»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.
pmurias QAST::VM.new(:jvm(QAST::SVal.new( :value('ModuleLoader.class')) ;) 00:00
jnthn Oh 00:01
yeah, OK, that bit is :)
s/nothing/few things/ :)
Maybe I should add a :generic to mark those out. 00:02
(as in, the ones that are likely generic)
pmurias if something is generic we should implement it on all vms and avoid QAST::VM 00:03
jnthn In the case of forceouter, yes, it should be an nqp::op implemented everywhere. 00:04
masak 'night, #perl6
timotimo night masak 00:08
stay positive :)
00:09 benabik joined 00:18 kst` is now known as kst
jnthn Grrr...something is still not quite right somewhere. Getting the odd serialization error. 00:21
Well, enough for today... 'night &
00:24 donaldh left
felher 'night, jnthn 00:24
dalek kudo-js: 669640b | (Paweł Murias)++ | / (5 files):
Update to work with nqp commit 3089165aa0646a334d24792cb7434051bdd4900f.
 00:36
00:39 Chillance left, cognominal left 00:40 pmurias left 00:44 cognominal joined 00:56 stevan_ joined 01:00 ggoebel_ joined 01:10 anuby joined 01:22 cognominal__ joined 01:23 cognominal left 01:44 ggoebel_ left 02:12 cognominal__ left 02:13 cognominal__ joined, FROGGS_ joined 02:15 dayangkun left 02:17 FROGGS left
colomon o/ 02:20
02:20 japhb_ left 02:31 cognominal__ left
sorear o/ colomon 02:49
colomon sorear: how's it going?
sorear good 02:50
colomon glad to hear it!
02:50 aindilis left
colomon my $work has been driving me insane the last week. 02:50
03:00 dayangkun joined 03:11 sivoais_ joined 03:14 felipe left, wtw left, xdbr left, crazedpsyc left, LoRe left, sjn left, telex left, Patterner left, xinming_ left, jtpalmer_ left, Pleiades` left, [Coke] left, tadzik left, Timbus left, Vlavv left, simcop2387 left, breinbaas left, rom1504 left, rafl left, sweet_kid left, nyuszika7h left, jdv79 left, krunen left, eternaleye left, perigrin left, Yappocall_ left, mehwork left, kst left, mathw left, Juerd left, colomon left, Ayiko left, renormalist left, lue left, betterwo1ld left, Gothmog_ left, araujo left, p5eval left, atrodo left, diakopter left, cxreg left, SHODAN left, Tene left, cotto left, amkrankruleuen left, [particle] left, baest left, TimToady left, raiph_ left, groky left, Bucciarati left, dayangkun left, FROGGS_ left, anuby left, stevan_ left, dalek left, jerome left, jaldhar left, mtk left, p6eval left, rhr left, sjohnson left, shachaf left, japhb left, chee left, preflex left, PerlJam left, alec__ left, szbalint left, raiph__ left, erkan left, Khisanth left, integral left, hoelzro|away left, rjbs left, nebuchadnezzar left, gfldex left, cosimo left, flussence left, steven__ left, marloshouse left, bapa left, BooK_ left, Ulti left, jlaire left, kshannon left, bakedb left, mls left, odoacre left, sivoais left, geekosaur left, mikemol left, xenoterracide left, djanatyn left, broquaint left, Exodist left, swarley left, imarcusthis left, Rix left, eiro left, DarthGandalf left, xy left, bonsaikitten left 03:16 sivoais_ left 03:24 perigrin joined 03:26 bonsaikitten joined 03:27 geekosaur joined 03:30 swarley joined
swarley You know you're lazy when you make an executable to add your license header to all of your source files that don't have it 03:46
bonsaikitten swarley: you're not lazy enough when that executable is more than a three-line shell script ;) 03:47
swarley Good point 03:48
gist.github.com/swarley/4736308 :p
bonsaikitten too much work! ;) 03:49
swarley Well I needed a reason to learn more Go anyway ;p
bonsaikitten I haven't had a motivation yet
OpenDylan is distracting me enough 03:50
swarley It's pretty hard to get into learning Go
But if you get past the top it's actually pretty interesting
It's compiled but it has garbage collection and polymorphism 03:51
rking Go is the most meh'tastic thing I've seen in a while.
swarley I'm on the fence about it in a lot of places
rking There are a few others vying for that title, though. Like Dart.
I think it's got a lot of good engineering for a lot of problems that I don't have.
swarley What is PAWN doing? 03:52
bonsaikitten rking: that describes most "innovation"
swarley I have no idea what the hell PAWN even is really. It looks like C to me based on the hello world
rking The one I'm waiting for an excuse to learn is in the channel name.
I still don't know how it'll fit into my "Used to use Perl5, use/like Ruby, dream of Smalltalk" world. 03:53
swarley lol. I use nqp more than I use perl6 tbh
bonsaikitten similar for me
I just think it's still an interesting experiment and package it for gentoo
swarley PAWN seems like a stupid language 03:54
bonsaikitten ... like OpenDylan, OpenCOBOL, J and a few others
swarley like exactly the same syntax as C
but it's scripted..
I don't know, nothing I would be motivated to use 03:55
I'd learn lua before any other new language
I still have C#, Perl6 to completeness (NQP included), and Go to finish learning 03:56
bonsaikitten swarley: as far as I can tell lua has some nice ideas, but is too much manual work
not sure if it's that useful ... but it seems to make lots of people happy
swarley I like the fact that Lua is embedded easily
But I'm iffy on the syntax
bonsaikitten yes, but it doesn't have any proper refcounting / GC as far as I remember 03:58
so it's very easy to get it wrong
swarley Ah. That's something that drew me into Go 03:59
bonsaikitten for me python covers most of that, and if I need things to be faster it's usually C
swarley They have a concurrent mark-sweep GC added in with the goal of minimal overhead
bonsaikitten with shell and perl5 covering the glue between things
swarley Yeah, I use ruby when speed isn't an issue, C extensions for ruby when its only a desire, and C++/C when it's a necessity 04:00
I prefer not to do large projects in C though. It feels like the code gets unruly quickly
bonsaikitten yes
takes lots of discipline to keep it nice 04:01
re: ruby, my brain can't parse it, it's just "wrong" in every aspect - plus the ecosystem is filled with bad code, bad advice and bad behaviour (who needs releases lol)
swarley ? 04:02
Please elaborate, I'm a fairly active member of the community and we don't really try to give bad advice
same for rking
I usually feel a lot less anxiety about asking a question in #ruby than i do in #perl 04:03
bonsaikitten well, most of the documentation I tend to stumble over is random blog posts
geekosaur ruby: bastard child of perl 5 and smalltalk. obj-c knowledge might help some...
swarley Ah. Well it depends on what you're looking up.
If it's the standard library, that's all documented and hosted on the same server
bonsaikitten swarley: there are some pretty smart people in the ruby cloud, and lots of people that found php too difficult and never figured out basic sanity checks
swarley If it's a 3rd party gem, that's up to the developer 04:04
oh god, yes that pisses me off
bonsaikitten but that's the added value, a module for everything
rking "Ruby is if Smalltalk and Lisp had a love child, but left it in the care of its nanny, Perl."
bonsaikitten well, at least they don't do error handling a la php - print a message to stdout *headdesk*
swarley I think there is a quote of me somewhere actually telling someone to "crawl back into the hole their frightened PHP emerged from"
bonsaikitten that's not how you handle errors :(
geekosaur the ruby community has some very good guidelines for how to handle versioning... which are not merely ignored but actively violated by most gem developers 04:05
swarley What? Module for everything?
I think you're confusing good ruby with common ruby ;)
04:06 dwoldrich joined
swarley Not to say, that I myself even write good ruby. In fact, I would say I'm about lower level advanced. So, not that special. I just keep my mouth shut when I see people coming in doing $i = 5; $i.times {|x| x += $i } and don't yell at them to at least read chapter one of any tutorial 04:07
brb
04:07 swarley left
dwoldrich Hey perl6 guys, I think Parrot was a great idea, I was hoping to see it succeed, am sad you all turned your backs on the project, and I think it's silly to target the JVM or invest man hours in it. 04:16
Just because the JVM can execute JVM bytecode faster than Parrot byte codes TODAY, doesn't make that so in the future. If you recall, the JVM was a sloth up until around JDK 1.5 or so. 04:17
And Oracle is going to tank the JVM, why oh why would you board a leaky, if not sinking, ship?? 04:18
Parrot should have been your glittering jewel, not Perl6. Parrot could have had legs, 20 year lifespan, developers and business porting AWAY from the JVM! 04:20
sorear dwoldrich: I think you have never tried to use Parrot
dwoldrich I am not saying that Parrot, or Perl6, were the best run or designed projects. 04:21
But what I am saying is, the vision was right on
If only you all could come together and be one team
not sure why the "throw it over the wall" deal existed 04:22
sorear hahahahaha
dwoldrich You just never saw yourself as competition with JVM, and that saddens me, because right around the time Parrot hit 1.0, the obsolescence and design cruft was setting in on the JVM 04:23
I watched the changelogs and was like, what the heck is the goal here?
bonsaikitten dwoldrich: the JVM has a lot more manpower behind it, so it won't just disappear
and just because someone experiments doesn't mean that everyone abandons their current project 04:24
dwoldrich you all said Parrot's dead
I'll take you at your word 04:25
bonsaikitten "all" ? pfff.
you must be a journalist
dwoldrich no, not even a perl developer 04:26
I'm a Java developer
and I don't like what I'm seeing in the java community
looking for options now
sorear you sound like a regular contributor to [email@hidden.address]
dwoldrich ?
