weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today
Set by moderator on 20 August 2010.
01:44 eternaleye left 01:45 eternaleye joined, eternaleye left 01:46 eternaleye joined 01:47 eternaleye left, eternaleye joined 05:31 moritz_ left 05:32 moritz_ joined 18:46 masak joined 18:55 TimToady joined 18:56 colomon joined, avuserow joined 18:58 mberends joined 19:01 masak` joined
masak` o/ 19:01
mberends o/
colomon quick report, because I've got to run: messed around with some timings this week, leading pmichaud++ to report a bug and do a nice optimization. also deleted List.batch on the theory it's obsolete. planning on continue timing things. .eor 19:02
19:02 masak`` joined
masak` colomon++ 19:02
moritz_ will probably re-appear in 5 to 10 minutes 19:03
jnthn colomon++
OK
19:03 masak` left
jnthn masak``: Anything to tell of? 19:03
19:04 masak`` left
jnthn ...fail :-) 19:04
19:04 masak` joined 19:05 masak left
masak` has an especially crappy neighbournet connection tonight 19:05
pmichaud I think he's just testing us. :-)
jnthn masak`: You should get that examined
sorear pmichaud: is there anything we can do to #cpan6sketch next week to get people to stay?
jnthn masak`: Now your back, have you a report? 19:06
*you're :-)
masak` yes. it's been a while since I participated in one of these meetings, during which I've finished my GSoC grant. 19:07
pmichaud sorear: I stuck around for an hour :-)
jnthn masak`++
masak` more exactly a nice foundation has been built, on top of which much work remains to be done. especially with pack/unpack.
I expect to spend parts of the rest of the year doing pack/unpack.
including porting ~18k pack/unpack tests from Perl 5 to the Perl 6 test suite. 19:08
the rest of my Rakudo-specific work has basically involved trying to track down the Parrot bug that causes my enums patch (needed for a chapter of the book) not to apply cleanly. 19:09
.eor
jnthn masak`: It'd be really nice to see a short tutorial on how to do binary IO stuff in Perl 6 so we can point people at it who ask. 19:10
masak`: Or maybe find a way to include it in book
moritz_ back
masak` jnthn: both good suggestions. will consider them well. 19:11
jnthn masak`: Excellent, thanks.
mberends: You're next alphabetically :-)
mberends I'm going to nom soon. Worked mainly in the java subtree of 6model. It stalled on circular dependencies, then jnthn++ showed me a good way out, so expecting more progress in the coming week.
.eor # nom, akshually
moritz_ guten appetit! 19:12
masak` mberends: I'm interested in hearing about that solution at some point.
jnthn mberends++
moritz_? 19:13
moritz_ I've did lots of small things
writing tests for tickets
blogging
applying patches
exploring a few possible enhancements and/or simplifications to rakudo - most turned out to be not that easy :-) 19:14
did more $book work
answered p6 questions on perlmonks
nothing big..
still looking for ideas for the weekly challenges 19:15
.eor
19:15 masak` left, masak` joined
moritz_ I guess It's PerlJam's turn now, by alphabet 19:15
jnthn :-) 19:16
PerlJam?
PerlJam nothing to report. eor.
jnthn pmichaud?
pmichaud in general, been trying to restore some sense of "order" to my office and house after a very hectic spring+summer (thus far) 19:17
also taking care of a few $otherjob tasks
specific rakudo & nqp things done:
* added && conjunctions back into regexes 19:18
* this enabled simplification of the <.dumbsmart> rule for smartmatching
* rewrote the Cool.chomp method to be far more efficient
** this speeds up line-oriented file I/O by about 67% (on my system)
* Nil is now undefined 19:19
masak` \\o/
pmichaud * added a very crude profiling capability to parse traces
moritz_ pmichaud++
pmichaud ** now we can get some idea of where things are taking time during parsing 19:20
* as a result of the profiling capability, discovered that quantification of character classes (e.g., \\h* and \\v*) was a bit slower than needed; optimized these
* also know of a few other regex optimizations to hopefully speed up Rakudo compilation a bit
* experimented with substring comparison operations -- didn't seem to result in significant speed improvement 19:21
* working on today's *+1 release
(expect that later tonight)
.eor 19:22
jnthn pmichaud++ # wow, so much stuffs 19:23
moritz_ indeed, pmichaud++
19:23 tylercurtis joined
moritz_ jnthn: did you report already? 19:23
masak` pmichaud++
jnthn next is...q....r... sorear
moritz_: No, we need a second cut at the alphabet. :-) 19:24
moritz_ right :-)
jnthn ETIMEOUT 19:26
jnthn can go
This week 19:27
* Mostly been hacking on 6model this week
* Did a first cut of the leixcal case of the "proto controls" multi-dispatch
* It's missing the syntactic sugar but appears to work
* Made lexpads smarter and shaved a third off signature binding time as a result. Suspect that teaching Parrot to do similar will be more than the 1 hour of effort it was in 6model...
