»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.
timotimo i think the spectests are clean so far and will stay clean until the end 00:00
00:01 rurban1 joined
japhb r: my $r = 1.969; $r.fmt("%.{$_}f").say for 0..3 00:02
camelia rakudo-jvm 7142b7: OUTPUT«1.␤1.␤1.␤1.␤»
..rakudo-parrot 7142b7: OUTPUT«1.␤1.0␤1.97␤1.969␤»
colomon :(\ 00:03
japhb Three bugs there, I think.
colomon three?
oh
yow
colomon hopes they are not his fault 00:04
\
japhb rakudo-jvm always wrong, r-p wrong for .1f, and failure to drop trailing . when no remaining after-. digits
Free karma to the person who turns that into spec tests, double if they actually fix at least the r-p .1f bug, triple if they fix all of them. I'm feeling expansive tonight. ;-) 00:05
colomon shouldn't first one be 2?
japhb Yeah, that's four bugs.
Sheesh.
colomon basically, it's just complete fail 00:06
japhb Sadly, yes, I think so.
:-(
colomon does sprintf fo the same thing
japhb Hadn't tried it. Just had golfed a fail in a larger problem down to that. 00:07
s/problem/program/; Well, I guess that depends on how you look at it.
colomon oh
Niecza's results:
1.969000
2.0
1.97
1.969
00:07 pjcj left
japhb Interesting. And much better. 00:07
colomon not sure if that's a niecza bug on the first or not. 00:08
japhb Though that first one is indeed odd, at least it's not complete fail.
I would think .0 should mean no decimals just like in sprintf, though.
r: my $r = 1.969; $r.sprintf("%.{$_}f").say for 0..3 00:09
camelia rakudo-parrot 7142b7: OUTPUT«Too few directives: found 0, fewer than the 1 arguments after the format string␤ in any TOP at gen/parrot/stage2/NQPHLL.nqp:2102␤ in any !reduce at gen/parrot/stage2/QRegex.nqp:724␤ in any !cursor_pass at gen/parrot/stage2/QRegex.nqp:686␤ in …»
..rakudo-jvm 7142b7: OUTPUT«Too few directives: found 0, fewer than the 1 arguments after the format string␤ in any TOP at gen/jvm/stage2/NQPHLL.nqp:2056␤ in any !reduce at gen/jvm/stage2/QRegex.nqp:724␤ in any !cursor_pass at gen/jvm/stage2/QRegex.nqp:686␤ in any parse at…»
japhb r: my $r = 1.969; sprintf("%.{$_}f", $r).say for 0..3
camelia rakudo-jvm 7142b7: OUTPUT«1.␤1.␤1.␤1.␤» 00:10
..rakudo-parrot 7142b7: OUTPUT«1.␤1.0␤1.97␤1.969␤»
japhb That was a way LTA error message, BTW.
I pretty much solved it with a WAG.
Looks like rakudo does the same thing with .fmt and with &sprintf
colomon probably .fmt just calls sprintf
okay, I'm going to add sprintf tests 00:12
japhb_ Thank you, colomon++! 00:13
colomon you're not out of this yet, you need to help me figure out what the correct answers are.
japhb_ Without any irony, "whatever perl5 says". 00:17
Going AFK for a bit
lue r: my $a = 4; say 32.fmt("%0{$a}X") 00:18
camelia rakudo-jvm 7142b7: OUTPUT«use of uninitialized value of type Any in string context␤Too few directives: found 0, fewer than the 1 arguments after the format string␤ in any TOP at gen/jvm/stage2/NQPHLL.nqp:2056␤ in any !reduce at gen/jvm/stage2/QRegex.nqp:724␤ in any !curso…» 00:19
..rakudo-parrot 7142b7: OUTPUT«use of uninitialized value of type Any in string context in block at /tmp/fHut4SfDu3:1␤␤Too few directives: found 0, fewer than the 1 arguments after the format string␤ in any TOP at gen/parrot/stage2/NQPHLL.nqp:2102␤ in any !reduce at gen/pa…»
lue fmt has a general problem with interpolation.
00:19 BenGoldberg left
dalek ast: 49d91df | (Solomon Foster)++ | S32-str/sprintf.t:
sprintf tests on rounding numbers like 1.969.

Tests issues found by japhb.		 00:23
00:23 BenGoldberg joined
colomon afk # time to make up a bedtime story for the 5yo 00:24
dalek ast: af65b62 | (Geoffrey Broadwell)++ | S32-str/sprintf.t:
Fix description copy pasta in new sprintf tests
 00:33
00:34 pjcj joined
BenGoldberg Copy pasta? 00:37
Is that served with copy pesto? 00:40
BenGoldberg clearly needs to eat, as his noodle is generating too many thoughts of food. 00:42
geekosaur usually spaghetti code :p
^part of 00:43
00:46 gfldex left 00:51 jeff_s1 left 00:52 jeff_s1 joined 00:59 gfldex joined 01:03 btyler joined
timotimo okay so i've run the spectests, but for jvm i got only "no subtests run" after a bunch of tests 01:05
preflex timotimo: you have 1 new message. '/msg preflex messages' to read it.
dalek kudo/chained_whatevercode: dfeff62 | (Timo Paulssen)++ | src/Perl6/Actions.nqp:
fix whatever currying for chain ops

When a WhateverCode gets created from a chain op, it'll be annotated with its original arguments and when a WhateverCode with a chain op ends up calling a WhateverCode that also had a chain in it, it will "flatten out" the arguments of that WhateverCode, creating a proper chain.		 01:10
timotimo i would appreciate spectesting this and other testings
01:13 hummeleB1 joined 01:14 thou left
colomon mmmm, pesto 01:18
dalek -pb: 13a19c0 | (Geoffrey Broadwell)++ | lib/PB/Binary/Reader.pm:
Continue work on PB::Binary::Reader

Getting closer, but still a bit more to go before it's testable.		 01:22
colomon peanut butter binary reader? 01:23
01:26 btyler left
lue colomon: I keep thinking PowerBook when I see PB in allcaps, for some reason. 01:26
01:53 tgt left
timotimo r: say sub { 5 }() 01:54
camelia rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/W1bbYGrU9C␤Two terms in a row␤at /tmp/W1bbYGrU9C:1␤------> say sub { 5 }⏏()␤ expecting any of:␤ postfix␤ statement end␤ statement modifie…»
..rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/KhRVNfW1kz␤Two terms in a row␤at /tmp/KhRVNfW1kz:1␤------> say sub { 5 }⏏()␤ expecting any of:␤ postfix␤ statement end␤ statement modifier…»
timotimo r: .say for < $< $> > 01:55
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«$<$>␤»
timotimo i'm not exactly sure how that's supposed to work anyway o_O
BenGoldberg r: .say < $foo > 01:57
TimToady it's counting brackets, but I don't know why it's not splitting into two words
camelia rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/ikwHnG_wtt␤Variable '$foo' is not declared␤at /tmp/ikwHnG_wtt:1␤------> .say < $foo⏏ >␤ expecting any of:␤ postfix␤»
..rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/O4mw2Goi7G␤Variable '$foo' is not declared␤at /tmp/O4mw2Goi7G:1␤------> .say < $foo⏏ >␤ expecting any of:␤ postfix␤»
BenGoldberg r: .say < $< >
camelia rakudo-parrot 455fbc: OUTPUT«Nil␤use of uninitialized value of type Nil in numeric context in block at /tmp/duhm53AajH:1␤␤»
..rakudo-jvm 455fbc: OUTPUT«Nil␤use of uninitialized value of type Nil in numeric context␤»
BenGoldberg r: .say < $> >
camelia rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/rTAD3C9adx␤Unsupported use of $> variable; in Perl 6 please use $*EUID␤at /tmp/rTAD3C9adx:1␤------> .say < $>⏏ >␤»
..rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/_fhYVsyPGL␤Unsupported use of $> variable; in Perl 6 please use $*EUID␤at /tmp/_fhYVsyPGL:1␤------> .say < $>⏏ >␤»
TimToady .say doesn't take arguments 01:58
BenGoldberg r: say < $< >
