»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.
00:14 dayangkun joined 00:17 lawrmcfa left 00:27 cog left 00:28 Chillance left 00:32 grondilu joined 00:33 cog joined 00:35 palakas joined 00:36 hash_table joined 00:39 palakas_ joined 00:41 palakas left 00:59 SHODAN left, SHODAN joined 01:07 anuby joined 01:17 hypolin joined 01:31 SmokeMachine left, SmokeMachine joined 01:36 SmokeMachine left 01:37 atrodo left 01:56 rom1504 joined 02:14 xinming left 02:15 xinming joined 02:39 hash_table left 02:42 Khisanth left
dalek p-jvm-prep: 6ef5e0a | coke++ | t/nqp/ (2 files):
2 more passing test files.
 02:48
[Coke] nwc10: Files=39, Tests=278 02:56
japhb w00t
02:57 Khisanth joined
dalek rl6-roast-data: 9ff7142 | coke++ | / (4 files):
today (automated commit)
 02:59
rl6-roast-data: 1d17226 | coke++ | / (4 files):
today (automated commit)
rl6-bench: e7f2167 | (Geoffrey Broadwell)++ | analyze:
Lower minimum time for comparisons from .001 to .0001 second (necessary for certain tests on fast computer)
 03:12
03:23 cog left 03:33 cog joined 03:37 xinming left, kaleem joined 03:38 xinming joined 03:46 cog left 03:53 cog joined 04:05 benabik joined 04:11 preflex left, preflex_ joined 04:12 preflex_ is now known as preflex 04:20 nebuchadnezzar left, nebuchad` joined 04:25 skids left 04:31 dalek left, dalek joined, ChanServ sets mode: +v dalek 04:33 Targen joined 04:57 cog left 05:00 balders_1od joined 05:02 balders_dod left 05:04 cog joined 05:09 jaldhar joined 05:10 sevin joined 05:17 dayangkun left 05:23 jaldhar left 05:26 jaldhar joined 05:27 quester joined 05:47 Patterner left 05:48 sevin left 05:54 Psyche^ joined, Psyche^ is now known as Patterner 06:17 sevin joined 06:20 dayangkun joined
moritz std: my int 42; 06:48
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Multiple prefix constraints not yet supported at /tmp/YPJPfDgvLA line 1:␤------> my int ⏏42;␤Malformed my at /tmp/YPJPfDgvLA line 1:␤------> my int ⏏42;␤ expecting any of:␤ multi_declarator␤
..type…
06:53 Targen left
diakopter std: \my ::my $my /my %my *my @my; 06:58
p6eval std 7deb9d7: OUTPUT«ok 00:00 43m␤»
07:03 sevin_ joined 07:05 nebuchad` is now known as nebuchadnezzar
diakopter std: \my &my ^my ::my $my /my %my *my @my; 07:05
p6eval std 7deb9d7: OUTPUT«ok 00:00 45m␤»
07:06 FROGGS left 07:07 sevin left 07:38 erkan joined, erkan left, erkan joined 08:35 xenoterracide left 08:40 xenoterracide joined 08:43 kresike joined
kresike hello all you happy perl6 people 08:43
08:46 hoelzro|away is now known as hoelzro
arnsholt o/ 08:46
hoelzro morning Perl6 folk 08:47
08:53 quester left
arnsholt Fun. My native attribute refactor passes NQP and Rakudo's test suites, but it's utterly, utterly broken Zavolaj 08:56
arnsholt investigates
Oh. Dependency bug in NQP's Makefile, sort of 08:59
moritz bug, or not optimized for your use case? :-) 09:03
arnsholt I changed stuff in sixmodelobject.h, but CStruct.c didn't get rebuilt 09:05
Oh, but none of the files probably changed because of sixmodelobject.h 09:06
I changed nqp.ops as well, which probably triggered the main REPRs, but not the dyncall REPRs
09:07 FROGGS joined 09:12 anuby left, anuby_ joined 09:15 spider-mario joined, djanatyn left, djanatyn1 joined 09:16 sqirrel joined 09:20 Khisanth left
dalek p/dyncall-sized-num: eb90ed8 | (Arne Skjærholt)++ | src/6model/reprs/P6int. (2 files):
Revert "Update P6int to properly read sized ints."

This reverts commit 0b7de0c193e25059b3e79d2d67d0f9e01d61ffbf.		 09:23
p/dyncall-sized-num: a7f7a84 | (Arne Skjærholt)++ | src/6model/reprs/P6int.c:
Better sized int handling in P6int.
p/dyncall-sized-num: d6eab55 | (Arne Skjærholt)++ | src/ (6 files):
Better getting/setting of natively typed attributes.

This brings natively typed attributes more in line with the approach used in NQP-JVM. Instead of get_attribute and bind_attribute dealing directly with pointers to the data, we use a NativeValue struct to pass values. This way, the class can use the attribute's REPR to get and set values.
FROGGS ohh, I need to look at that closely
arnsholt++
cxreg hm, wonder how that parrot threads nqp work is going 09:25
looks somewhat idle from github's pov 09:26
arnsholt FROGGS: It's a bit of an intermediate stage, but if jnthn is happy with it, I'll apply the changes to P6num and {bind,at}_pos_ref as well
At which point I think Zavolaj should have pretty decent sized value support 09:27
cxreg arnsholt++ 09:28
FROGGS I'll do some testing these days, even if it might not be the final interface 09:29
09:32 fhelmberger joined 09:34 Khisanth joined 09:36 cibs left, cibs joined 09:43 stevan_ left 09:47 balders_1od left, Patterner left 09:49 Psyche^ joined, Psyche^ is now known as Patterner 10:03 dakkar joined 10:06 hypolin left, daniel-s joined
jnthn o/ 10:08
arnsholt Morning 10:12
jnthn static void * default_get_attribute_native 10:13
arnsholt I've pushed the get_attribute refactor. I shouldn't do any more hacking today (barring serious procrastination), so no real hurry on reviewing it
jnthn Is the * there a fossil?
arnsholt Yeah
Where is it? 10:14
Oh, default_
Found another one in CStruct.c 10:16
10:16 erkan left
dalek p/dyncall-sized-num: 9b79426 | (Arne Skjærholt)++ | src/6model/repr (2 files):
Fix fossilized return types of two get_attribute_natives.
 10:16
jnthn arnsholt: This approach looks much better. 10:21
Better than any of the previous ones. :)
arnsholt Yeah, I like it better as well
This keeps the mucking about with pointers confined to the int/num REPRs, and the messing about with getting STables and then dispatching back to them was bass ackwards 10:22
10:24 dayangkun left 10:26 risou joined 10:28 sevin_ left
jnthn Introspection on a hot path is often a bad sign. 10:28
Su-Shee I've just gotten pasted this wonderful "perl 6 catpic" by mr shee ;) icanhascheezburger.files.wordpress....9f50f2.jpg 10:30
jnthn aww :) 10:35
10:36 hulu joined
moritz Su-Shee++ 10:36
hulu helo
moritz Mr-Shee++
helo hulu
hulu what means about "@^others"
moritz hulu: with ^ means it's a parameter to the current routine 10:37
s/with/the/
or block
and the @ means it's an array
hulu thx 10:38
moritz r: sub f { say $a; say @^others.perl }; f(1, <a b c>)
p6eval rakudo 9ec0a5: OUTPUT«===SORRY!===␤Variable '$a' is not declared␤at /tmp/GBMtlmboVb:1␤------> sub f { say $a⏏; say @^others.perl }; f(1, <a b c>)␤ expecting any of:␤ postfix␤»
moritz r: sub f { say $^a; say @^others.perl }; f(1, <a b c>)
p6eval rakudo 9ec0a5: OUTPUT«1␤("a", "b", "c")␤»
hulu helo 10:49
why sub f { say $^a; say @^b.perl; say $^others.perl }; f(1, <a b c>, 2) 10:50
is wrong
jnthn r: sub f { say $^a; say @^b.perl; say $^others.perl }; f(1, <a b c>, 2) 10:51
p6eval rakudo 9ec0a5: OUTPUT«1␤("a", "b", "c")␤2␤»
jnthn Looks right to me...what makes you thing it's wrong?
