas RT complaining about that, iirc 16:07
zengargoyle yeah, but it sorta breaks the rest of the re-install automagic
colomon jnthn: If someone can find the RT for me, I’m willing to take a look at fixing the code…
zengargoyle that is all the modules that happen to be after Linenoise don't get re-installed
ugexe it would seem linenoise is leaving panda in a odd directory
16:08 z448 joined, nightfrog left
masak pmurias: yes. I keep coming back to how useful it would be for create-your-own-control-flow hacking if there was a toolkit that gave access to and control over basic blocks. 16:08
16:08 z8 left
jnthn colomon: rt.perl.org/Ticket/Display.html?id=120895 16:09
16:09 nightfrog joined
colomon jnthn++ 16:10
masak 'night, #perl6
RabidGravy zengargoyle, it's a panda issue I actually fixed it with github.com/tadzik/panda/pull/221 but I keep forgetting about it
colomon “Which is bad for e.g. Makefiles.” # exactly my issue
jnthn 'night masak
zengargoyle ah, ok then. :P
mrf ugexe: I have a whole bunch of grammar tests for some of your Grammar::HTML grammar stuff. Will see if I can get them into a fork over the weekend for you if your interested. 16:11
raiph BooK: there are two searchers ( google and custom search engine) linked at top/bottom of the perlgeek.de web irclog pages (eg irclog.perlgeek.de/perl6/)
RabidGravy If someone who has a commit bit there would care to merge it, that would be cool (I don't) 16:12
arnsholt colomon: Sometimes, I wonder if Makefiles aren't the bane of my existence. Still, such a nice tool when it cooperates 16:13
ugexe mrf: sure!
colomon arnsholt: eh, they’re okay when you make nice Perl 6 scripts to generate them. ;) 16:14
mrf ugexe: I have been working on moveing most of teh IETF grammars into a seperate repo for easier use by others. But writting tests is causing it to take longer than I wanted so I am going to port the tests back so that you have them in the meantime.
16:14 nightfrog left, pierrot joined 16:15 pierrot_ left
arnsholt colomon: I have some long handwritten one. Looks like the longest one is 177 lines long =) 16:16
RabidGravy speaking of generating stuff, which of the many and various template modules would be the best for generating Perl 6 code (basically class definitions)
arnsholt <3 find
nine jnthn: file locking would help :) I've already run into problems with parallel test runs precomping the same file. With a bit of handwaving on your side, I may be able to expose them myself.
TimToady is not entirely sure that breaking the language definition to allow JVM to "catch up" is really progress towards Christmas...
16:16 nightfrog joined
TimToady jnthn: if we precompile a module on Unix and then use it on Windows, it'll have the wrong \n 16:16
arnsholt No wait, 158. The 177 one isn't mine 16:17
ugexe mrf: cool. i know how draining writing the tests for all those can get
BooK masak, RabidGravy: irclog.perlgeek.de/perl6/search/?ni...&q=PPI is not very conclusive :-)
mrf ugexe: yeah. I may have to make some adjustments as well based on the \r\n changes. 16:18
BooK ok, I should have used Alias_, that gives more results
TimToady woke up grumpy, apparently...
mrf ugexe: I quite enjoy writting tests so its not a big thing
ugexe mrf: i dont think so. when i fixed it before i didn't have to change the tests
RabidGravy TimToady, grumpy is fine 16:19
16:19 dwarring left, z448 left
TimToady but I'm sneezy too 16:20
[Coke] TimToady: are you referring to the patch that lets the jvm build again, or something else?
mrf ugexe: hmm. Ok. My CRLF tests are all failing at the mo. I don't know whether i need to change the tests or the Grammar. Will runn them against your updated ABNF grammar and see if they pass there.
TimToady [Coke]: yes, though admittedly it's a very minor regression
mrf TimToady: much more and you'll have a full set
[Coke] well, it's no impact on moar.
and "being able to build jvm" is a step up from the previous state, which was "horribly broken for days"
TimToady true nuff, just grumbling 16:21
jnthn TimToady: Since pre-compilation can only reasonably be used as cache anyway, that's not really a problem. Unless you think we'll be able to do better at that in the future, but I'm unconvinced given the meta-programming stuff we want to support.
[Coke] hokay. My goals eventually include "no more #?if in rakudo source, should all be hidden in nqp" and "tickets for everything where jvm is cheating or currently fudged." 16:22
TimToady fair enough
TimToady is also a bit dopey yet
JimmyZ what about packing to exe code ...
RabidGravy TimToady, it's send Snow White down the shops and get something to sort that out then ;-) 16:23
16:23 advwp left, diana_olhovik_ left
JimmyZ I mwan packing .moarvm code to .exe code 16:24
*mean
jnthn JimmyZ: That'll work fine, but again, if it's a .exe you know precisely what platform you're on :)
BooK masak: irclog.perlgeek.de/perl6/2012-02-03 # some of the arguments are there and I seem to understand what Alias_ says :-)
JimmyZ oh, yeah. I forgot it
pmurias TimToady: getting stuff to build at all seem like a very good first step for bringing stuff up to spec ;) 16:29
16:29 z8 joined
TimToady oh, what would you know about that? ;) 16:30
pmurias++ on recent progress, btw 16:31
16:32 jkva left
BooK now I'm stuck in the backlog of February 2012-02-03 16:32
I guess I have to read it all up to now to find out if TimToady ever wrote the talk about "developing a language when your raw materials are hares, tortoises, Achilles, and Zeno."