camelia rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/Z8_xiMDE1l␤Unable to parse expression in quote words; couldn't find final '>'␤ ␤at /tmp/Z8_xiMDE1l:1␤------> say < $< >⏏<EOL>␤ expecting any of:␤ arg…»
..rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/rE9vxqS7ZM␤Unable to parse expression in quote words; couldn't find final '>'␤ ␤at /tmp/rE9vxqS7ZM:1␤------> say < $< >⏏<EOL>␤ expecting any of:␤ …»
TimToady r: say < $< $> >.elems
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«1␤»
TimToady r: say < < > >.elems 01:59
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«1␤»
TimToady r: say < < > >.[0]
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«<>␤»
TimToady r: say < < > >.[0].SHAT
camelia rakudo-jvm 455fbc: OUTPUT«No such method 'SHAT' for invocant of type 'Str'␤ in block at /tmp/5b9FMqrvzb:1␤ in any eval at gen/jvm/stage2/NQPHLL.nqp:1086␤ in any evalfiles at gen/jvm/stage2/NQPHLL.nqp:1292␤ in any command_eval at gen/jvm/stage2/NQPHLL.nqp:1196␤ in any…»
..rakudo-parrot 455fbc: OUTPUT«No such method 'SHAT' for invocant of type 'Str'␤ in block at /tmp/yVxpzCfuJC:1␤ in any at /tmp/yVxpzCfuJC:1␤ in any at gen/parrot/stage2/NQPHLL.nqp:1146␤ in any eval at gen/parrot/stage2/NQPHLL.nqp:1133␤ in any evalfiles at gen/parrot/…»
TimToady r: say < < > >.[0].WHAT
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«(Str)␤»
TimToady I don't see any rational for turning < < > > into '<>' 02:00
timotimo r: my $foo = set <a b c d>; $foo (|)= (set <b c d e>); say $foo;
camelia rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/LqJHQYijeq␤Two terms in a row␤at /tmp/LqJHQYijeq:1␤------> my $foo = set <a b c d>; $foo ⏏(|)= (set <b c d e>); say $foo;␤ expecting any of:␤ postfix␤ …»
..rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/ByhJwEmQii␤Two terms in a row␤at /tmp/ByhJwEmQii:1␤------> my $foo = set <a b c d>; $foo ⏏(|)= (set <b c d e>); say $foo;␤ expecting any of:␤ postfix␤ …» 02:01
timotimo weird.
02:01 sunnavy left 02:02 thou joined
colomon timotimo: works in Niecza. ;) 02:06
TimToady r: my $foo = set <a b c d>; $foo ∪= set <b c d e>; say $foo 02:11
camelia rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/vPyKQfqLHP␤Two terms in a row␤at /tmp/vPyKQfqLHP:1␤------> my $foo = set <a b c d>; $foo ⏏∪= set <b c d e>; say $foo␤ expecting any of:␤ postfix␤ i…»
..rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/o6D4gdD4BV␤Two terms in a row␤at /tmp/o6D4gdD4BV:1␤------> my $foo = set <a b c d>; $foo ⏏∪= set <b c d e>; say $foo␤ expecting any of:␤ postfix␤ …»
TimToady that should be working 02:14
./src/core/set_operators.pm:only sub infix:<<"\x222A">>(|p) {
dalek p: 529c7b3 | (Solomon Foster)++ | t/hll/06-sprintf.t:
More tests based on bug found by japhb.
colomon r: my $foo = set <a b c d>; $foo = $foo ∪ set <b c d e>; say $foo 02:15
camelia rakudo-jvm 455fbc: OUTPUT«set(b, a, d, c, e)␤»
..rakudo-parrot 455fbc: OUTPUT«set(a, b, c, d, e)␤»
TimToady interesting
colomon TimToady: it's just using op= that breaks it
TimToady why should op= care whether op is unicode? 02:16
colomon it doesn't, (|)= doesn't work either
TimToady (well, it's all unicode, but non-ascii)
is rakudo not implement op= as a meta-operator?
*menting
or is this more "not quite" damage? 02:17
colomon TimToady: it should be, I think it's just wigging out on this
timotimo Result: PASS
t/spec/S02-types/whatever.rakudo.parrot (Wstat: 0 Tests: 83 Failed: 0) TODO passed: 79-80
so i just need someone to spectest chained_whatevercode on jvm. lizmat perhaps? 02:18
colomon r: sub infix:<XMAS>($a, $b) { $a * $b; }; my $r = 10; $r XMAS= 5; say $r;
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«50␤»
colomon TimToady: wonder if the problem has something to do with it being a list op?
timotimo bedtimes! 02:19
TimToady r: sub infix:<💩>($a,$b) { $a * $b }; my $r = 10; $r 💩= 5; say $r
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«50␤»
colomon what the heck is that?
.u 💩
128169, eh? 02:21
oh
My font is not showing that very clearly.
www.fileformat.info/info/unicode/ch.../index.htm 02:22
TimToady just as well :)
r: my @a = <a b c>; @a Z~= 1,2,3; say @a 02:26
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«a b c␤»
TimToady r: my @a = <a b c>; @a [Z~]= 1,2,3; say @a
camelia rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/WaQAHFKoEq␤Preceding context expects a term, but found infix = instead␤at /tmp/WaQAHFKoEq:1␤------> my @a = <a b c>; @a [Z~]=⏏ 1,2,3; say @a␤»
..rakudo-parrot 455fbc: OUTPUT«===SORRY!=== Error while compiling /tmp/of9Ftw8Pnn␤Preceding context expects a term, but found infix = instead␤at /tmp/of9Ftw8Pnn:1␤------> my @a = <a b c>; @a [Z~]=⏏ 1,2,3; say @a␤»
TimToady r: my @a = <a b c>; @a Z= 1,2,3; say @a
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«a b c␤»
TimToady r: my @a = <a b c>; @a Z[~=] 1,2,3; say @a
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«a b c␤»
02:44 _ilbot left 02:46 _ilbot joined 02:48 gfldex left 02:50 gfldex joined 03:13 gfldex left 03:15 gfldex joined 03:18 fridim_ joined 03:19 fridim_ left 03:34 araujo left, xinming_ joined 03:36 fridim_ joined 03:37 xinming left 03:39 hummeleB1 left
lue wonders if he should push S15 tests to a branch initially, since he's only got rakudo on-hand to fudge things from the start. 03:43
colomon warning to nqp folks, I added tests that fail in the NQP sprintf tests. I'd hoped to have them fixed tonight, but I'm giving up and going to bed. 03:44
lue niecza: say "hello! o/" 03:47
camelia niecza v24-108-g17d73e4: OUTPUT«(timeout)[auto-compiling setting]␤»
lue ♞ colomon o/
04:02 xinming joined 04:04 xinming_ left 04:07 preflex_ joined, ChanServ sets mode: +v preflex_, preflex left 04:08 preflex_ is now known as preflex 04:12 lumimies left 04:13 lumimies joined 04:14 ivanshmakov left, geekosaur left, ivanshma` joined 04:17 flussence joined 04:19 atrodo joined
dalek ast: 361d850 | lue++ | S15-strings/ (3 files):
Add the first S15 tests.

Not nearly enough, but hopefully serves as a start for others to chip in. At the very least, I expect those more versed in test writing to fix my blunders :) .		 04:19
04:19 geekosaur joined
lue I have to say, what I wrote wasn't at all impressive, but I already felt the power of Moar Unicode in Perl 6 :) 04:20
TimToady :) 04:21
lue TimToady: would it make sense for chr and chrs to return Uni objects, seeing as there's no telling what NF the numbers are in? Or would auto-conversion to NFG actually be appreciated? 04:23
(my thinking is that people with a specific list of codepoints would like to preserve that, even after conversion to a string-like object.) 04:26
japhb lue: FWIW, that thinking makes sense to me 04:27
lue It has a bonus of discouraging people from trying to feed a Str-making function negative numbers, because it won't be generating a Str in the first place. 04:29
04:35 atroxaper joined
TimToady can't see anything wrong with doing it that way offhand 04:35
dalek ecs: db99244 | lue++ | S15-unicode.pod:
[S15] Change .chr(s) to return Uni, not Str.

