Camelia, the Perl 6 bug

IRC log for #parrot, 2011-10-26

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:12 Themeruta joined #parrot
00:12 whiteknight joined #parrot
00:16 whiteknight good evening, #parrot
00:24 cotto hio whiteknight
00:24 whiteknight hello cotto
00:24 benabik o/ peoples
00:28 whiteknight hello benabik
00:28 bacek_at_work joined #parrot
00:29 aloha joined #parrot
00:33 whiteknight NotFound: any progress on that segv?
00:43 dukeleto ~~
00:43 * dukeleto waves from the SJC airport
00:53 schmooster joined #parrot
01:15 PerlJam joined #parrot
01:16 Coke joined #parrot
01:17 Util joined #parrot
01:43 tadzik joined #parrot
01:59 whiteknight left #parrot
03:02 alvis evening, #parrot
03:03 alvis cotto: I reviewed #ps earlier, and I will make every effort to join #ps in the future.
03:04 alvis cotto: Also, I saw bubaflub's remarks.  I will send out an email in the morning asking for input and direction from anyone interested in contributing to Parrot's docs.
03:13 schmooster joined #parrot
03:15 cotto ~~
03:16 cotto alvis, great!  Sorry I forgot to mention #ps in my response on parrot-dev.
03:16 cotto does the time work well for you in general?
03:27 bubaflub joined #parrot
03:35 alvis cotto: I can make the time.
03:38 cotto alvis, ok.  What timezone are you in?
03:39 alvis cotto: Central Standard.  I live in North Central Texas.  So, it'll be fine.  I'll arrange things to make it work out.
03:39 cotto ok.  That's easier than Europe or Australia.
03:40 alvis cotto: I also *try* to be more accessible here, in #parrot.
03:40 alvis cotto: Yeah :-)
03:40 cotto cool.  You can't do more than try.
03:41 soh_cah_toa_ alvis: great, i'm always so happy to see our docs being improved :)
03:42 alvis soh_cah_toa_: I'm just glad it's such a clear priority, i.e., it's on the agenda, so-to-speak.
03:43 alvis I was a'bit concerned there for a
03:43 soh_cah_toa_ avlis: something like that always is since it's never "done" ;)
03:43 alvis make that a'bit.
03:43 alvis soh_cah_toa_: Ha!  Too right! :-)
03:44 cotto alvis, have you signed a cla?  I don't know if Parrot was using them back when you were involved.
03:44 bubaflub alvis: are you working on getting documentation together?
03:45 alvis cotto: No, I haven't.  I saw something about it, but never followed up.  With whom do I get in touch, again?
03:45 cotto aloha, cla?
03:45 aloha cotto: cla is http://www.parrot.org/sites/www.​parrot.org/files/parrot_cla.pdf
03:45 cotto alvis, ^
03:46 alvis cotto: Ok, will check it in a few and follow directions.
03:46 cotto alvis, thanks.  I look forward to giving you a commit bit in the near future.
03:46 alvis bubaflub: Yes, with help, input, and direction of many, many others. :)
03:47 bubaflub alvis: i'd like to help out
03:48 alvis bubaflub: Sounds great!  I'll send out an email to parrot-dev in the morning, asking for help, direction, suggestions, yada, yada, for any and all interested.
03:48 cotto alvis, if the lack of a commit bit starts to keep you from getting anything done, let dukeleto or me know.
03:48 alvis cotto: Will do.
03:48 bubaflub alvis: great.  i'm not too familiar with parrot guts, but would be more than happy to read over code and learn.
03:49 cotto bubaflub, then you're the best kind of person to work on the docs
03:49 bubaflub cotto and alvis: i just need someone to point me in the right direction, copy edit, clarify and what not.  i have no idea where to start but i like what soh_cah_toa++ has done with the github wiki
03:49 alvis bubaflub: Much the same as my own thoughts.  I figure the least-harm, best way to get a handle on Parrot is through a doc project.
03:50 soh_cah_toa_ alvis: absolutely
03:51 bubaflub alvis: great.  we can coordinate after the parrot-dev email
03:52 alvis soh_cah_toa_: :)
03:52 alvis bubaflub: Sounds good.
03:54 cotto alvis, one thing I'd like to have as a starting point would be a goal for how the docs should be organized so that we can work toward that.
03:55 alvis cotto: Agreed.  I thought I'd make that a prominent point in the email, trying to solicit a consensus among the community.
03:55 cotto having (some subset of) them in a github wiki (possibly included in parrot as a submodule) would be contributions easier, especially since github speaks Pod.
03:55 cotto alvis, ok.  Just don't wait for too much consensus.  ;)
03:56 alvis cotto: Ha! Ok, got it! :)
04:00 alvis cotto: The CLA is perfectly acceptable.  I will execute and fax the same, sometime tomorrow afternoon, CST.
