Camelia, the Perl 6 bug

IRC log for #parrot, 2011-09-21

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:01 cotto_work dukeleto: ping
00:07 soh_cah_toa sorry, i was out
00:07 soh_cah_toa i'm starting now
00:07 soh_cah_toa again ;)
00:08 soh_cah_toa does it matter if i create a branch or just use master?
00:10 benabik soh_cah_toa: Right now, probably not.  :-D
00:10 soh_cah_toa ok
00:21 tty234 joined #parrot
00:33 soh_cah_toa darn, i thought i could include my changes in the release but i have to push them first
00:34 soh_cah_toa just improved upon the release manager guide and added my auto_release.pl script
00:34 soh_cah_toa do you think that warrants 'test' or 'fulltest'?
00:39 soh_cah_toa ugh...if it weren't for 'fulltest', i'd imagine that the entire release process would be about a third of what it is now ;)
00:40 plobsing soh_cah_toa: how powerfull is your machine? fulltest doesn't take very long on my 4 year old laptop
00:42 soh_cah_toa i have a quad core but i can't use -j5 b/c of the threads tests
00:42 soh_cah_toa it always fails w/ it
00:42 plobsing I use TEST_JOBS=n make fulltest
00:43 soh_cah_toa where n is the same number passed to -j?
00:43 soh_cah_toa or instead
00:43 plobsing n is however I feel at that time
00:44 plobsing if I'm leaving the machine alone for the duration, it is usually between 5 and 10 on my dual core (tests seem to not be cpu bound)
00:44 soh_cah_toa interesting
00:44 plobsing if I'm doing other stuff at the same time, 3 or 4
00:45 soh_cah_toa i'll try that for the next 'fulltest' ;)
00:45 plobsing try it next time you make any test
00:45 soh_cah_toa well yeah, of course
00:45 soh_cah_toa i should probably alias that
00:46 soh_cah_toa alias 4test='make fulltest TEST_JOBS=4'; alias 8test='make fulltest TEST_JOBS=8'
00:46 plobsing just set the env variable in your ~/$(SHELL_OF_CHOICE)rc
00:47 soh_cah_toa or that
00:49 plobsing it is documented in 'perldoc t/harness', but that's not terribly prominent.
00:51 jkitazawa joined #parrot
00:53 soh_cah_toa why don't we allow bareword filehandles?
00:56 soh_cah_toa i'm also getting a 'Two-argument "open" used' warning for the statement: open $REV_PARSE, 'git rev-parse master |' or stop();
00:56 soh_cah_toa that makes no sense
01:03 cotto ~~
01:08 soh_cah_toa plobsing++ # for the advice on $TEST_JOBS
01:09 soh_cah_toa wow :O
01:11 jsut joined #parrot
01:11 woosley joined #parrot
01:17 dalek parrot: 8ee72e6 | soh_cah_toa++ | tools/release/auto_release.pl:
01:17 dalek parrot: New tools/release/auto_release.pl script that automates most of release process. Still needs MANIFEST update.
01:17 dalek parrot: review: https://github.com/parrot/parrot/commit/8ee72e6953
01:17 dalek parrot: 78f0dab | soh_cah_toa++ | MANIFEST:
01:17 dalek parrot: Updated MANIFEST to include tools/release/auto_release.pl script.
01:17 dalek parrot: review: https://github.com/parrot/parrot/commit/78f0dab6a2
01:17 dalek parrot: b5fc8a0 | soh_cah_toa++ | docs/project/release_manager_guide.pod:
01:17 dalek parrot: Improved release manager guide.
01:17 dalek parrot: review: https://github.com/parrot/parrot/commit/b5fc8a08e9
01:17 dalek parrot: 157a27e | soh_cah_toa++ | ChangeLog:
01:17 dalek parrot: Added news about auto_release.pl and improved release manager guide to ChangeLog.
01:17 dalek parrot: review: https://github.com/parrot/parrot/commit/157a27e8e8
01:17 soh_cah_toa finally, thought i could kill two birds w/ one stone :/ no luck
01:18 benabik ?
01:18 soh_cah_toa i thought i could include that stuff w/ the release but it needed to be pushed first
01:18 soh_cah_toa and so i had to run 'fulltest', setting me back
01:23 cotto soh_cah_toa, alias mj='TEST_JOBS=9 make -j9'
01:23 cotto I use that more often than make
01:24 soh_cah_toa oh, nice :)
01:24 cotto It's fairly easy to mess with bash's tab completion so that it recognizes "mj" as a variant of "make"
01:25 cotto on ubuntu, at least
01:25 dalek parrot: cf64d47 | soh_cah_toa++ | tools/release/auto_release.pl:
01:25 dalek parrot: Forgot to include changes to auto_release.pl that fixed failures in t/codingstd/perlcritic.t.
01:25 dalek parrot: review: https://github.com/parrot/parrot/commit/cf64d47aab
01:42 soh_cah_toa wow. i think after the release, i'm gonna look into fixing some of these todo/skip tests. there are *way* too many of them
01:42 soh_cah_toa particularly, t/dynpmc/select.t
01:55 dalek parrot: 61dd38a | soh_cah_toa++ | / (10 files):
01:55 dalek parrot: Released 3.8.0 "Magrathea".
01:55 dalek parrot: review: https://github.com/parrot/parrot/commit/61dd38a6b4
02:03 cotto <3
02:04 soh_cah_toa almost there, just checking the tarballs
02:05 cotto soh_cah_toa, http://www.youtube.com/watch?v=X-nKHwJ3TyE
02:05 cotto you might appreciate that
02:06 soh_cah_toa ha! a fellow douglas adams fan i see ;)
02:07 cotto What gives you that idea?
02:10 cotto bugdays are a historical leftover.  We need to drop the pretense.
02:11 soh_cah_toa :)
02:11 soh_cah_toa yeah, i know
02:12 cotto soh_cah_toa, free karma if you take it out of the release manager guide.
02:12 soh_cah_toa pre-3.8.0 or post-3.8.0?
02:13 soh_cah_toa i could go back and make a new commit, but i'd have to start over
02:13 cotto soh_cah_toa, be lazy ;)
02:13 soh_cah_toa phew
02:13 cotto post
02:14 cotto I'm not a sadist.
02:15 dalek rakudo/nom: f1c805a | Coke++ | t/spectest.data:
02:15 dalek rakudo/nom: run fudged tests.
02:15 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f1c805aa50
02:17 soh_cah_toa doing a 'fulltest' on both the gzip and bzip2 tarballs is probably redundant, right? verifying one is probably enough
02:18 plobsing why do we make 2 different tarballs?
02:18 soh_cah_toa isn't that just standard practice?
02:18 soh_cah_toa stupid
02:18 soh_cah_toa but the usual ;)
02:18 benabik Because not everyone can use .bz2?
02:19 soh_cah_toa i mean, if you have either one, chances are you have the other
02:19 plobsing is there really a system out there that doesn't have bzip2 installed, and for some reason cannot build it from source, but for some reason *can* build parrot from source?
02:20 soh_cah_toa there probably once was years ago and now the practice has just stuck
02:21 plobsing bzip2 is floss, and relatively small
02:26 soh_cah_toa when i run the crow.pir script, i get:
02:26 soh_cah_toa Unable to open filehandle from path 'NEWS'
02:26 soh_cah_toa current instr.: 'parrot;Crow;get_news' pc 97 (runtime/parrot/library/Crow.pir:68)
02:26 soh_cah_toa called from Sub 'main' pc 115 (tools/release/crow.pir:63)
02:26 soh_cah_toa hey wait a second, what the heck happened to NEWS?
02:26 benabik Got rolled into ChangeLog.
02:27 benabik I thought someone updated crow when we nuked it...
02:27 soh_cah_toa damn, if i'd known i would've fixed it in 3.8.0
02:27 soh_cah_toa oh well
02:30 cotto soh_cah_toa, yeah.  the gzipped tarball is redundant.  I won't feel bad if it goes away, though it's probably best to bring it up at a #ps.
02:30 soh_cah_toa sure
02:32 soh_cah_toa can somebody run crow.pir on there system, i changed every instance of the word NEWS to ChangeLog and it still doesn't work
02:36 plobsing the backtrace mentions runtime/parrot/library/Crow.pir
02:36 plobsing did you fixup that too?