AlexDaniel Hey! It seems like #126561 is a duplicate of #126037, perhaps it should be marked as such.
[Coke] AlexDaniel: checking... 16:33
TimToady BooK: we didn't rename Perl to Gloop, as a datapoint
[Coke] they're related anyway. I'll link them but not merge them. 16:34
AlexDaniel [Coke]: I'd say they are identical, but feel free to do whatever you like 16:35
16:35 softmoth_ joined 16:36 softmoth left
andreoss what do i use to trace code? something like perl -d:Trace? 16:38
BooK I'm looking at github.com/moritz/json/blob/master...iny.pm#L59 and I wonder if that gives a useful message about where the parsing stopped (I think not, but I haven't tried the actual code) 16:39
moritz BooK: ~ used to throw exceptioins when it couldn't find the closing quoting character 16:40
BooK: it doesn't anymore, so I guess the error messages now are pretty shitty
jnthn m: grammar G { token TOP { '(' ~ ')' a } }; G.parse('(a') 16:41
camelia ( no output )
jnthn Ah, forgot we decided not to have it do so in grammars by default...
16:42 cognominal joined
moritz star: use JSON; say from-json('{') 16:42
camelia star-m 2015.09: OUTPUT«===SORRY!===␤Could not find JSON in any of:␤ file#/home/camelia/.perl6/2015.09/lib␤ inst#/home/camelia/.perl6/2015.09␤ file#/home/camelia/star-2015.09/share/perl6/lib␤ file#/home/camelia/star-2015.09/share/perl6/vendor/lib␤ file#/home/camelia/…»
moritz star: use JSON::Tiny; say from-json('{')
camelia star-m 2015.09: OUTPUT«Input (1 characters) is not a valid JSON string␤ in sub from-json at /home/camelia/star-2015.09/share/perl6/lib/JSON/Tiny.pm:59␤ in block <unit> at /tmp/XLRKnNKdfb:1␤␤»
jnthn m: grammar G { token TOP { '(' ~ ')' a }; method FAILGOAL(|) { die "Parse fail near " ~ self.pos } }; G.parse('(a')
camelia rakudo-moar 76ee2d: OUTPUT«Parse fail near 2␤ in method FAILGOAL at /tmp/pX4JSI_f3n:1␤ in regex TOP at /tmp/pX4JSI_f3n:1␤ in block <unit> at /tmp/pX4JSI_f3n:1␤␤»
moritz m: grammar G { token TOP { '(' ~ ')' a }; method FAILGOAL(|c) { say c.perl } }; G.parse('(a') 16:44
camelia rakudo-moar 76ee2d: OUTPUT«\("')' ")␤P6opaque: no such attribute '$!pos'␤ in regex TOP at /tmp/qFWRsXxnYo:1␤ in block <unit> at /tmp/qFWRsXxnYo:1␤␤»
BooK so someone needs to write a FAILGOAL method for JSON::Tiny::Grammar?
TimToady we should have a parse option that reports highwater on failure to parse
moritz BooK: aye
BooK well, I have to start somewhere if I want to actually learn Perl6
16:44 rurban left
moritz BooK: preferably throwing a customer error class 16:44
BooK the X::JSON::Tiny::Invalid class? 16:45
moritz BooK: yes; I forgot it existed :-) 16:46
16:48 zakharyas left, z8 left 16:50 advwp joined 16:53 Oatmeal joined
BooK moritz: should it work with moar? 16:53
moritz BooK: what "it"? 16:58
JSON::Tiny should certainly work with rakudo-moar 17:00
mr_ron m: use newline :crlf; given open 'abc.txt', :w {.say: 'abc'; .close}; dd slurp 'abc.txt'
camelia rakudo-moar 76ee2d: OUTPUT«open is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub open at src/RESTRICTED.setting:9␤ in block <unit> at /tmp/eBEf5IHeyJ:1␤␤»
BooK so this is unexpected: Unhandled exception: Missing or wrong version of dependency 'gen/moar/stage2/QRegex.nqp' 17:01
moritz BooK: it means that your rakudo installation is borked
BooK weird
TimToady usually we only see that while make is running 17:02
but that's an nqp error
moritz BooK: or there's an old, precompiled version of JSON::Tiny somewhere that depends on an old rakudo version which you have since overwritten
17:02 skids left, sprocket joined
BooK ok, will dig deeper. thanks 17:03
17:04 skids joined 17:06 xfix joined
sprocket hello again, p6 17:07
regarding this bug i had submitted: rt.perl.org/Public/Bug/Display.html?id=126560
if i just go ahead and patch the POSIX module to fix up $*USER and $*GROUP as liz suggested, it that good by all?
