»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.
FROGGS wow, what a mess 00:01
diakopter BenGoldberg: no, it's just the amount of significant digits provided in the code for pi 00:02
FROGGS BenGoldberg: if a certain precision of pi is specced, then it could be a n nieczabug
00:18 cognominal left 00:20 Ben_Goldberg joined 00:22 BenGoldberg left
TimToady n: say pi 00:23
camelia niecza v24-109-g48a8de3: OUTPUT«3.1415926535897931␤»
TimToady n: say pi.WHAT
camelia niecza v24-109-g48a8de3: OUTPUT«Num()␤»
TimToady well, nobody has done perlcabal.org/syn/S02.html#Allomorphic_Rats yet 00:29
timotimo yeah :( 00:30
jnthn timotimo: Weren't you looking for what to do next? ;) 00:32
timotimo enh :|
00:32 benabik left
timotimo for that i'd need to get the multiple inheritance thing going, no? 00:32
jnthn Well, we already ahve multiple inheritance
And RatStr is the easy case
'cus Rat doesn't have weird boxing stuff going on. 00:33
timotimo oh
well, that's fair
i don't think i'll get it in time for this release though.
jnthn uh, box target I mean.
That's fine :)
I'm happy enough if not too many more things land before the release :)
timotimo bitbucket has "sign in with github" and an "import ssh keys from github" button 00:37
also, there's a "clone from github" button 00:38
FROGGS nqp-m -e 'say("ABCabc123" ~~ /<+[C..Z]+:gc<Ll>>+/)'
Cabc
diakopter timotimo: that's cool
FROGGS I might add this after release... (I need to codegen for parrot and jvm) 00:39
timotimo what exactly did you fix there? 00:40
oh, now i get it, it's inside the < ... >
the :gc
FROGGS right 00:42
a colonpair, in theory 00:43
00:45 stevan_ joined
lue I just realized some people may want "use encoding" to specify the code file's encoding, not the encoding they wish to handle text in :/ . Something to mull over while I'm afk. 00:47
diakopter www.developermedia.com/wp-content/u...1H2014.pdf
2nd page
00:49 jnap left
jnthn diakopter: So...C is dead? :) 00:50
diakopter no.. this was a survey about technologies 00:51
jnthn The graph is a curious mix 00:52
00:52 benabik joined
diakopter (Java's not on there, but Eclipse is) 00:53
Mouq diakopter: I'm trying to find what the survey was, though. It could've just been a form of names wich checkboxed next to them
diakopter it's at the bottom 00:54
yeah, seems the case
Mouq *names with checkboxes
diakopter (the form
)
441 self-selected respondents mostly from among visitors to codeproject, which is very much .net/web centric
jnthn I got some survey recently that was so hip it didn't list any of the technologies I work with day to day aside from JavaScript :) 01:00
01:11 grondilu joined
grondilu sees on github.com/coke/perl6-roast-data/b..._rates.csv that moar has taken the lead?! Great news! 01:12
benabik I wonder why JVM regressed so much? 01:13
FROGGS grondilu: due to the fact that I messed up a huge test file for parrot and jvm
benabik FROGGS: let us have this moment. ;-)
FROGGS everything is back to normal tomorrow :o)
moar is then somewhere at >99%
grondilu ahh, ok. 01:15
01:20 jnap joined 01:24 jnap left 01:38 jnap joined 01:44 benabik left 01:53 psch left 01:56 xinming_ joined 01:59 btyler left 02:00 xinming left 02:06 FROGGS[mobile] left 02:09 grondilu left 02:13 araujo joined 02:33 rhr_ left 02:39 rhr joined 02:44 benabik joined
dalek href="https://perl6.org:">perl6.org: 5aabea9 | coke++ | source/index.html:
mention MoarVM
03:04
[Coke] hope that's not premature. 03:07
clsn p6: say <a b c>.WHAT; say <a <b> c>.WHAT; 03:16
camelia rakudo-parrot c5ca99, rakudo-jvm c5ca99, rakudo-moar c5ca99: OUTPUT«(Parcel)␤(List)␤»
..niecza v24-109-g48a8de3: OUTPUT«(Parcel)␤(Parcel)␤»
clsn Why Parcel and List I wonder. 03:17
03:17 bjz left, bjz joined
clsn p6: say qw[a b c].perl; say qw[a [b] c].perl; 03:19
camelia niecza v24-109-g48a8de3: OUTPUT«("a", "b", "c")␤("a", "[b]", "c")␤»
..rakudo-parrot c5ca99, rakudo-jvm c5ca99, rakudo-moar c5ca99: OUTPUT«("a", "b", "c")␤("a", "[b]", "c").list␤»
clsn It's having the (nested) delimiters in the qw. Makes it a list and not a parcel for some reason.
03:28 Alula left 03:31 Alula joined 03:39 mtj_ left
clsn p6: say rand.Rat.perl; say (9.5).Rat.perl 03:43
camelia rakudo-parrot c5ca99: OUTPUT«<659/1208>␤9.5␤»
..rakudo-moar c5ca99: OUTPUT«<375/922>␤9.5␤»
..rakudo-jvm c5ca99: OUTPUT«<53/1564>␤9.5␤»
..niecza v24-109-g48a8de3: OUTPUT«<450/769>␤9.5␤»
clsn Wonder why the two Rats get two different .perls (one with / and one as decimal). And the one with num/den shouldn't have <>, since that's not the perl input that produces them, right? I should probably take some of these little LHFs one of these days. 03:44
geekosaur Rat-s are specced with both those behaviors, IIRC 03:46
03:53 Biohazard left 03:56 Biohazard joined
diakopter TimToady: you seen this? github.com/tokuhirom/PSR 03:57
not to mention this github.com/tokuhirom/rockre 03:58
04:02 thou left 04:03 Biohazard left 04:05 Biohazard joined 04:25 SamuraiJack joined 04:26 Khisanth left 04:33 dayangkun joined 04:39 Khisanth joined 04:43 kaleem joined 04:46 alester joined 04:47 kaleem_ joined 04:49 kaleem left
Ben_Goldberg p6: my $r = pi.Rat; say $r.perl, $r, pi; 05:07
camelia niecza v24-109-g48a8de3: OUTPUT«<355/113>3.1415933.1415926535897931␤»
..rakudo-jvm c5ca99: OUTPUT«<355/113>3.1415933.141592653589793␤»
..rakudo-parrot c5ca99: OUTPUT«<355/113>3.1415933.14159265358979␤»
..rakudo-moar c5ca99: OUTPUT«<355/113>3.1415933.141593␤»
Ben_Goldberg wonders why conversion from Num to Rat truncates so much 05:08
p6: my $r = 3.1415926535897931.Rat; say $r.perl, $r, $r.denominator;
camelia rakudo-parrot c5ca99, rakudo-jvm c5ca99, rakudo-moar c5ca99, niecza v24-109-g48a8de3: OUTPUT«3.14159265358979313.141592653589793110000000000000000␤»
Ben_Goldberg p6: my $r = 3.1415926535897931.Rat; say join ' ', $r.perl, $r, $r.denominator;
camelia rakudo-parrot c5ca99, rakudo-jvm c5ca99, rakudo-moar c5ca99, niecza v24-109-g48a8de3: OUTPUT«3.1415926535897931 3.1415926535897931 10000000000000000␤» 05:09
Ben_Goldberg p6: my $r = 3.1415926535897931.Num.Rat; say join ' ', $r.perl, $r, $r.denominator;
camelia rakudo-parrot c5ca99, rakudo-jvm c5ca99, rakudo-moar c5ca99, niecza v24-109-g48a8de3: OUTPUT«<355/113> 3.141593 113␤»
05:09 alester left
TimToady diakopter: gee, viv converts P6 grammars into P5 grammars 05:12
05:18 Ben_Goldberg left 05:20 jnap left 05:27 [Sno] left 05:41 jeffreykegler joined
Mouq It's gonna look messed up until I drastically change the structure, but I think mouq.github.io/feast at least *looks* cool even the tables break sometimes and there's a lot of possible functionality missing 05:49
05:50 jnap joined
Mouq The name is meh, it's just what I used when I started working on it 05:53
06:01 zby_home_ joined, jnap left 06:04 bjz left, bjz joined 06:06 bjz left 06:18 jeffreykegler left 06:24 dayangkun left 06:32 [Sno] joined 06:37 dayangkun joined
moritz diakopter: it's not a matter of file size 06:44
06:45 AW3i joined 06:56 odoacre left 07:34 FROGGS left 07:46 Mouq left 07:47 daniel-s joined, daniel-s left 07:53 ivanshmakov joined 07:59 jnap joined 08:03 jnap left 08:09 Mouq joined 08:14 zakharyas joined 08:15 silug left 08:22 FROGGS joined, ribasushi left 08:25 ribasushi joined, LLamaRider joined 08:26 silug joined 08:35 pdcawley joined 08:39 daniel-s joined 08:41 Khisanth left 08:42 Khisanth joined 08:54 darutoko joined 08:55 bjz joined 09:00 jnap joined 09:03 Mouq left 09:04 jnap left 09:06 Khisanth left, araujo left, stevan_ left
FROGGS rakudo HEAD fail: gist.github.com/FROGGS/9677dc12d5a6cd38f73b 09:06
09:07 araujo joined
moritz eeks 09:07
09:11 Khisanth joined, stevan_ joined, drin_m left 09:13 Khisanth left, stevan_ left 09:14 Khisanth joined, pdcawley left, stevan_ joined, AW3i left 09:15 Vlavv` left 09:23 AW3i joined, dmol joined 09:26 dakkar joined 09:28 Vlavv` joined
FROGGS lue: about 'use encoding', I'd think that you would pass it like 'use Foo:enc<ucs2>' if you mean "modules" and 'use encoding "ucs2"' for strings/IO in your scope 09:30
09:35 pecastro joined
hoelzro morning #perl6 09:52
FROGGS o/( 09:54
moritz \o hoelzro
FROGGS -(
^--- sad cyclops
tadzik i.imgur.com/9jLBEsh.gif 09:55
somehow it seemed related 09:56
FROGGS O.<
tadzik: you are coming to Poznań too? :D 09:57
tadzik FROGGS: sure :)
FROGGS \o/
tadzik yay, meetup!
FROGGS sad that we have to wait until may :/
yeah
and it is awesome, it is just a bit more than 2.5 hours from berlin by train 09:58
tadzik I wonder how far from Warsaw by train..
FROGGS (though, I need about an hour from south berlin to that station)
tadzik yeah, it can be 2.5 hours too :)
hoelzro tadzik: where is that gif from? 09:59
FROGGS tadzik: 16 hours and one minute
tadzik hoelzro: uh, from the internet? P
hoelzro: I saw it on reddit a while back, and was able to google it just now
hoelzro ah ha
I thought it may have been a bug from a game or something =P
tadzik sth like "james bond opening get down gif"
it is a bug from a game :) 10:00
FROGGS lol, google says you travel via frankfurt >.<
10:00 fhelmberger joined
tadzik FROGGS: 16 hours what? :o 10:00
FROGGS from warsaw to poznań
(via frankfurt for some reason)
tadzik hoelzro: www.youtube.com/watch?v=DzwFpzdO3DY
10:01 jnap joined
hoelzro what the hell 10:01
tadzik it has a knowyourmeme article
hoelzro looks
tadzik FROGGS: surely not Frankfur Am Mein? 10:02
10:02 Mouq joined
FROGGS tadzik: no, Frankfurt (Oder), at least 10:02
tadzik but why 16 hours :o
with a 14 hour wait? 10:03
FROGGS warsaw to frankfurt is about 7 hours it seems
tadzik oh!
