Parrot 5.4.0 "Austin Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 1 June 2013.
05:24 ilbot2 joined
moderator Parrot 5.4.0 "Austin Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
ilbot2 joined
moderator Parrot 5.4.0 "Austin Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
06:17 he joined 07:34 myhrlin_ joined 07:42 cosimo_ joined 07:49 cosimo_ joined 08:52 bouncy joined 09:46 denisboyun joined 09:52 denisboyun joined 11:41 cogno joined 11:52 cogno joined 12:10 kid51 joined 12:24 gtodd joined 12:57 diakopter left 13:03 ThePoster62 joined, ThePoster62 left 13:17 bluescreen joined 14:21 PacoAir joined 14:46 rurban joined 14:53 cxreg joined 14:55 darbelo joined 15:31 darbelo joined 15:38 zby_home joined 15:47 benabik joined 16:28 contingencyplan joined 16:44 dukeleto joined
dukeleto ~~ 16:44
benabik o/ dukeleto
dukeleto benabik: wazzup
benabik: i saw a new parrot.git branch from you
benabik dukeleto: Pondering lunch.
dukeleto is thinking about breakfast 16:45
benabik There was some discussion of compressing PBCs, so I gave it a half-hour hack.
dukeleto is in some kind of superposition of timezones between PST and Central
benabik: cool!
benabik: what about parrot itself using fakecutables?
benabik After I finished I realized I wanted _library_ PBCs to be compressed and that involves tying gzip to the load_bytecode op.
dukeleto benabik: we could gzip the PBC and put it in a fakecutable, right? 16:46
benabik: i am not against adding a gzip opcode or whatever you need to make it reasonable
benabik: parrot needs big changes. Let's do interesting stuff. I am sick and tired of walking on eggshells
benabik Heh.
dukeleto Rakudo is actively moving away from Parrot. Let's break some shit.
The silly deprecation cycle is essentially a noose around our throat and legs right now. 16:47
the perl6/parrot/moarvm hackathon is happening now
PerlJam
.oO( "now"?!? )
dukeleto but i am in my hotel room
dukeleto lost a contact and is now a cyclops, because he forgot to bring any glasses or contacts
benabik wishes he was there.
dukeleto benabik: i wish you were here in my place. A lot more would get done. 16:48
benabik Bah.
dukeleto benabik: MoaVM is the spiritual succesor to M0
MoarVM, rather
i took many pictures
need to regurgitate them onto the interwebs
benabik MoarVM seems to match my list of "how I would improve/rewrite Parrot" to a ridiculous degree.
Wash them first. Ew. 16:49
;-)
dukeleto benabik: MoarVM is a VM specifically for NQP
benabik: it uses many lessons from Parrot and other systems
Util benabik: I wrote (or rewrote) a lot of the fakeucutable process code. Adding g(un)zip sounds good, if there has been a call for it. Let me know how|if I can help.
dukeleto but I see a bright future for Parrot
Util: oooh! I didn't know that, Util++ 16:50
benabik The use case for gzip was the *.setting.pbc files from NQP/Rakudo.
I think those are just called with load_bytecode, so that's where I'd try to hook it.
dukeleto why not bzip2? It is superior in every way, right? 16:51
benabik Although I suppose pervasive use of .gz wouldn't hurt. We store a _lot_ of 8b ints in 32/64b space.
Well, it started with "GzipHandle is already in Parrot". :-D
dukeleto benabik: gotcha
benabik: i thought fperrad wrote something that talked to bzip2 from parrot, but I could have dreamed that 16:52
Coke as an HLL author, I transitioned to nqp-rx (and am still in the middle of transitioning to nqp), so improved/faster/feature/whatever for nqp may still be a driving factor.
dukeleto Coke: what is that comment in response to? Just trying to understand.
benabik gzip compressing the setting _might_ be a speed improvement, although probably less so on an SSD.