This is because Uni is the only string type that won't convert strings given to it, which for people who are playing with numeric arrays of codepoints is probably important.		 04:38
lue gets the sense that he'll find himself reading the Unicode standard, the whole thing, in the near future :) . 04:43
I also wouldn't be surprised if, in the next couple of months, I finally send in the CLA and end up doing most of the implementation work on S15 :) . 04:46
japhb The whole Unicode standard? That's ... a lot. 04:48
lue japhb: I didn't say I'd read it *quickly* :) . 04:51
04:51 xenoterracide left
japhb lue: Well, you have to read it faster than they produce more of it, or you'll never get done. :-) 04:51
lue japhb: The Standard is good about not making drastic or previous-knowledge-deprecating changes most of the time. So at least that. 04:53
.oO(Unicode 6.3 is technically a just a list of changes since 6.2, since IIRC they didn't bother reprinting the whole deal for 6.3 . Next version will be 7.0 judging by the available UCD files.)
 04:54
BenGoldberg r: sub infix:<OTFLOL> ($a, $b) { $a, $b }; say [OTFLOL] 1..4; say [ROTFLOL] 1..4; 05:10
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«1 2 3 4␤4 3 2 1␤»
BenGoldberg r: sub infix:<OTFLOL> ($a, $b) { $a, $b }; .say for [ROTFLOL] 1..4; 05:14
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«4␤3 2 1␤»
BenGoldberg is confused.
What is the 4 seperate from 3 2 1?
05:17 xenoterracide joined 05:24 Pleiades` left 05:29 Pleiades` joined 05:32 jeff_s1 left 05:33 xenoterracide left, jeff_s1 joined 05:37 xinming_ joined 05:39 xinming left 05:43 xinming joined, xinming_ left 06:37 BenGoldberg left 06:40 atroxaper left 06:41 atroxaper joined 06:46 darutoko joined
lue
.oO(Fun Fact: despite what you might initially believe, mu/util/podhtm.pl only generates one HTML file in the synopses conversion process.)
 06:48
06:58 kaare__ joined 06:59 Tene left, djanatyn left 07:00 djanatyn joined, Tene joined
lue moritz: for the synopses, would you prefer that the new HTML generation toolchain be outside of mu much as possible? If so, I assume using panda to obtain the stuff is preferred? 07:07
♘ all o/
07:09 rurban1 left 07:14 atroxaper left 07:15 atroxaper joined
moritz lue: I don't care much where the stuff is 07:21
lue: though come to think of it, I'd prefer a repo in the perl6 organization 07:23
though I can fork repos into that org if necessary 07:24
07:26 atroxaper left 07:27 atroxaper joined 07:29 SamuraiJack_ joined 07:31 araujo joined 07:33 araujo left
TimToady r: sub infix:<OTFLOL> (\a, \b) { a, b }; .perl.say for [ROTFLOL] 1..4; 07:33
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«4␤3␤2␤1␤»
07:33 araujo joined, araujo left, araujo joined 07:34 slavik left
TimToady r: sub infix:<OTFLOL> (*@a) { @a }; .say for [ROTFLOL] 1..4; 07:36
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«4␤3␤2␤1␤»
TimToady or that
zzz &
07:42 camelia left 07:46 camelia joined 07:48 ChanServ sets mode: +v camelia 08:01 thou left 08:02 camelia left 08:03 raydiak_ left 08:10 rurban1 joined 08:14 nnunley left, rurban1 left 08:19 camelia joined, raydiak_ joined 08:20 ChanServ sets mode: +v camelia 08:28 atroxaper left
timotimo r: (say [X] ($(1, 2, 3) xx $_)) for 1..4; 08:45
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«1 2 3␤1 1 1 2 1 3 2 1 2 2 2 3 3 1 3 2 3 3␤1 1 1 1 1 2 1 1 3 1 2 1 1 2 2 1 2 3 1 3 1 1 3 2 1 3 3 2 1 1 2 1 2 2 1 3 2 2 1 2 2 2 2 2 3 2 3 1 2 3 2 2 3 3 3 1 1 3 1 2 3 1 3 3 2 1 3 2 2 3 2 3 3 3 1 3 3 2 3 3 3␤1 1 1 1 1 1 1 2 1 1 1 3 1…»
timotimo r: (say ([X] ($(1, 2, 3) xx $_)).perl) for 1..4; 08:46
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«((1,), (2,), (3,)).list␤((1, 1), (1, 2), (1, 3), (2, 1), (2, 2), (2, 3), (3, 1), (3, 2), (3, 3)).list␤((1, 1, 1), (1, 1, 2), (1, 1, 3), (1, 2, 1), (1, 2, 2), (1, 2, 3), (1, 3, 1), (1, 3, 2), (1, 3, 3), (2, 1, 1), (2, 1, 2), (2, 1, …»
timotimo that's pretty darn nifty.
although i think i have to force all these operators into awkward precedence :D
r: say ([X] ($(1, 2, 3) xx $_)).perl for 1..4; 08:47
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«((1,), (2,), (3,)).list␤((1, 1), (1, 2), (1, 3), (2, 1), (2, 2), (2, 3), (3, 1), (3, 2), (3, 3)).list␤((1, 1, 1), (1, 1, 2), (1, 1, 3), (1, 2, 1), (1, 2, 2), (1, 2, 3), (1, 3, 1), (1, 3, 2), (1, 3, 3), (2, 1, 1), (2, 1, 2), (2, 1, …»
timotimo r: say ([X] $(1, 2, 3) xx $_).perl for 1..4;
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«((1,), (2,), (3,)).list␤((1, 1), (1, 2), (1, 3), (2, 1), (2, 2), (2, 3), (3, 1), (3, 2), (3, 3)).list␤((1, 1, 1), (1, 1, 2), (1, 1, 3), (1, 2, 1), (1, 2, 2), (1, 2, 3), (1, 3, 1), (1, 3, 2), (1, 3, 3), (2, 1, 1), (2, 1, 2), (2, 1, …»
timotimo actually i don't. huh, pretty cool!
r: for ([X] $(1, 2) xx 4).tree -> \foo { say foo }; 08:50
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«1 1 1 1␤1 1 1 2␤1 1 2 1␤1 1 2 2␤1 2 1 1␤1 2 1 2␤1 2 2 1␤1 2 2 2␤2 1 1 1␤2 1 1 2␤2 1 2 1␤2 1 2 2␤2 2 1 1␤2 2 1 2␤2 2 2 1␤2 2 2 2␤»
timotimo r: for ([X] $(1, 2) xx 4) -> \foo { say foo };
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«1␤1␤1␤1␤1␤1␤1␤2␤1␤1␤2␤1␤1␤1␤2␤2␤1␤2␤1␤1␤1␤2␤1␤2␤1␤2␤2␤1␤1␤2␤2␤2␤2␤1␤1␤1␤2␤1␤1␤2␤2␤1␤2␤1␤2␤1␤2␤2␤2␤2␤1␤1␤2␤2␤1␤2␤2␤2␤2…»
08:53 rindolf joined
timotimo r: eval '{ 1 < $^a < 3 }' 08:54
camelia rakudo-jvm 455fbc: OUTPUT«Not enough positional parameters passed; got 0 but expected 1␤ in block at eval_0:1␤ in sub eval at gen/jvm/CORE.setting:684␤ in block at /tmp/1iFwqQEbXC:1␤ in any eval at gen/jvm/stage2/NQPHLL.nqp:1086␤ in any evalfiles at gen/jvm/stage2/N…»
..rakudo-parrot 455fbc: OUTPUT«Not enough positional parameters passed; got 0 but expected 1␤ in block at eval_0:1␤ in any at eval_0:1␤ in sub eval at gen/parrot/CORE.setting:692␤ in sub eval at gen/parrot/CORE.setting:683␤ in block at /tmp/SenabOqL7H:1␤ in any a…»
timotimo r: eval '-> { 1 < $^a < 3 }' 08:55
camelia rakudo-parrot 455fbc, rakudo-jvm 455fbc: OUTPUT«===SORRY!=== Error while compiling eval_0␤Placeholder variable '$^a' cannot override existing signature␤at eval_0:1␤------> -> { 1 < $^a < 3 }⏏<EOL>␤ expecting any of:␤ postfix…»
timotimo ... how do i ...