RabidGravy sprocket, works for me (I'd been thinking about something like this previously.) 17:10
BTW, sprocket++ thanks for the RPi module
sprocket i suppose my question is more, “if $*USER/$*GROUP currently rely on running `id` to get get uid/gid, should it be fixed in rakudo instead to use getuid/getgid"
RabidGravy: no prob - it was fun to get it up and going, even if it’s painfully slow to run :) 17:11
i’ll be diving into profiling next to see where things can be sped up
RabidGravy as regarding the default implementation of $*USER, I think it would want to go in NQP if it's a NativeCall 17:12
or so I think the conversation went last time it was discussed
sprocket RabidGravy: i would agree - do you know when it was last discussed? i’ll go back and take a look 17:13
RabidGravy "more than two months ago" 17:14
ilmari m: .version.say for $*PERL.compiler, $*VM
camelia rakudo-moar 76ee2d: OUTPUT«v2015.10.208.g.76.ee.2.d.8␤v2015.10.61.g.624.d.504␤»
sprocket hah! ok :)
ilmari is amused how it breaks the git describe output into components
RabidGravy sprocket, if you blame the file in question, I think I made a patch around the ssame time
sprocket RabidGravy: will do - i’ll check out the history 17:15
before i made the effort to go and patch it against nqp, just wanted to make sure it’d fly
RabidGravy I'm not in the cross platform zealot camp, but if you could get some input from a windows kind of person it might help ;-) 17:16
sprocket at the moment, it looks like it doesn’t work on windows (i think?) anyways, but that would be good, yes ) 17:17
RabidGravy I'm not going to find and then boot the only windows computer in the house to find out ;-) 17:18
17:19 ab6tract joined
sprocket i’ll track someone down, or set up a VM :) 17:19
17:20 telex left
sprocket on the subject of profiling code, is there an existing module/program i can use? 17:20
Skarsnik --profile ? 17:21
sprocket Skarsnik: thanks - still relatively new to the p6 world 17:22
jdv79 has anyone written a deep clone with a "twiddles"?
17:22 telex joined
Skarsnik sprocket, well it's just provide stats on time spend somewhere and such. I don't know if it's totaly what you want x) 17:23
sprocket skarsnik: it’d be a start - i’ll take a look now
jdv79 doing that on a largely immutable obj tree turns out to not be simple.
at least as far as i've gleaned
17:24 z8 joined
lucs Question: fpaste.scsys.co.uk/501088 17:24
andreoss how do i pass optional un-named parametr to MAIN? 17:25
Skarsnik I will blame [$_]
lucs Skarsnik: Similar result with that line commented out. 17:26
andreoss anything but *@a? 17:27
Skarsnik I don't really know lucs, I assume line 9 is the substitution? 17:28
lucs Skarsnik: Yes.
17:28 jkva joined
zengargoyle is there any hope for the REPL to just break to the prompt on ^C vs complete exit? 17:28
17:30 jkva left
zengargoyle and/or is there some option to build with readline or some other library that keeps history? 17:32
lucs Skarsnik: Lines 1 and 2 are actually: "#!/opt/prj/l/perl6/.rakudobrew/bin/perl6\n\n"
17:33 diana_olhovik joined
zengargoyle yet again golfing something that scrolled out of screen buffer and needed ^C and then is gone :/ 17:33
17:35 uruwi joined
[Coke] docs.perl6.org/type/Enum is showing "sub exit" as a routine supplied by any. 17:35
m: enum q <1 2 3>; say q.invert 17:37
camelia rakudo-moar 76ee2d: OUTPUT«Use of uninitialized value of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at /tmp/HvC9hW1f4P:1␤Cannot call invert(q); none of these signatures match:␤ in block <unit> at…» 17:38
[Coke] invert is also listed as an Enum method. I assume that predates the move to enum being immutable pairs?
trying to figure out how to answer the perl6-users question of how to map a string input from the user to the enum value of the same name (without an EVAL, I hope) 17:40
lucs A bug to report perhaps?: fpaste.scsys.co.uk/501094 17:41
(and my code would be happier with some kind of work-around :/ ) 17:44
17:44 ab6tract left 17:47 ^elyse^ joined
jdv79 deep clone 17:49
17:49 z8 left 17:50 pmurias left
andreoss where the features of MAIN are described? 17:52
only S06?
[Coke] andreoss: design.perl6.org/S06.html#Declaring...subroutine is probably the best at the moment.
ilmari all these proto method foo (|) {*} in the setting means that misspelled named parameters silently get swallowed 17:53
is that intentional?
andreoss i wonder if i could have unnamed optional arguments to MAIN?
ilmari m: "å".encode("ascii") 17:54
camelia rakudo-moar 76ee2d: OUTPUT«Error encoding ASCII string: could not encode codepoint 229␤ in block <unit> at /tmp/3Yh16iKEcu:1␤␤»
ilmari m: "å".encode("ascii", :replacment)
camelia rakudo-moar 76ee2d: OUTPUT«Error encoding ASCII string: could not encode codepoint 229␤ in block <unit> at /tmp/3mv5E6Zwlr:1␤␤»
ilmari m: "å".encode("ascii", :replacement).say
camelia rakudo-moar 76ee2d: OUTPUT«Blob[uint8]:0x<3f>␤»
17:55 rindolf left
[Coke] jnthn: did you see the followup comment on 126553 ? 17:56
17:56 tokuhiro_ joined, mr_ron left
[Coke] andreoss: first, see if you can do it on a normal sub. 17:57
but: why would you want that? 17:58
jnthn [Coke]: I don't see anything added to rt.perl.org/Ticket/Display.html?id=126553 ?