04:01 soh_cah_toa_ alvis: btw, gitster sent a patch a while ago regarding a few improvements that could be made to docs/project/git_workflow.pod: https://github.com/gitster/parrot/commit/​27a9f10d819ee63e5723821f8d616477b30ee7b1
04:01 soh_cah_toa_ alvis: just an example of an area that needs improvement
04:02 alvis soh_cah_toa_: Yeah, I saw that.  Agreed.
04:03 soh_cah_toa_ good
04:03 cotto soh_cah_toa_, some of his issues are because we use Git terminology in a way that's inconsistent with how Git hackers do.
04:03 alvis soh_cah_toa_: Do you know whether or not dukeleto resolved any questions in his mind about it?
04:03 soh_cah_toa_ yes, exactly
04:03 soh_cah_toa_ alvis: not sure
04:04 cotto he probably got to hang out with gitster earlier today
04:04 alvis soh_cah_toa_: Okay, I'll visit with him when he has a few more tuits.  I think he's a'bit occupied right now. :)
04:04 soh_cah_toa_ sure
04:05 alvis cotto: Oh, cool.
04:06 cotto alvis, are you alvis on github?
04:06 alvis cotto: Sorry, no.  ayardley
04:07 cotto there you are
04:07 alvis cotto: Fwiw, I'm fairly new to github.  Spent most of the last two years in asdf and quicklisp. :)
04:09 soh_cah_toa_ alvis: ah, you'll come to love it. github is great
04:09 soh_cah_toa_ it just doesn't compare to other git repos ;)
04:10 alvis soh_cah_toa_: yeah, it really looks like they did it right.
04:11 soh_cah_toa_ alvis: anyway, i gotta head out. i'll be around to help tomorrow
04:11 soh_cah_toa_ see ya
04:11 soh_cah_toa_ left #parrot
04:11 alvis soh_cah_toa_: Take care.
04:44 cotto http://www.jwz.org/blog/2011/10/jmc-rip/
04:50 alvis cotto: Wow! Missed that completely. Haven't checked Hacker News in a coupl'a days. (Really should use an RSS Feed).  Thank you.
04:53 alvis Truly, a brilliant man, imo. (But, truth-be-told, if it weren't for Perl, I'm not sure I'd've actually made my way back to Lisp. :)
05:17 SHODAN joined #parrot
05:26 alvis G'night, #parrot
05:27 cotto 'night, alvis
07:48 wagle joined #parrot
08:08 mj41 joined #parrot
08:08 mj41_nb joined #parrot
08:20 lucian joined #parrot
08:25 lucian_ joined #parrot
09:20 nopaste joined #parrot
09:29 lucian__ joined #parrot
09:36 dalek nqp: fd63797 | moritz++ | src/6model/reprs/P6 (2 files):
09:36 dalek nqp: fix copy&pastos
09:36 dalek nqp: review: https://github.com/perl6/nqp/commit/fd637972d3
09:57 ambs joined #parrot
10:27 ambs_ joined #parrot
10:35 jsut joined #parrot
11:04 lucian__ joined #parrot
11:43 Psyche^ joined #parrot
11:43 nine Good afternoon, #parrot
11:56 JimmyZ joined #parrot
12:07 whiteknight joined #parrot
12:09 nine Good morning, whiteknight
12:12 whiteknight hello nine
12:12 whiteknight last night fell apart for me. Sorry I didn't look at green_threads
12:13 nine Thought so :) Having a look myself right now
12:15 whiteknight did you see the email from moritz about how the green_threads branch is 6% slower than master?
12:17 nine whiteknight: yep. Disturbing... Will have a look at that as soon as I know with which tree to continue. May have something to do with alarms which seem to generate much more ticks than they should. Noticed that when debugging the (supposed) zero overhead patch.
12:18 whiteknight okay, so long as you have ideas about where to look
12:18 whiteknight running a whole cachegrind profile on it will be a bit of a pain, and I can't do it on this hardware
12:19 nine Too bad that Bulldozer is such a disappointment. Was planing to buy some mighty 8 core to speed up development...
12:20 whiteknight Bulldozer?
12:20 nine AMDs new processor architecture
12:21 nine Have been holding back on my system upgrade while waiting for it to arrive. But it would probably be a waste of money since it just cannot live up to the expectations.
12:21 moritz currently Intel seems to be ahead in the CPU race
12:23 nine How on earth can there be merge conflicts in src/hash.c when I never even touched that file in my branch?
12:24 whiteknight I'm booting my VM up now, I'll look at it as soon as it's up. What branch are you working in right now?
12:26 nine whiteknight: last week I tried merging current master into my local green_threads before seeing that you already did the same. Since all tests passed in mine, I tried to upgrade again to current master with lots of conflicts.
12:26 nine whiteknight: I think I'm missing some git foo. Working with different remote repos is quite new to me.
12:26 whiteknight I didn't have many conflicts in my merger.
12:27 nine last week it was just src/pmc/scheduler.pmc and src/scheduler.c. Same as in your merger.
12:27 whiteknight okay
12:27 whiteknight that's not many
12:28 whiteknight I just updated to master again, several new conflicts. Probably from the TT #1910 merge
12:28 nine Today I got conflicts in all kinds of files. Including the ones where I already fixed the conflicts. I think I screwed up git somewhere
12:29 nine whiteknight: including one in Changelog?
