»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
Mouq Hm. Exactly fc2a6e632efa? It would be useful to know when it did work 00:02
cxreg "The reduction metaoperator now recognizes operators with list infix precedence" # IRTA "redaction metaoperator"
00:03 cognominal left, cognominal joined, cognominal left, cognominal joined 00:19 FOAD joined 00:20 denis_boyun left
lue r: say uniprop("☃", "na"); # should be eqv. to uniname("☃") 00:24
camelia rakudo-parrot 02329b: OUTPUT«uniprop NYI on parrot backend␤ in sub uniprop at gen/parrot/CORE.setting:6757␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar 02329b: OUTPUT«1.1␤»
..rakudo-jvm 02329b: OUTPUT«uniprop NYI on jvm backend␤ in sub uniprop at gen/jvm/CORE.setting:6753␤ in block at /tmp/tmpfile:1␤␤»
Mouq doesn't completely understand these uni* things except uniname 00:25
lue Mouq: S15 update explaining the ones TimToady made up incoming :) 00:27
BenGoldberg uniprop, presumably, basically asks, for unicode string "Foo", what is it's "Bar" property. The "na" property is short for "name", I believe. 00:28
lue indeed. "Name" returns 0, as does "na1" and "Unicode_1_Name"
Mouq lue: Cool :)
lue "na" for some reason returns 1.1
Mouq m: say uniname("☃") 00:29
camelia rakudo-moar 02329b: OUTPUT«SNOWMAN␤»
00:29 btyler left
Mouq & 00:29
BenGoldberg say uniname("\c[PILE OF POO]")
r: say uniname("\c[PILE OF POO]")
camelia rakudo-jvm 02329b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name PILE OF POO␤at /tmp/tmpfile:1␤------> say uniname("\c[PILE OF POO⏏]")␤»
..rakudo-parrot 02329b: OUTPUT«uniname NYI on parrot backend␤ in sub uniname at gen/parrot/CORE.setting:6756␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar 02329b: OUTPUT«PILE OF POO␤»
dalek ecs: 3384acd | lue++ | S15-unicode.pod:
[S15] Refine meaning of uniprop, add unimatch, unival, unibool

An S15 updated informed by an implementation? Madness!
BenGoldberg How useful is it that property names ("na", "na1", "Unicode_1-Name", etc) be strings? Wouldn't an Enum be a better interaface? 00:32
lue BenGoldberg: enumerations, at least in my view, should be limited to the values of E-type properties. No other property value types, nor properties. 00:34
BenGoldberg In particular, it would allow most misspellings to be caught at compile time.
lue besides, there are variations in spelling that the consortium doesn't consider significant (e.g., IIRC, "na1" and "nA1" are equivalent) 00:35
Mouq ...why. 00:38
also lue++
lue why what? 00:39
Mouq »"na1" and "nA1"« 00:40
or is that just a general rule?
lue a general rule, not a special case :) Unicode_1_Name and uNICODE1nAME are also equivalent, if memory serves. 00:41
Mouq Ohh, ok, cool
It would be easiest just to standardize on e.g. wordcase, if that doesn't break anything 00:44
and collapse <[_-]>
but I know virtually nothing about this :P
segomos finally got perl6 on moar .. this is p cool
Mouq *standardize as in wordcase input like "na"
segomos: Yup :) 00:45
lue I don't know if _ and - collapse, actually
Mouq: there's nothing against internally normalizing property names before sending them off, I just don't want to enforce particular spellings on the enduser when Unicode has no such rules. 00:46
Mouq lue: I agree 00:47
lue Mouq: see www.unicode.org/Public/UCD/latest/u...liases.txt , "# Loose matching should be applied ..." 00:48
Mouq m: say uniprop('(','bmg') 00:57
camelia rakudo-moar 02329b: OUTPUT«0029␤»
Mouq needs to read www.unicode.org/reports/tr44/ in the first place 01:00
01:05 spider-mario left
Mouq nqp: my @named := [15,42]; sub foo (:$named) {say($_) for $named}; foo(:@named) 01:12
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«15␤42␤»
Mouq Coool
01:15 rurban left, redhands joined
redhands perl6: say 3; 01:16
camelia rakudo-parrot 02329b, rakudo-jvm 02329b, rakudo-moar 02329b, niecza v24-109-g48a8de3: OUTPUT«3␤»
dalek ast: a7ff499 | lue++ | S15-string (6 files):
S15-strings/ -> S15-string-types/

  "strings" is far too vague and general to be a meaningful directory name
  :) .		 01:17
ast: ece1435 | lue++ | S15-unicode-information/uniname.t:
Add tests for uniname()
Mouq Heh
lue__
lue++
lue^^ 01:18
Mouq raisin it to the roof
lue apparently _ and ^ aren't considered to create word boundaries by my client, for highlighting purposes :)
lue afk 01:19
Mouq p6: say 'lue__' ~~ /lue»/ 01:33
camelia rakudo-parrot 02329b, rakudo-jvm 02329b, rakudo-moar 02329b, niecza v24-109-g48a8de3: OUTPUT«Nil␤»
Mouq p6: say 'lue^^' ~~ /lue»/
camelia rakudo-parrot 02329b, rakudo-jvm 02329b, rakudo-moar 02329b, niecza v24-109-g48a8de3: OUTPUT«｢lue｣␤␤»
01:37 __thou left 01:46 bluescreen10 left 01:47 flussence joined 01:48 __thou joined, redhands left 01:52 tgt joined
Mouq almost has E<> working like he wants 01:52
grondilu r: constant @a = map {$_}, <foo bar>; constant @b = @a xx 2; say ~@b 01:53
camelia rakudo-parrot 02329b: OUTPUT«foo bar foo bar␤» 01:54
..rakudo-jvm 02329b, rakudo-moar 02329b: OUTPUT«␤»
grondilu r: constant @a = map {$_}, <foo bar>; constant @b = @a; say ~@b
camelia rakudo-parrot 02329b, rakudo-jvm 02329b, rakudo-moar 02329b: OUTPUT«foo bar␤»
grondilu r: constant @a = <foo bar>; constant @b = @a xx 2; say ~@b
camelia rakudo-parrot 02329b, rakudo-jvm 02329b, rakudo-moar 02329b: OUTPUT«foo bar foo bar␤» 01:55
Mouq "Cannot invoke this object (REPR: P6opaque, cs = 0)" That's a new one for me..
01:56 fridim__ left 01:57 tgt left
Mouq nqp: my %a := { a=>[1],b=>[2] }; say($a<b>[0]) 02:05
camelia nqp-parrot: OUTPUT«Use of undeclared variable '$a' at line 2, near "<b>[0])"␤current instr.: 'panic' pc 15934 (gen/parrot/stage2/NQPHLL.pir:5941) (gen/parrot/stage2/NQPHLL.nqp:425)␤»
..nqp-jvm: OUTPUT«Use of undeclared variable '$a' at line 2, near "<b>[0])"␤ in panic (gen/jvm/stage2/NQPHLL.nqp:378)␤ in (gen/jvm/stage2/NQP.nqp:2379)␤ in (gen/jvm/stage2/NQP.nqp:2370)␤ in variable (gen/jvm/stage2/NQP.nqp:2301)␤ in !reduce (gen/jvm/stage2/QRegex.nqp:…»
..nqp-moarvm: OUTPUT«Use of undeclared variable '$a' at line 2, near "<b>[0])"␤ at gen/moar/stage2/NQPHLL.nqp:369 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm:panic:120)␤ from gen/moar/stage2/NQP.nqp:2374 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/nqp.moarvm:…»
Mouq nqp: my %a := { a=>[1],b=>[2] }; say(%a<b>[0])
camelia nqp-jvm: OUTPUT«CodeRef representation does not implement at_key_boxed␤ in (/tmp/tmpfile:1)␤ in (gen/jvm/stage2/NQPHLL.nqp:1099)␤ in eval (gen/jvm/stage2/NQPHLL.nqp:1085)␤ in evalfiles (gen/jvm/stage2/NQPHLL.nqp:1291)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1195…»
..nqp-parrot: OUTPUT«get_pmc_keyed() not implemented in class 'Sub'␤current instr.: '' pc 65 ((file unknown):51) (/tmp/tmpfile:1)␤»
..nqp-moarvm: OUTPUT«This representation (MVMCode) does not support associative access␤ at /tmp/tmpfile:1 (<ephemeral file>::26)␤ from gen/moar/stage2/NQPHLL.nqp:1090 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm::94)␤ from gen/moar/stage2/NQPHLL.nqp:1076 (/ho…»
Mouq nqp: my %a := a=>[1],b=>[2]; say(%a<b>[0])
camelia nqp-parrot: OUTPUT«1␤»
..nqp-moarvm: OUTPUT«This representation (VMArray) does not support associative access␤ at /tmp/tmpfile:1 (<ephemeral file>::40)␤ from gen/moar/stage2/NQPHLL.nqp:1090 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm::94)␤ from gen/moar/stage2/NQPHLL.nqp:1076 (/ho…»
..nqp-jvm: OUTPUT«VMArray representation does not implement at_key_boxed␤ in (/tmp/tmpfile:1)␤ in (gen/jvm/stage2/NQPHLL.nqp:1099)␤ in eval (gen/jvm/stage2/NQPHLL.nqp:1085)␤ in evalfiles (gen/jvm/stage2/NQPHLL.nqp:1291)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1195…»
02:07 __thou left
Mouq is more disappointed with this than he wishes to describe.... 02:07
nqp: my %a; %a{$_} := $_.value for "a"=>[1],b=>[2]; say(%a<b>[0]) 02:09
camelia nqp-parrot: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near ">[1],b=>[2"␤current instr.: 'panic' pc 15934 (gen/parrot/stage2/NQPHLL.pir:5941) (gen/parrot/stage2/NQPHLL.nqp:425)␤»
..nqp-moarvm: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near ">[1],b=>[2"␤ at gen/moar/stage2/NQPHLL.nqp:369 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm:panic:120)␤ from <unknown>:1 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/n…»
..nqp-jvm: OUTPUT«Assignment ("=") not supported in NQP, use ":=" instead at line 2, near ">[1],b=>[2"␤ in panic (gen/jvm/stage2/NQPHLL.nqp:378)␤ in infix:sym<=> (gen/jvm/stage2/NQP.nqp)␤ in !protoregex (gen/jvm/stage2/QRegex.nqp:784)␤ in infix (gen/jvm/stage2/NQPHLL.nqp)…»
Mouq nqp: my %a; %a{$_.key} := $_.value for a=>[1],b=>[2]; say(%a<b>[0])
camelia nqp-jvm: OUTPUT«Method 'key' not found for invocant of class 'NQPArray'␤ in (/tmp/tmpfile:1)␤ in (/tmp/tmpfile:1)␤ in (gen/jvm/stage2/NQPHLL.nqp:1099)␤ in eval (gen/jvm/stage2/NQPHLL.nqp:1085)␤ in evalfiles (gen/jvm/stage2/NQPHLL.nqp:1291)␤ in command_eval (gen/…»
..nqp-parrot: OUTPUT«Method 'key' not found for invocant of class 'QRPA'␤current instr.: '' pc 229 ((file unknown):117) (/tmp/tmpfile:1)␤»
..nqp-moarvm: OUTPUT«Cannot find method 'key'␤ at /tmp/tmpfile:1 (<ephemeral file>::8)␤ from /tmp/tmpfile:1 (<ephemeral file>::47)␤ from gen/moar/stage2/NQPHLL.nqp:1090 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/NQPHLL.moarvm::94)␤ from gen/moar/stage2/NQPHLL.nqp:1076 …»
02:09 fridim__ joined
Mouq nqp: my %a; for "a",[1],"b",[2] ->$k,$v{ %a{$k} := $v }; say(%a<b>[0]) 02:11
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«2␤»
Mouq That's a good enough work-aroun
d
TimToady
.oO( Not Quite Productive :)
Mouq :)
02:13 woosley joined
Mouq TIMTOTDIAOOTWINQP 02:22
*TIMTOWTDIAOOWINQP 02:23
Looks good though
Just need to fix Pod::To::Text to support it... 02:25
perl6 --doc -e'=pod E<SNOWMAN;95;ntilde>' 02:26
☃_ñ
02:27 Sqirrel left 02:28 Sqirrel joined 02:33 klapperl left 02:34 __thou joined
dalek kudo/nom: eff1c27 | Mouq++ | / (4 files):
Do E<> properly.

