»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
Mouq m: X::AdHoc.new("foo").throw 00:00
camelia rakudo-moar d37353: OUTPUT«Default constructor for 'X::AdHoc' only takes named arguments␤ in method new at src/gen/m-CORE.setting:846␤ in block at /tmp/nzFXwrAX72:1␤␤»
Mouq m: X::AdHoc.new(:payload<"foo">).throw
camelia rakudo-moar d37353: OUTPUT«"foo"␤ in block at /tmp/nGyR99KsnZ:1␤␤»
Mouq Or that to get a backtrace
m: X::AdHoc.new(:payload<"foo">).throw; say 1 00:01
camelia rakudo-moar d37353: OUTPUT«"foo"␤ in block at /tmp/W9htYXLS9k:1␤␤»
Mouq m: X::AdHoc.new(:payload<"foo">).fail; say 42
camelia rakudo-moar d37353: OUTPUT«Unhandled exception: "foo"␤ at <unknown>:1 (/home/p6eval/rakudo-inst-1/languages/perl6/runtime/CORE.setting.moarvm:throw:4294967295)␤ from src/gen/m-CORE.setting:12649 (/home/p6eval/rakudo-inst-1/languages/perl6/runtime/CORE.setting.moarvm:sink:40)…»
Mouq Hm
Mouq isn't very good at this :(
TimToady p6: CONTROL { die "oops" }; return 00:04
camelia ( no output )
..niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: oops␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/tmpfile line 1 (ANON @ 5) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/tmpfile line 1 (mainline @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4595 (AN…»
00:04 rindolf left
TimToady niecza++ 00:05
that actually segfaults locally in rakudo
Mouq Wait, Rakudo has CONTROL... or at least, NQP does...
TimToady well, only r-m 00:06
(only r-m segfaults)
the others bypass the CONTROL and say Attempt to return outside of any Routine 00:07
Mouq And there are tests in Roast that presumably pass...
TimToady maybe it's a REPL thing 00:08
nope, segv's with -e too 00:09
Mouq is distracted
Mouq should probably sleep
jnthn Finally got the blog post done: 6guts.wordpress.com/2014/04/17/raci...ess-leads/ 00:10
jnthn accidentally the whole 3000 words
Mouq jnthn++ 00:11
00:12 zamolxes joined 00:14 benabik joined 00:19 benabik left
TimToady jnthn: did my $gerenuk := %animals; eat a <gerenuk> tag? 00:21
jnthn oh...maybe
TimToady: fixed, thanks 00:22
Mouq jnthn++, that certainly corrected my views of parallelization :) 00:28
jnthn Well, that's predicated on my sleep-deprived views on it being correct ;) 00:29
Mouq m: say [∪] await lines.map: {start { bag .words }} 00:30
jnthn oh, oops...I have to be at an important meeting in like 7 hours...
camelia rakudo-moar d37353: OUTPUT«(timeout)»
jnthn 'night
Mouq m: say [∪] lines.map: { bag .words }
'night jnthn!
TimToady o/
camelia rakudo-moar d37353: OUTPUT«(timeout)»
TimToady and yes, that would be a pretty nifty way to write a map/reduce :) 00:31
m: say [(+)] lines.map: { bag .words } 00:34
camelia rakudo-moar d37353: OUTPUT«(timeout)»
TimToady huh, what's so hard about that? it works locally...
Mouq Oh yeah, (+)/⊎ is what is wanted :)
TimToady m: say +lines 00:35
camelia rakudo-moar d37353: OUTPUT«(timeout)»
TimToady um...
Mouq m: .say for lines
m: 42
camelia rakudo-moar d37353: OUTPUT«(timeout)There were three men came out of the West␤Their fortunes for to try␤And these three men made a solemn vow␤John Barleycorn must die␤␤They've ploughed, they've sewn, they've harrowed him in␤Threw clouds upon his head␤And these three me…»
( no output )
TimToady m: say +$*IN.lines
camelia rakudo-moar d37353: OUTPUT«(timeout)» 00:36
TimToady huh
m: say lines[*-1]
camelia rakudo-moar d37353: OUTPUT«(timeout)»
TimToady how big is that song?
Mouq 49 lines :( 00:37
TimToady that behavior seems...suboptimal... 00:38
Mouq j: say +lines
camelia rakudo-jvm d37353: OUTPUT«0␤»
TimToady p: say lines[*-1]
camelia rakudo-parrot d37353: OUTPUT«Without a little Barleycorn␤»
TimToady 0 isn't exactly manifesting correctness either 00:39
j: say lines[*-1]
Mouq p: say +lines
camelia rakudo-jvm d37353: OUTPUT«Cannot use negative index -1 on List␤ in method gist at gen/jvm/CORE.setting:12584␤ in sub say at gen/jvm/CORE.setting:13497␤ in block at /tmp/lEbxtrre4q:1␤␤»
rakudo-parrot d37353: OUTPUT«49␤»
TimToady so lines is busticated on j
Mouq j: .say for lines
camelia ( no output )
TimToady maybe the eval server is not supplying stdin correctly 00:40
works fine locally 00:41
BenGoldberg m: .chars.say for lines 00:42
TimToady [∪] works fine as long as all args are bags
camelia rakudo-moar d37353: OUTPUT«(timeout)41␤25␤37␤24␤0␤55␤26␤37␤24␤0␤40␤35␤38␤22␤0␤42␤32␤46␤19␤0␤43␤26␤42␤28␤0␤43␤28␤48␤30␤0␤47␤26␤33␤23␤0␤43␤25␤48␤38␤0␤42␤28␤42␤32␤0␤35␤30␤43␤27␤0␤0␤0…»
TimToady ooh
Mouq BenGoldberg++
BenGoldberg j: .chars.say for lines 00:43
camelia ( no output )
TimToady guesses there's a non-blocking thing going on here in Moar 00:44
and eof is turning into 0 chars available
geekosaur that is pretty much what I concluded earlier 00:45
TimToady earlier++
Mouq $ perl6-j -e'say @*ARGS; .say for ([⊎] await @*ARGS».path.map: {start { bag .&slurp.words } if .f}).pairs.sort(*.value)' src/core/* 00:46
geekosaur jnthn++ was poking but had no conclusions, I think libuv ate him
Mouq The last few output are, unsurprisingly, {, }, my, method, =, multi, #, if, sub... 00:47
TimToady pity » has to reconstruct the order when the .map doesn't care 00:49
need some transitive race semantics there...
Mouq (of course, if I were really trying to optimize, a self-balancing tree would probably help more than parrellizing the files…) 00:51
*parallizing
I think...
*paralellizing
geekosaur parallelizing :p 00:52
TimToady is parallelyzed with fear
geekosaur wonders if anyone's coined that word to describe failed parallelization... 00:53
00:54 benabik joined
Mouq Just tried latest r-m with the one-liner above; The first say prints and then perl6 segfaults :( 00:57
Mouq is going to try sleeping, though 00:58
o/ #perl6
00:58 Mouq left 01:15 dayangkun_ joined 01:16 sjn_ left 01:19 dayangkun left 01:21 sergot_ joined 01:22 prammer_ joined 01:34 FROGGS_ joined 01:36 Sqirrel left 01:37 Sqirrel joined 01:38 FROGGS left
colomon panda failing tests in emmentaler?! :( 01:38
02:00 bcode_ joined 02:01 bcode is now known as Guest16748, Guest16748 left, bcode_ is now known as bcode, ggherdov left 02:04 ggherdov joined 02:11 djanatyn left, djanatyn joined, Alina-malina left 02:12 virtualsue joined, Alina-malina joined
lue I'd've thought parallelyzed (adj) : A multi-threaded application that is deadlocked :) 02:16
rjbs jnthn: I really liked your post on atomicity. 02:17
02:20 donaldh left 02:25 donaldh joined, bcode left 02:26 bcode joined 02:31 xragnar_ joined, xragnar is now known as Guest39371, Guest39371 left, xragnar_ is now known as xragnar 02:37 bcode_ joined, bcode left, bcode_ is now known as bcode
lue jnthn++ # making me realize that multithreading-ifying something involves a bit more than "put 'await do' in front of all the list-of-files loops" :) 02:50
02:51 virtualsue left 02:52 zz_nhayashi is now known as nhayashi 03:14 Alula_ joined
avuserow m: say "foo".encode("utf8"); say "foo".encode("utf8") ~~ Buf; 03:15
camelia rakudo-moar d37353: OUTPUT«Buf:0x<66 6f 6f>␤False␤»
avuserow m: say "foo".encode("utf8"); say "foo".encode("utf8").WHAT;
camelia rakudo-moar d37353: OUTPUT«Buf:0x<66 6f 6f>␤(utf8)␤»
avuserow is a bit puzzled
colomon how so? 03:16
oh, the last bit there? 03:17
avuserow admittedly I know very little about this but the Str docs imply that encoding returns something Buf-like?
looking at doc.perl6.org/type/Str#encode
colomon m: say utf8 ~~ Buf 03:18
camelia rakudo-moar d37353: OUTPUT«False␤»
avuserow I'm deep into yak shaving. Trying to fix up the openssl bindings so I can have a working MD5 on MoarVM so I can play with random web APIs that require it 03:19
colomon avuserow: I'm looking at the source, this looks like a weird Rakudo thing. (Unless it's a spec glitch.) 03:20
utf8 is a Blob[uint8]
Buf is also a Blob[uint8]
avuserow ah so maybe Blob is an okay type 03:21
lue ? I thought Buf was supposed to be a role, which would imply no parameterization
colomon errr… parameterized roles are a thing, and Buf is one
lue (no concrete-ifying of parameters, I should say)
colomon role Buf[::T = uint8] does Blob[T]
avuserow maybe I could just add some code to NativeCall to turn Bufs into CArray[uint8]... No clue how right that seems though. 03:22
colomon Or perhaps the real question here is why Blob.gist returns a string starting with "Buf"
avuserow: unfortunately I think most of the people with NativeCall experience are on European time, and so are asleep right now. 03:23
afk # time for bed
avuserow thanks for your help colomon :) 03:24
lue ♞ colomon o/
03:35 FROGGS_ left 03:41 water joined, hoverboard left, water is now known as hoverboard 03:49 FROGGS_ joined 03:52 hoverboard left 03:55 labster joined
[Coke] lizmat: I had no idea what the part(s) was. ensemble player; 5-ish different bits, 2 of which with a page or so of dialog. most surprising, however, was finding out I need to go in for the fight coach so someone can punch me in the head and then kick me in the face safely. 04:01
... beats writing cold fusion, though.