04:26 DarthGandalf joined, eiro joined, Rix joined, imarcusthis joined, felipe joined, wtw joined 04:27 mls joined, kshannon joined, bakedb joined, BinGOs joined, odoacre joined, mikemol joined, xenoterracide joined, djanatyn joined, broquaint joined, Exodist joined, scottp joined, telex joined, Patterner joined, xinming_ joined, jtpalmer_ joined, nyuszika7h joined, jdv79 joined, krunen joined, mehwork joined, eternaleye joined, Yappocall_ joined, spanner joined, sivoais joined, xdbr joined, crazedpsyc joined, LoRe joined, sjn joined, kst joined, mathw joined, Juerd joined, colomon joined, Ayiko joined, renormalist joined, lue joined, betterwo1ld joined, Gothmog_ joined 04:28 swarley joined, araujo joined, p5eval joined, atrodo joined, diakopter joined, cxreg joined, SHODAN joined, Tene joined, cotto joined, amkrankruleuen joined, nebuchadnezzar joined, [particle] joined, baest joined, TimToady joined, raiph_ joined, groky joined, Bucciarati joined, mtk joined, stevan_ joined, dayangkun joined, FROGGS_ joined, anuby joined, dalek joined, jerome joined, jaldhar joined, p6eval joined, rhr joined, sjohnson joined, shachaf joined, japhb joined, ponbiki joined, nwc10 joined, chee joined, moorcock.freenode.net sets mode: +vv dalek p6eval, gfldex joined, cosimo joined, flussence joined, steven__ joined, marloshouse joined, bapa joined, BooK_ joined, Ulti joined, jlaire joined, Maddingue joined 04:29 Pleiades` joined, [Coke] joined, tadzik joined, Timbus joined, Vlavv joined, simcop2387 joined, breinbaas joined, rom1504 joined, rafl joined, sweet_kid joined, integral joined, preflex joined, PerlJam joined, alec__ joined, szbalint joined, raiph__ joined, erkan joined, Khisanth joined, hoelzro|away joined, rjbs joined, sunnavy joined, tomaw joined, yves_ joined, felher joined, ivan`` joined, salparadise joined, raiph_ left, bapa left, sivoais left, ivan`` left, raiph__ left, ivan`` joined 04:30 bapa joined 04:31 raiph_ joined, raiph__ joined, xy joined 04:32 sivoais joined 04:33 shinobicl joined
swarley So what is the collective opinion on targeting Go for an nqp output? I figure.. Go has a lot of cool features going for it and it compiles very fast 04:35
I mean unless its not even possible I'll still have a whack at it, but id like to hear opinions 04:36
04:38 rindolf joined
dwoldrich What other VM's could you target that don't have a corporate overlord/owner? Doesn't PHP have an underlying VM? 04:40
Why do you have to go with Google or Oracle, seriously? 04:41
ain't there good FOSS options?
geekosaur *eyeroll* 04:42
04:42 Targen joined
dwoldrich :D 04:42
Just making the "business case" for Parrot, just sayin'.
geekosaur niecza uses mono, if that's not pure enough for you then you're welcome to go digging for a *nonsucky* alternative
dwoldrich yeah!!
mono's good
why not target that? 04:43
geekosaur ...
geekosaur points to the first three words of what he said
dwoldrich again, let me present, Parrot.
or the Parrot that could be
:P
geekosaur and let me repeat "nonsucky"
dwoldrich Make it so.
geekosaur consideringt he amount of times I see rakudo having to work around the parrot internals thrashing...
dwoldrich I demand a non-sucky Parrot immediately
rindolf dwoldrich: a VM is any interpreter or CPU. 04:44
geekosaur well, bully foryou
rindolf dwoldrich: even perl 4 was a VM.
dwoldrich: and Perl 5 has a VM too.
dwoldrich: VM == virtual machine.
dwoldrich I need a write once run anywhere byte code interpreter/JIT'er, plz
swarley There is gccgo..
And cgo 04:45
rindolf dwoldrich: Java? Mono? Parrot?
04:45 benabik left
swarley Go to an extent 04:46
If you code with portability in mind
Use os.FileSeparator instead of "/" etc 04:47
There is no VM mesiah. 04:48
The reason i keep jumping around is because i realize im attempting something unwise or better left to someone else 04:50
Llvm: unwise .. Yarv: unwise.. CIL: i bet someone can do it infinitely better than myself. 04:51
04:57 rindolf left
colomon rn: say 9.is-prime 04:57
p6eval rakudo 4fb07b, niecza v24-20-g93138cc: OUTPUT«False␤»
sorear dwoldrich: anyway, if you want parrot to become usable, take it to #parrot on irc.perl.net 04:58
*irc.perl.org 04:59
dwoldrich rindolf: I like the idea of Parrot. As I understand it, it didn't have a stable bytecode, and thus required a source language to feed it parrot assembly code, which to me makes it a non-starter. A VM should be responsible for verifying, loading, and running bytecodes, very well-specified bytecodes. 05:00
05:01 preflex_ joined
sorear dwoldrich: you just replied to someone who isn't even here. 05:01
dwoldrich darn
well bring him back 05:02
or her
sorear You do it.
05:02 preflex left
dwoldrich Alright I will! 05:02
05:02 preflex_ is now known as preflex
dwoldrich sorear: I want parrot to be valued as a project first, then treasured, then used (by Perl6), and THEN made usable. :D 05:03
05:03 shinobicl left
colomon rn: say (2, 3, * + 2 ... *).grep(*.is-prime)[^10] 05:05
p6eval rakudo 4fb07b, niecza v24-20-g93138cc: OUTPUT«2 3 5 7 11 13 17 19 23 29␤»
swarley It just doesn't seem wise to use a VM still being made stable with a language still waiting for a first full implementation 05:06
Let me clarify
A VM with the aims of parrot
dwoldrich swarley: that
that's a good point
But I think a byte code specification
iron clad
would have smoothed out a lot of wrinkles 05:07
colomon rn: my @primes := (2, 3, * + 2 ... *).grep(*.is-prime); say @primes[^10]
p6eval rakudo 4fb07b, niecza v24-20-g93138cc: OUTPUT«2 3 5 7 11 13 17 19 23 29␤»
sorear dwoldrich: we have no influence on parrot whatsoever here.
they do whatever they please.
if you want to be part of the parrot decision process, join #parrot on irc.perl.org
sorry.
dwoldrich Yeah, I detect your frustration
05:09 SamuraiJack joined
dwoldrich I think because parrot is yoked to perl6, though, your (perceived/real) abandonment of that project gives the impression that it is doomed. 05:09
and if you're not going to do your own parrot fork
swarley Parrot isn't really bound to perl6 05:10
dwoldrich then, let's all hope Oracle is a nice steward of the heart that pumps the blood through our veins. :/
I think it is 05:11
swarley Parrot is independent entirely of perl6
dwoldrich It wasn't always 05:12
swarley Without perl 6 parrot could still go on, its popularity is just to be determined by how well they can get it working
I really don't understand your logic
The way I understand it
Parrot is an all purpose vm
dwoldrich The flagship for Parrot was Perl6, it was billed as being an engine flexible enough to be a target for that language
dalek osystem: 92886d4 | colomon++ | META.list:
Remove obsolete Math::Prime from the ecosystem.
 05:13
dwoldrich Without the flagship, how will it know which way to sail?
swarley Uh, it seems to be doing much better with jvm and cil based on benchmarks
sorear dwoldrich: you seem to have heard very precious little of the news from the last five years 05:14
dwoldrich this is true
I don't soak in Perl news
But I am interested!
swarley Then you may want to start by finding out that parrot is almost completely removed from nqp i believe 05:15
sorear then I should tell you that #perl6 is now a place where we care about things that work
dwoldrich I just want to say that I believe the JVM, knowing what I know about the jvm byte codes and the new stuff in JDK 7, could be a fast execution engine for Perl6
sorear we've banished all the people with revolutionary fantasies 05:16
swarley Okay?.. that's what nqp-jvm is for
dwoldrich if, at the end of the day, all objects were just hashmaps to function pointers and values, it would still work pretty fast 05:17
right
I'm saying, that that course is going to perform okay, so-so, maybe fast in very specific scenarios where you can do more with dynamicinvoke on JDK7
swarley I think the main goal here is to have it just work 05:18
And then we worry about speed later after everything works
geekosaur thinks the chanel has been rather successfully trolled...
dwoldrich Right
Not a bad plan
sorear geekosaur: do not attribute to malice that which can be explained by stupidity 05:19
although I would use a gentler word than stupidity
dwoldrich I'm just saying Parrot, as a long term visionary, we can host the world's code thing, is a better plan
swarley Lol'd
dwoldrich don't be smug
I'm trying to be helpful
lue One advantage of JVM over parrot right now is that more people use JVM and JVM is (probably) more trusted, especially in business settings. 05:20
swarley Parrot, as of this moment, gives rakudo a reputation of a terribly slow and unusable interpreter
dwoldrich yeah, at some point someone said, this thing is unworkable, I bet 05:21
swarley Not functionally, just resource and speed
geekosaur you missed some of the earlier stuff
[07 23:42] <dwoldrich> I demand a non-sucky Parrot immediately
dwoldrich :)
I thought that was funny!
swarley Then go bitch to them about it
05:22 xinming_ left
dwoldrich You don't share my vision, then? 05:22
lue dwoldrich: Are you aware of the existing effort to put nqp (and as a result rakudo) on the JVM?
swarley No, I think parrot is a marvelous idea.
dwoldrich JVM byte codes are last century, man
swarley However
Wasting time with it right now is pointless
dwoldrich fair enough
geekosaur would quote more backscroll but doesn't see the point
dwoldrich perhaps, when Perl 6 settles out 05:23
and when JVM is a good performing target
swarley Perl6 isn't the problem
dwoldrich we can revisit the parrot thing
Oh, I think Perl6 is part of the problem
05:23 xinming joined
swarley Parrot is the problem 05:23
dwoldrich your team is problematic
design churn, rewrites, replatforming 05:24
c'mon
sorear you come in here and immediately act like you own the place
you _don't_
dwoldrich Just trying to be helpful
sorear and failing.
epically.
swarley I would hardly call the team problematic
Things are done with careful consideration
I'm not even on the team, I'm just here because I admire them 05:25
lue dwoldrich: about the rewrites: perlgeek.de/blog-en/perl-6/2012-sto...rites.html
sorear dwoldrich treats the JVM thing as some spur-of-the-moment decision
dwoldrich I don't know, can't say. Just know by the news stories I read
sorear we spent more than a year on that
so don't pretend you can make a difference
unless you want to spend a year arguing your point of view
au swarley: nqp-go is a fabulous idea, I'd love to see it. 05:26
swarley I think that using go routines would be a nifty optimization for certain areas
Concurrency without having to do anything special
au *nod* 05:27
dwoldrich lue: I have read that story before, is it inaccurate?
swarley I started writing some type definitions for the objects, its a lot more readable than I expected it to be
au yup. 05:28
...and when an implementor learns, we all learn :)
swarley I think its a plausible idea, I plan to give it some more work over the following week and see where it leads.
lue no, it's right. The part I take away from it is that sometimes rewrites are necessary (and also the last three subsystem rewrites didn't break everything)
au swarley++
swarley And since it has its own garbage collector and all of that fun stuff, no mess in that area. So I think its a reasonable candidate 05:29
05:30 Naga joined
sorear au: hey, I just noticed that you're not on the chanserv ACL since we rebuilt that. 05:30
want me to fix that now?