Util starts Hotel -> Hackathon commute 16:53
dukeleto benabik: most people don't have SSDs yet, so I would consider that a big win still
Util: see you soon
Coke just catching up with general "yay moarvm but what does this mean for parrot" chat.
dukeleto Coke: cool
i have many pics and notes to send to parrot-dev 16:54
but I am essentially wiped out
PerlJam Coke: doesn't it mean nothing for parrot that moarvm "exists"?
dukeleto I think I was hit by an emotional bus yesterday. Just too many things going on.
PerlJam: i would very much disagree that it means "nothing"
PerlJam: it is an important development
I am honored that MoarVM lovingly stole so many good ideas from Parrot.
and I have told all the MoarVM peeps that they have my help when they need it. 16:55
But I am excited about the non-Perl6 future of Parrot
now that MoarVM exists, we don't have to worry about losing Rakudo as a customer
THAT HAS ALREADY HAPPENED
Let's move on.
PerlJam dukeleto: oh, I see.
dukeleto I can find a new customer for Parrot. Somebody with a lot of $$$.
PerlJam dukeleto: I thought you were going to go the "healthy competition" route. 16:56
dukeleto If anybody has a problem with that, I humbly suggest they tap me on the shoulder in person, look into my eyes and tell me their darkness fears and wishes.
PerlJam: healthy competition? 16:57
PerlJam dukeleto: VMs competing for Rakudo's attention makes them both better.
dukeleto PerlJam: nope
PerlJam: but do what ye will 16:58
As a current Parrot Foundation board member I can tell you that the future of Parrot is not Perl 6.
PerlJam So ... which language(s) does Parrot seek to serve? Without the HLL, isn't Parrot kind of directionless?
dukeleto PerlJam: i don't care to argue with you and I will soon leave IRC 16:59
PerlJam: either send some commits into parrot.git or I really don't care
dukeleto doesn't care to argue anymore
PerlJam dukeleto: I'm not out to argue, just seeking understanding.
dukeleto PerlJam: same here.
PerlJam: i think your questions are good, but perhaps you need to find the answer yourself. 17:00
Or perhaps your question is nonsensical.
bacek is not welcome in the Parrot community unless he can be civil.
Just for reference.
If anybody doesn't like that, too bad. 17:01
dukeleto is done allowing negative people to poison communities
benabik dukeleto, our benevolent dictator?
dukeleto benabik: I don't know. I asked Larry for advice, but he only gives meta-advice
benabik Heh. My dad does that: Trys to tell people how to make the decisions he'd make rather than telling them what he'd do. 17:02
PerlJam sometimes meta-meta-advice ;)
dukeleto I am not sure everyone in here knows, but Larry is currently battling cancer. He talked about it in his keynote. 17:03
he will find out later this summer if things are good or bad
benabik :-/
17:05 kid51_ joined 17:17 denisboyun joined
Coke finds dukeleto's note about negative people ironical given the last email he saw from the duke. 17:30
but i certainly prefer a positive community.
masak I think it's important to talk about things when there's a crisis. if you're seeking to emulate Larry, that's definitely what he is doing. 17:41
but I also won't linger on the subject; it's been ages since I made a commit ;) 17:43
gtodd did something bad happen at YAPC 17:44
I mean "bad" for parrot? 17:45
dukeleto: oh
gtodd scrolled up
dukeleto: sheesh
sorear strictly speaking, moarvm happened over much of the last year and was released a week ago 17:49
released from bondage...
masak it wasn't *released*. 17:50
there hasn't been a moarvm release.
it was "unveiled" or "revealed". 17:51
Util gtodd: I think that *nothing* good about Parrot was said during *any* talk at this YAPC. I think that I attended all the talks that were likely to mention it. 18:02
PerlJam: Here is my perspective.
MoarVM is off to a strong start, and might someday completely replace Parrot's role as the Perl-community-owned backend for NQP&Rakudo.
(JVM/.Net/JS backends are great, but we will always need a backend of our own)
However, MoarVM is high-risk in several (undiscussed) ways.