02:36 soh_cah_toa yeah
02:37 plobsing why did we separate crow out into a library?
02:37 plobsing wtf
02:37 soh_cah_toa i know, it's dumb
02:37 benabik … so it can be reused?
02:37 soh_cah_toa this probably could be rewritten as a p5 script
02:38 cotto benabik, I lol'd
02:38 plobsing reuse the suffering caused?
02:38 benabik cotto: :-D
02:38 soh_cah_toa plobsing++
02:38 soh_cah_toa ha!
02:39 soh_cah_toa oh wait, i don't think i rebuilt after editing. whoops... :(
02:40 soh_cah_toa and it still doesn't make a difference
02:40 soh_cah_toa what a pain
02:40 soh_cah_toa error: can't find news on version 3.8.0 in 'ChangeLog'
02:41 cotto soh_cah_toa, you can copy/paste previous release announcements.
02:41 plobsing soh_cah_toa: you need to use "Release" not "release"
02:41 plobsing it is case sensitive
02:41 soh_cah_toa for what? what are we talking about?
02:42 plobsing in the changelog
02:42 soh_cah_toa oh i didn't do that
02:42 plobsing you have the text "2011-09-20    release 3.8.0"
02:42 plobsing who did?
02:42 soh_cah_toa dunno
02:42 plobsing tell them to fix it
02:42 plobsing ;)
02:42 soh_cah_toa 3.7.0 is like that too
02:42 soh_cah_toa git-blame
02:43 soh_cah_toa duke!
02:43 soh_cah_toa agh, does this mean i need to re-package the release?
02:43 cotto He's always up to no good.
02:43 plobsing soh_cah_toa: you could fix crow.pir
02:44 plobsing that's what I'm trying to do atm
02:44 soh_cah_toa yeah, i know but already pushed the bad version
02:44 cotto soh_cah_toa, I worked with a contractor who used to way "We aren't building a piano.".  The same principle applies here.
02:45 soh_cah_toa alright, i just don't want to get "release manager"-banned ;)
02:45 plobsing cotto: is there some insight about piano building that I am missing?
02:45 bubaflub joined #parrot
02:46 cotto soh_cah_toa, lulz.  There's no such thing.
02:46 cotto plobsing, unlimited attention to detail isn't needed in all areas.
02:47 dalek parrot: fb6470c | plobsing++ | runtime/parrot/library/Crow.pir:
02:47 dalek parrot: release announcements use all lower case, crow.pir should expect the same
02:47 dalek parrot: review: https://github.com/parrot/parrot/commit/fb6470c3c0
02:47 plobsing soh_cah_toa: pushed fix for crow.pir problem
02:48 plobsing soh_cah_toa: being "release manager"-baned might be considered a blessing by some
02:48 soh_cah_toa plobsing: shouldn't you change the case in ChangeLog instead?
02:49 plobsing soh_cah_toa: nah. the changelog is that way to begin with
02:49 plobsing crow should parse the changelog
02:50 plobsing and this works
02:50 soh_cah_toa alright
02:51 benabik soh_cah_toa: Yeah, if we wanted to punish you, we'd say you have to do it again next month until you get it right.
02:51 soh_cah_toa ha! i don't mind, it's kinda fun
02:51 soh_cah_toa plobsing: i haven't pulled those changes yet but don't you need to s/NEWS/ChangeLog/g in tools/release/crow.pir as well?
02:53 soh_cah_toa b/c i get a big blank spot @NEWS@ in the output
02:53 benabik What's new in the release?  @NEWS@!
02:54 soh_cah_toa :)
02:56 plobsing where is nopaste?
02:56 plobsing anyways, this is what *I* get from crow: http://nopaste.snit.ch/81862
02:57 soh_cah_toa alright, maybe it's just b/c i haven't pulled fb6470c in yet
02:57 soh_cah_toa b/c what i did was edit ChangeLog instead
02:57 Coke crow is ONLY used by the build manager. you don't have to redo the release if you fix crow to generate the release announcement post release.
02:58 Coke And my apologies, I thought that issue had been resolved.
02:58 soh_cah_toa :)
03:02 soh_cah_toa in the blog on parrot.org, how do i get a <code> block to preserve whitespace/tabs? i'm trying to format the ChangeLog
03:02 cotto probably <pre>
03:02 soh_cah_toa right, that's the one
03:16 japhb Hello all
03:16 soh_cah_toa japhb: hey
03:17 japhb I just wanted to extend a personal thank you to anyone and everyone who kept the OpenGL subsystem working all this time since I've been away from Parrot
03:17 benabik Something still works!  \o/
03:17 benabik (Has someone been working on that?)
03:18 japhb benabik, yeah, and it's been long enough that I'm really amazed it hadn't rotted.
03:18 Coke so, why did you change the style from Release to release?
03:18 benabik I think they didn't change the file, just changed what the script was looking for.
03:19 Coke Looks like 3.7.0 was an anomoly.
03:19 plobsing japhb: most credit goes to NotFound and his fly.pir
03:19 Coke benabik: every other release in that file is Release.
03:19 Coke better to fix Changelog.
03:20 soh_cah_toa that's what i said
03:20 japhb plobsing, fly.pir?
03:20 benabik Coke: Oh.  I haven't been paying enough attention.  I just felt like being peanut gallery tonight.
03:21 plobsing japhb: errr... fly.winxed
03:22 soh_cah_toa agh, how do i get text in a blog post to respect newlines?
03:22 Coke it's html, not text.
03:22 soh_cah_toa yeah
03:22 Coke IIRC.
03:22 Coke <p>stuff</p> <--- newline there.
03:23 plobsing japhb: https://github.com/NotFound/winxed/blob/0de3efd454​7ad133346f518c18d1505228bac369/examples/fly.winxed
03:23 benabik Or newline<br> here
03:23 soh_cah_toa for some reason, i don't find that <br /> always works
03:23 soh_cah_toa i hate it, our paragraphs are too close together
03:24 japhb plobsing, ah ... will try to get winxed running so I can take a look.  :-)
03:24 davidfetter joined #parrot
03:24 benabik japhb: It's part of Parrot these days.
03:25 japhb benabik, ah, but it's examples are not?
03:26 plobsing no, it is a minimalist import.
03:26 benabik japhb: I guess not?
03:27 japhb well shoot, that seems suboptimal.  Is winxed really part of parrot now, or is it being snapshotted every so often, like NQP-rx used to be?
03:27 soh_cah_toa japhb: https://github.com/NotFound/​winxed/tree/master/examples
03:28 benabik It's snapshotted at stable points so NotFound can break things whenever he wants.  :-D
03:28 japhb benabik, ah, OK
03:28 soh_cah_toa where is this "Publishing options" i'm supposed to look at on the website?
03:29 soh_cah_toa it also mentions a link "Administor" which i do not see either
03:30 soh_cah_toa i'm guessing that stuff is no longer relevant?
03:31 cotto soh_cah_toa, at the bottom of the page where you edit the post
03:31 soh_cah_toa it's not there
03:31 cotto lolwut
03:31 soh_cah_toa just "URL path settings" and "Input format""
03:32 cotto are you creating a story?
03:32 soh_cah_toa yes
03:32 benabik Incorrect user role?
03:32 soh_cah_toa most likely
03:32 cotto "Publishing options" doesn't appear anywhere on the page?
03:32 soh_cah_toa nope
03:33 cotto how special
03:33 japhb W00t, all winxed OpenGL examples work.
03:33 cotto looks like I don't have an admin bit for the drupal site
03:33 benabik I don't see it either.
03:33 soh_cah_toa no matter what type of content i try to create, all i have is those two: "Input format" and "URL path settings"
03:34 soh_cah_toa uh oh
03:34 cotto soh_cah_toa, go ahead and create it and I can clean it up
03:34 soh_cah_toa sure
03:35 cotto I have access to the vm.  I should be able to finagle myself an admin bit.
03:37 dalek website: soh_cah_toa++ | Parrot 3.8.0 "Magrathea" Released
03:37 dalek website: http://www.parrot.org/node/361
03:38 benabik cotto: Hacking your own web site for the win.
03:38 cotto benabik, it's hardly hacking if I have an account on the machine.
03:38 cotto soh_cah_toa, no quote?