au sorear: sure, thx much 05:31
swarley: there's some early work (not quite nqp, but similar) at github.com/fglock/Perlito/blob/mas...Emitter.pm which may be informative if not dircetly reusable 05:32
05:32 am0c joined
swarley Oh. What license is nqp/rakudo under? 05:32
sorear artistic 2, isn't it?
au Artistic 2.
swarley Thanks au, I'll take a read
sorear OTOH we have the CLA stuff to deal with 05:33
we'd need to get a signed attestation from fglock that that code is usable
swarley Alright, I wanted to make sure before I added any code to my repo so that I could change my license, the one I have in there now is MPL 05:35
au swarley: as long as you're the sole author you can always change the license at any point :) 05:36
au has bypassed this CLA & license compatibility business just by using CC0 for everything :p
swarley Yup, I almost never have help on my projects. So I tend to jump around occasionally from GPL to MPL 05:37
moritz \o morning
sorear o/ moritzning
japhb swarley, IANAL, but for the Perl 6 community, if you're not going to go au++'s route of CC0, you might consider Artistic 2. You'll find it probably the most common Perl 6 license, and it is GPL-compatible. 05:38
moritz what's CCO? Creative COmmons? 05:39
japhb creativecommons.org/publicdomain/zero/1.0/
lue Beefier public domain, IIRC 05:40
moritz woah, quite permissive
au ensured compatibility 05:41
05:43 Naga left
japhb au, are you still active in the Perl 5 community? 05:44
(I spend most of my "community time" here, so I've kindof lost touch with who's who on the Perl 5 side of late ...) 05:45
au japhb: I still maintain some of the modules and pull reqs from time to time.
japhb Fair enough.
swarley I just read Artistic 2, I think I like that the most out of any so far
au swarley++ # awesome 05:46
swarley That's a really good licensing strategy
And I'm assuming that applies to your code as a whole, so no header? 05:47
japhb I'm looking at building a list of modules and functionality that should be higher-priority for us to port from Perl 5, and I'm looking for good ways to figure out what people are relying on and is considered relatively state of the art.
sorear what does au do these days? 05:48
japhb The first thing that occurred to me was the CPAN 100: Volatile 100, which lists most-depended-on modules.
But that ends up being a lot of testing and low-level infrastructure.
sorear
.oO( Task::Weaken )
moritz swarley: IANAL, but I think the common stance is that using a header file or a Perl module isn't a derivative work, so you can do it in any project
japhb sorear, Oh yeah, I'd seen that one before, but forgotten about it, since I didn't happen to need it. Annoying as heck that someone had to go and do that, though. 05:50
swarley So just a COPYING and LICENSE file at the root would do? 05:52
au sorear: differs from week to week. last week was mostly hacking with folks on g0v.tw ( g0v.tw/about.html ) to build a unicode chinese-chinese dictionary by scraping government data with legacy encoding.
japhb Anyway, I'm looking for suggestions of places to investigate. Clearly things like the Perl 5 core modules, common sense stuff like DBI, modules that are packaged under lots of Linux distros and/or Strawberry Perl ...
au sorear: after a hackathon with 200ish folks, there's a longish list of repos in 3du.tw/ with ports that covers pretty much all the mainstream development platforms, including moedict.tw/ :) 05:53
sorear: and it'll be something entirely different next week, likely.
japhb: modules that are in Strawberry or ActivePerl but not in Perl Core is a good list to start imvho. 05:58
japhb au, thank you, good point. 05:59
sorear japhb: Task::BeLike::DOY?
au it's one of your points, I just highlighted that particular one :p
japhb au: Nodnod, I meant, good interpretation/angle on it. I was thinking about comparing across packaging systems, as opposed to diffing the Windows ones against core. 06:00
sorear, Hmmm, interesting. Anyone else that did that whose taste you trust? 06:01
sorear metacpan knows about rafl, rjbs, and sartak 06:02
I don't know any of the others
(i.e. I don't know avar very well) 06:03
(also note: rafl is spelled FLORA on cpahn)
japhb Thank you, sorear
sorear it's probably not a coincidence that this set of people overlaps strongly with the moose core team 06:04
japhb Moose-ers seem to be more likely to be CPANers. :-) 06:05
06:09 jeffreykegler joined 06:13 MikeFair joined 06:33 FROGGS_ left 06:34 dayangkun left 06:38 jeffreykegler left 06:42 xinming left 06:44 kaleem joined 06:50 Vas joined 06:51 Vas left 06:54 xinming joined 06:56 jtpalmer_ left, jtpalmer joined 07:08 MikeFair left 07:19 tomaw is now known as 13WAAIFYZ, FROGGS joined
FROGGS morning 07:21
07:26 am0c left
sorear o/ FROGGS 07:33
FROGGS hi sorear
07:35 swarley left 07:36 swarley joined 07:43 dayangkun joined 08:01 rindolf joined
arnsholt o/ 08:04
phenny arnsholt: 07 Feb 23:32Z <jnthn> tell arnsholt I think some of the offset compuation wasn't quite right after all... Gave it a tweak in 263332d.
arnsholt That makes sense
08:10 domidumont joined
FROGGS morning arnsholt 08:15
sorear o/ arnsholt
masak morning, #perl6
moritz \o masak 08:16
08:19 domidumont left 08:20 domidumont joined
FROGGS hi masak 08:21
hi moritz
moritz \o FROGGS, arnsholt, *
08:30 hoelzro|away is now known as hoelzro 08:35 daxim joined 08:40 bloonix joined 09:06 kivutar joined 09:09 cosimo left 09:14 13WAAIFYZ is now known as tomaw 09:35 fhelmberger joined 09:37 dwoldrich left 09:39 kst` joined 09:40 dakkar joined, cosimo joined 09:41 kst left 09:44 dayangkun left 09:45 dayangkun_ joined, dayangkun_ left 09:46 Patterner left 09:50 Psyche^ joined, Psyche^ is now known as Patterner 10:32 kaleem left 10:52 sqirrel joined 10:56 SmokeMachine joined
masak I love Deferreds. api.jquery.com/category/deferred-object/ Does the GoF book bring them up at all? I did a quick skim through the list of patterns but didn't find anything like it. 10:58
11:00 sizz left 11:01 sizz joined 11:07 rindolf left
moritz
.oO( DerfeRedHerrings? )
 11:17
but yes, looks like quite a handy concept 11:18
masak I keep seeing variants of it everywhere. it's in the air somehow.
it goes well with all this async/eventy stuff that's also in the air these days.
arnsholt Has anyone else tried to build Rakudo HEAD with NQP HEAD? 11:22
I get a compile error when compiling RESTRICTED.setting: "Serialization Error: Unimplemented object type 'ArrayIterator' passed to write_ref" 11:23
11:25 PZt left 11:30 pmurias joined
arnsholt pmurias: jnthn++ has possibly updated some stuff in NQP that can help your problem. Do you have time to test it again? 11:31
11:34 Targen left
pmurias arnsholt: sure 11:38
11:46 kresike joined
kresike hello all you happy perl6 people 11:46
FROGGS hi kresike 11:47
kresike FROGGS, o/
masak kresike! \o/ 11:48
kresike masak, o/
11:52 crab2313 joined 12:00 swarley left 12:05 not_gerd joined
not_gerd r: ("x" x 100000) x 100000 12:05
p6eval rakudo 4fb07b: OUTPUT«Failed allocation of 1410065440 bytes␤Parrot VM: PANIC: Out of mem!␤C file src/gc/alloc_memory.c, line 105␤Parrot file (not available), line (not available)␤␤We highly suggest you notify the Parrot team if you have not been working on␤Parrot. Use parrotbug (locate…
not_gerd see lists.parrot.org/pipermail/parrot-d...07332.html 12:07
pmurias arnsholt: seems to fix the problem
not_gerd with the right parameters, this gives a segfault instead of an OOM error
arnsholt pmurias: Excellent!
not_gerd r: "xxx" x (3 x 111) 12:11
p6eval rakudo 4fb07b: OUTPUT«(signal SEGV)»
not_gerd \o/
12:14 skids left 12:22 odoacre left
nwc10 jnthn++ # multi-dispatch. And more code in a higher level language 12:40
12:42 skids joined
masak jnthn++ 12:44
12:48 dakkar left, dakkar joined 12:50 xy left 12:54 dakkar left, xy joined 12:58 dakkar joined 13:03 dakkar left 13:05 felher left, dakkar joined 13:06 not_gerd left
pmurias why do we test deletepos with a negative value? 13:09
13:09 crab2313 left
tadzik because it can happen 13:09
and I'd rather if we know what happens when someone tries
pmurias an exception should be thrown? 13:10
tadzik existspos works for negative, for example
Just Works [tm]
pmurias by testing stupid usage you force all the implementations to implement it 13:11
and make it seems acceptable for users to depend on it
tadzik it didn't have tests before, and no one really knows how is it actually used 13:13
nwc10 what *is* using it? ie - if you locally change it to throw an exception, how many tests fail?
tadzik so I prefered to replicate Parrot as closely as possible, and _have_ tests for that
masak wait, in what way does it just work?
tadzik what, existspos?
masak either of them.
tadzik nwc10: we'll never know until we check with Rakudo
masak I agree that it's not a good idea to support negative positions out-of-the-box.
nwc10 that is what I mean.
tadzik existspos counts from the back, like *-n
nwc10 change it locally, rebuild rakudo, run the spectests
what fails?
tadzik deletepos removes the 0th element, always 13:14
masak it's internal leakage like that that causes hard-to-predict bugs up at the user level.
yeah, I don't like it.
tadzik well, feel free to change it
masak I'm at $dayjob.
but if no-one has done it by tomorrow, I will.
nwc10 answer that, and you get to a pretty good idea of what (if anything) in the compiler need to change to avoid it
tadzik I've been reimplementing Parrot opcodes in Java, I wanted them to work alike, especially because they're not tested
masak nwc10: anything that does use it should be changed, too :) 13:15
tadzik well, maybe they're tested in Parrot
masak it's good that we have tests for it.
it's bad that we accept silly inputs.
nwc10 parrot opcodes directly? rather than something related to NQP and the JVM prototype?
tadzik nwc10: nqp::opcodes. Which were directly mapped to pir::opcodes 13:16
nwc10 ah OK
I still think the answer is to figure out whether NQP is even using the behaviour
pmurias it's not used by rakudo and nqp at all (according to ack)
tadzik look, maybe it's even better than they're changed. I didn't have any strong opinion, and wasn't really trying to come up with one. The task was "implement nqp::existpos and nqp::deletepos on JVM". I did. The same way that they works 13:17
pmurias the opcode itself
tadzik * worked
nwc10 oh, right. OK
tadzik I found the usage in rakudo, fwiw
masak jnthn says that existspos and deletepos are used in so few places, that "we don't need to worry [about the change being breaking] in this case".