*htink
moritz hulu: which version of rakudo are you using? 10:53
hulu 3.6.0
moritz that is not a rakudo version. 10:54
jnthn That's a Parrot version...from 2011...
hulu rakudo 0.1 10:55
moritz hulu: that's also not a rakudo version
hulu how to check version
nwc10 did you build Rakudo yourself, or install it from the OS or packaging system?
hulu install from os 10:56
nwc10 aha. thanks. :-(
it will be a very old version.
10:57 cog left
moritz hulu: you might have more fun with rakudo if you install a new version (which you have to compile yourself), like 2013.01 10:57
there's a lot of bugs we fixed since 2011
jnthn Indeed. There's been a LOT of improvements...
nwc10 parrot 3.6.0 is 18 months old, so likely it will be a Rakudo version from 18-15 months ago 10:58
and, OS packaging of fast moving software considered harmful
(I continue to assert that I'm right until proven wrong)
(If I state an opinion strongly)
10:58 anuby_ left
nwc10 (sorry, can't be more directly helpful) 10:59
hulu i c thx
11:03 cog joined 11:07 erkan joined, erkan left, erkan joined, Su-Shee left
hulu helo 11:10
what's means of "Placeholder variable cannot override existing signature"
moritz a placeholder is basically a signature for the surrounding block 11:11
11:11 Su-Shee joined
moritz but if that block already has a signature, you'll get that error 11:11
r: sub f() { $^x }
p6eval rakudo 9ec0a5: OUTPUT«===SORRY!===␤Placeholder variable '$^x' cannot override existing signature␤at /tmp/fTS4o98edn:1␤------> sub f() { $^x }⏏<EOL>␤ expecting any of:␤ postfix␤ infix or meta-infix␤ infix stopper␤ statem…
moritz the () already is an empty signature 11:12
hulu what's a signature 11:13
moritz a specification what kind of arguments a function expects 11:15
perlcabal.org/syn/S06.html#Signatures
sergot hi! o/ 11:20
11:28 Liz__ joined 11:29 Liz_ left 11:41 kaleem left 12:00 kaleem joined 12:07 isBEKaml joined 12:16 kaleem left, kaleem joined
nwc10 [Coke]: cool 12:25
12:37 dayangkun joined
jnthn Should be able to break the 300 passing test cases mark tonight :) 12:38
12:40 FROGGS left 12:41 isBEKaml left 12:57 hulu left 13:05 cog left 13:06 cog joined 13:07 domidumont joined
nwc10 I wasn't actually thinking of Rakduo-atop-the-JVM when I typed AbstractPonyFactory 13:08
but it does seem to (mostly) fit
jnthn If I needed a name for the project, that would be a cute one :P 13:11
moritz will take the ConcretePonyFactory then :-) 13:13
jnthn I used to think the "use all the pattern names in every Java codebase!" thing was a joke. Then I started looking at the BCEL API and sure enough... InstructionFactory. Visitor. :) 13:14
moritz it's the waterbed pushing up on the side of type indirection, because other language features are missing 13:17
nwc10 ie this? steve-yegge.blogspot.co.at/2006/03/...nouns.html 13:19
(Google leaks the location (or at least, its inferred location) of my IP address) 13:20
(was fun when I was appearing from something.earth.li)
13:24 kaleem left 13:40 bluescreen10 joined 13:53 pjcj_ joined 13:55 BooK_ joined, xenoterracide left, pjcj left, pjcj_ is now known as pjcj 13:57 xenoterracide joined 13:59 bluescreen10 left, bluescreen10 joined 14:05 cog left 14:06 PacoAir joined 14:11 cog joined 14:18 kaare_ joined, atrodo joined 14:20 mtymula joined
mtymula hello 14:20
??
i have a problem with installing panda 14:21
[Coke] tadzik, you're up.
14:21 mtk joined
jnthn
.oO( sounds unbearable 0
 14:21
[Coke] What's the problem?
jnthn mtymula: Please describe it a bit :)
mtymula perl6 bootstrap.pl ==> Bootstrapping Panda No such method 'path' for invocant of type 'Any' in block at bootstrap.pl:20
i have fedora F17
i've installed git 14:22
and cloned from git repository
than i got inside the panda folder
and perl6 bootstrap.pl 14:23
throws error
No such method 'path' for invocant of type 'Any' in block at bootstrap.pl:20
14:23 mjwhitta joined
moritz mtymula: what's your rakudo version+ 14:24
s/\+/?/
mtymula This is perl6 version 2012.08 built on parrot 4.7.0 revision 0
timotimo wow, that's quite old already 14:25
14:27 sqirrel left
jnthn Such is the problem...Rakudo moves fast, Panda tracks its latest module installation details as we flesh them out, and then it doesn't work for older Rakudos that are shipped with distros. 14:27
I guess if R* was packaged instead you'd at least get a matching panda...
moritz mtymula: if you use panda from git, you also have to use rakudo from git 14:28
mtymula moritz: thanks for humanish explanation;)
i will do that 14:29
moritz jnthn: yes, though I don't know if 2012.08 included panda
jnthn moritz: oh, good point.
14:32 stevan_ joined
mtymula i would like to make a propositon... 14:33
there is a perl Configure.pl --gen-parrot --gen-nqp
i thing good idea is to make a perl Configure.pl --gen-parrot --gen-nqp --gen-panda 14:34
so lame people like me could have easier live;)
moritz mtymula: if you download the latest rakudo star tarball, 'make install' will give you a working combination of rakudo and panda
it's not quite 'from git', but new enough
timotimo the newest rakudo star is from yesterday, for instance :) 14:35
mtymula i am installing new rakudo as we speak...
yes i just saw that
hoelzro does Rakudo * have packaging for some of the more popular distributions? 14:36
moritz not that I'm aware of. 14:38
hoelzro someone should get on that =) 14:39
aur.archlinux.org/packages/rakudo-star/ =(
moritz just make archlinux a more popular distribution => problem solved :-)
hoelzro the problem there is the package is nearly two years out of date 14:40
and my offer of a new PKGBUILD (which I update monthly) was ignored
I'm going to see if I can get control of the primary one
14:41 benabik left
timotimo .o( "assuming direct control" ) 14:41
hoelzro haha
I was thinking about that game on the tram this morning
and how much I liked the first two
arnsholt hoelzro: If there's been nothing for two years, your case that it's an orphan package should be pretty strong 14:42
But I guess there might not be much in the way of procedures for that?