(...He said in a room full of MoarVM hackers)
I, for one, will stay-the-course with Parrot, serving the needs of NQP and Rakudo, stealing ideas from MoarVM, unless&until MoarVM *actually* supplants Parrot.
dukeleto and other board members have a larger/longer-term vision than I on the matter of the future of Parrot. In the short-term, I (and other Parrot contributors) continue to make improvements.
gtodd there seem to have been a few vms and probably more to come.. that might be good? 18:04
dalek rrot: f40728a | coke++ | docs/project/release_manager_guide.pod:
remove ponie-era relic
Util gtodd: I wish MoarVM, and P2, and Moe, all the best. I may even lend a hand; MoarVM looks really cool, and Moe is in SCALA!. We just can't wait on them, and the tension to do so is bubbling up in some slightly contentious ways. It *is* good, quite good overall, to have this plethora; just not good for everybody in every way. 18:09
masak Util: I can stand by that. 18:10
gtodd I'm just someone who just wants to use perl6 all over the place ... so ... moarvm moe lua p2 are all cool but parrot seems ahead in practical terms like using perl6
masak aye. 18:11
Util For example, some MoarVM people *could* be working on Parrot. If fact, some *did* in the past. Does the loss of those people make me happy, from a Parrot perspective? Of course not; I ache for their loss. But I will not pretend that they should return to Parrot, nor that their new efforts are in any way a waste.
gtodd but if they borrow amongst themselves via liberal licenses and friendly competition then yay!
Util: well put ... 18:12
masak "open source is not a zero-sum game. it is a positive-sum game!" -- someone, I don't remember who, at YAPC::NA 18:15
Util gtodd: Borrowing is great, but does not happen without effort. Nor without knowledge, for that matter. jnthn described some awesome lock-free tactics he uses in MoarVM. *I* am not qualified to retrofit them into Parrot; hopefully an active Parrot contributor *is*, but the fewer people are working on Parrot, the longer such borrowing will take. 18:16
gtodd I have to say I like the potion/p2 website :-P 18:17
Util masak: True! Unfortunately, it is not an infinite-sum game :)
gtodd Util: yeah mulitiple VMs != TIMTOWDI 18:18
PerlJam masak: Whoever said that (and I don't actually remember who) reminded me of the talk that Clay Shirky did about "Open Source is Love" or whatever and then Larry (and others) actually *said* it. 18:19
masak well, in some sense it's Darwinian evolution. the genotypes survive whose phenotypes have utility. 18:20
PerlJam: I have not seen Clay Shirky's talk, I think.
Util Stevan Little pointed out that supporting a JVM backend opens up whole *worlds* to us. Phones, JVM-only shops, all the classfiles only implemented on JVM, etc. People are *scared* of fragmentation; I heard it repeatedly during YAPC. They don't understand that frag is the price of sufficiently-fast evolution today. 18:21
PerlJam masak: www.youtube.com/watch?v=Xe1TZaElTAs 18:23
Util Personally, I can't stand the taste of the fragmentation; the single implementation of `perl` has been a point of strength in my own career. I recognize the need for the dispora, though, and will do my best to be a positive voice for it. 18:24
PerlJam Util++
gtodd so we shouldn't fear fragmentation ... cuz we can fix it later with ... perl 18:25
dalek p: 3887f45 | sorear++ | src/vm/jvm/ (2 files):
Fix compilation of \\n on JVM
 18:26
masak Util: really, the trick isn't going out and exploring, as many are doing. the trick is coming back and usefully reintegrating the findings in the mainline community.
Util: the parallel with git and branching/merging is striking. and not a coincidence, of course, since it's the same kind of thing going on. 18:27
Util: if we as a community can make subcommunity merging a solved problem, then we'll really have something.
Util masak: Good point
PerlJam masak: indeed.