The @contents should contain the actual sequences, with the actual codes used in the E<> being stored correspondingly in @meta in case the Pod formatter needs it		 02:35
02:35 klapperl joined
Mouq (Also pushed to Pod::To::HTML) 02:37
02:38 kaare__ joined
dalek rl6-roast-data: 532ad2e | coke++ | / (6 files):
today (automated commit)
 02:38
Mouq [Coke]++
I'll fix the formatting code tests 02:39
02:41 kaare_ left 02:49 FOAD left
dalek ast: ac4bb81 | Mouq++ | S26-documentation/08-formattingcodes.t:
Use latest formatting code semantics
 02:50
[Coke] ok. we should be back to daily runs. they're going to run later in the day. 02:51
Mouq \o/
02:53 tgt joined 02:55 __thou left 02:57 tgt left 03:02 FOAD joined
dalek kudo/nom: 3c97ad6 | Mouq++ | src/Perl6/Grammar.nqp:
Don't panic on bad Pod formatting
 03:07
Mouq lue: You should be able to generate S26 again without modifications 03:08
As in, I was able to do that
[Coke] moritz: any idea how to fix this: 03:12
pugs: user error ((0,"this version of PCRE is compiled without UTF support"))
camelia pugs: OUTPUT«*** No such subroutine: "&user"␤ at /tmp/nwJKpSVo4T line 1, column 1 - line 2, column 1␤»
03:14 skids left
dalek rl6-roast-data: 7922c11 | coke++ | / (2 files):
Very slight improvement to pugs

  (needed to setup cabal on the new box)		 03:15
lue uh-oh: "java.lang.NoClassDefFoundError: org/objectweb/asm/tree/MethodNode" 03:16
This happens immediately after "/usr/local/bin/nqp-j --target=jar --output=blib/Perl6/Pod.jar --encoding=utf8 \ 03:18
src/Perl6/Pod.nqp"
Mouq ...O.o 03:20
That's my fault, but... I'm not sure what's going wrong 03:21
03:22 grondilu left
Mouq :| I get it too 03:30
dalek p: 4809918 | larry++ | / (2 files):
bump MOAR rev
 03:33
lue
.oO(all aboard the bisect train!)
 03:34
Mouq nqp: class A { my %x; for 1,2,3,4 { %x{$_} := 42 }; }
camelia ( no output )
Mouq lue: It's my penultimate commit
lue Mouq: I figure, I just want to confirm for myself :) [also, bisect practice, don't get to do it often] 03:35
Mouq github.com/rakudo/rakudo/commit/eff1c2
dalek kudo/nom: b1b1d52 | larry++ | src/core/Str.pm:
refactor uniprops with better caching
kudo/nom: 87399b9 | larry++ | tools/build/NQP_REVISION:
bump NQP rev
Mouq lue: k :) I'll see if I can figue out what's going on
lue Mouq: it's definitely something to do with your huge list of entities. 03:41
Mouq lue: I tried getting around it some other way, but it only works if I remove almost all the entities from the list -_- 03:45
lue Mouq: could you possibly assign the list to a @var and then iterate over that? Would make the for loop look nicer :) 03:46
Mouq lue: Didn't help :/ 03:48
lue (Note: if it's impossible/too difficult to get E<> handling working correctly within rakudo, then it's OK for some Pod::Entities module to handle it. I only ever didn't like the idea of every converter module needing to write its own handlers.) 03:49
Mouq lue: Looks like it's the only practical choice 03:51
I'm just going to remove it for now 03:52
lue Yeah. I'm not too concerned with relegating to modules; POD apparently has such helpers for it, and Pod::To::HTML wants a treewalker helper module too :)
03:54 tgt joined
dalek kudo/nom: beeafc1 | Mouq++ | src/Perl6/ (2 files):
HTML entities list made JVM fail make install. removing
 03:56
Mouq So. That's the quick fix.
I'm seriously out of tuits to do it right :/
03:57 boks joined
boks hi guys 03:57
Mouq o/ boks
lue Mouq: I'm perfectly fine if we defer E<> handling to some external module. 03:58
03:58 tgt left
dalek kudo/nom: fa46c8e | Mouq++ | src/Perl6/Actions.nqp:
Put parsed number from E<> into @.meta
 03:59
Mouq lue: Well there ^^
boks ei guys do you have any available links to share for sybase perl materials 04:00
04:00 molaf joined
Mouq boks: Like this? perl6.org/documentation/ :) 04:00
lue Mouq: that collapses 0x, 0o, etc. into a number, but nothing else, right? 04:01
Mouq lue: Yup
lue "nothing else" wrt before attempting to handle E<>, that is.
Mouq So you can reinstate what Pod::To::HTML used to do, just use @.meta instead of @.contents 04:02
If that makes sense.
I need to go..
o/ #perl6
boks ei guys, i need your inputs. how do you study perl? how things motivated you to learn perl quickly. kinda hard to absorb perl coz i'm not in a programming side i'm more on administration side. 04:11
04:11 jnap left, Mouq left 04:23 woosley left 04:25 woosley joined 04:26 woosley left, woosley joined 04:34 __thou joined 04:42 jnap joined 04:47 jnap left
Timbus boks, it sounds like you are probably looking for perl 5 material, not perl 6. 04:47
also to answer your question, i guess 'make something with it' is the best advice I can personally give. 04:48
boks do we have channel for perl5? 04:54
04:55 tgt joined
Timbus well, #perl has nearly 600 people in it. but i think there's a separate irc/channel for specifically learning perl 04:58
ask around :) 04:59
04:59 tgt left
masak good early morning, #perl6 05:03
boks: on the irc.perl.org server, there's #perl-help 05:04
05:11 cxreg joined 05:22 [Sno] left 05:25 telex left 05:26 telex joined
moritz good morning 05:30
[Coke]: no idea, sorry
05:36 fridim__ left 05:42 jnap joined 05:46 BenGoldberg left 05:47 jnap left 05:55 tgt joined 05:59 tgt left 06:17 xinming_ joined, xinming left 06:18 go|dfish left 06:20 go|dfish joined
TimToady m: say 'A'.&uniprop-bool('Any') # oddly, this works locally 06:31
camelia rakudo-moar fa46c8: OUTPUT«(signal SEGV)»
TimToady something must be different in how it was compiled 06:32
06:37 [Sno] joined 06:43 jnap joined 06:48 jnap left 06:49 colomon left 06:56 tgt joined 07:00 tgt left 07:20 tgt joined 07:23 tgt left, xinming joined 07:26 xinming_ left 07:27 FROGGS joined 07:35 woosley left, woosley joined 07:40 woosley left, woosley joined 07:44 jnap joined, woosley left, woosley joined 07:48 jnap left 07:50 Rotwang joined 07:59 kaare__ is now known as kaare_ 08:02 darutoko joined 08:05 zakharyas joined, JimmyZ_ joined, perigrin_ joined 08:07 colomon joined 08:08 kaleem joined 08:11 klapperl left, salv0 left, Alina-malina left, vendethiel left, pdcawley left, itz left, JimmyZ left, perigrin left, FROGGS left, Sqirrel left, gfldex left, camelia left, dayangkun left, aindilis left, wooden left, Alina-malina joined, __thou left, vendethiel joined 08:21 tgt joined 08:23 salv0 joined, tgt left 08:24 virtualsue joined 08:26 nebuchadnezzar joined 08:27 woosley left 08:28 woosley1 joined 08:33 arnsholt left 08:43 pdcawley joined 08:44 Rotwang left 08:45 jnap joined 08:49 jnap left 08:52 itz joined 08:53 boks left 09:01 dmol joined, wooden joined, aindilis joined, dayangkun joined, camelia joined, gfldex joined, Sqirrel joined, klapperl_ joined, FROGGS joined, barjavel.freenode.net sets mode: +v camelia 09:02 woosley joined 09:03 kbaker joined, isacloud_ left 09:04 woosley left 09:05 isacloud__ joined, isacloud__ left, isacloud__ joined, isacloud__ is now known as isacloud_ 09:06 woosley joined, woosley1 left 09:10 woosley1 joined 09:11 JimmyZ_ is now known as JimmyZ 09:12 JimmyZ left, JimmyZ joined, woosley left, Adriaaan joined
timotimo nqp: my @things := ["foo", [1], "bar", [2], "baz", [1, 2, 3]]; my %test := nqp::hash(|@things); say(%test<baz>[0]); say(%test<baz>[1]); say(%test<foo>); 09:16
camelia nqp-jvm: OUTPUT«Method 'type' not found for invocant of class 'NQPMu'␤ in coerce (gen/jvm/stage2/QAST.nqp:4191)␤ in as_jast (gen/jvm/stage2/QAST.nqp:2974)␤ in (gen/jvm/stage2/QAST.nqp:3771)␤ in compile_all_the_stmts (gen/jvm/stage2/QAST.nqp:3759)␤ in as_jast (gen/jvm…» 09:17
..nqp-parrot: OUTPUT«Error while compiling op hash (source text: "nqp::hash(|@things)"): Unknown QAST node type NQPMu␤current instr.: '' pc 57002 (gen/parrot/stage2/QAST.pir:21033) (gen/parrot/stage2/QAST.nqp:3629)␤»
..nqp-moarvm: OUTPUT«Error while compiling op hash (source text: "nqp::hash(|@things)"): Iteration past end of iterator␤ at gen/moar/stage2/QAST.nqp:4471 (/home/p6eval/rakudo-inst-1/languages/nqp/lib/QAST.moarvm::17)␤ from gen/moar/stage2/QAST.nqp:4470 (/home/p6eval/rakudo-ins…»
timotimo oh, whoops :)
nqp is ticklish
09:29 kurahaupo_mobile joined 09:30 Adriaaan left 09:32 rindolf joined 09:34 woosley joined, woosley1 left 09:39 woosley left 09:41 woosley joined 09:45 jnap joined 09:50 jnap left 09:52 kurahaupo_mobile left 09:54 dakkar joined 09:55 kurahaupo_mobile joined, kurahaupo_mobile left 10:05 Pleiades` left 10:11 Pleiades` joined 10:29 woosley left 10:35 dakkar left 10:43 arnsholt joined 10:46 jnap joined