14:42 mjwhitta left
hoelzro I don't know 14:42
mtymula timotimo ME3??
hoelzro I might have to reach out in a different way 14:43
maybe I should package up Rakudo * for Fedora and Debian?
that might spur some interest?
jnthn The problem with packaging is that Rakudo improves fast 14:46
And packages rarely keep up
So we end up with lots of "X is broken!" queries we'd not have if people always build from recent source releases. 14:47
hoelzro jnthn: with my AUR package, I currently need to change two lines
hmm, that brings up a good point
jnthn And Linux is an utter mess for this. For Windows it's easy: you build one installer, done. For Linux there's a gazillion distros that all want something different. :/
14:47 nnunley left
hoelzro jnthn: well, if I were to do this, I would build an RPM for a popular RPM distro and a deb for a popular APT distro 14:48
jnthn *nod*
hoelzro (along with an Arch Linux, because I'm an Archer)
jnthn See, there's 3 already ;)
hoelzro I'm hoping that if I did that 14:49
other distro maintainers that use RPM/deb would use my spec files
either verbatim or with slight modifications
and perhaps someone in Fedora/Debian would take over my spec files
arnsholt Might be an idea
hoelzro this is a way I can *actually* contribute to Rakudo
arnsholt Definitely not a bad idea to have a Linux-packager-in-chief I think 14:50
hoelzro because I haven't sat down long enough to wrap my head around the internals =)
the question remains: how often do I build a new package?
every month?
every 3?
moritz every 3 months would be fine, IMHO
jnthn I don't want to discourage contribution or making it easy to get Rakudo. I'm just aware that we get a lot of questions or have people running into problems that they'd not hit with a recent release.
timotimo hoelzro: i just skimmed the perl6 vim syntax and ... oh crap
there's just no way i could even make a dent in that thing 14:51
hoelzro timotimo: I know, it's intimidating
jnthn: so the problem is that people aren't using a recent enough release? or that new releases are breaking old code?
moritz the former, mostly 14:52
jnthn hoelzro: Often that the distros have very old releases.
hoelzro I figured
hmm
jnthn Very old in Rakudo time is quite short :)
timotimo oh, isn't today "jnthn has many tuits for jvm" day? :D
jnthn timotimo: Got moved to tomorrow. 14:53
hoelzro well, that's a good point
timotimo OK
hoelzro Ubuntu is (currently) on a 6 month release cycle
6 months in Rakudo time is ancient
jnthn timotimo: Ended up reviewing a bunch of security course material today.
Should get some tuits again this evening...
hoelzro I'm open to suggestions
and I'm willing to maintain the spec files for Debian/Fedora
timotimo tadzik: i'm thinking i should create a String::Levenshtein or something package rather than putting my code right into panda. would that still be useful to you? 14:54
jnthn hoelzro: I'm not sure what to suggest. Simply, we're just at a point where even a few months of changes can make a huge difference to a user's experience of Rakudo.
timotimo tadzik: would you prefer trying to load levenshtein at runtime when it's needed and using it if it's there or ignoring it if it isn't?
jnthn Consider how much better parse error reporting between November and now.
hoelzro jnthn: what's worse? exposure to an older version, or no exposure at all?
jnthn: too true (also, that change is *great*)
timotimo hoelzro: no version at all could mean "this version is 2 years old, maybe i should look beyond my packaging system?" 14:55
hoelzro timotimo: right
jnthn hoelzro: Well, would it be "no exposure" or "oh, I'll install the latest release from source".
hoelzro: I know on Windows "install from source" is a huge impediment. But I got the impression it's less so for various other platforms.
hoelzro jnthn: how many users would be uncomfortable installing something like Rakudo outside the context of their package manager, though?
I (personally) like everything on my system tracked by my package manager 14:56
jnthn hoelzro: I don't know. I know the Windows issue is mostly that you simply have no build environment on most machines even if you were willing to compile from source.
*nod*
hoelzro and I feel that having OS packages would instill some more confidence in users
jnthn OK.
hoelzro but I definitely see where you're coming from; it's not as simple as "create some packages and throw them out there" =S
jnthn Yeah. Well, at least things get better from this angle with time. 14:57
It's not that Rakudo will stop improving quickly, just that with time the things people are likely to hit and find really nasty decrease.
mtymula can anyone tell me how to embed perl6 script into html file??
hoelzro also, people should upvote this: www.reddit.com/r/perl6/comments/17m..._released/ 14:59
;)
Su-Shee mtymula: how do you mean? like javascript? 15:00
15:00 nnunley joined
mtymula kinda 15:00
i run server in perl6
Su-Shee mtymula: that's not possible, which browser is supposed to understand it? ;) 15:01
mtymula: do you mean a cgi script?!
mtymula ok i will tell you what i want to do
Su-Shee that would help a lot ;) 15:02
mtymula i have RequestHandler module fr perl6
*for
and i start a http server and run some templates prepared in html files
Su-Shee ah you need a template module. 15:03
mtymula i got it
it works...:)
what i need is:
15:03 stopbit joined
mtymula what i need is to for example change content between tags <title> something </title> 15:04
Su-Shee mtymula: write a parser.
mtymula: or use an xml module.
moritz there are template modules already on modules.perl6.org 15:05
mtymula bu... i was hoping someone made it simpler;p to call me aprogrammer would be and overstatement
ok, i will try to "make it so:
thx 15:06
Su-Shee mtymula: sadly it's a pretty nasty problem so you can't get away with regex in the long run.
mtymula: that's why you use an xml module for example which gives you a node title and then you can usally just $node.text('foobar') or something similar.. 15:07
mtymula: behind it usally already is a parser.
15:07 skids joined
timotimo it's strange that you can't use the template engine you already have and do the title replacement with that 15:07
Su-Shee good point, shouldn't it look like <title>$title</title> basically anyways? 15:08
mtymula yes <title>$title</title> would be great 15:09
15:09 sjn left 15:10 sjn joined
Su-Shee mtymula: uhm that's the entire idea of your template thingie :) 15:10
mtymula: the passing in of parameters is usally what your web framework does. so that is possibly the part you have to do. 15:11
mtymula oh... crap... lots of work ahead of me...
timotimo i think i misunderstood mtymula when they wrote "i got it","it works...:)"
15:11 arlinius joined
Su-Shee mtymula: well that's one of the main features of web frameworks.. connecting "web framework" with "templating engine" 15:12
15:14 sjn left 15:15 sjn joined 15:19 sjn left 15:20 sjn joined, SmokeMachine joined 15:24 sjn left 15:25 sjn joined 15:30 sjn left, sjn joined 15:31 daniel-s left, lue left 15:32 hoelzro is now known as hoelzro|away, hoelzro|away is now known as hoelzro 15:34 sjn left 15:35 sjn joined 15:38 hash_table joined 15:39 sjn left 15:40 sjn joined 15:43 lue joined 15:48 Targen joined 15:49 kaare_ left 15:51 marloshouse left, marloshouse joined 15:54 kaare_ joined
dalek rl6-bench: 7f9a331 | (Geoffrey Broadwell)++ | bench:
Work around for fail
 15:55
rl6-bench: 17eb310 | (Geoffrey Broadwell)++ | timeall:
Make it possible to detect and skip tests using code; in particular, being able to check whether a particular compiler build contains a particular commit
rl6-bench: bba8568 | (Geoffrey Broadwell)++ | microbenchmarks.pl:
Add for_push_1e5 test, guarded by code that makes sure Rakudo is new enough not to eat all of RAM running the test
jnthn om nom nom ram 15:56
japhb Tasty, tasty chips
jnthn
.oO( lamb flavor? )
 15:57
16:02 jaldhar left, benabik joined 16:07 mtymula left
nwc10 hoelzro: what Chromium did (but I didn't get it to work, for some stupid reason) on Ubuntu was to create a packed for the "nightlies" that actually installed a new source for APT 16:09
and in turn *that* pointed to a repository that had a new .deb every night
hoelzro nwc10: so have two sets of packages?
nwc10 (this was some years ago)
the idea being that it's probably worse to have an obsolete package in the base OS, than not to have a package 16:10
but this way, people could still use the package manager to get something current
not sure why it didn't work for me. That says something about it being "not actually that good"
I realise that I'm sounding like "Stop energy", but I feel that it's counterproductive having *at best* several-month old packages in an OS (ie, when that OS version is at the day of release) 16:11
at least, it's certainly not a "no brainer obviously brilliant" thing to have it packaged 16:12
eg, as far as I can make out, macports has parrot 3.3.0
(and no Rakudo)
what use is parrot 3.3.0 to anyone?