17:59 colomon left 18:01 rindolf joined, tokuhiro_ left 18:04 lizmat left
[Coke] jnthn: huh, it's not on the ticket itself. there was an email reply from Taiki Kawakami 18:04
... oh. nevermind. 18:05
It took his email that long to make it through the spam filters.
[Coke] derps. 18:06
18:07 diana_olhovik left 18:12 mr_ron joined
lucs Please, am I misusing $0 or something, or is this a bug?: fpaste.scsys.co.uk/501094 18:15
18:16 duncan_dmg left
AlexDaniel lucs: it works for me 18:19
lucs :(
AlexDaniel lucs: what's your version of rakudo?
lucs Hmm... How can I tell?
[Coke] perl6 --version 18:20
AlexDaniel lucs: well, it seems like you are using rakudobrew. When was the last time I built it?
lucs This is perl6 version 2015.09-195-ge5a266b built on MoarVM version 2015.09-39-g1434283
AlexDaniel or that, yes
interesting. That's not too old
m: for 1 .. 3 { say "[$_]"; my $s = "123"; say "bef <$s>"; $s ~~ s/ (\d+) /$0$0/; say "aft <$s>"; } 18:22
camelia rakudo-moar 76ee2d: OUTPUT«[1]␤bef <123>␤aft <123123>␤[2]␤bef <123>␤aft <123123>␤[3]␤bef <123>␤aft <123123>␤»
lucs Can I somehow just upgrade my current installation, or do I need to reinstall from scratch?
AlexDaniel lucs: ./rakudobrew build moar 18:23
lucs: which is exactly what you did the first time
18:23 uruwi left
lucs AlexDaniel: Thanks, trying... 18:24
18:24 z8 joined
[Coke] m: say so "\r\n" ~~ /[";"|"\r\n"]/ 18:24
camelia rakudo-moar 76ee2d: OUTPUT«False␤»
[Coke] m: say so "\r\n" ~~ /"\r\n"/ 18:25
camelia rakudo-moar 76ee2d: OUTPUT«True␤»
18:25 z8 left
TimToady rakudobrew --help shows that build has options [tag|branch|sha-1] 18:25
[Coke] m: say so "\r\n" ~~ /";"|"\r\n"/
camelia rakudo-moar 76ee2d: OUTPUT«False␤»
[Coke] jnthn: ^^ RT #126573
18:25 lizmat joined
TimToady m: say so "\r\n" ~~ /";"||"\r\n"/ 18:26
camelia rakudo-moar 76ee2d: OUTPUT«True␤»
TimToady ltm bug
18:26 colomon joined
TimToady well, nfa 18:26
jdv79 any more acronyms? 18:27
TimToady ama?
jnthn
.oO( Needs Fixing, Argh )
[Coke]: "next week" :)
AlexDaniel jdv79: perhaps you can count his nickname as one?
lucs AlexDaniel: The rebuild fixed it. Thanks! 18:28
AlexDaniel lucs: Yeah! Great!
lucs: though recently there were some changes to \r\n handling. There is a chance that you will run into some issues 18:29
lucs I hadn't realized how easy it was to keep my version current.
AlexDaniel: Okay, thanks for the heads up.
[Coke] m: endprotoent 18:30
camelia rakudo-moar 76ee2d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/XIY4c3EZgZ␤Undeclared routine:␤ endprotoent used at line 1␤␤»
AlexDaniel lucs: if you run into some, then perhaps you can downgrade to 2015.10
lucs: it is newer than what you had but no \r\n stuff yet :)
[Coke] .tell SZABGAB cpansearch.perl.org/src/SZABGAB/Syn...e/Perl6.pm looks more like a Perl 5 highlighter, no?
yoleaux [Coke]: I'll pass your message to SZABGAB.
TimToady lizmat: the current split needles algo cannot be made lazy; did you consider doing a parallel search rather than serial? an additional advantage is that you could examine each character position only once, which might be faster 18:32
dalek ecs: 05f46f7 | (Geoffrey Broadwell)++ | S29-functions.pod:
Fix repeated 'will be thrown'
 18:33
TimToady also, what's the story on needle overlap?
lucs AlexDaniel: Using linux only, I'll probably avoid \r\n problems, but, noted anyway :) 18:34
lucs & # Lunch... 18:35
andreoss [Coke]: on normal subs i can do it with multies 18:36
[Coke] you can multi MAIN 18:37
andreoss with multi MAIN help message woudn't be as nice
18:40 sufrostico left
japhb [Coke]: In 8b6077968b361faea5bdeffe4d90b5619ecf6c80, did you intend to delete the line setting $!nl-out, or just include it in the #?if !jvm guard? 18:47
18:48 _mg_ joined 18:51 zakharyas joined
[Coke] Didn't mean to delete anything. 18:51
double checking my mischief... 18:52
18:52 TimToady left
[Coke] dammit. good catch. fixing. 18:52
18:52 pippo joined, yqt joined
dalek kudo/nom: ba34287 | coke++ | src/core/IO/Handle.pm:
Add line accidentally deleted during jvm triage.

