Camelia, the Perl 6 bug

IRC log for #parrot, 2011-10-25

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 cotto dukeleto, how's the git conf so far?
00:03 cotto doug lea is here.  It's weird.
00:15 cotto kiss advice (for any x): every 6 months, try to teach it to someone who doesn't know it
00:21 cotto There was also an entertaining talk about how it's impossible to microbenchmark in a way that avoids statistically meaningful non-determinism.
00:21 Coke joined #parrot
00:22 cotto one graph had 5-6 overlapping steady states
00:22 cotto (though if it's 2x faster, it's 2x faster)
00:23 Coke joined #parrot
00:27 cotto http://www.cs.iastate.edu/~de​sign/vmil/2011/program.shtml (see the microbenchmarking pdf)
00:50 whiteknight in perl5, is there a way to do s/// without modifying the string in place?
00:51 benabik my $copy = $str;  $copy =~ s/pat/rep/
00:51 whiteknight okay, that's not a reference copy?
00:51 nbrown joined #parrot
00:51 benabik I don't think so?
00:52 whiteknight whatever, I'll try it
00:52 benabik No, it copies the string.
00:55 whiteknight I can't find a gui tool that does batch conversions from .mov to .ogv, so I'm writing a small wrapper around ffmpeg
00:57 benabik Handbrake?
01:01 plobsing If you are using perl 5.14 there's s///r for non-destructive substitution. Otherwise a fairly common (used many times in parrot core) idiom is "(my $copy = $str) =~ s///".
01:56 dalek parrot: b1d625e | petdance++ | src/dynpmc/select.pmc:
01:57 dalek parrot: fixed indentation and consted some vars
01:57 dalek parrot: review: https://github.com/parrot/parrot/commit/b1d625e9fc
01:57 dalek parrot: e12d045 | petdance++ | include/parrot/packfile.h:
01:57 dalek parrot: re-headerized
01:57 dalek parrot: review: https://github.com/parrot/parrot/commit/e12d045d36
01:57 dalek parrot: 29aefd9 | jkeenan++ | src/ (2 files):
01:57 dalek parrot: [codingstd] Rebreak two lines to follow linelength standard.
01:57 dalek parrot: review: https://github.com/parrot/parrot/commit/29aefd9175
01:57 dalek parrot: ec17023 | jkeenan++ | / (2 files):
01:57 dalek parrot: Merge branch 'master' of git@github.com:parrot/parrot
01:57 dalek parrot: review: https://github.com/parrot/parrot/commit/ec170230a1
02:20 jlaire joined #parrot
03:13 preflex joined #parrot
04:09 dngor joined #parrot
04:37 cotto ~~
04:55 dalek TT #2201 created by petdance++: This product has only a moose, not a pony
04:55 dalek TT #2201: http://trac.parrot.org/parrot/ticket/2201
05:03 cotto I'm glad petdance++ hasn't become a troll.
05:07 wagle joined #parrot
05:12 SHODAN joined #parrot
05:14 alester joined #parrot
05:17 sorear alester: how much should I prioritize writing a dalek plugin for GitHub issues?
05:21 alester What is dalek?
05:21 alester Oh, IRC bot?
05:21 alester I don't know, but we'll have github issues pretty soon.
05:22 alester https://github.com/petdance/scraps/blob/master/tix
05:22 alester Now I just have to dump into Github format
07:48 mj41_nb joined #parrot
07:49 mj41 joined #parrot
08:10 jsut_ joined #parrot
08:18 lucian_ joined #parrot
08:27 perlite_ joined #parrot
08:30 contingencyplan joined #parrot
08:59 lucian joined #parrot
10:14 moritz fwiw I just tested rakudo on top of the green_threads branch, all tests PASS
10:18 moritz ... and it's slower than on master.
10:32 benabik moritz: How much slower?  1, 10, 100%?
10:32 moritz benabik: 6%. Just sent a mail to parrot-dev
10:32 moritz lunch&
10:36 benabik moritz: Hm.  That's LTA.  I thought nine had pushed the no-overhead patch.  Hope someone has the tuits to diagnose it.  I have to do thesis work today.
10:45 ambs joined #parrot
10:51 smash joined #parrot
10:51 smash hello everyone
11:31 benabik joined #parrot
11:38 Psyche^ joined #parrot
11:40 benabik o/ #parrot
11:42 lucian_ joined #parrot
11:44 mtk joined #parrot
11:50 JimmyZ joined #parrot
12:05 lucian joined #parrot
12:07 p6eval joined #parrot
12:10 whiteknight joined #parrot
12:14 ambs joined #parrot
12:14 jsut joined #parrot
12:15 whiteknight good morning, #parrot
12:15 benabik morning, whiteknight
12:16 whiteknight hello benabik, how are you doing today?