12:31 whiteknight no
12:31 whiteknight just src/embed.c, src/scheduler.c and src/pmc/scheduler.pmc
12:32 nine I think I'll just throw away my green_threads.merged and continue with your's :)
12:32 whiteknight give me a few minutes, I'll push the updated version as soon as it builds
12:32 nine ok, thanks
12:33 bubaflub joined #parrot
12:38 dalek rakudo/nom: cbc652c | moritz++ | docs/ChangeLog:
12:38 dalek rakudo/nom: add some ChangeLog entries
12:38 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cbc652c834
12:39 dalek Heuristic branch merge: pushed 38 commits to parrot/green_threads by Whiteknight
12:40 whiteknight that last commit message won't inspire any confidence
12:45 nine :)
12:46 whiteknight I did a "git pull origin green_threads" to update it, then I did a "git merge master" which gave me several of the same exact conflicts I had seen the other day, then I did a "git pull origin green_threads" again, which updated more stuff and gave me more conflicts
12:46 whiteknight so I have no idea what the hell happened
12:49 nine so you got pretty much the same git strangeness I got
12:53 moritz btw I strongly recommend   git config --global rerere.enabled 1
12:53 moritz that will record your resolution to merge conflicts
12:53 moritz very handy
12:53 * nine is looking forward to the new fulltest harness
12:55 nine wow... moritz++ # thanks
12:55 moritz (it won't explain the weirdnesses, but makes the handling of the weirdnesses much less hassle)
12:55 nine whiteknight: your current green_threads passes fulltest for me :)
12:56 whiteknight does it? awesome
12:56 whiteknight okay, so let's get to the hard work of figuring out where that slowdown came from
12:57 nine on it :)
12:57 whiteknight nine++
13:04 bluescreen joined #parrot
13:20 lucian_ joined #parrot
13:27 benabik joined #parrot
13:30 benabik I think some of the green_threads conflicts are because kid51 did some codingstd fixes in the branch and I did the same ones in master.  (Could be wrong)
13:30 benabik (Good morning, BTW)
13:31 whiteknight good morning, benabik
13:31 whiteknight I think we have the conflicts sorted out. I don't think they are codestd-related
13:32 benabik Ah.  Thought they might have been since src/hash.c was one of the places I had to sort out.
13:35 nine Funny how some 20 years of programming can change one's thinking from "no need to waste a character, I know the language!" to "oh, it might be more clear if I state the > 0 explicitely"
13:36 whiteknight nine: yeah, once your harddisk is bigger than you could ever possibly fill, adding in a little extra clarity suddenly makes a lot of sense
13:36 benabik nine: I tend to find that attitude changes very quickly when you give them someone else's code to work on.  :-D
13:37 nine benabik: oh yeah, that help tremendously :)
13:37 whiteknight of course, you still run into the exact opposite situation when coding JavaScript: "Every single extra byte means more bandwidth and slower loading times. I must squeeze the crap out of this code above all else!"
13:38 nine whiteknight: that will pass as well...
13:38 whiteknight some of the tricks people use to smoosh javascript are rediculous
13:38 benabik Most people seem to at least be using JS minimizers for that instead of doing it by hand.
13:39 whiteknight benabik: you'd be surprised, especially for small JS snippets included inline in HTML pages
13:39 nine whiteknight: some JavaScript code people produce is ridiculous to begin with
13:39 benabik whiteknight: That's funny given how much space HTML takes up by itself…
13:39 benabik Really they should just enable compression in apache and move on with their lives.
13:40 nine benabik: and how much more bandwith is wasted by non-optimized images
13:40 whiteknight it's the same mindset as the people who say "C/assembly is always faster, therefore I'm going to write all my software in C/assembly"
13:41 whiteknight there are plenty of people who don't know their craft well enough that they blindly follow rules like this without understanding them.
13:41 benabik nine: Or those little 5x5 images people use for tricks they can't get CSS to do.  Or the people who use flash applets for headers so they can have "just the right font".
13:41 * benabik HATES people who use flash for <h2> tags.
13:42 * whiteknight hates people who build 100% flash websites for restaurants
13:42 nine Speaking of ridiculous: how could I ever push such code as my original zero overhead patch? This code is awful!
13:42 moritz it (seemd to) work
13:43 nine moritz: it's an ugly hack patched to sort of work before really understanding whats going on, done in a hurry just to get it finished.
13:46 whiteknight nine: you just described 99% of all software
13:46 whiteknight at least, all commercial software
13:47 benabik Only like 95% of F/OSS.  :-P
13:47 whiteknight FOSS at least has the benefit of having fewer "do this or you're fired" deadlines to meet
13:47 benabik And more having to show your code to others.
13:50 nine Well, I guess I'm lucky to have a dayjob where there hardly is any time pressure and the manager mandates code reviews for every patch and is quite anal about code quality. Mostly by being said manager :)
13:50 benabik I wish Parrot became vital to somebody so they could pay someone work to full time on it as maintainer.