Probably_Tadzik Juerd: ping 10:49
10:50 jnap left 10:52 xinming left
FROGGS an imposter! kill it! 10:54
timotimo to be honest, as long as the imposter-tadzik does the same cool work as the original tadzik, i wouldn't mind keeping them around 10:55
at least they're somewhat up front about the (possibility of) imposterization 10:56
Probably_Tadzik :) 10:58
10:58 xinming joined
FROGGS I even start to like the new one better :o) 10:59
11:11 xenoterracide left 11:26 sftp left 11:27 sftp joined
arnsholt Probably_Tadzik: How slow was your asteroids thing on JVM? 11:31
(Or not slow, as the case may be) 11:32
I've been reading a bit about invokedynamic and there might be some speedups to be had there
11:32 denisboyun joined
timotimo what's the name for the kind of typing we have? optional typing? incremental typing? 11:33
colomon 's overnight smoke run was seriously messed up. :\ 11:34
11:35 sftp left
Probably_Tadzik arnsholt: frankly, it feels about as fast as the C verison :) 11:40
I just played it until almost-winning, and there was only one noticable GC pause 11:41
timotimo almost-winning meaning that the win condition is NYI? ;) 11:42
11:42 virtualsue left, aindilis left
Probably_Tadzik hard to say, I never reached it :P 11:43
11:43 tadzik1 joined
tadzik1 ...great, even weirder nickname 11:43
timotimo: hard to say, I never reached it :P
also, the typing is called Gradual afaik
11:43 sftp joined, d4l3k_ joined 11:45 aindilis` joined, dalek left, Probably_Tadzik left 11:46 d4l3k_ is now known as dalek 11:47 jnap joined 11:49 Adriaaan joined, nebuchad` joined
timotimo :D 11:51
timotimo is afk for a bit
probably_tadzik should update his weechat.
11:51 sivoais_ joined, jnap left, pochi joined 11:52 Adriaaan left, nwc10_ joined, ruoso joined 11:53 zamolxes joined 11:54 arnsholt_ joined 11:56 nebuchadnezzar left, timotimo left, go|dfish left, japhb_ left, arnsholt left, timotimo joined 11:57 ggherdov joined, japhb joined
moritz ... probably. 12:00
12:02 spider-mario joined 12:06 bjz left 12:07 bjz joined
timotimo yes, i just remembered that while taking a shower 12:07
hm. what disconnected me?
FROGGS some sort of netsplit
timotimo irc: reading data on socket: error -9 A TLS packet with unexpected length was received.
lizmat g
timotimo o_O
lizmat ra 12:08
dua
l ty
ping
FROGGS lizmat: all is well? :o)
lizmat yes, just busy offline
FROGGS yeah, me too these days :/ 12:09
lizmat glad to see the channel is busy
timotimo hey lizmat, nice to see you :)
lizmat and the commits flowing in
timotimo yes, you saw TimToady contributing lots of things during the last days? :D
lizmat TimToady++
jnthn o/ lizmat :) 12:11
lizmat jnthn, FROGGS, timotimo o/ 12:12
jnthn also hasn't had much Perl 6 time for a little while... 12:13
timotimo o/ jnthn
jnthn Things will be back to normal from Friday evening / this weekend, though :)
lizmat will also get back to rakudo in preparation for the QA hackathon in Lyon next week 12:14
jnthn This weekend I'll look into serialization issues on Moar that block some modules. 12:15
And then get back to moar-conc.
lizmat is looking forward to doing conc on moar
lizmat is looking forward to doing conc on moar 12:16
FROGGS ++jnthn :o) 12:17
timotimo m: say "i'm looking forward to doing concurrency".comb.pick(*).join("") 12:20
camelia rakudo-moar fa46c8: OUTPUT«nctmnoknroir dieo alfgcur gy'o wrdic noo␤»
jnthn noo!
moritz timotimo: fwiw '' is the default for .join 12:21
timotimo oh good point
12:26 tgt joined 12:30 xenoterracide joined
colomon lizmat++ 12:35
12:39 xenoterracide left
lizmat colomon o/ 12:40
afk again&
12:46 cognominal left, cognominal joined 12:48 dayangkun left, jnap joined
tadzik1 moritz, ping 12:48
moritz tadzik1: pong 12:49
tadzik1 moritz: do you have root access on feather1? 12:50
moritz tadzik1: aye
tadzik1 could you help me regain access to my account?
neither the password nor the ssh key works
moritz tadzik1: should I set a new password and send it by email?
12:51 dayangkun joined
tadzik1 moritz: let's see if that works 12:51
moritz tadzik1: also, what's your login? there doesn't seem to be a 'tadzik'
tadzik1 it's tjs
I've no idea why I'd suddenly lose access, weird
moritz tadzik1: mail sent 12:52
tadzik1: if it doesn't work, please bother me again
12:52 jnap left 12:53 Sqirrel left
tadzik1 moritz: nope, doesn't let me in :/ 12:54
moritz tadzik1: seems your login shell was lost during the upgrade
tadzik1 uh 12:55
moritz sorry about that
moritz installs zsh again
12:56 pecastro left
moritz urks, lots of unconfigured packages on feather... 12:57
12:57 pecastro joined
moritz tadzik1: done, please try again 12:59
13:01 xinming left 13:07 tgt left, tgt_ joined
tadzik hooray 13:24
13:24 tadzik1 left
tadzik moritz++ 13:24
timotimo greetings original tadzik 13:25
tadzik The Original Tadzik 13:26
13:27 sisar joined
moritz The Original Tadzik, Formerly Known As tadzik 13:27
sisar i once came across a website which listed design flaws, or things people dislike about a programming language. It had a lot of programming language in the list. For the life of me, I can't remember the name of the website, and google did not help. Does anyone by chance knows what I'm talking about ? 13:29
moritz my memory only comes up with the C++ FQA, but that's C++ only 13:31
sisar hmm 13:36
FROGGS .oO( To boldly go where no tadzik has gone before ) 13:38
moritz
.oO( Probably Tadzik = Just Tadzik | Nothing )
 13:44
13:44 xinming joined
hoelzro timotimo++ # weekly post 13:47
sad to see that that precomp bug is still present =(
moritz hoelzro: jnthn said he'll look into it on the weekend
hoelzro \o/
moritz (the one on moarvm, that is) 13:48
13:48 jnap joined
hoelzro right 13:49
if I had to wager a guess
I'd guess it's in the NQP ModuleLoader
because it occurs on the JVM as well
and I'm betting there's some copy pasting between the two impls 13:50
13:50 jnap left, jnap1 joined 13:51 go|dfish joined 13:54 tgt_ left 14:04 nebuchad` is now known as nebuchadnezzar 14:05 kaare_ left 14:06 PZt left 14:13 guru joined, guru is now known as Guest61973 14:14 Guest61973 is now known as ajr_ 14:16 xragnar_ joined, xragnar is now known as Guest20324, Guest20324 left, xragnar_ is now known as xragnar 14:19 PZt joined 14:20 PZt left 14:21 skids joined, PZt joined
xfix sisar, wiki.theory.org/YourLanguageSucks 14:25
This?
sisar xfix: hmm, interesting. But the site I remember had a different interface. I wonder if the site changed interface. 14:28
xfix: thanks, by the way :-)
xfix "hashes don't stay in order like they do in most other languages" 14:33
I just decided to reread this website
uh
14:33 hoverboard joined
xfix Who would depend on hash order? 14:34
jnthn I'm not sure that's true of "most languages" at all...
xfix Yeah, that too.
The only language I can think of is PHP, and possibly JavaScript (but only in current implementations, not specification).
14:34 fridim__ joined
xfix (and in JS, many implementations break the order when you have numeric keys) 14:35
(I'm almost sure that ordered keys in JS are just a compatibility feature)
moritz what about lua? 14:43
14:49 sisar left 14:58 __thou joined 15:02 btyler joined
vendethiel "Perl is worse than Python because people wanted it worse. Larry Wall, 14 Oct 1998" 15:13
15:15 treehug88 joined 15:16 kaleem left, kaleem joined
timotimo xfix o/ 15:17
15:20 virtualsue joined 15:21 xinming left 15:29 kaleem left
timotimo do you like my blog? :) 15:30
15:31 dmol left 15:33 xinming joined 15:40 arnsholt_ left, siruf joined 15:42 genehack joined 15:44 genehack left 15:49 dayangkun left 15:50 grondilu joined
grondilu can't compile rakudo on MoarVM. fails during configure. 15:51
grondilu tries again to show you the error message
FROGGS k
PerlJam I tried compiling a little while ago and got a failure too. 15:53
It didn't like \« in a regex in src/Perl6/World.nqp. I need to use '«' for it to compile 15:54
oh, sorry, it was », not «
line 2507
grondilu paste.siduction.org/20140305155437
"Whitespace character is not allowed as a delimiter at line 1555, near "«==» { <"" 15:55
PerlJam huh ... that's very similar to the error I got. 15:56
moritz PerlJam: so rakudo needs patching?