I didn't look at google :)
FROGGS I don't know otehr pages to look for trains outside of germany :o) 10:04
tadzik :)
FROGGS my connections shows up pretty accurate though
tadzik pkp.pl is the only train webpage I ever neded
it's a valley of tears sometimes, but it can be nice
last time I was going to Kraków, and Geothe instutute left free books inside the train 10:05
FROGGS well, I've got the problem that I cant read it :/
10:05 jnap left, grondilu joined
tadzik like a preview books, and I can bring it to the Geothe Institute and get a full-sized book of a German author 10:05
FROGGS a bit like matrix... the blue or the red button?
tadzik hehe
grondilu p6: say pi.FatRat.perl
camelia rakudo-parrot c5ca99, rakudo-jvm c5ca99, rakudo-moar c5ca99: OUTPUT«FatRat.new(355, 113)␤»
..niecza v24-109-g48a8de3: OUTPUT«FatRat.new(884279719003555, 281474976710656)␤»
tadzik FROGGS: the green one :P
www.rozklad-pkp.pl/bin/query.exe/pn? is what you care about
it's pretty confusing though, even in Polish 10:06
FROGGS ha, that is the same software as we use in berlin it seems :o)
tadzik the main page also says: Recommended: Train Delays
10:07 Mouq left
tadzik I don't think that people riding a train for 24 hours instead of 3 appreciate the humour 10:07
FROGGS ohh dear... I will bug you about train station namen in may then, okay?
tadzik sure
FROGGS names*
tadzik you're most likely interested in Poznań Główny
FROGGS it is rediculous that Poznań does not show up when you type Poznan 10:08
at least to me 10:09
tadzik haha
indeed
try "pozna" :P
FROGGS that worked, yes
dalek p: be6a8f1 | jnthn++ | t (2 files):
Bump to Moar 2014.01 in readiness for release.
10:10
10:11 AW3i left 10:20 erkan joined, erkan left, erkan joined 10:29 sqirrel joined
sergot masak: [Coke]: I meant passwords on others sites, if you have the same. :) 10:33
10:41 denisboyun joined 11:01 jnap joined 11:06 jnap left 11:09 daniel-s left
timotimo o/ 11:14
FROGGS o/ 11:20
dalek ast: 67c03a7 | (Tobias Leich)++ | / (2 files):
adjust expectation after off-by-one was fixed
11:23
11:33 mtj_ joined, mavcunha joined 11:35 erkan left 11:37 erkan joined 11:50 Mouq joined 11:55 Mouq left 12:02 jnap joined 12:07 jnap left
[Coke] sergot++ 12:12
12:22 atroxaper joined 12:37 atroxaper left, atroxaper joined 12:43 atroxaper left 12:45 atroxaper joined 12:50 kaleem_ left 12:59 rindolf joined 13:02 kaleem joined 13:03 jnap joined
FROGGS only one worrying this left: 13:04
t/spec/S32-io/IO-Socket-INET.t (Wstat: 0 Tests: 41 Failed: 21)
Failed tests: 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23
25, 27, 29, 31-32, 34, 36, 38, 40
timotimo that's for moarvm?
FROGGS no
parrot
timotimo oh!
that's not good
FROGGS I know
jnthn FROGGS: Does this show up in the daily runs? 13:05
FROGGS jnthn: yes, since I fixed the test file yesterday
it was skipped
jnthn Oh, darn...
FROGGS yeah 13:06
it was meant to be skipped for jvm, but was skipped for all ... for five months
jnthn So we have no idea when it started failing :( 13:07
FROGGS correct
jnthn otoh, if sockets were actually busted then we'd probably have heard about it...iirc, Panda relies on them to obtain its stuff. 13:08
13:08 jnap left
moritz and people have been developing Net::IRC 13:08
jnthn Busted as in "don't work at all"
Right
Doesn't mean nothing is wrong, though.
colomon jnthn: last I looked, panda just called curl (or something like that)
13:08 yves joined
FROGGS it times out I think 13:09
but just started now
colomon hmmm, Panda does have one call to IO::Socket
IO::Socket::INET 13:10
jnthn my $s = IO::Socket::INET.new(:host<feather.perl6.nl>, :port(3000)); 13:11
yeah
it's using it to get the module list
colomon is trying it now, but pandabrew insisted on rebootstraping, so it may take a while
also, I think I'm running it on JVM at the moment. ;) 13:12
um 13:13
FROGGS not ok 1 - [44s] echo server and client 13:15
# got: '===SORRY!===
# Could not find symbol '&INET'
# (timeout)
I am not sure I understand that
r: say IO::Socket::INET 13:16
colomon I can confirm that rakudo parrot panda updated projects.json I my machine just now without apparent issue
camelia rakudo-moar c5ca99: OUTPUT«===SORRY!===␤Could not find symbol '&INET'␤»
..rakudo-parrot c5ca99, rakudo-jvm c5ca99: OUTPUT«(IO::Socket::INET)␤»
FROGGS aha
jnthn No sockets in Moar yet 13:17
FROGGS I know, but the TAP I pasted is from perl6-p 13:18
jnthn FROGGS: Is it accidentally invokving a perl6 (that happens to be perl6-m) at some point? 13:19
FROGGS >.<
jnthn++
but my configure line was: perl Configure.pl --backends=parrot,jvm,moar --prefix=../nqp/install && make install 13:20
k, now it is only a timeout 13:22
jnthn Does it timeout on JVM too?
(I know it's skipped, but if you try to run it...)
FROGGS yes, but it invokes ./perl6 anyway... I fixing that 13:24
jvm does work now, at least the first test 13:27
jnthn k
FROGGS p: say $*EXECUTABLE_NAME.path.absolute
jnthn I seem to remember something about timeouts...
camelia rakudo-parrot c5ca99: OUTPUT«IO::Path</home/p6eval/rakudo-inst/bin/perl6-p>␤»
FROGGS j: say $*EXECUTABLE_NAME.path.absolute
camelia rakudo-jvm c5ca99: OUTPUT«IO::Path</home/p6eval_eval/perl6-j>␤»
FROGGS assum 13:28
and now it does not time out for parrot? 13:29
the answer came in 3s
13:29 skids left
jnthn What did you change? 13:31
FROGGS just gave it the right binary name
but I think the moar runner was to blame
jnthn ah
k
So now we're clean? 13:32
FROGGS not sure yet
in roast there is that: github.com/coke/perl6-roast-data/b....out#L2582
13:33 sqirrel left
FROGGS yeah, test 40 fails here too 13:33
dalek ast: 0faa8bb | (Tobias Leich)++ | S32-io/IO-Socket-INET. (3 files):
use correct executable for running client/server
13:36
13:38 Mouq joined 13:43 Mouq left
dalek ast: 8f066b8 | (Tobias Leich)++ | S32-io/IO-Socket-INET.t:
correctly interpolate executable name
13:43
FROGGS gah, test 40 fails not because the binary data differs, no, one of them is a Buf[uint8] and the other one is a Buf 13:50
dalek ast: 2620680 | (Tobias Leich)++ | S32-io/IO-Socket-INET.pl:
just check for the contained data
13:51
FROGGS jnthn: spectest clean sir
colomon FROGGS++
FROGGS (p-spectest taht is)
[Coke] IWBNI MAIN() would tell you the types of arguments when you make a mistake. (declare that MAIN takes an Int arg, but give it a string - it will show you the usage, but won't mention Int vs. Str)
FROGGS++ cleaning up my mess. 13:52
moritz so, all clear for masak++ doing the release today? :-)
FROGGS moritz: well, doing a j-spectest now 13:53
jnthn FROGGS++ # awesome
moritz now is a good time to add entries to docs/ChangeLog
13:53 araujo left
jnthn Indeed. 13:53
FROGGS t/spec/S02-magicals/env.rakudo.jvm ........................... Failed 1/17 subtests
jnthn The MoarVM monthly tarball is already cut for masak too :)
moritz is it uploaded somewhere? 13:54
13:54 araujo joined, araujo left, araujo joined
jnthn moritz: jnthn.net/MoarVM/MoarVM-2014.01.tar.gz 13:54
moritz \o/
[Coke] there have been jvm failures for several weeks.
jnthn moritz: I'm planning to get a site up at moarvm.org and then have it there in a bit 13:55
dalek kudo/nom: 64e6517 | moritz++ | docs/ChangeLog:
add a ChangeLog entry
jnthn Started on that last night, but it got to 3am and I wanted to sleep. And got $dayjob things to do at the moment
moritz FROGGS: have you written any spectests for <alias=:alpha> etc. ? 13:56
FROGGS moritz: yes
moritz \o/
FROGGS++
FROGGS moritz: just grep for "foo" and "bar" in t/spec :P
moritz :-) 13:57
t/spec$ git grep --word foo|wc -l
2632
bar has only 1239 hits
FROGGS m-runner-default: $(M_RUNNER) 13:59
$(CP) $(M_RUNNER) perl6
that should have a $(BAT) or so, right?
dalek kudo/nom: c2967a0 | jnthn++ | docs/ChangeLog:
A few more ChangeLog entries.
moritz FROGGS: yes, I think so
jnthn FROGGS: Yes, think so
jnthn makes a cup of tea and gets back to course writing
FROGGS ahh, it has a @runner_suffix@ in the Makefile.in, nvm then 14:00
14:04 jnap joined 14:08 jnap left 14:11 kbaker joined 14:18 mtk joined 14:20 woolfy left 14:24 sqirrel joined 14:34 bluescreen10 joined 14:39 telex left 14:40 telex joined, immortal joined, immortal left, immortal joined 14:42 erkan left 14:46 rindolf left 14:49 rindolf joined 14:54 sctt joined 14:56 btyler joined 14:57 bluescreen100 joined 14:58 sctt left 15:01 bluescreen10 left 15:03 skids joined 15:04 jnap joined
nwc10 forgets - who is doing the release today? 15:11
oh, scrollback implies masak
jnthn el sengor masak
Maybe starting on it in a couple of hours, I guess 15:12
15:14 kaare__ left, kaare__ joined 15:19 kaare__ left 15:25 sqirrel left, FROGGS left 15:27 Mouq joined 15:32 Mouq left
[Coke] r: say 1,2…9 15:34
camelia rakudo-parrot c2967a, rakudo-jvm c2967a, rakudo-moar c2967a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Two terms in a row␤at /tmp/tmpfile:1␤------> say 1,2⏏…9␤ expecting any of:␤ postfix␤ infix stoppe…»
[Coke] is that something we can module?
(or should we add a non texas variety?)
jnthn module imo 15:36
colomon is that the unicode … ? 15:39
jnthn .u … 15:43
yoleaux U+2026 HORIZONTAL ELLIPSIS [Po] (…)
[Coke] just wondering how we'd provide that... as an infix? 15:46
r: sub infix<<…>> {$a...$b} ; say 1,2…10 15:47
camelia rakudo-parrot c2967a, rakudo-jvm c2967a, rakudo-moar c2967a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Missing block␤at /tmp/tmpfile:1␤------> sub infix⏏<<…>> {$a...$b} ; say 1,2…10␤ expecting any of:␤ state…»
[Coke] r: sub infix<<…>> {$^a...$^b} ; say 1,2…10
camelia rakudo-parrot c2967a, rakudo-jvm c2967a, rakudo-moar c2967a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Missing block␤at /tmp/tmpfile:1␤------> sub infix⏏<<…>> {$^a...$^b} ; say 1,2…10␤ expecting any of:␤ sta…»
PerlJam r: sub infix:<…> { $^a...$^b }; say 1,2…10 15:48
camelia rakudo-parrot c2967a, rakudo-jvm c2967a, rakudo-moar c2967a: OUTPUT«12 3 4 5 6 7 8 9 10␤»
PerlJam (note the relative precedence) 15:49
IS there a doc somewhere that explains the rationale behind having some unicode+texas ops and not others ? 15:50
skids was thinking of doing a module that mapped most existing core ops to most appropriate unicode
Mainly just to provide a place for people to argue about which ones were most appropriate. 15:51
Though IIRC 'is tighter'/'is equiv'/etc are NYI.
moritz PerlJam: the "core" Unicode operators are all in the Latin-1 range
skids: I think they are implemented 15:52
PerlJam: only the set operators fall out of that schema
PerlJam moritz: Aye, but is that "the" answer. (I honestly never gave it much thought until now). 15:54
skids r: sub infix:<⊶> (\a, \b) is equiv(&infix<^..>) { \a ^.. \b }; .say for 1 ⊶ 5;
15:54 araujo left
camelia rakudo-parrot c2967a, rakudo-jvm c2967a, rakudo-moar c2967a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Can't use unknown trait 'is equiv' in a sub+{precedence} declaration.␤at /tmp/tmpfile:1␤------> ␤ expecting any of:␤ rw parcel hi…» 15:54
PerlJam (i.e., when people ask why we don't have … and →, what do we say?)
colomon I'm in favor of … if for no better reason that my IRC client seems to decide to substitute it when I type three . 15:55
15:58 Psyche^_ joined 16:00 FROGGS joined
colomon well, and it would be a Bad Thing to let people define it to mean something else. 16:00
FROGGS *brrrr*, it is pretty damn cold on a bike these days
Juerd colomon: That means you'd have to predefine all the right arrows...
colomon (I mean, I guess we would allow it anyway, but having it default to the right thing seems sensible)
Juerd I think that any character that is not present on any standard keyboard layout, should not be used :) 16:01
colomon p6 wildly disagrees with you on that one.