03:39 soh_cah_toa at the bottom
03:39 cotto and I thought I'd picked a long one
03:39 soh_cah_toa yeah, that's why it's at the bottom :)
03:40 soh_cah_toa what's the domain for the perl6-language and perl6-announce mailing lists? perl.org?
03:40 cotto yup
03:40 cotto though you should use perl6-compiler
03:41 soh_cah_toa instead or in addition?
03:41 cotto istr -language being dead
03:42 benabik p6l at least seems to not have useful people on it, judging from convos on #perl6, although that seems to imply that it's not dead.
03:42 soh_cah_toa so...message? no message?
03:43 cotto message
03:43 plobsing I can clobber master yet?
03:43 benabik More active than some lists..  A message every few days...
03:44 benabik Although a lot of it just seems to be github.
03:44 soh_cah_toa yes, go forth and break master
03:45 soh_cah_toa uh, i don't have a use perl or perl monks account (actually tried to few weeks ago but never got a reply)
03:45 soh_cah_toa so how do i submit the story?
03:46 cotto istr use perl being read-only
03:46 soh_cah_toa oh yeah
03:46 cotto you'll probably have to hack the site to submit something.
03:47 cotto we should factor that into the release manager guide
03:47 soh_cah_toa yeah, i'll fix it in a minute
03:47 soh_cah_toa what about perl monks though?
03:49 cotto my hero: http://stackoverflow.com/questions/55863​58/any-decent-php-parser-written-in-php
03:49 cotto one of many steps on the way to world domination
03:49 soh_cah_toa lol
03:51 soh_cah_toa oh whoops, i never subscribed to parrot-users so it's being held
03:52 cotto I don't see any notification
03:53 benabik Our release manager isn't one of our users?  Seems suspicious.
03:53 soh_cah_toa lol
03:53 soh_cah_toa i never bothered w/ it
03:54 nopaste "soh_cah_toa" at 192.168.1.3 pasted "parrot-users email" (15 lines) at http://nopaste.snit.ch/81863
03:55 Coke \o/, muddle can now accept POST uploads.
03:55 benabik muddle?
03:55 Topic for #parrot is now Parrot 3.8.0 "Magrathea" | http://parrot.org | Log: http://irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
03:55 benabik Oh, new smolder?
03:55 benabik soh_cah_toa: Wow, I'm tired…  I nearly visited the cancel e-mail to see the post.
03:55 benabik I don't think that would have worked well.
03:56 soh_cah_toa ha!
03:56 Coke parrot releases are off topic for p6l, IMO.
03:57 soh_cah_toa what does the guide mean when it says that i should update the wiki frontpage? there's nothing here that's release-specific
04:01 dalek muddle: d2b6432 | Coke++ | / (2 files):
04:01 dalek muddle: Extract the guts of this script into a module.
04:01 dalek muddle: review: https://github.com/coke/muddle/commit/d2b6432733
04:01 dalek muddle: 17c501d | Coke++ | / (4 files):
04:01 dalek muddle: Add barebones upload facility via web
04:01 dalek muddle: review: https://github.com/coke/muddle/commit/17c501dffd
04:02 soh_cah_toa cotto: just realized i forgot to set the url to "news/[year]/Parrot-[release number]". should i just use the http://parrot.org/node/361 link on wikipedia then?
04:03 Coke Please fix the link.
04:03 soh_cah_toa i can't
04:03 soh_cah_toa i don't have the permissions
04:03 soh_cah_toa i had to skip a few things b/c of that
04:03 Coke what's your parrot.org login?
04:03 soh_cah_toa soh_cah_toa
04:04 Coke did you read the step that you were supposed to do that bit weeks ago? ;)
04:04 soh_cah_toa for ssh access? that's done
04:05 cotto soh_cah_toa, you can set it after the fact
04:06 Coke giving you access now.
04:06 soh_cah_toa ok
04:06 Coke and no, the part where you need to get admin access on the website.
04:07 soh_cah_toa i'm not sure it mentions anything like that
04:08 Coke bleargh. I'm fairly certain it used to.
04:08 Coke done.
04:08 Coke you have all privs-but-director.
04:08 soh_cah_toa thanks
04:10 soh_cah_toa there we go, now i see the links
04:10 Coke is there a way to git-grep old revisions of a file?
04:10 benabik git log -Stext -- file
04:10 benabik (Kinda)
04:10 benabik Might want to add -p to that to see the patch.
04:13 soh_cah_toa i still don't have the "administer" -> "site building" -> "url redirecting" links though
04:13 soh_cah_toa just "url redirecting" but it's not as described in the guide
04:13 Coke d5ecb29e8ddcb0aa3ae27866f3712ba73067af7b
04:13 Coke (that's got a note about admin privs.)
04:14 soh_cah_toa it's not there anymore though
04:15 soh_cah_toa it's way old
04:19 Coke soh_cah_toa: "docs/project/release_manager_guide.pod" line 375 of 535 --70%-- col 51
04:19 Coke I am now removing your admin rights. ;P
04:19 Coke (but leaving editor)
04:19 nbrown joined #parrot
04:20 Coke drupals user management sucks, man.
04:20 Coke I have to page through 30 pages of users to find you before I can edit you.
04:21 Coke done.
04:23 * Coke should preface this with soh_cah_toa++ # Thanks for doing the release!
04:23 * Coke is bad about that sort of thing. Sorry.
04:27 cotto Coke, it's not that bad.  We're just using it wrong.
04:27 benabik "using it wrong" seems to be the theme of the day...
04:28 cotto srsly
04:30 dalek muddle: ebcf1da | coke++ | Makefile:
04:30 dalek muddle: change default port
04:30 dalek muddle:
04:30 dalek muddle: also allow make target to succeed first time.
04:30 dalek muddle: review: https://github.com/coke/muddle/commit/ebcf1daa0b
04:30 dalek muddle: 9957803 | coke++ | templates/projects/upload.html.ep:
04:30 dalek muddle: fix whitespace bug
04:30 dalek muddle: review: https://github.com/coke/muddle/commit/99578037b0
04:30 Coke http://feather.perl6.nl:2653/parrot/ has bee updated, feel free to dump some test runs in there so I can start working on the visualizations.
04:30 Coke cotto: how do I edit user 'foo' ? there's no "jump to this user" option.
04:32 TimToady you mean like sudo su foo?
04:32 Coke in drupal.
04:32 TimToady btw, #perl6 seems to be very badly netsplit at the moment
04:32 cotto oh my
04:33 soh_cah_toa release/current and release/developer should point to the tarball or the ftp page to download the tarball? the guide says the tarball but the 3.7.0 had it point to the ftp page
04:33 cotto Coke, if you know the user's uid, you can go to user/999
04:33 Coke cotto: that is a crappy interface. ;)
04:33 cotto Coke, I don't know where it should be by default
04:33 Coke Search: [ Enter Username here...]
04:34 dalek parrot: 3014457 | plobsing++ | / (63 files):
04:34 dalek parrot: Merge branch 'master' into kill-useless-defaults
04:35 dalek parrot: review: https://github.com/parrot/parrot/commit/30144578ac
04:35 dalek parrot: 0bf02d9 | plobsing++ | / (12 files):
04:35 dalek parrot: Merge branch 'kill-useless-defaults'
04:35 dalek parrot: review: https://github.com/parrot/parrot/commit/0bf02d9e07
04:35 dalek Heuristic branch merge: pushed 27 commits to parrot/kill-useless-defaults by plobsing
04:35 benabik plobsing++ # killing useless things.
04:35 soh_cah_toa plobsing++ indeed
04:36 cotto Coke, care to give me an admin bit?
04:36 plobsing benabik: wait a few days, I'm sure the --'s will come rolling in
04:36 cotto I spend all day staring at Drupal.
04:38 soh_cah_toa i'm not quite sure how to do step 6 of updating the website
04:39 soh_cah_toa soh_cah_toa@parrot.org has no directories, just a few hidden files
04:39 cotto soh_cah_toa, the ftp site, not the vm
04:39 soh_cah_toa then the guide is wrong
04:40 benabik The Guide is never wrong.  The galaxy sometimes needs to be updated.
04:40 soh_cah_toa ha! clever
04:40 benabik soh_cah_toa: you started it.
04:41 soh_cah_toa so what do i login to? parrot@ftp-osl.osuosl.org?