04:04 anaeem1_ joined 04:12 Teratogen left 04:18 djanatyn left, djanatyn joined 04:34 SamuraiJack__ joined 04:43 thou joined 04:59 dragon_d left 05:07 BenGoldberg left 05:10 punter joined 05:12 kaare__ joined 05:33 punter left
masak good normal-tz morning, #perl6 06:27
raydiak \o 06:29
06:32 SamuraiJack__ left, SamuraiJack joined
jnthn o/ #perl6 06:34
urgh, morning
lue++ # parallelyzed 06:35
raydiak morning jnthn 06:37
jnthn avuserow: A utf8 is a blob8 rather than a buf8 because a buf8 would be mutable, and the mutation might cause it to become invalid utf8, which is what the utf8 type is meant to imply it ain't 06:39
o/ raydiak
avuserow okay, that makes sense. I'm guessing that it makes the most sense to have algorithms like SHA and CRC32 take input as Blobs rather than Bufs or Strings in that case... 06:45
jnthn r: say Buf ~~ Blob 06:47
camelia rakudo-parrot d37353, rakudo-jvm d37353, rakudo-moar d37353: OUTPUT«True␤»
jnthn So yes, doing so doesn't harm the ability to take a Buf.
avuserow well, some of them do things like: multi sub foo(Blob $b) {foo($b.decode("utf8"));}; multi sub foo(Str $s) { # do actual work }; 06:49
and for these algorithms, it seems like it'd be better to do the work in the one that takes the blob instead of the string
jnthn For the ones you just described, yes, certainly.
06:54 raiph joined
jnthn commute & 06:56
06:57 FROGGS_ left, darutoko joined 07:06 Ven joined
dalek rlito: 6ead80f | (Flavio S. Glock)++ | src5/lib/Perlito5/Grammar/Regex5.pm:
Perlito5 - Perlito5::Grammar::Regex5 wip
 07:08
Ven jnthn++ # blag
07:08 zakharyas joined 07:15 klapperl joined 07:21 FROGGS joined
moritz jnthn: "Having spent a bunch of years working with and teaching about parallel, asynchronous, and parallel programming outside of the Perl world" -- I guess the duplication of 'parallel' isn't intentional 07:22
raydiak parallel parallels?
07:23 pecastro left 07:26 dmol joined
moritz sounds like Euclidean Geometry to me :-) 07:26
jnthn++ # blog 07:28
FROGGS ohh, I haz to read... 07:29
gmorning @all
raydiak \o FROGGS 07:30
Ven o/, *
07:31 thou left
raydiak hi Ven 07:31
07:33 beastd joined 07:48 SamuraiJack left
FROGGS jnthn: typo? CPS (Communicating Sequential Processes) 07:49
lizmat good *, #perl6! 07:54
and jnthn++ for an excellent, excellent blog post :-)
FROGGS if I also do jnthn++, do we have to lock him (away)? 08:03
moritz FROGGS: nah, the IRC servers serialize the lines that can contain ++
tadzik oh, the honk-manatee www.youtube.com/watch?v=s-33i2b17nc 08:06
08:09 FROGGS left
sergot_ morning! o/ 08:09
tadzik o/
lizmat o/ 08:10
Ven (when I git pull rakudo, I only need to make, right ?)
08:10 SamuraiJack joined
moritz Ven: if Configure.pl or something in tools/build/ changes, you need to reconfigure too 08:11
Ven noted, thanks
08:15 SamuraiJack left
Ven (actually, `make` never failed, it was `make install`) 08:16
lizmat does an immediate "make install" again work? 08:19
Ven Can't test with that user, I'll need the server owner to try it :p
Ven will fix his mess now 08:20
do I just add `is copy` for the 4 params (2 for sub + 2 for the method) or should I go with another strategy ? 08:21
moritz adding 'is copy' should be fine (just test it before you submit the pull request :-) 08:22
Ven can do that now 08:23
Ven ... expect that github broke down 08:27
lizmat afk for a few hours&
08:28 thou joined
nwc10 ooh, they're talking about PERL: github.com/blog/1819-write-line-no...your-phone 08:28
Ven Hahaha, that's some nice irony 08:29
nwc10 wonders how hard it would be to port Acme::Inline::PERL to Perl 6 08:32
tadzik they are?
08:33 thou left
nwc10 is joking. They don't mention any particular programming language 08:33
08:34 benabik left 08:35 FROGGS joined
raiph jnthn++ # I've always loved your posts but this is my favorite so far 08:38
08:39 kaare__ left
raiph (I mean your latest blog post, in case it isn't obvious!) 08:39
08:39 kaare__ joined 08:41 FROGGS left 08:42 dakkar joined, FROGGS joined 08:43 falk0n joined 08:47 FROGGS left 08:51 raiph left 08:54 pecastro joined 08:57 falk0n left 08:59 FROGGS joined 09:01 kaare__ left 09:03 rindolf joined 09:06 FROGGS left 09:13 falk0n joined, grondilu joined
Ulti btw have you guys seen Untrusted? alexnisnevich.github.io/untrusted/ 09:21
raydiak wow, that's...pretty awesome 09:31
09:33 sergot_ is now known as sergot
sergot Ulti++ 09:34
Ven didn't manage to get past lv21 ... 09:48
raydiak is up to 7 so far 09:51
eh, just realized the time...g'night #perl6 09:52
Ulti run 9 I waisted a lot of time 09:54
night raydiak
masak .oO( hula hoops are a big waiste of time ) 10:17
10:19 snarkyboojum joined 10:28 telex left 10:30 telex joined, telex left
Ven spent 15 minutes trying to find the bug before he realized the issue with "$" in a linuxterm 10:31
10:32 telex joined
Ven github.com/rakudo/rakudo/pull/264 10:34
Which makes me ask, now that I have a perl6 install and I'm not totally blind anymore, are there low-hanging fruits ? 10:35
10:38 denis_boyun joined 10:51 falk0n left 10:54 FROGGS joined
dalek kudo/nom: 88a084e | Nami-Doc++ | src/core/IO.pm:
Add , fix open :rw
 10:54
kudo/nom: a4e5caa | (Tobias Leich)++ | src/core/IO.pm:
Merge pull request #264 from Nami-Doc/fix-open-rw

Add , fix open :rw
10:55 hummeleB1 joined
FROGGS Ven: perhaps RT reveals some LHFs? 10:56
Ven FROGGS: marked as "easy" ? 10:57
eeh. Possible cross-site request forgery hehe
FROGGS I don't think that it is common practice to mark the tickets that way here...
:/
Ven can't actually understand how rt.perl works 10:58
or do you *need* an account to browse stuff ?
FROGGS Ven: if you want to get way deeper into the rabbit hole you might enjoy: github.com/rakudo/rakudo/blob/nom/docs/ROADMAP 10:59
Ven: no, you can read tickets without an account
Ven Ok, "use the public interface"
FROGGS about that roadmap, even when there is a nick assigned to something, it does not mean that the guy/girl is really working on that already
11:09 dmol left 11:10 dmol joined
jnthn afternoon o/ 11:12
masak \o 11:14
nwc10 danger, UGT violation. Imposter detected :-)
lizmat after* all!
FROGGS *g* 11:16
jnthn hey, I had my UGT morning earlier :P 11:18
rindolf jnthn: afternoon. 11:22
11:23 Ven left 11:27 clkao joined
dalek rlito: f13f6cf | (Flavio S. Glock)++ | / (2 files):
Perlito5 - Perlito5::Grammar::Regex5 wip
 11:32
11:41 ldthien0 joined
jnthn moritz, FROGGS: blog typos fixed, thanks! 11:43
FROGGS jnthn++ # that is a very ncie article btw 11:44
jnthn FROGGS: Thanks, hope it was informative :) 11:45
colomon jnthn: gist.github.com/colomon/10976822 ;) 11:46
FROGGS it is... not only the fact about the cpu cycles when it comes to shared memory
also the part about the implicit locks and about the solution at the end...
jnthn colomon: Cute!!
colomon good bit slower, I fear
jnthn colomon: Well, also, keeping the code simple keeps the post accessible to those less 6-skilled. :) 11:47
colomon sure
11:47 Ven joined
FROGGS Rakudo has been built and installed. # (r-m* that is) 11:48
Ven r: my ($a is readonly) = 5; say $a; $a = 42; say $a 11:51
11:51 dmol left
camelia rakudo-jvm a4e5ca: OUTPUT«(timeout)» 11:52
..rakudo-parrot a4e5ca, rakudo-moar a4e5ca: OUTPUT«5␤42␤»
Ven What are "logical cascades" ?
or "coercion types" ? 11:54
moritz iirc S02 defines coercion types 11:55
and for logical cascades, look for "andthen" and "orelse" in S03 11:56
Ven To coerce your return value, use a coercion type: sub get_pet() returns Pet(Animal) {...} # coerce any Animal to Pet
(do we have any reason to keep `returns` around, with `of` ?)
11:58 vimal2012 joined 12:01 guest40 joined 12:03 snarkyboojum left, vimal2012 left 12:05 thou joined, guest40 left
dalek rlito: 89b7aa0 | (Flavio S. Glock)++ | src5/lib/Perlito5/Grammar/Regex5.pm:
Perlito5 - Perlito5::Grammar::Regex5 wip
 12:05
jnthn .tell masak MoarVM release done, and at www.moarvm.org/releases/MoarVM-2014.04.tar.gz 12:06
yoleaux jnthn: I'll pass your message to masak.
12:09 thou left 12:14 xenoterracide_ joined 12:17 SamuraiJack joined
Ven Is there already modules that exposes grammars/roles with regexes for quoted strings ? 12:22
FROGGS Ven: what would be the use case? 12:23
Ven FROGGS: grammars dealing with quoted strings ? 12:24
instead of copy-pasting these lines :o
FROGGS hmmm, I still do not seem to get it 12:25
Ven FROGGS: I mean, rules to parse ``
'hey\' you'
(and the double-quoted equivalent)
FROGGS well, rakudo has parsers for quoting languages of course 12:26
Ven JSON::Tiny uses its own regexes for parsing "" quoted strings
FROGGS yes, because JSON has its own understanding what a q/qq-string is
Ven yeah, and I'd like to get the same rules in a module 12:27
for "simple, stupid" quoted strings
FROGGS so, you want to expose a grammar that would let one specify its own set of escape sequences... something like that?