13:50 whiteknight at my job, we put all our code review tasks into a queue, so we can ignore them all at once after launch
13:50 benabik :-D
13:51 whiteknight we don't want to waste time ignoring them when there is real work to be done
13:51 whiteknight what's funny about code review is that the people who need to spend time reviewing are the most productive and most experience coders. The inexperienced coders writing shit code don't do any review
13:52 benabik +1
13:55 nine whiteknight: I'm handling this problem by making the one who reviewed the code responsible for it and encouraging all team members to do reviews. So people learn how to write good code by really trying to find faults in other's code.
13:56 moritz nine: you don't happen to be looking for more developers? :-)
13:56 whiteknight nine: that's a good system, I think. You still have the situation where the people who find the most bugs are the people with the most experience
14:03 nine moritz: well I guess the drawback is that we don't pay that well ;)
14:04 moritz nine: and that you're probably not anywhere near my $home
14:05 nine whiteknight: of course. But I was still surprised at the quick progress. People started to getting it right the first time more often than not after just two months or so.
14:12 nine moritz: where are you based?
14:12 moritz nine: Germany
14:13 nine Well Linz/Donau is not that far away ;)
14:13 benabik Heh.  "GitHub employees have gotten very good at writing irrelevant code."
14:14 moritz nine: ~270km it seems :-)
14:18 nine moritz: http://www.atikon.com/content/webdes​ign/atikonisten/jobs/index_ger.html (Webprogrammierer) just in case...
14:19 moritz nine: the job description doesn't sound very exciting, to be honest
14:21 benabik aloha: dalek?
14:21 aloha benabik: dalek is run by sorear.
14:21 nine moritz: not very exiting? To be honest I would never apply for a job described like that... I never managed to convey the essentials to our HR person.
14:21 benabik aloha: aloha?
14:21 aloha benabik: aloha is simple Bot::BasicBot::Pluggable hosted on https://github.com/bacek/aloha
14:22 * benabik is poking at hubot.  Likely nothing will come of it.
14:24 nine moritz: we're doing Catalyst applications mostly. Most interesting project has been replacing our Python/ZoDB based CMS with a Perl/PostgreSQL based one using (and fixing) Inline::Python to maintain compatability with the existing > 1000 websites. A real two year stunt and I cannot imagine many companies where one would be allowed to try such things.
14:25 moritz nine: now that does sound better :-)
14:28 nine If only we would write such things into the job description :) We've been looking for a new programmer unsuccessfully for two years before simply managing to get by with the four we have.
14:29 whiteknight I've never applied to a job where the description was even remotely what the job actually entailed
14:29 whiteknight The first job I had was advertised as "Embedded C Developer". The actual job was to write web services in C#
14:30 whiteknight somewhere, I think the essence of the job was miscommunicated to the HR people
14:31 benabik whiteknight: Did you find out at the interview or after being hired?  :-D
14:31 Coke I think I've only ever applied based on a job description 2x ever. First job, and the scramble once when I was laid off.
14:31 JimmyZ most at the interview
14:31 Coke Mostly, it's been word of mouth.
14:32 whiteknight benabik: after being hired. The real obnoxious part was that I both enjoy embedded programming more, and consider myself better at it than I do at web development
14:32 whiteknight and I eventually had to be the interviewer for new hires who were taking the job I wanted
14:34 benabik whiteknight: sadface
14:35 whiteknight benabik: tell me about it.
14:42 whiteknight regex question: I want a regex that matches all html-alike tags EXCEPT <sub></sub>
14:43 whiteknight is there a way to put such a thing into a single regex? I can't think of anything obvious
14:43 benabik Uhm...
14:43 benabik Perl/pcre, I assume?
14:43 whiteknight for our purposes, close enough
14:44 benabik /<(?!sub)[^>]*>/ ?
14:44 moritz whiteknight: <(?!sub)\w+>
14:44 moritz depends on whether you allow attributes etc.
14:44 whiteknight moritz: (?! ) is a zero-width, negative match assertion?
14:44 benabik whiteknight: Yes.
14:45 moritz whiteknight: yes
14:46 whiteknight thanks, I'll try that. moritz++, benabik++
14:46 benabik I'd love a p6cre…  /'<' <!before 'sub'> \w+ '>'/ is a bit more verbose, but easy to understand.
14:49 benabik Is there a dalek repo somewhere?  I'd like to poke at getting issues/pull requests on channel.
14:50 moritz benabik: https://github.com/Infinoid/dalek-plugins
14:51 benabik moritz++
14:52 benabik Oh.  Dalek uses the RSS feeds instead of the service hooks?
14:53 moritz benabik: it can use both
14:53 JimmyZ does it also not match subXXXX ?