I guess, right now, having monthly updating .debs and .rpms on a distribution site is useful 16:13
geekosaur I had considered doing rakudo for macports but decided against taking on a packaging responsibility I wasn't really in a position to maintain. I'm still not sure I'm in a position to maintain it... 16:14
16:14 sjn left, sjn joined 16:15 jaldhar joined
kresike bye folks 16:19
16:19 kresike left 16:23 cog left
jnthn decommute & 16:23
16:24 sjn left 16:25 sjn joined 16:27 cog joined 16:29 balders_dod joined 16:30 snobbo joined
masak good evening, #perl6 16:45
phenny masak: 30 Jan 23:30Z <lue> tell masak his blog's RSS feed is failing to load
masak :/
lue: thanks for noticing. 16:46
16:52 domidumont left 16:54 Chillance joined 16:56 kaleem joined 16:57 d4l3k_ joined 16:58 dalek left, atrodo_ joined, atrodo_ left 16:59 d4l3k_ is now known as dalek, atrodo left
tadzik timotimo: that sounds sane, yes 17:05
and it definitely should be a module
but I think we do have Text::Levenshtein, don't we?
feather.perl6.nl/~sergot/modules/mo...htein.html
timotimo we do? oh! 17:06
17:07 hoelzro is now known as hoelzro|away
masak only now notices blogs.perl.org/users/stevan_little/...5-mop.html 17:08
17:15 sjn left 17:16 sjn joined
Liz__ wonders what masak thinks of it 17:17
PerlJam clearly he's going to learn scala and make Perl 5 a better place too ;) 17:18
17:19 arlinius left
tadzik mayhaps we could have a Moe hackathon at PLPW and invite the Scala guys 17:19
17:21 cog left 17:22 xinming left, kaleem left 17:23 xinming joined 17:25 cog joined
PerlJam wonders what stevan's plans are for that time period ... 17:25
maybe you could have a colocated hackathon if he's not goign to be around for plpw 17:26
17:26 benabik_ joined, benabik left, benabik_ is now known as benabik
Liz__ wonders whether it wouldn't make more sense to put more effort into STD5 and get Perl 5 running on JVM because of jhntn's work ? 17:28
jnthn o/ :) 17:30
Liz__: Well, that approach would certainly allow easier integration :) 17:31
17:31 MayDaniel joined
jnthn Note that there's a step between the grammar and having a QAST tree ;) 17:31
17:31 xinming left 17:32 xinming joined
Liz__ where would this need to be done, and in what language? 17:32
jnthn The action methods that construct the QAST are written in NQP 17:33
Liz__ I assume the STD action methods could be used as inspiration? 17:34
jnthn STD_P5 just parses, afaik.
17:35 domidumont joined
Liz__ but the parse is pretty complete ? Or is it still a small subset? 17:35
17:35 domidumont left, domidumont joined
TimToady it can parse itself :-) 17:35
it doesn't parse indirect objects
jnthn TimToady: Didn't those turn out to be Very Tricky? 17:36
TimToady yeah, a lot trickier than Perl 6's
Liz__ since I personally hate indirect object syntax, I can live with that
but I guess "print STDERR 'foo'" also doesn't work then, right? 17:37
TimToady well, it's completely unambiguous in P6, due to the extra colon
it could be taught to do some of the simpler forms like that one
std: use v5; print STDERR 'foo' 17:38
p6eval std 7deb9d7: OUTPUT«ok 00:00 52m␤»
TimToady hmm, probably thinks STDERR is a function
flussence well, it's not by default an object :)
(in the sense of pointy arrows working on it)
TimToady std: use v5; STDERR->print('foo') 17:39
p6eval std 7deb9d7: OUTPUT«ok 00:01 52m␤»
TimToady it's fine, syntactically :)
Liz__ hehe
flussence oh, it needs the ()s
Liz__ just some actions methods missing to get it to work, eh? 17:40
TimToady just like P6
(or a colon, in the case of P6)
if you mean getting STD to 'work', we haven't done that with P6, let along P5
*alone
flussence if you want a "working" STD, niecza is pretty close... 17:41
Liz__ understood
TimToady and rakudo keeps gettin closer
jnthn Rakudo ain't so far off these days either, mind :)
Especially since we nabbed much of the error goodness :)
There's a few big things to figure out... 17:42
Well, only so big.
I guess we don't get indirect object syntax right in Rakudo yet :)
moritz nor labels
jnthn Oh, I can *parse* labels very easily :)
moritz nor null regexes, afaict
jnthn Labels have never been a parsing issue :) 17:43
Liz__ so, if I would want to use some figures in a presentation at FOSDEM
moritz r: /[a | ] /
p6eval rakudo 9ec0a5: ( no output )
TimToady std: /[a | ] /
p6eval std 7deb9d7: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/frw_aJKynN line 1:␤------> /[a | ⏏] /␤Parse failed␤FAILED 00:00 42m␤»
Liz__ about Perl 6 / Rakudo / Niecza, what would they be?
jnthn Yeah, that also wants a fix.
Need to update to do the new sigspace rules also... 17:44
jnthn is patiently waiting to see if some Pm tuits appear for that :)
flussence r: say 'abc' ~~ /[a | ] / 17:45
p6eval rakudo 9ec0a5: OUTPUT«｢a｣␤␤»
flussence r: say 'bcd' ~~ /[a | ] /
p6eval rakudo 9ec0a5: OUTPUT«｢｣␤␤»
moritz it matches the empty string 17:46
timotimo what do labels do in perl6?
TimToady it's outlawed for user confusion, not because we can't match the thing
n: L10: say "hi"; goto L10;
jnthn timotimo: Enable us to have a comefr^Wgoto keyword :)
flussence timotimo: loop break targets
p6eval niecza v24-18-gaf64300: OUTPUT«(timeout)hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi␤hi
moritz timotimo: they serve as destinations for next/goto/last/redo
timotimo ah, loop break targets would indeed be nice
jnthn Yeah, I pondered how to do those a bit ago
Then go distracted with other things
Like, this JVM port and stuff :)
And native arrays...
diakopter stuff++ 17:47
17:47 cog left
flussence
.oO( use Acme::LineNumbers )
 17:47
jnthn Though I didn't actually implement the native arrays yet, just figured out much of how to do 'em.
TimToady as long as jnthn++ keeps knocking off the impossible bits, we can probably find someone else to tackle the merely hard bits
diakopter waves from the dungeon 17:48
TimToady
.oO(He says "It can be proven..." and falls back asleep.)
17:53 cog joined 17:55 benabik_ joined, benabik left, benabik_ is now known as benabik 18:00 dakkar left
dalek p-jvm-prep: cea09c4 | jnthn++ | src/org/perl6/nqp/ (5 files):
Ensure CodeRefs get a valid STable.
 18:07
p-jvm-prep: 20d3bd9 | jnthn++ | t/nqp/40-lists.t:
We now pass 40-lists.t.
timotimo do i understand correctly that before nqp can be cc'd onto the jvm, one needs to implement a/the regex/grammar engine? 18:08
jnthn timotimo: yes
timotimo: Including porting the NFA stuff for doing LTM
timotimo oh my :|
jnthn timotimo: It's a fairly sizable chunk of work. 18:09
timotimo: otoh, getting this far was even more sizable and chunky :)
18:09 cog left
timotimo why can't it be ported to nqp and then cross-compiled, too? :D 18:09
jnthn Well, about of much of it as can be already is written in NQP 18:10
TimToady wonders if anyone is planning to propose any Perl 6 talks for OSCON
jnthn TimToady: When is OSCON this year?