Juerd I know
PerlJam doesn't see any guillemets on his keyboard
Juerd PerlJam: They are on some standard keyboard layouts though. 16:02
(Yes, there is a → on my keyboard, but that's a cursor key :))
16:02 Psyche^ left
nwc10 I have a ⌘ key, but if I press it nothing seems to happen :-) 16:03
to be fair, I also have a ⏏ which does do something.
PerlJam Juerd: maybe if programming langauges start using unicode more, the "standard keyboard layouts" will start including unicode characters 16:04
skids I think it's important to keep most of the non-typable characters out of the core distro as a warning that they should only be used in domain-specific devel environments, but also to stake claim to some obvious ones so they don't get abused. Thus, module. 16:06
(One module to bind them all)
Juerd PerlJam: Well, tbh I've just mappend compose + arrow to their respective symbols 16:07
skids But as to what to name it, that's where I'm hung.
PerlJam heh, I was just thinking about remapping some of my keys too :)
Juerd So typing → is now 2 keystrokes, one less than ->.
16:07 zby_home_ left
Juerd So does Perl 6 support ≥ and such? 16:08
skids No I think you can still count them on two hands, including the set operators. 16:09
Juerd Two hands... that's 0..1023 :)
skids Kindercounting. :-)
And it's 3**10 if you use trinary with bent knuckles. 16:10
16:13 [Sno] left
colomon to me, the interesting distinction between … and → is that the former really looks like the ASCII version, while the latter does not. At least in the font I'm using, I'd never confuse the latter, and really can't tell the difference between the former. 16:13
16:13 treehug88 joined 16:14 rindolf left
skids Problem with any arrow is there are tens of nearly identical options. 16:14
Juerd In my terminals, … looks like one character whereas ... looks like 3 :) 16:15
colomon Juerd: in mine, they look almost exactly the same. The latter is slightly darker, but that's it. and that's only easy to compare because you put them both on the same line.
Juerd juerd.nl/i/e110ad53785506342d9c8285eb4a711b.png 16:16
skids xchat-gnome looks like what colomon is seeing. 16:17
16:20 zakharyas left
colomon harmonyware.com/pictures/dots.png 16:20
FROGGS that is why I use a fixed-width font :o) 16:21
skids FROGGS: I bet you even have (setq show-trailing-whitespace t) in your .emacsrc, you snob :-) 16:23
FROGGS skids: I am using SciTe, but yes, it shows whitespace :P
16:24 zby_home_ joined
FROGGS maybe I am too german to tolerate tab/spaces mixups 16:27
16:30 atroxaper left
PerlJam FROGGS: what does being german have to do with tab/spaces mixups? 16:41
FROGGS: and does this affect the number of german python programmers?
skids
.oO(ALLCAPS programming language)
16:42
FROGGS PerlJam: I'd say germans tend to be picky of that kind of stuff
PerlJam: for example we have like ten times more road signs than ppl living here... we like it when everything is ordered well 16:43
PerlJam interesting. 16:44
skids I'd have to get out and walk to read that many road signs. 16:45
16:45 bluescreen100 left
PerlJam you made me think of a trip to Mexico I made about 10 years ago though. There were road signs all over the place. The problem was that they were small and not at eye-level, so you really had to look around to see them (and most locals seemed to just ignore them anyway) 16:45
16:47 bluescreen10 joined
FROGGS sad, I can't find the pic... 16:48
but I'Ve seen one near a forest 16:49
and there was a street and a walkway through the forst
forest*
and another walkway in parallel to the street
where the walkways met where at least 30 road signs 16:50
skids Here we just paint colored rectangles on trees and hope some high school kids don't go and mess with it.
timotimo should we reword the changelog entry for moar to read ">99%" already? 16:51
jnthn timotimo: Yeah, go for it. :) 16:52
timotimo: We were 100% yesterday :P
dalek kudo/nom: 0e89264 | timo++ | docs/ChangeLog:
rakudo-moar actually has >99% now

edited with no small amount of pride
colomon \o? 16:55
\o/
\o? is really the kind of day I'm having
17:03 denisboyun left
dalek kudo-star-daily: 0ab14e7 | coke++ | log/ (5 files):
today (automated commit)
17:05
timotimo tadzik: you didn't have a look at the warnings in the panda tests :( 17:06
17:07 panchiniak joined
jnthn timotimo: There's time for it before the month's Star release still 17:08
moritz which tests does moar pass that the JVM doesn't? 17:13
timotimo the test of time. 17:14
jnthn Probably some Unicode related things also
We don't do NFG yet, but we don't have UTF-16 brain damage either. 17:15
17:15 Mouq joined
timotimo what exactly is the 32 bit integer representation we have? does it cover all and for realsies all unicode characters ever? 17:17
jnthn It's basically UTF-32 at the moment. But note they're signed; we'll use negatives for NFG 17:18
It does cover all the Unicode *codepoints* ever.
Character is a much fuzzier notion.
timotimo ah, of course
FROGGS the terms are described in the first 50 pages of the unicode spec btw :o) 17:19
17:20 Mouq left
[Coke] someone wants to write some p5 to help out star, it'd be nice if the 'make modules-test' actually had a single harness rather than one per module. 17:20
17:21 djanatyn left, djanatyn joined 17:22 ajr joined 17:23 ajr is now known as Guest31072, Guest31072 is now known as ajr_
timotimo jnthn: fwiw, the async file class still has a nasty bug in which it closes the $PIO and later tries to boolify it, which doesn't give back false for a closed file descriptor 17:25
and i don't know how to properly set the boolification spec.
17:26 Tene left, kaleem left, Tene joined 17:27 djanatyn left, djanatyn joined
jnthn timotimo: oh, hmm 17:27
timotimo despite spending lots of time on moarvm this month, the rakudo team has collected a sizable amount of changes visible to every perl6 user; i'm very happy 17:28
17:29 djanatyn left, djanatyn joined
moritz well, there is overlap between the moar and rakudo teams, but also some folks that work on just one them 17:30
jnthn Some significant improvements on JVM also. 17:31
[Coke] also I ate a donut.
timotimo does the usage of a single classloader for all classes warrant a changelog entry?
that was for all P6Opaque, right?
jnthn timotimo: Did that actually go in yet?
timotimo i should have a look 17:32
jnthn ah yes, seems so
6ee48136a
[Coke] r: "Moar is at {100*28656/28786} of parrot today."
camelia ( no output ) 17:33
[Coke] r: say "Moar is at {100*28656/28786} of parrot today."
camelia rakudo-parrot 0e8926, rakudo-jvm 0e8926, rakudo-moar 0e8926: OUTPUT«Moar is at 99.548392 of parrot today.␤»
timotimo Only use a single ByteClassLoader, attached to the GlobalContext. got in
[Coke] is there -ever- a case where the three rakudos will be running different versions?
jnthn timotimo: Mark it as JVM specific in Changeog
[Coke] (and if so, can we only report the hash once in cases where it's the same?)
timotimo [Coke]: i had been working on a script that would compress this text down to a readable but succint format in all cases
colomon 99.5%?! \o/
jnthn [Coke]: I think that they are built/deployed at different times
colomon: That's against Parrot, not against JVM... 17:34
timotimo jnthn: did anybody check if it actually makes a big difference in anything?
colomon oh
jnthn timotimo: Well, it made a differnce in that arnsholt++ could unbust stuff
timotimo :D
jnthn timotimo: Ohh!!! Of course, we must mention NativeCall on JVM should work much better now in the ChangeLog too!
PerlJam one of these days it'll be "Moar is at 102.23434% of parrot today" :)
[Coke] timotimo: danke. 17:35
timotimo [Coke]: but i've been writing it in perl6, because that's all i know :P
someone will have to port it for me
[Coke] timotimo: can't we guarantee that we always have a working perl6 about? ;) 17:36
timotimo who knows :)
[Coke] but yah, someone can port it.
dalek kudo/nom: 4850a74 | timo++ | docs/ChangeLog:
a few more JVM changelog entries

  arnsholt++ is to thank for these
17:37
17:44 psch joined 17:45 fhelmberger left 17:47 djanatyn left 17:48 djanatyn joined 17:50 immortal left, erkan joined, erkan left, erkan joined
timotimo i've found something where perl6-m runs fine and perl6-p double-frees :) 17:54
jnthn
.oO( bug free )
17:55
17:55 dakkar left
timotimo rakudo: my $foo = 0 but ""; say $foo ~ " hey!" 17:56
17:56 FROGGS[mobile] joined
camelia rakudo-parrot 0e8926, rakudo-jvm 0e8926, rakudo-moar 0e8926: OUTPUT« hey!␤» 17:56
timotimo hm. 17:57
masak evening, #perl6
TimToady one would expect ~ to prefer the stringy bits
psch hi masak, hi #perl6 17:58
timotimo yes, it does in this example, but locally i have a fail that i need to golf
psch (or the other way around..?)
timotimo r: class Test { has $.prefix = 0 but ""; has $.postfix = 0 but "" }; given Test.new { say .prefix ~ " - " ~ .postfix } 17:59
17:59 rindolf joined
masak I had this idea that I would cut a Rakudo release tonight. what do you think of that, #perl6? 17:59
camelia rakudo-parrot 0e8926, rakudo-jvm 0e8926, rakudo-moar 0e8926: OUTPUT« - ␤»
timotimo masak: i think that would be splendid
masak in fact, I'm planning to cut a Rakudo release *while on a train*.
grondilu rakudo: my $foo = 0 but ""; say $foo ~~ " hey!"
camelia rakudo-parrot 0e8926, rakudo-jvm 0e8926, rakudo-moar 0e8926: OUTPUT«False␤»
geekosaur rolling release? :p
masak .oO( :P
)
yeah, it's continuous delivery. 18:00
jnthn Rakudo on Rails?
timotimo ah, i know what i've done wrong
masak .oO( Perl 6 on Pails )
timotimo i've been initialising the attributes when calling the constructor
and *then* i put in default values
18:03 stevan_ left 18:10 immortal joined
masak hm, so what should we name this release... 18:10
PerlJam Fred 18:11
TimToady Three Musketeers
18:11 erkan left
PerlJam wishes we would do away with the names as that ends up being the hardest part of releasing :) 18:11
TimToady Three Stooges 18:12
Three Blind Mice
Three Little Pigs
the latter is pretty accurate :)
timotimo but it's two-thousand and *four*!
TimToady
.oO( off by 10 error? )
18:13
PerlJam TimToady: which one has the house of bricks?
TimToady well, if we go with Musketeers, rakudo can be D'Artagnan, the 4th musketter 18:15
*teer
Triple Play 18:17
But Wait, There's Moar 18:18
jnthn Just don't call it "Three Way" and we're all good...
PerlJam trifecta!
masak ok, how about "HongKong"?
TimToady PerlJam: that's not bad 18:19
jnthn Are there any place names with a 3 in them, if we want to pretend there's a Pm group there? :) 18:20
TimToady especially if we don't commit to whether the "*fecta" part refers more to "per" or "de"
masak .oO( one country, two systems, three backends )
TimToady I sure there must be a city somewhere with 三 in its name 18:21
masak yes, but does it have a .pm group? :)
jnthn masak: I'm not sure that's stopped us choosing cities before ;) 18:22
TimToady Three Rivers Stadium is in Pittsburgh, fwiw
18:23 rindolf left
masak ok, I'll leave you to settle this. if I hear a convincing argument for a city, I can change from HongKong to something else. 18:23
jnthn There are multiple cities called Trinity, it turns out.
timotimo are there three of them?
jnthn At least 4 18:24
benabik Some of them need to go.