tadzik at least existspos
pmurias tadzik: I'm not really blaming you, I just want to avoid porting crazy parrot behavior
tadzik and I have mixed feelings about deletepos anyway 13:18
pmurias: sure, that's reasonable
I'm all for reasonable behaviour :)
masak tadzik: I'm also not blaming you. good work. :)
tadzik heh, no offence taken :)
masak tadzik: you'll find that when you do good work, people will have opinions ;)
tadzik so, who wants to come up with sane existspos and deletepos tests? :) 13:19
masak tadzik: happy path: non-negative inputs. sad path: negative inputs. 13:21
there, done.
pmurias r: my $foo = [0,1,2,3,4];say $foo.elems;$foo[3] = Mu;say $foo.elems;$foo.delete(4);say $foo.elems; 13:22
p6eval rakudo 4fb07b: OUTPUT«5␤5␤3␤»
pmurias I have found a rakudo bug
13:25 araujo left
pmurias how do I add an issue on github? 13:28
tadzik it may be closed for Rakudo
email [email@hidden.address]
13:28 araujo joined
pmurias forgot his email password recently :( 13:29
my $foo = [0,1,2,3,4];say $foo.elems;$foo[4] = Mu;$foo[3] = Mu;say $foo.elems;$foo.delete(4);say $foo.elems; 13:30
masak we cannot help you there, sorry.
r: my $foo = [0,1,2,3,4];say $foo.elems;$foo[4] = Mu;$foo[3] = Mu;say $foo.elems;$foo.delete(4);say $foo.elems
p6eval rakudo 4fb07b: OUTPUT«5␤5␤3␤»
masak submits rakuodbug 13:31
pmurias++
tadzik don't know about nqp, but on JVM we keep a constant value in QRPA which says how many elements are there
it probably needs to be updated when we set Mu somewhere
13:32 PZt joined
moritz the problem is that if you delete an element that's not the last, it set to Mu 13:32
because we don't have an offchannel way to store an array element as deleted
pmurias no
moritz no? 13:33
tadzik we have to settle of something: do we count the number of defined elements in the array, or do we count "the longest streak" of elements?
pmurias the problem is that if you delete the last element all the undefined elements before it get deletedf
13:33 SmokeMachine left
moritz r: my @a = <a b c d>; @a.delete(2); say @a.perl 13:33
p6eval rakudo 4fb07b: OUTPUT«Array.new("a", "b", Any, "d")␤»
moritz tadzik: or do we simply keep the count separate, and don't do any magic
pmurias: I'm (nearly) sure there's a spectest that requires that 13:34
13:34 wk joined
pmurias r: class Foo {};my $foo = [0,1,2,3,4];say $foo.elems;$foo[4] = Foo;$foo[3] = Foo;say $foo.elems;$foo.delete(4);say $foo.elems; 13:34
p6eval rakudo 4fb07b: OUTPUT«5␤5␤3␤»
pmurias moritz: it seems to be a bug in the code
moritz fwiw I agree that's undesirable
but I can't quite imagine that it happened without any intention
13:35 SmokeMachine joined
masak adds this to the ticket 13:35
pmurias the spec seems to hold an distinction between elements that aren't in an array and those that are set to an undefined value 13:36
it should be very easy to fix using nqp::deletepos 13:39
dalek p: 55ae367 | (Paweł Murias)++ | t/nqp/59-nqpop.t:
Avoid testing for negative arguments to nqp::deletepos.
nwc10 did that actually get designed in? Or was it just adopted unchanged from Perl 5?
tadzik pmurias++ 13:40
13:41 cognominal joined
[Coke] hurls xkcd.com/1171/ for masak. 13:46
pmurias deletepos moves values?
tadzik yes
masak [Coke]: I... I think I know all the requisite memes involved to get this joke. 13:47
pmurias why?
masak [Coke]: but I'm not tickled by it. so maybe I'm missing something.
[Coke] masak: then I feel bad for you, son.
masak :P
[Coke] hits random, and gets this gem instead: xkcd.com/639/ 13:48
tadzik pmurias: uh, because Parrot folks wrote it that way
[Coke] so, yes, copied from p5. 13:49
pmurias so it's a duplicated splice, parrot--
masak [Coke]: :P
cognominal masak: about xkcd, may be the pop up explains it. I can't make sense of it either
masak cognominal: no, the pop up didn't explain it. so maybe I am missing something. 13:50
tadzik it's a mixture of 2 jokes
nwc10 It's not a good thing to copy. Sparse arrays should be sparse
and normal arrays not
tadzik 1) I have one problem. Oh, I'll use regexes, now I have two problems
2) I've got 99 problems but bitches ain't one
well, "jokes"
[Coke] memes.
tadzik ayep
[Coke] it's a mememe.
... wow.
cognominal now the jay-z makes sense knowyourmeme.com/memes/i-got-99-pro...h-aint-one 13:51
arnsholt 2) is from a song by Jay-Z
masak tadzik: yes, I knew both of those.
nwc10 otherwise, to be consistent, popping the last element of an array should shrink the array further if the new last element does not exist
arnsholt Thus the alt-text
13:51 mtk left 13:52 mtk joined
pmurias do Perl 5 arrays have a way to check for existance of an element? 13:54
nwc10 yes, hangon, writing a crazy test case demonstration
pasta.test-smoke.org/437 13:58
14:00 SmokeMachine left 14:01 dakkar left
pmurias thanks, I forgot to scroll down the perldoc manpage ;) 14:01
TimToady: is deleting elements from an Array using .delete something we want to keep, as it seems to be deprecated in Perl 5? 14:03
14:03 daxim left
nwc10 it's useful for sparse arrays 14:04
kind of necessary, really, to make them useful
moritz would tend to introduce a separate class for sparse arrays
our arrays are already resizable AND lazy 14:05
and making them sparse too... *shudder*
oh, and typable
14:05 FROGGS left
nwc10 I *think* (but don't hold me to this) that having sparse arrays as an option is useful. But not the only thing. Or the default. 14:06
masak +1 14:11
JavaScript does sparse arrays as the default. and then all people do with JavaScript all day is handle (dense) DOM lists of child elements. d'oh.
14:12 SamuraiJack left
arnsholt Oooooh, Dreamfall chapters kickstarter... I might actually back my first Kickstarter now 14:18
14:19 PacoAir joined 14:20 dakkar joined 14:24 SmokeMachine joined, fhelmberger left 14:25 SmokeMachine left
masak arnsholt: what's Dreamfall? 14:31
arnsholt The sequel to The longest journey 14:32
I really liked TLJ, and Dreamfall was intriguing (but flawed) as well 14:33
Oh, and they're computer games 14:34
masak oh 14:38
14:39 FROGGS[mobile] joined 14:40 bluescreen10 joined 14:45 stopbit joined 14:46 mikemol left 14:51 fgomez left 14:56 mikemol joined 15:08 wk left 15:15 dd070 joined
dd070 is it ready ? 15:16
tadzik my tea? Not anymore, I drank it all. What is that you're asking about? :)
dd070 perl6
tadzik the implementations are not quite finished 15:17
they're quite usable though
having, what, 20 or 30 monthly releases already
go on and try it
r: say "Hi dd070 "
p6eval rakudo 4fb07b: OUTPUT«Hi dd070 ␤»
dd070 cool 15:18
tadzik if that's any measure, we're passing about 24 000 tests of the official test suite
there are some things still missing: proper threads, proper async IO, mature module ecosystem 15:19
performance too, although that's being worked on
but it's suitable for playing around and writing some stuff in it
dd070 initially I hated implicit 'use strict' and 'use warnings'. but I would love it now. 15:20
its good in quick-dirty scripts too. 15:21
masak dd070: it's supposed to be off by default on the command line. 15:22
dd070: no implementation does that yet, though.
dd070: I'm never quite clear what people mean when they ask "is it ready". can I test a thing with you? I want to show you a document, and you can tell me if it answers your question better than just "yes" or "no". 15:23
dd070: gist.github.com/masak/43f95c5ba980dfee8cee 15:24
is that helpful?
15:24 mtk left
dd070 yes. 15:26
masak \o/
dd070: people have been working for years on Perl 6. quite a bit of things are there already. but it's not, you know, something I'd say "sure, you can put it into production" to people about, without lots of qualifiers. 15:27
it goes without saying we're working towards that point. it takes a bit of time to get all of the pieces into place.
dalek kudo-js: e56716e | (Paweł Murias)++ | / (2 files):
Add nqp-jvm-prep as a submodule.
 15:28
kudo-js: 3f5c935 | (Paweł Murias)++ | .gitignore:
update .gitignore
kudo-js: 3800636 | (Paweł Murias)++ | run_tests:
Add passing qast_relational.t to run_tests.
kudo-js: 0ad92a1 | (Paweł Murias)++ | / (3 files):
Pass an updated t/qast_string.t directly from nqp-jvm-prep.
15:28 am0c joined
dalek kudo-js: 1b5f733 | (Paweł Murias)++ | / (3 files):
Pass an updated t/qast_math.t directly from nqp-jvm-prep.
 15:31
15:31 dd070 left, pupoque joined
masak dd070: come back soon! 15:33
[Coke] perl6: tuits welcome. 15:34
p6eval niecza v24-20-g93138cc: OUTPUT«===SORRY!===␤␤Confused at /tmp/3HBejTPyWT line 1:␤------> tuits welcome⏏.␤␤Undeclared routines:␤ 'tuits' used at line 1␤ 'welcome' used at line 1␤␤Parse failed␤␤»
..rakudo 4fb07b: OUTPUT«===SORRY!===␤Confused␤at /tmp/ffcmC7SMFl:1␤------> tuits welcome.⏏<EOL>␤ expecting any of:␤ dotty method or postfix␤»
[Coke] perl6: our $marketing is bot-confusing;
p6eval niecza v24-20-g93138cc: OUTPUT«===SORRY!===␤␤Trait bot-confusing not available on variables at /tmp/Y0bOXea32w line 1:␤------> our $marketing is bot-confusing⏏;␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1…
..rakudo 4fb07b: ( no output )
[Coke] snrks.
15:36 cognominal left
masak [Coke]: that's what you get for calling Perl 6 "perl6" :P 15:36
15:37 SmokeMachine joined 15:43 SmokeMachine left, dakkar left, SmokeMachine joined 15:44 domidumont left
[Coke] We have so many marketing problems, perl6 vs. Perl&nbsp;6 is the least of my worries. ;) 15:44
15:47 SmokeMac_ joined, SmokeMachine left
cxreg jnthn seems to be tearing through that jvm nqp port 15:47
masak aye.