PerlJam hold on a sec, I'll paste the error I got
gist.github.com/perlpilot/9369982
moritz: perhaps? \» worked before, but now fails. Not sure why. 15:57
Seems like there's some problem parsing » characters
moritz locale-dependent Unicode/encoding fuckup? 15:58
FROGGS nqp: say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, '«==»', 0))
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«0␤»
FROGGS nqp: say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, '«==»', 1)) 15:59
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«0␤»
FROGGS hmmm
grondilu: is one of the two true on your box?
PerlJam -> conference call & 16:00
FROGGS nqp-m: say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, 'sym«==»', 0)) # grondilu: please test this 16:01
camelia nqp-moarvm: OUTPUT«0␤»
FROGGS hmmm, if you had a nqp-m ... :/
nqp-m: say(nqp::ord('sym«==»')) 16:02
camelia nqp-moarvm: OUTPUT«115␤»
grondilu FROGGS: hang on 16:03
not sure how to run nqp now :/ 16:04
ok, got it 16:05
$ nqp-m -e 'say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, '«==»', 0))' 16:06
0
$ nqp-m -e 'say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, '«==»', 1))'
0
16:06 dayangkun joined
FROGGS and this? say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, 'sym«==»', 0)) 16:06
16:07 denisboyun left, xragnar left
grondilu $ nqp-m -e 'say(nqp::iscclass(nqp::const::CCLASS_WHITESPACE, 'sym«==»', 0))' 16:07
0
TimToady m: say '«==»' ~~ /<space>/
FROGGS weird
camelia rakudo-moar fa46c8: OUTPUT«Nil␤»
TimToady m: say '«==»' ~~ /\s/ 16:08
camelia rakudo-moar fa46c8: OUTPUT«Nil␤»
moritz well, if it thought any of that was whitespace, it wouldn't have build
16:11 [Sno] left, xragnar joined 16:22 xinming left, xinming joined
dalek p: e882556 | (Tobias Leich)++ | src/HLL/Grammar.nqp:
improve error message about not allowed delims
 16:24
kudo/nom: 4345b6a | (Tobias Leich)++ | tools/build/NQP_REVISION:
bump nqp revision for error message improvement
FROGGS grondilu: can you try that and paste the error message? 16:25
TimToady I'm wondering whether we're somehow running into a signed vs unsigned character problem
PerlJam is back 16:26
grondilu FROGGS: what do you want me to try exactly? 16:27
PerlJam TimToady: I was thinking perhaps it was a character/byte problem. Like something is getting out of sync when reading characters and interprets » as bytes
FROGGS grondilu: just pull and build until it crashes
PerlJam Dunno though, that was just a guess when I first noticed the problem
TimToady though with the current code in MVM_string_is_cclass, it would not explain weird latin-1 behavior unless it was to not recognize ws in the latin-1 range; a negative cp would force all of latin1 to return false for WHITESPACE 16:28
jnthn If there's anything going on latin1 wise, it'll have gone wrong way earlier (like, in the latin-1 decoder) 16:29
TimToady otherwise, it'd pull it from the unicode tables, which is the likelier spot for a failure at the point
grondilu ok
TimToady *this point
jnthn Because by the time you've an MVMString, it's turned the latin-1 into the NFG-ish thing.
And since we don't really NFG yet, there should simply never be a negative at present...
TimToady but I should check for that and carp there 16:30
since nobody should be calling that code with an NFG char
(of a negative persuasion)
still, my money is on the generated tables somehow 16:36
though everyone should have the same tables, at any given version
16:37 hoverboard left 16:38 hoverboard joined 16:40 dmol joined 16:44 molaf left
grondilu FROGGS: paste.siduction.org/20140305164433 # pretty much the same as before I think 16:45
16:47 sivoais_ is now known as sivoais 16:48 sivoais left, sivoais joined
TimToady I wonder if it's really complaining about the ws after the » 16:48
is this with the most recent version? 16:51
16:52 xinming left
TimToady oh, yeah, reading backlog, so nm 16:52
16:53 jnap1 left, xinming joined 16:54 arnsholt joined 16:56 molaf joined
TimToady are you doing --gen-moar or --gen-moar=master? 16:56
that is, what's your Configure line? 16:57
timotimo jnthn: are you going to be tuited tomorrow?
grondilu $ perl Configure.pl --gen-{nqp,moar} --backends=moar
TimToady so no =master 16:58
lemme try that
16:58 xinming_ joined 16:59 xinming left, zakharyas left
jnthn timotimo: No, 'fraid not. 16:59
timotimo: Will be Friday evening at earliest. Depends how tired I am. 17:00
FROGGS grondilu: you are not using HEAD
TimToady should be close to HEAD, I bumped last night 17:01
FROGGS the current msg should be like: Whitespace character (0x73) is not allowed as a delimiter at ...
timotimo jnthn: that's okay. do you have suggestions for what i could do to feel useful? ;) 17:02
except finish up the block inlining code
17:03 bowtie left, xinming joined, xinming_ left
TimToady FROGGS: I see no such message in my nqp stages 17:04
(that include the char)
FROGGS TimToady: I pushed that 40min ago 17:05
TimToady ah, so I see
FROGGS so we know what char is exploded on grondilu's box
TimToady but it's in the stage0, so is that gonna help?
FROGGS ohh, hmmm 17:06
17:07 bluescreen10 joined
TimToady I just did the headless configure and it built fine 17:07
but that's after rm -rf install nqp
grondilu: how violently did you clean up the previous version?
or anyone else who's repro'd this? 17:08
FROGGS it built fine for me too, without rm -rf'ing anything
17:09 kurahaupo_mobile joined 17:10 bowtie joined
PerlJam I didn't nuke the nqp dir before I did my build, but I did a git clean -dxf 17:10
TimToady grumbles about error messages that say "near" 17:13
17:16 jnap joined 17:17 hoverboard left
TimToady I guess it's really at stage1 17:17
I note that that is the first occurance of « in the file 17:19
grondilu TimToady: I did not clean it at all. I thought it was automatic. 17:28
grondilu removes the install dir
timotimo i seem to think my apartment gets cleaned automatically, too. that would explain why it's so dirty :P 17:29
17:29 [Sno] joined
grondilu well, now that you mentio it, I do remember that removing install and blib are required from times to times. 17:30
17:32 kbaker left
TimToady in this case, removing nqp/ might be necessary 17:34
colomon rebooted the server and reran the smoke test, and we're back to the same state as yesterday, except someone++ fixed CSS::Grammar.
17:38 hoverboard joined, hoverboard is now known as Guest96733 17:39 virtualsue left
grondilu well, the build failed again, despite the fact that I had removed blib and install. Did I forget something? 17:39
.oO( mabe nqp )
 17:40
grondilu emoves nqp
17:42 Guest96733 left, Guest96733 joined, Guest96733 is now known as hoverboard
masak "emove" should totally be a word. 17:42
itz_ yeah its odd "make clean" doesn't rm npq 17:43
maybe it should
masak itz_: is there a "make realclean" -- there used to be one that did that. 17:44
moritz nope
'make realclean' also cleans out the Makefile
17:44 kurahaupo_mobile left
timotimo it really shouldn't - except if you --gen-nqp'd 17:44
moritz but not NQP, which is a different project
itz_ I meant the "clean" target in the nqp Makefile 17:49
which doesn't rm nqp .. maybe there should be a realclean like rakudo and MoarVM which does rm nqp
grondilu hum... got the error again. 17:50
FROGGS :( 17:51
17:52 ajr_ left 17:53 guru joined
TimToady did it include did it include the hex code in the error message this time? 17:53
17:54 guru is now known as Guest71724, Guest71724 is now known as ajr_
grondilu TimToady: not sure what hex code you're talking about. Here is the error message: paste.siduction.org/20140305175524 17:55
TimToady no, it didn't, probably because it's using the old stage 0 to try to compile the stage 1 17:57
grondilu isn't the old stage zero in nqp/MoarVM? If so, since I did remove it, why doesn't it get updated to the latest? 17:58
17:58 ajr_ left
FROGGS grondilu: no, stage0 is in nqp 17:59
TimToady what kind of a machine are you running on? which OS? 18:00
grondilu Linux redkey 3.13-3.towo-siduction-686 #1 SMP PREEMPT Fri Feb 14 13:57:25 UTC 2014 i686 GNU/Linux 18:01
TimToady so little-endian 18:02
grondilu indeed
TimToady any special locale settings? 18:03
dalek p: 90f2864 | larry++ | tools/build/MOAR_REVISION:
bump MOAR to get negative char test
 18:05
TimToady m: say "«==»".encode("utf-8").decode("latin-1") 18:06
camelia rakudo-moar fa46c8: OUTPUT«Can not decode a utf-8 buffer as if it were latin-1␤ in method decode at src/gen/m-CORE.setting:5355␤ in block at /tmp/m0mews4FLi:1␤␤»
TimToady !!!
masak hee 18:07
jnthn .encode('utf-8') returns a utf8, not a generic Buf.