;-)
TimToady why "HongKong"?
jnthn benabik: Well, Trinity, TX shows up on Google maps with a winter storm warning, so there's hope ;) 18:25
PerlJam TimToady: hopefully the *fecta part is more "in" or "per" than "de" :)
18:26 btyler_ joined, ponbiki left, btyler_ left
TimToady maybe we should name it Plano because that implies at least three dimensions 18:27
jnthn oooh!
PerlJam +1
jnthn TimToady++
dalek kudo/nom: c6739ec | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] added

Still need to fill in changes in this release, and contributors.
masak ok, Plano it is.
hold on, changing.
18:29 btyler left
dalek kudo/nom: d32e089 | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] code name "Plano"
18:29
kudo/nom: 3ecf7ee | masak++ | README:
[README] update copyright year

It's 2014! The year of Perl 6 on the desktop, surely.
18:31
jnthn adds "make Rakudo Star MSI install a desktop icon" to his todo list, to make sure it is so 18:32
18:32 SamuraiJack left
timotimo do we have a proper icon? 18:32
just camelia?
once we have rakudo-star JVM, we can even ship a little gui with the repl 18:33
and perhaps a gui for panda, too
18:35 ponbiki joined, ponbiki is now known as Guest63085
timotimo it might be time to write a perl6-highlighter in perl6 for that repl :P 18:36
masak Use of uninitialized value $lines[0] in pattern match (m//) at tools/update-tai-utc.pl line 18. 18:37
still getting this error message, like, a million times.
*sigh* -- can I just assume there were no leap seconds this month?
jnthn Well, see ftp://hpiers.obspm.fr/iers/bul/bulc/TimeSteps.history 18:38
Last entry on there is for 2012 Jul 1
18:39 jeffreykegler joined
masak looks like the ChangeLog has been properly kept up to date. timotimo++ jnthn++ moritz++ 18:39
TimToady thinks they should just install a few leap-minutes every 1000 years or so
masak someone is very welcome to find out what's wrong with tools/update-tai-utc.pl 18:40
jnthn Well, it looks for this: 18:41
shift @lines until $lines[0] =~ /\A 1972 Jul\. 1/;
But the line it's looking for now reads:
dalek kudo/nom: 2d29ae5 | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] contributors
18:42
jnthn oh...just like that.
TimToady wow, that's different :)
jnthn \A is start of this...
argh
strart of *string*?
masak yep.
jnthn wonders why not ^ :) 18:43
raydiak b/c ^ is start of line :)
TimToady because someone might put /m on it someday :)
PerlJam doesn't get any such message when he runs tools/update-tai-utc.pl
TimToady and then it would be misleading, despite still working right :) 18:44
jnthn rylinaux: Only with the m modifier... :)
timotimo masak: "Timo Paulssen" and "timotimo" is a dupe
jnthn I always found the naming of the s and m modifiers fun to explain. "s makes . match *m*ultiple lines, while m makes ^ and $ refer to a *s*ingle line" :) 18:45
In .Net they're spelled SingleLine and MultiLine or so and it once took me 10 minutes to convince somebody that yes, they really did want the one called SingleLine in order to make . match over multiple lines... 18:46
TimToady m makes ^ and $ match multiple times in a multi-line string; s treats the multi-line string as a single line
jnthn TimToady: Yeah, that's the correct interpretation. It's just easy to guess or mis-remember the other one, I guess.
TimToady or if you prefer, 's' makes it use string semantics rather than line semantics
dalek kudo/nom: f81a636 | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] some changes
18:47
TimToady but yeah, the whole thing is a bad thunk
raydiak (oh and good morning all)
timotimo hullo raydiak :) 18:48
dalek kudo/nom: a489283 | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] de-dupe timotimo

  timotimo++ (or was it Timo Paulssen?) for pointing it out.
masak raydiak: \o
jnthn masak: The other one I'd call out in there is the fixing of the JVM stack overflow bug
masak: It was one that was easily run into and a real pain.
timotimo oh yes definitely
masak jnok, will do. 18:49
raydiak [Coke]: re single test harness, do you mean just calling prove once with all the module test dirs passed to it?
jnthn afk for a bit
dalek kudo/nom: c0c6a6b | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] include SO JVM bug

By popular demand.
TimToady should contribute someday... :) 18:51
dalek kudo/nom: a365a23 | duff++ | CREDITS:
Add another alias for timo
timotimo oh good call 18:52
dalek kudo/nom: bbbebb3 | masak++ | docs/release_guide.pod:
[docs/release_guide.pod] update release dates/names
18:53
masak down the call stack I slide. time for the NQP release process. 18:54
masak .oO( two release processes for the price of one! ) 18:55
timotimo how so?
PerlJam the rakudo release isn't going to mention moar yet?
timotimo at least not the rakudo star release ...
masak PerlJam: forgiveness > permission 18:56
feel free to come up with sump'n.
by the rule outlined in NQP's step 1, I should *not* update the PARROT_REVISION.
dalek p: 283f8f2 | masak++ | VERSION:
bump VERSION to 2014.01
18:57
timotimo i think it's supposed to give us regex test passes
18:59 treehug88 left
masak well, I'm just following the (very clear) instructions. 18:59
timotimo sure, i'm not complaining
19:00 darutoko left 19:01 treehug88 joined 19:02 rindolf joined 19:03 woolfy joined, Mouq joined 19:04 LLamaRider left 19:06 Ayiko joined 19:07 Mouq left 19:14 btyler joined
pmichaud FWIW, DFW.pm hasn't been picked for a release name yet. :) 19:14
masak pmichaud: maybe add it to the release guide for next time? :) 19:15
jnthn In the first para, "Rakudo is an implementation of Perl 6 on the Parrot Virtual Machine and the Java Virtual Machine." May want to tweak that. ;) Or I'll do it after shopping if nobody beats me to it. :)
bbi15
raydiak how about this: "We've worked extra-hard this month to offer you a world-exclusive special. Starting right here, right now, Rakudo supports revolutionary MoarVM technology. That's right, folks! You get not one, not two, but *three* rakudo backends for the low, low price of FREEEEEEEE! Download now! Offer only valid while supplies last." 19:16
masak I don't think I will be able to upload things. I don't have the SSH keys on this machine.
maybe someone can help me with that?
raydiak: "...except in Nebraska!" *wide eyes*
19:17 stevan_ joined, panchiniak left
pmichaud I'm not sure we want two Texas cities in back-to-back releases... but yes, I'll add DFW to the list. 19:17
psch r: $_ = "aa"; .++.say while 1; # RT99732
camelia rakudo-moar 4850a7: OUTPUT«(timeout)aa␤ab␤ac␤ad␤ae␤af␤ag␤ah␤ai␤aj␤ak␤al␤am␤an␤ao␤ap␤aq␤ar␤as␤at␤au␤av␤aw␤ax␤ay␤az␤ba␤bb␤bc␤bd␤be␤bf␤bg␤bh␤bi␤bj␤bk␤bl␤bm␤bn␤bo␤bp␤bq␤br␤bs␤bt␤bu␤bv␤bw␤bx…» 19:18
..rakudo-jvm 4850a7: OUTPUT«(timeout)aa␤ab␤ac␤ad␤ae␤af␤ag␤ah␤ai␤aj␤ak␤al␤am␤an␤ao␤ap␤aq␤ar␤as␤at␤au␤av␤aw␤ax␤ay␤az␤ba␤bb␤bc␤bd␤be␤bf␤bg␤bh␤bi␤bj␤bk␤bl␤bm␤bn␤bo␤bp␤bq␤br␤bs␤bt␤bu␤bv␤bw␤bx…»
..rakudo-parrot 4850a7: OUTPUT«(timeout)aa␤ab␤ac␤ad␤ae␤af␤ag␤ah␤ai␤aj␤ak␤al␤am␤an␤ao␤ap␤aq␤ar␤as␤at␤au␤av␤aw␤ax␤ay␤az␤ba␤bb␤bc␤bd␤be␤bf␤bg␤bh␤bi␤bj␤bk␤bl␤bm␤bn␤bo␤bp␤bq␤br␤bs␤bt␤bu␤bv␤bw␤…»
dalek kudo/nom: e7280d0 | duff++ | docs/announce/2014.01.md:
Add more Moar to release announcement
19:19
timotimo psch: if you put a space after the RT, you'll get a link for all of us in the chat, too: # RT 99732
actually: RT #99732
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=99732
psch timotimo: last time i did it with a # befo... yes
PerlJam Someone vet the Moar stuff I just added
psch i forgot this time
my point being: this seems fixed, at least if the $_ semantic inside the while is corrent 19:20
*correct
timotimo PerlJam: i'd suggest using moarvm.org instead of the long github link
especially as that's currently a redirect to github
pmichaud It's a little weird that footnote [^2] comes before [^1].
timotimo oh hey pm :) 19:21
masak vets
timotimo i'm glad you got a codename suggestion for next month's release already; that way i won't have such a hard time :P 19:22
[Coke] raydiak: I think you need 1) a list of .t files across modules, 2) a script that prove can use that will say "oh, this .t is in *this* module, I'll chdir there...".
pmichaud I don't follow the "Plano implies three dimensions" reference, though.
s/reference/meme/ #whatever
masak PerlJam: looks good.
[Coke] plane is 2, I get that far... 19:23
masak I'm not sure there is more to it than that.
a plane is usually embedded in a 3-space, I guess...
19:23 btyler left
dalek kudo/nom: 0a9e40e | duff++ | docs/announce/2014.01.md:
reorder footnotes and use moarvm.org
19:24
pmichaud change text to "because Rakudo now exists in multiple dimensions" maybe?
PerlJam how about we paste in the IRC discussion from earlier and say "Plano" was *waaaay* better than "Three Blind Mice" :)
masak :)
pmichaud PerlJam: +1
maybe as a footnote / postscript
psch there's 11 american Planos and 1 croatian Plano and the digit sum of 12 is 3...?
pmichaud well, I'm only familiar with the Plano I live in :) 19:25
[Coke] I don't think we need to 'splain it.
pmichaud Maybe also put in a note saying that we're looking for pm group names to use in future releases. :)
*Or*, perhaps this is the point at which we adopt an entirely new naming scheme. 19:26
PerlJam pmichaud: +1 to that!
pmichaud since we're now multi-backend with n>2
PerlJam (new naming scheme)
moritz how about simply dropping the naming scheme?
timotimo moritz: i'll have a short script that'll prepare strings for all combinations of backends that'll be minimized; you'd have to run it once for each rebuild. it's in p6. is that acceptable?
PerlJam moritz: And that's even better IMHO
19:27 tgt joined
moritz timotimo: prepare strings? what strings? 19:27
timotimo: I feel I'm missing a bit of context here
pmichaud We could also keep the names but drop the explanation.
timotimo rakudo-{parrot, jvm, moar} c8ec1d
for example
dalek kudo/nom: f4b0fb4 | masak++ | docs/announce/2014.01.md:
[docs/announce/2014.01.md] explain "Plano" better
masak pmichaud, PerlJam: ^ 19:28
timotimo instead of rakudo-parrot c8ec1d, rakudo-jvm c8ec1d, rakudo-moar c8ec1d
masak didn't go with the footnote/postscript idea. this feels better. I could still change it, though.
pmichaud release manager gets to decide :)
[Coke] my 2¢: I hate release names.
timotimo i don't care much about them 19:29
pmichaud proposed release name: "Coke hates release names."
PerlJam reiterates ...
* PerlJam wishes we would do away with the names as that ends up being the hardest part of releasing :)
pmichaud It is good to be able to talk about previous milestone releases, though... e.g., "Beijing"
benabik Perhaps we should just start naming releases after people who hate release names. 19:30
masak I like the release names.
I don't mind looking for a pm group and a rationale.
it bonds us together somewhat.