  japhb++		 18:53
[Coke] ^^
japhb :-)
18:54 TimToady joined, leont joined
[Coke] thanks for the eyeballs 18:54
japhb np
18:56 virtualsue left 18:57 andreoss left, baest_ joined, Peter_R joined 18:58 brrt left 18:59 baest left 19:02 baest_ left, baest joined, baest left
mr_ron any windows programmers out there? I claim that under windows .say always outputs \r\n at the end of line regardless of 'use newline'. It looks like a possible bug. 19:07
m: given open 'abc.txt', :w {.say: 'abc'; .close}; dd slurp 'abc.txt'
camelia rakudo-moar 76ee2d: OUTPUT«open is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub open at src/RESTRICTED.setting:9␤ in block <unit> at /tmp/25LeKhF_aK:1␤␤»
mr_ron If you can run it under windows the dd outputs Str $var = "abc\r\n" 19:08
it doesn't seem to do it under linux
should say always output \r\n as a line terminator under windows?
Should the "say" routine always output \r\n as a final terminator under windows??? 19:14
RabidGravy well that is the line separator on windows yes
mr_ron m: use newline :lf; given open 'abc.txt', :w {.say: 'abc'; .close}; dd slurp 'abc.txt'; # say finishes with \r\n on windows here too ... 19:16
camelia rakudo-moar ba3428: OUTPUT«open is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub open at src/RESTRICTED.setting:9␤ in block <unit> at /tmp/g_r6NXopY1:1␤␤»
jnthn open should respect newline when picking nl-out, but doesn't yet
But yes, \r\n is the correct default 19:21
19:23 ChoHag left 19:24 ChoHag joined
mr_ron I think I understand. In my first example without use newline the behavior was ok but in my second example with use newline :lf the behavior is off. 19:24
19:27 spider-mario joined
mr_ron Sorry about not giving a good example. I was having trouble installing panda on windows because it uses .say to write a json file with \n line ends and later dies on the \r\n the .say adds at the end. Still working on problem including why \r\n seems to choke JSON::Fast. 19:29
lizmat .botsnack 19:30
yoleaux :D
RabidGravy yeah, I'd say it's the JSON parser - don't think it should care
lizmat TimToady: yes, I considered that, but felt that *that* would be trying to rewrite grammars 19:31
I mean, if you want NFA's, you should give it a regex :-)
19:31 _mg_ left 19:42 ^elyse^ left
AlexDaniel I wonder, is there any way to print random unicode symbol but excluding most of the boring stuff (letters, kanji, combiners and other stuff)? :) 19:47
m: say (0..0xFFF0).chrs.join('') ~~ m:g/<:NumericValue({+$_ > 0})>+/ # that question reminds me this 19:48
camelia rakudo-moar ba3428: OUTPUT«(｢123456789｣ ｢²³｣ ｢¹｣ ｢¼½¾｣ ｢١٢٣٤٥٦٧٨٩｣ ｢۱۲۳۴۵۶۷۸۹｣ ｢߁߂߃߄߅߆߇߈߉｣ ｢१२३४५६७८९｣ ｢১২৩৪৫৬৭৮৯｣ ｢৴৵৶৷৸৹｣ ｢੧੨੩੪੫੬੭੮੯ੰੱ…»
19:48 nick25 joined 19:50 rindolf left, kmel joined 19:52 nick25 left
kmel m: 'salut les filles'.say; 19:52
camelia rakudo-moar ba3428: OUTPUT«salut les filles␤»
kmel ;0
19:56 Begi1115 joined 20:00 espadrine left
kmel where's everyone? 20:02
helloooooooo
*waves hand*
bartolin_ o/ hello kmel 20:03
AlexDaniel o/
kmel hi bartolin_ AlexDaniel 20:04
dalek ast: 818ec3d | usev6++ | S02-names/pseudo.t:
Tests for RT #126523 do not die on Moar anymore
 20:07
20:12 raiph left
RabidGravy Hmm 20:15
if I have a module that exports some subroutine, and I use that module in another module can I simply re-export the subroutine from the second module? 20:17
lizmat not sure whether the "is export" trait is transitive 20:22
20:24 _mg_ joined 20:26 darutoko left
dalek kudo/nom: d273e91 | ugexe++ | src/core/Buf.pm:
Fix `$Blob1 ~= $Blob2` on jvm

Also fixes `IO::Socket::INET` on jvm
Resolves RT126529		 20:26
kudo/nom: 5fda5d2 | lizmat++ | src/core/Buf.pm:
Merge pull request #577 from ugexe/patch-7

Fix `$Blob1 ~= $Blob2` on jvm
20:30 azawawi joined
azawawi hi 20:30
20:31 nightfrog is now known as culb
stmuk jvm? 20:31
:O
20:32 azawawi left
geekosaur m-j: "hi".say 20:32
whoops 20:33
timotimo hello everyone
geekosaur r-j: "hi".say
camelia rakudo-jvm ba3428: OUTPUT«hi␤»
geekosaur derp
20:33 culb is now known as MiserySignals, MiserySignals is now known as culb
timotimo let's port moar to the jvm 20:35
20:38 abaugher left 20:39 abaugher joined, firstdayonthejob joined
RabidGravy decides to put that aside and document 20:40
20:41 abaugher left 20:42 kaare_ left, abaugher joined 20:45 Erendis42 joined 20:50 sufrostico joined, sufrosti1o joined 20:57 eliasr left 21:01 ZoffixW joined
ZoffixW moritz, modules.perl.org builds fine on my 5.22 install. IO::Socket::SSL might be the cause, but it is listed in Build.PL, so I would think it's installed. 21:01
Erendis42 i love you too, honey 21:03
diakopter !!