TimToady deadline is Feb 4, my executive secretary informs me
timotimo wow, that sentence took me more work to parse than expected
jnthn TimToady: I am somewhat pondering doing YAPC::NA for the first time this year...
TimToady July 22-26, it would appear 18:11
jnthn timotimo: Thing is, the NFA can be assembled just fine in NQP code, and Cursor can be implemented there. However...
The NFA runner really wants to be in something low level since it's very hot path.
timotimo i see
jnthn And the regexes themselves are actually compiled.
timotimo well, the JVM can do wonderful things to hot paths, i hear ...
jnthn Which means implementing QAST::Regex compilation. 18:12
timotimo i shan't bother you with work that's "far" in the future
jnthn Well, it's kinda the next big thing to worry about.
timotimo oh my :|
i feel entirely powerless to help, unfortunately
jnthn We're already up to 40 out of the 65 tests from t/nqp 18:13
spider-mario then we can make the JVM the primary platform for rakudo and rename it to JPerl
… just kidding. ;)
jnthn 9 of those remaining are regex related.
And 3 of them are Parrot specific. 18:14
r: say 65 - 40 - 9 - 3
p6eval rakudo 9ec0a5: OUTPUT«13␤»
tadzik lucky you!
oh, wait
jnthn :P
tadzik gee, the compilers exam is an exhausting beast
18:14 xinming_ joined
jnthn So it's probable that by tomorrow, if not today, I'll be down to a single digit number of tests left before we block on regexes. 18:15
Liz__ after regexes, what would be the next block? (if any?) 18:16
TimToady I think a more likely name is 'rakujo 18:17
or rakujou
18:17 xinming left
TimToady comfort child, of comfort shape, or comfort passion 18:18
jnthn Liz__: Probably lots of small things rather than any one huge thing. 18:19
Liz__: More pir:: ops to make go away, IO to abstract...
Liz__ ok… sounds familiar...
:-( 18:20
wrt OSCON: if I would sign up for a presentation, what would be a good thing to present about Perl 6 ?
suggestions? 18:21
jnthn Well, Rakudo should be on the JVM by summer conf season at least... :) 18:22
Liz__ thoughts?
PerlJam Liz__: "Real World Perl 6"
jnthn And who knows, maybe more...
Liz__ jnthn: you're thinking potion there ?
jnthn Liz__: No, more of the pmurias++ JavaScript porting 18:23
Liz__ aha… v8 as a backend ? 18:24
jnthn I believe so. I've only followed it so closely, but each time I look it seems to be heading in an interesting way :) 18:25
Liz__ url?
tadzik Liz__: github.com/pmurias/rakudo-js
jnthn github.com/pmurias/rakudo-js/
jnthn likes to hope the work he's doing to further detangle NQP from Parrot is helping that effort also :) 18:26
Liz__ I would think so...
jnthn dinner, bbiab 18:31
18:32 cog joined, atrodo joined
Liz__ same here 18:35
18:37 cog left
[Coke] (we should add a perl6 --check-for-updates so users can know how out of date their local copy is. :) 18:44
nwc10 and then the option to read e-mail? 18:45
[Coke] (macports parrot) I reaaaaly hope you don't look at the maintainer there. :P
jnthn: arrrgh. if you're going to YAPC::NA I feel like I should make an effort to go. 18:46
18:47 cog joined 18:48 balders_1od joined
[Coke] nwc10: do you think it would be more helpful to get parrot-latest in there, or to remove it entirely? 18:48
nwc10 OK, I'm biased, I think. I'm only interested in parrot as a means to an end. The end being Perl 6. 18:49
so, this week, parrot-latest is more useful 18:50
[Coke] ok, with that end in mind, is it more helpful to have parrot as a separate install? I know package maintainers would say yes.
18:50 balders_dod left 18:51 arlinius joined
nwc10 I'm not sure what I should say. :-) 18:51
[Coke] or is it easier/better/whatever to have parrot as a bundled part of installing rakudo?
eh. I feel guilty enough I will probably get that sorted this weekend now.
nwc10 I'm building from source from git, but not actually using it
ie "does it still build" "Oh, nice"
[Coke] Last time I did it was on my PPC macbook, so it's been a while. 18:52
nwc10 I didn't mean to cause you to do work
ie, to my mind a macports parrot isn't very interesting without a macports Rakudo
[Coke] as a user of perlsix, I "just" grab the latest rakudo every once in a while and reinstall it somewhere I can get at it.
nwc10 and I don't need either.
(and I couldn't find Rakudo in macports, so either I suck at searching, or it's not there) 18:53
[Coke] catch-22; parrot wants to be used by more than just rakudo, but it needs users from more than just rakudo.
nwc10: it's not there, no.
I might be able to be convinced to get that working also. 18:54
so, for macports, the /ideal/ would be having latest parrot release and rakudo release uploaded into ports every month. 18:55
hai?
nwc10 Personally, I'm far more curious to see how Rakudo-on-JVM stacks up against Rakudo-on-Parrot. Which, obviously is not going to happen this weekend.
Yes, I think that that would be the ideal. That's a lot of work.
[Coke] well, yes, me too.
the first time, yes.
nwc10 Likely I wouldn't even bother installing Rakudo from macports. So I am not the end users you seek. 18:56
[Coke] next time I need to stop myself from having a snack, I'll look into it. :)
[Coke] feels momentarily confused.
nwc10 Rakudo - part of a calorie controlled diet
[Coke] sends nwc10 on his way.
nwc10 ie, please don't do this to please me, as I'm not seeking to use it from macports 18:57
given infinite resources (ie not true), having current in macports, updated monthly, would be ideal.
[Coke] right. that is lower hanging fruit for me than whatever voodoo jnthn is doing at the moment. 18:58
nwc10 if you ask the more direct question of "which is best, which is worst - leave macport parrot alone at 3.3.0, update to current parrot, or delete the port?"
I don't know.
[Coke] ... though really I should follow that more closely, as I know enough java to be dangerous.
eh. someone opens a ticket on that port, I'll be more concerned. :) 18:59
nwc10 because at the rate Jnthn's going, Christmas will be before Ascention.
("you wish")
dalek : b647daf | larry++ | docs/Perl6/Cheatsheet/cheatsheet. (2 files):
Minor tweaks
 19:01
moritz woah, I've written 20 lines of NQP code, and it compiled on the first try 19:05
nwc10 ship it!
moritz well, first I should see if rakudo still compiles :-) 19:06
TimToady and then see if it complies 19:07
moritz compliance is overrated :-)
\o/ it ran
TimToady well, fudging certainly overrates compliance 19:08
19:09 aindilis` left
dalek kudo/constant-folding: 622595b | moritz++ | src/Perl6/Optimizer.pm:
first shot at constant folding

so far it does nothing, but also breaks nothing. It seems that the multi dispatch call optimization comes first, and interferes with this one		 19:12
19:13 cog left
[Coke] moritz: you trying to make it so that my $x = 2 * 3 skips the * at runtime? 19:15
(e.g.) ?
moritz [Coke]: correct
timotimo ah, it will do things like compile-time-create parcels, too, won't it? 19:17
19:18 fhelmberger left
colomon does nothing, breaks nothing! 19:18
moritz the current plan is to simply call all routines at compile time that have the 'pure' trait applied, and whose arguments are all known at compile time
benabik GNDN
moritz I have no idea if that creates parcels at compile time :-)
subroutine calls, I should say
19:19 cog joined
moritz not method calls 19:19
timotimo Nothing In Nothing Out, NINO?