04:41 soh_cah_toa there's no docs dir there either
04:41 cotto yup
04:42 cotto actually, let me make sure I'm not lying to you
04:42 soh_cah_toa :)
04:42 cotto ok.  parrotvm is the right place
04:42 soh_cah_toa but there's nothing there
04:42 cotto do you have write access to /var/www/html/docs.parrot.org?
04:43 soh_cah_toa nope
04:43 soh_cah_toa /var/www/html, that is
04:44 Tene japhb: looks like I got on the wrong side of you on the freenode netsplit.  I've been meaning to try to get rakudo's HLL interop somewhere reasonable again, including loading parrot libraries, but I haven't had the time or energy lately to help get tha work started again.
04:44 cotto should be /var/www/docs.parrot.org/htdocs/
04:45 soh_cah_toa ah, there it is
04:45 soh_cah_toa gotta fix the guide some more
04:45 japhb Tene, ah,thanks for the info.  I'd love to see it back.  Being able to use my Parrot OpenGL library from Rakudo was half the reason I wrote it in the first place.  :-)
04:45 Tene japhb: similarly, language interop was my primary interest in Parrot.
04:46 soh_cah_toa well, it's right it just doesn't say what dir webroot is
04:46 Tene japhb: anyway, uhh... plausibly I might have time tomorrow, but i've said that before...
04:46 japhb Tene, nodnod.  I care most about all Parrot languages being able to use common libraries -- for me, being able to directly call between the languages is gravy.  But very tasty gravy.  ;-)
04:47 Tene exactly
04:47 Tene +1
04:47 soh_cah_toa agh!
04:47 soh_cah_toa i can't write to webroot
04:47 Coke cotto: you're on the board?
04:48 cotto Coke, nope
04:48 soh_cah_toa /var/www/docs.parrot.org/htdocs/parrot
04:48 Coke cotto: updated.
04:48 Coke anyone else as long as I'm in here?
04:49 Coke hey, particle logged into the website 2 weeks ago.
04:49 soh_cah_toa is anyone able to give me write access to /var/www/docs.parrot.org/htdocs/parrot?
04:49 benabik soh_cah_toa: I'm sure someone cann.
04:49 cotto Coke, can you add users to the parrot group on parrotvm?
04:49 Coke benabik: you're an editor now, claim a release or something. ;)
04:49 benabik What?  Who?
04:49 * benabik hides.
04:50 Coke cotto: I have no access to parrotvm at the moment due to recent ssh wonkery.
04:50 cotto ok
04:50 benabik Coke: School is threatening to eat all my free time.  :-/
04:50 benabik For that matter, I should really be asleep now.
04:50 soh_cah_toa yeah, me too
04:51 benabik soh_cah_toa: 7:30 AM to 6 PM today.
04:51 soh_cah_toa !!!
04:51 soh_cah_toa wow
04:51 Coke (sleep) me 3.
04:52 benabik Tomorrow's only 9:30-5, but I can't get a ride anywhere until 6…  and that'll be to my MMA class.  Won't get home until probably 9:30.
04:52 Coke benabik: *pfwa* I have 3 kids (elementary, middle school, high school) and a wife and a job with roughly those hours.
04:52 Coke I do, however, have a new car. So I win there. ;)
04:52 benabik Things like this is most of my week.  Fortunately I have three days "off", by which I mean to do homework in.
04:52 Coke benabik: whereabouts are you?
04:52 benabik Coke: That sounds fun.
04:52 benabik Coke: We have a pretty new car, but only one of them.
04:52 benabik Coke: Rochester, NY
04:53 benabik soh_cah_toa: Many of those hours at school are occupied by work.
04:53 soh_cah_toa ah
04:53 benabik They gave me root access, the fools.  Bwa-hahahahaha…  Time to install parrot on all the workstations.
04:53 benabik >.>  <.<
04:54 plobsing joined #parrot
04:54 Coke benabik: a /smidge/ out of my way to proffer a ride.
04:55 Coke (only about 4 hours west of here.)
04:55 soh_cah_toa cotto: can you give me write access?
04:55 soh_cah_toa or add me to 'parrot' group, whatever
04:55 benabik Coke: We're practically neighbors.  (If you ask jnthn.  ;-) )
04:56 cotto soh_cah_toa, I don't think I can.
04:56 soh_cah_toa stupid guide...
04:57 soh_cah_toa wait a sec, you're in the parrot group
04:57 cotto yeah
04:57 soh_cah_toa so doesn't that mean you can?
04:57 soh_cah_toa add
04:57 soh_cah_toa me ;)
04:59 benabik "Code Hero is a game that teaches you how to make games and save the world with a code gun that shoots Javascript"
04:59 benabik I've wanted to shoot Javascript.
05:05 cotto soh_cah_toa, I seem to not have the permissions needed to do that.
05:05 soh_cah_toa ugh
05:05 benabik group membership != group administration
05:05 soh_cah_toa only root can add people to a group?
05:06 benabik There's probably a way to make it more liberal, but generally yes.
05:06 Tene soh_cah_toa: you can also set a group administrator.
05:06 benabik Barring exciting things like LDAP, it's just editing /etc/groups
05:06 soh_cah_toa i've never actually had to manage groups on a machine before
05:06 soh_cah_toa other than me:me
05:06 benabik Sorry /etc/group
05:07 Tene soh_cah_toa: gpasswd -A tene somegroup
05:07 Tene then tene can add/remove people from somegroup
05:07 benabik Tene: gpasswd?
05:07 soh_cah_toa interesting
05:07 Tene benabik: yes
05:07 Tene benabik: gpasswd is for administering /etc/group and /etc/gshadow
05:08 Tene see also: group passwords
05:08 benabik Ah.
05:08 benabik Tene: Well, it's a suid binary, so it's still root editing it.  ;-)
05:09 soh_cah_toa msg whiteknight it seems that i don't have write access to /var/www/docs.parrot.org/htdocs/parrot. apparantly, i'm not in the 'parrot' group. can you see if you're able to add me? it's the last step of the release and i can't update the docs :)
05:09 aloha OK. I'll deliver the message.
05:09 Tene benabik: sure
05:09 benabik IIRC, "updating the docs" was one of the more problematic things for kid51 as wll.
05:09 soh_cah_toa ah
05:09 cotto The parrot group doesn't show up in /etc/group
05:10 benabik The beginning of the release guide should probably start with "get the following permissions and do ___ to make sure you have them"
05:10 soh_cah_toa whiteknight has his one group too. weird
05:10 soh_cah_toa benabik: that's exactly what i'm adding tomorrow
05:10 benabik soh_cah_toa++ # karma on credit
05:11 soh_cah_toa stupid guide just costed my calculus homework
05:11 benabik Bah calc.
05:11 benabik ,/` x dx = I don't care.
05:11 cotto soh_cah_toa, I can take care of updating the docs
05:11 cotto go be a responsible student
05:12 soh_cah_toa btw, if anybody knows how to solve for x when e^x = 2, do tell ;)
05:12 benabik Don't mind me…  I just dislike that people made me learn calc before getting to the interesting stuff like discrete and algebra.
05:12 soh_cah_toa cotto: great thanks :)
05:13 benabik Only time I use calc is when I forget area or volume formulas and want a challenge.
05:13 benabik soh_cah_toa: log_e e^x = x
05:13 soh_cah_toa log_e = ln though right?
05:13 benabik soh_cah_toa: Yes.
05:14 soh_cah_toa ah, that's right
05:14 benabik and log_2 = lg (which I tend to use more than ln, but that's CS for you)
05:14 soh_cah_toa oh cool, never heard of that before: lg
05:14 benabik Or at least that's a notation I've seen more than once.  Not sure how standard it is.
05:15 benabik WolframAlpha seems to think it's an alias for log.
05:16 soh_cah_toa ah, good ol' wolfram alpha
05:16 benabik Ah.  Wikipedia to the rescue…  log_2 was often spelled ld or lg, but ISO says it's lb.
05:16 benabik My algorithm teacher said it was just spelled log.  :-D
05:17 benabik Good ol' ISO making arbitrary decisions.
05:17 soh_cah_toa :)
05:18 benabik ISO says log requires base, lb = log_2, ln = log_e, lg = log_10
05:18 benabik http://en.wikipedia.org/wiki/ISO_31-11
05:18 soh_cah_toa right
05:18 benabik Man, there's a ISO standard for everything.
05:18 benabik Although I guess that's the point.