[Coke] A thought about parrot vs. perl6, btw. I'd like to see both projects succeed, but I don't think that in order for either to succeed, the other must, and I don't think it's reasonable that tuits spent on one project are fungible to the other.
arnsholt The idea that tuits are fungible in general annoys me alot 15:48
s/alot/a lot/
[Coke] I'm not suggesting that's the cause of any current discussion issues, but I've seen it historically. (when hacking on partcl, e.g. I often wanted to USE parrot, not figure out how to fix it.) 15:49
arnsholt There are so many reasons why a person decides to spend his spare time on a particular project, most of them probably non-technical, that it's hard to see them as fungible in any sensible way I think
[Coke] jvm/nqp... aye. I keep meaning to catch up and then there's 50 more commits.
nwc10 when hacking on PONIE, I most definately didn't want to have to hack on parrot. I'd not budgeted time for that. 15:50
15:51 ggoebel_ joined, FROGGS joined 15:55 benabik joined 15:57 kivutar left
atrodo I think it's safe to say people want to use parrot, but not many people want to hack on parrot 15:58
benabik Hacking on parrot is a thankless job.
atrodo benabik: it really can be 15:59
[Coke] agreed.
15:59 SmokeMac_ left
[Coke] ... ew, I just had a horrible idea. 16:00
kresike bye folks
16:00 kresike left
benabik Have to find a new project to work on this summer. Given the complete lack of activity, I'm not hopeful for Parrot being in GSoC. 16:00
nwc10 hauge grants? 16:01
atrodo [Coke]: I've had a few of those past couple days too 16:02
pmurias nwc10: for parrot hacking?
masak atrodo: I'm curious what the context is of twitter.com/atrodo/status/299620421496627200 (and if there's something I/we can do to improve)
tadzik I doubt that the hague grants will be granted for parrot hacking 16:03
[Coke] tadzik: there are hague grants, regular GC grants...
tadzik sure
[Coke] PaFo has no infrastructure for getting/disbursing grants, so going through TPF is not a horrible idea. 16:04
You'd just need a good grant proposal with clear benefits to the community.
s/just//
benabik [Coke]: s/ for.*/./
atrodo masak: The context would be the drama that happened yesterday 16:05
[Coke] Having been on the board, I would disagree with the letter of your substitution, but agree with the spirit.
masak atrodo: yes, that much is clear. I don't like drama either. 16:06
atrodo: if you just meant "drama doesn't make me want to use Perl 6", than I understand you perfectly already.
atrodo masak: More or less, yes. 16:08
benabik I know it's early, but is #perl6 planning on GSoC? Really need to find something this summer to keep me occupied/fed. 16:09
masak atrodo: if only there were a way to make Perl 6 without causing heat online. I wish I knew a way.
pmurias would like to take part of GSoC too
tadzik me too, if that'd be possible 16:12
colomon what's the deadline for getting that organized? it seems like we ought to be able to find things for people to do...
I mean, if not core stuff, much-needed modules. 16:13
benabik Last year it was announced 2/4 and org deadline was 3/9
colomon so we need to get on the ball pronto.
benabik Well, AFAIK the announcement hasn't happened this year. 16:14
(yet)
colomon afk but definitely interested in this topic # time to snowblow the driveway!
pmurias what's the difference between QAST::Stmts and QAST::Stmt? 16:22
benabik One's plural?
pmurias me found the difference in the docs
benabik I think Stmt is designed to limit things like temporary generation for a single statement, while Stmts is a sequence of code. (Someone else can probably explain better.)
benabik is far more familiar with PAST than QAST. A little too familiar sometimes. 16:23
16:24 cognominal joined, domidumont joined
benabik ponders if VMKit could actually be a reasonable way to do a "custom" NQP VM. 16:26
Although their GC simultaneously impresses and scares me. It uses MMTk, which is written in Java, by translating it through their VMKit-JVM into native code. 16:28
16:33 hoelzro is now known as hoelzro|away 16:34 benabik left 16:38 benabik joined 16:46 SunilJoshi joined
pmurias benabik: don't we have a surplus of VMs? ;) 16:47
benabik pmurias: I'm a systems researcher? 16:48
I'd like to play with a VM that uses 6model as its core object system. I keep wondering if it would have the flexibility that Parrot didn't quite manage. 16:49
16:51 rindolf joined
arnsholt A 6model-based VM would indeed be interesting 16:51
PerlJam pmurias: there's always room for one more :) 16:52
16:55 jeffreykegler joined
rindolf Hi all. 16:56
PerlJam: Jammin!!! 16:57
16:57 SmokeMachine joined
FROGGS I hate SOAP ó.ò 16:57
16:58 tail_ joined
rindolf FROGGS: join the support group - www.shlomifish.org/open-source/anti/SOAP/ 16:58
tail_ hi
rindolf tail_: hi. 16:59
tail_ I have a problem
'cabal install pugs' failed on my win7 17:00
I found a messeage 'Perhaps you need to add `haskell98' to the build-depends in your .cabal file' 17:01
[Coke] tail_: Not sure if we can help, but: what's the error message?
ah.
dalek kudo-js: 7ccf1fd | (Paweł Murias)++ | / (4 files):
Pass a bunch of tests directly from nqp-jvm-prep.
kudo-js: 271e02f | (Paweł Murias)++ | runtime.js:
Remove p6_num.
kudo-js: 00de30f | (Paweł Murias)++ | / (4 files):
Pass t/qast_conditional.t. Switch from using p6_true() to .Bool().
kudo-js: 581b72b | (Paweł Murias)++ | / (2 files):
Handle QAST::Stmt properly.
kudo-js: 77be5f0 | (Paweł Murias)++ | run (2 files):
Pass qast_terms.t
[Coke] and if you do that does it work? 17:02
tail_ see pastebin.com/2y2f97Fv 17:03
[Coke] pastebin blocked here. 17:04
tail_ Oh, I'm sorry
17:05 hoelzro|away left 17:06 hoelzro|away joined 17:07 kaleem joined 17:09 felher joined
pmurias if I want to use some BSD 3 licence code in nqp-js what do I have to do to meet all the licence requirements? 17:09
17:10 domidumont left, domidumont joined, kaare_ joined 17:11 hoelzro|away left 17:12 hoelzro|away joined
pmurias maybe using a library would be better then copy and paste ;) 17:14
[Coke] easier, aye. 17:17
er, from a licensing perspective.
17:19 cognominal left, uvtc joined
FROGGS rindolf++ 17:22
rindolf FROGGS: :-)
17:22 Chillance joined 17:23 hoelzro|away left, hoelzro|away joined 17:26 cognominal joined
stevan_ jnthn++ # irclog.perlgeek.de/perl6/2013-02-07#i_6423586 thank you 17:29
17:33 cognominal left 17:38 kivutar joined
dalek gs.hs: d3a6617 | au++ | pugs-compat/src/Pugs/Compat/Posix.hs:
* Workaround #25 by disabling opendir() on Windows.
 17:38
gs.hs: d467adb | au++ | pugs-compat/ (2 files):
* 0.0.6.20130209.0 as CC0.
gs.hs: 50efb57 | au++ | / (2 files):
Merge branch 'master' of github.com:perl6/Pugs.hs
tail_ something went wrong with pugs-compat-0.0.6.20120717.204800 17:39
17:39 cognominal joined
au tail_: yeah that's github.com/perl6/Pugs.hs/issues/25 ; please "cabal update" and retry with 0.0.6.20130209.0 17:39
released about 120 seconds ago
tail_ !!
okey, trying now 17:40
au you're welcome, too :) please re-open the issue if you run into any problem and I'll check next lunar year (that is to say, less than 24 hours from now)
tail_ thank you. pugs-compat successfully installed 17:44
au woot 17:45
tail_++ for reminding me about this :) 17:46
-> dreamland, with a happy smile &
tail_ au++
pugs-compat is ok, but it still says "Pugs-6.2.13.20120717 failed during the configure step." 17:50
* Missing (or bad) header file: perl5/p5embed.h
uvtc Is someone working on a Perl 6 --> Javascript compiler?
benabik uvtc: pmurias seems to be doing something with rakudo-js (I think still very early stages) 17:52
uvtc benabik: thanks. 17:53
tail_ it seems EXTERN.h is missing
17:55 cognominal left
uvtc benabik: you mentioned some reasons on the 5th here about the difficulties with using LLVM for Perl 6. 17:58
irclog.perlgeek.de/perl6/2013-02-05#i_6416952
benabik uvtc: It's much like building something from scratch.
uvtc I was going to add an Implementations FAQ to the wiki. Ok if I add in those points you mention? 17:59
benabik uvtc: Sure.
uvtc thanks
[Coke] au++ random pugs commits! 18:01
uvtc Is github.com/perl6/Pugs.hs the official github home of Pugs?
18:02 SmokeMachine left
[Coke] as official as it gets, aye. 18:02
18:02 cognominal joined
uvtc Is it's purpose primarily experimental at this point? 18:03
18:03 SmokeMachine joined
uvtc I notice that it is no longer listed in the feature comparison matrix. 18:03
n/m 18:04
[Coke] it was experimental years ago. 18:05
uvtc understood
thanks
If someone wants to port Rakudo to another backend, what are the main pieces involved?
benabik QAST, rules engine, and REPRs? 18:06
uvtc Are those all parts of nqp?
[Coke] uvtc: port nqp.
benabik uvtc: Yes.
uvtc Cool. Thanks.
PerlJam uvtc: what other backend are you thinking about porting it to?
18:07 benabik left
uvtc I am not. But a few days ago someone (pmurias?) mentioned that maybe Perl 6 could use an "Implementations FAQ". So I'm quickly writing up something for the wiki, and asking questions now as I type. 18:07
FROGGS nqp is the "glue language", but there are still some parrot specific thing in rakudo itself (I'm thinking of IO::Socket::INET)
18:07 SamuraiJack joined
FROGGS ++uvtc 18:08
au tail_: try "cabal install Pugs --flags=-Perl5" to install without Perl5 bindings. it's possible your perl installation had no compatible EXTERN.h
cabal should actually retry that... not sure if Pugs.cabal can be tweaked to make that happen. pull requests welcome :) 18:09
-> zz for real &
uvtc REPRs == regexes?
PerlJam uvtc: no
uvtc What are REPRs?