TimToady and if you happen to be doing it in a sink context, you can carp about anything pure you find as "Useless use of"
benabik GNDN = Goes Nowhere, Does Nothing. It was a vary common label on pipes in the background of Star Trek: The Next Generation
TimToady eval: 42; 43
moritz TimToady: oh, and I could war about variables used in sink context 19:20
*warn
TimToady enobuu?
flussence eval bacon
(colon check, guess it's gone)
moritz timotimo: do you know if there's a way to recompile the optimizer and then run it without recompiling the whole setting first? 19:37
19:38 mtk left
Liz__ I see "race" got added to the Perl 6 cheat seet by dalek just now 19:41
19:41 cog left
masak dalek, you racist. 19:41
moritz well, TimToady added it, dalek reported it 19:42
Liz__ not sure what it means, and googling it is bad 19:43
masak Liz__: do you know 'eager'?
moritz Liz__: it means "evaluate a list in parallell, and don't care about the return order"
PerlJam Perl 6 has built-in race conditions? 19:44
diakopter the starting gun is a few miles away though
masak right. 'race' is like 'hyper', except it doesn't order the results.
Liz__ aha, hyper, but on a single list as it were?
aha,
ok…. somehow I never envisioned hyper ordering results 19:45
masak Liz__: for more info, see S02:4680
Liz__ thanks!
masak Liz__: references like that turn into links in the IRC log.
(is it just me, or do you have more underscores than yesterday?) :P 19:46
moritz accumulative underscoring! 19:48
masak where will it end? 19:49
diakopter it's scores all the way down 19:50
19:52 FROGGS joined
masak .oO( all scores are under, but some are underer than others ) 19:54
19:58 Liz__ is now known as Liz
Liz better? 19:58
19:58 Liz is now known as Guest64872 19:59 cog joined
jnthn
.oO( apparenlty, not better :P )
 20:00
timotimo moritz: github.com/timo/junction-opt - see how i did it here 20:01
moritz timotimo: that looks... scary 20:02
timotimo it's not so bad 20:03
masak Guest64872: you might need to have a talk with NickServ, registering a nick or sump'n. 20:05
jnthn moritz: It just adds another stage into the compiler then tries running some code :)
moritz jnthn: currently I'm checking if all arguments of a QAST::Op(:op<call>) have a $_<has_compile_time_value> (more) 20:07
20:07 Guest64872 is now known as lizmat
lizmat I guess Liz is taken 20:07
moritz jnthn: but QAST::Want and QAST::WVal don't have that
jnthn: what's the best approach? special-case them? 20:08
afaict QAST::WVal always has a compile-time known value
jnthn moritz: Oh, I thought it was .has_compile_time_value these days... 20:10
moritz oh.
that explains a lot :-)
jnthn moritz: And that the <...> had gone away
iirc on QAST::Want it automatically delegates to the zeroth child
moritz yes, you're right
moritz recompiles 20:11
nr: say Q'\\share\a\b' 20:12
p6eval rakudo 9ec0a5, niecza v24-18-gaf64300: OUTPUT«\\share\a\b␤»
moritz now my optimization thingy actuallly does something 20:15
... and promptly fails, because I misspelled a method
20:15 domidumont left 20:21 SmokeMachine left 20:22 SmokeMachine joined 20:26 SmokeMachine left 20:27 benabik left
timotimo hm. i don't think i've got an answer to this question yet: how do i launch a process from perl6 and communicate with its stdin/stdout/stderr? 20:27
masak I second that question. 20:28
moritz you can open a pipe with open(:p, $program)
but that's only one-way communication
timotimo so i'd have to mkfifo three files and use a shell to hook them up? 20:29
20:29 benabik joined
jnthn I guess we need an open-3way or something... :) 20:30
timotimo an idea how that could look? would it return a parcel of three pipes? 20:31
moritz nah, a proper object
jnthn Or...*gasp*...an object! 20:32
timotimo i was wondering if instead of launching the perl6 ipython kernel from the python wrapper, i could launch the python wrapper from inside the perl6 ipython kernel code
that way there would be no problems if someone ended up opening stdout and writing to it without permission, because stdout is the way i currently handle communication between the two processes
other than that i suppose i could indeed mkfifo two pipes outside of both processes and use that 20:33
if i have an external pipe for communicating the protocol, the wrapper could totally just read stdout and stderr and pass that on as "output data" to ipython, that would even be more or less elegant 20:35
well, not really elegant, but it should work better than replacing $*OUT
20:47 vividsnow joined, Khisanth left 20:48 Khisanth joined 20:49 raiph joined 20:50 arlinius left, djanatyn1 is now known as djanatyn 20:58 hoelzro|away is now known as hoelzro
hoelzro o/ #perl6 20:58
raiph lizmat: you can google just selected p6 sources using tinyurl.com/googlep6
lizmat cool, thanks! 20:59
20:59 Khisanth left
hoelzro does it make sense to build Rakudo * with an existing Rakudo? 21:01
I'm thinking about packaging again, and there's an up-to-date rakudo package for Arch Linux; I'm wondering if it would make sense to package against that instead of having a different perl6 binary altogether
moritz yes, it does
21:02 sftp joined 21:03 bluescreen10 left
dalek kudo/constant-folding: cd64560 | moritz++ | src/ (2 files):
move constant folding code to the correct location; some small fixes
 21:06
hoelzro excellent 21:08
timotimo wat. my $a = open(:p, "cat") does NOT work well within the repl 21:09
21:09 sftp left
moritz $ ./perl6 -e 'say 42 + 38' 21:10
Constant-folding &infix:<+> with 2 arguments:
... done.
80
jnthn woo
timotimo cool :)
moritz and --target=optimize shows that there is no call to infix:<+> anymore
timotimo moritz: try a parcel with infix:<,>?
jnthn moritz++ 21:11
moritz timotimo: no dice
timotimo OK
[Coke] moritz: how will this work in modules wher eyou might ahve loaded fmore infix:<+>'s before you run the precompiled module?
moritz timotimo: ah, I haven't marked infix:<,> as 'pure' in the setting
21:12 sftp joined
timotimo you could, though :) 21:12
moritz [Coke]: fine, I hope :-)
jnthn [Coke]: Any "use" statements that bring in operator overloads happen at compile time.
moritz timotimo: or you could. It's a public branch :-)
moritz runs a spectest first
tadzik whoa, that's awesome!
moritz++
timotimo i don't have a commit bit on rakudo yet ;)
but i could indeed try a pull request 21:13
jnthn [Coke]: So they're already in scope (and provided moritz++ looked up in the current scope, we're all good)
tadzik is that -O3 or wht?
moritz timotimo: aye
21:13 Khisanth joined
moritz tadzik: it's... dunno, whatever proto inlining is 21:13
jnthn That's -O2, the default
moritz please note that this is a very, very early prototype 21:14
lots of stuff can still go wrong
tadzik yeah, right
you know us ;)
jnthn perl6 -e "say 1/0" :) 21:15
moritz jnthn: I've disabled 'is pure' on infix:</>
there's... stuff going on
constructing of Rats and stuff
masak did someone say "very, very early prototype"? :>
jnthn ah
moritz and some bootstrappy problems too, I think
jnthn masak: No, they said "very very early jQuery" :P
moritz YUI!
tadzik masak: that's Perlish for "production ready" :>
21:16 p5eval joined
moritz ... and I wanted to be in bed half an hour ago 21:16
jnthn Hacking > Bed...well, sometimes :)
moritz good night folks
jnthn 'night, moritz
lizmat night, moritz
tadzik gnight
[Coke] jnthn: let's say you define a + in your .pl - you include a module that uses + - does your defined + impact how the module works? 21:17
jnthn [Coke]: No
[Coke]: Due to separate compilation.