Ven Not really escape sentences, really, just rules to parse correctly "a\"b\\" 12:28
12:28 molaf left
FROGGS Ven: there is nothing like that out there atm AFAIK 12:32
Ven Hmm, can we call p6 grammars "parser combinators" ? 12:33
lizmat Ven: wrt to my $a is readonly:
afaik the problem is currently that if we *do* mark $a as read-only, the assignment "=5" will not work
FROGGS m: grammar QString { token number { 23 } }; say "1234" ~~ /<QString::number>/ 12:34
camelia rakudo-moar a4e5ca: OUTPUT«｢23｣␤ QString::number => ｢23｣␤␤»
12:34 grondilu left
lizmat as the marking as read-only is compile time, and the assignment is runtime 12:34
FROGGS would perhaps be nice to have a collection of patterns for mail addresses and stuff
Ven FROGGS: yeah, exactly :_
lizmat Regexp::Common comes to mind ;-)
retupmoca perl6-debug is spitting out 'Unknown compilation input 'optimize'' anytime I try to use it on anything that's loading .pm6 files o.O
Ven can't quite :) on qwerty keyboards
tadzik :) is easier on colemak, 'cos the keys are next to each other 12:37
12:37 punter joined
masak good afternoon, dear #perl6ers. 12:39
yoleaux 12:06Z <jnthn> masak: MoarVM release done, and at www.moarvm.org/releases/MoarVM-2014.04.tar.gz
lizmat masak o/
and how's your jetlag?
masak excellent :>
I'll be away for a few whiles, but I'll be doing some Rakudo release management later today.
stand by for a stunning release. 12:40
&
Rounin Hm... The QWERTY keyboard apparently favours lefthanded typers, though, and I've found it's the easiest one to type "David" on
Important qualities for a keyboard layout
(My name is David) 12:41
tadzik Hello David
Rounin Hello um
hi
tadzik yes, tadzik is my name
Rounin Oh :D
Well, then hello, Tadzik :D
tadzik makes it easy to remember now that my irc nick is the same
Rounin Very true 12:42
Wow, if you love one-handed typing, "tadzik" is pretty easy on QWERTY :P
"tadz" + "ik"
We have an ISP called "Powertech"... "po" "wert" "e" "c" "h" 12:43
12:47 anaeem1_ left 12:48 kbaker_ joined
Ven I just noticed `is copy` in my command got executed -.- 12:48
s/command/commit message/ 12:50
12:51 dmol joined 12:54 xenoterracide_ left
moritz knows folks who only use passwords that can be typed one-handed 13:00
Juerd Every password can be typed one-handed. 13:03
There's a shift key on either side of the keyboard :P
huf if you can type it without your left hand, and type it without your right hand 13:04
13:04 stevan__ left
huf surely you can type it with no hands 13:04
Juerd Nose typing is hard 13:05
huf yeah, i know that too
Juerd typing-speed-test.aoeu.eu/ # 4 cpm 13:07
Ven Your score: 492 CPM (that is 98 WPM) # I can't do that on a Mac qwerty keyboard 13:10
Ven will try with his nose
Juerd: Your score: 46 CPM (that is 9 WPM) ! HA ! 13:12
13:12 guru joined 13:13 guru is now known as Guest90732, Guest90732 is now known as ajr_
Rounin :D 13:18
timotimo what a useful skill 13:23
tadzik Look, ma, no hands! 13:24
it seems like Y::E organizers want speakers to have slides ready almost 2 months ahead 13:26
crazy :o
Juerd Ven: Impressive. 13:28
tadzik: Solution: present without slides. Do live programming. 13:29
tadzik :) 13:31
that actually sounds more reasonable to me than doing slides more than 48 hours before the talk :D
in other news: yapc::na is expensive and call for papers is over :| 13:32
FROGGS jnthn: that is the current state of r* gist.github.com/FROGGS/0ea5537eb675588baaa2 13:34
timotimo no problems on moar \o/ 13:35
seems like it'll be time for moar to smoke all modules as well 13:37
tadzik wonders if task::star is up to date
FROGGS r: say Buf:0x<0c 30 03 26 74 22 a3 03>.decode('UTF-16') 13:40
camelia rakudo-parrot a4e5ca, rakudo-jvm a4e5ca, rakudo-moar a4e5ca: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Two terms in a row␤at /tmp/tmpfile:1␤------> say Buf⏏:0x<0c 30 03 26 74 22 a3 03>.decode('UTF␤ expecting any of:␤ …»
moritz m: say Buf.new(<0c 30 03 26 74 22 a3 03>.map({ :16($_) })).decode('UTF-16') 13:41
camelia rakudo-moar a4e5ca: OUTPUT«「☃≴Σ␤»
FROGGS j: say Buf.new(<0c 30 03 26 74 22 a3 03>.map({ :16($_) })).decode('UTF-16')
camelia rakudo-jvm a4e5ca: OUTPUT«0&t"£␤»
FROGGS p: say Buf.new(<0c 30 03 26 74 22 a3 03>.map({ :16($_) })).decode('UTF-16')
camelia rakudo-parrot a4e5ca: OUTPUT«「☃≴Σ␤»
jnthn FROGGS: Hmm, thanks. 13:42
13:46 fhelmberger left
FROGGS r: say Buf.new(0x03, 0x26).decode('UTF-16') 13:47
camelia rakudo-jvm a4e5ca: OUTPUT«&␤»
..rakudo-parrot a4e5ca, rakudo-moar a4e5ca: OUTPUT«☃␤»
FROGGS r: say utf8.new(0x03, 0x26).decode('UTF-8') 13:48
camelia rakudo-parrot a4e5ca, rakudo-jvm a4e5ca, rakudo-moar a4e5ca: OUTPUT«&␤»
FROGGS r: say nqp::decode(Buf.new(0x03, 0x26), 'utf16') 13:53
camelia rakudo-jvm a4e5ca: OUTPUT«&␤»
..rakudo-parrot a4e5ca, rakudo-moar a4e5ca: OUTPUT«☃␤»
13:53 thou joined, btyler joined, anaeem1 joined 13:54 haroldwu joined 13:57 thou left 13:59 haroldwu left, haroldwu joined
jnthn hm, 443cmp (89wpm) on my laptop keyboard :) 13:59
14:00 nhayashi is now known as zz_nhayashi 14:04 xragnar left
Ven jnthn: yeah, but what can you do with your nose :P ? 14:04
14:04 xragnar joined 14:06 thou joined
jnthn Ven: Sneeze? :P 14:09
Ven jnthn: how much WPM/CPM is that :P ? 14:10
tadzik depends on the intensity 14:11
Ven tadzik: did we just create another PhD subject again :P ?
tadzik heheh
maybe something for my masters.. ;P
Software Supported Selective Sneezing 14:12
14:12 treehug88 joined 14:13 anaeem1 left
masak Ven: grammars are usually not called "combinators", no. biggest reason is that they are not combinators. :P 14:15
a combinator is "a function whose argument is also a function". grammars are not that. 14:17
14:20 donaldh left
geekosaur Ven: parser combinators are a mechanism detail, not an essential part of parsing or lexing. in theory combinators are more flexible, but in practice keeping a language comprehensiblle by humans leads to highly formulaic combinators which can be folded into a framework instead 14:23
parser generators (e.g. yacc) do that, as does perl 6 by hiding it in what amounts to a generalization of the regex framework 14:24
14:24 donaldh joined
masak I think the big innovation of Perl 6 grammars is to *not* talk about lexing and parsing as two different phases. 14:26
sometimes it is, but not always. definitely not with Perl 6 itself. 14:27
geekosaur enh, most parser combinator systems also lose that distinction
masak interesting. did not know that.
in fact, grammars go even further and allow declaration-y stuff to happen at parse time.
geekosaur mostly that distinction goes along with parser generators and lexer generators, and that because they tend to be distinct domain-specific languages
there is no particular reason to separate those except efficiency; note that the rise of lex and yacc was driven specifically by the need for memory efficiency on pdp11-class machines 14:28
masak noted. 14:29
it's interesting to find that pattern in many parts of programming history.
geekosaur you can do lexing in yacc but the parser table will be much larger and use more working memory than the equivalent one generated by lex
and its DSL is more painful to work with at that level 14:30
masak we used to separate things for memory efficiency, but then that concern goes away and later generations remove the separation. (like the client/server separation in CVS/SVN, removed in Git)
14:35 bluescreen10 joined 14:39 donaldh left
cognominal I don't think parser combinators can deal easily with highly contextual grammars. Perl 6 has many mechanisms for that (slangs, dynamic variables...) 14:44
14:44 mtk joined 14:45 plobsing left 14:46 rindolf left 14:48 Ven left
geekosaur it depends on the combinator framework but it does require some extra work. 14:50
note that, if you're using parsec as the example of a combinator framework, its limitations come more from the monadic underpinning; monads have a certain inflexibility that can be avoided (at the cost of a different kind of inflexibility) by using an applicative-based parser framework instead 14:51
14:53 aindilis joined 15:06 denis_boyun left, rindolf joined
jnthn beer, curry & 15:10
15:11 guru joined, guru left 15:13 guru joined, guru is now known as Guest625, Guest625 left 15:14 ajr joined, ajr_ left, ajr is now known as ajr_
ajr_ Don't forget the major constraint on the size and complexity of programs; the limitations of the human brain. (E.g. magic # 7+/-2) The Unix philosophy of small programs that do one thing well contrasts with Microsoft's enormous all-embracing monoliths. 15:17
geekosaur note that in some ways that is a false dichotomy because you must also remember how to hook the little pieces together; many people have more trouble with that part and benefit from a Microsoft-style fixed framework 15:21
moritz ...until they get lost in the framework :-)
geekosaur ("useless use of cat" is arguably a symptom of this) 15:22
moritz I think the real difference is that if you look at excel, for example, you can reach many functions just by exploring the GUI
15:22 pecastro left
moritz and the UNIX approach lacks this explorability 15:22
geekosaur that is actually a large part of what I meant 15:23
moritz geekosaur: I actually often like "useless use of cat", because it makes data flow clearer
geekosaur the unix approach, you need to learn where all those little tools are and what they do and how to hook them together. excel can guide you through them
moritz now I wonder if it's possible to add discoverability to the UNIX approach 15:24
s/if/how/ 15:25
cognominal Pipe is a nice but limited form of combination. That's probably why I jumped to Perl.
moritz well, there are two limitations 15:27
one is that it only allows text
powershell solves that, I believe
the second is that the shell doesn't allow you to open several pipes to different process (at least I never managed to make it work)
Juerd moritz: There are shells like fish that improve discoverability a little 15:28
geekosaur you can do the multiple pipes thing but it's generally painful
the one that works best (and is most comprehensible generally) is the >(command) / <(command) mechanism in bash and zsh 15:29
ksh had a notion of "coprocesses" but they were difficult to keep track of
Juerd moritz: All you'd need to pipe to different processes is a forking tee(1) fork.
moritz Juerd: cmd1 | tee >(cmd2) | cmd3 ? 15:30
geekosaur zsh also has MULTIOS which is pretty powerful and flexible
Juerd moritz: More like tee -c -- cmd2
moritz Juerd: my tee doesn't have a -c option 15:31
Juerd moritz: Indeed. That's why I said "tee(1) fork"
Although it'd be easier to rewrite it in perl 15:32
Just don't call it pee
moritz tee >(cmd2) does seem to work though
Juerd But then it doesn't pipe to cmd3 anymore?
moritz why wouldn't it? 15:33
Juerd tee without a filename is just cat
moritz Juerd: >(cmd2) creates /dev/fd/42 file name
so it's called with a (pseudo-virtual) file name
Juerd cat | tee >(perl -ple's/^/1: /') | perl -ple's/^/2: /' 15:34
Doesn't work for me. Only 2: prints.
moritz seems cmd3 receives the output from both cmd1 and cmd2
$ echo foo | tee >(tr o ua) | tr f g
goo
guu
moritz should go home 15:35
Juerd Hm, it does if I pipe to it
Hm, it does if I pipe to it with echo, but not if I pipe to it with cat?! 15:36
Oh, it buffers.
Suboptimal :)
geekosaur cat -u?