18:10 SmokeMachine left
uvtc Sounds like something one would catch down by the river and fry up under the stars over a campfire. 18:10
PerlJam uvtc: REPRs are the "clostest to the metal" representation of things.
18:10 arlinius left
geekosaur cabal doesn't so much try things, in the sense of try the compile and then try it again on fail. it's just a constraint solver and if you can't describe the constraint directly in the cabal file (as when you need to do configure search for perl5 guts) you can't have it automatically set that flag 18:10
tail_ It works! 18:18
Thank you for helpful advice au++
[Coke] \o/
uvtc wiki.perl6.org/Implementations%20Faq 18:23
18:25 cognominal left, cognominal joined
tail_ good night (3:30 here) 18:28
18:28 tail_ left
uvtc After using a language on the JVM (other than Java), my two general complaints are: (1) slow start-up time, and (2) contant gentle pressure to use Java-equivalents of libs that you'd really rather use instead. Hoping these aren't too big of a deal with Rakudo-on-JVM. 18:29
Er, Rakudo+JVM. Rakudo⌁JVM 18:32
18:32 kst` is now known as kst 18:33 cognominal left, kaleem left
[Coke] .u ⌁ 18:33
phenny U+2301 ELECTRIC ARROW (⌁)
18:33 fgomez joined
uvtc I like that one. :) 18:34
18:34 cognominal joined
rjbs just wanted to pop in and say how much he loves you all. »hugs« 18:40
PerlJam hugme: hug rjbs 18:42
hugme hugs rjbs
PerlJam hugs by proxy :) 18:43
rjbs good call. I've been a bit sniffly, you don't want to catch anything
18:45 SamuraiJack left
[Coke] ~~ @ rjbs! 18:48
rjbs We going to see good p6 talks in Austin? 18:49
[Coke] Argh. I really wish I could convince work to pay for travel. :| 18:53
jnthn will likely be there, so yes. 18:54
19:05 rindolf left 19:06 benabik joined
diakopter laughs at dwoldrich in the backlog 19:08
uvtc s/dwoldrich/dwoldrich's comments/ 19:09
:)
rjbs [Coke]: Where are you?
19:11 japhb_ joined
[Coke] rjbs: near Albany, NY 19:11
the whole family drove to Houston a few years back, but I doubt that will happen this year. YAPC seems to have firmly settled on dates in the school year.
rjbs That's only 208 miles from here! Coming to ABE.pm this week?
moritz backlogs
sorear++ # "then I should tell you that #perl6 is now a place where we care about things that work"
[Coke] rjbs: ... chrome does not want me to search for ABE.pm in the omnibox. 19:12
rjbs [Coke]: abe.pm.org in Bethlehem, PA
[Coke] rjbs: not this week, but if you have a big event, sure, I can swing by. 19:14
rjbs We tend to just get beers and burgers, but I'll let you know. :-) Peter Martini, author of the p5 work on core sub sigs, may be here this week. 19:15
uvtc (Actually, it's not very nice to laugh at folks' comments either, unless they're meant to be funny. Anyhow.) 19:16
19:17 thou_ is now known as thou 19:18 vmspb joined, Targen joined 19:24 cogno joined
diakopter uvtc: lol? 19:24
uvtc I just meant that it seemed that maybe dwoldrich had other problems. That's all. And that I shouldn't have added that smiley above. 19:25
19:26 kivutar left
arnsholt Argh. Building Rakudo HEAD with NQP HEAD fails on both my work machine and laptop. Has anyone else gotten that combo to work? 19:26
19:31 kivutar joined
nwc10 no, I meant Hauge Grants for hacking on something Perl 6 instead of GSoC for hacking on Parrot. (If that's fun) 19:32
19:32 am0c left
uvtc In pmichaud's "perl 6 lists and arrays" talk from June 2012 (it was actually a two-parter), he mentioned near the beginning (3:20) that Synopsis 7 & 8 are very wrong as of that date. 19:37
In fact, he says "don't read them", and that they'll just confuse you.
Have those issues been cleared up since then?
This talk: www.youtube.com/watch?v=mrfK1dtyD14
moritz uvtc: I don't think so
19:38 Targen left
uvtc Has the design been straightened out, just not the spec? 19:38
moritz uvtc: pmichaud has a grant for rewriting S07, but he got distracted by lots of real-life stuff
19:38 cogno left
uvtc I understand about pmichaud. 19:38
Has the design/implementation settled down though? (that is, is it just the spec that is still behind?) 19:39
TimToady we were very clear here that we expected him to prioritize his family over Perl 6
uvtc Absolutely.
TimToady it's very close to right; there may be a few glitches yet in self-referential lists 19:40
19:42 cognominal left, FROGGS[mobile] left
uvtc TimToady, at perl6.org/compilers/features , I see that "LoL" (list of lists, I presume) is marked yellow. Is that one of the bits still being designed? 19:42
Or is it just not implemented yet? 19:43
TimToady partially implemented
19:43 cognominal joined
TimToady binding to ** works, but lol as a verb isn't there yet, last I knew 19:43
nr: say ((1..2),(3..4)).lol.perl 19:44
p6eval niecza v24-20-g93138cc: OUTPUT«Unhandled exception: Unable to resolve method lol in type Parcel␤ at /tmp/BUDGjTfIKt line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4227 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4228 (module-CORE @ 580) ␤ at /home…
..rakudo 4fb07b: OUTPUT«LoL.new(1..2, 3..4)␤»
uvtc Sorry, right. I should've written s/not implemented/only partially implemented/ above.
TimToady I guess it's there-er than I thought
nr: say ((1..2),(3..4)).lol.[1].perl 19:45
p6eval rakudo 4fb07b: OUTPUT«3..4␤»
..niecza v24-20-g93138cc: OUTPUT«Unhandled exception: Unable to resolve method lol in type Parcel␤ at /tmp/I_FAHGgUoW line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4227 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4228 (module-CORE @ 580) ␤ at /home…
TimToady nr: say ((1..2),(3..4)).lol.[1].WHAT
p6eval niecza v24-20-g93138cc: OUTPUT«Unhandled exception: Unable to resolve method lol in type Parcel␤ at /tmp/E8h7rwLnqq line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4227 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4228 (module-CORE @ 580) ␤ at /home…
..rakudo 4fb07b: OUTPUT«Range()␤»
TimToady nr: say ((<a b c>),(<d e f>)).lol.[1].WHAT
p6eval rakudo 4fb07b: OUTPUT«Parcel()␤»
..niecza v24-20-g93138cc: OUTPUT«Unhandled exception: Unable to resolve method lol in type Parcel␤ at /tmp/R0Pmpq3fY9 line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4227 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4228 (module-CORE @ 580) ␤ at /home…
19:45 xenoterracide left
TimToady okay, enough spamm 19:45
ing 19:46
uvtc Where can I find the docs for `lol` at doc.perl6.org/ ?
Didn't see them under doc.perl6.org/type/List .
19:46 xenoterracide joined 19:47 xenoterracide left
TimToady S29:368, for one 19:47
uvtc Thanks.
FROGGS TimToady: btw, both niecza and rakudo have the same .gist for Matches
TimToady but it lies slightly
insofar as we don't have a Seq type
FROGGS: I saw that, FROGGS++ 19:48
FROGGS it's awesome that one can almost copy+paste from one to the other
working on niecza is pretty nice though, since it compiles so damn fast
19:48 arlinius joined
TimToady sorear++ thinks it's too slow :) 19:49
FROGGS thats the only thing I dont like about rakudo, it takes minutes to check changes to code when you have to recompile nqp and rakudo
heh
TimToady well, the nice thing about eating one's own dogfood is that when the dogs are happier, so are you 19:51
19:52 cognominal left
rjbs rakudo: my $x = sub {}; say $x.^methods 19:54
p6eval rakudo 4fb07b: OUTPUT«No such method 'gist' for invocant of type 'Sub'␤ in method gist at src/gen/CORE.setting:5128␤ in sub say at src/gen/CORE.setting:7595␤ in block at /tmp/u4sD3jGU1D:1␤␤»
19:54 SunilJoshi left
rjbs wonders where his rakudo checkout went. 19:54
benabik r: my $x = sub {}; say $x.^methods.map: *.name
p6eval rakudo 4fb07b: OUTPUT«Unmarshallable foreign language value passed for parameter '$x0'␤ in block at /tmp/GZgng_XLLx:1␤␤»
FROGGS there are rt tickets for that 19:55
rjbs FROGGS: yeah, I filed it
I wanted to see if it was still an issue, and I can't find a local rakudo, so I cheated and asked the bot. :)
rjbs is compiling in the background now. 19:56
FROGGS :o) 19:57
20:01 fgomez left, atrodo_ joined, atrodo_ left
tadzik PLPW registration is now open! \o/ 20:02
join the fun at act.yapc.eu/plpw2013/
20:03 cognominal joined 20:06 bruges_ joined 20:08 bruges left 20:10 fgomez joined 20:12 cognominal left
benabik gets segfaults compiling NQP. :-( 20:14
FROGGS tadzik too 20:15
err
arnsholt too
20:17 cognominal joined
japhb_ pulls and tries on a 32-bit Intel Linux box 20:18
dalek c: c57c8d0 | (Gabor Szabo)++ | index.pl:
add listing of the words
 20:21
[Coke] hugme: hug chromatic 20:26
hugme hugs chromatic and blushes
[Coke] ... that's odd.
benabik hugme has a crush?
tadzik what, you wouldn't blush?
20:28 pupoque left 20:29 pupoque joined
uvtc hugme: hug tadzik 20:29
hugme hugs tadzik
uvtc Hm.
There is more to hugme than meets the eye. 20:30
tadzik the real beaty is not screen deep
* beauty
uvtc Beauty is in the eye of the behugger.
rjbs Hooray, I have a perl6 again.
FROGGS \o/
[Coke] rjbs: I challenge you to add the behugger to the monster manual. 20:31
rjbs One sec.
japhb_ benabik, FROGGS, arnsholt: I got a successful build of NQP on 32-bit Intel Linux. FWIW, the command I used was: make realclean; rm -rf install; git clean -dXf; perl Configure.pl --gen-parrot; make; make test; make install 20:32
AKA "Paranoia R Us"
20:33 cognominal left
[Coke] if you're reall paranoid, you also need "rm parrot nqp" in there with install. 20:33
since git clean doesn't do that. 20:34
benabik builds Parrot separately.