PerlJam pmichaud: but being able to say the 2011.07 release gives more information
pmichaud PerlJam: yes, if I can remember that date. 19:31
lue same goes for the release name :)
PerlJam If it's significant, you'll remember it :)
pmichaud It's much easier for me to remember Beijing than to remember 2011.08 (and wonder if it really was 2011.08 or 2011.07 or 2012.08 or ...)
PerlJam Or maybe *only* name significant releases?
pmichaud We don't always know a-priori which ones are significant.
PerlJam Well, mostly we get to decide if we think they will be. 19:32
and if they become so after-the-fact, we can post-name it.
pmichaud I'm with masak++ ... I like having the names. I don't know that we need to continue with .pm groups, but there are bound to be more for a while yet. 19:33
PerlJam masak: re bonding ... is that like a shared pain? :)
masak PerlJam: heh. 19:35
lue We could just use a random town name picker in the future :) 19:36
masak angularjs has pretty cool release names. they look generated.
PerlJam "What do you want to name the release?" "I dunno, what do *you* want to name the release?"
masak we shouldn't copy them straight off, but it's an idea for the future.
lue
.oO(Relase #87 "Frotzlexity")
19:37
PerlJam maybe we just pick a name and vet it here with whoever is around and go with it. No "scheme"
geekosaur generate random syllables 19:38
benabik Train a Markov chain with the #perl6 backlog.
psch SORRY 19:39
pmichaud use the #perl6 backlog but strip out common words.
leaving only the weird words as potential names.
skids Pick a city from Middle Earth and pretend there was a Perl con there.
pmichaud "Whipuptitude" and "Manipulexity" ought to be release names somewhere. :-)
PerlJam skids: why do we have to pretend? ;) 19:40
timotimo i think whipuptitude is for when NativeCall is on all three backends (or four at that point?) and manipulexity when we have everything for macros
PerlJam timotimo: sounds good to me.
lue timotimo: I like that.
skids Because actually having a BOF session with the elven delegates about unicode representations of their alphabet would be hard.
19:41 tgt left
skids And forget the dwaves and their runish hyperoperators. They are just cantankorous. 19:41
lue I don't think we'll get to talk to the dwarves. They still haven't pulled the lever to the drawbridge :/ 19:44
pmichaud hah!
This month's release should be "TimToady"
since we now really do have more than one way to do it :-)
(can also use TMTOWTDI to avoid the nick pun) 19:45
benabik There's more than two ways to do it?
pmichaud TimTwody
masak we should ask quietfanatic to change to that nick :P 19:46
PerlJam heh
pmichaud I'm afk -- more housework and prep for later. 19:47
I'm okay with adjusting the naming scheme by consensus 19:48
bbl
skids
.oO(If Apple can get away with it...)
19:49
masak I think we can adjust the naming scheme, but we should do it at some major shift into a "new era" for Rakudo.
moritz moar areas?
PerlJam like the introduction of a new VM?
jnthn The point when we've a multi-backend star, maybe...
timotimo introduction of rakudo into the browser? ;)
masak jnthn: yes, maybe. 19:50
timotimo that'd be fine by me, jnthn
how likely is february moar star? :)
because february jvm star sounds quite likely
jnthn Dunno. I've no idea how bad the ecosystem looks on Moar so far.
moritz does moar have nativecall?
jnthn I know we've work to do on NativeCall. The debugger shouldn't be hard.
timotimo moritz: arnsholt noted it'd mostly be copy-pasting from parrot
jnthn moritz: Well, we've got dyncall into the build; not sure how far along the REPRs are, but we can...what timotimo said :) 19:51
PerlJam Still sounds more like April to me (for moar R*)
timotimo oh, really?
well, sockets will be big, too
jnthn I may take on sockets when I do the Moar IO do-over
I'll put in the async IO and I want to get async sockets in too.
With an API that works on Moar and JVM. 19:52
s/may/almsot certainly will/
timotimo that sounds lovely 19:53
19:57 tgt joined, treehug88 left 19:59 treehug88 joined 20:00 Rotwang joined
FROGGS .u ARMENIAN DRAM SIGN 20:02
yoleaux U+058F ARMENIAN DRAM SIGN [Sc] (֏)
20:06 locsmif joined
masak ok, the NQP release can now be found at masak.org/carl/tmp/nqp-2014.01.tar.gz -- could I ask for someone's help (pmichaud? jnthn? tadzik? moritz? PerlJam?) to upload it to the real place? 20:06
(I don't currently have the SSH key)
besides the uploading step, this concludes the NQP release process. back to the Rakudo release process. 20:07
20:07 denisboyun joined
timotimo ✔ NQP Release prepared 20:07
20:09 rindolf left 20:11 tgt left
PerlJam masak: uploaded. 20:12
dalek kudo/nom: 560d347 | masak++ | tools/build/NQP_REVISION:
[release] bump NQP revision
20:13
kudo/nom: 2834f75 | masak++ | VERSION:
[release] bump VERSION
masak PerlJam++
PerlJam: hope you're around for the Rakudo .tar.gz in an hour or so ;) 20:14
20:14 Mouq joined
masak there's no 'realclean' target in the Makefile. 20:15
docs/release_guide.pod:162 expects there to be one.
PerlJam huh, the 2014.01 release of nqp is about 100K smaller than the 2013.12 release. Until now we've had steady growth. 20:16
masak interesting.
timotimo very cool! :) 20:17
the stage0s may have shrunken
i think since last month we got rid of base64 in the moarvm stage0s
masak oh! 20:18
PerlJam masak: Are you sure there's no "realclean"? I've got one.
(well ... I had one until I ran "make realclean" ;) 20:20
20:20 crazedpsyc left
skids now that's clean! 20:20
timotimo :D 20:21
psch .oO( sub undo { ::undo := Nil } )
huh, could that actually work? obviously not like that
r: sub foo { &OUTER::foo := sub { 1 } }; foo.say; foo.say # seems like it? 20:22
camelia rakudo-parrot bbbebb, rakudo-jvm bbbebb, rakudo-moar bbbebb: OUTPUT«sub() { ... }␤1␤»
psch realizes more and more how little he actually knows about perl6 20:23
masak psch: what are you trying to accomplish? 20:24
20:24 crazedpsyc joined
psch masak: i was just thinking, seeing as realclean removes it self, if perl6 can have a function that undefined itself 20:24
which apparently it can
s/it self/itself/ 20:25
but i might misunderstand camelia here, explainations welcome :)
20:27 denisboyun left
timotimo huh, here moar seems to have gotten stuck trying to get .combinations 20:28
masak psch: no, seems you got it right.
psch: but I'm kinda surprised that works.
20:28 japhb__ left
masak wonder how that interacts with optimizations. 20:28
20:29 japhb left
jnthn I think it's a bug that it works 20:29
The optimizer woulda been free to inline it
20:30 FROGGS[mobile] left
lue jnthn: that might be considered an optimizer bug in that case, were it to happen :) 20:30
jnthn lue: No, buy spec the optimier is allowed to inline it.
*by 20:31
If you want to suppress inlining you need to say "use soft;" or so
lue or perl6 --optimize=0 , right?
jnthn Well, you could do that too, but that means you're depending on everybody running your code with --optimize=0 20:32
Not to mention giving up every other optimization.
masak Perl 6 is already fairly geared towards static subroutines, which is why it's surprising that it works.
lue oh, I thought 'use soft' would mean the same thing
timotimo obviously there's a bit more performance to had that way? 20:33
jnthn lue: No, use soft just gives up some things
timotimo r: print "a" for ([<a>], [<b>], [<c>], [<a b>], [<b c>], [<a c>], [<a b c>]).combinations 20:34
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb: OUTPUT«No such method 'combinations' for invocant of type 'Parcel'␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar f4b0fb: OUTPUT«Cannot find method 'combinations'␤ in block at /tmp/tmpfile:1␤␤»
timotimo r: print "a" for ([<a>], [<b>], [<c>], [<a b>], [<b c>], [<a c>], [<a b c>]).list.combinations
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb: OUTPUT«aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa»
timotimo r: my %foo{Any}; %foo{$_} = 1 for ([<a>], [<b>], [<c>], [<a b>], [<b c>], [<a c>], [<a b c>]); print "a" for %foo.keys.combinations 20:35
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb: OUTPUT«aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa»
PerlJam masak: are you updating the release guide to talk about jvm or moar? It still seems highly parrot-centric
timotimo so ... what am i doing wrong locally?
PerlJam (maybe that's best to do after the release)
timotimo maybe i'll just have to get the newest version or something
20:38 stevan_ left 20:44 immortal left
masak PerlJam: yeah, I'm not updating it now. 20:45
just pointing out points where it doesn't work.
make: *** No rule to make target `stresstest'. Stop. 20:46
there's a p-stresstest
PerlJam p-stresstest only exists if you've listed parrot as one of the backends 20:48
20:49 jeffreykegler left
masak I suppose I must have, at some point. 20:49
as in --backends=parrot
so yeah, the build instructions are not in sync with reality. 20:50
I suggest fixing after release.
PerlJam yep, sounds sane 20:51
masak++ :)
20:52 zby_home_ left
diakopter wut happened in kiev 20:57
masak diakopter: lots of fire and protest.
and strife :(
colomon :( 21:01
21:02 dayangkun left
masak govt there is calling the demonstrators terrorists. 21:04
locsmif not just there
diakopter sry for OT 21:06
colomon doesn't seem to have made the headlines on CNN.com? :\ 21:09
psch colomon: click "International"
colomon is sure Justin Bieber's legal status is vastly more important.
psch it was there at least
not sure if it still is 21:10
PerlJam colomon: if you want real news, you need to read bbc.com.
colomon PerlJam: honestly, I don't much want real news. or the crap cnn.com covers.
tadzik I watched an episode of DS9 yesterday, about a military guy taking over the entire earth because of a terrorist threat. To make it more relevant, a name "Snowden" occured in the episode
colomon psch: if you click "Ukraine" at the top, you get there. 21:11
itz_ is there any short cut to build all backends at once?
timotimo prn: say "hello"
itz_: of course, --backend=parrot,jvm,moar --gen-moar --gen-parrot
perl6: say "hello"
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb, niecza v24-109-g48a8de3: OUTPUT«hello␤»
colomon Bieber still more prominent.
itz_ I was thinking of --backend=all :) 21:12
lazy
tadzik well volunteered :P
psch .oO( --everything )
moritz itz_: shouldn't be hard for you to patch :-)
lue
.oO( perl Configure.pl --* )
itz_ :)
psch lue: that would clash with shell expansion, wouldn't it? 21:13
lue whatever.
jnthn Is that "my eyes are closed 'cus it's snowing"?
PerlJam itz_: I would suggest using some special designation for things that aren't real VMs either ALL CAPS or some symbology.
TimToady just type * at the command line 21:14
it *might* run Configure.pl
PerlJam lue: was that an intentional pun? :)
lue PerlJam: yes, of course :)
jnthn Yeah, if you do --backend=all, then somebody will go and make an allvm :)
21:14 Rix left
lue $ perl6 -e 'say @*ARGS.perl' --* 21:14
Array.new("--*")
tadzik mostvm
lue Apparently it doesn't hurt. Hm. 21:15
jnthn OK, I have some time for Perl stuffs...what to do...
tadzik starts up on moar, and warms up jvm just in case
diakopter jnthn: specializer/ :)
lue Apparently, putting a dash in front of the asterisk stops shell expansion. So we could have perl Configure.pl --* :)
jnthn diakopter: but SSA haaarrd :P
TimToady though in my nom directory it ends up printing out the 12 days of Christmas, since 12days sorts first :) 21:16
diakopter do it in reverse then
timotimo jnthn: write a small amount of code to subclass JavaBootInterop inside rakudo and give me that so i can continue my work? :)
jnthn timotimo: uh...we already have a subclass of JavaBootInterop in Rakudo?
timotimo: src\vm\jvm\runtime\org\perl6\rakudo\RakudoJavaInterop.java
timotimo last time i tried that it didn't seem to want to work 21:17
is it hooked up to an nqp:: op?