Juerd -u (ignored) 15:37
geekosaur gnuish? they came up with a different horrible hack
man stdbuf 15:38
15:38 kivutar joined 15:39 cognominal left, cognominal joined 15:40 zakharyas left 15:43 sqirrel_ joined, SamuraiJack left 15:48 kivutar left 15:53 raiph joined 16:02 kurahaupo joined 16:03 bluescreen100 joined 16:07 bluescreen10 left 16:09 denis_boyun joined
TimToady ven, lizmat: wrt "is readonly", that's why we have ::= instead, which marks readonly after binding 16:12
16:14 mtk left
perigrin :::= turns it into a domino set after binding 16:15
16:15 Rotwang joined
FROGGS looks more like straciatela ice on a stick 16:15
16:18 FROGGS[mobile] joined 16:20 dwarring left 16:23 FROGGS left 16:39 rindolf left 16:42 punter left
TimToady our treatment of failure and exceptions has become somewhat fragmented, I fear 16:47
p6: say (unlink "nosuch.file").WHAT; say $!
camelia rakudo-jvm a4e5ca: OUTPUT«(Bool)␤Nil␤»
..rakudo-parrot a4e5ca, rakudo-moar a4e5ca: OUTPUT«unlink is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:2␤ in sub unlink at src/RESTRICTED.setting:6␤ in block at /tmp/tmpfile:1␤␤»
..niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: unlink may not be used in safe mode␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (unlink @ 0) ␤ at /tmp/tmpfile line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4595 (ANON @ 3) ␤ at /home/p6eval/…»
16:47 rindolf joined
TimToady oh yeah, but locally I get (Bool) and Nil 16:47
no way to get an errno out of that, apparently
$! was supposed to reflect errno 16:48
and it should really be returning a failure, not Bool
ajr_ discoverability; "man man", "apropos x" 16:49
TimToady also, the out-of-bandness of $! wrt 'try' and 'orelse' is a problem currently
m: say ((try die "oops") orelse $_).WHAT 16:50
camelia rakudo-moar a4e5ca: OUTPUT«Nil␤»
TimToady that needs to capture the (X::AdHoc) somehow 16:51
TimToady is wondering why we don't just fail from the try, instead of forcing people to rely on an out-of-band (and potentially race conditionish) $! 16:52
I suppose the issue was whether the failure was considered handled or not, but mandatory out-of-band $! is pretty smelly 16:53
contrariwise, if we decide to stick with out-of-band-ness on try, we need to make 'orelse' treat $! robustly on the LHS 16:55
set it to 'no error' before evaluating the thing on the left, and capturing $! after evaluating it
but as it is, try doesn't work with orelse correctly 16:56
s/correctly/as intended/
16:56 ajr_ left 16:57 dakkar left
masak this looks interesting. www.greghendershott.com/fear-of-mac...index.html 16:59
in the HN discussion, one poster talks about "pass[ing] macros to functions". news.ycombinator.com/item?id=7601506 17:00
I haven't thought much about that. is that useful?
seems to me functions are at runtime, when (usually) most of the macros have already finished what they should do. 17:01
TimToady but we're running lots of functions and methods at compile time
masak though I guess it could actually be a BEGIN-time thing (a macro, for example) that passes a macro to a function.
right.
TimToady HOM! 17:02
17:02 kivutar joined
TimToady but from the ToC it looks like they're trying to get at the same sort of syntactic issues that I'm trying to get at with macros for different syntax categories 17:02
we need to be able to insert macros into any grammatical category and have correct expectations both before an dafter 17:04
TimToady avoids making the required mention of the freudian typo, "dafter"
masak :) 17:06
TimToady and "is parsed" could be viewed through a HOP lens
masak TimToady: I'm just back from a two-week (ish) vacation (ish). by the end of it, a weird and possibly interesting alternative take to Perl 6's macros emerged from an unexpected direction.
TimToady: I have no name for it yet, but let's call it "control flow builder API", because it pretty much captures what I see it aiming for. 17:07
TimToady: I'll try to develop the idea in writing in the next few days/weeks.
TimToady well, just remember we avoid a lot of explicit HOP programming in P6 because mere mortals run away screaming 17:08
'is parsed' is one of those not-quite-a-combinator things
17:09 treehug8_ joined
TimToady and if we're gonna turn ';' into a monad operator, we need to make sure that at least appears to be optional :) 17:09
17:10 thou left
masak sure, sure. 17:11
that's not where I'm heading with this. :)
17:11 treehug88 left
masak the igniting idea was something more like "if the atomic thing is a straight sequence of statements, what's an intuitive set of combinators that give us our current set of control flow constructs?" 17:12
as in, what primitive operations would be needed to build control flow graphs that we can already build with Perl 6 keywords? 17:13
those are probably a good starting point for building custom keywords, too. maybe not a complete set, but a good start.
TimToady all without reinventing contuations? :)
*continuations
masak I... don't know yet. that's one of the things I don't know. 17:14
one of the things that led me down this path was reinventing CPS (poorly). so I can't in all honestly answer "no" to that question. 17:15
TimToady well, continuations might just be one implementation technique for it
masak right.
you can emulate most things with a switch statement in a while loop, so should be fine :P
colomon pandabrew/bin/pandabrew panda install Shell::Command 17:16
===SORRY!===
Missing or wrong version of dependency 'lib/Shell/Command.pm'
:\
Rounin They implemented dependency hell... Perl 6 is going places 17:17
How about BSOD next 17:18
TimToady :D
Rounin ;)
17:18 FROGGS joined
TimToady maybe take down the world's financial infrastructure for an encore? :) 17:19
oh wait, that's a job for the SSL folks these days 17:20
FROGGS yay, looks like I fixed the jvm utf16 bug 17:21
Rounin :>
FROGGS btw, Java is odd
Rounin With regards to Unicode or just in general 17:22
FROGGS if (n % 2) { # expected Bool but got Integer?? really??
Rounin Ah yes, it requires explicit bool types
FROGGS and its understanding of unicode is like the wrong approach also :P
but perhaps it was a good idea when they decided how to do it 17:23
TimToady well, it's at least a 20-years-out-of-date approach
Rounin I think maybe everybody thought Unicode was going to be 16 bits wide when they chose UTF-16 for Java and Windows
TimToady "16 bits should be good enough for anyone!"
Rounin I'm glad everybody else chose UTF-8 though
TimToady well, for certain values of "everybody" 17:24
Rounin Haha yes
TimToady tends to get blank looks when he talks about Unicode in the Far East... 17:26
Rounin Every .java file should have the word "char" replaced by "DO NOT USE THIS" before compilation
And Char of course
Or whatever its boxed type is called
Character
17:26 thou joined, thou left
Rounin Haha, that's a shame... That would probably be where it's most usseful 17:27
You could talk about GB18030 instead and not mention Unicode :P
17:27 guru joined
Rounin "We convert it to uh... an internal format... before processing" 17:28
17:28 prammer_ is now known as prammer, guru is now known as ajr_
TimToady well, there's more to the Far East than just GB-land 17:28
Rounin True... And they all seem to have more than one encoding per country
EUC-KR, codepage 949.. or was it 948... Shift-JIS 1, 2, 3 and I've lost count... 17:29
One would think they'd beg for Unicode
colomon okay. If I pandabrew rebootstrap, I can install Shell::Command and it installs successfully 17:31
Rounin \o/ 17:32
colomon If I try running my code script (which does "use Shell::Command"), it fails with Missing or wrong version of dependency 'lib/Shell/Command.pm'
If I try to install Shell::Command again, that process ALSO fails with Missing or wrong version of dependency 'lib/Shell/Command.pm'
Rounin Could it just be that it's searching in the wrong base directory? 17:33
It doesn't say where "lib" is in the file system
17:34 xinming__ joined
colomon gist.github.com/colomon/11000133 # same command works first time, fails the second 17:36
17:37 xinming left 17:38 sqirrel_ left
colomon tadzik: ^^ 17:43
17:44 kivutar left
arnsholt You know it's CS when the build system for the thesis has several layers of Makefiles 17:45
moritz makefiles that generate makfiles that generate makefiles? 17:49
TimToady sounds smarter than configures that write configures that write configures... 17:52
17:54 rindolf left, mtk joined 17:57 SamuraiJack joined
arnsholt moritz: Not quite that bad (yet) 17:57
But I do have a whole directory dedicated to the figures for a chapter, and associated scripts for generating data files and plot files 17:58
17:58 btyler left
colomon 's $work has at least one makefile that fires off a Perl script which generates another makefile... 18:01
nwc10 I'm suspicious that we shoul actually change to do that in the perl 5 core, to exploit parallel makes 18:02
specifically for building man pages 18:03
18:05 treehug88 joined
vendethiel FROGGS: I'm actually one in favor of "force if to take bool" :-p 18:05
TimToady well, that's what we do in P6, it's just that most everything has a Bool meaning :) 18:06
hoelzro jnthn: wonderful post 18:07
18:07 treehug8_ left
hoelzro I'm still reading it, but I'm really enjoying it =) 18:07
18:07 btyler joined
vendethiel ... which I'd fix with a module (though that might break inner workings) 18:07
18:09 kivutar joined
tadzik colomon: please remind me an hour from now :) 18:09
colomon tadzik: I'll see what I can do
raydiak good morning #perl6 18:12
timotimo hello raydiak :) 18:13
did you have the chance to look at p6bench yet?
raydiak yes
my fork has proper "fastest" calculations
now what do I do with it?
timotimo pullrequest it and i'll merge
raydiak actually, I should test it with a larger dataset...do you want to do it, or can I d/l some data somewhere? 18:14
timotimo i can give you data, hold on
t.h8.lv/p6bench/timings.tar.gz 18:15
otherwise i can just pull your changes and test it locally as well
oh, another thing 18:16
i've previously tried to ensure that at least 3 data points are always available, no matter what
18:16 SamuraiJack left
timotimo in the other branch 18:16
i never got it to work, though
especially for the forest fire and parse-json benchmarks this is interesting 18:17
i need to re-build a working rakudo first :) 18:18
raydiak cyberuniverses.com/perl6-bench/ 18:20
looks like it works, at first glance
timotimo that's exactly how i had hoped it would look! :D
raydiak \o/ 18:21
PR to japhb or your repo?
timotimo japhb please
i think i can delete my repository 18:22
everything that was in there has already been merged
18:23 hoverboard joined, thou joined
dalek rl6-bench: 13f386a | raydiak++ | analyze:
Separate "fastest" rate for each X step on plot
 18:24
rl6-bench: e51d359 | raydiak++ | analyze:
Re-add global fastest comparison to plot under scale fastest
rl6-bench: 489e92e | timo++ | analyze:
Merge pull request #6 from raydiak/master

  "Fastest" plot info improvements
18:24 denis_boyun_ joined 18:25 denis_boyun left
raydiak another issue: initially drawing all the charts on the page on a slow enough computer can freeze for long enough that it asks if I want to kill the script 18:26
timotimo oh? wow. 18:27
i have no idea how you could break that up into pieces
though
if the computer is slow enough for *that* to break, they're probably not interested in running any rakudo anyway :P
raydiak wit's my computer :P 18:28
timotimo oops :|
raydiak haha
nothing comes to mind for fixing it unless we switch to different plotting lib or don't have all the plots next to each other on the same page...guess I'll just file that away in the back of my mind for now 18:30
I was thining about having just 1 chart, and then having controls for which test(s) it shows, because that would make the merge feature work too...but in that incarnation, you wouldn't be able to just scroll down 1 page and see all the test results 18:32
timotimo aye.