14:54 whiteknight I needed to do <(?!sub)(?!/sub)[^>]*> to get the beginning and end tags
14:54 moritz benabik: there's a short explanation of how to set up the service hooks in the perl6/mu repo, somehwere in misc/dalek-*
14:54 moritz whiteknight: <(?!/?sub)  is just as good
14:54 whiteknight moritz: whatever. I'm just happy that it works at all
14:55 whiteknight moritz: the regex is in an SQL query, which calls an SQL function, which calls the String.Replace method from C#.
14:55 whiteknight so, we're already in a mountain of ugly code
14:55 moritz "fun"
14:56 whiteknight yes, with quotes
14:57 whiteknight despite the fact that I rarely use perl anymore, at work I'm known as the "perl guy" and the "regex expert"
14:58 moritz for the outsider, it doesn't take much to become "the perl guy"
14:58 moritz if you are at YAPC::Somewhere, it takes a lot more to be recognized as an expert :-)
14:59 whiteknight yes, I would never refer to myself as an expert.
14:59 whiteknight in terms of perl, I wouldn't refer to myself as "competent"
15:01 dmalcolm joined #parrot
15:13 dukeleto ~~
15:14 Coke whiteknight: I also am both those things here, despite a dire lack of perl coding for work in the last 10 years.
15:18 nine Anyone missing a reason to bang his head to the table? http://forums.devarticles.com/c-c-h​elp-52/how-to-pause-in-c-25320.html
15:22 whiteknight ...srsly?
15:22 benabik | less ?
15:22 whiteknight system("pause")?
15:22 benabik Oh, is this in MSWin?
15:23 dukeleto THE HORROR
15:23 moritz looks like
15:25 nine I also very much liked the busy waiting on getch()
15:26 lucian joined #parrot
15:26 whiteknight the busy wait on getch() eats more ram. getchar() might yeild to the OS
15:26 whiteknight a system that doesn't busy wait on a modern system would always be better
15:27 benabik Yeah…  I was wondering why get char never got suggested.
15:28 moritz just configure the damn window to not terminate upon program completion
15:29 benabik Heh.
15:29 moritz so that the users of sane OSes don't have to deal with your read-from-stdin crap
15:30 nine moritz: that's the answer I was missing most :)
15:30 moritz nine: thought so.
15:30 whiteknight but that solution doesn't let the window close on keypress
15:30 whiteknight if he wants the window to pause then close on keypress, that's a different thing
15:31 moritz whiteknight: there's always Alt+F4
15:31 moritz which is also a nice keypress
15:31 moritz which has the advantage of closing most windows on windows, not just that one
15:31 whiteknight moritz: no question. But "press any key" is different from "Press ALT+F4"
15:31 nine whiteknight: the stated problem was that the window closes before he can see the output.
15:33 lucian_ joined #parrot
15:34 whiteknight nine: true. but different solutions to that problem have very different sideffects. He wasn't clear enough about what he wanted
15:35 * moritz found it very amusing that declaring a separate variable and then doing cin >> m; was found to be superior to simply reading a line from cin
15:38 whiteknight moritz: it's the blind leading the blind
15:38 moritz ... into the blind
15:38 whiteknight :)
15:38 moritz erm, dark
15:38 moritz whatever
15:42 NotFound joined #parrot
15:46 dalek parrot: 393e5e7 | NotFound++ | src/pmc/bytebuffer.pmc:
15:46 dalek parrot: one more partial fix for issue 182
15:46 dalek parrot: review: https://github.com/parrot/parrot/commit/393e5e7081
15:50 dalek parrot: 038c46f | NotFound++ | t/op/stringu.t:
15:50 dalek parrot: print some info in non-passing case of ucs4 tests
15:50 dalek parrot: review: https://github.com/parrot/parrot/commit/038c46ffb4
15:50 nine Ok....fulltest passes. Time to build rakudo and start benchmarking
15:54 whiteknight nine++
15:55 alvis Hello, #parrot
15:56 whiteknight hello alvis!
15:56 dukeleto alvis: wazzup
15:57 alvis I've just fired off an email to parrot-dev, regarding Parrot's Documentation set and would sincerely appreciate your input/feedback.
15:57 whiteknight alvis++ on the email
15:57 NotFound Talking about documentation, there is something about encoding partial_scan ?
15:58 alvis whiteknight & dukeleto: Hey.
15:59 alvis NotFound: I'm sorry, I don't understand.
16:00 moritz alvis++  means giving alvis karma
16:00 moritz alvis: karma alvis
16:00 moritz sorry
16:00 moritz aloha: karma alvis
16:00 aloha moritz: alvis has karma of 4.
16:01 NotFound alvis: is a general question, not directly to you.
16:01 moritz so, that karma is counted, and is a coarse measure of appreciation
16:01 moritz oh, I completely misunderstoof context
16:01 dukeleto alvis: are you in CREDITS ?
16:01 moritz never mind me
16:01 alvis moritz: Ha! That's kind'a cool. :-)
16:02 NotFound Please don't use html mail in messages to parrot lists.
16:02 alvis dukeleto: Naw, not yet.  Haven't accomplished anything yet.  Just some words here and there.
