»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:05 cognominal left 00:06 cognominal joined 00:18 kaare_ left
lue rakudo: class Present { has $!item; method look() { return $!item; };}; say Present.new(:item("sock")).look; 00:27
p6eval rakudo eb6c4b: OUTPUT«Any()␤»
lue I wonder what I'm doing wrong there... (is it not setting $!item like I think?) 00:28
tadzik default .new doesn't set private attrs
lue ah, ok. 00:29
tadzik that's new in nom
lue bless isn't cooperating. I think I'll just go to a pubic $.item (it's just part of an example for something entirely unrelated anyway) 00:36
Util tadzik: Unsure what you mean. the `#` divider is below the two Perl 5 lines, and above the two Perl 6 code lines and one Perl 6 comment line. 00:43
tadzik Util: oh, okay. I must have got some old version in RSS 00:44
I have # before ( $x, $y ) .= reverse;
Util tadzik: yes, I corrected it from a comment from japhb++ 00:45
For great freshness: perl6advent.wordpress.com/2011/12/2...ic-perl-6/
tadzik alrighty 00:47
01:00 drbean left 01:06 buubot_backup left, drbean joined
lue Would describing |$ as something that means "all possible arguments" in the context of proto sub something(|$){*} be correct? 01:08
01:11 ksi left 01:18 docdurdee joined
docdurdee perl6: say 3; 01:19
p6eval pugs b927740, rakudo eb6c4b, niecza v12-153-g7139c15: OUTPUT«3␤»
01:22 shinobicl_ joined 01:27 buubot_backup joined 01:33 cognominal_ joined 01:37 cognominal left 01:42 docdurdee left
lue There's a preview of my advent post available, if anyone's logged in and willing to look it over. 01:51
01:52 benabik joined 02:08 wolfman2000 joined 02:12 cooper joined
Util lue: Your advent post looks great, accurate and informative 02:16
lue++ 02:17
lue thank you, just published: perl6advent.wordpress.com/2011/12/...ulti-ples/ 02:18
[ that last bit in the post would'nt've happened had I not learned of the official humorous answer from your post, Util :) ] 02:19
s/\'nt/n\'t/ 02:20
02:20 leprevost joined
Util :) 02:24
02:25 shinobicl_ left
lue turns to ponder how long it would take to re-implement ☃☄ for the holidays 02:28
02:30 stepnem left, Chillance left
sorear interesting; of the test files niecza currently does not run at all, the longest is 468 lines 02:31
all the 500+ line test files are being run now
colomon By plan, the longest niecza doesn't run is 200 tests. 02:32
there are only 12 files it doesn't run with 100 or more tests
sorear yeah, well, there was a movement a while back to eliminate plans :[
colomon and most of those are S05-modifier/perl5_*.t 02:33
S02-literals/listquote-whitespace.t # look further 02:34
02:34 stepnem joined
colomon array-shapes.t 02:34
int-unit.t 02:35
S19-command-line-options 02:36
okay, I just eyeball scanned all the plans, and those are the only funny guys that I don't know we're handling 02:37
I only saw one plan * left in the entire list
sorear plan * is not longer valid 02:39
the way to do no_plan now is, just have done; at the end
a test file with done; does not need a plan
sorear could start on the release changelog, that might take a while ;) 02:40
colomon you're kidding me. that's like the worst of all possible worlds?
s/\?/!/
sorear no. 02:41
colomon ...actually, I guess I can easily tweak the LHF finder to look for that. 02:42
but I have to disinfect some eBay christmas toys first. 02:43
benabik … disinfect? 02:45
colomon lysol disinfecting wipes, applied to higglytown heroes 02:46
benabik Oh. Yes. eBay + Children = disinfect plz
geekosaur 's immediate reaction to that was a thought that some "toys" one simply should not get via eBay
02:47 hundskatt left
lue I can't actually augment Perl6::Grammar in rakudo yet, can I? (Not even use MONKEY_TYPING worked) 02:49
sorear Perl6::Grammar uses old regex, anything you write will use QRegex, interop is not possible 02:50
so no, even if you could add a method, modifying the grammar per se is impossibvle
colomon The thing is, the higglytown heroes show went off the air a few years ago, so no more toys are being made. my son got addicted watching DVDs from the library. thus ebay hunts to find toys for him 02:51
okay, looks like there are about 40 files with no plan that we're not running atm 02:52
I can get to work on them, see what we can handle 02:55
niecza: constant foo = 42; say foo; 03:04
p6eval niecza v12-153-g7139c15: OUTPUT«42␤»
colomon niecza: constant foo = 42; say foo == 42
p6eval niecza v12-153-g7139c15: OUTPUT«Bool::True␤»
colomon that's not passing in constant.t .... :\ 03:06
dalek ast: 9548c68 | (Solomon Foster)++ | S04-declarations/constant.t:
Fudge for niecza.