05:19 moritz has the kill-useless-defaults branch already landed?
05:19 soh_cah_toa moritz: yessir
05:19 benabik moritz: I thought I saw dalek say so.
05:20 cotto d.p.o is updated
05:20 soh_cah_toa cotto++
05:20 soh_cah_toa great
05:21 soh_cah_toa who should i talk to about that for future releases though?
05:21 woosley joined #parrot
05:21 cotto osuosl, I guess
05:21 benabik soh_cah_toa: Just update the guide to say "ask cotto to update docs.p.o" ;-)
05:22 soh_cah_toa :)
05:22 cotto benabik, -1
05:22 benabik cotto: You're no fun.  ;-)
05:22 cotto No I'm not.
05:22 cotto It violates laziness.
05:23 soh_cah_toa this place gets really weird at night :D
05:23 benabik My boss said "lazy sysadmin" like it was a bad thing today.  Confused me greatly.
05:23 benabik soh_cah_toa: I get weird at night and feel talkative.  :-D
05:23 benabik *tonight
05:24 soh_cah_toa i know, there's like a little window of time you can go to sleep at and if you miss it...bam! wide awake, up, and doing stuff
05:26 soh_cah_toa and on that note, i'm outta here
05:27 benabik 'night
05:32 JimmyZ joined #parrot
05:56 dalek nqp/kill-useless-defaults: b7b1019 | plobsing++ | src/ (3 files):
05:56 dalek nqp/kill-useless-defaults: use new get_id pirop to generate unique IDs for PMCs
05:56 dalek nqp/kill-useless-defaults: review: https://github.com/perl6/nqp/commit/b7b10190f8
05:56 dalek nqp/kill-useless-defaults: 1b2d991 | plobsing++ | src/stage0/ (7 files):
05:56 dalek nqp/kill-useless-defaults: rebootstrap
05:56 dalek nqp/kill-useless-defaults: review: https://github.com/perl6/nqp/commit/1b2d9915d3
05:56 dalek nqp/kill-useless-defaults: c846f63 | plobsing++ | src/ (16 files):
05:56 dalek nqp/kill-useless-defaults: Merge branch 'master' of git://github.com/perl6/nqp
05:56 dalek nqp/kill-useless-defaults: review: https://github.com/perl6/nqp/commit/c846f63933
05:56 dalek nqp/kill-useless-defaults: 2ea19fd | moritz++ | tools/build/PARROT_REVISION:
05:56 dalek nqp/kill-useless-defaults: bump PARROT_REVISON
05:56 dalek nqp/kill-useless-defaults: review: https://github.com/perl6/nqp/commit/2ea19fd688
05:57 benabik Is get_id added in kill-useless-defaults?
05:58 cotto it replaces get_addr
06:00 moritz fwiw rakudo has some troubles on that branch
06:00 moritz get_bool() not implemented in class 'Bool
06:00 moritz '
06:00 moritz while building the setting
06:01 benabik boo
06:05 cotto moritz, plobsing sent an nqp pull request.  Does that help?
06:05 moritz cotto: that's what I'm using as the base
06:06 cotto then that's a problem
06:06 moritz no, wait
06:06 moritz my rakudo copy was out of date
06:06 moritz nevermind
06:06 * moritz tries again
06:07 moritz (working in too many directories at once)--
06:07 cotto yeah
06:08 moritz sorry for the noise
06:08 moritz I've also given plobsing++ a commit bit for nqp
06:08 moritz does anybody else want one?
06:10 cotto moritz, no worries.  I should test it anyway
06:13 moritz build successful
06:13 moritz now spectesting...
06:13 cotto that's a good sign
06:18 cotto moritz++
06:19 cotto I'll take a false positive over a real breakage any day.
06:40 moritz spectest clean
06:43 rfw joined #parrot
06:46 cotto moritz, that makes me very happy.  Thanks for testing.
06:54 moritz soh_cah_toa++ # parrot release
06:54 moritz soh_cah_toa++ # parrot release name
07:28 mj41 joined #parrot
08:04 lucian joined #parrot
08:22 JimmyZ joined #parrot
09:04 dalek parrot/mls/sub-profiler: 825c836 | (Michael Schroeder)++ | / (3 files):
09:04 dalek parrot/mls/sub-profiler: fix off-by-one error, free all profile memory after dumping, export just one init function
09:04 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/825c836782
09:07 dalek parrot/mls/sub-profiler: 9820278 | (Michael Schroeder)++ | CREDITS:
09:07 dalek parrot/mls/sub-profiler: add myself to CREDITS
09:08 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/9820278941
09:13 dalek parrot/mls/sub-profiler: c9f3e26 | (Michael Schroeder)++ | / (27 files):
09:13 dalek parrot/mls/sub-profiler: Merge branch 'master' into mls/sub-profiler
09:13 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/c9f3e26fb8
10:53 ambs joined #parrot
11:21 nbrown_ joined #parrot
11:35 Psyche^ joined #parrot
11:54 whiteknight joined #parrot
11:55 whiteknight good morning, #parrot
11:55 whiteknight soh_cah_toa++ on the release
12:03 lucian joined #parrot
12:10 * whiteknight is testing whiteknight/kill_threads now in a fresh update from master. Will merge soon if tests look good
12:14 dalek parrot/whiteknight/kill_threads: 8e8aa75 | Whiteknight++ | / (75 files):
12:14 dalek parrot/whiteknight/kill_threads: merge in from master. Fix small whitespace conflicts
12:14 dalek parrot/whiteknight/kill_threads: review: https://github.com/parrot/parrot/commit/8e8aa75831
12:16 lucian joined #parrot
12:35 lucian joined #parrot
12:46 redicaps joined #parrot
13:00 whiteknight NotFound: ping
13:03 dalek parrot/whiteknight/kill_threads: 991bead | Whiteknight++ | / (18 files):
13:03 dalek parrot/whiteknight/kill_threads: Merge the NotFound/fix_sleep branch, to get fixes for the Timer PMC tests
13:03 dalek parrot/whiteknight/kill_threads: review: https://github.com/parrot/parrot/commit/991beada14
13:04 whiteknight msg NotFound: I merged the NotFound/fix_sleep branch into Whiteknight/kill_threads for the test fixes. If you don't mind, I would like to merge both into master soon.
13:04 aloha OK. I'll deliver the message.
13:12 whiteknight msg mls do you have a timeline for the subprofiler branch? If it's looking good we can do a merge
13:12 aloha OK. I'll deliver the message.
13:31 dalek TT #2194 created by Andy Dougherty <doughera@…>++: Alignment errors on Solaris/SPARC since 3.7.0
13:31 dalek TT #2194: http://trac.parrot.org/parrot/ticket/2194
13:39 dalek nqp: 3faca64 | moritz++ | tools/build/PARROT_REVISION:
13:39 dalek nqp: bump PARROT_REVISION to 3.8 release
13:39 dalek nqp: review: https://github.com/perl6/nqp/commit/3faca64beb
13:43 dalek rakudo/nom: e7ed52f | moritz++ | t/02-embed/01-load.pir:
13:43 dalek rakudo/nom: TODO 02-embed test
13:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e7ed52fafd
13:43 dalek rakudo/nom: 0cbc528 | moritz++ | src/Perl6/Compiler.nqp:
13:43 dalek rakudo/nom: use .gist instead of prefix:<~> to stringify REPL autoprint output
13:43 dalek rakudo/nom:
13:43 dalek rakudo/nom: Now when the return value of an expression is a type object, you get Int()
13:43 dalek rakudo/nom: or so, and no warning
13:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0cbc528cd1
13:43 dalek rakudo/nom: 28f0696 | moritz++ | tools/build/NQP_REVISION:
13:43 dalek rakudo/nom: bump NQP revision to get some testing on released parrot
13:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/28f0696523
13:46 mls whiteknight: IMHO it's good enough for merging
14:05 whiteknight mls: if you're ready, do it
14:05 whiteknight the kill_threads branch will probably merge later today. I would rather do yours first
14:05 whiteknight I don't suspect a lot of overlap, but I'll cleanup the wreckage if there is any
14:15 redicaps left #parrot
14:29 cotto ~~
14:33 lucian joined #parrot
14:34 whiteknight good morning, cotto
14:39 benabik joined #parrot
14:40 benabik o/
14:40 cotto benabik, hio
14:40 dalek parrot: cac743e | cotto++ | docs/project/release_manager_guide.pod:
14:40 dalek parrot: add a new batch of slots for future release managers
14:40 dalek parrot: review: https://github.com/parrot/parrot/commit/cac743e9ff
14:40 cotto whiteknight, good morning
14:40 cotto benabik, what's the status of newPOST?