And it's "smart enough" to "help" you not decode it wrongly...
masak :/ 18:08
jnthn Made sense at the itme I worte it that way :P
detrain &
TimToady m: say Buf.new("«==»".encode("utf-8")[]).decode("latin-1")
camelia rakudo-moar fa46c8: OUTPUT«Â«==»␤»
masak \o/
pmichaud good afternoon, #perl6
masak pmichaud! \o/ 18:09
TimToady it would be complaining about a spurious » if it were taking  as the delimiter
18:10 tgt joined 18:15 treehug8_ joined
dalek kudo/nom: 5d084b8 | larry++ | tools/build/NQP_REVISION:
bump nqp to propagate negative char check to nom
 18:16
TimToady so it seems like it has to be a CCLASS_WHITESPACE issue, not an encoding issue 18:17
18:17 treehug88 left
TimToady but it seems to be an issue in the stage0 compiler 18:18
18:29 Rotwang joined
grondilu is back 18:30
TimToady which, at the time, was purely using MVM_UNICODE_PROPERTY_WHITE_SPACE, so shouldn't have any locale dependencies, nor should it vary from machine to machine 18:31
grondilu: is this a 32-bit machine? 18:33
or 64?
grondilu 32
TimToady there's a difference
grondilu (it's an intel 686, so 32 bits) 18:34
TimToady I've noticed some confusion in some of the opcodes where oplist/interp thinks there are 64-bit args and the C code says MVMint32 for the type 18:35
so maybe something of that nature is messing up
TimToady doesn't know what moar does with a C MVMint32 that gets mapped to a .i64 arg in interp.c 18:36
well, MVMCodepoint32 18:38
[Coke] ahahahaha. "oh, Maybe I'll bump up the daily runs on the new box to start now instead of whenever I kicked them off yesterday, slow move them earlier in the day." 18:45
Login to the box. crontab -l. date. ... oh. only have to wait 13 seconds.... there.
FROGGS :o) 18:46
18:46 laidback_01 left
[Coke] now only like 4-5 hours to get the result! 18:47
TimToady but the codepoint for testing whitespace doesn't suffer any 32/64 skew that I can see offhand, unless someone changed how characters are pulled out of strings at an offset 18:49
18:52 guru joined, guru is now known as Guest74452, Guest74452 is now known as ajr_
TimToady if somehow the semantics of MVM_string_get_codepoint_at changed, that might have a version skew effect 18:53
m: say 'A'.&uniprop-bool('Any') 18:55
camelia rakudo-moar 4345b6: OUTPUT«(signal SEGV)»
TimToady I keep wondering if that's a 32-bit problem too
is camelia running on 32-bits?
18:57 yoleaux left
TimToady p: say $*VM<config><archname> 18:57
camelia rakudo-parrot 4345b6: OUTPUT«x86_64-linux-gnu-thread-multi␤» 18:58
TimToady well, so much for that theory :)
but I don't see any 32/64 version skew in the codepath down to MVM_string_is_cclass 19:01
19:01 hoelzro joined, jtpalmer joined
TimToady nothing changed in those signatures 19:01
19:01 retupmoca joined
FROGGS I've got an ubuntu 12.04 x86 19:01
I can perhaps test it on that box
19:02 atrodo joined
TimToady that seems like the best plan at this point; I don't think I have a 32-bit handy 19:02
FROGGS ohh, I love VM's that boot within 3s
19:03 yoleaux joined, ChanServ sets mode: +v yoleaux
TimToady I'll look through unicode_db.c's types to see if there's anything that may be assuming 64-bit 19:03
19:05 ivan`` joined, tchoury joined
FROGGS src/core/interp.c: In Funktion »MVM_interp_run«: 19:06
src/core/interp.c:2106:42: Warnung: Typkonvertierung von Zeiger auf Ganzzahl anderer Breite [-Wpointer-to-int-cast]
it is in GERMAN!! but maybe you understand it :o) 19:07
hmmm, but that is op getwhere 19:08
19:08 darutoko left
TimToady which looks like it is intentionally casting a pointer to an integer 19:11
a 32-bit pointer should fit into a 64-bit integer though 19:12
FROGGS yeah, true
TimToady I don't suppose there are 32-bit machines with busted 64-bit integer support?
FROGGS nqp built fine
TimToady huh
FROGGS but rakudo assplodes: 19:13
/home/froggs/dev/nqp/bin/nqp-m --target=mbc --output=blib/Perl6/World.moarvm --encoding=utf8 \
src/Perl6/World.nqp
No unspace allowed in regex; if you meant to match the literal character, please enclose in single quotes ('»') or use a backslashed form like \xbb at line 2507, near "/ {\n "
TimToady is this with HEAD? 19:14
FROGGS I cloned it just now
TimToady maybe it thinks n is a space? 19:15
FROGGS all three repos
TimToady or misses, and gets the space after?
FROGGS or it thinks / is a backslash
TimToady or that
FROGGS that is the line: if $next ~~ /\)|\]|\}|\»/ 19:16
TimToady you didn't change anything regarding the semantics of string offsets somehow?
FROGGS no
maybe it is timotimo string optimization? 19:17
timotimo++'s*
TimToady maybe
but why did it wait till rakudo this time, if it's the same bug?
FROGGS dunno 19:18
weird is that » is involved in both cases
TimToady ja 19:20
FROGGS ahh, that "('»')" is not an example, it is the thing it is complaining about 19:22
TimToady yes
it a less than LTA message :)
FROGGS so again it thinks that this thingy is whitespace
TimToady yeah
the current code hardwires spaces under '~', maybe we should hardwire everything up to 0xff 19:24
19:25 Sqirrel joined
TimToady w: .ord.base(16).say if /\s/ for "\x0"..."\xff" 19:26
umm 19:27
oh, duh 19:28
masak 'w:' ?
FROGGS w:
hehe
TimToady m: .ord.base(16).say if /\s/ for "\x0"..."\xff"
camelia rakudo-moar 4345b6: OUTPUT«9␤A␤B␤C␤D␤20␤85␤A0␤»
TimToady lysdexics are us
FROGGS that is upsido down then :o)
m: say "»".ord 19:29
camelia rakudo-moar 4345b6: OUTPUT«187␤»
FROGGS m: say "»".ord - 128
camelia rakudo-moar 4345b6: OUTPUT«59␤»
TimToady it's ubsipe pown
m: .ord.base(16).say if /<:space/ for "\x0"..."\xff" 19:30
camelia rakudo-moar 4345b6: OUTPUT«===SORRY!=== Error while compiling /tmp/bJd42q3w0T␤Unable to parse expression in metachar:sym<assert>; couldn't find final '>' ␤at /tmp/bJd42q3w0T:1␤------> .ord.base(16).say if /<:space⏏/ for "\x0"..."\xff"␤ …»
TimToady m: .ord.base(16).say if /<:space>/ for "\x0"..."\xff"
camelia rakudo-moar 4345b6: OUTPUT«20␤»
TimToady m: .ord.base(16).say if /<:whitespace>/ for "\x0"..."\xff"
camelia rakudo-moar 4345b6: OUTPUT«9␤A␤B␤C␤D␤20␤85␤A0␤»
TimToady m: .ord.base(16).say if /<space>/ for "\x0"..."\xff" 19:31
camelia rakudo-moar 4345b6: OUTPUT«9␤A␤B␤C␤D␤20␤85␤A0␤»
19:31 aindilis` left
FROGGS nqp-m -e 'say( "»" ~~ /\»/ )' 19:32
No unspace allowed in regex; if you meant to match the literal character, ...
geekosaur I would ask if the problem is perhaps deeper, like \ is taken only as an unspace introducer
FROGGS nqp-m: say( "»" ~~ /\»/ )
camelia nqp-moarvm: OUTPUT«»␤»
geekosaur like, oh, that
FROGGS weee ud
TimToady nqp-m: say( "»" ~~ /'»'/ ) 19:33
camelia nqp-moarvm: OUTPUT«»␤»
FROGGS my bet is this commit: github.com/perl6/nqp/commit/8aab21...dd618d25b5
though it does not make sense
nqp-m -e 'say( "»" ~~ /"»"/ )' 19:34
»
TimToady nqp-m: say( "»" ~~ /\s/ )
camelia nqp-moarvm: OUTPUT«Substring length (-4) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from gen/moar/stage2/NQPCORE.setting:674 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPCORE.setting.moarvm:…»
TimToady say what
that looks like String Offset damage 19:35
so I think our bug is in pulling the wrong character out to match
FROGGS that -4 is about a failed match
nqp-m: say( "a" ~~ /b/ )
camelia nqp-moarvm: OUTPUT«Substring length (-3) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from gen/moar/stage2/NQPCORE.setting:674 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPCORE.setting.moarvm:…»
FROGGS kinda 19:36
nqp-m: say( "a" ~~ /bb/ )
camelia nqp-moarvm: OUTPUT«Substring length (-3) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from gen/moar/stage2/NQPCORE.setting:674 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPCORE.setting.moarvm:…»
TimToady nqp-m: say( " » " ~~ /<?before '»'>\s/ )
camelia nqp-moarvm: OUTPUT«Substring length (-14) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from gen/moar/stage2/NQPCORE.setting:674 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPCORE.setting.moarvm…»
moritz nqp-m: my $m := 'a' ~~ /bb/; say($m.from); say($m.to); 19:37
camelia nqp-moarvm: OUTPUT«0␤-3␤»
moritz nqp-m: my $m := 'a' ~~ /bb/; say($m.from); say($m.to); say($m.Str)
camelia nqp-moarvm: OUTPUT«Substring length (-3) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from /tmp/vY4FdcFWVY:1 (<ephemeral file>::83)␤ from gen/moar/stage2/NQPHLL.nqp:1091 (/home/p6eval/rakudo-in…»
moritz so it's only .Str that's broken (at least in this example)?
FROGGS moritz: yes
parrot is igno^\Wtolerant about that 19:38
nqp: say( " » " ~~ /<?before '»'>\s/ )
moritz but probably not by substringing, but rather by special-casing
camelia nqp-moarvm: OUTPUT«Substring length (-14) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from gen/moar/stage2/NQPCORE.setting:674 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPCORE.setting.moarvm…»
..nqp-jvm, nqp-parrot: OUTPUT«␤»
TimToady is something in moar not checking for negatives where it should?
FROGGS nqp: my $m := 'a' ~~ /bb/; say($m.from); say($m.to); 19:39
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«0␤-3␤»
moritz TimToady: I'd say Match.Str
FROGGS nqp: my $m := 'a' ~~ /bb/; say($m.from); say($m.to); say($m.Bool)
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«0␤-3␤0␤»
TimToady something that could influence the \s test for «==»?
TimToady needs brane nutrients 19:40
FROGGS perhaps we have to check that: github.com/perl6/nqp/blob/master/s...T.nqp#L990
(because I've written it) 19:41
nqp: say( "»" ~~ /\W/ ) 19:43
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«»␤»
19:43 ruthless joined
timotimo oh, did my optimizations break something again? 19:44
FROGGS timotimo: dunno, was just a wild guess
timotimo i hope it proves untrue
moritz try the build step with --optimize=off?