 03:18
colomon sorear: when you get a chance, look at the first couple of todo tests in that one ^^
03:19 cognominal_ left 03:20 cognominal joined
sorear niecza: { constant foo = 42; say foo == 42 ] 03:24
niecza: { constant foo = 42; say foo == 42 }
p6eval niecza v12-153-g7139c15: OUTPUT«===SORRY!===␤␤Unable to parse block at /tmp/MlMSDksbaH line 1:␤------> {⏏ constant foo = 42; say foo == 42 ]␤Couldn't find final '}'; gave up at /tmp/MlMSDksbaH line 1:␤------> { constant foo = 42; say foo == 42…
niecza v12-153-g7139c15: OUTPUT«Bool::True␤»
sorear colomon: was that the error you got?
wait, nvm
colomon constant foo = 42;
ok foo == 42 # this comes out false
sorear niecza: use Test; { constant foo = 42; ok foo == 42, "bar" } 03:25
p6eval niecza v12-153-g7139c15: OUTPUT«ok 1 - bar␤»
sorear niecza: use Test; { constant foo = 42; ok foo == 42, "bar"; { foo = 3 } }
p6eval niecza v12-153-g7139c15: OUTPUT«ok 1 - bar␤Unhandled exception: assigning to readonly value␤ at /tmp/s4Zs2kw9sT line 1 (ANON @ 2) ␤ at /tmp/s4Zs2kw9sT line 1 (mainline @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 2498 (ANON @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line …
colomon sorear: you are seeing it at your end, too, right? 03:28
sorear haven't tried
colomon it's so weird, almost all the harder tests pass, but that first basic test fails 03:29
03:40 Patterner left 03:42 Psyche^ joined, Psyche^ is now known as Patterner
dalek ecza: e0216c6 | sorear++ | src/CClass.pm6:
Fiddle \w and \d to match the UTS18 definitions
 04:04
sorear niecza: { constant foo = 42; say foo; }; { constant foo = 5 } # colomon
p6eval niecza v12-153-g7139c15: OUTPUT«5␤»
sorear colomon: I'm calling the test wrong, since it tries to define two constants GLOBAL::foo (constants are "our" by default). Still, LTA failure mode. 04:06
colomon ooooo... so foo gets 5, eh? 04:07
sorear yes
colomon I'll fix the test
dalek ast: 01a2c6d | (Solomon Foster)++ | S04-declarations/constant.t:
Fix tests (constants are our, so you can't declare the same one twice) and unfudge.
 04:13
ecza: cf4faee | (Solomon Foster)++ | t/spectest.data:
Turn on S04-declarations/constant.t.
 04:14
04:15 benabik_ joined 04:16 benabik left, benabik_ is now known as benabik 04:17 leprevost left, leprevost joined 04:19 thou left
colomon sorear++ 04:20
dalek ast: 20ccb4a | (Solomon Foster)++ | S06-advanced/callsame.t:
Fudged for niecza.
 04:23
04:24 aindilis left 04:28 bkolera joined
colomon sorear: does niecza have "handles" delegation yet? 04:28
04:32 silug_ joined
sorear colomon: no 04:32
04:35 silug left, cognominal left
dalek ecza: b3fda07 | sorear++ | / (5 files):
Fix \w handling, update <<>> to use new \w, add a bunch of predefined rules like <punct>
 04:36
colomon sorear: don't suppose you could get Action method post_constraint running for Christmas? ;)
04:36 cognominal joined
sorear niecza: sub foo($x where * > 0) { } 04:38
p6eval niecza v12-154-ge0216c6: OUTPUT«===SORRY!===␤␤Action method post_constraint not yet implemented at /tmp/rMIFf5Ja3q line 1:␤------> sub foo($x where * > 0⏏) { }␤␤Unhandled exception: Unable to resolve method sorry in class Match␤ at /home/p6eval/niec…
sorear you mean that?
colomon yes 04:39
it keeps coming up in these test files tonight. :)
04:46 drbean left 04:50 cognominal left 04:51 drbean joined 04:52 cognominal joined 04:55 lestrrat left 04:56 lestrrat joined
sorear rakudo: sub foo(Mu $x [ $y ]) { }; foo([ any(1,2,3) ]) 05:09
p6eval rakudo eb6c4b: OUTPUT«(timeout)maximum recursion depth exceeded␤»
sorear thought so
worth filing a bug over?