14:41 benabik cotto: Bytecode works for very very simple things.  But PIR generation is approx 2x slower.
14:41 cotto ow
14:41 cotto same for pbc generation?
14:41 cotto (speedwise)
14:42 benabik PBC generation is infinitely faster than before…  It not working previously.
14:42 benabik ;-)
14:42 cotto this is true
14:43 benabik I only got very very simple PAST trees to build to bytecode, so I never compared speed vs PIR.
14:47 dalek rakudo/nom: c2194d6 | moritz++ | docs/ChangeLog:
14:47 dalek rakudo/nom: copy ChangeLog from 2011.07
14:47 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c2194d6022
14:48 cotto benabik, where's the code live?
14:48 benabik nqp_pct branch, IIRC
14:53 NotFound whiteknight: pong
14:53 whiteknight NotFound : I sent you an msg. I want to merge your fix_sleep branch at the same time as the kill_threads branch
14:54 NotFound whiteknight: go ahead
14:54 whiteknight okay, awesome
14:54 cotto benabik, what future do you see for the branch?
14:55 NotFound whiteknight: please add a comment about that in the TT
14:55 whiteknight okay, will do
14:55 benabik cotto: Sadly, not much unless we can get some serious speed improvements on it.  Which probably means getting nqp-rx to generate more optimized code.
14:55 benabik cotto: I suspect, but have no proof, that much of the added time is find/store_lex
14:56 cotto benabik, jnthn mentioned that he's going to be working on an nqp optimizer in the foreseeable future.  Would that help, or is the branch stuck with nqp-rx?
14:57 NotFound aloha:
14:57 NotFound aloha hasn't sent me the msg
14:57 whiteknight that stinks
14:58 NotFound Uhhh... maybe it go to "NotFound:" instead of "NotFound"
14:58 benabik cotto: I don't think there's anything that would prevent the code from running on nqp instead of nqp-rx, other than not wanting PCT to depend on 6model before it's core parrot.
14:59 whiteknight msg NotFound test
14:59 aloha OK. I'll deliver the message.
14:59 cotto so it comes back to 6model yet again
14:59 NotFound ~~
14:59 NotFound Got it
14:59 cotto msg NotFound: test
14:59 aloha OK. I'll deliver the message.
14:59 NotFound ~~
14:59 NotFound Nope
14:59 cotto silly bot
15:00 benabik cotto: Mostly because nqp uses 6model.  PCT is used in parrot.git itself.
15:00 whiteknight maybe aloha doesn't like NotFound. I got a message from it today
15:00 benabik cotto: I'd love to work on PACT or further refactors of PCT, but I don't have a lot of spare time.
15:01 benabik cotto: A direct port of PCT as-is to Winxed might be interesting though.  nqp_pct would be a nice reference for that.
15:01 benabik PCT needs to be more hackable.  Massive PIR files are…  unfriendly.
15:02 cotto yeah
15:02 cotto That is a possibility now that winxed is in core.
15:02 benabik Exactly.
15:02 cotto well, distributed as a core component
15:03 benabik Plus Winxed is gaining multi and other features PCT uses…  It would have been literally impossible to write PAST::Compiler in Winxed 3 months ago.
15:05 benabik Okay, not impossible…  Just a lot of very painful manual typechecking.
15:05 cotto very demotivating
15:05 whiteknight I suspect we would have been able to rewrite most of it to use polymorphism and inheritance instead of multis
15:05 whiteknight but, that's not trivial effort either
15:05 benabik I don't think moving the PAST/POST generation into the nodes is a good idea, if that's what you mean.
15:06 dalek rakudo/nom: 313b687 | moritz++ | docs/ChangeLog:
15:06 dalek rakudo/nom: add some few ChangeLog entries
15:06 dalek rakudo/nom:
15:06 dalek rakudo/nom: This just scratches the surface of all the nom changes; feel free to
15:06 dalek rakudo/nom: add many more
15:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/313b687d98
15:06 whiteknight benabik: I'm not saying it was a great idea, I'm saying it's an alternative to multis and type checking
15:07 TonyC joined #parrot
15:08 benabik whiteknight: NotFound: Actually, did type-based multi get into Winxed?  I remember you two talking syntax, but don't remember if it went anywhere.
15:08 NotFound Do you think it will be possible right now to write a PCT based compiler in winxed?
15:09 PerlJam Is the only advantage of porting PCT to Winxed that hacking it becomes nicer?
15:09 NotFound benabik: right now it only uses int, string, float and var as types.
15:09 benabik NotFound: Yes.  Minor amounts of NQP wrapper might be needed if Winxed can't inherit nicely from something using P6metaclass
15:10 benabik PerlJam: Yes.
15:10 benabik PerlJam: But making it nicer to hack on means it can be improved faster.
15:10 cotto One very nice feature is that the process could be done iteratively.
15:10 NotFound Also, I think winxed code is less error prone than hand written pir.
15:11 NotFound But that may count as "nicer"
15:11 nopaste joined #parrot
15:13 benabik Although if I had spare time to hack on Parrot right now, I'd probably work on improving lexicals…  Make it possible to create new nested scopes inside subs, create some kind of reference PMC that lets you get/set lexicals with only having to do the lookup once, etc.
15:13 PerlJam If there are opportunities for optimization or improved code generation simply because winxed is nicer to deal with than PIR that would be very nice.
15:14 benabik PerlJam: I think PCT codegen could be vastly improved.
15:14 benabik Parrot uses registers…  we should be able to take advantage of the piles of compiler knowledge out there for register machines.
15:15 benabik But PCT is very tied to using PIR right now, which is somewhat sad.
15:16 whiteknight extremely sad, considering PIR is what we want to remove and destroy
15:16 dukeleto ~~
15:16 cotto good morning, dukeleto
15:17 dukeleto cotto: howdy
15:17 benabik IMNSHO, we need a PIR-like.  Or PASM-like.  Something that shows a stream of literal opcodes.
15:19 NotFound benabik: shows?
15:19 benabik NotFound: Yes, shows.  Primary use is seeing what a compiler outputs or decompiling existing code.
15:20 benabik People could write it, but it's not really human friendly.
15:22 NotFound You mean a new PASM, able to be used as compiler target instead of pir?
15:22 benabik Yes.
15:22 NotFound And a disassembly complete enough to be able to be recompiled?
15:23 NotFound Will be good, yes.
15:23 benabik PIR stands for Parrot Intermediate Representation.  I'd be fine with something like LLVM IR.
15:23 JimmyZ M0?
15:23 benabik I view M0 as 100% orthogonal to HLL development.
15:24 benabik M0 is something to implement Parrot in, to get us inside the managed code as soon as possible.
15:25 benabik NotFound: Disassembly complete enough to be compiled in most cases…  Since PBC allows arbitrary objects to be stored, 100% round-trip is unlikely to be feasible.
15:25 NotFound If we allow high level compiler to directly emit M0, security and validity checks will become extremely hard or impossible.
15:28 * cotto heads to $work
15:30 dukeleto M0 is not for HLLs
15:30 dukeleto HLLs should be completely ignorant of M0
15:30 dukeleto because it is an internal implementation detail
15:31 benabik dukeleto++
15:31 dukeleto only parrot core internals should know about m0, and then, as little as possible
15:31 benabik Heh.
15:32 benabik Aren't opcodes supposed to move to M0?  Wouldn't that mean HLLs who want to poke into parrot guts will still see it?
15:33 whiteknight benabik: depends on the poking
15:33 whiteknight a good interface should obscure the internal implementation details
15:33 benabik whiteknight: I'm thinking Rakudo levels.  :-D
15:33 JimmyZ i.e:   rakudo ops
15:33 benabik (of poking)
15:52 benabik afk: class
15:59 cotto_work ~~
16:04 ligne hello #parrot
16:05 dukeleto ligne: howdy
16:06 * ligne brings bad news of g++ breakage.