TimToady there seems to be some problem with « or » matching on 32-bit machines
FROGGS k
moritz: does not help 19:45
timotimo: is there a way to turn off the optimization in moar? 19:46
timotimo you'll have to comment it out, or put false && into the if statement or something
in src/string/ops.c in MVM_flatten_strings
the one that has memcpy stuff in it
FROGGS k
no, still fails 19:48
timotimo yay 19:49
TimToady nqp: say( "»" ~~ /\w/ )
19:49 nwc10_ is now known as nwc10
camelia nqp-moarvm: OUTPUT«Substring length (-4) cannot be negative␤ at gen/moar/stage2/QRegex.nqp:1155 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/QRegex.moarvm:Str:23)␤ from gen/moar/stage2/NQPCORE.setting:674 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPCORE.setting.moarvm:…» 19:49
..nqp-jvm, nqp-parrot: OUTPUT«␤»
TimToady nqp: my $m := '»' ~~ /\W/; say($m.from); say($m.to); say($m.Bool) 19:50
camelia nqp-jvm: OUTPUT«(timeout)» 19:51
..nqp-moarvm, nqp-parrot: OUTPUT«0␤1␤1␤»
FROGGS same on my x86 box
timotimo m: "🗻⛽">>.&uniname.say 19:54
camelia rakudo-moar 4345b6: OUTPUT«MOUNT FUJI␤»
timotimo oh, wrong
m: "🗻⛽".comb>>.&uniname.say
camelia rakudo-moar 4345b6: OUTPUT«MOUNT FUJI FUEL PUMP␤»
timotimo m: "🐉🗻⛽".comb>>.&uniname.say
camelia rakudo-moar 4345b6: OUTPUT«DRAGON MOUNT FUJI FUEL PUMP␤»
TimToady nqp: my $m := '»' ~~ /\s/; say($m.from); say($m.to); say($m.Bool)
camelia nqp-moarvm, nqp-jvm, nqp-parrot: OUTPUT«1␤-3␤0␤» 19:55
tadzik what 19:56
Those might just start being my favourite absurd unicode characters
ah, these are 3 charcters :(
I thought DRAGON MOUNT FUJI FUEL PUMP is a single one
TimToady nqp: my $m := '»' ~~ /<space>/; say($m.from); say($m.to); say($m.Bool) 19:57
camelia nqp-jvm: OUTPUT«(timeout)»
..nqp-moarvm, nqp-parrot: OUTPUT«1␤-3␤0␤»
19:57 Zaydon joined
TimToady and a new game is born, spelling things with uninames 19:57
timotimo :3 19:58
vendethiel tadzik: well, having a mount fuji character isn't really useful though :o)
19:58 Guest94499 joined
Guest94499 تحذير 19:58
warning you may be watched
do usa & israel use the internet ( youtube, facebook, twitter, chat rooms,skype,ect...)to collect informations,,,,can we call that spying ?
do they record&analyse everything you type????
why they ask for ur name,age,gender,occupation,place of work,ur mobile number.............ect, can they use these informations harming you?!!
warning you may be watched
do usa & israel use the internet ( youtube, facebook, twitter, chat rooms,skype,ect...)to collect informations,,,,can we call that spying ?
do they record&analyse everything you type????
why they ask for ur name,age,gender,occupation,place of work,ur mobile number.............ect, can they use these informations harming you?!!
warning you may be watched
do usa & israel use the internet ( youtube, facebook, twitter, chat rooms,skype,ect...)to collect informations,,,,can we call that spying ?
do they record&analyse everything you type????
why they ask for ur name,age,gender,occupation,place of work,ur mobile number.............ect, can they use these informations harming you?!!
warning you may be watched
19:58 Guest94499 left
FROGGS um 19:58
vendethiel well.
FROGGS that escalated quickly
19:58 Guest94499 joined 19:59 ChanServ sets mode: +o moritz
FROGGS hehe 19:59
19:59 moritz sets mode: +b *!*m@41.153.216.*, Guest94499 was kicked by moritz (Guest94499)), moritz sets mode: -o moritz
vendethiel hehe++ 20:01
moritz: do you know how the configuration option not to auto-op is called ? I was looking for it the other day, couldnt find it 20:02
moritz vendethiel: dunno, not auto-opping is default 20:03
vendethiel moritz: ah, might be default on freenode, I see
geekosaur nickserv defaults to opping you on channel join if you have op access
nickserv "help set noop"
vendethiel geekosaur++ # "set noop", thanks, I'll take note 20:04
geekosaur (well, maybe more recently created accounts default to it turned off; mine certainly did not)
timotimo hmm 20:06
r: use noop;
camelia rakudo-jvm 5d084b: OUTPUT«(timeout)»
..rakudo-parrot 5d084b: OUTPUT«===SORRY!===␤Could not find noop in any of: /home/p6eval/rakudo-inst-1/lib/parrot/6.1.0-devel/languages/perl6/site/lib, /home/p6eval/rakudo-inst-1/lib/parrot/6.1.0-devel/languages/perl6/vendor/lib, /home/p6eval/rakudo-inst-1/lib/parrot/6.1.0-devel/lang…»
..rakudo-moar 5d084b: OUTPUT«===SORRY!===␤Could not find noop in any of: /home/p6eval/rakudo-inst-1/languages/perl6/site/lib, /home/p6eval/rakudo-inst-1/languages/perl6/vendor/lib, /home/p6eval/rakudo-inst-1/languages/perl6/lib, /home/p6eval/.perl6/2014.02-68-g5d084b8/lib␤»
20:06 tchoury left 20:07 telex left
TimToady m: say "\e•≬⧧🌀⛓"».&uniname 20:08
camelia rakudo-moar 5d084b: OUTPUT«ESCAPE␤»
20:08 telex joined
TimToady m: say "\e•≬⧧🌀⛓"».ords».&uniname 20:08
camelia rakudo-moar 5d084b: OUTPUT«ESCAPE BULLET BETWEEN THERMODYNAMIC CYCLONE CHAINS␤»
FROGGS hehe 20:09
vendethiel say "•≬⧧🌀⛓"».ords».&uniname
m: say "•≬⧧🌀⛓"».ords».&uniname
camelia rakudo-moar 5d084b: OUTPUT«BULLET BETWEEN THERMODYNAMIC CYCLONE CHAINS␤»
timotimo m)
vendethiel TimToady++ 20:10
20:10 kurahaupo_mobile joined
vendethiel TimToady: idea : what unicode poetry 20:11
FROGGS TimToady: I am going to bisect moarvm now, seems to be between 2014.02 and HEAD
timotimo unihaikode
vendethiel timotimo++
lue FROGGS: what issue are you looking for?
TimToady m: say "⊨🍄👸👘"».ords».&uniname 20:12
camelia rakudo-moar 5d084b: OUTPUT«TRUE MUSHROOM PRINCESS KIMONO␤»
FROGGS lue: this: nqp-m -e '/\»/'
Unspace not allowed in regex at line 2, near "/"
lue ah, OK
TimToady but apparently only on 32-bit machines 20:14
m: say "\cX💣\cG"».ords».&uniname 20:16
camelia rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/25INwcMLAZ␤'X' is not a valid number␤at /tmp/25INwcMLAZ:1␤------> say "\cX⏏💣\cG"».ords».&uniname␤»
tadzik vendethiel: well, I think it's more useful than FRONT FACING BABY CHICK 20:18
FROGGS 26e596f1f0127e3ab3b6c694f922e4df602761e8 is the first bad commit 20:19
only fetch char once for built-in cclasses
TimToady m: say "\x18💣\x7"».ords».&uniname
camelia rakudo-moar 5d084b: OUTPUT«CANCEL BOMB BELL␤»
TimToady hmm, so it's my fault after all :) 20:20
tadzik I remember a channel where people were kicked if they joined and didn't say Hi =>
or a Hi equivalent
FROGGS TimToady: of course, you were the only active hacker :o)
vendethiel tadzik: and the hardcore version of this channel had r9k bot activated? 20:21
tadzik I'm not familiar with r9k bot
vendethiel tadzik: you're only allowed to say a sentence one time. I believe it was initiated on xkcd's irc 20:22
tadzik heh
FROGGS p: say '»' < '~'
camelia rakudo-parrot 5d084b: OUTPUT«Cannot call 'Real'; none of these signatures match:␤:(Mu:U \v: *%_)␤ in any at gen/parrot/BOOTSTRAP.nqp:1219␤ in any at gen/parrot/BOOTSTRAP.nqp:1210␤ in method Real at gen/parrot/CORE.setting:1032␤ in method Real at gen/parrot/CORE.setti…»
tadzik I have a weechat plugin that does that in a hardcore way
FROGGS p: say '»'.ord < '~'.ord
camelia rakudo-parrot 5d084b: OUTPUT«False␤»
tadzik you know those people who press return instead of every other space?
20:22 xenoterracide joined
masak who would 20:24
ever do
such a
thing?
tadzik way too many people :) 20:26
gist.github.com/tadzik/9375874, if you ever feel a need for such thing :P
vendethiel tadzik: I know I do that :( 20:31
FROGGS m: say '~'.ord 20:33
camelia rakudo-moar 5d084b: OUTPUT«126␤»
FROGGS m: say '»'.ord
camelia rakudo-moar 5d084b: OUTPUT«187␤»
FROGGS TimToady: cp in your patch when checking for WHITESPACE is never 187 20:34
TimToady interesting 20:36
so the problem is uphill and not downhill from there
tadzik vendethiel: now you have a plugin to do that for you :D 20:37
but you have to write it all in one line in order to use it, so it's kind of a therapy :P 20:38
FROGGS it almost looks like as if it checks the \ instead of the »
vendethiel tadzik: I don't have perl5 installed
TimToady the old MVM_string_offset_has_unicode_property_value used MVM_string_get_codepoint_at_unchecked, but this code uses MVM_string_get_codepoint_at
could that make a difference?