21:03 Erendis42 left
ZoffixW 0.0 21:03
21:08 _mg_ left
geekosaur gesundheit 21:08
21:10 softmoth_ left 21:11 woolfy left 21:12 leont left
ZoffixW starts commute home, but will read the log 21:12
21:12 ZoffixW left 21:19 ^elyse^ joined
moritz ZoffixW: ah, my bad; I simply ran the updater and waited for error messages 21:19
instead of checking Build.PL
so yes, it might be IO::Socket::SSL missing
ugexe r: say "a=b".split(/"="/,2) 21:28
camelia ( no output )
..rakudo-moar 5fda5d: OUTPUT«(a b)␤»
timotimo r: say "hi" 21:29
camelia rakudo-moar 5fda5d, rakudo-jvm ba3428: OUTPUT«hi␤»
21:30 ^elyse^ left
timotimo what can come after "say" that'd cause "no output"? there'd be at least a newline, no? 21:30
ugexe its really an unwind exception
timotimo oh, damn
21:31 xfix left 21:32 Begi1115 left
ugexe r: my @lines = "key=value","key2=value2"; say @lines>>.split(/"="/, 2); 21:32
camelia rakudo-jvm 5fda5d: OUTPUT«Use of uninitialized value $result of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at /tmp/tmpfile:1␤Use of uninitialized value $result of type Any in string context␤Any of…»
..rakudo-moar 5fda5d: OUTPUT«[(key value) (key2 value2)]␤»
lizmat ugexe: will look at that now 21:34
m: "key2=value2".split(/"="/,2).say
camelia rakudo-moar 5fda5d: OUTPUT«(key2 value2)␤»
21:36 uruwi joined
lizmat m: ("key2=value2",).split(/"="/,2).say 21:36
camelia rakudo-moar 5fda5d: OUTPUT«(key2 value2)␤»
lizmat m: ("key2=value2",)>>.split(/"="/,2).say
camelia rakudo-moar 5fda5d: OUTPUT«((key2 value2))␤»
lizmat m: ("key=value","key2=value2")>>.split(/"="/,2).say 21:37
camelia rakudo-moar 5fda5d: OUTPUT«((key value) (key2 value2))␤»
lizmat m: say ("key=value","key2=value2")>>.split(/"="/,2)
camelia rakudo-moar 5fda5d: OUTPUT«((key value) (key2 value2))␤»
lizmat m: say (my @ = ("key=value","key2=value2"))>>.split(/"="/,2) 21:38
camelia rakudo-moar 5fda5d: OUTPUT«[(key value) (key2 value2)]␤»
lizmat ah, it's only on the JVM :-(
j: ("key2=value2",).split(/"="/,2).say
camelia ( no output )
lizmat j: say ("key2=value2",).split(/"="/,2) 21:39
camelia ( no output )
lizmat j: say ("key2=value2",)>>.split(/"="/,2)
camelia rakudo-jvm 5fda5d: OUTPUT«Use of uninitialized value $result of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at /tmp/fsucEr9kjf:1␤()␤»
ugexe i thought it was strange the bare split is an unwind exception but on the hyper it doesnt seem to get thrown
lizmat so this doesn't have anything to do with split, or does it ?
ugexe yes, split is an unwind exception 21:40
21:40 zakharyas left
lizmat and why is that ? 21:40
ugexe i dunno
21:40 Oatmeal left
lizmat well, I'm overhauling split atm to implement :k/:v.:kv,:p,skip-empty on all candidates 21:41
so it might well change, so I suggest not spending any time on JVM split until I'm done :)
21:42 vendethiel joined
diakopter m: say ("==")>>.split(/"="/,2) 21:42
camelia rakudo-moar 5fda5d: OUTPUT«(( =))␤»
diakopter o_O 21:43
oh, it's a limit of things to split into, including empty string <shrug> 21:44
21:44 raoulvdberge joined 21:46 muraiki left 21:53 Oatmeal joined 21:59 uruwi left
psch .split($re, $limit) calls .match(:ex), which throws the UnwindException internally 22:00
last i checked it was the try inside <if $multi> in Str.match 22:01
so, something about exception handling
22:02 kmel_ joined 22:03 skids left, kmel_ left
psch r: say "abbb" ~~ m:ex:x(2)/ab+/ # actually not only :ex, but :ex with :x is what unwinds 22:04
camelia rakudo-moar 5fda5d: OUTPUT«(｢abbb｣ ｢abb｣)␤»
( no output )
psch ...i think i'm gonna patch the evalserver to print those exceptions :S
22:05 kmel_ joined 22:06 kmel left
lizmat psch: ah, yeah, not looked inside .match just yet 22:06
22:08 kmel_ left 22:09 kmel joined
raydiak m: class C { multi method m () { "A" }; multi method m () { "B" } }; C.*m 22:10
camelia rakudo-moar 5fda5d: OUTPUT«Ambiguous call to 'm'; these signatures all match:␤:(C $: *%_)␤:(C $: *%_)␤ in block <unit> at /tmp/Uz81rcg0a8:1␤␤»
raydiak ^ should have worked, right? 22:11