raydiak why is it important to have >= 3 data points? 18:33
timotimo because it looks bad otherwise :) 18:34
also, if you can come up with something clever to get rid of spikes farther to the left, that'd be pretty cool, too :)
raydiak ?
timotimo t.h8.lv/p6bench/2014-04-11-rakudo_opt.html - look at for example while_empty_native for rakudo-jvm or rakudo-parrot, there's big bumps to the left 18:35
same with while_bind
and there's sometimes bumps downward, too
postwhile_nil_native has a pretty mean bump for rakudo-parrot there
and loop_empty_native for the jvm
raydiak what causes those? 18:36
masak builds Rakudo
timotimo i suppose timing inaccuracy
hopefully not background tasks going insane :)
raydiak they seem to be bumps *up* tho? 18:37
timotimo sometimes up, sometimes down
might be errors during the "find out how long this test takes to compile and start up" time thingie 18:38
raydiak it's not run in a vm like a vps or something, is it?
timotimo o
nope, it's on my desktop computer
masak "Whenever you need to control the order of evaluation, macros are what you want." -- from the HN discussion. I like that rule of thumb.
timotimo i make a point to close as much background software as possible and then i don't touch the desktop at all any more
until it's done
raydiak well, we could throw the outliers out, but mysterious missing data points would look kinda weird too 18:39
lizmat jnthn: looking at the Supply.merge pb: it seems that with "on", the case with Supply => EnumMap is not working 18:45
masak hey, something looks/feels different with the "Stage parse:" reporting :)
18:45 sqirrel_ joined
lizmat It shows the stage while doing it 18:46
masak schweet.
lizmat this was changed some weeks ago, afaik
masak whoever++
TimToady I think we're still reporting stages that will always be 0 though 18:47
masak aye.
ok, let's do this.
FROGGS arnsholt++ # he did the stage print thingy
masak release process starting.
masak throws a lever for dramatic effect
TimToady <clunk!>
lizmat ketjing! 18:48
18:48 mtk left
masak looks like the last two releases were nicknamed with the .pm in their name. 18:49
I'm going to break that tradition, because it doesn't make much sense. 18:50
the tags are still .pm-less.
lizmat jnthn: if I rework SupplyOperations.zip to use the EnumMap syntax, it also fails (spectacularly, with a segfault)
18:51 rindolf joined, kaare__ joined
lizmat hates it she couldn't get Supply.merge to work before the release 18:51
masak I hear there's another release within the next month or so, though. 18:54
dalek kudo/nom: 8d7a505 | masak++ | docs/announce/2014.04.md:
add stubbed release announcement
masak no suggestions for .pm names, I presume? 18:55
[Coke] , after having a year long release cycle nearly at work on the last revision, is now trying to compensate by doing weekly releases.
masak proposes "Rio" 18:56
if I hear no counter-proposals, that's what we're getting.
lizmat well, the other day, we had an Echt.pm meeting, the first in ~5 years :-) 18:57
FROGGS can I ask a thing in Java what it is an instance of?
like, without knowing it?
masak lizmat: "Echt" it is. 18:58
lizmat :-)
masak FROGGS: what are you, a pervert? :P
FROGGS :/
Rounin Hmmmmm... 18:59
masak FROGGS: obj.class, IIRC.
18:59 darutoko left
FROGGS I have a buf8, which is an instanceof VMArrayInstance_u8 for example 18:59
Rounin docs.oracle.com/javase/7/docs/api/j...lass%28%29
FROGGS but buf16 is not an instanceof VMArrayInstance_u16...
dalek kudo/nom: ff24385 | masak++ | docs/announce/2014.04.md:
release, I christen thee "Echt"!
 19:01
19:02 dwarring joined
masak tools/update-tai-utc.pl is still full of warnings. 19:04
[Coke] echt? seems legit.
masak I'm just going to disregard it and assume there were no leap seconds. someone correct me if that's a risky/wrong assumption.
tadzik colomon: ok, looking at it now
colomon oooo, good timing. :) 19:05
tadzik ah, yes, known :(
colomon :(
tadzik panda should really rebuild itself when one of its dependencies are updated
I dunno if I have a ticket about it though
I can probably look at this tomorrow
colomon it's not just panda, though 19:06
everything that tries to use Shell::Command is borked
19:07 kaare__ left
colomon (not saying you can't look at it tomorrow, just that it's a bigger problem than panda being awkward 19:07
)
dalek kudo/nom: b39dd0b | masak++ | docs/announce/2014.04.md:
tentative list of changes in release
 19:08
masak as usual, it's good to have multiple pairs of eyes on this list.
anything missing? anything that's too trivial to include?
comments/suggestions welcome.
19:09 kivutar left, FROGGS[mobile] left
colomon tadzik: …. errr, or maybe it is just that. After rebootstrap, I can run my script after all. WTF? Could have sworn that didn't work this morning. 19:09
lue masak: looked at the wp page for leap second, none yet for 2014 it seems. 19:10
tadzik colomon: well, panda depends on the old version of shell-command (old precompiled version). If you update it, the binary compatibility goes to hell and panda bails out
masak lue: thanks, buddy.
tadzik I should really fix these sort of things
FROGGS masak: there are alos jvm+nativecall fixes
colomon tadzik: but definitely if I try to install Shell::Command, it appears to work, but now Shell::Command DOESN'T work any more 19:11
lizmat masak: I would remove the Bag.xkkv
lue (according to the chart, that question doesn't matter until June 30 anyway. So unless we missed the one from June 2012 (or Dec 2008, or...), rakudo should be good.)
lizmat or spell it correctly: .kxxv :-)
masak FROGGS: should they be in docs/ChangeLog, too?
tadzik colomon: hm, that's news
19:11 rindolf left
lizmat but prefer removal from announcement, as the name is still under debate 19:11
masak lizmat: understood; removing :) and fixing spelling in ChangeLog.
lizmat masak++
vendethiel .oO( why is masak removing that smiley from the ChangeLog ? ) 19:12
dalek kudo/nom: 3b20616 | masak++ | docs/ChangeLog:
fix ChangeLog typo
FROGGS masak: NativeCall is kinda important, so I would mention that NativeCall passes all its tests on jvm backend too
lizmat FROGGS++
dalek kudo/nom: 2916bff | masak++ | docs/announce/2014.04.md:
remove .xkkv from release announcement
masak FROGGS: if you add it to docs/ChangeLog, I'll add it to the release announcement. deal? :) 19:13
I'll buy a beer to whoever fixes whatever's wrong in tools/update-tai-utc.pl.
dalek kudo/nom: 3801758 | (Tobias Leich)++ | docs/ChangeLog:
add note about NativeCall passing all its tests
 19:14
FROGGS :o)
masak FROGGS++ 19:15
19:16 kaare__ joined
dalek kudo/nom: 3340712 | masak++ | docs/announce/2014.04.md:
copy NativeCall changelog entry to release announcement
 19:16
timotimo FROGGS: except moarvm 32bit 19:17
dalek kudo/nom: a0e7e89 | masak++ | docs/announce/2014.04.md:
I present to you the contributors

Hoopy froods, all of them.		 19:19
masak the release announcement is now "ready". I invite all people with eyes to review. kthx.
dalek kudo/nom: eaece08 | masak++ | docs/release_guide.pod:
append current release to end of release guide
 19:22
masak descending into NQP release sub-process. 19:23
awwaiid y'all are invited to the DC-Baltimore Perl Workshop + Hackathon on May 3-4, btw.
segomos i saw that place on the wire 19:24
lizmat awwaiid: too bad I can't make it: too close to the NLPW :-(
masak the rule in the NQP release guide tells me *not* to upgrade the Parrot dependency. 19:25
moritz masak: sounds correct to me
masak (it's currently 6.1.0)
dalek p: ccbebe3 | masak++ | tools/build/MOAR_REVISION:
bump moar dep to 2014.04

Thanks, jnthn++.		 19:26
PerlJam Has anyone tried building against parrot 6.3.0?
vendethiel PerlJam: I tried to build rakudo with parrot yesterday. Did not go well. 19:27
moritz vendethiel: with 6.1.0 or 6.3.0?
or something else entirely?
vendethiel moritz: no idea :-). Just cloned rakudo and used --gen-parrot 19:28
dalek p: bdb3259 | masak++ | VERSION:
bump VERSION to 2014.04
moritz vendethiel: that'll be 6.1.0
vendethiel (I cloned it literally 1 minute ago)
(I mean, 1 minute ago yesterday before building)
timotimo i ought to contribute more so i move up on that list again :P 19:30
colomon "NativeCall passes all its tests on jvm backend too"?!!! \o/ 19:34
lue I currently have 6.2.0, IIRC
lizmat timotimo: perhaps the author order logic should include MoarVM commits as well 19:35
that would put me lower on the list for sure :-) 19:36
masak um. problem. 19:38
gist.github.com/masak/11006944
halp.
lizmat doesn't ring a bell with me :-( 19:39
FROGGS install something that provides "jar"?
masak I'm on Mint. what should I install?
is this documented somewhere?
FROGGS dunno, type "jar" and look for the suggestions?
masak I have *javac*; how can I not have *jar*? 19:40
FROGGS it is something about java archives
I dunno
masak does `default-jdk` sound good?
FROGGS I'd say so 19:41
masak installing. 19:42
TimToady doesn't remember what he installed on mint 19:43
masak yeah, now it works. 19:44
FROGGS++
TimToady which reminds me I was gonna upgrade to Mint 16 19:45
FROGGS pheeew
TimToady since 15 isn't sending out patches any more (*cough* openssl *cough*)
lue TimToady: OpenSSL is preventing them from sending Mint 15 patches‽ I had no idea the bug was that serious :) 19:46
TimToady though seems a little silly when 17 is coming out so soon
19:50 perigrin joined
TimToady lue: I suppose that depends on whether they're using https to fetch the patches... 19:51
thou r: say $*OUT.t ?? 'yes' !! 'no' 19:59
camelia rakudo-moar b39dd0: OUTPUT«Cannot find method 'isatty': no method cache and no .^find_method␤ in method t at src/gen/m-CORE.setting:13750␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm b39dd0: OUTPUT«No such method 'isatty' for invocant of type 'BOOTIO'␤ in method t at gen/jvm/CORE.setting:13693␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-parrot b39dd0: OUTPUT«no␤»
19:59 gdey joined 20:01 kurahaupo left
masak building NQP. taking a loooong time. 20:02
20:03 sqirrel_ left 20:08 kurahaupo joined
TimToady
.oO(masak needs a stage coach)
 20:08
lue TimToady: ooc, do you have any objections/concerns to the number literals section I added to S15 recently? 20:10
TimToady well, some of it seems like a solution looking for a problem 20:11
20:11 kbaker_ left
TimToady circled number as big digits seems to fall into that category just a bit 20:12
lue TimToady: agreed, though that mainly comes from wanting to use a general "integral Number Value" rule to decide what digits are. 20:13
[Coke] perlcabal.org/syn/S15.html - 404
What's the blocker on getting that rendering? 20:14
timotimo TimToady: only the server side of HTTPS connections are vulnerable i thought?
lue (though whether :radix<> should support radices > 36 is of course debatable, and that's the only place where ㊿ would appear)
[Coke]: I'm working on the replacement for the current synopsis generator, which is what's needed to automate generation of S15.