TimToady btw, rosettacode.org/wiki/Fractran#Perl_6 fails under moar for some reason
jnthn timotimo: Hm, good question...
timotimo: It's used when you say "use such.enterprise:from<java>;"
timotimo ah. hum. 21:18
that'll be a good hint
jnthn timotimo: ah, nqp::jvmrakudointerop()
TimToady: Explodes, or wrong answer? 21:19
TimToady wrong answer, 2 15 15 15 15 15 15 15 15 15 15 15...
timotimo i think our % operator with rats may still be busted in some kind of way?
FROGGS TimToady: on my box too
and it hangs after that
TimToady well, you have to ^C it
the final loop is over 0..* :)
FROGGS gah 21:20
:o)
TimToady which, by the way, I couldn't write as .say for FT2
because for some reason it tries to calculate the entire sequence first...
[Coke] r: my int $a = 2**50;
camelia ( no output )
..rakudo-parrot f4b0fb: OUTPUT«===SORRY!===␤Error while compiling op bind: Cannot infer type from '1125899906842624'␤»
[Coke] masak: ^^ 21:21
TimToady but I suspect the latter is just for loops not being correctly lazy 21:22
jnthn TimToady: I'm suspecting it tries to batch it
TimToady: Unconditionally
masak [Coke]: ugh. feel free to report.
TimToady yes, well, it'll take a while to generate the first 100000 members of *that* sequence...
jnthn yes ;) 21:23
itz_ if (defined $options{backends}) {
+ $options{backends} = join ",",keys %known_backends if $options{backends} eq 'ALL';
TimToady wonders what we could do about overbatching...
jnthn TimToady: I think Pm may well have thought this one through a bit. 21:24
TimToady: Not that more thinking from you isn't also a good idea ;)
TimToady give or take a negative here or there
21:24 ggoebel1112 left 21:26 ggoebel1112 joined
TimToady but if it were just batching, why wouldn't it do the same for FT2[$_]? 21:26
jnthn reproduces the ft bug locally...
TimToady: I think it's MapIter that batches 21:27
TimToady: oh, but you have a map...hm
FROGGS realizes that he does not understand that code
TimToady well, that bit of the code is grondilu++'s
jnthn TimToady: Me either, I'm just gonna try and make it work :P 21:28
uh, FROGGS
timotimo FROGGS: why? it seems understandable to me :P
FROGGS timotimo: maybe I am lacking some sort of CS degree? I dunno
timotimo hah :)
r: say 16.msb; say 16.lsb; 21:29
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb: OUTPUT«4␤4␤»
FROGGS what does map (* * n).narrow, LIST ?
timotimo that's pretty handy!
well, * * n is the whatevercode that multiplies with n
and narrow is our method to try to get the narrowest numeral type for the number
r: .perl.say for 10.0, 10.0.narrow, 10.0e, 10.0e.narrow; 21:30
TimToady also, I'm pretty sure that .lsb == .msb is not going to be the most efficient way to detect a power of two
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Confused␤at /tmp/tmpfile:1␤------> .perl.say for 10.0, 10.0.narrow, 10.0⏏e, 10.0e.narrow;␤ expecting any of:␤ …»
timotimo TimToady: i remember some really neat trick from the hacker's delight; i think you have to xor it with the number itself - 1 or something
masak darn, train will arrive soon, and I'm still just stresstesting S12...
timotimo so 100000 xor 011111?
that's not quite it, though
21:31 Rix joined
TimToady well, I tried +& and looked for 0, but that doesn't work 21:31
21:31 locsmif left
timotimo oh, that was my next thought 21:31
jnthn masak: Why do the delays never happen when you wnat them... 21:32
TimToady oh, maybe the precedence is wrong
timotimo that would probably be it
FROGGS ahh, I think I understand...
timotimo r: say 16 +& 15
masak jnthn: hey, don't jinx my train in the last 10 minutes! :P
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb: OUTPUT«0␤»
timotimo r: say 16 +& 16 - 1
camelia rakudo-parrot f4b0fb, rakudo-jvm f4b0fb, rakudo-moar f4b0fb: OUTPUT«15␤»
timotimo yup, it's the precedence.
jnthn r: sub x(\n) { say (^5).map: (* * n).narrow }; .&x for ^2
camelia rakudo-moar f4b0fb: OUTPUT«0 0 0 0 0␤0 0 0 0 0␤»
..rakudo-parrot f4b0fb, rakudo-jvm f4b0fb: OUTPUT«0 0 0 0 0␤0 1 2 3 4␤»
TimToady yeah, it was precedence 21:33
jnthn ah...nested whatever curries don't correctly nested their closures... 21:34
Which we must mystically get away with on other backends
No doubt at some performance cost.
grondilu TimToady: to detect a power of two you can test whether the logarithm in base 2 is integer
jnthn Moar just get's 'em wrong so there's no choice but to fix it properly... :) 21:35
FROGGS ahh, I love that when jnthn++ says these sentences :o)
timotimo oh, huh
FROGGS: now there's something i realise i don't understand :)
masak detrain &
FROGGS timotimo: do tell :o)
timotimo how exactly moar fails there. what the closure nesting stuff means
jnthn timotimo: I think you worked on whatever_curry recently? :)
timotimo yes, i did
that's what makes me scratch my head even more :P 21:36
jnthn timotimo: I'm almost certain you didn't introduce the issue though, I think it predates it...
timotimo: Can you explain what the chain stuff is doing?
timotimo i only changed whatever handling if there's a chain
21:36 Guest1338 joined
grondilu an other way would be .base(2) ~~ /^10*$/ 21:36
TimToady I suspect the AND method is the fastest 21:37
colomon grondilu: if you've got an Int, you can detect powers of 2 with !($n +& ($n - 1))
timotimo well, it remembers all the arguments to the operator, remembering exactly which ones were whatever-arguments, if it's a chain so that one level up if it finds such a chained whatevercode it can flatten it out properly
colomon … which someone already posted back there
grondilu oh, ok
TimToady and, in fact, the RC entry is already updated for 21:38
colomon needed this for ABC
jnthn timotimo: Yeah, I think I see it 21:39
TimToady thinks about shouting "First bug!" 21:40
21:40 dayangkun joined
[Coke] anyone care to volunteer to give feedback on a "what's perl6" presentation to a handful of p5 folks? 21:42
timotimo i'll volunteer to look at it :P 21:43
TimToady should we have lower standards because it's only a handful? 21:44
tadzik we're a handful too :)
TimToady quite a handful
hoelzro I would look
TimToady what is Perl 6, anyway? 21:45
tadzik I'd happily welcome a second handful
moritz would love to know that too
psch a DWIM-y pile of *..?
FROGGS nah, it is the last language one learns :o) 21:46
tadzik then you just tolerate the rest :P
FROGGS yeah...
TimToady It'll either kill you or cure you...
psch is painfully reminded of all those accidentally omitted returns at $phpdayjob
FROGGS psch: exactly 21:47
but there are worse things out there...
[Coke] (lower standard) yes.!
FROGGS my colleagues hack to hack ABAP!
PerlJam Perl 6 is kind of like a FPGA only the "gates" are way more powerful
diakopter -_+
TimToady other than the field, the gates, and the array, that's pretty accurate
hoelzro [Coke]: are you going to show 1..* examples of cross pollination between Perl 5 and Perl 6? 21:48
PerlJam :-P
diakopter field lol
[Coke] well, smartmatch was the first one that came to mind. :P
hoelzro heh
TimToady alas, that one didn't work out so well
PerlJam yeah, it's always good to start off with a bad example :)
hoelzro I'm thinking use feature qw(say state); use Moose
[Coke] but yes, specifics are good. 21:49
tadzik it's a kind of a... field
TimToady // worked out better
21:49 araujo joined, bluescreen10 left, bluescreen10 joined
TimToady and maybe some of the parameter backports 21:49
skids r: say not !0 +& !-1 # 0 is not a power of 2 :-) 21:50
camelia rakudo-parrot 2834f7, rakudo-jvm 2834f7, rakudo-moar 2834f7: OUTPUT«True␤»
TimToady where'd those extra !s come from? 21:51
[Coke] timotimo++ convinced me: docs.google.com/presentation/d/1ls...sp=sharing is the link, anyone can comment.
GAH. I see the hordes joining.
colomon skids: it's 2** -Inf ;)
[Coke] HAHAHA. I love anonymous llama and anonymous chinchilla. 21:52
21:52 mavcunha left
jnthn TimToady: Is this the right (start of the) output: 2 15 825 725 1925 2275 425 390 330 290 770 910 170 21:52
[Coke] I only have 30m. so there are things that are intentionally brief.
TimToady jnthn: yes 21:53
skids oh. duh. well anyway It was just pedantry.
jnthn TimToady: good :)
PerlJam [Coke]: on slide 11: "In Perl 5, you have a limited # of types." I'm not sure that's true or what you mean by that. 21:54
colomon [Coke]: page 14, what's the last line got to do with it?
diakopter pedantrics
skids diakopter++ for meta-pedantrics
jnthn [Coke]: On 8, not sure if worth calling out that NQP is a subset of Perl 6 ideal for compiler writing, to give an idea of what subset 21:55
[Coke]: In the "everything is an object" on 10, I tend to express that as "Everything can be treated as an object" these days 21:56
TimToady (4/6).Rat # (stringifies via Num) <-- I don't think this is true 21:57
jnthn [Coke]: On 13, I'd put { ... } there rather than the ; after the signature; it confused me at first
TimToady also, why .Rat something that already is
[Coke] PerlJam: I was thinking: scalar, array, hash, glob (and a few others)
jnthn: updated 8 & 10. 21:58
TimToady assigning 2**70 to an int ought to blow up, not produce 0 21:59
PerlJam [Coke]: I'd probably differentiate via Perl 5 - type systems on CPAN, Perl 6 - built-in type system. (or something like that)
TimToady slide 15, superstitious semicolons 22:00
psch [Coke]: i feel slide 25 should be somewhere before 24, maybe right before 21. you're introducing sub infix:<hi> before mentioning sub *fix:*
TimToady a rule is a method, really, not a sub
22:00 treehug88 left
TimToady and the comparison to lex/yacc is kinda wrong 22:01
[Coke] TimToady: removed the comment.
colomon: last line is showing you can invoke via locals with the same name as the params.
dalek kudo/nom: 75f0284 | jnthn++ | src/Perl6/Actions.nqp:
Fix code-gen of nested whatever curries.

Makes this construct work on MoarVM, and may be more efficient on other backends.
22:02 treehug88 joined
dalek kudo/nom: fa1e956 | moritz++ | Configure.pl:
[Configure] --backends=all, itz_++
22:02
[Coke] TimToady: leftover .Rat removed. CNP issue, I thin.
jnthn TimToady: 75f0284 should fix it
timotimo jnthn: will that still make it into our current release?
jnthn timotimo: Doubt it.
PerlJam [Coke]: slide 16 also seems out of place (or is an opportunity for more context wrt Grammars and Regex)
timotimo OK
jnthn timotimo: I think masak may have already tagged it 22:03
[Coke] PerlJam: wanted to mention rules. will really not have time to delve into them.
jnthn timotimo: Or at least, be spectesting what he intends to tag/release.
PerlJam [Coke]: okie.
timotimo right.
fair enough
timotimo is always on the bleeding edge anyway :)
jnthn Just means we've got our first ChangeLog entry for next month :P
timotimo oh, btw. a person from karlsruhe.pm is going to make sure soon that debian will have all dependencies of moarvm packaged properly
PerlJam [Coke]: how much time will you have? (I've been thinking either you've got lots of time or you're going to cover alot of stuff in a short time)
jnthn timotimo: oh, cool :) 22:04
[Coke] PerlJam: 30m 22:06
PerlJam [Coke]: on slide 20, the juxtaposition of ^20 and a list that goes from 1 to 20 weirds me out :)
timotimo oh, whoops :)
[Coke] PerlJam: yah, but then I can talk briefly about 0-indexing.
TimToady slide 23, comment should start at 0, not 1 22:07
timotimo [Coke]: for syntax highlighting, would you prefer shades of gray or colors? 22:08
[Coke] TimToady: fixed.
timotimo: never used the projector at this place. something contrasty.