colomon I'm not sure I understand it? 05:10
rather, I'm sure I don't understand it
sorear I'm reading Rakudo's binder and I noticed a case where the auto-threader could be made to recurse without limit
colomon file a bug, I'd say
sorear++ 05:11
afk # sleep
sorear sleep well
rakudobug@ still?
sent. 05:12
05:24 thou joined 05:25 kaleem joined
moritz Util: in p6, grep takes a smart-matcher 05:25
nom: say (0..10).grep(4|8)
p6eval nom eb6c4b: OUTPUT«4 8␤»
moritz and basically anything can serve as a smart-matcher 05:44
Util moritz: thanks! I will re-read that part of the Spec in the morning. Midnight approacheth... 05:49
05:53 kaleem left
moritz Util++ # advent post 06:01
lue++ # advent post
06:02 bkolera left
moritz nom: say "\x[FFFE]" 06:06
p6eval nom eb6c4b: OUTPUT«===SORRY!===␤Invalid character for UTF-8 encoding␤␤»
dalek ast: afa5414 | moritz++ | S05-mass/properties-general.t:
fudge properties-general.t again for rakudo
 06:24
sorear starts the process of turning sig_i and sig_r into proper objects 06:27
moritz: do you think nom might be right here? 06:32
moritz sorear: I have no idea 06:34
geekosaur I would say it's wrong, but that depends on your conceptual level 06:37
since it's a low-level "character" whose purpose is defining the interpretation of the rest of the character stream
sorear was under the impression that all values from 0 to 0x10FFFF would be legal 06:38
geekosaur it can be argued that its proper place is whatever handles translating between internal and external encodings, and for it to be visible internally is an error
that's the thing. it's *not* operable at that level, it's more like the surrogates used for encoding in utf16 06:39
sorear geekosaur: how many BMP chars are guaranteed to stay Cn forever? 06:40
geekosaur which are not really characters in the 0 - 0x10fff sense, but an artifact of how those characters are encoded
sorear geekosaur: the old code was using "one past the 5.1 CJK UNIFIED IDEOGRAPHS block" as "Cn character". Which...no.
06:42 cognominal_ joined
geekosaur is not familiar with said code and isn't really talking about that level 06:42
IIRC the spec says that fffe is a reserved noncharacter which allows a program reading particular unicode encodings to determine the byte order used in the encoding and thereby its correct interpretation. they are explicitly *not* actual characters, just reserved "slots" 06:44
sorear that was my reasoning
fffe was a reserved slot, so the tests could use it as a noncharacter
06:45 cognominal left 06:46 kaare_ joined
sorear hah, I had forgotten how many places I had duplicated the sig_i parsing logic :/ 06:55
07:04 thou left 07:32 cognominal joined 07:34 cognominal_ left 07:36 bkolera joined
dalek ast: 21b8e66 | moritz++ | S05-mass/properties-script.t:
fudge properties-script.t again for rakudo
 08:21
ecza: acc5e83 | sorear++ | / (2 files):
Factor Niecza::Benchmark out of perf/

Useful for timing one-liners
ecza: 10ba66d | sorear++ | lib/ (6 files):
Turn Signatures into real objects

Instead of the old weird int[] and object[] thing. Actually reduces code! Should have no visible effects; preparation for some signature improvements, like "where", sub-signatures, maybe even signature introspection.