12:16 benabik bleh.  I dislike 6 AM.
12:22 whiteknight yeah, it's one of the worst of all the AMs
12:23 moritz my lovely daughter decided a few days ago that she wants to get up between 5 AM and 6 AM
12:23 moritz and today I found out that she's just hungry, and will continue sleeping after she had her breakfast
12:24 moritz so it's getting up at 5am, preparing food for her, feeding her and then straight back to bed
12:25 whiteknight ouch. What's she eating now?
12:31 moritz cereal pulp with fruits (for breakfast)
12:34 whiteknight oh, for my son we would have just jammed a bottle in his face and all gone back to bed
12:34 whiteknight the gourmet fruit puree can wait until the sun comes up
12:36 moritz she has never liked bottles at all
12:55 bluescreen joined #parrot
13:07 preflex joined #parrot
13:11 pbaker joined #parrot
13:12 dalek rakudo/nom: 481a027 | jnthn++ | src/ (4 files):
13:12 dalek rakudo/nom: Give initialization of state variables START semantics.
13:12 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/481a027027
13:15 bluescreen joined #parrot
13:47 dalek rakudo/nom: 3e8f571 | jnthn++ | src/Perl6/SymbolTable.pm:
13:47 dalek rakudo/nom: Make uninitialized int default to 0 and an uninitialized num default to NaN.
13:47 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3e8f571ea4
13:55 lateau joined #parrot
13:56 lateau winxed —version
13:56 lateau 1.3.-1
13:56 benabik lateau: ?
13:56 benabik Oh.
13:57 lateau ;)
13:57 benabik I think winxed uses -1 for "development version"
13:57 lateau i got it
13:57 benabik Might be nice to show 1.3-dev instead of 1.3.-1 though.  :-D
13:59 lateau i think so :)
14:02 SHODAN joined #parrot
14:03 lucian joined #parrot
14:06 lucian_ joined #parrot
14:11 lateau joined #parrot
14:14 mj41 joined #parrot
14:14 mj41_nb joined #parrot
14:25 benabik Every time I discover a bit of winxed syntax, I go "that is both useful and should have been obvious."
14:25 benabik (Today's discovery:  int version[3])
14:27 benabik Ack!  11k lines in winxed stage 1.
14:33 benabik My winxed --version: Winxed 1.4.0-dev
14:35 lucian joined #parrot
14:35 lucian joined #parrot
14:35 * JimmyZ more likes 6model in parrot :)
14:36 benabik JimmyZ: From hacking Rakudo, hacking on non-6model PMCs, or just in general?
14:36 benabik JimmyZ: Or, more hopefully, working on adding 6model to core parrot?
14:36 benabik ;-)
14:37 JimmyZ porting 6model to core parrot, or PCC refactor, hehe
14:45 lucian joined #parrot
15:08 dalek rakudo/nom: 9897b01 | moritz++ | t/spectest.data:
15:08 dalek rakudo/nom: run new test file for native types
15:08 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9897b01e3a
15:09 dalek TT #2199 closed by coke++: [PATCH] 3.9.0 Fails timer tests on NetBSD 5.1
15:09 dalek TT #2199: http://trac.parrot.org/parrot/ticket/2199
15:19 jsut_ joined #parrot
15:40 dmalcolm joined #parrot
16:03 cotto ~~
16:24 dukeleto ~~
16:24 dalek rakudo/nom: 3fb9155 | jnthn++ | src/binder/bind.c:
16:24 dalek rakudo/nom: Fix a bug in the binder related to unboxing to native types. Also add a just-in-case check for something that should 'never happen'.
16:24 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3fb9155dce
16:25 cotto hio dukeleto
16:25 dukeleto cotto: mornin'
16:25 * dukeleto waves from the second day of the Git Together
16:25 dukeleto cotto: how was SPLASH?
16:26 cotto dukeleto: pretty cool.  I got a couple ideas and ran into tewk.
16:26 dukeleto that is awesome
16:27 benabik SPLASH?
16:27 cotto he says he's been keeping up with #ps but hasn't had tuits to do much
16:27 cotto specifically, vmil (virtual machine and intermediate language workshop)
16:28 cotto the dart talk was curious.  Apparently they store a token stream and compile lazily, so there's no bytecode representation.
16:28 benabik O.o
16:29 cotto and the type system is more like a set of heuristics that attempt to be helpful
16:31 cotto The first speaker said that they were trying to make something that'd be familiar to web developers.  It was *really* hard not to shout "like javascript?".
16:32 cotto to be fair, it doesn't look like it'd be a bad language, just unnecessary
16:32 benabik cotto: There's a lot of grumbling about the type system in PLT circles.  They made the same mistakes as Java wrt generics.