... I feel compelled to add a note about the community now. :) 22:09
PerlJam [Coke]: do you want to mention Z and S on slide 24?
timotimo colors actually don't look good on white background i don't think 22:10
22:10 Guest1338 left
timotimo boldface looks bad, too 22:10
22:12 panchiniak joined
TimToady Z maybe, but S is probably not worth mentioning 22:12
grondilu this thing on google docs when we can see other people editing a document in real time is pretty cool. I wish github would do that.
[Coke] timotimo: should I change the background color?
timotimo [Coke]: only if you want to :)
[Coke] PerlJam: Sure.
PerlJam S is obscure, but maybe that's why it needs to be mentioned ;)
TimToady not in an overview
22:12 japhb joined
timotimo i'm actually not sure how to do the coloring without doing it by hand 22:13
PerlJam I dunno ... I'd at least say "and some other cool meta-ops are ..."
TimToady much more important to get the bikeshed the right color
PerlJam and then "for more information, check out the synopses"
[Coke]: oh, you should have a slide with some useful links 22:14
timotimo i'm not sure about "parsed to QAST (NQP)", actually
PerlJam or maybe just "Everything is linked from perl6.org"
[Coke] PerlJam: added urls on #30 22:15
PerlJam yeah, I see that now :)
[Coke] timotimo: would a gray work?
PerlJam I was thinking you could put the slide with links last so that you could leave it up during any question/answer period 22:16
masak is back online 22:17
up to S32 now.
PerlJam [Coke]++
22:17 kbaker left
[Coke] perljam: good point. 22:18
TimToady needs a camelia, to flush out the patriarchs
masak jnthn--'s jinx made the train stop completely :/ 22:19
22:20 treehug88 left
jnthn masak: ...what?! 22:20
masak: fail...did it break down?
22:21 skids left
TimToady oh, there's a camelia at the beginning 22:21
[Coke] TimToady: I gotcher back.
PerlJam There's enough white space on each slide that a small camelia could appear on every one. :) 22:22
diakopter r: »ö«'
[Coke] the meeting is in about an hour, btw. thanks for the last minute feedback, MOST appreciated.
camelia rakudo-jvm 2834f7: OUTPUT«(timeout)»
..rakudo-parrot 2834f7, rakudo-moar 2834f7: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Bogus statement␤at /tmp/tmpfile:1␤------> <BOL>⏏»ö«'␤ expecting any of:␤ statement list␤ prefix or term␤ p…»
masak jnthn: it stopped in le middle of nowhere, because there was only one usable track, and other trains got precedence.
jnthn masak: d'oh
timotimo [Coke]: you're very welcome :)
diakopter heh jvm timeout
masak jnthn: SJ...
jnthn Shouldn't Jinx.... 22:23
22:23 treehug88 joined
jnthn Shaln't Journey 22:23
[Coke] PerlJam: done.
diakopter Slight Jolt 22:24
22:25 woolfy1 joined, woolfy left
PerlJam [Coke]: heh, now the first slide looks a little odd with 2 Camelias :) 22:26
timotimo [Coke]: i made a copy to try out the dark theme with code "highlighting" 22:27
[Coke] timotimo++
22:27 BenGoldberg joined
[Coke] PerlJam: how about now? 22:28
timotimo [Coke]: have a look and see if i should continue working on it
PerlJam better.
timotimo it should be shared to you, too
22:28 airdisa joined
[Coke] timotimo: I'm not sure. anyone else have luck with white on black presentations with projectors of dubious origin? 22:29
timotimo worst case if it don't look right, switch over to the other copy :) 22:30
[Coke] ok. there have been some changes since you forked, I think.
masak ok, I'm on the integration tests now.
[Coke] masak++ #release work
22:30 btyler joined
[Coke] masak: oh! what number release is this? 22:30
masak read the release announcement! :P 22:31
(#72)
PerlJam [Coke]: Good luck and have fun with your presentation
jnthn [Coke]: Hope it goes well!
masak [Coke]: good luck! 22:32
[Coke] masak: added note about 72 releases.
r: say 72 / 12
camelia rakudo-jvm 2834f7: OUTPUT«(timeout)»
..rakudo-parrot 2834f7, rakudo-moar 2834f7: OUTPUT«6␤»
masak Result: PASS
[Coke] meeting is here: www.techvalleycenterofgravity.com/ 22:34
masak Rakudo tarball now available at masak.org/carl/tmp/rakudo-2014.01.tar.gz -- PerlJam, still there? ready for another upload? :)
dalek kudo/nom: 79079be | (Pepe Schwarz)++ | src/ (2 files):
Added X::TypeCheck::Argument and means for the Optimizer to throw typed exceptions.
kudo/nom: 6c83d6c | jnthn++ | src/Perl6/Optimizer.nqp:
Correct private method exception construction.

Also remove "not sure how to get it" note; you can get it with:
   class A { trusts GLOBAL; }; A!A::nope();
But only with --optimize=3 at present.
kudo/nom: 743ebf7 | jnthn++ | src/Perl6/Optimizer.nqp:
Replace comment with better explanation.
timotimo my glob this is strenuous :P
masak arrrgh!
who *commits* during release engineering!?
PerlJam masak: aye, I'll upload it 22:35
diakopter let alone pushes
jnthn masak: Somebody who knows the release engineer is (a) using git, and (b) knows how to make a branch :P
masak yo, I'm going to tag the commit I tested/tarballed, just so you know.
for all I care, your commits are post-release.
you shoulda made 'em yesterday. 22:36
PerlJam masak: uploaded.
jnthn masak: Oh, I fully expected these to go into the Feb release. :)
masak: Everything I wanted in the release, I landed *yesterday*. :) 22:37
masak PerlJam: thank you!
jnthn: ok, we are in agreement, then.
PerlJam It's all about the teamwork! :)
[Coke] jnthn++ masak++
PerlJam: what's gonna work?
PerlJam: TEEEEEAMWORK
masak p6c email sent. 22:38
proceeding to defile Wikipedia. 22:39
PerlJam masak++ thanks for being release manager :)
PerlJam now wanders off to get some food
BenGoldberg p6: say pi.Rat.denominator
masak whoever released December forgot to update the Wikipedia article...
psch jnthn: should i update the PR for roast with test cases for X::Method::NotFoud(:private)?
camelia rakudo-jvm 2834f7: OUTPUT«(timeout)»
..rakudo-parrot 2834f7, rakudo-moar 2834f7, niecza v24-109-g48a8de3: OUTPUT«113␤»
BenGoldberg Why is that number so small, rather than being close to 2**64 or whatever? 22:40
psch *NotFound
jnthn psch: Well, thing is that it's a level 3 (off by default) opt at the moment, but given we're post-release...let's make it a level 2 provided it spectests
psch: And then yes, a test would be great 22:41
psch: Do you have a roast commit bit?
psch: I can give you one so you don't need to go thorugh the pull request mechanism there...
psch jnthn: no, i do not
masak Wikipedia updated.
standing down release status.
feel free to, I dunno, start *committing to rakudo/nom* again... :P
masak celebrates with the appropriate amount of going to bed 22:42
'night, #perl6
diakopter o/
jnthn masak: 'night o/
masak++ # release 22:43
psch: You do now
22:43 kaare__ joined
[Coke] timotimo: made sure all the text was in sync enough. 22:44
psch jnthn: so i'll close the PR and push my changes myself i assume
jnthn psch: plesae
psch: I'll let you know about the NotFound thingy in a moment
psch i'll do that, thanks
[Coke] heads to the locale. 22:49
jnthn psch: haha :) When I turn on the opt, it realizes one of the method calls in S12-methods/private.t can never possible work at compile time. 22:50
psch: And so refuses to compile the test :)
psch: Can I leave updating the test to use eval and look for the typed exception to you?
dalek kudo/nom: 717a008 | jnthn++ | src/Perl6/Optimizer.nqp:
Add private method devirtualizing to default opts.
22:51
timotimo [Coke]: how much time do i have left? :)
22:51 bluescreen10 left
jnthn timotimo: I think none, given [Coke] just left... :) 22:52
psch jnthn: i'll have a look at S12-methods/private.t, although i won't promise i can figure it out; i'll get back to you
timotimo oh no!
but he has the internets
jnthn timotimo: oh, true :)
psch: Lemme know if you have trouble; I think it wants to become a throws_like and move the code into an eval or something. 22:53
clsn: Is github.com/rakudo/rakudo/pull/235 ready for me to try/review? Or was the fail I found still to fix? 22:54
psch jnthn: afair he commented on PR 237 that he fixed it 22:56
psch still has to rebuild from fiddling with <.unsp>
jnthn psch: oh, thanks. 22:57
psch no problem, i got it per mail after all... :)
timotimo done 22:58
yay, coke put a thank you onto the last slide for me %)
that was fun
23:03 araujo left, araujo joined
timotimo docs.google.com/presentation/d/1dk...sp=sharing ← this is the one with colors :) 23:03
23:05 tgt joined 23:06 Rotwang left 23:07 mavcunha joined 23:10 tgt left
dalek kudo/nom: 5527210 | (Mark Shoulson)++ | src/core/ (2 files):
Fix Parameter.perl and do Signature smart-matching.
23:13
rakudo/nom: 574613b | (Mark Shoulson)++ | src/core/Signature.pm:
rakudo/nom: Catch another case of Signature matching with slurpies.
23:13 dalek left 23:14 dalek joined, ChanServ sets mode: +v dalek
jnthn There's two pull requests down :) 23:14
timotimo clsn++ # signature work :) 23:15
i'm thinking about making introspection easier all the way down by giving gist and perl methods to the BootstrapFoobar classes
jnthn Indeed
psch oh boy, i just now notice the extend of my git-mess
i can't push a rebase of rakudo/nom into my own nom because there's a patch missing that doesn't introduce difference... or something like that 23:16
jnthn timotimo: I'd rather not
timotimo jnthn: fair enough. bloat reasons?
psch but seeing as everything i have is merge i can just trash my fork and fork again, right?
*merged
jnthn psch: Yes
BenGoldberg p6: Rat.new((pi * 2**59).Int, 2**59).denominator.say;
camelia rakudo-parrot fa1e95, rakudo-jvm fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«281474976710656␤» 23:17
BenGoldberg p6: Rat.new((pi * 2**60).Int, 2**60).denominator.say;
timotimo also, you can just make a new branch at those "borked" commits and they'll never be lost
camelia rakudo-parrot fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«281474976710656␤»
..rakudo-jvm fa1e95: OUTPUT«288230376151711744␤»
BenGoldberg doesn't get why jakudo's output is different there.
jnthn timotimo: That and you can't easily get at such things unless you go meta-programming at least a level deep anyway. 23:18
psch timotimo: the problem is i reverted my commit that should have gone into a branch in nom, and now i cant merge upstream that has that commit into my nom
jnthn timotimo: All the things you get through .HOW on a Perl 6 object have a .perl and .gist, for example.
timotimo psch: then git reset --hard nom origin/nom
psch timotimo: that might have worked, i nuked it already though 23:19
lue BenGoldberg: different math libs, I'm guessing
23:19 jeffreykegler joined, Mouq left
psch the problem i had happened when pushing into my fork, in any case, i got a 'rejected' because my HEAD was behind 23:19
timotimo oh
then push -f it :)
psch ...right, that exists 23:20
BenGoldberg The docs for Rat, at doc.perl6.org/type/Rat, say that the denominator is supposed to be a UInt64. It ought to be able to hold values up to 2**63 with no problem.
timotimo BenGoldberg: i think currently it's falsely implemented as Int64 instead
BenGoldberg Even so, funny things happening at 2**60 is odd
timotimo oh, right 23:21
psch only readily uses -f on rm $dirwithsubdirs, which might actually the worst readily use of -f possible
s/y t/y be t/
BenGoldberg p: say 281474976710656.log / 2.log 23:23
camelia rakudo-parrot fa1e95: OUTPUT«48␤»
grondilu perl6: Rat.new(1, 2**60).denominator.say;
camelia rakudo-parrot fa1e95, rakudo-jvm fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«1152921504606846976␤»
grondilu perl6: Rat.new(3**20, 2**60).denominator.say;
camelia rakudo-parrot fa1e95, rakudo-jvm fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«1152921504606846976␤»
lue That reminds me. TimToady: UInt64 looks to me to be unnecessary. There was a discussion here a couple days back on it.