[Coke] ok. my concern was that, if it did, and you precompiled the module outside of your .pl, boom.
jnthn [Coke]: It used to; that was a bug fixed around October last year :)
[Coke] jnthn++ 21:18
jnthn Yeah, the spec says such things should not leak into the compilation of modules.
oh, it wasn't actually the multis that leaked
We've not had that bug in years, if ever
It was user-defined operators.
And only the parsing, not the multi candidates.
dalek rl6-roast-data: b13834d | coke++ | / (4 files):
today (automated commit)
 21:21
[Coke] looks like rakudo has several new aborts.
niecza has some old ones. 21:22
pugs needs a hug.
masak aww
21:22 Khisanth left 21:23 sftp left 21:28 hash_table left
[Coke] at this point, I'm just trying to keep tests fudged for pugs. very unlikely haskell shaped tuits will appear. 21:31
jnthn: is docs/LHF uptodate for nqpjvmprep 21:32
? 21:33
21:35 cog left 21:36 xinming_ left 21:37 cog joined, geekosaur left 21:38 xinming joined 21:39 Khisanth joined, geekosaur joined 21:41 kaare_ left 21:43 spider-mario left 21:45 [pp]sromanov joined
[pp]sromanov moritz: (or anyone) sorry for bothering, but what it takes to add a channel to irclog.perlgeek.de/? 21:46
we really miss some nice logging at #moe :)
jnthn [pp]sromanov: Asking moritz should be enough, but he's sleeping at the moment. 21:47
[pp]sromanov oh, thanks!
jnthn [Coke]: just checked. Yes.
21:48 benabik left, Khisanth left
masak hm, I can see strangelyconsistent.org/blog/feed.atom and it looks legit... apart from all those weird characters at the end of every post. 22:03
what are those?
jnthn masak: Doesn't work for me 22:04
masak: Also, seems planetsix gets them
masak hexdump says they're NUL bytes.
wtf.
yeah, I can see that being an issue in an *utf-8* file.
jnthn masak: oddly, if I get the source it looks OK 22:05
oh
no, has the null bytes it seems
22:05 skids left
masak I generated this using Rakudo. so I suspect I can make a rakudobug out of this. 22:05
jnthn which prevents firefox's thingy from displaying the feed
masak but need to investigate the cause.
flussence takes a look 22:06
.oO( this is the strangest compiler I've ever debugged )
 22:14
jnthn It is compiling the strangest language :P 22:15
22:17 PacoAir left
masak .oO( strangely compiling ) 22:19
huf :)
flussence it's adding the nulls when doing a completely normal-looking slurp/print around line 84... (whee, another IO bug)
22:20 Khisanth joined 22:21 gaussblurinc_ joined 22:24 gaussblurinc_ left
flussence hmm, I'm not seeing any obvious pattern to the length... 22:28
dalek p: c4434db | jnthn++ | src/core/IO.pm:
Add a join sub to the NQPCORE.setting.

The nqp::join op should be low-level and assume it's being given a set of strings to join. This sub provides the convenient, coercing form.		 22:30
p: 55631fa | jnthn++ | t/nqp/57-construction.t:
Update a test to use join, not nqp::join.
p-jvm-prep: 07762be | jnthn++ | / (2 files):
Implement nqp::join; code from thecabinet++.
 22:34
p-jvm-prep: 0cd4828 | jnthn++ | nqp-src/NQPCORE.setting:
Add a join sub.

The nqp::join op expects what it is provided with to already be stringified; add a join sub that makes sure of this.
p-jvm-prep: a3fb625 | jnthn++ | / (2 files):
Implement not_i.
p-jvm-prep: 0c31e28 | jnthn++ | lib/QAST/JASTCompiler.nqp:
Support qlist.

We now pass 57-construction.t.
jnthn dalek: you missed two! :P
tadzik did dalek die in the mi... right :)
jnthn fail 22:35
tadzik failek
diakopter silent floo
jnthn nyetlek
flussence aha 22:37
[Coke] нетlek
ooh, we should totally rename him to да 22:38
er,
flussence psyde line 212-213 eats a few lines, the length looks almost the same as the amount of null bytes. I suspect parrot's readall() is screwed up and allocating a file-sized buffer instead of a remaining-file-sized one.
[Coke] даlek. curse you, newlines!
hoelzro heh
22:41 Khisanth left
timotimo having trouble with Rakudo::Debugger 22:41
22:42 cog left
jnthn Try debu...oh. :) 22:42
masak xkcd.com/1163/ 22:44
jnthn :P
[Coke] s/brain/rakudo/g 22:45
22:46 nnunley left
p-jvm-prep: db47d4e | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Fix nqp::substr out-of-range handling.
p-jvm-prep: 39b0324 | jnthn++ | src/org/perl6/nqp/runtime/Ops.java:
Fix nqp::cmp_s semantics.

Should always give -1, 0 or 1.
p-jvm-prep: 3a441f5 | jnthn++ | t/nqp/59-nqpop.t:
Run 59-nqpop.t (which has 94 tests).
hoelzro can one install the Rakudo * modules without installing Rakudo? 22:48
for example, if I already have Rakudo itself installed?
tadzik install Task::Star
should work-ish
hoelzro tadzik: with panda? 22:49
tadzik yeah
github.com/tadzik/Task-Star/
hoelzro hmm
ugh, I'll have to think about this packaging stuff more tomorrow, I think 22:51
22:51 sftp joined 22:53 hoelzro is now known as hoelzro|away 22:55 Khisanth joined 22:59 cog joined
timotimo how can i use the perl6 grammar to parse some string? 22:59
i would like to try to use perl6-debug to interactively parse somethign
(i suppose that won't be possible, though)
flussence null bug testcase: gist.github.com/4687526 23:03
masak: you might have more success generating that atom feed with niecza ;) 23:10
23:10 LoRe joined, lizmat left 23:11 sftp left
jnthn for hell's sake. 23:11
I can't wait to be on the JVM.
So I don't have to tolerate such crap.
23:12 MayDaniel left
jnthn And nor will Rakudo's users. 23:12
23:12 sftp joined
jnthn *sigh* 23:12
23:13 Liz_ joined
tadzik *back pat* 23:13
thanks to you we won't have to
you're a hero man
23:14 Liz_ is now known as lizmat
sorear hurray for jnthn 23:14
23:14 stopbit left
colomon \o/ jnthn \o/ 23:14
23:15 [Coke] left
flussence ooh, maybe once it's on JVM we can get DBDI for nearly free, since it's intentionally a clone of Java's API... 23:16
23:17 benabik joined 23:19 geekosaur left 23:20 geekosaur joined
jnthn flussence: Perhaps so, otoh it's nice if we can have a sufficiently good NativeCall on JVM to do it that way too :) 23:20
tadzik jvm is teh silver bullet :) 23:21
jnthn heh
tadzik takes long to load though ;)
jnthn Not quite, but it's a step towards somewhere better.
tadzik one big barell
diakopter or stupendous integration with a perl5
jnthn The first port is the hardest.
diakopter looks around 23:22
tadzik then plunder gets easier
timotimo what were you just upset about, jnthn?
the null bug test suite thing?
jnthn timotimo: Yeah.
timotimo: I suspect also introduced in the IO refactor in Parrot that has caused us various other bits of pain (like, the infamous stdin reading bug) 23:23
masak lol, I wrote a post! \o/ strangelyconsistent.org/blog/i-am-g...-to-fosdem
tadzik lol
jnthn ooh, I have 80% of a blog post...
