weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today
Set by moderator on 5 January 2011.
03:42 _ilbot joined
moderator weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today
17:59 tadzik joined 18:22 takadonet joined
takadonet hey all 18:22
tadzik hey
btw, is "IR clogs at" some inside-joke? :) 18:23
moritz_ yes
it's "irc logs", but sinde the subdomain is irclog.perlg...
some people read that as "IR clogs"
sorear report: been hacking on self-hosted niecza. It is starting to work but still far from releasable. EOR 18:37
moritz_ report: I have implemented non-spec Devel::Trace in Rakudo. Been hacking on masak's challenges a bit more; made t/harness barf when spectest_smolder is run without TAP::Harness::Archive present. EOR (I think) :-) 18:39
18:45 diakopter joined
tadzik report: thought about and worked on the META.list idea from S22. Created an ecosystem fork and poked module authors to move their modules to the new infrastructure. Refactored neutro, spread some interest in Perl 6 between Dancer devs. EOR I think 18:46
diakopter tadzik++
tadzik oh, and my CLA is on the go to USA
diakopter also you did some perl6.org patches
moritz_ right, a new page for STD.pm6 + viv 18:47
tadzik oh well, that was #perl6 requests
diakopter (I meant a couple patches tadzik did, but moritz_++ too)
oh wait, yeah, superseded by moritz_'s changes; forgot 18:48
sorear is currently trying to fix heredoc parsing 18:51
18:52 shortcircuit joined
Util Pre-report: No work done this week. Today, will commit (first draft) script to build Rakudo Star OS X .dmg binary. 18:57
EOR
colomon o/ 19:00
diakopter ooo
Util Hello
jnthn o/
moritz_ \\o
diakopter sorear: so... #phasers is *officially* beginning now ;) 19:01
o/
jnthn lol :)
diakopter Util reported, tadzik reported
jnthn has a report ready Really Soon... 19:02
tadzik moritz_ and sorear to
moritz_ let's roll the drums for jnthn
tadzik drmdrmdrmdrmdrm
moritz_ *thud*
:-)
jnthn In the last week I...
* Had a throat infection :-/
* Had another crack at the nominal type cache idea for Rakudo. Shaves 30% off tools/benchmark.pl.
* Played with making scalar and rw flags cheaper in the common case. Work's in a branch. Fails various tests; can probably fix.
* Ported some of the latest 6model/nqpclr bits to nqp-rx/nom
* Eliminated nqp-rx/nom use of PCT::HLLCompiler by moving stuff into nqp-rx repo. Translated some bits to NQP. Much of the rest is LHF.
* Added class_6m to nqp-rx/nom to start transition to the new meta-model
* So far, all Actions classes and Compiler classes are moved over! \\o/ 19:03
In the next week I'll...
* Get the nqp-rx/nom equivalent of Mu in place
* Eliminate class_6m by switching classes to always use 6model
* Start to port 6model/nqpclr's multi-dispatch to nqp-rx/nom
* Write an architecture overview doc for 6model
* Hack with mberends++ on 6model/java
Blockers
* None yet, but soon will need to try and co-ordinate with Pm and then Parrot folks on a nqp-rx => nqp-rx/nom migration plan
EOR
diakopter yay for 'Write an architecture overview doc for 6model'
colomon \\o/
moritz_ \\o/
diakopter report: added a few more bits to perlesque; I want to add hash literals, and then finally return to viv_psq after.. what's it been... 4-5 months? EOR 19:06
moritz_ so 19:07
do we need to discuss anything?
diakopter (and by "return to", I mean, go back to trying to figure it out enough to take sorear's work and extend it) EOR4REALZ
tadzik I have a few questions about releasing Rakudo, as I'm to be a release manager. What is the usual/expected time for a commit freeze? 19:08
more like an hour before the release, or rather a day? 19:09
moritz_ more like an hour
19:09 masak joined
masak is late 19:10
PerlJam tadzik: There's a commit freeze? (does that help answer your question? :-) 19:11
tadzik: seriously, I tend to agree with moritz_
moritz_ PerlJam: typically only announced in #perl6
jnthn tadzik: The commit rate isn't *that* high and committers are often about on the channel.
What moritz_ said :)
tadzik PerlJam: well, I think I recall something like this in before
masak sorear: what'll be the big benefit from the self-hosting? &eval? also, what are the chances the self-hosting part will be the default in the next release? 19:12
moritz_ yes, people just says "I'm cutting the release now, please hold your commits for a while" 19:13
PerlJam Even still. Once you pick a revision for release, it doesn't matter much if anyone commits.