sorear -> sleep
08:23 molaf joined 08:26 dsafjn348 joined 08:37 cognominal left 08:38 cognominal joined 08:43 snearch joined
masak oh! good morning, hushpuppies. :) 08:53
moritz \o 08:54
08:59 pkkm joined
dsafjn348 hi 09:21
kshannon I need a bit of help cleaning up some monkey-see monkey-do copy and paste work I've done to fix Rakudo bug RT#53814. I made a copy of token term:sym<miscbad> from STD.pm6 and merged it into Perl6::Grammar as token nullterm (I tried to just copy it but couldn't get it to work) 09:36
token nullterm { {} <!{ $*QSIGIL }> [ || <?.infixish> <.panic: "Preceding context expects a term, but found infix instead"> || <?> ]
} 09:37
09:37 mj41 joined
kshannon It compiles and passes stresstest (+long) 09:37
It also passes the TODO test in t/spec/S06-signature/optional.t for the bug. 09:38
moritz waits for the "but" 09:39
kshannon Well, I don't really know all the regex stuff yet, so I'm just wondering if anyone can see any style issues with it. 09:40
moritz well, the leading {} is probably not necessary right now, because it's just to limit LTM, and rakudo doesn't do that yet 09:41
but it doesn't hurt to keep it in there either
kshannon: but if it works, and comes from STD, it's fine to include it in rakudo's grammar 09:42
kshannon Well it comes from STD, but as an entirely different rule, essentialy as a fallback term which could never match but would give an error if matching an infixish would work. It used CURSOR.suppose rather than a lookahead. 09:46
I've added it into the nullterm instead which is used by nulltermish to possibly match nothing. The comma operator has the :nexterm<nulltermish> property which is how it allows trailing commas. We inherit token nullterm { <?> } from HLL::Grammar. I think I've just ended up reducing the scope of the infixish lookahead to after infix:<,> which is good enough for the bug but won't give the nice error message which STD (and Niecza) give for any two consecutive 09:53
10:00 dsafjn348 left 10:10 mj41 left
tadzik hello #perl6 10:22
anyone with a minute in handy and a zavolaj commit bit can update its README and possibly Makefile to not run tests? I have cookies 10:23
10:23 nebuchadnezzar left
tadzik bbs 10:23
I can do it in 30-60 as well 10:24
10:25 nebuchadnezzar joined
moritz hugme: add tadzik to zavolaj 10:30
hugme moritz: You need to register with freenode first
moritz hugme: add tadzik to zavolaj
hugme moritz: You need to register with freenode first
10:31 hugme left, hugme joined, ChanServ sets mode: +v hugme
moritz hugme: add tadzik to zavolaj 10:31
hugme moritz: You need to register with freenode first
moritz swears loudly
why did it work yesterday, but not today?
10:39 donri joined 10:57 mucker joined
masak API change somewhere? 10:59
11:02 mucker left
moritz there was no notification from freenode 11:06
11:09 snearch left
tadzik moritz: I'm added :) 11:26
I was just afk, will poke it in a while
11:28 tokuhirom joined
tadzik Incorrect pre-compiled version of <unknown> loaded 11:34
-_-
I'll just leave that to jnthn :) 11:39
11:46 pmurias joined 11:48 abercrombie left 11:49 cognominal_ joined 11:51 mishin joined, mj41 joined 11:52 cognominal left 11:57 fsergot joined 11:58 risou_awy is now known as risou 12:01 MayDaniel joined 12:07 hundskatt joined 12:10 drbean left 12:11 mj41 left
masak lue: "What has to be different between these subs is their signature, or list of arguments." -- arguments are on the caller side, parameters are on the callee side. 12:13
masak fixes
lue: it took me quite a while to get used to that distinction, too ;) 12:14
12:17 drbean joined
masak I fail to see why lue's proto open was needed in the blog post. 12:17
also, if I may grumble a bit, I'm not sure that post ends on the most PR-tactical of notes. 12:18
(we use the 24th post to tell the world "guess [Perl 6] isn't [completed]"? huh.)
12:21 pkkm_ joined 12:23 pkkm left 12:33 mkramer joined, mkramer left
kshannon Ok, refactored to give error message for all consecutive infixes: gist.github.com/1517254 12:45
moritz kshannon: that looks good; would you care to submit that as a pull request on github? 12:48
kshannon The last thing I'd like to do with this is include the offending infix in the error mesaage like the original in STD did (and Niecza does too), but I'm having troubles.
Fails to parse: <?before <infixish> <.panic: "Preceding context expects a term, but found infix " ~ $<infixish>.Str ~ " instead"> > 12:49
If I try to use a code block instead, it fails to find the end of the block :(
moritz kshannon: that's a known limitation in the grammar engine
kshannon What is the limitation? 12:50
moritz that it's hard to include the content of variables in error messages
it's in nqp, in which Perl6::Grammar is written
12:53 fglock joined
pmurias fglock: hi 12:53
fglock pmurias: hi!