r: eval 'sub { 1 < $^a < 3 }'
camelia ( no output )
timotimo ah!
moritz or simply without the ->
timotimo no, see above 08:57
turns out my chained_whatevercode branch still has issues 08:58
08:58 kivutar joined
timotimo brought to light by a "test suite" that automatically turns code like 1 < $^a < $^b < 10 into 1 < * < * <10 and such 08:58
hm, now it would be super helpful if i could a) tell eval directly to dump the ast or b) have --target=ast that doesn't stop after the first ast stage 09:01
i guess i can still nqp::getcomp, though
09:10 rindolf left 09:12 fridim_ left 09:22 [Sno] left 09:31 atroxaper joined 09:45 bakedb left 09:46 bakedb joined 09:47 atroxaper left 09:48 spider-mario joined, spider-mario left 09:49 kivutar left 09:50 spider-mario joined
timotimo suspects something messing with a qast node that's shared, and so distributes some nqp::clone calls throughout the code ... >_< 09:57
diakopter QASHARED 09:58
timotimo QASH? :) 09:59
is pm working on a more efficient hash implementation perchance?
diakopter he's otherwise occupied for the time being still 10:00
afk&
lizmat good *, #perl6 10:03
dalek : 3311b3d | (Elizabeth Mattijsen)++ | misc/perl6advent-2013/schedule:
20: Defaulting to Nililism & other variable traits
 10:08
lizmat need to go out just about now: if I get back tonight and there is still no advent blog post, I'll move mine up to today 10:09
10:12 bloonix joined
jnthn Morning-ish, #perl6 10:20
preflex jnthn: you have 1 new message. '/msg preflex messages' to read it.
jnthn japhb: Looks so...will update it...
lizmat: It's my turn today. I'll be working on it shortly. :) 10:21
jnthn maybe should sign up for the day *after* the one he'll have time to write posts, just to prevent panic. ;) 10:22
timotimo if it's acceptable to deliver the blog post for a day *on* the day of the post, i won't have to ask for mine to be moved a day back! :) 10:23
bleeeeh, where is that "Unknown QAST node type NQPMu" coming from >:( 10:24
something where $thing.HOW.name($thing) returns NQPMu, but ... what?! 10:25
timotimo hopes for a better error message on jvm
FROGGS hi jnthn 10:26
jnthn timotimo: Well, I suspect the answer may be to make NQP's error detection more robust...dunno how hard that'll be. :)
o/ FROGGS
10:28 Piers_ left
timotimo No such method 'jast' for invocant of type 'NQPMu' d'ohpe 10:28
jnthn Does looking at the --target=ast output not show something up 10:30
?
10:33 gfldex left
timotimo no :( 10:36
let me paste you a bit
FROGGS someone wanna port a core module? :o) github.com/rakudo-p5/v5/blob/maste...modules.md 10:37
timotimo gist.github.com/timo/fac77c791de77a5dab79
port? are these written in C? 10:38
FROGGS timotimo: don't think so 10:39
at least, the majority should be written in perl 10:40
timotimo can't you just compile them with v5 then?
FROGGS timotimo: btw, is your gist complete?
timotimo that's all the output there is
10:40 rindolf joined
FROGGS timotimo: I just need to provide their api, how I/someone does that is up to that someone 10:40
timotimo ah, ok
jnthn timotimo: The chain at 65 looks suspect 10:42
It only has one child
But a chain needs two
timotimo oh!
how did i overlook that? >_>
thanks
jnthn Well, maybe let's improve the error too :)
FROGGS yeah, that probably is it
timotimo sure, the op "chain" needs to complain if the number of arguments is <= 1 10:43
right?
FROGGS I think the weird "cannot call .jast" might be fixed
because sometimes it takes ages to figure out what is wrong 10:44
jnthn: btw, have you seen the msg in #moarvm? could this one be related to our exploding frame? 10:45
10:46 SamuraiJack_ left
jnthn FROGGS: Maybe. 10:47
10:48 denisboyun joined 10:49 ivanshma` left 10:50 ivanshma` joined
felher Somebody here from rumania? 10:54
jnthn
.oO( The country where they drink a lot of rum? )
felher :D 10:56
yeah, romania I meant :) 10:57
FROGGS felher: not that far from romania, but no :o)
lunch & 10:58
arnsholt timotimo: We could very well resume work on the ZMQ stuff 11:01
You starting to look into your ipython stuff again?
timotimo not sure 11:05
with JVM NativeCall around the corner, i could try
the problematic thing was the heartbeat that was supposed to run in parallel
for now, the blasted chained_whatevercall work is keeping me sufficiently busy :P
arnsholt Yeah, I remember the heartbeat being a problem 11:07
timotimo (which is funny, because i recently saw a cardiologist about my own heartbeat)
(the thing i went there about was completely harmless, though) 11:08
11:14 djanatyn left 11:16 Tene left 11:17 gfldex joined, Tene joined 11:18 djanatyn joined 11:29 gfldex left 11:33 rurban1 joined 11:35 xenoterracide joined
timotimo oh yeah! my code is proper now! :) 11:35
dalek kudo/chained_whatevercode: a3ccefb | (Timo Paulssen)++ | src/Perl6/Actions.nqp:
fix remaining problems with whatever_curry and chain ops
 11:36
kudo/nom: dfeff62 | (Timo Paulssen)++ | src/Perl6/Actions.nqp:
fix whatever currying for chain ops

When a WhateverCode gets created from a chain op, it'll be annotated with its original arguments and when a WhateverCode with a chain op ends up calling a WhateverCode that also had a chain in it, it will "flatten out" the arguments of that WhateverCode, creating a proper chain.		 11:37
kudo/nom: a3ccefb | (Timo Paulssen)++ | src/Perl6/Actions.nqp:
fix remaining problems with whatever_curry and chain ops
kudo/nom: 40c8114 | (Timo Paulssen)++ | src/Perl6/Actions.nqp:
Merge branch 'chained_whatevercode' into nom

This fixes WhateverCode generation when chain ops are involved.
timotimo does a little dance to celebrate 11:41
jnthn timotimo: Can you give me a quick summary of what you just fixed? :) 11:44
timotimo what was wrong or how i fixed it?
jnthn What was wrong
timotimo sure
Whatever currying would find, that the "* < 5" part of "1 < * < 5" was a WhateverCode and would emit a call to the code object 11:45
however, for chain to properly function, it had to be inlined
so i had to turn -> $outer { 1 < (-> $inner { $inner < 5 })($outer) } 11:46
into -> $outer { 1 < $outer < 5 }
(actually, i think the comparison operators are left-assoc) 11:47
(but you get the idea)
i think this warrants an entry into the changelog? :)
wowza. this dumb benchmark i made for the benefits of inlining WhateverCode into ifs (like when 10 < * < 20) would give a speed improvement of 2x 11:50
11:50 rurban1 left
timotimo i should find out how inline_info works 11:50
jnthn timotimo: Yes, worth a mention 11:53
timotimo: And glad this is fixed. :)
timotimo wow, add_inlining_info_if_possible is pretty useful for this! :)
it will even help in all other cases, not just when a chain is involved 11:54
jnthn Well, something needs to make use of the info too
So we're careful where we put it. :)
timotimo the optimizer already tries to
at least i hope so.
jnthn Oh :)
Then it's worth it :)
Well, don't assume...check :)
timotimo i'll see if it makes a difference and if not i'll look into why the optimizer doesn't
i'll be afk for a bit first. 11:55
(and let the test run in the background)
jnthn For example, inlining methods isn't something we can do at compile time, and so best left to runtime.