16:02 dukeleto alvis: gotcha
16:02 dukeleto alvis: also, just to be clear, our CLA does not involve copyright assignment. Our CLA is to protect developers from being sued by nasty companies
16:03 dukeleto alvis: so you still own the copyright on all your contributions to Parrot. Just wanted to make sure that was explicitly explained to you.
16:03 bluescreen joined #parrot
16:03 moritz alvis: I have some experiences to share regarding the state of documentation, but will need to commute first
16:04 NotFound dukeleto: about the comment on #182: I don't know how to write a test for a segfault I can't reproduce.
16:05 dukeleto NotFound: fair enough
16:05 * dukeleto can dream, can't he?
16:05 alvis NotFound: I'm guessing that's to me.  Very sorry.  Hadn't seen anything to the contrary (but, I reckon' I should figured that.)
16:06 dukeleto alvis: parrot.github.com is the new parrot.org
16:07 dukeleto alvis: we haven't had the resources to put towards it, though
16:07 alvis dukeleto: Yes, it was clear.  I'll fax it in later this afternoon.
16:07 dukeleto alvis: can you give it some love?
16:07 whiteknight dukeleto: I don't think that's "official" yet, but I would like to see it explored seriously
16:07 dukeleto whiteknight: there is no "official". Only what actually happens. And I know how much we all love Drupal...
16:07 alvis moritz: Ok, sounds good.
16:08 whiteknight dukeleto: I'm just saying, we've not taken a vote or anything
16:08 dukeleto whiteknight: when was the last time we voted on something?
16:08 dukeleto whiteknight: i thought it was "make stuff better and ask for forgiveness" ?
16:08 alvis dukeleto: Really?  I did not understand that. So, parrot.org is going away?  Is the foundation dropping the domain?
16:08 whiteknight dukeleto: We vote on things all the time at #ps
16:08 NotFound Can someone take a look at src/string/encoding/ucs4.c:223 ? That >> 1 shouldn't be >> 2 ?
16:09 dukeleto whiteknight: indeed. I didn't mean to say "parrot.github.com will replace parrot.org website", just that I want to work on it more, so we can ask people to vote between them
16:09 dukeleto whiteknight: i am not quite awake yet, forgive me
16:10 whiteknight dukeleto: strong agreement from me. That's why I started putting stuff there in the first place
16:10 * dukeleto went into sleep deprivation this weekend between the GSoC mentor summit and the Git Together
16:10 dukeleto whiteknight++ # dragging us screaming and kicking into the future
16:10 dukeleto alvis: https://github.com/parrot/parrot/parrot.github.com is the repo
16:10 dukeleto alvis: so every time you commit to that repo, the website at parrot.github.com magically updates
16:11 dukeleto alvis: consider it your playgroud/sandbox for designing a new parrot web documentation experience
16:12 alvis dukeleto: Oh, ok.
16:13 alvis dukeleto: As for love, I *am* willing to contribute some $$$ to revising/redesigning the parrot.org website, IFF
16:13 dukeleto alvis: can you do graphics?
16:13 alvis dukeleto: that's appropriate and there's some kind'a consensus established on it.
16:14 nine moritz: what did you use to measure runtime of mandelbrot-color.pl? Just time?
16:15 alvis dukeleto: middlin', at best.  Once upon a time, I owned a mid-sized ISP with a graphics design shop.
16:16 alvis dukeleto: But ... I live 4.5 miles for a University with starving, graphics-design students. ;-)
16:16 whiteknight I don't think we need many graphics
16:17 alvis whiteknight: I absolutely agree.  But that's just my pov. :)
16:18 alvis whiteknight: *My* preferences lean towards something like what the LLVM folks have done.  But, again, that's just me.
16:18 dukeleto cotto++ and I would like a new logo, but that is just, like, our opinion, man
16:19 alvis dukeleto: Agreed.  The pigeon's gott go.
16:19 whiteknight new logo?
16:20 alvis make that, the pigeon has got to go.  (Sorry, just can't type, again.)
16:21 dukeleto i like the color scheme, but we need to mix it up a bit
16:21 whiteknight we don't need to go playing with CSS twiddling just for the sake of change
16:21 whiteknight the problem with our documentation isn't the colorscheme
16:21 dukeleto whiteknight: i am talking about our logo, nothing to do with our docs
16:21 dukeleto whiteknight: logo is a separate issue from our docs
16:21 NotFound It will be nice to have it in a scalable format, BTW.
16:22 whiteknight dukeleto: okay, I misunderstood. Yes, a new logo would be awesome
16:22 whiteknight do we have any candidates?
16:22 dukeleto NotFound: we have that, but it is in the parrot foundation dropbox. I could put it in the parrot.git repo if people want it
16:22 dukeleto whiteknight: nope, just wishes for unicorn farts at the moment
16:22 whiteknight something like this, but stylized and digitized would be awesome: http://www.redbubble.com/people/bevvie​/art/1934337-sulphur-crested-cockatoo
16:23 dukeleto whiteknight: oooh!
16:23 dukeleto i like
16:23 dukeleto i imagine a robotic-ish version of that
16:23 NotFound dukeleto: not in the repo, but in some accesible location.