Util Now that we are using Git, the release manager *could* just not pull/fetch during the release process. 19:14
(oh, like PerlJam was saying)
PerlJam Util: yep
sorear masak: &eval is part of it yes, although that won't be in the next release 19:16
masak sorear: ok. 19:17
diakopter moritz_: where's your 2010 summary url?
sorear masak: more immediately, it seems to be significantly faster, and eliminates the requirement to install Perl 5 extension modules
moritz_ diakopter: perlgeek.de/blog-en/perl-6/perl-6-in-2010.html 19:18
masak \\o/
diakopter thanks
sorear diakopter: could you look at feather.perl6.nl/~sorear/niecza-v6-...110111.zip on the Microsoft CLR? 19:19
I doubt it will work; it's from the same codegen you had issues with last week
diakopter (yes I'm trying when I get a minute) :) 19:25
sorry to block #phasers if I was
masak small report: haven't done much since Christmas, but slowly getting in enough order to be able to hack on simple things again. the Contest will end this Friday -- looking forward to summarize that. eor 19:26
moritz_ Friday already... 19:28
masak $time.fly( :like(Arrow.new) ); 19:29
PerlJam :-( I haven't even started on the contest problems
masak PerlJam: er. did you sign up? 19:31
PerlJam: I don't see you among the contestants.
PerlJam I may have forgotten to sign up. 19:32
In any case, it doesn't much matter to me to be part of the contest as much as to solve interesting problems and see how other solved them and so forth
masak there's still approximately four days left to do that :)
PerlJam unfortunately I only really have time in the evenings after I put the kids to bed to work on fun coding projects and lately I've been falling asleep myself 19:35
Makes me feel like I'm getting old :) 19:36
masak aww :) 19:40
moritz_ any more phases we need to cross? 19:41
tadzik masak: can I ask you to move your modules to the META.info format? I'd like to merge the neutro/metainfo branch and continue investigating the module case 19:42
jnthn doesn't have any questions/discussion items, beyond "scream at me if the nomtype cache busted your code"
tadzik and yours are quite often used and important
jnthn But nobody seems to have, so I'm optimistic. :)
masak tadzik: sure, just let me know what changes I should make.
takadonet jnthn: i noticed the speed up
tadzik masak: you need to add a META.info file in the root directory of a repo 19:43
looks like this: github.com/tadzik/neutro/blob/meta.../META.info
jnthn takadonet: Really?
takadonet: That's nice to hear. :)
I was wondering if it was any real-world help :)
takadonet takadonet: for the module that have lots of fcn calls
tadzik oh, one more thing 19:44
jnthn takadonet: Nice :)
Now if only we could get iteration not to be dog slow... :)
takadonet jnthn: and startup
jnthn takadonet: Yes. 19:45
That'd help too.
jnthn will keep hacking away :)
tadzik I see Kodi++ closing many tickets, but there are still dozens of those opened, but fixed, I remember finding like a dozen of them once, but there was no one to close them. What is the best thing to do to move it forward?
bah, I never thought of replying to a tickit until now :) 19:46
moritz_ tadzik: get write access to the RT queue
and do it :-)
tadzik moritz_: I asked pmichaud for it like 2 months ago ;)
masak ask pmichaud again.
moritz_ tadzik: ask again when he's actually online
tadzik ok :) 19:47
shortcircuit I don't know whether it's my place to talk, but I've taken a toy interest in looking at the PIR/NQP layer of Rakudo, and seeing where I might be able to replace Parrot with, e.g. C++.
To that end, it would be helpful to have a relativaly defined boundary between NQP and Perl 6.
tadzik or we can call a "ticket closing week, yay!", encourage people to look through the RT and prepare a nice prize for the winners, like The Smile Of Jonathan or something :) 19:48
jnthn ...my smile is enough to motivate people to do stuff? :)
tadzik oh, that's symbolic ;)
I can donate a cookie for a prize 19:49
it may even be a cookie from rakudo.org :)
[Coke] can grant bugadmin access.
PerlJam shortcircuit: I think there was talk of moving all of the PIR stuff behind the NQP:: namespace so that all you have to do is re-wire the NQP::* routines
masak ooh, the Smile of Jonathan! \\o/
jnthn The aim is certainly to have less PIR, more NQP 19:50
Especially because you *really* don't want to write PIR when dealing with the new meta-model stuff. :)
shortcircuit I'd point out that it's presumably helpful to have a fluidic boundary between VM and PIR, to allow for optimizations at the expense of maintainer effort.
I'm going for minimum-effort, but others might not. 19:51
er.
fluidic boundary between VM and NQP.
jnthn shortcircuit: Certainly I want a cleaner interfce there.
PerlJam shortcircuit: are you volunteering to work on it ? :)
jnthn shortcircuit: It's important to me too, as I want Rakudo on other VMs. It's kinda just not quite there yet.
Help on it is most welcome and certainly going in the right direction. 19:52
sorear One thing I'd like to discuss is whether ($x) and ($x,) should be the same 19:53
shortcircuit PerlJam: Don't tell anyone, but I already am. ;) At least in the sense that I'm looking at it and figuring out the parts that would make it easier for me to do, which may or may not ultimately have any results. 19:54
PerlJam sorear: Yes, they should be the same.
Util sorear: why would they differ? 19:55
shortcircuit Even if I did get it working, I wouldn't maintain it. At best, the effort would provide experience toward future, more persistent efforts. 19:56
sorear PerlJam: good to get a second opinion on that. 20:02
now I just need to figure out how ($x) ,= 1 works 20:03
(nevermind the fact that it's practically useless)
jnthn agrees with PerlJam fwiw :)
sorear basically I'm plotting to make 1-element parcels not exist in niecza 20:04
so sub return(\\|$things) is return-pass-special { } can work
20:08 masak left 20:40 eternaleye left, eternaleye 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!