So no point sticking inline info onto those.
12:01 tgt joined 12:04 gfldex joined
colomon timotimo++ 12:04
12:07 denisboyun left 12:12 gfldex left
timotimo Null PMC access in find_method('add_inlining_info_if_possible') ... ?!?! 12:14
oooh!
whatever_curry is a sub
12:18 gfldex joined
timotimo i wonder how close it'll get to the $_-based solution after inlining 12:19
12:22 denis_boyun joined
timotimo apparently the inlining info wasn't used 12:23
maybe it's because i'm operating on the result of block_closure(reference_to_code_object($code, $block)), whereas i should be working on... maybe just the inner part? 12:25
12:28 atroxaper joined 12:30 ivanshma` left 12:38 denis_boyun__ joined 12:39 denis_boyun left
timotimo maybe it's something for jnthn to look at. he should know how that's supposed to work :P 12:40
timotimo adds lots of debug prints :D 12:46
12:47 denis_boyun__ left, denis_boyun joined 12:53 rurban1 joined, denis_boyun left, PacoAir joined 12:55 ivanshmakov joined 12:57 hummeleB1 joined 12:58 denis_boyun joined 13:06 denis_boyun_ joined, denis_boyun left
colomon r: my $r = 1.969; say $r.fmt("%.0f"); say $r.fmt("%.1f"); say $r.fmt("%.2f"); say $r.fmt("%.3f"); 13:06
camelia rakudo-parrot 40c811, rakudo-jvm 40c811: OUTPUT«1.␤1.0␤1.97␤1.969␤»
timotimo oh, when did that happen! 13:07
colomon timotimo: that's still broken
13:08 rindolf left
colomon should be 2 2.0 1.97 1.969 13:08
timotimo oh, whoops
colomon that seems to happen on both parrot and jvm
jnthn Well, they share a sprintf implementation, afaik :)
colomon there is also a second bug involving $_ that japhb found on jvm only.
13:09 rurban1 left
colomon r: say "$_" for 0..3 13:09
camelia rakudo-parrot 40c811, rakudo-jvm 40c811: OUTPUT«0␤1␤2␤3␤»
colomon r: my $r=1.969; say $r.fmt("%.{$_}f") for 0..3
camelia rakudo-jvm 40c811: OUTPUT«1.␤1.␤1.␤1.␤»
..rakudo-parrot 40c811: OUTPUT«1.␤1.0␤1.97␤1.969␤»
colomon r: say "{$_}" for 0..3 13:10
camelia rakudo-parrot 40c811: OUTPUT«0␤1␤2␤3␤»
..rakudo-jvm 40c811: OUTPUT«␤␤␤␤»
colomon binog
bingo
oh, Time::Duration started failing tests overnight, dagurval 13:11
timotimo ah, ok, is inlinable is only defined for Routine 13:13
r: say WhateverCode ~~ Routine
camelia rakudo-parrot 40c811, rakudo-jvm 40c811: OUTPUT«False␤»
13:14 denis_boyun__ joined, denis_boyun_ left 13:18 denis_boyun_ joined 13:19 denis_boyun__ left 13:23 FROGGS left 13:27 Mouq joined 13:37 ivanshmakov left
jnthn Advent post is done. :) 13:39
13:44 ivanshmakov joined 13:46 FROGGS joined
moritz \o/ 13:47
13:47 denis_boyun__ joined, denis_boyun_ left, dmol joined 13:53 denis_boyun joined, Mouq left 13:54 denis_boyun__ left 14:07 jnap joined
timotimo r: say WhateverCode.^mro 14:07
camelia rakudo-parrot 40c811, rakudo-jvm 40c811: OUTPUT«===SORRY!===␤No such method 'arity' for invocant of type 'QAST::WVal'␤»
14:07 SamuraiJack_ joined
timotimo er ... 14:08
jnthn star: say WhateverCode.^mro 14:09
camelia star 2013-09: OUTPUT«WhateverCode.new()␤»
14:10 jnap left
jnthn That ain't right either... 14:10
14:10 rurban1 joined, jnap joined
timotimo yeah, we are whatever_currying that it seems 14:13
m)
14:13 kivutar joined 14:14 ivanshmakov left 14:20 Mouq joined 14:21 rurban1 left
colomon whatever_currying is when you can't decide between red or green, right? 14:21
timotimo i think that's actually junctions 14:31
i can't seem to get the inlining info to work properly so far
arnsholt Whatever code is what you get when a whatever is part of a larger expression 14:33
jnthn timotimo: What are you trying to inline?
moritz tries to inline the inliner 14:34
timotimo i'm trying to add inlining info to all WhateverCode objects 14:36
14:36 ivanshmakov joined, ivanshmakov left
timotimo (by first providing an anonymous role that provides an inline_info sub to the WhateverCode and then calling add_inlining_info_if_possible from whatever_curry 14:38
14:38 ivanshmakov joined, rurban1 joined
jnthn Yes, but why do you want to do that? 14:38
When will we ever have opportunity to use that info?
timotimo given $foo { when 10 < * < 20 { ... } } 14:39
(hopefully)
jnthn OK, but you don't need inline info for that.
timotimo right; but that mechanism was already in place :) 14:40
well, mostly
but i just found a flaw in my benchmark, i think
jnthn inline_info mostly exists to enable cross-compilation unit inlining.
timotimo oh?
jnthn It's overkill for such local things.
timotimo i did not know that
jnthn Yeah, it's there so we can have operators declared in the setting inlined, for example. 14:41
timotimo so, is there a simpler method of doing what i want to achieve, except for using the info i already use to inline WhateverCode inside WhateverCode?
jnthn Well, all inline_info does is persist AST
And you already have the AST to hand... 14:42
timotimo mhm
say, could it be that we don't have to have a block_closure for our whatever_curry?
jnthn Typically you need it to be a closure. 14:43
timotimo i'm guessing block_closure does the Op( clone, BVal )
okay
jnthn @x.map: $a + *
It matters which $a
timotimo right
i don't know exactly how to make that work properly with inlining, then 14:44
jnthn Needs analysis 14:45
timotimo writes a changelog entry for the feature
jnthn It's probably not easy to do at the moment
timotimo that's fine then 14:46
huh, not that many interesting commits to rakudo/nom this month so far 14:49
14:49 dmol left, dmol joined
timotimo should the indy-work be mentioned in the changelog? 14:49
i think so; but left out of the rakudo start short changelog 14:50
jnthn Yeah, it got us a win 14:51
14:51 jeff_s2 joined
dalek kudo/nom: b11daf9 | (Timo Paulssen)++ | docs/ChangeLog:
A few changelog entries for this month
 14:52
14:54 jeff_s1 left
timotimo jnthn: you get a $p = promise, then a $v = $p.vow and call break/keep on the original $p then? 15:06
or is that a typo?
jnthn typo 15:07
Fixed, thanks 15:08
timotimo yw
is there a reason you're not showing AsyncFile.slurp here? 15:10
well, i guess it makes 0 difference here
15:12 Rotwang joined, Rotwang left
jnthn timotimo: Well, article wasn't menat to be an exhaustive guide :) 15:12
15:12 Rotwang joined
timotimo right :) 15:12
15:15 nebuchadnezzar left, jeckyhl joined 15:17 nebuchadnezzar joined 15:21 ggoebel115 joined, rindolf joined
jeckyhl hi, had anybody experienced problems building rakudo / jvm on Windows ? I've got an error when running tools\build\create-jvm-runner.pl : on line 32, $install_to happends to be "\C:\Dev\perl6\rakudo\install\bin\perl6-j.bat" (with a backslash on the beginning) which make the next instruction fail 15:22
15:23 jnap left 15:24 jnap joined, ggoebel114 left, xinming_ joined 15:26 xinming left 15:28 jnap left 15:30 denis_boyun_ joined 15:31 denis_boyun left 15:34 nnunley joined 15:35 jeff_s1 joined 15:37 jeff_s2 left 15:40 denis_boyun joined 15:41 denis_boyun_ left
timotimo FROGGS: do you have an overview what exactly is missing for loop labels to work in nqp and then rakudo? 15:41
jnthn I think me reviewing the branch so far was one of the things... :)
FROGGS timotimo: there is nothing missing for using them in nqp :o) 15:46
timotimo even for jvm and moar?