16:06 nopaste "ligne" at 192.168.1.3 pasted "duplicate const declaration" (11 lines) at http://nopaste.snit.ch/81886
16:07 ligne but also an (incomplete) patch
16:10 whiteknight ligne++
16:10 whiteknight g--
16:12 ligne hehe
16:21 dmalcolm joined #parrot
16:32 Coke aw, no one uploaded any parrot test runs for me!
16:33 Coke http://feather.perl6.nl:2653/parrot
16:34 cotto_work $ make upload_test_results_for_coke
16:34 cotto_work make: *** No rule to make target `upload_test_results_for_coke'.  Stop.
16:35 Coke try "make smolder_coretest" and then manually upload it. :P
16:35 cotto_work might be a bit
16:35 Coke though honestly, it should be pretty easy to take over the submit-to-smolder script and give it a new URL.
16:36 Coke no rush
16:39 alester joined #parrot
17:07 fperrad joined #parrot
17:33 dalek rakudo/nom: a979ded | jnthn++ | docs/ChangeLog:
17:33 dalek rakudo/nom: Add some ChangeLog entries.
17:33 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a979ded1e6
17:34 whiteknight cotto_work: do you have any further reservations about merging the subprofiler branch? I would like to get that in to master ASAP for testing and usage
17:34 cotto_work whiteknight: is mls done with it?
17:34 whiteknight He said this morning that he thought it was mergable
17:35 cotto_work lemme give a quick review
17:35 whiteknight if it's not perfect, in master it will start to get real users and more eyes on it
17:35 whiteknight it will get fixed up pretty quick
17:35 cotto_work It already has real users. ;)
17:35 cotto_work or really good imaginary ones
17:45 cotto_work whiteknight: let's merge
17:45 whiteknight +1
17:46 whiteknight I've got the kill_threads branch setup locally for a merge. It's just waiting for a push. I'll do that first since it's here then start testing the profiler
17:47 cotto_work whiteknight: so you'll take care of merging?
17:47 whiteknight yeah, I can do that
17:47 cotto_work ok
17:47 whiteknight any objections to kill_threads merge?
17:47 cotto_work less talking, more merging
17:48 cotto_work master isn't going to break itself
17:48 dalek parrot: f4017d5 | Whiteknight++ | docs/project/release_manager_guide.pod:
17:48 dalek parrot: Merge branch 'master' of github.com:parrot/parrot
17:48 dalek parrot: review: https://github.com/parrot/parrot/commit/f4017d5bb9
17:48 cotto_work whiteknight++
17:49 whiteknight testing sub-profiler now
17:49 dukeleto whiteknight: MURDER DEATH KILL THREADS
17:49 whiteknight if there are no obvious explosions, I'll push that too
17:50 dukeleto Today is the day for nothing to work.
17:50 whiteknight I'll grab the campaign, somebody else grab some red plastic cups,
17:50 whiteknight champagne* Stupid spell-correction
17:51 mj41 joined #parrot
17:57 moritz hi, I have a question about ByteBuffer
17:58 NotFound 42
17:58 moritz Perl 6 has a Str.encode($encoding) function, which I try to implement as a String -> ByteBuffer transition
17:58 moritz what's the best way to do that?
17:59 moritz do I have to recode the string, and then use the set_string vtable in ByteBuffer?
17:59 moritz or can I somehow directly encode a String to a ByteBuffer in a given encoding?
18:00 whiteknight moritz: that's a very good question
18:00 NotFound moritz: ByteBuffer is low level. If you have a binary string whose content you know is appropiately encode, you can just use the set_string
18:01 moritz NotFound: well, I don't. Unless I find a way to do that
18:01 moritz which would be recode, if I understand recode correctly
18:01 NotFound moritz: if what you want is to transcode, I don't see the need for a ByteBuffer
18:02 moritz NotFound: I don't want to transcode, it's just a means to an end
18:02 whiteknight what is the end?
18:02 moritz go from String (with an encoding that could be anthing) to a ByteBuffer, using a specified encoding
18:03 moritz for example I want to encode 'foo' in UTF-16LE, and store the result in a ByteBuffer
18:03 NotFound "anything" in that context is a tricky word.
18:03 moritz NotFound: "anything that parrot currently supports". Better?
18:03 NotFound If the encoding is "binary" there is no correct way, you sould know what you want to do.
18:04 moritz ok, not 'binary' then
18:04 moritz ("binary" doesn't make sense as an encoding to me; it's rather a non-encoding)
18:05 NotFound moritz: if you want to specify endianess, AFAIK parrot does not suppor it, you must set the ByteBuffer content byte by byte.
18:05 moritz NotFound: then let's say UTF-8 or $other_encoding
18:05 NotFound Then you need to transcode the string and then set the ByteBuffer to it.
18:05 moritz ok, thanks
18:08 whiteknight is string transcoding a common operation in Rakudo?
18:09 whiteknight I mean, is it something that could potentially be a bottleneck?
18:13 whiteknight okay, sub-profiler branch passes the handful of tests I've thrown at it. I'm going to merge
18:13 whiteknight DUCK AND COVER
18:13 dalek parrot: 6e67840 | Whiteknight++ | / (9 files):
18:13 dalek parrot: Merge remote-tracking branch 'origin/mls/sub-profiler'
18:13 dalek parrot: review: https://github.com/parrot/parrot/commit/6e67840309
18:13 whiteknight BOOM!
18:14 atrodo BAM!
18:14 whiteknight POW!
18:15 Coke cotto_work: I got a ton of rakudo reports, no need for parrot reports now.
18:16 Coke I thought rakudo transcoded source to utf-16  on startup.
18:19 whiteknight Coke: transcoding is a pretty tight loop over a bunch of function pointers. I think we could optimize that a little bit if it was a very common operation.
18:20 whiteknight if it's very common, it might be worth exploring
18:30 PerlJam mls++
18:34 benabik nqp.git/src/HLL/Compiler.pm:376 $in-handle.encoding($encoding);
18:35 dalek winxed: d1abd4f | NotFound++ | / (2 files):
18:35 dalek winxed: builtin 'sleep'
18:35 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/d1abd4fef8
18:35 benabik And rakudo uses ":encoding('utf8'), :transcode('ascii iso-8859-1')
18:35 benabik So, yes, Rakduo very commonly transcodes.
18:37 NotFound whiteknight: optimization attempts may lead to a lot of special casing code hard to maintain and error prone.
18:38 Coke see also: parrot.
18:39 whiteknight NotFound: that's my point. We can special-case a few common transcodings if it's worth it.
18:39 whiteknight we won't just do it unless it is an improvement
18:39 NotFound I'd say don't optimize it unless hard evidence of being a serious performace bottleneck.
18:40 whiteknight that's what I'm asking about
18:42 Coke whiteknight: need better profiling tools.
18:43 Coke which I think are getting here.
18:43 Coke Do we have any docs on what that sub profiler merge means to an HLL user?
18:43 whiteknight Coke: no docs yet. It's really been driven by mls and Rakudo users. We'll write docs up as we go
18:43 Coke "big merge to master" should probably be followed by "post to parrot-dev". (/me realizes he has not checked his mail today)
18:44 PerlJam Coke: with a call for volunteers to document it?  :)
18:45 whiteknight Coke: I'll send out an email soon
18:46 cotto_work I mentioned the sub profiler merge in #phasers
18:46 cotto_work they seemed happy
18:46 whiteknight thanks
18:46 whiteknight i forgot it was moved to wednesdays
18:47 cotto_work I like the new time.  It was hard to try focusing on #ps and #phasers at the almost same time.
18:53 benabik Of course, Rakudo won't build on master right now because NQP still uses get_addr, right?
18:53 benabik Or did the branch of NQP get merged into master?
18:54 Coke PerlJam: a spherical merge would contain its own docs.
18:55 benabik Coke: Spherical merges?  Are we physicists now?
18:55 PerlJam heh
18:56 PerlJam The first thing I thought of was "Assume a spherical cow of uniform density..."
18:56 Coke benabik: aren't we all?
18:56 Coke PerlJam: exactly!
18:56 * Coke needs a vacation.
18:57 whiteknight did we trash the get_addr op? I didn't think that happened yet
19:01 benabik I thought so?
19:03 NotFound I think get_addr must be changed to not use vtable get_pointer, if we want to get rid of Default.get_pointer
19:04 benabik I think get_addr is get_pointer, and there's a new get_id op.