TimToady it doesn't really seem to buy much over the current :radix[] notation 20:15
[Coke] can we not show the link to the 404 in the meantime?
retupmoca timotimo: if the client is using a bad version of openssl, it's vulnerable as well
timotimo oh crap
if the server is malevolent?
i have an older mint, too 20:16
TimToady timotimo: and I run a web server too
timotimo i wanted to wait for a new hard drive to arrive so i can have a fresh installation and the old one for disaster recovery
TimToady though I'm actually more worried about my DSL router, which uses https to program it, sign 20:18
*sigh even
lue TimToady: does Unicode define what characters belong to certain digit ranges? (E.g. 0..9 is arabic, ߀..߉ is NKo ...)
TimToady it distinguished valid decimal digits from other numeric values 20:19
or if you mean which block they're in, you can do that too
m: say uniprop('߉', 'block') 20:20
camelia rakudo-moar b39dd0: OUTPUT«NKo␤»
lue I don't think there are any non-decimal digit ranges, right?
TimToady nope
I mean, yup
masak just pushed NQP tag 2014.04 20:21
TimToady m: say uniprop('߉', 'decimal')
camelia rakudo-moar b39dd0: OUTPUT«Decimal␤»
TimToady m: say unimatch('߉', 'decimal') 20:22
camelia rakudo-moar b39dd0: OUTPUT«True␤»
TimToady m: say unimatch('Ⅵ', 'decimal') 20:23
camelia rakudo-moar b39dd0: OUTPUT«False␤»
masak I don't have the proper private ssh key on this computer it turns out. ++someone is welcome to put masak.org/carl/tmp/nqp-2014.04.tar.gz in the right place. (on [email@hidden.address] )
20:23 gtodd joined
lue The rule could change from "integral NV" to "Category Nd", though that leaves out certain 0..9,10,20,30,... sets in No. 20:23
PerlJam masak: done. 20:25
timotimo raydiak: when will you have time free for perl6-bench again? 20:26
lue ( :51<㊿⑫③> is a cool trick, but that's about it, so I don't mind losing it and letting a module implement it :) . ) 20:27
masak PerlJam: thank you.
this concludes the NQP release sub-process.
back to Rakudo's.
[Coke] that's not a cool trick, it's unreadable.
lue [Coke]: that's clearly :51[50, 12, 3] in Unicode Form™. :) 20:28
[Coke] would the literals allow pulling digits from different blocks together, like: 12③ == 123 ?
timotimo it looks pretty garbled in my terminal :P
flussence
.oO( weird how I've got fonts to display ㊿ and ③ in my terminal, but not ⑫... )
lue [Coke]: that's a natural consequence of Unicode not providing explicit numeric ranges (even if/when the spec switches to a "Nd category" rule, that'll still be possible.) 20:29
20:29 vendethiel left
masak +1 on losing that feature and letting a module implement it 20:30
[Coke] lue, the 12③ thing? I would say we wouldn't want to allow that sort of thing
dalek kudo/nom: 6e22eff | masak++ | tools/build/NQP_REVISION:
[release] bump NQP revision
 20:31
kudo/nom: 6f0b1d3 | masak++ | VERSION:
[release] bump VERSION
lue wonders if Unicode would implement numeric ranges as an explicit property, rather than something you have to figure out manually. 20:32
gtodd Hello perl6 ... :-) 20:33
20:34 vendethiel joined
lue [Coke]: that would involve constructing the ranges ourselves, like we do for .succ/.pred, but in the grammar instead of the setting. I'd much rather just let people mix them together, with the understanding that you might confuse people. 20:34
gtodd Q: about a 5-ism ... how to translate or think about it in perl6 ...
$num="2947341"; ($lowest_odd) = sort $num =~ /[13579]/g
masak gtodd! \o/ 20:35
PerlJam gtodd: $lowest_odd = 1; # there! ;)
gtodd $lowest_odd ends up as 1 ... this may be a form of context abuse
masak gtodd: you might want to have a look at .comb
r: $num = "2947341"; say for $num.comb(/<[13579]>/) 20:36
r: $num = "2947341"; .say for $num.comb(/<[13579]>/)
gtodd masak: it works on evens and when the lowest value is > 1 too ;-)
oh cool
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '$num' is not declared␤at /tmp/tmpfile:1␤------> $num⏏ = "2947341"; say for $num.comb(/<[13579␤ expecting an…»
rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '$num' is not declared␤at /tmp/tmpfile:1␤------> $num⏏ = "2947341"; .say for $num.comb(/<[1357␤ expecting an…»
masak r: my $num = "2947341"; .say for $num.comb(/<[13579]>/)
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«9␤7␤3␤1␤»
gtodd I have the perfect stackoverflow codegofl entry for this :-D
masak :)
gtodd I was amazed to see perl6 there and looking *very* spiffy and readable ! 20:37
PerlJam But don't use sort to find the smallest, just use .min
gtodd hmm yes strings are objects 20:38
timotimo r: .say for "2947341" ~~ m:ex/(\d+) { $0 !%% 2 }/
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«False␤»
timotimo shouldn't that give me all odd numbers that are in that string?
gtodd r: my$x="343277";say $x;if $x~~m/2|4|6|8/ {say "$/,",$/.from+1} else {say "-1,0"}
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«343277␤4,2␤» 20:39
gtodd that is what I want to fix :-)
timotimo like 29 2947 29473 2947341 47 473 47341 ... ?
20:40 vendethiel left 20:42 vendethiel joined
vendethiel timotimo: with that +; it should check if the whole is odd, no ? 20:43
timotimo yes
but it should backtrace and try from every position, because i put :ex there
vendethiel r: .say for "2947341" ~~ m:ex/[(\d) { $0 !%% 2 }]+/
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«False␤»
vendethiel can't perl6 anyway 20:44
20:44 kaare__ left
lue r: say "abcde" ~~ m:ex/<alpha>+/ 20:45
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«False␤»
timotimo did we break that?
if so, why didn't the tests catch it?
PerlJam looks like it's broken to me.
lue I'm not sure if :ex and/or :ov ever worked in rakudo, but I could be wrong.
timotimo r: .say for "foobarbaz" ~~ m:i:ex/<alpha>+/ 20:46
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«False␤»
timotimo r: .say for "foobarbaz" ~~ m:i/<alpha>+/
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«｢foobarbaz｣␤ alpha => ｢f｣␤ alpha => ｢o｣␤ alpha => ｢o｣␤ alpha => ｢b｣␤ alpha => ｢a｣␤ alpha => ｢r｣␤ alpha => ｢b｣␤ alpha => ｢a｣␤ alpha => ｢z｣␤␤»
timotimo r: .say for "foobarbaz" ~~ m:exhaustive/<alpha>+/
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«False␤»
timotimo r: .say for "foobarbaz" ~~ m:overlap/<alpha>+/
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«False␤»
timotimo o_O
github.com/perl6/roast/blob/master....t#L9-L150 - well, we have this 20:47
FROGGS masak: is it allowed to push to nom again?
timotimo it's not in our spectest.data
dalek ast: cbeb43b | (Elizabeth Mattijsen)++ | S17-concurrency/supply.t:
Tests Supply.map; Supply.merge breaks still

But found the reason. But no fix yet :-(		 20:48
FROGGS err, actually I want to push to nqp 20:49
timotimo has the release been tagged the right way this time?
masak FROGGS: allowed to push to either, but you won't make the release.
timotimo: please be more specific?
timotimo i think -a is the flag we want? 20:50
masak timotimo: -m implies -a
FROGGS masak: that is the point, I do not want to have it in the current release
timotimo there was something about the tags we've been making recently that was causing the tag not to properly work with "git describe" or something?
masak timotimo: so if you just follow the instructions in the release guide, everything should be fine.
timotimo: git describe by default doesn't see lightweight tags.
timotimo unfortunately i don't really remember :(
masak man, it's good to have Git at my fingertips. :) 20:51
$dayjob++
timotimo ah, so lightweight was the wrong thing and you're already doing it right if you have -m?
masak timotimo: an annotated tag is created by either -a, -m, or -s
timotimo: release guide says to use -m, which is a bit meaningless in our case, but doesn't hurt.
timotimo fair enough
masak timotimo: I used it (for NQP).
and will, for Rakudo.
still building from realclean. 20:52
timotimo good :)
FROGGS and I also see the tags for nqp+moar 20:53
20:53 LLamaRider joined
masak hm, the release guides even write -a explicitly. though that's redundant when -m is there. 20:53
vendethiel m: [*]1..3
camelia rakudo-moar eaece0: OUTPUT«===SORRY!=== Error while compiling /tmp/jUgWp0Wgoy␤Two terms in a row␤at /tmp/jUgWp0Wgoy:1␤------> [*]⏏1..3␤ expecting any of:␤ argument list␤ postfix␤ infix stopper␤ inf…»
FROGGS redundant is fine for releases me thinks
PerlJam redundancy++ :)
masak agreed. it doesn't hurt. 20:54
vendethiel m: say'foo'
camelia rakudo-moar eaece0: OUTPUT«===SORRY!=== Error while compiling /tmp/j_BXIajAea␤Two terms in a row␤at /tmp/j_BXIajAea:1␤------> say'foo⏏'␤ expecting any of:␤ argument list␤ postfix␤ infix stopper␤ in…»
masak vendethiel: spaces are good, m'kay? :) 20:56
vendethiel yes sir :-) 20:57
vendethiel just looks at codegolf p6 answers
timotimo i wonder if we should start signing our commits with gpg keys? :3 20:59
20:59 adu joined
gtodd jeez 20:59
dalek ecs: de83968 | lue++ | S15-unicode.pod:
[S15] Change what a numeric literal can be.