FROGGS timotimo: ahh, no
that was parrot only
timotimo oke
i'm just a bit bummed out by the small amount of things i could write in the changelog for this month so far 15:47
FROGGS don't wanted to continue when the approach was maybe wrong or so (even when I don't believe I took the wrong one)
timotimo: btw, there is a sized-arrays branch that would need reviewing and merging
but I was unable to squish the 20+ commits into one, so it would be reviewable at all 15:48
timotimo oh? 15:49
i didn't know that's an operation that can fail :)
FROGGS was the first and only time I did that 15:52
(or tried to)
timotimo in the worst case, you can git diff the branch against its fork-off-point, write the diff to a file, check out the fork-off-point and apply the resulting patch with "patch -p1" 15:54
16:02 jeckyhl left
FROGGS timotimo: and how do I preserve the author? 16:03
jnthn Note that if we're doing this just to make it easier to review, it's easy to read the diff to review it and then apply the original commits... :) 16:04
16:04 Rotwang left
jnthn So long as the branch is up to date with latest nom 16:04
FROGGS yeah, that'd work too 16:07
jnthn: I can do that in a few hours if you will have time to review this night or tomorrow 16:08
jnthn FROGGS: Yeah, I can find time to do that. 16:13
FROGGS k
timotimo FROGGS: git commit will accept --author and --commiter iirc 16:17
16:18 Rotwang joined 16:19 ico joined 16:20 ico left 16:24 rurban1 left 16:29 zakharyas joined 16:43 kivutar left 16:54 rurban1 joined 16:58 fridim_ joined 17:00 pmurias joined 17:08 grondilu joined 17:10 rurban2 joined 17:11 rurban1 left 17:14 rurban2 left, denis_boyun_ joined, denis_boyun left 17:22 Timbus left
TimToady is supposed to be singing tenor in our Christmas musical tomorrow, and will probably be singing basso profundo instead, sigh... 17:23
17:23 Timbus joined 17:25 atroxaper left 17:29 ivanshmakov left, denisboyun joined 17:30 denis_boyun_ left 17:40 kaare__ is now known as kaare_
TimToady jnthn: asyncrhonously 17:42
jnthn TimToady: ? :)
TimToady s/rh/hr/ 17:43
jnthn oh... 17:44
Stared at it trying to spot the spello for a whole minute :/
fixed
17:45 gdey joined 17:47 isBEKaml joined
moritz succeeded in gdb 17:48
timotimo jnthn: i'm silly. i figured out why it didn't inline the WhateverCodes for "when"; because when makes a .ACCEPTS($_) rather than a normal call
moritz erm, wc
timotimo thus, i can teach the optimizer to look at the invocant of ACCEPTS($_) if it's a WhateverCode that's annotated with the information from the whatever_curry thingie and then inline it 17:49
17:50 dronmbb joined, dronmbb left
timotimo but i'm not entirely sure what i have to analyse to make sure i don't break closures that are necessary for correct functioning 17:50
maybe if the only vars that are inside are the arguments (the only argument) to the whatevercode object
17:56 xenoterracide left
timotimo jnthn: ^^ does that sound like the necessary analysis? 17:57
jnthn timotimo: Kinda, but we really need much more general analysis of these things 17:58
timotimo OK, it'll wait until much later then
jnthn And putting in optimizations like this that special-case it will only make it harder later.
The optimizer is already harder to maintain than I'd like.
timotimo oh no! what did i do! ;_; 17:59
jnthn Just added more things into a less-than-sustainable design. :)
It needs to grow and do more, but I think it'll need a bit of restructuring first to keep things sane. 18:00
In general, it needs a bunch of changes to start being able to track variables and their usage.
18:01 gdey left, isBEKaml left
timotimo right 18:02
i once had "a" design based on hashes and stuff
18:02 nnunley_ joined 18:03 [particle]1 joined
TimToady knows from personal experience that it's really easy to get stuck at the wrong abstraction level 18:04
jnthn I'm also a little wary of special-casing something that I'm not sure comes up a load, or at least, I wonder if this special case actually falls out of two things
That is, ACCEPTs on a known code object with a single arg in a boolean context can be first reduced to just a call. 18:05
And then an inline falls out more naturally.
18:05 itz_ joined, _ilbot left, PacoAir left, nnunley left, atrodo left, itz left, estrabd_ joined, estrabd left, rurban left, Fatalnix left, Pleiades` left, [particle] left, japhb_ left, Juerd left, dylanwh left, hoelzro left, Juerd joined 18:06 dylanwh joined, rurban joined, Fatalnix joined, hoelzro joined, atrodo joined, Pleiades` joined, japhb_ joined, Pleiades` left 18:07 Pleiades` joined, _ilbot joined
TimToady welcomes everyone back, including _ilbot 18:07
jnthn: you have $dest.last(); but I believe the current speci is $dest.close 18:11
18:11 rurban1 joined 18:13 SamuraiJack_ left 18:14 gdey joined 18:16 rurban1 left
TimToady and we don't seem to have a specced $dest.fail, which should probably be $dest.quit anyway, for consistency with supplies 18:16
I see that the current implementation has .close and .fail 18:22
so the .done is definitely fossilic
er, .last
timotimo i should finally get my simplify_before branch ready for merging 18:23
and write my advent calendar post
also, i think i need about 6 naps right now
moritz sleep faster! :-)
TimToady do them concurrently! 18:24
timotimo i should employ a few people to sleep fo rme
moritz SaaS, Sleep as a Service 18:25
Cloud Sleeping!
18:32 darutoko left 18:33 ajr_ joined 18:34 ivanshmakov joined
moritz kinda enjoyed www.nosql-vs-sql.com/ 18:35
timotimo yeah
18:38 spider-mario left 18:39 jeff_s1 left 18:40 jeff_s1 joined 18:41 Rotwang left, snoopy_ joined, snoopy_ left 18:42 dwarring joined 18:43 Rotwang joined
dalek kudo/nom: a05e24d | moritz++ | docs/ChangeLog:
two more changelog entries
 18:48
18:59 SamuraiJack_ joined
dwarring yawns 19:01
preflex dwarring: you have 1 new message. '/msg preflex messages' to read it.
dwarring aha thanks raiph 19:02
colomon dwarring: PDF::Grammar failing in my smoke tests.
