Perl 6 - the future is here, just unevenly distributed

IRC log for #phasers, 2011-05-03

| Channels | #phasers index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
06:21 [particle] left #phasers
10:23 [particle] joined #phasers
18:53 moritz sorry, I'll miss (at least the start of) #phasers today
18:53 moritz did: started to put exceptions stuff in S32/Exceptions.pod
18:53 moritz .EOR
18:54 masak joined #phasers
18:54 Util Pre-report: Nil. Still intend to work on OS X packaging of R* this week, just dry of tuits. EOR
18:56 takadonet joined #phasers
18:56 masak pre-report: released Yapsi. submitted a number of rakudobugs. tuits might be thawing again, too soon to tell. enjoyed watching tadzik gearing up, and talking to him about his GSoC grant. EOR
19:01 masak o/
19:02 jnthn o/
19:02 colomon \o
19:03 Util o/
19:04 masak ok, anyone else wants to report?
19:04 colomon will in a sec
19:04 masak colomon?
19:05 masak ok, goodie.
19:05 colomon * Perl 6 Presentation at Penguicon
19:05 colomon * Think there were 25-30 people in the audience
19:05 colomon * Feel like I did okay for my first-ever presentation -- pulled it in exactly on time, had at least 5 people in the crowd very engaged
19:05 colomon * justrakudo.wordpress.com (with the notes for the presentation) had a noticeable bump in visitors over the weekend
19:05 colomon * Possible improvements:
19:05 colomon * I feel bad for forgetting to do a section on how to get involved
19:05 colomon * Ribbons very popular at the con, having Perl 6 ribbons to pass out would have been a PR boon
19:05 colomon * Use a QR code to make it easier for smartphone users to get to the web page
19:05 colomon * In other news, have a crude version of FatRat implemented on Math::BigInt working at the momenet
19:05 colomon * Plans: Get FatRat to the point it works for the infinite stream of Pi project
19:05 colomon * Blog on that and on abstraction versus encapsulation
19:06 colomon EOR
19:06 masak colomon++
19:06 masak more reports?
19:06 jnthn colomon++
19:06 jnthn haz
19:06 masak jnthn? tadzik?
19:07 masak jnthn: ok, shoot.
19:07 jnthn and leave?
19:07 jnthn ;)
19:07 jnthn Did NQP stuff...
19:07 jnthn * Attributes added through the compile time meta-object
19:07 jnthn * Detection of undeclared attributes at compile time
19:07 jnthn * Natively typed attributes
19:07 jnthn * More cleanup of package stuff
19:07 PerlJam colomon++  (though I read "engaged" as "enraged" at first and wondered what you did to them  ;)
19:07 jnthn * Methods added through the compile time meta-object - tricky as we need to do fixup in the compile-n-run case. But...it works now. :-)
19:07 jnthn * Refactored regex prefix method handling and got regexes added through compile time meta-object too
19:07 jnthn * Resolve more type names at runtime, so we detect issues with undeclared types better
19:07 jnthn * Build signatures in general deserialization, which got rid of a bunch of loadinits
19:07 jnthn * Finished getting everything going through compile time meta-objects; tossed $*PACKAGE-SETUP stuff
19:07 jnthn * Fixed role regression...hard, because things need the correct type args, but it works now :)
19:07 jnthn Over the next week...
19:07 jnthn * Fix last two multi-method test regressions
19:07 jnthn * Fix installation of new NQP
19:07 jnthn * Work on rakudo/nom
19:07 jnthn * Prepare for YAPC::Russia
19:08 jnthn EOR
19:08 masak jnthn++
19:08 colomon jnthn++
19:08 PerlJam jnthn++
19:08 takadonet jnthn++
19:08 Util colomon++ , jnthn++
19:09 masak any more reports?
19:11 masak any other points to discuss? questions, stuff?
19:11 benabik joined #phasers
19:11 wknight8111 joined #phasers
19:11 masak if not, we'll just go back to hacking :)
19:11 colomon jnthn: do you think you'll get rakudo/nom started properly before YAPC::Russia?
19:12 jnthn colomon: Yeah, hope to have some time for that during this week and at the weekend and probably a little next week also.
19:12 jnthn Tuits aren't in too short supply right now :)
19:13 jnthn The first phase is "rip out loads of stuff" :)
19:13 colomon I'm hoping you can get it to the point where moritz and I can poke away at it some even if you're not around.  :)
19:13 jnthn me too :)
19:13 pmichaud hello, #phasers
19:13 colomon pmichaud \o/
19:13 pmichaud I can report whenever you all are ready
19:14 masak pmichaud: please go ahead.
19:15 pmichaud * issued the 2011.04 star release
19:16 pmichaud * played with rakudo and parrot benchmarks
19:16 pmichaud * noticed we didn't get an overall speed win from 2011.01 to 2011.04
19:16 pmichaud * putting together a matrix of timings at http://pmichaud.com/rakbench
19:17 pmichaud * as expected, the weekend got fairly busy thus I didn't get many tuits for hacking
19:17 pmichaud EOR
19:17 jnthn pmichaud++ # benchmarking
19:18 masak pmichaud++
19:18 colomon pmichaud++ # what jnthn said
19:18 masak rakbench++
19:19 pmichaud right now I'm trying to chase down the variability in timings in the 2011.01 release
19:19 pmichaud what the table doesn't show is that earlier today I had core.pm timings as low as 200 seconds and as high as 850 seconds
19:19 pmichaud (for just 2011.01)
19:19 jnthn That wsa the only one where I got a significant variance.
19:20 pmichaud so, I'm running 2011.01 four times each with four separate parrot builds