16:23 dukeleto NotFound: shall I make some kind of graphics repo in the parrot github org?
16:24 whiteknight there are lots of types of birds that are technically "parrots", but don't look like popular parrots
16:24 cotto ~~
16:24 alvis whiteknight: that's nice.
16:24 NotFound dukeleto: not a bad idea, that way people can add proposal and variants.
16:26 cotto I was thinking something like what http://sidneyeileen.deviantart.co​m/art/Minimalist-Cat-11-54382075 but of a parrot in flight.
16:28 dukeleto NotFound: https://github.com/parrot/art
16:29 dukeleto NotFound: i will put stuff in there Real Soon Now
16:29 NotFound I have something in that style, but not flying: http://blassic.org <--> http://es.wikipedia.org/wiki/Vanellus_chilensis
16:29 NotFound A friend designed it for me.
16:30 alvis NotFound: I really should just go back to pine (or alpine, now).  I will --html and ++text in a few and resend.
16:31 dukeleto alvis: we should probably standardize on README.pod in parrot repos.
16:31 dukeleto alvis: all Parrot docs in parrot.git should be in POD
16:31 moritz nine: just time
16:32 dukeleto alvis: i am answering some of your email questions here, since my tuits are low today
16:32 NotFound (The reason for choosing a "tero" for the Blassic logo was some absurd jokes and puns on irc in spanish hard to translate)
16:37 alvis dukeleto: Ok.
16:38 alvis Grrr! Too helpful email clients.  (I'm going back to alpine this weekend.)
16:38 NotFound dukeleto: novels? Will be nice.
16:41 * dukeleto really wants a Parrot VM graphic novel
16:42 * dukeleto has asked if it is ok to put all the logo/graphics from PaFo in the new art.git repo.
16:42 NotFound Including a chapter about the pie in the face of some guy? ;)
16:42 dukeleto NotFound: yes, please
16:44 nine New benchmark results: Patched version is slower than the original by 0.84%, signficance 4.2σ
16:45 nine moritz: would that be acceptable for Rakudo?
16:46 fperrad joined #parrot
16:47 ambs joined #parrot
16:55 moritz nine: much better :-)
16:55 moritz no idea about "acceptable"
16:59 NotFound Has someone tested winxed examples/fly.winxed with that bramch?
17:02 whiteknight no
17:03 dalek rakudo/nom: 6b78ad0 | jnthn++ | src/Perl6/Actions.pm:
17:03 dalek rakudo/nom: Fix code-gen for assignment to native types.
17:03 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6b78ad0e37
17:09 pbaker joined #parrot
17:11 Coke me dislikes html formatted email formatted to look like monospace.
17:15 whiteknight I really dislike html formatted email where the text color is set to black, because I use a mail-reader with a black background
17:15 * Coke sees notfound covered this more succinctly earlier.
17:16 NotFound Fixes all problems in one shot
17:19 dalek rakudo/nom: a4602de | jnthn++ | src/Perl6/Metamodel/NativeHOW.pm:
17:19 dalek rakudo/nom: Helps to actually compute the MRO before using it. :-) This fixes various issues with native types.
17:19 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a4602ded5f
17:19 alvis Tough crowd.  I'll get'er fixed.
17:19 whiteknight alvis: no big deal
17:20 dukeleto alvis: mostly we use plain text email on parrot-dev
17:20 dukeleto alvis: we should probably state that somewhere in our docs ... ;)
17:21 alvis dukeleto: yeah, i see that. :-)
17:21 alvis dukeleto: yeah, i was just greppin' through the docs 'bout that.
17:21 whiteknight that's hardly a community standard. It just happens to be what most individual users choose
17:22 whiteknight html-formatted emails are fine
17:24 NotFound I said "please" ;)
17:26 alvis NotFound: Yes, you did.  Thank you. (Don't worry folks. I'm much more thick-skinned than that. I *was* a Criminal Defense attorney for five years.)
17:27 cotto I don't care as long as the plaintext version doesn't look like garbage or say "your email client doesn't support html email, you need to use a bettar outlook lulz".
17:27 whiteknight alvis: still, we shouldn't be in the business of telling other people how to format their own emails
17:27 whiteknight it's a waste of time and good-natured energy
17:29 alvis whiteknight: :)
17:29 whiteknight if we want to criticize people we should do it the old-fasioned way: criticize the code they produce, and make humorous exaggerations about each others mothers
17:30 NotFound Or just good-old direct insults.
17:30 NotFound You assholes!
17:31 whiteknight :)
17:32 alvis Ok, folks. gtg.  (Btw, it's Game 6 of the World Series tonight, so I'm liable to be ... unavailable.)