[Coke] (ah, unless -X does when -x doesn't)
japhb_ [Coke], "Oh god I'm being crushed to death ... but man, it feels so fuzzy and soft and warm .... aaaugghhzzcckkk."
rjbs [Coke]: www.evernote.com/shard/s2/sh/e5ae1...9159eb3379
[Coke] rjbs: (perl6 again). and don't you feel better now?
rjbs [Coke]: Mostly confused. :) 20:35
20:35 domidumont left
[Coke] hit dice? isn't that old school? 20:35
or is it also new school? I stopped playing in '89. :( 20:36
rjbs [Coke]: I play the 1981 rules.
[Coke]: (I also play 4E, but with less gusto.)
benabik Monsters still have hit dice.
rjbs benabik: Not in 4E they don't.
[Coke]: You can get the 1981 Basic rules for $5 now at dndclassics.com! :)
benabik rjbs: Ah. I went 3->3.5->Pathfinder, not to 4. :-/
[Coke] wishes he had not sold his old school books for cash. 20:37
20:37 cognominal joined
tadzik did you learn d&d in school? 20:38
japhb_ Pathfinder is the first d20 game I've played since the original AD&D (and old skool D&D before that).
20:38 jeffreykegler left
japhb_ I played a lot of GURPS in college 20:38
[Coke] AD&D, paranoia, call of c'thulhu. even larped once. 20:39
moritz loved paranoia 20:40
colomon AD&D 1 -> GURPS -> Amber Diceless
tadzik play-perl.org/about looks cool 20:42
20:44 vmspb left
rjbs /win/win 77 20:44
Oops.
20:44 cognominal left 20:45 jeffreykegler joined
japhb_
.oO( It's a /win/win! )
 20:45
moritz arnsholt: fwiw I can't compile rakudo on newest nqp either 20:47
also while compiling the restricted setting
though compiling lib/Test.pir produces the same error 20:48
benabik gets segfault Parrot/NQP HEAD on 'nqp src/stage2/gen/NQPP5QRegex.pm'
arnsholt Right, so it's probably not something particular to my setup then 20:49
We should probably ping jnthn about it when he's around
20:54 benabik left
FROGGS fwiw: nqp 2013.01-95-ga0bba92 with rakudo Sonoma-109-g4fb07b1 is fine 20:56
jnthn good evening o/ 20:57
FROGGS hi jnthn
masak jnthn! \o/
tadzik jnthn!
jnthn Yes, I noticed the Rakudo not building on latest NQP last night. At first I thought it was the allocation mis-calcs, but I patched that and...no pony.
tadzik did you know you can sign up for PLPW already? :) 20:58
20:59 fgomez left
jnthn tadzik: omg omg no I didn't! 20:59
uvtc hehehe
jnthn tadzik: url? :)
tadzik act.yapc.eu/plpw2013/
last minute, before we run out of seats :P 21:00
uvtc jnthn: be the first one. It'll put you in the ... what's the term... the "first position" in the workshop.
In the ...
tadzik POLE POSITION! \o/
uvtc Pole Position
!
Oh man.
tadzik uvtc: looks like I got to Pole Position first ;P
uvtc hahaha
masak why is it that every time au++ helps someone on the channel, she does it in such an awesome way? how can the rest of us be awesome like au?
TimToady some of us are called to be LTA to make the awesome people stand out more :) 21:02
21:03 cognominal joined
jnthn tadzik: I signed up :) 21:03
tadzik awesome! \o/
jnthn Moja prezentacja bude po angielsku, though :P
PerlJam masak: strength through diversity! :) 21:04
tadzik będzie :) That's cool :)
masak moritz: if we were to set as a goal to play Paranoia when we next meet, would that goal be realizable? :)
jnthn tadzik: heh, I guess Polish entirely from a poor recollection of Slovak :)
tadzik they're close enough :) 21:05
masak Moja prezentacja tiez bude po angielsku...
tadzik masak: oh yes, rpg! 21:06
masak jnthn: where does the adverb go?
jnthn masak: There seems OK to me... :)
masak: Though I'm pretty sure it coulda gone on the end too. 21:07
masak so typical of a cased language. they don't really care :P
arnsholt jnthn: From the error message I get compiling, it looks like Rakudo tries to serialize an object NQP doesn't like? 21:08
masak Latin is worse, though. they flaunt their nonchalance of word order.
jnthn Well, just means you get to use word order for nuance. :)
I never got to the level of being able to do that terribly well in Slovak. Maybe I will with $slavic_language.next :) 21:09
21:09 fgomez joined
jnthn arnsholt: Yes, but the object it's getting passed is...bizzare. 21:09
21:09 jeffreykegler left
arnsholt Oh, dang 21:09
masak jnthn: don't tell me... Bulgarian!? say it ain't so! :P
jnthn masak: :P 21:10
tadzik Bulgarian is cool :)
jnthn No, but I'd like to make an effort to learn one of the languages from that sub-family.
masak tadzik: it is, actually. 21:11
timotimo it is basically advanced Russian, no?
jnthn Russian often feels like pretty advanced Russian to me :P 21:12
timotimo would evolved be a better term?
masak timotimo: not really.
geekosaur "like Russian, only more so" 21:13
timotimo i talked to a Bulgarian and she told me Russian sounds like archaic Bulgarian to their ears
21:13 pupoque left
[Coke] ooh, let's learn bulgarian! I already have a starter book and some cd's. "ciao!" 21:13
21:14 swarley joined
swarley I cannot seem to answer this physics problem 21:14
21:15 cognominal__ joined
[Coke] does it involve pushing on a rope 21:15
?
masak :P
swarley No :( electrostatics
21:15 cognominal left
masak swarley: I always get static shocks from a metal staircase in a $work building... ;) 21:16
that's also a physics problem, I think.
jnthn hah, I was thinking exactly the same :)
masak jnthn: that's because you get them too! :P
jnthn I think that's a problem of them using a HUGE piece of iron, then people rubbing their huge winter coats against it :) 21:17
The glass stairs probably also don't help... :)
swarley Two point charges, q1 and q2 are placed 0.30m apart on the x-axis, as show in the figure above. Charge q1 has a value of -3e-9 C. The net electric field at point P is zero. Calculate the magnitude and charge of q2.
masak swarley: where's point P?
swarley 0.1m left of point q1 21:18
masak yes, then it's solvable.
[Coke] hey, don't we have a community rule about not doing people's homework? ;)
masak I didn't give him the solution! :P
tadzik physics? Do help him :)
TimToady since it's really just electromagnetism...
tadzik I remember my struggle with physics
jnthn -3e-9 C...that's some really cold weather.
swarley I thought I got the value for q2, but in theory if I got it right then if I calculate it the electric field should be zero
moritz masak: re paranoia: possible
swarley but it's 0.000005
masak jnthn: with wind chill, it's 20 degrees below absolute zero! 21:19
uvtc swarley: figure out the field at P due to q1.
masak yeah, uvtc has it. field.total = field.q1 + field.q2
swarley uvtc; I started with Eq1 = Eq2 then culoumbs law, k(-3e-9)/(0.1)^2 = k(q2)/(0.4)^2 21:20
k cancels, simplify etc
I came down to 4.8e-8 nC
err
-nC 21:21
But if I try to prove that with the electric field equaling zero it doesn't turn out as expected
uvtc Eq1 = -Eq2
swarley Electric field of q1 must be equal and opposite of q2 at point P 21:22
since they cancel
Jesus, you don't need to take a standard physics class before AP physics they said, he's only a little bit of a hard teacher they said 21:23
uvtc hahaha
swarley Seriously, his motto is "If I make it harder than the exam, you won't have any problems"
masak I like him already. 21:24
moritz swarley: just remember that the electric field decreases quadratically with distance
swarley I know, because of radius squared
Inverse proportionality
inverse square*
moritz so you just have to square the ratio of distances to get the ratio of charges to make them cancel out
swarley but that doesn't change that they cancel
ohh 21:25
masak it feels to me that you need to plug your values into the same equation twice: the one that relates point charge, electric field, and distance.
I would even write up a nice solution in Perl 6, if I could think of a nice way to run the same formula both backwards and forwards :)
hrm, I guess I could just implement both directions... 21:26
arnsholt Maybe we need a computer algebra system O:)
swarley I have one, I really do love math, so I spoiled myself to a TI-Nspire CX CAS 21:27
with a*
moritz was good as physics as long as constant factors could be ignored :-)
swarley I'm good with physics 2 weeks after the unit has finished. It's just that our teacher assumes that we have taken calculus already
And I'm a calculus student stuck in a trig students curriculum.. 21:28
Oh man, seeing vectors for the first time in kinematics, that was not even a learning curve 21:29
That was a learning wall
jnthn Seems I can go all the way back before the multi-dispatch changes and still see problems with serializatin. 21:30
21:30 rking left
swarley I guess it was for the best though, quick way to jam a lot of math into your brain 21:30
jnthn *serialization
21:31 uvtc left
arnsholt Hmm. So quite possibly a bug somewhere in my branch code? 21:32
masak swarley: gist.github.com/masak/4742101 21:33
swarley: didn't need much math at all ;)
r: say -3e-9 * -sqrt(2)
p6eval rakudo 4fb07b: OUTPUT«4.24264068711929e-09␤»
masak swarley: (was this review helpful to you? yes/no) 21:36
21:37 spider-mario joined 21:39 kivutar left 21:40 bluescreen10 left 21:41 donaldh joined
swarley oh, sorry. I was finishing another one 21:41
Let me take a look
jnthn arnsholt: I have got a case in a commit right after we merged it when Rakudo does work, but if I --target=pir on t/nqp/55-multimethod.t it explodes with exactly the same kind of error. 21:42
swarley and 1/(4 pi e0) is 9e9
r: say 9e9 21:43
p6eval rakudo 4fb07b: OUTPUT«9000000000␤»
masak swarley: you don't need to know what 1/(4 pi e0) is. 21:44
swarley r: say 9e9 * ((-3e-9)/(0.1**2))
p6eval rakudo 4fb07b: OUTPUT«-2700␤»
masak all you're interested in is q2.