* Cleared up some bootstrapping issues
(to make things easier to port)
* Removed some .Net-isms
* Added some missing bits to the STable which brings the design a little closer to complete-ish there
* Having good profiling tools has been very informative
Plans
* Wrote a blog post laying out the way forward
* Hack on 6model a bit further
* Then start porting to Parrot
* Then branch the whole world and start re-shaping it 19:28
Vacation
* Leaving tomorrow
* Back on the first Monday in September
* The laptop is *not* coming, so I'll just not be around
* May find time to design a good multi-dispatch cache while I'm gone though...or maybe I'll just hike, nom and beer all the time :-)
* I'll miss all of you much more than you miss my bad puns, I'm sure :-)
.eor
moritz_ :-)
jnthn++
19:30 masak` left
jnthn Anyone else have a report? 19:30
[Coke] o/ 19:31
jnthn o/ [Coke]
19:31 masak` joined
jnthn If that meant "I have a report" go for it :-) 19:31
[Coke] I just went through some RT tickets and tried to verify/duplicate/whatever, also helped track down a leak in rakudo.
. 19:32
jnthn [Coke]++
That leak will be very nice to have patched :-)
[Coke] tylercurtis to the rescue. 19:33
TimToady [Coke]++ also reminded me that I'm already very good at sniping and hence don't really need to practice every day :) 19:36
pmichaud afk # kid pickup
jnthn Anything else for discussion? 19:37
moritz_ can't think of anything 19:38
jnthn either, but mostly because I'm thinking about what I'm going to forget to pack for vacation this time. :-) 19:39
masak` why will people want to upgrade from R* to R(*+1)?
what's the main goal for R(*+2)?
moritz_ improved performance (a few percent), regex and substitution adverbs, updated module list 19:40
masak` good, good. 19:41
moritz_ goals for R(*+2): module installer (IMHO)
masak` agreed. 19:42
jnthn Aye, that would be nice.
masak` and quite acheivable.
19:42 masak` is now known as masak
jnthn Yes, I think so 19:43
masak thanks everyone for their excellent work in the past month. it's great to be a part of all the developments happening all the time. 19:44
jnthn Yes, it's really exciting to see so much happening.
There was a time when I felt like I could keep abrest of pretty much all the bits and pieces going on in the Perl 6 world. Happily, there's too much of it for that now. :-) 19:45
colomon backlogs 19:46
Is neutro's module smoker still up and running? 19:47
(perhaps not automatically, but at least working?)
PerlJam BTW, I was working on a Dist::Zilla based rakudo distribution builder. If anyone is interested in playing with that, it's at git://github.com/perlpilot/rakudo-dist.git Read the README.dist file 19:51
It's not polished yet, but I think it's useful (especially if you already have dzil)
masak PerlJam++ 19:52
PerlJam Edit the dist.ini to change the name to rakudo-star and I think you'll have a good R* builder 19:53
20:01 masak left 20:03 patrickas joined 20:08 patrickas left
sorear jnthn: sorry I'm late, what do you want? 21:27
mberends sorear: jnthn asked during the meeting if you wished to report anything. 21:30
sorear I was ... out
jnthn sorear: No worries. :-)
sorear anyways, I haven't done anything interesting with Rakudo this week 21:31
mostly been trying to get niecza to accept STD.pm6 21:34
I've found and fixed a couple STD bugs (mosty stuff like "&$name" being expected to not interpolate) in the process, that'll get merged back at some point 21:35
jnthn Nice 21:39