psch m: class C { multi method m (:$a) { "A" }; multi method m (:$b) { "B" } }; say C.*m(:1a, :2b) # .* dispatch seems underimplemented 22:13
camelia rakudo-moar 5fda5d: OUTPUT«(A)␤»
22:13 mr_ron left
psch m: class C { multi method m (:$a) { "A" }; multi method m (:$b) { "B" } }; say C.+m(:1a, :2b) # .+ too 22:13
camelia rakudo-moar 5fda5d: OUTPUT«(A)␤»
psch raydiak: i don't know if it should allow ambiguous dispatch, but it seems to only call one method anyway 22:14
22:14 pmurias joined
raydiak this form works: 22:16
m: class B { multi method m () {"B"} }; class C is B { multi method m () {"C"} }; C.*m
camelia ( no output )
raydiak m: class B { multi method m () {"B"} }; class C is B { multi method m () {"C"} }; say C.*m
camelia rakudo-moar 5fda5d: OUTPUT«(C B)␤»
psch m: role R { multi method m { "R" }; class A does R { multi method m { "A" } }; say A.*m 22:21
camelia rakudo-moar 5fda5d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/RBiEysL4UR␤Cannot declare our-scoped class inside of a role␤(the scope inside of a role is generic, so there is no unambiguous␤package to install the symbol in)␤at /tmp/RBiEysL4UR:1␤------> 3{ multi …»
psch m: role R { multi method m { "R" } \; class A does R { multi method m { "A" } }; say A.*m 22:22
camelia rakudo-moar 5fda5d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1BN98DsN0O␤Strange text after block (missing semicolon or comma?)␤at /tmp/1BN98DsN0O:1␤------> 3role R { multi method m { "R" }7⏏5 \; class A does R { multi method m { "A␤ expecting any of:␤ …»
psch argh
m: role R { multi method m { "R" } }; class A does R { multi method m { "A" } }; say A.*m
camelia rakudo-moar 5fda5d: OUTPUT«Ambiguous call to 'm'; these signatures all match:␤:(A $: *%_)␤:(A $: *%_)␤ in block <unit> at /tmp/bM9Pgw8k6G:1␤␤»
psch m: role R { multi method m (:$a) { "R" } }; class A does R { multi method m { "A" } }; say A.*m
camelia rakudo-moar 5fda5d: OUTPUT«(R)␤»
Skarsnik that pretty mess up x) 22:23
22:23 vendethiel left, raiph joined, MadcapJake joined
ugexe named argument isnt required 22:24
22:24 Zoffix joined, Ven joined 22:25 kmel left
Skarsnik that still should not create ambiquous call like this 22:25
ugexe order matters
psch notable is that .* should call all fitting methods 22:27
(actually, i'm not sure it's limited to "fitting", but i think that probably makes sense to assume...)
Skarsnik probably say cut to the first one?
ugexe i could have swore .* worked not long ago
22:27 MadcapJake left
psch ugexe: it works vertically, apparently. see raydiak++'s example a bit earlier 22:28
Skarsnik worked here
<raydiak> m: class B { multi method m () {"B"} }; class C is B { multi method m () {"C"} }; say C.*m
<camelia> rakudo-moar 5fda5d: OUTPUT«(C B)␤»
psch but it doesn't work among multis in the same class or across class and role
ugexe ah
RabidGravy you learn something new every day
22:30 Ven left 22:36 MadcapJake joined
AlexDaniel m: say 5 × 20; 22:42
camelia rakudo-moar 5fda5d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/6hvnVFlMJc␤Confused␤at /tmp/6hvnVFlMJc:1␤------> 3say 57⏏5 × 20;␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement modifi…»
AlexDaniel m: say 5 · 20;
camelia rakudo-moar 5fda5d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/A4prr7SuJ9␤Confused␤at /tmp/A4prr7SuJ9:1␤------> 3say 57⏏5 · 20;␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement modifi…»
AlexDaniel m: say 5 * 20;
camelia rakudo-moar 5fda5d: OUTPUT«100␤»
AlexDaniel
.oO( not sure why star means multiplication )
 22:43
22:44 tokuhiro_ joined 22:48 tokuhiro_ left 22:49 rurban joined
RabidGravy it always has done 22:53
AlexDaniel RabidGravy: :) 22:54
m: my @a = 4, 8, 15, 16; @a grep= * < 10; 22:55
camelia rakudo-moar 5fda5d: OUTPUT«5===SORRY!5=== Error while compiling /tmp/FrRBETrFEn␤Two terms in a row␤at /tmp/FrRBETrFEn:1␤------> 3my @a = 4, 8, 15, 16; @a7⏏5 grep= * < 10;␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ …»
AlexDaniel is there any way to do that without typing @a = @a.grep… ?