vendethiel (and I don't use weechat :p)
FROGGS I am going to check 20:39
the string is /\»/, and the offset is 1 btw
TimToady er, nocheck rather
that sounds right
FROGGS but should I see the backslash in a C string? 20:40
TimToady the only difference bounds checking, I think
why wouldn't you see a backslash? 20:41
tadzik heh 20:43
FROGGS hmmm
PerlJam FROGGS: what value does cp have if not 187? Is it 92? 20:44
FROGGS the last one before the exception is 92, yes 20:45
PerlJam an offset of 0 should give 92 (that's the \ character) 20:46
20:46 hoverboard is now known as lovetsunami 20:47 bluescreen10 left
FROGGS the string has the slashes of the regex too 20:47
so offset 1 would be the backslash
PerlJam oh! I thought those were meta purely for quoting. 20:48
20:48 lovetsunami is now known as hoverboard
PerlJam then somehow the offset if borked? 20:48
s/if/is/
FROGGS no, it needs to check the backslash too perhaps
TimToady well, maybe just a skip-ws test in the regex compiler, but if so, what tests for \<ws>? 20:49
FROGGS token backslash:sym<unsp> { 20:50
[\s|'#'] {}
TimToady I think that's later. you're still just looking for non-significant ws in the regex to find the \, I think
after all, / \» / is also valid 20:51
unless the ordinary ws skipper is calling unsp
do you know that we've actually entered the unsp rule? 20:52
FROGGS it shows up in the backtrace 20:55
but this might even be because we did not run the correct rule 20:56
20:57 kaare__ joined 20:59 Zaydon left 21:03 dayangkun left
rurban_ parrot's Configure.PL --optimize is now recommended again. I just fixed the outstanding PPC floating point optimizer bug 21:05
FROGGS that is still turned on by default 21:10
21:16 dayangkun joined
FROGGS TimToady: the old check MVM_unicode_codepoint_has_property_value(tc, cp, MVM_UNICODE_PROPERTY_WHITE_SPACE, 1) returned true for cp 92 21:18
no, wait 21:22
okay, these two things which are meant to do the same do infact different things: 21:25
MVM_unicode_codepoint_has_property_value(tc, cp, MVM_UNICODE_PROPERTY_WHITE_SPACE, 1),
MVM_string_offset_has_unicode_property_value(tc, s, offset, MVM_UNICODE_PROPERTY_WHITE_SPACE, 1)
masak 'night, #perl6
FROGGS gnight masak
TimToady but the latter is just a wrapper around the former, and doesn't appear to do anything weird 21:29
the parameter types look consistent to me 21:30
FROGGS cries 21:31
why does a printf statement hide the error?
TimToady threading?
sensitive dependence on GC status? 21:32
jnthn If you're in master, not moar-conc, there's certainly only gonna be one thread...
TimToady it's just master
FROGGS okay, casting the return value of MVM_string_offset_has_unicode_property_value to int does unhide the error again
21:33 panchiniak joined 21:34 stevan_ left, bluescreen10 joined
TimToady aha, looks like MVM_unicode_codepoint_has_property_value is missing from a .h file 21:34
so maybe defaulting to int?
FROGGS ohh 21:35
TimToady which is different on a 32
FROGGS yeah, I'm going to test that
TimToady patch posted 21:37
FROGGS that seems to do it! 21:38
TimToady++ # \o/
dalek p: 0a80816 | larry++ | tools/build/MOAR_REVISION:
bump moar
 21:39
kudo/nom: f492fb8 | larry++ | tools/build/NQP_REVISION:
bump nqp
 21:40
FROGGS what a blunder... I am never going to be a C pro
TimToady yes, well, we see how small seemingly dwimmy decisions at the start (let's just default the return type to int) can come back to bite you over and over 21:41
that's why we're so strict about return types in Perl...oh wait...
FROGGS and that is where I love Perl :o)
bah :P
TimToady protip: when adding a function to a .c, add it to the .h too, unless it's static 21:42
FROGGS well yes, I usually did that
TimToady wonders if he's made that blunder elsewhere in the last week
FROGGS and I remember that we already had this issue half a year ago 21:43
sad that my brane didn't tell me
TimToady actually, I know I've made that mistake elsewhere, because it caused issues until I rectified it, but I guess I misse3d one
FROGGS 70.9s for stage parse of a 32bit in a virtual box is quite nice :o) 21:44
perl6-m installed
21:45 Rotwang left
FROGGS on a* 21:45
21:46 aindilis joined
timotimo why aren't we running our C stuff with a warning for that? 21:46
TimToady grondilu: we think we may have fixed your issue
FROGGS also pippo and lue I think
timotimo don't we have -Wall, mister Wall? 21:47
FROGGS *g* 21:48
we'd need a +Wall though
timotimo Wall all the things! 21:49
21:49 daniel-s_ joined
PerlJam reads scrollback 21:49
21:50 xenoterracide left
FROGGS nah, just pull and make :o) 21:50
PerlJam I'm on it! :)
C's little foibles have caused so much grief over the years. It's a wonder people still use it! ;) 21:51
TimToady m: say "\x18🚑℟"».ords».&uniname
camelia rakudo-moar 5d084b: OUTPUT«CANCEL AMBULANCE RESPONSE␤»
TimToady n: say "\cX".ord 21:52
camelia niecza v24-109-g48a8de3: OUTPUT«24␤»
TimToady r: say "\cX".ord
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤'X' is not a valid number␤at /tmp/tmpfile:1␤------> say "\cX⏏".ord␤»
TimToady rakudobug
[Coke] TimToady: I just saw a spec test that was specifically testing to insure that what followed \c was a number. 21:53
timotimo that awkward moment when you look at an X and wonder if it's just an uppercase X or some strange unicode glyph
21:54 hoverboard is now known as stankyfunk
lue [Coke]: but, what about one-word unicode names? Sounds like a weird test to me. 21:54
timotimo OH!
21:54 BenGoldberg joined
timotimo nodyn! 21:54
you can run node.js applications on JVM 21:55
[Coke] digs for it, but doesn't see it, and chalks it up to this cold.
timotimo so we can get rakudo onto the JVM in two exciting ways!
FROGGS hrhr 21:56
err
hehe
[Coke] # RT #83952
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=83952
[Coke] there we go. feel free to fix that test if needed.
nqp: "\cla"
camelia nqp-moarvm: OUTPUT«Unrecognized \c character at line 2, near "la\""␤ at gen/moar/stage2/NQPHLL.nqp:369 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPHLL.moarvm:panic:120)␤ from <unknown>:1 (/home/p6eval/rakudo-inst-2/languages/nqp/lib/NQPHLL.moarvm:charspec:366)␤ from <…»
..nqp-jvm: OUTPUT«Unrecognized \c character at line 2, near "la\""␤ in panic (gen/jvm/stage2/NQPHLL.nqp:378)␤ in charspec (gen/jvm/stage2/NQPHLL.nqp)␤ in quote_escape:sym<chr> (gen/jvm/stage2/NQPHLL.nqp)␤ in !protoregex (gen/jvm/stage2/QRegex.nqp:784)␤ in quote_escape (…»
..nqp-parrot: OUTPUT«Unrecognized \c character at line 2, near "la\""␤current instr.: 'panic' pc 15934 (gen/parrot/stage2/NQPHLL.pir:5941) (gen/parrot/stage2/NQPHLL.nqp:425)␤»
[Coke] p6: "\cla" 21:57
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized \c character␤at /tmp/tmpfile:1␤------> "\c⏏la"␤ expecting any of:␤ statement list␤ pre…»
..niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Unrecognized \c character at /tmp/tmpfile line 1:␤------> "\c⏏la"␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1502 (die @ 5) ␤ at /home/p…»
lue I don't even know:
===SORRY!===
Error while compiling op if (source text: "$backlinks {\n my $didReplace; # save time by not rewriting unchanged files\n my $curf..."): Iteration past end of iterator
(running the command perl6-m -Ilib syngen.p6 --synopses=../specs --outdir=html --backlinks=../roast in my synopsis generator) 21:58
timotimo somewhere in the code for "if" in the qast compiler there must be some iteration code that b0rks 21:59
maybe the optimizer is destroying the qast tree?
lue -j and -p work, so it's moar specific. 22:01
perl6-m --optimize=0 didn't help :(
dalek p: 65cf12e | larry++ | tools/build/MOAR_REVISION:
bump moar some moar, sigh
 22:02
FROGGS lue: try --optimize=off 22:03
dalek kudo/nom: b6fed1a | larry++ | tools/build/NQP_REVISION:
bump, bump, bumpity, bump
 22:04
FROGGS timotimo: that makes a different, right?
lue FROGGS: nope :/
FROGGS k :o(
timotimo -0 should also make a difference, but --off will make an even bigger difference 22:05
FROGGS that wants golfing down then I fear
timotimo if it doesn't change anything, it means it's not my fault :)
lue FROGGS, timotimo: gist.github.com/lue/2d85ada3a258d22669c0
timotimo that also means i've contributed hardly anything besides optimizations, and that makes me a bit sad :|
PerlJam timotimo: why sad?!? That's *awesome* :)
FROGGS lue: is $backlinks an array? 22:06
lue FROGGS: no, it's a commandline option
FROGGS ahh, I see
timotimo PerlJam: it's not exactly bringing perl 6 much closer to "production readiness" so far ;) 22:07
lue (I have no clue how I would golf this, and bisecting would be slooow (unless I could cheat and use -I~/.perl6/lib instead of reinstalling modules with panda all the time, then it could possibly be a bit faster) 22:08
PerlJam timotimo: of course ... having inavertently become the "optimization pumpking" for Perl 6, you'll then be the blame receptacle for why it's so slow.
FROGGS lue: just comment out stuff until the problem vanishes
lue it dies on a simple "if $backlinks" statment, I don't know what I would comment out aside from the conditional :/ 22:09
timotimo hahaha
lue I commented out the if conditional, and I get a less descriptive error out of it: Error while compiling op call: Iteration past end of iterator
FROGGS lue: well, then provide your few lines in a gist or so
timotimo someone really needs to push me so that i finish the block inlining and cascading lexical-to-local transformation finally 22:10
FROGGS hmmm
timotimo i think all i need to succeed in that is a rubber duckie :)
FROGGS timotimo: I wanna see that done until the weekend! ò.ó 22:11
lue Commenting out the for statement works! Which is unfortunate, 'cos that's kinda necessary.