19:20 pmichaud we'll see if the variation is due to parrot build or what
19:21 colomon memory issues, maybe?
19:21 pmichaud possible, but doesn't seem likely.  The machine has 4GB
19:21 moritz I'm doing some benchmarking too
19:21 pmichaud and I hardly ever hit swap
19:22 moritz on an otherwise idle 2-core machine
19:22 masak hitting swap ought to cause big variations.
19:22 PerlJam 2011.02 and 2011.01 both use gc=ms2, right?
19:22 pmichaud yes.
19:23 [Coke] left #phasers
19:23 jnthn I can pretty much promise the machine I ran on won't have hit swap. It's got 8 GB physical and 7 GB free when idling... :)
19:23 colomon A 4x difference that doesn't involve swap seems kind of scary..
19:23 pmichaud well, I only had the one 850s run... so I'm willing to call it an outlier
19:23 pmichaud it's entirely possible there was some other process (chrome, cron, etc.) that hit it badly
19:24 pmichaud since then I've been pretty careful to keep the machine unloaded
19:24 pmichaud (I also rebooted since then to clear out any background things that might've been running)
19:24 pmichaud anyway, the rakbench times are much more reliable
19:24 pmichaud I'll probably run the entire set again from scratch later today, and put it in a directory where I can keep clean archives of everything
19:25 [Coke] joined #phasers
19:28 jnthn pmichaud: How high on your list is the PAST::Compiler work you were pondering recentlyish?
19:31 jnthn pmichaud: The reason I ask is that I'm seeing more and more use for the PAST::ContextSensitive or whatever it would be called node
19:31 pmichaud it's high-ish, but it also depends on having a dedicated block of $tuits
19:31 pmichaud noted on ContextSensitive
19:32 jnthn It'd fix some sink issue for one
19:32 pmichaud yup
19:32 jnthn But the other use I have is that if I can go on register type,
19:32 pmichaud okay, I'll bump that up a bit
19:32 jnthn then it's rather useful for situations where we make, e.g. a boxed Int today, but if it's going into a natively typed attr we'd really rather have never boxed it only to unbox it again
19:33 jnthn The code I ended up with generating for native attrs is kinda...awful
19:33 jnthn Because many things seem to assume they'll be working with a boxed type.
19:34 tadzik oh oh!
19:34 tadzik am I very late?
19:35 jnthn Terribly. Awfully.
19:35 pmichaud afk, kid pickup
19:35 tadzik I may be out of my mind, I've been doing MIPS assembly all day long :|
19:35 PerlJam tadzik: that'll do it.
19:35 pmichaud (side note:  2011.04 does do better than 2011.01 on the rx.t spectest)
19:36 tadzik also, I've been going back home through the snow (yes!)
19:36 tadzik can I still report?
19:36 jnthn snow?!?!
19:36 tadzik yep. In may. In Poland
19:36 tadzik and I didn't even have a coat
19:36 benabik jnthn: pmichaud: Please note that parrot has someone working on PAST for GSoC (namely me)
19:36 [Coke] left #phasers
19:37 pmichaud benabik: yes, I know.  I think we'll just end up duplicating some efforts and merge together later
19:37 pmichaud also, I think your focus is a bit different from mine
19:37 pmichaud you're of course welcome to steal anything I write :)
19:37 pmichaud afk -- bbi15
19:38 tadzik My report:
19:38 tadzik Added tests for Benchmark and a new shiny META.info for it
19:38 [Coke] joined #phasers
19:38 tadzik Started work on Shell::Command and started deprecating File::Tools
19:39 tadzik Put some thoughts in the gsocmess repo, did a bit of inspiring research with moritz++ and masak++
19:40 masak ;)
19:40 tadzik Fixed a number of RT tickets (moritz++ for the actual closing), we were down to 848 for a short moment when masak-san was not around to open a new ones
19:40 tadzik also started work on Module::Starter: it can even create a github repo for you!
19:41 * masak mumbles about people closing bugs while he's asleep ;)
19:41 tadzik that'd be it methinks
19:41 masak tadzik++
19:41 tadzik and sorry for being late :(
19:41 moritz tadzik++ indeed
19:41 colomon tadzik++
19:41 masak tadzik: well... snow. o.O
19:42 tadzik no snow in Sweden?
19:43 masak no, just colder than in the past week or so.
19:44 tadzik it's 2 degrees Celsius right now
19:44 moritz we had some snow in .de, but not where I live
19:44 masak forecast says we'll go below zero tonight.
19:45 tadzik a friend of mine was talking on the phone today "snow? (lauhgs) No, we just have rain". Ha-ha...
19:46 jnthn Apparently it did snow a little in Stockholm today.
19:46 [Coke] left #phasers
19:46 jnthn But that's like, almost at the north pole, so not surprisng really.
19:47 masak :P
19:48 masak they have twice the amount of polar bears we have down here in southern Sweden!
19:48 [Coke] joined #phasers
19:49 jnthn Only twice? I thought it was, like, three times or so.
19:50 masak could be more.
19:53 pmichaud back
19:54 pmichaud it got below 5degC here last night
19:54 pmichaud Yes, In Texas.  In May.
19:55 tadzik the times, they are a'changing
20:12 [Coke] left #phasers
20:14 [Coke] joined #phasers
20:28 wknight8111 left #phasers
20:57 benabik left #phasers
22:17 [particle]1 joined #phasers
22:17 [particle] left #phasers
22:38 masak left #phasers

| Channels | #phasers index | Today | | Search | Google Search | Plain-Text | summary