22:56 vendethiel joined
dalek href="https://modules.perl6.org:">modules.perl6.org: a9d5a49 | (Zoffix Znet)++ | web/lib/P6Project.pm:
Explicitly `use` IO::Socket::SSL

It is needed by Mojo::UserAgent for HTTPS, and ::UA is a bit shy reporting when the ::SSL is not installed		 22:56
RabidGravy lots, but they'd almost certainly involve more typing ;-) 22:57
psch m: my @a = 4, 8, 15, 16; @a .= grep: * < 10; # this is probably the shortest
camelia ( no output )
AlexDaniel psch: oh yeah!
bartolin_ r: my $foo = do { fail "oops" }; $foo.defined; say $foo # Failure is reported as handled
camelia ( no output )
bartolin_ oops 22:58
r: my $foo = do { fail "oops" }; with $foo { 42 }; say $foo
camelia ( no output )
RabidGravy I thought failures did that
i.e. checking them somehow makes them handled 22:59
bartolin_ hmm, locally the second evaluation is not reported as handled.
though it should be, because used with 'with'
22:59 pmurias left, Zoffix left
bartolin_ sorry, only on jvm it is not reported as handled in the 2nd example 23:00
psch bartolin_: that's probably the bandaid
RabidGravy o/ blame it on the bandaid o/ 23:01
bartolin_ psch: could be. but maybe that's behind some test failures
psch i guess it might also be whatever makes that bandaid neccessary...
i haven't yet looked closely at how we do all of this exceptional stuff 23:02
bartolin_ looks like 'method defined()' in 'src/core/Failure.pm' is not reached in some cases (where it should)
dalek kudo/nom: 774f0a3 | lizmat++ | src/core/Str.pm:
Implement :k/:v/:kv/:p/:skip-empty on Str.split

Well, all candidates except the Regex one, that will be for tomorrow		 23:03
ast: 815631e | lizmat++ | S32-str/split.t:
Remove two now bogues tests
 23:04
bartolin_ psch: as you said before, it feels like a lot of the jvm breakage is related to failure/exception handling 23:06
23:08 nys joined, jevin joined
bartolin_ r: sub oops { fail "oops" }; my $foo = oops(); with $foo { 42 }; say $foo 23:09
camelia rakudo-jvm 5fda5d: OUTPUT«oops␤ in sub oops at /tmp/tmpfile:1␤ in block <unit> at /tmp/tmpfile:1␤␤Actually thrown at:␤ in block <unit> at /tmp/tmpfile:1␤␤»
..rakudo-moar 5fda5d: OUTPUT«(HANDLED) oops␤ in sub oops at /tmp/tmpfile:1␤ in block <unit> at /tmp/tmpfile:1␤␤»
23:09 jevin_ left
bartolin_ r: sub oops { fail "oops" }; my $foo = oops(); $foo.defined; say $foo 23:09
camelia rakudo-{moar,jvm} 5fda5d: OUTPUT«(HANDLED) oops␤ in sub oops at /tmp/tmpfile:1␤ in block <unit> at /tmp/tmpfile:1␤␤»
bartolin_ ^^ that's what I wanted to see/show
RabidGravy in S22 it say "provides" is optional, but I thought the way that installers work means this is pretty much mandatory 23:10
psch i wonder how much of that is actually GLR-related and how much just brought it forward that it was wonky all along...
jdv79 its not opt
psch s/much just/much the GLR just/
RabidGravy jdv79, yeah that's what I thought 23:11
jdv79 as far as cpan is concerned
yeah
RabidGravy fixes that
dalek ecs: c71fa23 | RabidGravy++ | S22-package-format.pod:
"provides' is Mandatory
 23:15
23:16 pippo left
RabidGravy it would be good if people working on the CompUnit stuff took a look at S22 'cause y'know .... 23:17
I'm interested because I'm writing a thing to test the META file in a distribution right now 23:18
23:19 sprocket left 23:20 sprocket joined
dalek ecs: 6d75e2d | lizmat++ | S15-unicode.pod:
Remove the idea of negative codepoints
 23:22
23:24 sprocket left 23:29 sprocket joined 23:34 pippo joined
lizmat good night, #perl6! 23:34
pippo Is there a way to do "for @a -> $first, @rest {...};" ?? 23:35
night lizmat
lizmat just guessing: for @a.shift, @a -> $first, @rest {...} 23:36
sleep&
23:36 skids joined
pippo thank you lizmat! 23:36
RabidGravy toodlepip
23:37 kid51 joined, vendethiel left
psch m: my @a = ^10; for $@a -> [$first, *@rest] { say $first; say @rest } 23:37
camelia rakudo-moar 774f0a: OUTPUT«0␤[1 2 3 4 5 6 7 8 9]␤»
pippo psch: Cool! TY! 23:38
I think it would be nice to put it in the perl6 documentation for the for control flow. 23:39
RabidGravy carry on then :)
23:40 sprocket left
pippo Hem. If someone can explain me how to do that. 23:40
psch doc.perl6.org/type/Signature#Destru...Parameters 23:41
pippo: ^^^ that's what's happening, it's already documented
gfldex m: my @a = 1..10; my $s = sub([my $first, my @rest]){ say $first, ':', @rest }; $s(@a);
camelia rakudo-moar 774f0a: OUTPUT«5===SORRY!5=== Error while compiling /tmp/DCZI7TXTD2␤Undeclared routine:␤ sub used at line 1␤␤»
dalek oblem_solver_tutorial: 9727a65 | lichtkind++ | chapter/text0.md:
extend Prerequisites
RabidGravy pippo, fork the github.com/perl6/doc clone your fork locally, make the change, push to your fork, make pull request 23:42
it involves github account and local git 23:43
pippo RabidGravy: Exciting. I'll try that.
RabidGravy I'm beered out now, bed calls 23:44
play nicely y'all
23:46 spider-mario left
pippo o/ 23:48
23:49 RabidGravy left 23:50 rurban left 23:55 pippo left

Powered by Cro and the Raku® Programming Language.

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

Thank you!