I just released perlito #8
moritz fglock++ 12:54
pmurias just saw that perlito now also compiles perl5 to javascript 12:56
fglock yes, the original idea was a perl5 that could be embedded in a perl6 compiler 12:57
but I found it better to write the new perl5 compiler in perl5 12:58
so perl5 people can contribute more easily
also it makes the bootstrap more interesting
it can still be embedded in a perl6 compiler: just add a perl6 backend 12:59
13:03 drbean left 13:04 bluescreen10 joined 13:10 drbean joined
pmurias fglock: it's already bootstraped? 13:10
fglock yes, it is 13:13
13:14 mishin left
fglock it contains some non-perl5 extensions, which I'm cleaning up (these make it easier to write the compiler) 13:15
13:16 am0c joined
pmurias and how are you planning to deal with Perl6 rules being used for parsing? 13:16
add them as a CPAN module to perl5? ;)
fglock something like: *myrule = eval Rule->compile( "..." ) 13:17
as in PCR
Pugs::Compiler::Rule 13:18
pmurias the alternative would be to use Devel::Declare to add the token keyword to real p5 13:21
fglock I think it can be ignored for now, because there is a perl5 backend which "desugars" the extensions 13:23
but all other extensions should be cleaned up 13:24
13:25 bluescreen10 left
fglock re Devel::Declare, I can only support pure-perl5 modules 13:32
masak fglock++ # perlito 13:33
fglock: I've been toying with the idea of compiling some Perl 6 code to one of the many JS web graphics libraries out there (such as Paper.js), and using Perlito for that. haven't attempted it yet, though. 13:34
13:35 pkkm_ is now known as pkkm
fglock masak: I've tried it with titanium - www.appcelerator.com/products/titan...velopment/ 13:36
node.js also works, you only need to tweak the missing runtime stuff 13:37
masak ooh
sounds eminently doable, then.
13:39 Guest91318 left
fglock titanium is particularly difficult because it contains a syntax checker doesn't like automatically generated javascript 13:40
*that* doesn't
masak sounds odd.
or rather, sounds like a not-particularly-good syntax checker, then. 13:41
13:41 bkolera left
moritz somehow precompilation is rather busted 13:44
when I run panda from outside the panda dir, I get "Incorrect pre-compiled version of /home/moritz/p6/panda/ext/JSON/Tiny.pm loaded" 13:45
from inside the panda dir, I get "Method 'parse' not found for invocant of class 'Signature' in block <anon> at /home/moritz/.perl6/bin/panda:12 in sub from-json at lib/JSON/Tiny.pm:50
from-json calls JSON::Tiny::Grammar.parse(...)
so, JSON::Tiny::Grammar ends up as Signature? wtf? 13:46
masak moritz: consider submitting those findings as an RT ticket. 13:50
using Titanium to build nontrivial Perl 6 apps that run on Android or iPhone would be *awesome*. 13:52
13:58 pkkm left
pmurias fglock: what javascript implementation does titanium use underneath? 14:00
fglock hmm - no idea 14:01
this was my test case for titanium: 14:03
nopaste.info/22d897c4fd.html
the Titanium object was hacked in the emitted js code
14:07 tokuhirom left
pmurias fglock: it runs on your mobile phone? 14:07
pmurias only has a symbian phone 14:08
fglock no, this was something nferraz++ was testing 14:11
masak it's something like that I'd like to try out, yes. 14:27
flussence_ omg! github.com/flussence/specgraphs/ra.../impls.png 14:28
tadzik flussence_: nice. Does it get autocommited or what? 14:37
flussence_ nah, I do most of it manually. I'm not all that confident in my ability to write automatic things interacting with remote stuff which don't go horribly wrong :) 14:38
tadzik :) 14:39
perldoc perlsyn | perl -ne 's/^.*?"|\\.*$//g,print if/1\.\.15/' 14:40
odyniec++
masak wow. 14:41
arnsholt Nifty =D 14:42
masak flussence_: that graph represents how I feel about Niecza. quite capable and fast approaching the level of Rakudo. 14:45
oh, that perlsyn thing was less magical than I thought. there's the actual substring "Merry Christmas" in perldoc perlsyn. 14:48
still, creative :)
15:04 am0c left 15:18 x3nU joined
dalek ast: 0dee370 | (Solomon Foster)++ | S32-list/map.t:
Fudge for niecza.
 15:23
ast: 03b1456 | (Solomon Foster)++ | S32-list/grep.t:
Fudge for niecza.
 15:27
15:27 cognominal_ left 15:29 cognominal joined
dalek ecza: 1075f3e | (Solomon Foster)++ | lib/CORE.setting:
Enable Whatever argument to roll, add sub roll.