Instead of "integral Numeric Value", it's now "category Nd". This leaves out some potentially useful digit ranges in category No, which can be dealt with as needed.		 21:00
gtodd I had spent 6 months forgetting about it but now I have to have .comb :-|
must ... have
timotimo in perl5? :) 21:01
21:01 dayangkun_ left
lue ponders a proposal for Unicode to have a "digit set" property or similar, so Perl 6 wouldn't have to do it itself with, e.g., .succ :) 21:03
gtodd heh yeah ;-)
masak .oO( not having digit set kinda .succ ) 21:04
lue You can't use script or block properties for it. e.g.: ② and ㊿ are of different blocks, despite seeming related, yet 2 and ㊿ are of the same script, despite being wildly different :) 21:06
timotimo hmm 21:07
what piece of documentation on doc.perl6.org should i devote my attention to as soon as i'm in the doc improvement mood? 21:08
21:10 bluescreen100 left
dalek p: 41b67f4 | (Tobias Leich)++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
[jvm] fix decoding of bufs for bufs that do not match encoding bit width

When we want to decode a buf8 to utf16, we have to take two elems, assemble what Java thinks is a codepoint and append that to our StringBuilder. For buf32->utf16 we do the opposite and for buf16->utf16 we get away with appending the elems to the StringBuilder. Note: Since StringBuilder does not distinguish between utf16 and utf32, we treat them equally until we know better.		 21:15
gtodd r: $_='33392';$n=.comb(/<[2468]>/).min;/$n/;$/.from//-1,",$/"
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '$n' is not declared␤at /tmp/tmpfile:1␤------> $_='33392';$n⏏=.comb(/<[2468]>/).min;/$n/;$/.from//-1,␤ expec…»
gtodd k 21:16
r: $_="333992";$n=.comb(/<[2468]>/).min;/$n/;$/.from//-1,",$/"
camelia rakudo-parrot eaece0, rakudo-jvm eaece0, rakudo-moar eaece0: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Variable '$n' is not declared␤at /tmp/tmpfile:1␤------> $_="333992";$n⏏=.comb(/<[2468]>/).min;/$n/;$/.from//-1,␤ expe…»
gtodd :P
21:16 dayangkun_ joined
gtodd almost there ... works in the REPL ;-) cheers 21:17
21:17 LLamaRider left
gtodd goes away for awhile 21:17
21:17 adu left 21:18 dragon_d joined
FROGGS jnthn: the post release nqp patch fixes gist.github.com/FROGGS/0ea5537eb67...t-j-sh-L30 21:18
masak gtodd: it probably works in the REPL because there you had already declared 'my $n' 21:19
21:20 BenGoldberg joined
[Coke] mouq++ #ticket maint. 21:35
21:36 Sqirrel_ joined, Sqirrel left 21:40 denis_boyun_ left 21:42 kurahaupo left
timotimo who agrees with me that the exceptions in doc.perl6.org ought to go in their own section, preferably even after routines? 21:46
lue +1 21:48
21:49 kurahaupo joined 21:51 BenGoldberg left
timotimo interesting 21:52
index.pl in perl6/doc seems to rely on dir("lib") returning paths that need to have "lib/" prepended in order to work 21:53
21:54 treehug88 left 21:55 BenGoldberg joined
masak that has shifted a bit during the existence of &dir 21:55
timotimo p: .say for dir()
camelia rakudo-parrot 6f0b1d: OUTPUT«IO::Path<.lesshst>␤IO::Path<rakudo-star-2014.03.tar.gz>␤IO::Path<.viminfo>␤IO::Path<star-2014.03>␤IO::Path<.ghc>␤IO::Path<rakudo>␤IO::Path<rakudo-inst-2>␤IO::Path<rakudo-inst-1>␤IO::Path<nohup.out>␤IO::Path<std>␤IO::Path<.bash_history…»
timotimo p: .say for dir(".ghc/")
masak though I don't remember how many times or in which directions, or how recently.
camelia rakudo-parrot 6f0b1d: OUTPUT«IO::Path<.ghc/x86_64-linux-7.6.2>␤»
timotimo p6: .say for dir(".ghc/") 21:56
21:56 kurahaupo_mobile joined
camelia rakudo-jvm 6f0b1d: OUTPUT«Failed to get the directory contents of '.ghc/': nqp::opendir: unable to get a DirectoryStream␤ in block at gen/jvm/CORE.setting:13966␤ in method contents at gen/jvm/CORE.setting:13972␤ in sub dir at gen/jvm/CORE.setting:14002␤ in block at /tm…» 21:56
..niecza v24-109-g48a8de3: OUTPUT«Unhandled exception: dir may not be used in safe mode␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (dir @ 0) ␤ at /tmp/tmpfile line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4595 (ANON @ 3) ␤ at /home/p6eval/niecza…»
..rakudo-parrot 6f0b1d, rakudo-moar 6f0b1d: OUTPUT«IO::Path<.ghc/x86_64-linux-7.6.2>␤»
21:56 kurahaupo left
timotimo so at least parrot and moar agree on this 21:56
21:57 kurahaupo_mobile is now known as kurahaupo 21:58 Rotwang left
dalek c: 4db9132 | (Timo Paulssen)++ | htmlify.pl:
move all exceptions into their own category at the very end
 22:01
timotimo that's much cleaner IMO
masak make: *** No rule to make target `stresstest'. Stop. 22:03
indeed, there's p-stresstest, j-stresstest, m-stresstest, but no stresstest. 22:04
the release_guide is charmingly outdated on this point.
22:04 kurahaupo_mobile joined
masak hm, or the Makefile should have a target that combines those. 22:04
masak temporarily adds one
22:04 kurahaupo left, hummeleB1 left 22:07 kurahaupo joined
lue syngen is currently 10 minutes faster on feather than on my machine (≈30m vs. ≈20m) :) 22:07
22:09 kurahaupo_mobile left
timotimo what's the difference between those machines? 22:09
i didn't expect feather to be beefy 22:10
lue Well, it's not like my computer is the newest, so... :) 22:11
timotimo OK 22:13
lue For reference, it took me "1730 wallclock secs" to run m-spectest a short while ago. 22:14
timotimo how many concurent test jobs? 22:16
22:16 ajr_ left
timotimo how do i refer to a private attribute of a non-self instance of my own class again? 22:17
lue the line was just 'make m-spectest', I didn't think -j did anything
22:17 Mouq joined
Mouq o/ 22:17
lue \o
Mouq timotimo++ are you planning on generating a new doc.perl6.org? I'm all set up to right now
timotimo -j won't change that
Mouq if you're not 22:18
timotimo Mouq: i'm not, please do it for me :)
Mouq timotimo: Ok :)
lue timotimo: how do you change it then, ooc? 22:21
timotimo TEST_JOBS=n
lue so, -j to run multiple backends' tests, and TEST_JOBS to run each spectest in a concurrent fashion. :) 22:22
timotimo yup 22:23
lue Now that I know, it'll be interesting to see what TEST_JOBS does for me. 22:24
gtodd masak: doh ... seriously I can barley read the output in this IRC client :-)
22:24 adu joined
gtodd barely even 22:25
22:26 treehug88 joined, treehug88 left 22:29 treehug88 joined, treehug88 left
Mouq timotimo: Done :) 22:29
timotimo much better, aye?
22:30 treehug88 joined
masak gtodd: get a better IRC client? 22:31
22:32 colomon left
timotimo doc.perl6.org/language/objects#Role_Application - is the R1, R2 thing really the right way around? 22:32
masak masak.org/carl/tmp/rakudo-2014.04.tar.gz ready for upload to the appropriate area. PerlJam++, still there? 22:34
(and rakudo tagged) 22:35
timotimo: in `class C does R2 does R1 { }`, order doesn't matter between the roles. either there is a conflict or there isn't. 22:36
timotimo: having said that, I think writing it that way (essentially backwards) is confusing.
gtodd Mouq: I just commented on your SO codegolf ... I translated my perl5 with Masak and Camelia ;-) 22:37
lue It gives off an expansion feel though: C does R2 in the first case could be seen as something like C does (R2 does R1)
22:38 btyler left
timotimo OK 22:38
i'm about to patch that file, masak. so should i reverse R1 and R2 there?
masak IMO, yes. 22:39
anyone who is competent enough to manage the build system of Rakudo -- please add a 'stresstest' target. it is currently missing, but the release guide assumes it's present. 22:40
gtodd Mouq: not really astonished at how unreadable so many golfed languages are (errm "linenoise anyone? but idioms are like that) still though perl6 really stays very readable in golf mode
timotimo gtodd: or rather: it doesn't get much more unreadable ;)
Mouq gtodd: Thanks! :)
gtodd haha
Mouq m: $_="333992";/<{.comb(/<[2468]>/).min}>/;say $/.from//-1,",$/"
lue timotimo: I personally think the current order better conveys a sense of substitution, like I demonstrated above. But I guess it doesn't matter. 22:41
camelia rakudo-moar 6f0b1d: OUTPUT«5,2␤»
lue (would Door #3 be to mention that the order doesn't matter?)
dalek c: 91f0dca | (Timo Paulssen)++ | lib/objects.pod:
elaborate a bit on the objects tut.
masak I'm currently blocking on someone uploading the tarball.
timotimo feel free to review
22:41 colomon joined
timotimo (and Mouq, feel free to rebuild and re-upload) 22:41
gtodd Mouq: \o/
Mouq gtodd: That fails on all-odd, though
bbiab 22:42
gtodd Mouq: doh ... I think the one I put up worked but was quite long ... if else etc
and I just cargo coded ;-) 22:43
timotimo i can't access rakudo.org via ssh right now
masak: otherwise i'd've done it for you
gtodd .comb should have a unicode short cut ... .UNICODE_COMB_DRAWING 22:44
22:46 LLamaRider joined
dalek kudo/nom: 4b9d8c6 | lue++ | Configure.pl:
Add all-backends 'stresstest' target for Makefile.

Thanks to masak++ for noticing.		 22:46
masak lue++
I'm done with the Rakudo release, except I'm blocking on the upload. 22:47
still a bit jetlagged, and not really eager to stay up for much longer.
parting instructions:
(a) anyone who has the SSH keys, feel free to upload masak.org/carl/tmp/rakudo-2014.04.tar.gz to the appropriate place 22:48
timotimo and also the nqp tarball?
masak PerlJam++ uploaded that one already. 22:49
22:49 dmol left
masak (b) sending out the p6c announcement and updating the Wikipedia article -- *not* before (a) has been done -- can either be done by other people while I'm asleep, or I'll get to it first thing tomorrow morning, jetlag-time. 22:49
dalek c: 976bd5f | (Timo Paulssen)++ | lib/operators.pod:
typo: $++x instead of ++$x
 22:50
masak that's all, I think.
'night, #perl6
lue ♞ masak o/
timotimo gnite masak, and good work!
sleep well!
masak thanks, I most certainly will. 22:54
Mouq 'night masak!
timotimo: I'll regen doc in a bit
gtodd: Gimme a little while and I'll look at it 22:55
timotimo hold on
r: say 1 cmp 5
camelia rakudo-parrot 6f0b1d, rakudo-jvm 6f0b1d, rakudo-moar 6f0b1d: OUTPUT«Less␤»
timotimo r: say 5 cmp 5
r: say 7 cmp 5
camelia rakudo-parrot 6f0b1d, rakudo-jvm 6f0b1d, rakudo-moar 6f0b1d: OUTPUT«Same␤»
rakudo-parrot 6f0b1d, rakudo-jvm 6f0b1d, rakudo-moar 6f0b1d: OUTPUT«More␤»
timotimo the docs are outdated! :)
dalek c: 4082c42 | (Timo Paulssen)++ | lib/operators.pod:
increase/decrease have been renamed less/more.