* moritz was good as physics as long as constant factors could be ignored :-)
swarley oh, i see what you're saying now 21:45
arnsholt jnthn: Right. I'll provisionally blame myself, in that case
masak in this case, the constant factors could be ignored. :)
jnthn arnsholt: Well, doing a built of before any of the merging for now
That should help answer it 21:46
arnsholt Yeah, definitely
masak swarley: basically all I'm using in the end is "it's an inverse square law -- if it's twice the distance, the charge needs to go up by a factor of sqrt(2)"
swarley Yeah, I understand it now. I just was caught up in a slur of variables
jnthn arnsholt: Hm, seems not... 21:47
masak the trick to solving these problems is to abstract away the unnecessary parts :)
swarley I swear I'm not bad at physics, I got one of the headers
errr
Harder ones
Sorry, someone asked me a question as I was typing
arnsholt jnthn: No error pre-merge, or error pre-merge?
jnthn arnsholt: Error pre-merge. Very odd.
masak swarley: don't spend time classifying yourself. just practice the problems until they feel easy :) 21:48
moritz swarley: perlpunks.de/paste/show/51157269.9f7.1a4 # there's the math, formally
swarley I got one of the variable only ones 100% correct and I was so excited lol. Those usually have the lowest average points earned
masak er, moritz is right, I'm wrong. 21:49
the factor is -4, not -sqrt(2). of course.
arnsholt jnthn: Weird, weird 21:50
swarley OH! so I actually had the right answer when I rewrote it :)
masak gist.github.com/masak/4742101 updated.
moritz++
arnsholt But bedtime now
swarley I just said 'screw physics' let's treat this as a proportionality
arnsholt See you guys tomorrow o/
arnsholt &
masak arnsholt: \o
swarley o/
21:51 fgomez left
jnthn o/ arnsholt 21:52
masak I should have checked myself somewhere along the way. the field strength is divided by 4 every time we double the distance. d'oh :)
21:53 xenoterracide joined 21:54 cognominal joined, cognominal__ left
moritz wonders if there's a weird but beautiful way to solve that problem with Gauss' theorem 21:57
pmurias jnthn: what do we need deletepos for? 21:58
jnthn Array.delete in Perl 6.
21:58 cognominal left
jnthn Though I wonder if we couldn't instead use splice or something... 21:58
pmurias it's not used in rakudo atm 21:59
jnthn: a few people were unsure if we want to keep the current Array.delete behavior 22:00
jnthn: as it's copied over deprecated perl5 behavior 22:01
masak moritz: well, it looks that way. Wikipedia says that Coloumb's law and Gauss' law are equivalent.
moritz masak: yes, just very different formulations
masak I'm looking at the math and realizing that I've forgotten too much of this to easily take it in. 22:02
pmurias jnthn: pasta.test-smoke.org/437 - how the current way is crazy
22:03 cognominal__ joined
pmurias and the way nqp::deletepos is implemented in nqp-parrot is not really suitable for implementing Array.delete as it moves the elements 22:04
22:08 fgomez joined 22:20 cognominal__ left 22:21 cognominal joined, kaare_ left
dalek kudo-js: a017f5a | (Paweł Murias)++ | run (2 files):
Implement nqp::sha1, pass qast_serialization.t.
 22:22
jnthn pmurias: Sorry, very tied up debugging a weird problem here... 22:26
pmurias: Anyway, if .delete's current behavior is crazy, I'm most certainly not attached to it.
22:28 donaldh left
rurban I'm trying to design a better perl macro system. `statements` looks better than quasi{statements} or {{{statements}}}. qx should be used instead of `` for systems commands. macro declaration arguments should be possible structures ("destructuring-bind"), scalar, lists ie. () or hashes i.e. blocks. See e.g. perl11.org/p2/p2-extensions.html 22:29
@`` for late evaluation in list-context. 22:30
flussence I wouldn't complain if shell-`` went away - I can never remember what function it maps to... 22:34
tadzik shell `` is gone 22:35
masak rurban: by the use of 'expr' and 'block' there, are you conflating rule names with parameter names?
flussence oh well, you didn't see me complain :)
pmurias rurban: it would be great to be able to specify the language for quasi quotes 22:36
rurban well, {block} could be matching a lexed block or a hash
masak rurban: what exactly can occur within the `` ?
rurban I want to be able to matcbh both, blocks or hashes (of hashes). That's where I'm currently struggling 22:37
(in lisp it was much easier)
masak rurban: it doesn't say {block} on that page, it says '{'block'}' 22:38
rurban That's why I'm thinking a macro definition should have the same definition as in the parser definition. parsed at compile-time. macro while '('expr')' '{'block'}' {...}
masak rurban: I find the syntax confusing. things are not marked out as rules in any way.
rurban {block} might want to match a hash key named block 22:39
confusing, yes
masak rurban: ah, well. looks like the beginning of something, but I don't see it working as-is. 22:40
what if you wanted a macro with two blocks in it?
like a loop with a continue block, for example?
pmurias wouldn't it make sense to optimize for clarity, and not for conciseness
masak pmurias: yes. the kind of clarity that makes things stand out. 22:41
rurban I'd need a public API for the lexer, so that the structural macro arguments are well defined.
masak pmurias: that's why we have {{{$ast}}} currently. it stands out.
22:41 sqirrel left 22:43 pmichaud joined 22:44 stevan__ joined
pmichaud good afternoon, #perl6 22:44
diakopter hi :D
jnthn o/ pmichaud
pmurias hi 22:45
masak pmichaud! \o/
pmichaud o/ jnthn
22:45 stevan_ left
PerlJam pmichaud: greetings 22:45
pmichaud interesting discussion taking place on #parrot; anyone have anything they want me to know/add? 22:49
swarley gist.github.com/swarley/4742592 22:51
\o/
PerlJam swarley: you're making a Go backend for rakudo? 22:52
swarley nqp
but yea
PerlJam swarley++ cool.
swarley I need to rethink the base object 22:53
pmichaud swarley++ awesome!
swarley I'm not really sure how perl6 objects are, my brain wants them to be like ruby objects. I looked at nqp-jvm's SixModelObject.java but that has no real information other than the STable, and I pretty much already had that. 22:54
22:58 cognominal left
swarley I can't wait to see regular NQP without the parrot code in it 22:58
That would make it a lot easier to read without having to look back and forth between the source trees
jnthn swarley: The main thing to realize about 6model is that it does representation polymorphism. 23:00
swarley: There's not a single "how to represent an object"
There's just an STable pointer and you follow that to the REPR.
swarley Oh, I understand that.
I'm not planning on restricting it, or at least if i did I didn't mean to
jnthn swarley: I didn't think you were. :) I was just trying to explain it. :) 23:01
pmurias swarley: have you seen nqp/docs/serialization_format.markdown? 23:02
swarley Oh okay, I would love to hear an explanation or read a paper on the subject from you. I'm quite certain you know exactly what you're talking about
jnthn P6opaque is perhaps the most interesting thing to look at but I'm afraid any implementation of it is quite...crazy.
In C you do it by calculating a struct on the fly.
swarley pmurias; I have, but right now I'm just trying to get a basic object system that I can make a test suite for to start
jnthn On the JVM it's done by calculating a class on the fly...
I'm not quite sure how it is in JavaScript :)
swarley I think a lot of things I'm going to just end up making interfaces for.
What is the opaque exactly?
pmurias jnthn: it maps cleanly to the javascript object system 23:04
23:04 cognominal joined
pmurias which is really primitive 23:04
jnthn swarley: It's the representation the majority of objects in NQP and Perl 6 use
swarley Oh okay, so it's a higher level object representation? 23:05
23:05 LordVorp joined
jnthn pmurias: Nice :) 23:05
swarley: It's a representation that knows how to lay out object attributes.
pmurias jnthn: for each stable I just create an object which is a prototype for it's all instances 23:06
swarley Oh, okay. I'm glad you said that now. Because the lowest level object I just prototyped has an attributes array 23:07
err, map
jnthn swarley: Yeah, the only thing that's common to everything is the STable and an SC. 23:09
23:09 jferrero joined, cognominal left
swarley Alright. I'm omiting the sc for now, just to get basic functionality 23:09
jnthn *nod*
swarley To make sure everything works conceptually
pmurias jnthn: the .symtable attributes of blocks is no longer important? 23:12
swarley See, a really cool thing about doing this with go is the ability of using goroutines
jnthn pmurias: It's still used
swarley This is from golang.org/doc/effective_go.html 23:13
jnthn pmurias: And really it's just exposing something already there
swarley They're called goroutines because the existing terms—threads, coroutines, processes, and so on—convey inaccurate connotations. A goroutine has a simple model: it is a function executing concurrently with other goroutines in the same address space. It is lightweight, costing little more than the allocation of stack space. And the stacks start small, so they are cheap, and grow by allocating (and freeing) heap storage as required. Goroutin
es are multiplexed onto multiple OS threads so if one should block, such as while waiting for I/O, others continue to run. Their design hides many of the complexities of thread creation and management.
Oh my, I'm sorry
I didn't mean to copy that second line
23:14 arlinius left
pmurias swarley: please don't paste stuff into the channel 23:19
swarley Yes, I'm sorry. I only meant to paste the first sentence. I apologize 23:20
23:21 swarley-freenode joined 23:22 spider-mario left 23:25 xilo_ joined
jnthn finally tracks down at least one of the strange things going on 23:29
(with the serialization stuff)
[Coke] general community note, inspired by diakopter in #parrot. - when we actually have parrot people on the line to talk to, it helps to actually talk about technology and code and bugs rather than he said/she said crap. It really doesn't help either project along. 23:36
23:36 stevan__ left
[Coke] I know this can be very frustrating, esp. if we feel we are on the receiving end. 23:36
sorear also can we please talk about code rather than mental illnesses? 23:40
23:42 stopbit left
swarley Okay, I've definitely finished playing around with the spec structs I believe 23:44
23:46 alec__ left
swarley-freenode o-O 23:49
23:49 alec__ joined
pmichaud [Coke]: I agree; I was seriously considering invoking the "bring in the relationship managers" rule... except afaik diakopter wasn't really speaking for Rakudo 23:51
[Coke] no, nothing official, I just knew it wasn't going to end well. :( 23:52
diakopter I don't see how it didn't end well 23:54
or proceed well, for that matter
23:54 stevan_ joined
[Coke] diakopter: really? did you get an answer to your question? 23:55
are you any closer to getting an answer to your question?
diakopter [Coke]: all my questions were rhetorical.
23:56 benabik joined
[Coke] ok. The way you approached it was unhelpful. the fact that it happened over something you didn't expect an answer to makes it even more puzzling to me. 23:56
unhelpful - in this case not in the case that it didn't further your goal, since it seems you didn't have one there, but in that it helped to sour the relationship between the two projects. 23:57
diakopter I disagree completely
I see no relationship soured
[Coke] You keep doing that.
I'm not sure it means what you think it means.
pmichaud diakopter: if you're speaking on behalf of rakudo, then I'm not sure you get to decide if it helped or lessened the relationship. :-) 23:58
diakopter but .. I wasn't doing so 23:59
[Coke] diakopter: so, what was your point in the conversation then, if not to get your technical question finally answered after 2 months?

Powered by Cro and the Raku® Programming Language.

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

Thank you!