18:31 danaj_ joined 18:32 darbelo joined 18:37 rurban joined
Coke (jvm) my main reason for supporting it is because it's the only way I will ever be able to perl for work. 18:38
darbelo Ah. Good ol' "Use what you want, so long as it can run on the JVM" 18:39
Coke exactly. :|
sorear Coke: down to 3 failures in the nqp test suite for jvm
Coke We're about 50/50 Java/ColdFusion here. I think one group bothered to try out groovy - but everything has to run in an j2ee container.
sorear++ 18:40
sorear that's been my hackathon project
Coke awesome. 18:41
darbelo I thought they had killed ColdFusion some time ago. I might be confusing it with something else. 18:42
Coke <cfquery name="stuff">select * from dual</cfquery><cfdump var="#stuff#"> 18:43
darbelo: I've been doing CF on and off since CF4 or so.
I had a dream once of getting it running on parrot. 18:44
darbelo Not nightmare?
18:45 denisboyun joined
Util Coke & darbelo: (jvm) exactly! Coke's situation is a lot more common than I had known before Stevan enlightened us. 18:45
go, sorear, go! 18:47
darbelo I've had to deal with people who absolutely *did not care* if you wrote the code in a dead basic dialect from the 70s, so long as they could deploy it as a WAR file. 18:48
benabik WAR?
Coke j2ee deployment file format. 18:49
it's a basically a zip of zips with some special files in
benabik Sounds like "fun"
sorear do they care how big the WAR file is? 18:50
if we have a 30-MB JRakudo... 18:51
Coke 30MB is pretty small, relatively speaking.
I'd be more worried about memory utilization.
gtodd did the scala MOE person give a talk at YAPC::NA? 18:52
oops found it
sorear you mean stevan? 18:53
gtodd yeah :)
sorry heh
well known for some object thingie ;-) 18:54
sorear yeah, he's around. 18:55
don't remember if he actually gave a talk, but I've talked to him plenty :)
sartak's also here.
PerlJam he gave at least 2 talks
sorear doy's busy getting married
PerlJam sartak already left. He did the advanced moose class this morning and then headed for the airport before lunch 18:56
darbelo sorear: (JRakudo WAR) You'd need to buy into (read implement) the servlet API and other environment stuff as well, but as Coke said, 30MB is not much on the Java world.
Coke darbelo: I would love to end up with SixMojo or something I can run at work. 19:00
darbelo I haven't done anything web on Java in quite a while. You could say I've moved to a different ghetto :) 19:02
darbelo is doing mobile apps this days. 19:03
gtodd I guess the talk on Moe was somewhere else ... 19:07
Moe sounds neat esp. if is it true that scala can be built and run with LLVM etc 19:27
19:36 denisboyun_ joined
dalek kudo/nom: d690ac6 | (Elizabeth Mattijsen)++ | src/core/Any.pm:
Implement the whole list of valid adverb combinations from S02:2533
 19:50
p: 3382c85 | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
Use a more correct definition of CCLASS_WHITESPACE for the JVM
 20:02
20:15 perlite_ joined
dalek kudo/nom: e769409 | (Elizabeth Mattijsen)++ | src/core/Any.pm:
Need to revert xor logic to be correct
 20:22
20:42 denisboyun_ joined 21:11 benabik joined
dalek p: 527419e | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/ (3 files):
Allow flattening native-typed arrays into argument lists on JVM
 21:15
Coke Should we stop sending rakudo commit messages here? 21:22
(and or nqp?) 21:23
21:32 bluescreen_ joined
PerlJam Coke: seems reasonable (to stop) to me. 21:35
darbelo Did we stop sending commits from other HLLs or did we just stop committing to other HLLs? 21:43
dalek kudo/nom: 0dceace | (Elizabeth Mattijsen)++ | src/core/Bool.pm:
Booleanify ?^, as discussed with Pm

As shown at irclog.perlgeek.de/perl6/2013-06-07#i_7170609		 21:59
22:35 myhrlin joined 22:42 denisboyun joined 22:44 danaj_ joined 22:54 Khisanth joined 23:00 contingencyplan joined 23:11 danaj_ joined 23:22 benabik joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!