 22:56
timotimo r: say 1, 2, 4, 8, 16 ... 8 23:01
camelia rakudo-parrot 6f0b1d, rakudo-jvm 6f0b1d, rakudo-moar 6f0b1d: OUTPUT«1 2 4 8␤»
23:04 hoverboard left
dalek c: f8540de | (Timo Paulssen)++ | lib/operators.pod:
... checks initial elements against endpoint, too
 23:05
lue Mouq: there's no such scheme specified (AFAIK), but I'd fully support syn: for the synopses (L<syn:S02>, L<syn:A12>, ...) 23:10
Mouq timotimo: That '$++x' should probably be '$x++' based on the comments 23:14
Yeah "postfix ++"
As in timotimo++ :}
*:)
timotimo er, it should? oops :)( 23:15
:)
Mouq lue: Oh, I think you're right that it's not in S26… I'll add it
timotimo there we go
dalek c: 417e4c4 | (Timo Paulssen)++ | lib/regexes.pod:
fix a typo'd formatting code
c: a6ff140 | (Timo Paulssen)++ | lib/operators.pod:
oops. this is the correct fix.
Mouq m: $_="33399";/<{.comb(/<[2468]>/).min}>/;say $/.from//-1,",$/"
camelia rakudo-moar 6f0b1d: OUTPUT«This type cannot unbox to a native string␤ in sub MAKE_REGEX at src/gen/m-CORE.setting:12959␤ in block at src/gen/m-CORE.setting:12897␤ in method reify at src/gen/m-CORE.setting:7686␤ in method reify at src/gen/m-CORE.setting:7549␤ in metho…»
Mouq m: $_="33399";/<{.comb(/<[2468]>/).min||Any}>/;say $/.from//-1,",$/" 23:17
camelia rakudo-moar 6f0b1d: OUTPUT«This type cannot unbox to a native string␤ in sub MAKE_REGEX at src/gen/m-CORE.setting:12959␤ in block at src/gen/m-CORE.setting:12897␤ in method reify at src/gen/m-CORE.setting:7686␤ in method reify at src/gen/m-CORE.setting:7549␤ in metho…»
Mouq m: "abcd"~~/<{Any}>/ # Arguably Rakudobug, though I'm guessing it's undefined behaviour
camelia ( no output )
Mouq Oh
timotimo m: say "!!!" ~~ / <[!]>+ /
camelia rakudo-moar 6f0b1d: OUTPUT«｢!!!｣␤␤»
timotimo m: say "!!!" ~~ / <[\!]>+ / 23:18
camelia rakudo-moar 6f0b1d: OUTPUT«｢!!!｣␤␤»
Mouq m: "abcd"~~/<{Nil}>/
camelia ( no output )
timotimo m: say "..!!!." ~~ / <[.]>+ /
camelia rakudo-moar 6f0b1d: OUTPUT«｢..｣␤␤»
Mouq m: say "abcd"~~/<{Nil}>/
camelia rakudo-moar 6f0b1d: OUTPUT«Nil␤»
Mouq :?
timotimo m: say "..!!!./\\" ~~ / <[/]>+ /
23:18 LLamaRider left
camelia rakudo-moar 6f0b1d: OUTPUT«｢/｣␤␤» 23:18
Mouq m: say "abcd"~~/<{+"14"}>/
camelia rakudo-moar 6f0b1d: OUTPUT«This type cannot unbox to a native string␤ in sub MAKE_REGEX at src/gen/m-CORE.setting:12959␤ in block at src/gen/m-CORE.setting:12897␤ in method reify at src/gen/m-CORE.setting:7686␤ in method reify at src/gen/m-CORE.setting:7549␤ in metho…»
timotimo i'm not 100% sure what exactly must and what must not be escaped between <[ ]>
23:18 adu left
Mouq m: say "abcd"~~/<{14}>/ 23:18
timotimo can somebody be un-lazy for me while i type up the rest of the section?
camelia rakudo-moar 6f0b1d: OUTPUT«This type cannot unbox to a native string␤ in sub MAKE_REGEX at src/gen/m-CORE.setting:12959␤ in block at src/gen/m-CORE.setting:12897␤ in method reify at src/gen/m-CORE.setting:7686␤ in method reify at src/gen/m-CORE.setting:7549␤ in metho…»
Mouq THAT is a definite Rakudobug
timotimo it should .Str it by itself? 23:19
m: say "abacabadabacaba" ~~ / <[ a .. c ]>+ / 23:20
camelia rakudo-moar 6f0b1d: OUTPUT«｢abacaba｣␤␤»
Mouq timotimo: "<nibble($¢.cursor_fresh( %*LANG<Q> ).tweak(:cc).unbalanced("]"))>" 23:21
timotimo Mouq -v # for tired timotimo 23:22
Mouq So... ']', '\', '-', and '..' I think
timotimo why -? 23:23
Mouq Though '..' shouldn't practically appear in a character class
timotimo: Backward incompatability
timotimo yeah, the second . would be redundant already
oh, right
it'll complain if it's in a weird spot
Mouq std: /<[a-z]>/
camelia std e347792: OUTPUT«===SORRY!===␤Unsupported use of - as character range; in Perl 6 please use .. (or \- if you mean a literal hyphen) at /tmp/HkzjQUSA6c line 1:␤------> /<[a-z⏏]>/␤Parse failed␤FAILED 00:01 123m␤»
Mouq m: $_="33399";/<{!.comb(/<[2468]>/).min}>/;say $/.from//-1,",$/" 23:24
camelia rakudo-moar 6f0b1d: OUTPUT«This type cannot unbox to a native string␤ in sub MAKE_REGEX at src/gen/m-CORE.setting:12959␤ in block at src/gen/m-CORE.setting:12897␤ in method reify at src/gen/m-CORE.setting:7686␤ in method reify at src/gen/m-CORE.setting:7549␤ in metho…»
Mouq Oops
m: $_="33399";/<{~.comb(/<[2468]>/).min}>/;say $/.from//-1,",$/"
camelia rakudo-moar 6f0b1d: OUTPUT«-1,␤»
Mouq m: $_="333994";/<{~.comb(/<[2468]>/).min}>/;say $/.from//-1,",$/"
camelia rakudo-moar 6f0b1d: OUTPUT«5,4␤»
timotimo std: /<[az-]>/
camelia std e347792: OUTPUT«ok 00:01 123m␤»
Mouq Looks good to me
gtodd++
thou This has been a pet peeve of mine for a while: github.com/rakudo/rakudo/pull/265 23:25
timotimo oh, nice 23:26
Mouq thou: thou =:= softmoth?
dalek c: 8fe2c15 | (Timo Paulssen)++ | lib/regexes.pod:
a bit on character ranges
 23:28
timotimo Mouq: i think i've had enough documenting for the moment - need to rest my wrists a bit
feel free to re-generate
(review welcome as always)
we've been talking about giving perl6.org AAAA records; who actually has the permissions to change the DNS entries? 23:30
Mouq timotimo++++ 23:32
timotimo i'm not sure i get any karma for that :P
Mouq you get karma for documentation 23:33
:)
timotimo mhm
Mouq ok
gtodd: Post updated
Actually, I'll say as much in the comments 23:34
lue I like how using TEST_JOBS gets me a shiny progress-bar-thing at the bottom of the screen :) 23:35
dalek kudo-star-daily: de9ae78 | coke++ | log/ (5 files):
today (automated commit)
 23:36
Mouq doc/htmlify.pl should really be caching
PerlJam Did anyone upload masak's rakudo release? 23:37
timotimo i don't think so, PerlJam 23:38
lue Mouq: syngen has a pretty dumb cacher, if you dare to attempt to read some really-needs-to-be-cleaned-up code :)
lue is somewhat in the process of said cleanup, actually.
PerlJam uploaded.
timotimo PerlJam++
not sure if the announcement should go out before moarvm.org has its download link updated, too? 23:39
23:40 adu joined
PerlJam how is moarvm.org updated exactly? 23:41
Mouq PerlJam: github.com/MoarVM/moarvm.org
timotimo www.dlugosz.com/Perl6/ - should perl6.org/specification continue to link to this page? 23:42
PerlJam Is someone cutting a moarvm release?
timotimo it is already cut, afaict
hm
but i don't see an url for a tarball yet
Mouq masak++++, totally missed that you did the release, and PerlJam++ for the upload 23:44
timotimo: doc updated 23:46
dalek href="https://perl6.org:">perl6.org: a051871 | (Timo Paulssen)++ | source/community/index.html:
fix a leaky anchor tag
timotimo who can update perl6.org from the sources?
jnthn evening o/ 23:48
lue timotimo: I don't see much point to link to that anymore myself.
Mouq o/ jnthn
lue \o jnthn
timotimo hey jnthn
about to upload & link a moarvm tarball, are you? :) 23:49
Mouq lue: I think planetaria.org/perl6 is good, especially for someone new to the community
23:49 adu left
lue whaddya know, TEST_JOBS=4 brings me down the hundreds 23:50
jnthn timotimo: Uh...I uploaded the 2014.04 moarvm earlier?
lue Mouq: I was talking www.dlugosz.com/Perl6/
Mouq lue: Ohhh
timotimo jnthn: didn't see a link to it; so i guess only the website needs updated?
seen szagab
!seen szagab
.seen szagab
yoleaux I haven't seen szagab around.
lue Mouq: your link doesn't even work :)
Mouq lue: ooc, is syngen yet concurrent?
jnthn timotimo: I .tell'd it to masak earlier today, and mentioned it on #moarvm
timotimo which of our bots currently tracks karma btw?
Mouq *planeteria :P 23:51
jnthn 14:52 < jnthn> www.moarvm.org/releases/MoarVM-2014.04.tar.gz
timotimo ah, i only searched on #moarvm
jnthn timotimo: Website update is todo, yes.
lue Mouq: I threw some awaits in there once, but it didn't help much, and jnthn's recent post made me realize it needs more thought put into it anyway :)
jnthn timotimo: Above was from #moarvm
lue (Also, code cleanup feels more important atm)
timotimo huh? weird. my backscroll search thingie didn't find it
lines that contain [off] won't come up in the irc log, right? 23:52
23:52 adu joined
Mouq lue: Ok :) I've been following you're commits but I haven't really taken a good look at the details of syngen, and was just curious 23:53
*your
timotimo except the second logger, the one that logs to colabti.org, does feature those.
oh, [off] isn't the right trigger for that apparently
lue If you look at the code, it seriously needs to be modularized by now :) 23:54
dalek href="https://perl6.org:">perl6.org: 4803e39 | (Timo Paulssen)++ | source/community/irc.html:
aloha is gone, mention ilogger2.
 23:55
href="https://perl6.org:">perl6.org: ea6c5ae | (Timo Paulssen)++ | source/community/index.html:
a positive note about mailing list traffic and response time.
23:55 beastd left
thou Mouq: Yes, thou == softmoth
(Sorry, got pulled away for a minute.)
lue would really love it if he could go two git commits without having to reinstall panda and assorted modules
23:56 xenoterracide_ joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!