 15:40
ecza: a1c4b26 | (Solomon Foster)++ | t/spectest.data:
Turn on S32-list's grep.t, map.t, and roll.t.
colomon niecza: say [min] 2, 4, 6, 10 15:54
p6eval niecza v12-158-g10ba66d: OUTPUT«Unhandled exception: Excess arguments to infix:<min>, used 2 of 4 positionals␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (infix:<min> @ 0) ␤ at /home/p6eval/niecza/lib/CORE.setting line 1511 (reduceop @ 8) ␤ at /tmp/zSeumtDVPB line 1 (mainline @ 1…
colomon niecza: say [+] 2, 4, 6, 10 15:59
p6eval niecza v12-158-g10ba66d: OUTPUT«22␤»
16:19 fsergot left 16:27 cognominal left 16:28 cognominal joined
TimToady if we can get away with it, I'd like to see fffe and surrogates be valid at a codepoint level, but not at a grapheme level 16:41
phenny TimToady: 19 Dec 10:52Z <moritz> ask TimToady if he wants to the 24th post for the advent calendar
TimToady: 19 Dec 10:53Z <moritz> tell TimToady erm, missing verb. Insert 'write' where appropriate :-)
TimToady er, no, I don't :)
only just starting to get my brane back after the zombies munched it overmuch... 16:42
flussence_
.oO( and lue++ might object to having his day 24 post replaced :)
.oO( unless it's declared `multi post` )
 16:43
16:54 leprevost left 17:11 JimmyZ joined, mkramer joined, mkramer left
JimmyZ Merry Christmas, #perl6 17:22
17:22 MayDaniel left 17:29 MayDaniel joined 17:32 drbean left 17:38 drbean joined 17:39 Ingis left 17:54 fglock left 18:00 drbean left
masak merry Midwinter Wish-it-were-summer festivities, JimmyZ! 18:03
JimmyZ well, here is summerish. I'm in Shenzhen 18:05
;)
masak :D
JimmyZ: I was in HK a week ago. it was alright, too.
18:06 drbean joined
JimmyZ masak: Happy vacation 18:06
masak yeah, it was a happy vacation. ;)
JimmyZ hadn't been in other countries 18:08
masak it was interesting getting a sense of Chinese culture.
JimmyZ: you consider HK to be another country? :)
JimmyZ HK
masak 香港
JimmyZ I thought USA, Maybe it's too late here 18:09
2:00 AM
masak ah :)
well, US has that kind of two-letter combinations for states, so it's a good guess, I guess.
JimmyZ yeah
HK looks like US to me 18:10
masak there's a HI for Hawaii; that seems to be the closest one. 18:11
[Coke] gist.github.com/1476841 - niecza now at 75.29%. up from 49.67% 10 days ago.
colomon++ sorear++
colomon [Coke]: something's wrong with those numbers
niecza is not getting 1264 failures
JimmyZ I still more like rakudo :) 18:12
colomon [Coke]: I'm pretty sure we actually hit 80% yesterday 18:13
I did a run with 14814 passing tests 18:18
[Coke] colomon: did you commit your fix to test_summary in spec?
colomon yes
[Coke] hurm.
colomon and even without it, we were showing ~300 failures, not ~1200
[Coke] double checks. 18:19
perhaps I'm using my "limited" perl6 for that (restricts memory & time) 18:21
yup, that's part of it, at least. rerunning.
dalek ecza: 34200c9 | coke++ | t/spectest.data:
run S03-smartmatch/any-hash.t
 18:23
flussence_ [Coke]: what time limit are you using on that? 120s seems to work for me, while 30s caused a few thousand test fails 18:24
er, on rakudo
[Coke] #exec "ulimit -t 45; ulimit -v 256072; nice -20 mono ./run/Niecza.exe @ARGV";
I had cut the memory down quite a bit after juerd's recent request. 18:25
(doing a run now with no limits)
flussence_ you probably want -m instead of -v... 18:26
[Coke] flussence_: taken directly from tools/rakudo/*limit* 18:32
er, rakudo's tools/*limit*, that is.
18:34 JimmyZ left
[Coke] gist.github.com/1476841 - niecza now at 82.12%, up from 49.67%, in 10 days. colomon++ for requesting a recount. ;) 18:37
flussence_ oh, I just remembered I need to do some recounting of my own... 18:39
[Coke] nom: say 18376-18222, 15090-9050 18:42
p6eval nom eb6c4b: OUTPUT«1546040␤»
[Coke] nom: say 18376-18222, ":", 15090-9050
p6eval nom eb6c4b: OUTPUT«154:6040␤»
[Coke] (not bad, 6040 more passing tests, that's 604 a day. ;)
sorear good * #perl6 18:43
sorear wonders why we don't see fglock around anymore 18:44
[Coke] sorear: o/
aloha: seen fglock
aloha [Coke]: fglock was last seen in #perl6 4 hours 33 mins ago saying "no, this was something nferraz++ was testing".