That I forgot that I have left to finish from yesterday :)
tadzik masak: this is awesome 23:24
5 word sentences
masak :)
tadzik I'm surprised with how nicely that does read
masak yeah.
that took some serious doing. :) 23:25
jnthn It's fortunate the talk title has five owrds :D
*words
tadzik admit it, this is why you chose 5
diakopter I wonder what the other language community was
masak no.
tadzik otherwise it would be hard
formulating the sentence about talk 23:26
including talk title in it
masak tadzik: the talk title actually has 6 words :P
tadzik where's doesn't count, does it
masak "Dude,"
23:26 Khisanth left
tadzik r: "Dude, where's my flying car".words.perl.say 23:26
p6eval rakudo 9ec0a5: OUTPUT«("Dude,", "where's", "my", "flying", "car").list␤»
masak :P 23:27
tadzik r: "Dude, where's my flying car".words.elems.say
p6eval rakudo 9ec0a5: OUTPUT«5␤»
jnthn Oh, I thought it was "where's" :)
masak actually I count expanded words.
tadzik you see, I was right
masak note that there are none.
in the blog post itself.
jnthn "Also, Perl 6 is everywhere. Especially in the flying cars."
dalek p-jvm-prep: bd5c5ee | jnthn++ | src/org/perl6/nqp/sixmodel/KnowHOWMethods.java:
Add missing checkarity calls in MOP code.

In their defense, maybe it didn't exist when they were written...		 23:28
p-jvm-prep: 31fcacd | jnthn++ | src/org/perl6/nqp/runtime/ (3 files):
Fix a flattening thinko, caught by the tests.
p-jvm-prep: 1bf287f | jnthn++ | t/nqp/41-flat.t:
We pass 41-flat.t.
23:28 cog left
lizmat masak: looking forward to seeing you at FOSDEM 23:28
lue feels tempted to work on a native machine code backend for nqp when he has more time whenever he hears about the jvm backend
23:29 cognominal joined
masak lizmat: looking forward to it, too. :) 23:29
lizmat lue might be interested in looking at rurban's work on p2 and potion
diakopter mmmppphhhfffhppphh
masak tadzik: I like constrained writing. it's oddly liberating :)
tadzik you broke the rules now 23:30
masak oh shut up :P
tadzik you should use 5 words ;)
masak diakopter: you too! :P
lizmat diakopter: am I missing a translation here?
masak lizmat: no, diakopter is just being indisrete. ignore him. :P
indiscrete* 23:31
diakopter so, continuous
masak smooth as an android's butt.
jnthn Didn't Apple patent being so smooth? 23:32
tadzik silky smooth
masak early findings: both the atom feed and the generated static html files contain the null bytes. 23:33
I am so not amused.
diakopter your mom contains the null bytes
jnthn masak: You saw flussence++'s gist too?
23:33 cognominal left
masak no. 23:34
masak looks for it
tadzik it's in the part where jnthn grumbles :P 23:35
jnthn :P
Yeah, I...really should not be doing much besides sleeping when this tired. :/
I'm not sure I should be this tired when I slept 9 hours last night either.
OK, there's native types fixed up... 23:36
masak found it.
yeah, that fucking sucks.
Parrot--
masak angrily ponders a workaround
jnthn .subst(/\x00/, '', :g) :P 23:37
OK, we're now passing 44 test files.
masak yeah, that might be it, actually.
jnthn r: say 65 - 44; say 65 - 44 - 9;
tadzik -3 for parrot? 23:38
jnthn uh, ah yeah
And...where is the evalbot?
tadzik p6eval: WAKEUP
jnthn bah, I can do it with nqp-jvm
tadzik p6eval: GRABHMMHMHMHMAKEUP
23:39 skids joined
dalek p-jvm-prep: 91bb86d | jnthn++ | src/org/perl6/nqp/sixmodel/SerializationReader.java:
Cope with the lack of a method cache.

Fixes deserialization of native types.		 23:39
p-jvm-prep: a0bcc37 | jnthn++ | nqp-src/NQPCORE.setting:
Uncomment native type decls in NQPCORE.setting.
p-jvm-prep: 23578d4 | jnthn++ | t/nqp/64-native.t:
We now pass 64-native.t.
jnthn So, 18 test files to go, but 9 of them need the regex work.
23:39 s1n left
lue lizmat: seems interesting. I just want a faster rakudo though :P (also, faster CORE.setting step) 23:39
benabik is converting null test case to Winxed to see if it is Parrot proble.m
I suspect it is.
p6eval rakudo 9ec0a5: OUTPUT«21␤12␤»
lizmat lue: "libp2 should be usable as a fast backend for nqp/perl6", implies rakudo to me 23:40
diakopter to say that is wildly optimistic makes even an expansive-thinking manic mutter 23:42
tadzik sleepytime for me it is. Good knight #perl6!
lue good ♞, tadzik o/
jnthn OK, enough nqp-jvm today. Especially given that to make more progress means doing one of roles, multiple dispatch, exception handlers or the regex engine...all of which are far too much effort to dig into at half past midnight. :) 23:43
lizmat diakopter: hehe, but at least he's thought about it
masak jnthn: yes, that fixed the problem. jnthn++
diakopter cry
masak jnthn: I would still like to find the cause.
oh wow. now getting "Unaligned end in UTF-8 string". never got that error before. 23:44
masak investigates
lizmat at least it will direct rurban's attention away from parrot, no?
flussence points masak at parrot/src/io/api.c:749
tadzik tee-hee
flussence that's the bottom of the turtle stack, and the rest of the layers look pretty empty to me
23:45 xinming left
lizmat is calling it a say, gnight #perl 6 23:45
s/say/day
:-)
jnthn 'night, lizmat
masak 'night, lizmat 23:47
see you on the other side.
flussence: thanks. that doesn't tell me much, but maybe it'll help others debugging this. 23:48
I am currently having an unrelated Unicode string error. :(
jnthn :( 23:49
23:51 xinming joined
masak so, basically, this is what happens. 23:52
I have a file which is *pure* ASCII.
23:52 Chillance left
masak I do qx[cat the-file] in Rakudo. 23:52
I get "Unaligned end in UTF-8 string".
this is bad. I can't render my blog now. :( :( :( 23:53
masak considers urgently downgrading his Rakudo
does anyone know which one is the latest Parrot that wasn't broken?
oh heck, I'll just try monthly decrement until stuff works again. :( 23:54
decrements*
benabik masak: If you can tell me when it breaks, I'll look into it. 23:57
flussence niecza: say Date.today 23:58
p6eval niecza v24-18-gaf64300: OUTPUT«===SORRY!===␤␤Undeclared name:␤ 'Date' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1443 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) ␤ at /home/p…
masak benabik: gist.github.com/4687526
benabik: I can also gist you the ASCII-only file that causes the "Unaligned end in UTF-8 string" error message.
benabik masak: Yes, I'm looking into that. Is that the same bug as your blog?
rurban diakopter: p2 is far easier to target than jvm. I support dynamic, typing jvm not.
masak benabik: both of those bugs are simultaneously affecting my blog.
rurban and its also much faster and smaller than the jvm 23:59
sorear what's p2?

Powered by Cro and the Raku® Programming Language.

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

Thank you!