16:32 NotFound cotto: or maybe "like winxed" ;)
16:33 cotto there was a group of plt hackers at the conference.
16:33 cotto NotFound++
16:33 cotto NotFound++
16:33 cotto NotFound++
16:33 benabik NotFound: I meant Programming Language Theory, not PLT scheme, sorry.
16:33 cotto ah
16:33 cotto lots of those too
16:33 benabik And I meant cotto, not NotFound.  Bah.
16:34 cotto benabik: there's plenty to grumble about.  They seem to be explicitly ignoring the nerds and going for the average developers.
16:34 benabik List<Car> l1 = new List<Car>(car, van, sports_car);  List<Object> l2 = l1;  l2.push(new Integer(2);  l1.pop() // Explode
16:36 NotFound Some days ago I looked at a language mentioned in a web mentioned in some message, don't remember the name, that was another javascript-alike. Looks like a trend.
16:36 cotto http://try-dart-lang.appspot.com/
16:36 cotto trend.set('winxed');
16:36 NotFound The funny part is that at the same time other people are trying to develop non-javascriptish languages for use in the browser.
16:37 benabik In-browser haskell.
16:37 dukeleto dart is looking to sway the people that think javascript is too hard
16:37 NotFound And coffeescript
16:37 NotFound And a lot others.
16:38 cotto I like coffeescript.  It seems to have the right idea.
16:38 PerlJam cotto: me too
16:38 NotFound So maybe in a few years javascript-alike languages will be used everywhere... except in the browsers X-)
16:38 PerlJam cotto: although it seems a little behind the times in a few areas.
17:01 dukeleto we probably need to apply to Google Code-In soon
17:01 * dukeleto is trying to not forget
17:05 Coke the tech teacher at my local middle/high school will appreciate that.
17:11 cotto https://github.com/blog/967-github-secrets
17:11 cotto that's something I want to bring up at #ps
17:11 cotto there are some significant changes this year
17:11 benabik cotto: github secrets?
17:14 cotto all of them
17:18 bluescreen joined #parrot
17:29 GodFather joined #parrot
17:35 ambs joined #parrot
17:36 dukeleto we need an ideas page to apply to GCI
17:36 * dukeleto is filling out the form right now
17:36 benabik dukeleto++
17:37 benabik What scale of project is GCI aiming for?
17:38 dukeleto https://github.com/parrot/parrot​/wiki/Google-Code-In-Task-Ideas
17:38 benabik We have https://github.com/parrot/parrot/wiki/Projects
17:38 dukeleto benabik: small tasks that would take a crazy smart high-school student that has never touched parrot before a few hours or days to accomplish
17:40 dukeleto this is the old page: http://trac.parrot.org/parro​t/wiki/GoogleCodeIn2010Tasks
17:40 dukeleto trac.parrot.org uses an invalid security certificate.
17:40 dukeleto The certificate expired on 11-09-05 12:54 AM. The current time is 11-10-25 10:40 AM.
17:41 dukeleto (Error code: sec_error_expired_certificate)
17:41 dukeleto BLARG
17:41 cotto one thing is that gci will require orgs to have >= 5 tasks in each category
17:42 cotto that ui doesn't make sense for some projects didn't seem to be a big concern
17:42 cotto http://code.google.com/opens​ource/gci/2011-12/index.html
17:44 cotto also, tasks will be released in two sets, one at the beginning of the contest and one at the middle.
17:44 cotto the idea is to keep orgs to catering to a particular class of student
17:45 cotto that said, they'd be fine if we had a ridiculous number of tasks unevenly distributed across categories
17:47 dukeleto just migrated content to https://github.com/parrot/parrot​/wiki/Google-Code-In-Task-Ideas
17:47 cotto note that this is what I heard from Carol (who's in charge of gci).  I haven't found where it's spelled out online yet.
17:47 benabik Carol++
17:48 pbaker joined #parrot
17:49 dukeleto cotto: the >= 5 tasks is needed to get accepted
17:49 dukeleto can i get some help adding tasks ideas?
17:49 dukeleto the task template is on the above URL
17:49 cotto dukeleto: put it on the wi... nm
17:49 cotto ;)
17:50 cotto for translations, we might have better luck focusing in a single language.
17:53 Coke I was unsatisfied with the translation stuff from last year.
17:53 cotto me too
17:54 cotto I'm not 100% convinced that gci is the best use of our time this year, but if we want to do it, we can do better than we did last year.
17:55 dukeleto well, i just submitted parrot to GCI, so it's gonna happen :)
17:56 dukeleto msg whiteknight what is your melange ID? Can I add you as the backup admin for Google Code-In ?
17:56 aloha OK. I'll deliver the message.