18:53 Ingis joined
dalek ecza: e919c7b | coke++ | t/spectest.data:
re-enable S04-phasers/begin.t
 18:53
pmurias sorear: he can't irc from work anymore, so he can be here as often as he used to 19:09
19:10 rsimoes joined
[Coke] niecza: list(4) 19:12
p6eval niecza v12-162-ge919c7b: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'list' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 880 (die @ 2) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1143 (P6.comp_unit @ 32) ␤ at /hom…
[Coke] nom: list(4)
p6eval nom eb6c4b: ( no output )
[Coke] nom: list(4).say
p6eval nom eb6c4b: OUTPUT«4␤»
[Coke] nom: list(4).HAT
p6eval nom eb6c4b: OUTPUT«Method 'HAT' not found for invocant of class 'List'␤ in block <anon> at /tmp/1C9adP8GYz:1␤ in <anon> at /tmp/1C9adP8GYz:1␤»
[Coke] nom: list(4).WHAT
p6eval nom eb6c4b: ( no output )
[Coke] gives up. ;)
flussence_ I just noticed the PHP advent calendar... day 16's a long essay complaining about the language and detailing why it can't be fixed. Ouch. 19:15
19:21 molaf_ joined 19:24 molaf left 19:25 fglock joined
[Coke] niezca: say 33 ~ Failure 19:28
niezca: say 33 ~~ Failure
niecza: say 33 ~~ Failure
p6eval niecza v12-162-ge919c7b: OUTPUT«===SORRY!===␤␤Undeclared name:␤ 'Failure' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 880 (die @ 2) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1143 (P6.comp_unit @ 32) ␤ at /hom… 19:29
fglock sorear: I'm hacking offline - except when it's holidays :)
pmurias fglock: should i seperate Runtime.js into Runtime.js and Runtime/D8.js so we can have Runtime/Node.js? 19:30
fglock I think it's a good idea - but only the i/o subs are different, so maybe keep the generic Runtime.js and add the extra backend-specific ones 19:32
dalek ast: 004612d | coke++ | S32-array/pop.t:
niecza fudge
 19:34
ecza: 4ffb3b3 | coke++ | t/spectest.data:
run S32-array/pop.t
pmurias fglock: yes, that's what i want to do
fglock: Runtime/D8.js would contain the d8 specific io part
sorear pmurias: ah, that makes sense :| 19:37
o/ fglock
19:37 Trashlord left
fglock sorear: hi \o/ 19:37
sorear how does perlito do on spectests? 19:38
fglock sorear: it... doesn't - I haven't tried the spectests for a while 19:39
pmurias fglock: src5/lib/Perlito5/Javascript/Runtime.js is editable?
sorear how close to feature-complete is it?
fglock pmurias: yes, everything in src/ is editable
sorear hah, this is a large repository
pmurias fglock: should i remove the "do not edit warning" from that? 19:40
sorear is trying to figure out what the relationship between niecza and perlito is/should be 19:41
fglock sorear: hmm - let me think about it
pmurias: ah, yes - that message would be copied to the compiled files, but it is no longer needed 19:42
sorear: perlito goal is to have as much as possible perl6 source code, with minimal backend-specific parts 19:46
also as many features as possible implemented as extensions, with a small core 19:47
the small core allows it to be executed in the browser and small devices 19:48
I understand that niecza is also retargetable, right?
up to some point
pmurias fglock: it is retargetable, backends are much more complex then perlito ones, but allow more features 19:52
fglock does niecza embed perl5 for 'use v5'? 19:57
alternately, it could embed a perlito5 19:58
pmurias it doesn't support 'use v5' yet 19:59
[Coke] sorear: where would one define Array's delete() method?