19:03 lizmat left, lizmat joined
dwarring colomon: will check it out 19:06
colomon dwarring++
masak evenin', #perl6 19:08
19:08 culture joined, culture left
moritz it's a masak! 19:09
masak \o/ 19:12
19:13 telex left, grondilu left 19:15 rurban1 joined 19:16 telex joined 19:17 sftp_ joined 19:18 PZt left, sftp left, aindilis left
raydiak_ this typo was too amusing not to share: "tmux attack" 19:19
19:19 rurban1 left, raydiak_ is now known as raydiak
timotimo :D 19:19
colomon masak! \o/ 19:20
moritz raydiak: the attach of the clone warriors :-) 19:23
19:23 kivutar joined 19:27 djanatyn left 19:28 Tene left
dalek Heuristic branch merge: pushed 22 commits to rakudo/sized-arrays by FROGGS 19:29
19:30 Maddingu1 joined, gdey left, betterwo1ld joined
lue hello world o/ 19:31
FROGGS jnthn: here is the diff gist.github.com/FROGGS/50d58158001696e89327
(for sized-arrays)
19:32 betterworld left, Maddingue left
FROGGS jnthn: btw, this should be done in rakudo too: github.com/rakudo-p5/v5/commit/3105451 19:35
raydiak moritz: I haven't seen that one...can you mail it to me as an attackment? :) 19:39
preflex raydiak: you have 1 new message. '/msg preflex messages' to read it.
raydiak morning lue \o
mathw off-topic-ish: good real-world-ish example of a simple class for introducing someone to OOP? 19:47
timotimo what a huge ass-diff
TimToady FROGGS: looks like a chain thing snuck in there 19:52
19:52 Tene joined, Tene left, Tene joined
FROGGS hmmm 19:53
19:53 djanatyn joined
FROGGS I just merged nom into sized-arrays branch, and made a diff 19:54
but yeah, that does not belong there
ohh well
moritz mathw: maybe the Task example from the OO chapter of the p6 book? 19:55
FROGGS maybe I really fake a diff like `git am` produces?
so I could just strip that stuff
timotimo huh, why are the spectests clean? 19:56
weren't they nonclean on my simplify_before branch recently?
well, that's cool i guess o_O
moritz FROGGS: did you try git diff nom...origin/sized-arrays (note the three dots)
TimToady doesn't see offhand where multi-dimensional shapes are handled
FROGGS moritz: no
moritz: that seems to be identical 19:57
moritz ok
might be simpler than doing the merge locally 19:58
timotimo: do we have any spectests for the chained *-curry thing? 19:59
jnthn TimToady: gist.github.com/FROGGS/50d58158001...s-diff-L26 explicitly says it doesn't yet
FROGGS: Do you understand the map thingy that's going on in gist.github.com/FROGGS/50d58158001...s-diff-L17 ? 20:00
FROGGS jnthn: ummm, no 20:01
jnthn oh, maybe it's to handle my @seasons[4, :map( * % 4 )];
FROGGS ahh, my nom missed two commits.... the chainy stuff is gone now 20:02
jnthn It looks like a fragile way to do that...
FROGGS wth is that even O.o?
timotimo moritz: yes, two todos that now pass
moritz: but this spectest run was for simplify_before, which does regex optimizations 20:03
jnthn FROGGS: See S09 for what it is, but I can see that code being at risk of exploding in various ways...
FROGGS k
jnthn It doesn't look for there being a Pair and then look if the first key is map, it just assumes... 20:05
I can see that going horribly wrong somehow
I don't understand what gist.github.com/FROGGS/50d58158001...s-diff-L61 is trying to achieve either 20:06
gist.github.com/FROGGS/50d58158001...-diff-L288 probably makes some wrong assumptions about flattening 20:09
timotimo moritz: questhub.io/realm/perl/quest/52a5f3...de150000f1 - does this seem like something that could go into the spectests? 20:10
FROGGS jnthn: the commit does not explain either github.com/rakudo/rakudo/commit/a9...14aa5700d3
moritz timotimo: yes
with some say() replace by ok/is calls, of course :-) 20:11
jnthn Also, this patch will slow things down, since it makes every array pay for the :map feature when relatively few will do that.
timotimo can someone do spectests on jvm for me? :( 20:15
once for simplify_before/regex_optimizer and once for master/nom? 20:16
moritz timotimo: I think I can, but only with a single process 20:17
timotimo so, only if you have time then :)
deprecations / isDEPRECATED still has known failures?
why don't we fudge those?
(i can do it if it's okay with all of you) 20:18
moritz ask lizmat
timotimo my spectest run of simplify_before/regex_optimizer got to S05 now 20:21
lue is a bit alarmed at the irregularity between L<Syn/Section/"Text fragment"> and L<Syn/Section/Text Fragment> and L<Syn/"Section"/Text Fragment> and ... 20:22
dalek ast: e34f95c | moritz++ | S02-types/whatever.t:
unfudge whatever-chain tests, timotimo++
moritz has just fnished compiling j-all
jnthn FROGGS: Anyway, those are my comments. I'm a bit uncomfortable about merging something that will slow down every array access. 20:23
moritz daugher sleeps, now I can take my laptop to the living room again, where it has a power supply :-)
*daughter
jnthn FROGGS: And doubly so when it contains code I know is fragile.
moritz though it says it still has more than an hour to live
FROGGS jnthn: of course... I'll msg Arcterus a link to to this here
dwarring promises & channels advent post - jnthn++ 20:24
20:25 Juerd left
jnthn FROGGS: The mixin approach taken by typed ones may be better. 20:25
lue briefly ponders replacing the L<> stuff in roast with M<> during his making of the Pod6 toolchain, so as to not confuse the point of L<> 20:26
20:26 Juerd joined 20:27 SamuraiJack_ left
moritz can't decide if that's scope creep or not 20:28
20:29 zakharyas left
dalek ar: 8d120c2 | moritz++ | modules/ (2 files):
update two modules
 20:31
timotimo already preparing for the star release? :)
lue moritz: I definitely wouldn't do any theoretical L<> -> M<> change until the very end. It was just me wondering if calling it a "L"ink is exactly correct. (fwiw, it feels like it's better describe as a *reverse* link, but there's no RL<>) 20:32
moritz lue: it's actually a link, the tools just use it as a reverse link 20:36
20:36 wooden left 20:37 djanatyn left, wooden joined, Tene left
lue Fair enough. Still feels a bit of a stretch to call it a link, maybe because I'd feel better with L<syn:S02/Sect/"fragm"> (protocols are important!) :D 20:37
20:38 rindolf left
dalek ar: 7d731a0 | moritz++ | README:
update README. Closes #15
 20:41
lue My worry over the use of "" in the links earlier was because I forgot that /Section/ doesn't care about hierarchies. (i.e. to reach a =head4 it doesn't do /head1/head2/head3/head4 but rather /head4/) 20:42
timotimo lue: use ˥<...> instead 20:45
lue
.oO(λ<>)
timotimo t/spec/S06-other/main-eval.rakudo.jvm ........................ No subtests run 20:46
is the first one that fails like that when i do TEST_JOBS=2
lue & 20:47
dalek ar: 42daf9f | moritz++ | tools/star/release-guide.pod:
Update release guide to avoid out-of-date submodules
 20:48
20:56 Tene joined, Tene left, Tene joined 20:57 djanatyn joined 20:58 ajr_ left 20:59 ajr joined, ajr is now known as Guest49853, Guest49853 is now known as ajr_
masak so... 21:08
Rakudo release on Thursday, and no-one signed up as release manager, huh? 21:09
here, I'll make you a deal.
someone either grap Thursday and I'll grab January, or someone grab some other month and I'll grab Thursday.
moritz masak: I'll do Thursday 21:10
masak moritz++
dalek kudo/nom: 1069590 | moritz++ | docs/release_guide.pod:
split December and January releases between masak++ and me
 21:11
moritz it would be awesome to have a volunteer for the star release too
colomon moritz++ masak++
masak grab*
man, why feather so slow again? :/
moritz low load average 21:12
low CPU usage 21:13
low ping times
moritz has no idea 21:15
maybe cross-talk from another VM? IO?
timotimo yeah, VMs are problematic that way 21:16
isn't there some standardised protocol to tell VMs how much load their host has? 21:17
moritz though it's not usually that bad :-)
timotimo moritz: were you able to start the spectest run yet?
moritz timotimo: the nom one is t/spec/integration/error-reporting.rakudo.jvm right now
timotimo: I'll start the second one before going to bed 21:18
timotimo thanks!