17:33 whiteknight later
17:56 nopaste joined #parrot
17:59 TonyC joined #parrot
18:41 bluescreen joined #parrot
18:47 contingencyplan joined #parrot
19:27 dalek rakudo/nom: 158bd07 | moritz++ | src/Perl6/Grammar.pm:
19:27 dalek rakudo/nom: protect "grammar" etc with <.end_keyword>
19:27 dalek rakudo/nom:
19:27 dalek rakudo/nom: cognominal++ noted that it is applied too late (after the package_declarator:sym<...>
19:27 dalek rakudo/nom: action method has fired)
19:27 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/158bd07eef
19:49 bluescreen joined #parrot
20:00 nine whiteknight: pushed my zero overhead rework. fulltest passes. overhead reduced from 6.4 % to 0.84 %. Rakudo fulltest passes
20:00 benabik nine: I guess it's a <1% overhead patch now?
20:00 whiteknight nine: awesome. I think we definitely talk about merging soon
20:00 nine benabik: well, the number at least starts with a 0 :)
20:00 whiteknight We can optimize master by 1% to offset
20:01 whiteknight I don't know *how*, but we'll find a way
20:01 nine Maybe I can find a way to optimize further.
20:02 whiteknight I'll start the wheel turning on a merger tonight. Anything you can do before then will be candy
20:14 benabik Whee…  that ended up far more wordy than I expected.
20:14 nine Well, bed is calling. Good night #parrot
20:20 jlaire joined #parrot
20:23 whiteknight goodnight nine
20:24 whiteknight I'm out too. Later
20:44 jlaire joined #parrot
20:53 aantn joined #parrot
20:54 nbrown joined #parrot
20:54 perlite_ joined #parrot
20:56 Lma joined #parrot
21:17 aantn joined #parrot
21:29 soh_cah_toa joined #parrot
21:48 soh_cah_toa msg whiteknight hey, i was wondering if it were possible for you to work on the missing sections in your winxed tutorial. once i'm done w/ my featherspec work, i'd like to rewrite plumage in winxed. however, i'm not too experienced w/ winxed. i really like your winxed tutorial so it'd be great if it were more complete when the time comes :)
21:48 aloha OK. I'll deliver the message.
22:01 cotto soh_cah_toa: there's no better way to learn than by doing
22:02 soh_cah_toa cotto: agreed, but i'd like to have some kind of headstart ;)
22:15 mudkip joined #parrot
22:19 dalek winxed: 873047e | NotFound++ | winxedst1.winxed:
22:19 dalek winxed: convert to appropiate type in compile time evaluation of indexof builtin, and add functions for such conversions
22:19 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/873047eb68
22:20 NotFound The way to find the missing sections is to tell when you don't find them ;)
22:21 whiteknight joined #parrot
22:22 NotFound soh_cah_toa: we don't require years of experience with winxed for this job. Mostly because nobody has ;)
22:23 soh_cah_toa yeah, it just seems like there's a few "gotchas" that need to be understood
22:23 jsut_ joined #parrot
22:23 NotFound I remember a job offer that required 5 years of experience with java, when java had 3 years of public life,
22:23 soh_cah_toa ha! that's great :)
22:23 NotFound "We need a liar"
22:28 NotFound soh_cah_toa: if you want to be even more confused you can take a look at winxedxx
22:32 soh_cah_toa NotFound: what's that?
22:32 NotFound soh_cah_toa: C++ backend for winxed
22:33 soh_cah_toa oh yeah, i've already gone through some of that. it's *way* beyond my skill level
22:33 NotFound For a subset of winxed, better said.
22:44 bubaflub NotFound: is winxedxx on github?
22:45 NotFound bubaflub: https://github.com/NotFound/winxedxx
23:14 jlaire joined #parrot
23:16 jlaire joined #parrot
23:26 nbrown_ joined #parrot
23:28 eternaleye_ joined #parrot
23:29 rfw joined #parrot
23:29 AzureStone joined #parrot
23:29 whiteknight good evening, #parrot
23:29 TiMBuS joined #parrot
23:29 perlite joined #parrot
23:29 Themeruta joined #parrot
23:30 simcop2387 joined #parrot
23:30 Hunger joined #parrot
23:30 smash joined #parrot
23:30 autark joined #parrot
23:31 whiteknight soh_cah_toa: sure, I'll gladly get back to work on that tutorial. Any particular portions you want me to focus on first?
23:33 GeJ joined #parrot
23:45 AzureStone joined #parrot
23:46 rfw joined #parrot
23:47 TiMBuS joined #parrot
23:47 perlite joined #parrot
23:48 simcop2387_ joined #parrot
23:48 ascent joined #parrot
23:48 Psyche^ joined #parrot
23:48 smash joined #parrot
23:48 plobsing joined #parrot
23:49 aloha joined #parrot
23:49 jjore_ joined #parrot
23:49 tewk joined #parrot
23:49 bacek_at_work joined #parrot
23:49 jsut joined #parrot
23:50 atrodo joined #parrot
23:51 bubaflub joined #parrot
23:52 soh_cah_toa whiteknight: well, perhaps you could explain what each of the builtin functions does in the "builtins" section. the entire "winxed and parrot" section is empty so i could definitely benefit from info on that; in particular, using nci in winxed
23:52 whiteknight ok

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

Parrot | source cross referenced