pmurias fglock: i implemented some basic eval(:language<perl5>,...) support using an embedded perl5
fglock: how do we distinguish if perlito5 should emit code for d8 or node? 20:02
20:02 molaf_ left
fglock pmurias: how about -Cjs-node / -Cjs-v8 options in perlito.pl 20:03
perlito5.pl
sorear: I'd like to see more code shared between implementations, but that's really really difficult to achieve 20:04
the architectures are pretty different 20:05
[Coke] is perlito running any spec tests? 20:07
fglock Coke: v6.pm used to pass over 50% of the spec tests, but it was unmaintainable 20:08
perlito now uses a small set of tests,
these tests are features required for bootstrapping 20:09
pmurias fglock: js-node and js-d8? node also uses v8 internally
fglock running the bootstrap is the ultimate test for perlito
pmurias: hmm - the default now is d8; maybe just add js-node 20:10
Coke: it is not optimal, but it keeps things simple 20:11
20:11 drbean left, mkramer joined
pmurias fglock: ok 20:12
pmurias doesn't like node very much 20:13
fglock pmurias: don't do it then :P
it should be fun 20:14
20:14 mkramer1 joined 20:16 pothos left, y3llow left
sorear [Coke]: in the Array class? you mean for @arr.delete($ix) ? 20:16
20:17 mkramer left, drbean joined 20:18 y3llow joined, pothos joined 20:20 y3llow left, y3llow joined 20:21 pothos left, pothos joined 20:22 y3llow left, y3llow joined
[Coke] aye. 20:23
fglock hmm - maybe runtime-loadable modules is the key to keep perlito core simple - it only needs some way to map the missing features to modules 20:24
moritz \o 20:25
moritz wishes everybody of the appropriate time zones a happy christmas eve
fglock moritz: happy christmas there 20:26
perlito5.js is 1.2Mbytes, that's too big already
pmurias fglock: is it possible to sanely split up the compiler into parts? I can't see how that would be done for the parser? 20:32
fglock: both minifiying the js and emitting more consise js should greatly reduce the size of the executable 20:35
s/executable/js file/
fglock: what do we need CallSub for? 20:36
fglock yes, it is possible to split up
ah, that's a js workaround
because in js you can't tell a method from a sub,
if the first arg is a CallSub then you know it's a sub - otherwise, it needs to add the object to @_ 20:37
method-call is the default 20:38
20:39 cognominal_ joined
fglock re split up: consider something like - <rule> { require Stuff } <parse_more> 20:40
20:42 cognominal left
fglock the problem is how to get the "right" granularity - and if the compiler itself requires some feature, then splitting doesn't save anything 20:43
moritz -> sleep again 20:44
pmurias fglock: i think it would make sense for more complex features like Moose support etc. 20:47
fglock yes - but I'm thinking about perlito6, where all features are "core"
[Coke] sorear: also need Array.exists
pmurias fglock: perlito6 and perlito5 will have seperate backends? 20:48
20:49 snearch joined
fglock they have separate backends now, because it is easier to implement the small differences 20:51
keeping everything together can be confusing - but this might change later 20:52
20:53 cognominal_ left 20:55 cognominal joined 20:57 pmurias left 20:59 Trashlord joined 21:00 fridim_ joined 21:09 benabik left 21:12 fglock left 21:13 benabik joined 21:22 fridim_ left
[Coke] niecza: say "3".exp 21:22
p6eval niecza v12-163-g4ffb3b3: OUTPUT«20.085536923187668␤»
[Coke] niecza: say 3.exp
p6eval niecza v12-163-g4ffb3b3: OUTPUT«20.085536923187668␤»
21:24 bluescreen10 joined
[Coke] nom: say 17.roots(4) 21:24
p6eval nom eb6c4b: OUTPUT«2.03054318486893 + 0i 3.6446153367886e-09 + 2.03054318486893i -2.03054318486893 + 7.28923067357719e-09i -1.09338460103658e-08 - 2.03054318486893i␤»
[Coke] nom: say +(17.roots(4)) 21:25
p6eval nom eb6c4b: OUTPUT«4␤»
21:33 MayDaniel left
[Coke] thinks S32-num/cool-num.t has some suspicious "floar" tests. 21:36
dalek ast: badeafc | coke++ | S32-num/cool-num.t:
fix plan
ast: c03f823 | coke++ | S32- (2 files):
niecza fudge
21:36 drbean left
dalek ecza: 8275ae4 | coke++ | t/spectest.data:
run S32-array/shift.t, S32-num/cool-num.t
 21:37
21:43 drbean joined
dalek ast: ee8a33c | coke++ | S32-hash/kv.t:
niecza fudge
 21:45
ecza: e0913f8 | coke++ | t/spectest.data:
run S32-hash/kv.t
21:50 DarthGandalf left 21:54 sudokode left, integral left 21:56 sudokode joined 21:57 jrockway left 21:58 mikemol left, mikemol joined 22:00 DarthGandalf joined, integral joined, integral left, integral joined 22:15 jrockway joined 22:31 mkramer joined, mkramer1 left, MayDaniel joined 22:33 ksi joined 22:40 mkramer left 22:53 PacoLinux_ joined 22:55 mkramer joined 23:00 mkramer left 23:07 kaare_ left 23:08 fridim_ joined 23:12 packetknife joined 23:17 PacoLinux_ left 23:29 mkramer joined 23:31 mkramer left, mkramer1 joined 23:33 Mowah left 23:44 Trashlord left 23:45 ponbiki joined 23:47 MayDaniel left, mkramer1 left 23:50 mkramer joined 23:51 thou joined, mkramer left, mkramer joined 23:56 bluescreen10 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!