17:56 dukeleto our task page does not have enough tasks to actually get accepted, so we need to get that done Real Soon Now
18:02 benabik_ joined #parrot
18:47 Util joined #parrot
18:55 zpmorgan joined #parrot
19:06 NotFound Why you don't like -1 release number?
19:06 NotFound "Version numbers like 1.4.-1 are somewhat unexpected for users" -> but devel versions are not for users.
19:06 benabik NotFound: I wrote it up in response to lateau wondering why winxed --version gave a -1.
19:07 benabik I don't feel strongly about it.
19:07 NotFound The answer is simple: -1 < any non-neagtive regular release number.
19:09 mj41 joined #parrot
19:09 cotto #ps in 21
19:09 mj41_nb joined #parrot
19:16 jsut joined #parrot
19:24 bubaflub joined #parrot
19:31 cotto #ps time!
19:36 whiteknight joined #parrot
19:36 whiteknight is #ps now?
19:36 cotto yup
19:38 SHODAN joined #parrot
19:52 nbrown joined #parrot
19:52 dukeleto crud. too busy for #ps
19:53 benabik dukeleto: Amusingly, in #ps we were just talking about how busy we all are.
20:02 alester joined #parrot
20:11 nine Too busy to even talk about beeing too busy... Busyness on another level :)
20:12 benabik cotto: We might want to keep our reference documentation in github pages rather than wiki.  Wiki may be good for intro/general user docs so people can update them as needed.
20:14 dalek parrot: 65e6ab7 | NotFound++ | src/pmc/bytebuffer.pmc:
20:14 dalek parrot: don't store value of strstart in ByteBuffer, get it in each usage - issue 182
20:14 dalek parrot: review: https://github.com/parrot/parrot/commit/65e6ab7b61
20:15 whiteknight benabik: it's trivial to mirror docs between the two formats
20:16 whiteknight benabik: they're both git repos. the github pages could be exactly the same stuff, included into pages with formatting
20:16 benabik whiteknight: Do we want to replace the perldoc in .ops and .pmc files with the contents of the gh wiki?
20:19 benabik whiteknight: Moving the contents of doc/ into the wiki, possibly with mirroring, might be useful though.
20:19 SHODAN joined #parrot
20:19 * benabik likes having as few copies of things around as possible.  Copying things is just asking for it to get out of sync.
20:19 NotFound whiteknight: Have you seen winxed issue 12?
20:20 * nine likes having docs with the code
20:20 nine makes grepping useful
20:23 whiteknight NotFound: no. I'll look at it now
20:23 nine whiteknight: have you had time to look at the green_threads merge issue?
20:23 whiteknight nine: no, I've been hella busy
20:23 whiteknight nine: but I'll devote some time to it tonight
20:24 whiteknight NotFound: looks like he's missing a string() cast
20:25 NotFound whiteknight: the problem is that currently string + var is interpreted as var(string) + var, not as string concat. I'm not sure if this is the more intuitive semantic.
20:25 whiteknight NotFound: the rules are self-consistent, but they are different from JavaScript and Perl
20:25 whiteknight NotFound++ on #182 fix
20:26 whiteknight GH #182
20:26 whiteknight how are we referring to Gh tickets now?
20:26 benabik Github itself recognizes gh-182
20:26 benabik Or just #182
20:26 benabik (In comments and commit messags)
20:26 whiteknight gh-182 sounds good to me
20:26 whiteknight eitherway, NotFound++
20:27 NotFound whiteknight: yes, and we risk to break things if changing. I think I'm going to close as not-a-bug.
20:27 nine I'll have a couple of hours for Parrot tomorrow
20:27 whiteknight NotFound: good plan
20:27 * whiteknight has to sign off and go home now.  Will look at nine++'s green_threads issue tonight
20:27 whiteknight later
20:45 dalek nqp: 29d8aed | jnthn++ | tools/build/PARROT_REVISION:
20:45 dalek nqp: Bump to a Parrot version with a fix for ByteBuffer segfaults, which caused programs using Buf to sometimes segfault.
20:45 dalek nqp: review: https://github.com/perl6/nqp/commit/29d8aedaf8
20:45 dalek rakudo/nom: 97aa736 | jnthn++ | tools/build/NQP_REVISION:
20:45 dalek rakudo/nom: Bump to latest NQP.
20:45 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/97aa7365d6
21:02 alester_ joined #parrot
21:02 alester__ joined #parrot
21:05 tadzik hello #parrots
21:05 benabik hellos tadziks
21:07 preflex joined #parrot
21:08 alester ok back
21:56 alester joined #parrot
22:21 soh_cah_toa joined #parrot
22:27 soh_cah_toa_ joined #parrot
22:52 mudkip joined #parrot
22:58 bubaflub joined #parrot
23:10 jsut_ joined #parrot
23:18 benabik joined #parrot

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

Parrot | source cross referenced