grondilu perl6: say (pi*2**60).Int; 23:24
camelia rakudo-jvm fa1e95: OUTPUT«3622009729038561300␤»
..rakudo-parrot fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«3622009729038561280␤»
grondilu BenGoldberg: I think that's your explanation. The difference is in the numerator, not in the denominator. 23:25
but it appears in the denominator after simplification of the fraction. 23:26
perl6: say (pi*2**59).Int;
camelia rakudo-parrot fa1e95, rakudo-jvm fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«1811004864519280640␤»
grondilu see? all rakudos agree with pi*2**59, but not with pi*2**60 23:27
which is related with the different ways they store the pi constant, I guess. 23:28
lue p6: say pi
camelia rakudo-jvm fa1e95: OUTPUT«3.141592653589793␤»
..rakudo-parrot fa1e95: OUTPUT«3.14159265358979␤»
..niecza v24-109-g48a8de3: OUTPUT«3.1415926535897931␤»
..rakudo-moar fa1e95: OUTPUT«3.141593␤»
jnthn r: say pi.WHAT
camelia rakudo-parrot fa1e95, rakudo-jvm fa1e95, rakudo-moar fa1e95: OUTPUT«(Num)␤»
timotimo i can see which slide coke is on :D
BenGoldberg perl6: say pi - 3.141592653589793 23:29
camelia rakudo-parrot fa1e95, rakudo-jvm fa1e95, rakudo-moar fa1e95, niecza v24-109-g48a8de3: OUTPUT«0␤»
timotimo at least i think that's what i'm seeing
lue p: say π
camelia rakudo-parrot fa1e95: OUTPUT«===SORRY!=== Error while compiling /tmp/HJgZ8TFYRC␤Undeclared routine:␤ π used at line 1␤␤»
BenGoldberg perl6: say pi - 3.14159265
camelia niecza v24-109-g48a8de3: OUTPUT«3.5897929073769319E-09␤»
..rakudo-jvm fa1e95: OUTPUT«3.589792907376932E-9␤»
..rakudo-parrot fa1e95: OUTPUT«3.58979290737693e-09␤»
..rakudo-moar fa1e95: OUTPUT«0␤»
BenGoldberg perl6: say pi - 3.14159265.Num
camelia niecza v24-109-g48a8de3: OUTPUT«3.5897929073769319E-09␤»
..rakudo-jvm fa1e95: OUTPUT«3.589792907376932E-9␤»
..rakudo-parrot fa1e95: OUTPUT«3.58979290737693e-09␤»
..rakudo-moar fa1e95: OUTPUT«0␤»
grondilu p6: say (pi - 3.14159265)/pi 23:30
camelia rakudo-parrot fa1e95: OUTPUT«1.14266657177053e-09␤»
..niecza v24-109-g48a8de3: OUTPUT«1.1426665717705302E-09␤»
..rakudo-jvm fa1e95: OUTPUT«1.1426665717705302E-9␤»
..rakudo-moar fa1e95: OUTPUT«0␤»
BenGoldberg perl6: say (pi - 3.14159265) * 1e9
camelia rakudo-moar fa1e95: OUTPUT«3.589793␤»
..rakudo-jvm fa1e95: OUTPUT«3.589792907376932␤»
..niecza v24-109-g48a8de3: OUTPUT«3.5897929073769319␤»
..rakudo-parrot fa1e95: OUTPUT«3.58979290737693␤»
BenGoldberg is annoyed at how few significant digits moar prints out by default. Sadly, it's not specced to anything in particular, yet 23:31
grondilu m: say pi.perl
camelia rakudo-moar 717a00: OUTPUT«3.141593e0␤»
BenGoldberg One would think that Num.perl would print out a long enough decimal representation that .Num-ing the string could reconstruct a value equal to the original Num, within one or two bits. 23:33
23:33 dmol left
psch m: my $x = 0.12345678910; $x.perl.say 23:34
camelia rakudo-moar 717a00: OUTPUT«0.1234567891␤»
lue BenGoldberg: what makes you say it doesn't?
psch i think this more about how many significant digits are present in the implementation 23:35
BenGoldberg If moar, internally, had tiny floats for Num, then:
perl6: say (pi - 3.14159265) * 1e9 23:36
colomon BenGoldberg: pi is a Num, I believe
camelia rakudo-moar 717a00: OUTPUT«3.589793␤»
..rakudo-jvm 717a00: OUTPUT«3.589792907376932␤»
..niecza v24-109-g48a8de3: OUTPUT«3.5897929073769319␤»
..rakudo-parrot 717a00: OUTPUT«3.58979290737693␤»
colomon p: say pi.WHAT
camelia rakudo-parrot 717a00: OUTPUT«(Num)␤»
BenGoldberg ... would produce a zero for moar, instead of 3.58793
colomon r: say pi.perl 23:37
camelia rakudo-moar 717a00: OUTPUT«3.141593e0␤»
..rakudo-jvm 717a00: OUTPUT«3.141592653589793e0␤»
..rakudo-parrot 717a00: OUTPUT«3.14159265358979e0␤»
colomon though yeah, that's a moar-bug
lue m: say pi; say pi - 3.14159265; say (pi - 3.14159265) * 1e9
camelia rakudo-moar 717a00: OUTPUT«3.141593␤0␤3.589793␤»
colomon r: say (1/3).Num.perl
camelia rakudo-moar 717a00: OUTPUT«0.333333e0␤»
..rakudo-jvm 717a00: OUTPUT«0.3333333333333333e0␤»
..rakudo-parrot 717a00: OUTPUT«0.333333333333333e0␤»
lue That second number should not be zero.
BenGoldberg perl6: my $p = (pi - 3.14159265); say $p.gist, ' ', $p.perl, ' ', $p.fmt('%.15f') 23:39
camelia niecza v24-109-g48a8de3: OUTPUT«3.5897929073769319E-09 3.5897929073769319E-09 0.000000003589793␤»
..rakudo-moar 717a00: OUTPUT«0 0e0 0.000000003589793␤»
..rakudo-parrot 717a00: OUTPUT«3.58979290737693e-09 3.58979290737693e-09 0.000000003589790␤»
..rakudo-jvm 717a00: OUTPUT«3.589792907376932E-9 3.589792907376932E-9 0.000000003589793␤»
timotimo google docs does show coke on the first slide constantly apparently 23:40
psch 30 minutes camelia :)
BenGoldberg perl6: my $p = (pi - 3.14159265); say $p.fmt('%.15e') 23:41
camelia rakudo-parrot 717a00: OUTPUT«3.589792907376930e-09␤»
..rakudo-jvm 717a00, rakudo-moar 717a00: OUTPUT«3.589792907376932e-09␤»
..niecza v24-109-g48a8de3: OUTPUT«3.589792907376932e-009␤»
timotimo :D
now coke dropped off the 'net
jnthn BenGoldberg: Got a MoarVM patch coming up that gives us some more digits.
colomon will stop searching through the NQP source trying to figure out where that code is. 23:42
BenGoldberg perl6: my $p = (pi - 3.14159265358979290737693); say $p.fmt('%.15e')
camelia niecza v24-109-g48a8de3: OUTPUT«4.440892098500626e-016␤»
..rakudo-parrot 717a00: OUTPUT«4.440892098500630e-16␤»
..rakudo-moar 717a00: OUTPUT«4.440892098500626e-16␤»
..rakudo-jvm 717a00: OUTPUT«0.000000000000000e+00␤»
BenGoldberg perl6: my $p = (pi - 3.14159265358979290737693444089209850062); say $p.fmt('%.15e') 23:44
camelia rakudo-parrot 717a00, rakudo-jvm 717a00, rakudo-moar 717a00: OUTPUT«0.000000000000000e+00␤»
..niecza v24-109-g48a8de3: OUTPUT«0.000000000000000e+000␤»
BenGoldberg :)
I think that perl6 should define pi to be 3.14159265358979290737693444089209850062, as a Rat ;) 23:45
BenGoldberg is only slightly joking 23:46
colomon BenGoldberg: that's too precise for a Rat, isn't it?
timotimo the question is: what do you use pi for normally and how well do rats work in that case?
colomon FatRat, now...
timotimo do our trigonometric operations handle rats well?
BenGoldberg r: say 3.14159265358979290737693444089209850062.nude 23:47
camelia rakudo-parrot 717a00, rakudo-jvm 717a00, rakudo-moar 717a00: OUTPUT«157079632679489645368846722044604925031 50000000000000000000000000000000000000␤»
raydiak why use so many digits if the last 2/3rds are wrong? :)
BenGoldberg r: say pi.Rat.nude
camelia rakudo-parrot 717a00, rakudo-jvm 717a00, rakudo-moar 717a00: OUTPUT«355 113␤»
lue I think we should define a π :) 23:48
BenGoldberg r: say pi.Rat.Num.fmt("%.30e")
camelia rakudo-jvm 717a00: OUTPUT«3.00000000001.415929203539825E29e+00␤»
..rakudo-parrot 717a00: OUTPUT«3.000000000001.4159292035398e+29e+00␤»
..rakudo-moar 717a00: OUTPUT«3.000000000009223391921479614464e+00␤»
BenGoldberg r: say pi.Rat.Num.fmt("%.15e")
camelia rakudo-jvm 717a00, rakudo-moar 717a00: OUTPUT«3.141592920353983e+00␤»
..rakudo-parrot 717a00: OUTPUT«3.141592920353980e+00␤»
jnthn BenGoldberg: MoarVM HEAD has the patch now. :)
timotimo great! :) 23:49
colomon jnthn++
BenGoldberg r: say (pi - pi.Rat.Num).fmt("%.15e")
camelia rakudo-parrot 717a00: OUTPUT«-2.667641894049670e-07␤»
..rakudo-jvm 717a00, rakudo-moar 717a00: OUTPUT«-2.667641894049666e-07␤»
jnthn Also one spectest fail went away thanks to it :)
timotimo jnthn: you could draft up a prototype for the re-structured optimizer perhaps? 23:50
23:50 grondilu_ joined, ajr_ left
jnthn timotimo: Perhaps, though I think there's a few more pressing things than that. 23:51
timotimo sure 23:52
you were asking for things to do, so i just grabbed that off the top of my head :)
jnthn timotimo: Well, there's plenty to do, I'm just kinda tired today after staying up quite late working on Moar release today, so was looking for LHF :)
timotimo heh. that's absolutely fair 23:53
23:54 grondilu left
jnthn If anybody is looking to take a fairly manageable task off my list: bring the perl6-debug frontend (not the Perl 6 UI module, just the perl6-debug program itself) into the Rakudo repo and get it so we build it on all 3 backends, generating and installing a fakeexecutable on Parrot and a runner on Moar and JVM. 23:55
grondilu_ pi has no unversally accepted numeric approximation, so honestly I'm not even sure it's a good idea to put one in the core. If we do, we might as well put one that is voluntarliy not too accurate. Like constant pi = <355/113>.
lue 22/7 is the conventional fractional form of pi :) 23:56
jnthn Bonus: you most likely get to be the person to thank for us having perl6-debug on MoarVM :)
lue jnthn: "not the Perl 6 UI module" what do you mean by that? Is there a GUI I missed? 23:57
23:57 grondilu_ is now known as grondilu 23:58 denisboyun joined
jnthn lue: The debugger is broken up into a small NQP program that inserts debug hooks into AST, and a larger Perl 6 module that subscribes to them and does the actual user interaction. 23:58
23:58 grondilu left, grondilu joined
timotimo jnthn: how do we feel about loading .moarvm files from any kind of archive, compressed or otherwise? 23:58
jnthn lue:It's a command line interface at the moment. The design is such that we can build other ones.
timotimo: Negative.
timotimo: Ruins mmap
lue ah, so you want the debugger hooks in rakudo, but not the specific CUI program used for 'perl6-debug' ? 23:59
timotimo then we should just make sure to make the moarvm files small-ish :3
jnthn timotimo: Well, the integer thing in the serialization blob may be a big win.
timotimo right. only in the serialized blob though, yes?
jnthn lue: Well, perl6-debug is the thing I want brought in.
timotimo: right