FROGGS lue: can you paste your code?
lue here's the offending loop: github.com/lue/synopsis-generator/...en.p6#L137
timotimo FROGGS: pff. i don't need pressure, i need assistance ;)
FROGGS timotimo: damn :o)
22:11 skids left
FROGGS lue: it still explodes when you comment out the stuff inside the loop? 22:12
lue FROGGS: no, when I comment out the entire loop (with the help of embedded comments), it runs. Unfortunately that's not an actual solution. 22:13
FROGGS lue: no, I mean just to comment out the content of the loop
lue Commenting out the stuff in the loop (making for an empty loop) works, so something in there is breaking. 22:14
FROGGS k
that is what I wanted to hear
lue O, I think I can guess where the break is happening. 22:15
FROGGS r: my $foobar = ''; $foobar.=subst(/"<p>"\x10FFFB(<-[􏿼]>+)\x10FFFC(<-[􏿽]>+)\x10FFFD"</p>"/, {q:c:to/EOBL/}, :g);␤x␤EOBL␤
camelia ( no output )
..rakudo-moar 5d084b: OUTPUT«===SORRY!===␤Error while compiling op call: Iteration past end of iterator␤»
lue yep, github.com/lue/synopsis-generator/...en.p6#L142 22:16
r: say "\x10FFFC" ~~ /􏿼/
camelia rakudo-jvm 5d084b: OUTPUT«Unhandled exception: java.nio.charset.MalformedInputException: Input length = 1␤ in print_exception (gen/jvm/CORE.setting:11097)␤ in <anon> (gen/jvm/CORE.setting:11139)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1230)␤ in command_eval (src/Perl…»
..rakudo-parrot 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized regex metacharacter 􏿼 (must be quoted to match literally)␤at /tmp/tmpfile:1␤------> say "\x10FFFC" ~~ /⏏􏿼/␤ expecting any of:␤ p…»
..rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Null regex not allowed␤at /tmp/tmpfile:1␤------> say "\x10FFFC" ~~ /􏿼/⏏<EOL>␤»
FROGGS r: my $foobar = ''; $foobar = $foobar.subst(/"<p>"\x10FFFB(<-[􏿼]>+)\x10FFFC(<-[􏿽]>+)\x10FFFD"</p>"/, {q:c:to/EOBL/}, :g);␤x␤EOBL␤
camelia ( no output )
..rakudo-moar 5d084b: OUTPUT«===SORRY!===␤Error while compiling op call: Iteration past end of iterator␤»
lue r: say "\x10FFFC" ~~ /\x10FFFC/ 22:17
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b: OUTPUT«｢􏿼｣␤␤»
lue r: say "\x10FFFC" ~~ /\􏿼/
camelia rakudo-jvm 5d084b: OUTPUT«Unhandled exception: java.nio.charset.MalformedInputException: Input length = 1␤ in print_exception (gen/jvm/CORE.setting:11097)␤ in <anon> (gen/jvm/CORE.setting:11139)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1230)␤ in command_eval (src/Perl…»
..rakudo-parrot 5d084b: OUTPUT«｢􏿼｣␤␤»
..rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤No unspace allowed in regex; if you meant to match the literal character, please enclose in single quotes ('􏿼') or use a backslashed form like \x10fffc␤at /tmp/tmpfile:1␤------> [3…»
FROGGS m: my $foobar = ''; $foobar = $foobar.subst(/a/, {q:c:to/EOBL/}, :g);␤x␤EOBL␤
camelia ( no output )
FROGGS m: my $foobar = ''; $foobar = $foobar.subst(/"<p>"\x10FFFB(<-[􏿼]>+)\x10FFFC(<-[􏿽]>+)\x10FFFD"</p>"/, '', :g);␤
camelia rakudo-moar 5d084b: OUTPUT«===SORRY!===␤Error while compiling op call: Iteration past end of iterator␤»
lue FROGGS: it's the line above that that causes the failure, so it's to do with weird characters in the regex.
r: say "A" ~~ /<-[􏿼]>/ 22:18
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b: OUTPUT«｢A｣␤␤»
..rakudo-moar 5d084b: OUTPUT«===SORRY!===␤Error while compiling op call: Iteration past end of iterator␤»
FROGGS lue++
22:18 rindolf left
FROGGS r: say "A" ~~ /<![􏿼]>/ 22:18
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b: OUTPUT«｢｣␤␤»
..rakudo-moar 5d084b: OUTPUT«===SORRY!===␤Error while compiling op call: Iteration past end of iterator␤»
lue r: say "􏿼" ~~ /\􏿼/ # this breaks on JVM and Moar too
camelia rakudo-jvm 5d084b: OUTPUT«Unhandled exception: java.nio.charset.MalformedInputException: Input length = 1␤ in print_exception (gen/jvm/CORE.setting:11097)␤ in <anon> (gen/jvm/CORE.setting:11139)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1230)␤ in command_eval (src/Perl…»
..rakudo-parrot 5d084b: OUTPUT«｢􏿼｣␤␤»
..rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤No unspace allowed in regex; if you meant to match the literal character, please enclose in single quotes ('􏿼') or use a backslashed form like \x10fffc␤at /tmp/tmpfile:1␤------> [3…»
PerlJam wonders if that rules out TimToady breakage ;) 22:19
lue JVM is probably the fault of the JVM, but moar is just being weird there. parrot++
FROGGS r: say "􏿼" ~~ /<!before 􏿼>/ 22:20
camelia rakudo-jvm 5d084b: OUTPUT«Unhandled exception: java.nio.charset.MalformedInputException: Input length = 1␤ in print_exception (gen/jvm/CORE.setting:11097)␤ in <anon> (gen/jvm/CORE.setting:11139)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1230)␤ in command_eval (src/Perl…»
..rakudo-moar 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Null regex not allowed␤at /tmp/tmpfile:1␤------> say "􏿼" ~~ /<!before 􏿼⏏>/␤ expecting any of:␤ postfix␤ infix stopper␤ infix …»
..rakudo-parrot 5d084b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized regex metacharacter 􏿼 (must be quoted to match literally)␤at /tmp/tmpfile:1␤------> say "􏿼" ~~ /<!before ⏏􏿼>/␤ expecting any of:␤ …»
22:20 ruthless left
FROGGS r: say "􏿼" ~~ /<!before '􏿼'>/ 22:20
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b: OUTPUT«｢｣␤␤»
22:21 tgt left, stankyfunk is now known as hoverboard
lue I wonder why moar thinks 􏿼 is a space char, and thus can't be unspaced in regexes. 22:21
22:22 bluescreen10 left
TimToady say "\x10fffc".&uniprop 22:23
m: say "\x10fffc".&uniprop
camelia rakudo-moar 5d084b: OUTPUT«Lo␤»
lue r: say "" ~~ /<-[]>/
TimToady odd, it said Ll here
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b: OUTPUT«Nil␤»
lue m: say "".&uniprop
camelia rakudo-moar 5d084b: OUTPUT«Co␤»
22:26 pippo joined
pippo Hello #perl6 22:26
FROGGS hi pippo
pippo Hi FROGGS
My perl6-m is back! Timtoady++ for solving the issue on 32 bit machines !! 22:27
:-))
FROGGS \o/ 22:28
yay!!
22:28 kivutar joined
timotimo anyone want to continue work on my perl 6 code to turn "rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b" into a much shorter version? 22:28
i found it on my hard drive :)
lue Rakudo's being kinda weird with PUA characters I find :) 22:29
timotimo gist.github.com/timo/c2119931370ee31360e7
22:31 panchiniak left
lue oh, nevermind. *I'm* being weird with interpreting what I typed into camelia :P 22:31
r: say "" ~~ /\/ 22:32
camelia rakudo-parrot 5d084b, rakudo-jvm 5d084b, rakudo-moar 5d084b: OUTPUT«｢｣␤␤»
lue r: say "𰀀" ~~ /\𰀀/
camelia rakudo-jvm 5d084b: OUTPUT«Unhandled exception: java.nio.charset.MalformedInputException: Input length = 1␤ in print_exception (gen/jvm/CORE.setting:11097)␤ in <anon> (gen/jvm/CORE.setting:11139)␤ in command_eval (gen/jvm/stage2/NQPHLL.nqp:1230)␤ in command_eval (src/Perl…»
..rakudo-parrot 5d084b, rakudo-moar 5d084b: OUTPUT«｢𰀀｣␤␤»
timotimo (now the gist even has syntax highlighting) 22:33
BenGoldberg .u no-break
yoleaux U+0083 NO BREAK HERE [Cc] (<control>)
U+00A0 NO-BREAK SPACE [Zs] ( )
U+2011 NON-BREAKING HYPHEN [Pd] (‑)
lue So yeah, rakudo needs some help with the more unusual codepoints :) [esp. JVM, which doesn't like !BMP]
timotimo does that code seem clear to anybody? :P 22:35
22:37 treehug8_ left
PerlJam timotimo: seems overly complicated. 22:38
timotimo lue, like mysql? :P 22:41
PerlJam: there is probably a nicer algorithm and probably even nicer spellings of what i wanted to do
22:41 ajr_ left
timotimo feel free to improve upon it 22:41
(and then port it to perl 5)
lue timotimo: yeah. (Fun fact: when setting up my personal mediawiki for testing stuff, I specifically went for the option that went past the BMP, just on principle :P)
22:47 dmol left
TimToady the problem with classifying all the properties by type is that UNIDATA doesn't seem to carry property type info such as is listed in wp 22:58
so I think we might have to make our own list 23:00
timotimo i'll go make my own list!
with hooks
and, er, properties
actually, forget the hooks! 23:01
lue TimToady: it's in UAX#44 for some of them at least.
TimToady the current getprop has some memory of what worked last time (str, int), but nothing official
question is whether rakudo, nqp, or moar is the right level to install such a list 23:02
the other vms might not make this info available readily 23:03
so probably rakudo or nqp level
23:04 stevan_ joined
lue TimToady: I'd go with the nqp level, it's as close as we can get without entering the vms themselves. 23:04
pippo good night #perl6 23:05
23:05 pippo left
TimToady hmm, does that mean we should really be hoisting some or all of moar's unicode lookups into nqp too? 23:05
23:06 Psyche^_ joined
TimToady will nap on it & 23:07
lue The JVM demonstrates that NQP has to do the unicodeing for some things, at least.
ashleydev qg/mark
23:10 Psyche^ left
lue For some weird reason, this script I made that tests how every unicode character functions inside regexes is slowing eating up memory :) 23:13
23:19 stevan_ left 23:20 hoverboard is now known as adabsurdum 23:26 virtualsue joined 23:31 btyler left
timotimo nap on nqp 23:34
23:43 virtualsue left, skids joined 23:44 stevan_ joined 23:45 hinachan joined
hinachan I gess I see some familiar nicks in here.. 23:46
23:52 Celelibi joined 23:59 adabsurdum left

Powered by Cro and the Raku® Programming Language.

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

Thank you!