i'll try a one-job-in-parallel run now, too
though today is feuerzangenbowle in my hackspace, so ... :)
%)
at least it does get past the S05 tests just fine 21:20
masak
timotimo
lizmat timotimo: those deprecations tests failing on jvm, is because backtraces fail, afaik
and that seems pretty important to me
timotimo can we todo them so that they don't show up as "not ok" on comparison runs for branches and features and stuff?
masak moritz: I remember having this issue a few weeks back, and no-one else having it then either.
moritz: maybe this laptop is simply giving up the ghost.
moritz timotimo: gist.github.com/moritz/2375e30d8c3bf35bcf04 master/nom result 21:21
masak: no, it's slow for me too
timotimo i thought fudging is for "we know this is bad and we should fix it, but don't worry, it's not *your* particular fault"
21:21 rurban1 joined
lizmat timotimo: ok, mark them as todo (or skip, I think, because they die) 21:21
timotimo perl.rakudo.jvm fails for me, too
er, it doesn't have a failure when i call it on its own. great. 21:23
masak moritz: oh! 21:26
21:29 rurban1 left 21:32 denisboyun left 21:35 jeff_s1 left 21:36 jeff_s1 joined
dalek ast: 1f062e9 | (Timo Paulssen)++ | S02-types/ (2 files):
fudge jvm problems with deprecation tracebacks
 21:37
kudo-star-daily: 0321033 | coke++ | log/ (5 files):
today (automated commit)
 21:38
kudo-star-daily: 16bd544 | coke++ | log/ (5 files):
today (automated commit)
kudo-star-daily: b1ab6d4 | coke++ | log/ (5 files):
today (automated commit)
kudo-star-daily: cb0ce3f | coke++ | log/ (5 files):
today (automated commit)
rl6-roast-data: 8ce3a8a | coke++ | p (2 files):
today (automated commit)
rl6-roast-data: ba57761 | coke++ | p (2 files):
today (automated commit)
rl6-roast-data: 348552e | coke++ | p (2 files):
today (automated commit)
timotimo moritz: i don't get the failure in S32-test/roll.t
and "no subtests run" again 21:39
21:46 kaare_ left 21:53 ajr_ left
dwarring colomon: have fixed PDF::Grammar 21:54
colomon \o/
dwarring++
21:55 PZt joined
dalek kudo/moar-support: 20cf1d2 | jnthn++ | src/vm/moar/ops/container.c:
Fix RakudoScalar as having non-invoking fetch.
 22:09
kudo/moar-support: 2a7cd2f | jnthn++ | src/core/List.pm:
Optimization that avoids a Rakudo Moar bug.

Doing $some_int == Inf ends up coercing $some_int and re-dispatching, which we can avoid much more quickly with a type check for the very common case of passing an integer to gimme. This helps all backends, but also avoids a Moar bug; hopefully sanity tests will show it up in a clearer way.
22:28 Pleiades` left 22:29 colomon left, colomon joined 22:33 Pleiades` joined 22:34 rurban1 joined
dalek kudo/moar-support: df50def | jnthn++ | src/core/Version.pm:
Disable version object construction on Moar.

Gets in the way of running any sanity tests. Can get it back in place later.		 22:36
22:36 [Sno] joined 22:38 xenoterracide joined
pmurias PDF::Grammar is a grammar for pdfs? 22:54
22:55 rurban1 left
woolfy jnthn++ for nice perl 6 advent blog. 22:55
Looking forward to the Day 15 Perl 6 Advent Blog (which I hope will appear any moment now) 22:56
masak oh, seems it's my turn tomorrow...
woolfy I thought it would be every day around 23:00... which was 56 minutes ago... 22:57
22:59 webchat_masak joined
webchat_masak grumbles over feather's non-responsiveness 23:00
woolfy: I would have had a post ready by now, but I missed that I'm on for tomorrow, and not Monday...
lue hello webchat_masak o/
webchat_masak I'll draft a post now, go to bed, and then publish it first thing in the morning.
woolfy masak... :-) (was I subtle enough to get you to work?) 23:01
webchat_masak yeah, writing it now.
woolfy masak++
webchat_masak though I actually independently checked the schedule, curious who was on for tomorrow...
woolfy resulting in? 23:02
webchat_masak mild surprise, and a decision not to go to bed yet.
woolfy hehe
lizmat tried me to write a blog for the advent, and I really have not enough Perl 6 knowledge to do so. It would be a social thing, like something about Tuits, or Fosdem, or how I enjoyed seeing Perl 6 people perform at a hackathon, a workshop or a YAPC. 23:03
And I am so not going to write something egotistical like that! 23:04
webchat_masak woolfy: maybe you should write a few Perl 6 programs ;) 23:05
woolfy masak: Maybe. Probably not a good idea. Would be bad for the reputation of Perl 6. :-)
23:08 laidback_01 joined
raydiak apparently I dodged a bullet...I almost wrote my advent post about something social :) 23:10
webchat_masak woolfy: you keep underselling yourself :)
woolfy: I for one would love to see some woolfy Perl 6 code. 23:11
colomon woolfy! woolfy! woolfy!
dwarring have my stocking filler ready
woolfy webchat_masak: indeed. But I know myself as a programmer. I did nice scripts in the 90's, but mostly worked as a manager, marketeer, project manager, account manager, human resource manager... I hired people that really were good programmers and tought them Perl. Perl 5. 23:12
webchat_masak woolfy: sounds good enough to me. 23:13
woolfy: I would also have been a recovering Perl 5 programmer, if only I had discovered Perl 6 later and Perl 5 sooner :P
lue only picked P6 because trying to use a Qt module without knowing P5's classes discourages you :) 23:15
woolfy Let me do the tuits, the leaflets and the other simple marketing stuff and I leave the programming to you, lizmat, jnthn, TimToady, pmichaud, colomon, timotimo, moritz, lue, FROGGS, sorear, tadzik, and everybody else that does real programming. 23:16
oO(was not meant to exclude people, you all are heroes!) 23:17
dwarring didn't pick Perl6 - just started doing advent exercises last year and hasn't stopped yet
woolfy dwarring++
webchat_masak dwarring++
I didn't even know we had advent exercises :)
23:18 beastd joined
dwarring neither did I - just started playing with the code 23:19
then tired something on my own own on day 25
I followed previous calendars 23:21
diference last year was I kept going afterwards 23:22
woolfy again, dwarring++ 23:23
By the way, this advent of yesterday perl6advent.wordpress.com/2013/12/1...kudo-star/ is very useful for marketing purposes.
It's one of the most asked questions / snide remarks given... when is it christmas / is Perl 6 ready yet. This blog is helpful in answering those. 23:24
Coke++
lue
.oO(JSON::Tiny's roundtrip test has a nasty habit of slowing my computer to a near-halt.)
dwarring Perl 6 is ready at Christmas - its just that the Christmases are unevenly distributed 23:27
lue does anyone else get this, or should I upgrade my r-p? "build stage failed for panda: Failed building lib/Panda/Project.pm" 23:33
tadzik: any ideas what could be causing this? ^^^ 23:35
FROGGS lue: did you pass the --prefix to rakudo like ./something? 23:36
lue no...? It's been a while since I played with installing r-p, and it always goes straight to /usr/local 23:37
23:38 MikeFair__ joined
lue even cleared out his ~/.perl6 to make sure panda wasn't borking on old directories there. 23:38
23:39 MikeFair left
lue (everything else builds fine, it's just panda. Running rebootstrap, IIRC, makes it so File::Find won't build. Somehow. 23:40
)
FROGGS: are you suggesting running sudo bootstrap or install the r-p toolchain into a non-sudo-required place would work? Because I'm not sure what makes ~/.perl6 broken. 23:42
FROGGS I dont know what to suggest :/
lue I just ran the build command (after reading through panda's code), and it seemed to run fine. I've no clue why it errors then :/ 23:51
23:51 rurban1 joined 23:56 xinming_ left 23:59 xinming 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!