19:04 benabik Or something.
19:05 NotFound 'get_addr and set_addr' These opcodes are being repurposed. They will always return  a unique memory address of the PMC.
19:05 NotFound From api.yaml
19:06 benabik I feel icky about giving memory addresses to things inside the VM.
19:06 NotFound benabik: is not a memory address, is a unique id
19:06 Coke I would feel much better if it said "unique id"
19:07 NotFound Being the address is just an implementation detail that can change.
19:07 Coke ... but I'm not sure why.
19:08 dalek parrot: 55a6542 | fperrad++ | tools/dev/mk_inno_language.pl:
19:08 dalek parrot: [inno] handles rakudo-nom which needs its own nqp
19:08 dalek parrot: review: https://github.com/parrot/parrot/commit/55a6542822
19:11 NotFound In the ticket I suggested (19 months ago, sigh) a new opcode name. No answers.
19:12 benabik I think the problem is that get_addr actually does something for some PMCs.
19:12 NotFound http://trac.parrot.org/parrot/ticket/218
19:13 NotFound See also TT #1857
19:14 NotFound I still see a lot of set_addr lying around
19:15 NotFound Is someone maintaining nqp-rx?
19:16 dalek jaesop: e8dd18a | Whiteknight++ | Makefile:
19:16 dalek jaesop: Add a rudimentary clean target to the makefile
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/e8dd18a349
19:16 dalek jaesop: 93905ae | Whiteknight++ | stage0/js/ (2 files):
19:16 dalek jaesop: Simple break and continue implementations
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/93905ae34c
19:16 dalek jaesop: 2685688 | Whiteknight++ | t/stage0/loops.t:
19:16 dalek jaesop: Add in a quick test file to show correct working of basic 3-clause for loop, continue, break
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/268568888b
19:16 dalek jaesop: 802ff79 | Whiteknight++ | stage0/js/wast.js:
19:16 dalek jaesop: Variable declarations sometimes use .name or .value for the name. I don't know why, but add a small fix.
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/802ff791fe
19:16 dalek jaesop: 684aec9 | Whiteknight++ | / (6 files):
19:16 dalek jaesop: Merge branch 'master' of github.com:Whiteknight/jaesop
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/684aec9d4a
19:16 dalek jaesop: c981309 | Whiteknight++ | / (6 files):
19:16 dalek jaesop: Re-add several tests files that were accidentally deleted. Cleanup Makefile after merge. Fix harnesslib to delete the temporary files, not the test files. Fix harness
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/c981309bb4
19:16 dalek jaesop: ccb4874 | Whiteknight++ | stage0/README.md:
19:16 cotto_work I thought the change from get_addr to get_id got merged.  plobsing++ submitted an nqp patch that gets Rakudo passing spectest, and which will be merged when nqp's ready.
19:16 dalek jaesop: update stage0/README now that we have continue/break
19:16 dalek jaesop: review: https://github.com/Whiteknig​ht/jaesop/commit/ccb48746a9
19:17 NotFound Oh, we have get_id ?
19:18 cotto_work someone does
19:18 cotto_work I think it's us.
19:18 cotto_work yup
19:18 NotFound And no mention of it in TT #218? Great.
19:18 whiteknight our collective knowledge of the software we work on every day is depresssingly low today
19:19 whiteknight on the bright side, we're setting ourselves up for pleasant surprises
19:20 cotto_work That got merged very recently.\
19:21 benabik NotFound: Going back and updating old TTs seems to be generally low priority.  I don't think get_id was added because of #218, just because plobsing wanted to kill get_addr in PMCs that aren't pointerish but people needed some value.
19:21 NotFound benabik: Looks like the ticket talks exactly about such concerns.
19:22 benabik Probably plobsing ran across the same issues separate from the TT.
19:22 whiteknight it's very plausible he didn't know about the ticket
19:23 NotFound No wonder, given the attention it seems to have received.
19:24 NotFound Anyway, this is orthogonal with the problem of set_addr being still used instead of set_label
19:39 zby_home joined #parrot
20:30 dalek rakudo/nom: 3b1fe55 | jnthn++ | src/Perl6/Actions.pm:
20:30 dalek rakudo/nom: Fix placeholder parameter sorting bug.
20:30 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3b1fe5500d
20:30 NotFound $ winxed -e 'float f0 = floattime(); sleep(0.0002); float f1 = floattime(); say(f1- f0);'
20:30 NotFound 0.000218868255615234
20:30 NotFound Looks like sleep does well.
20:40 dalek winxed: 5f5e69d | NotFound++ | winxedst1.winxed:
20:40 dalek winxed: fix emision of float that looks like int
20:40 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/5f5e69dd46
21:07 soh_cah_toa joined #parrot
21:08 cotto_work dukeleto: ping
21:15 dukeleto cotto_work: pong
21:15 cotto_work dukeleto: it seems that I have to pick between the VLIW thing and GitTogether.
21:16 cotto_work er, vmil
21:16 dukeleto cotto_work: oh yeah
21:16 dukeleto cotto_work: i guess i already made up my mind :) Bought my ticket a few hours ago.
21:16 cotto_work dukeleto: for which?
21:16 dukeleto cotto_work: GitTogether. I went last year and it was a blast
21:17 dukeleto cotto_work: hung out with lots of Git core devs and Github dudes
21:17 cotto_work I guess we'll have Parrot representation at both events then.
21:17 dukeleto cotto_work: works for me.
21:17 cotto_work stupid scheduling conflicts
21:17 dukeleto cotto_work: indeed. Perhaps it is better this way, anyway.
21:17 dukeleto cotto_work: please take lots of good notes at the VLIW thing and see if there are any good ideas to steal :)
21:17 dukeleto cotto_work: i will do the same :)
21:18 cotto_work dukeleto: deal
21:37 dukeleto https://github.com/blog/936-one-million
21:41 dukeleto Also, this looks like an interesting job for somebody near Oakland, CA: http://www.rootlabs.com/developer-job.html
21:57 dalek winxed: f72098c | NotFound++ | winxedst1.winxed:
21:57 dalek winxed: allow specifying types in 'multi' modifier, only simple types by name for a now
21:57 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/f72098c0a2
22:09 athomason joined #parrot
22:21 dalek rakudo/nom: a26660d | jnthn++ | src/Perl6/Metamodel/ (2 files):
22:21 dalek rakudo/nom: Fix diamond role composition.
22:21 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a26660d7bc
22:21 dalek rakudo/nom: b0b9b78 | jnthn++ | src/Perl6/Actions.pm:
22:21 dalek rakudo/nom: Ensure that we compose a ParametricRoleHOW before adding it to the group, so the group gets the fully composed view of it.
22:21 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b0b9b78a5a
22:45 dalek rakudo/nom: 1dd4914 | jnthn++ | src/Perl6/Metamodel/BOOTSTRAP.pm:
22:45 dalek rakudo/nom: Oops, a26660d made us accidentally miss some real conflicts. Fix that.
22:45 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1dd491459c
22:45 dalek rakudo/nom: b72d678 | jnthn++ | src/Perl6/Metamodel/ (4 files):
22:45 dalek rakudo/nom: Roles pass any classes they're made to inherit along.
22:45 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b72d678a50
22:45 dalek rakudo/nom: df4db06 | jnthn++ | t/spectest.data:
22:45 dalek rakudo/nom: We now pass S14-roles/composition.t.
22:45 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/df4db06261
22:59 whiteknight joined #parrot
23:01 cotto_work NotFound: ping
23:02 whiteknight good evening #parrot
23:02 cotto_work hi whiteknight
23:02 tadzik good night whiteknight
23:02 cotto_work aloha: clock?
23:02 aloha cotto_work: LAX: Wed, 16:02 PDT / CHI: Wed, 18:02 CDT / NYC: Wed, 19:02 EDT / UTC: Wed, 23:02 UTC / LON: Thu, 00:02 BST / BER: Thu, 01:02 CEST / TOK: Thu, 08:02 JST / SYD: Thu, 09:02 EST
23:03 cotto_work Hmph.  NotFound's probably sleeping.
23:03 dalek rakudo/nom: e606bec | jnthn++ | src/binder/bind.c:
23:03 dalek rakudo/nom: Toss dead code.
23:03 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e606bece68
23:03 dalek rakudo/nom: 5440fde | jnthn++ | src/pmc/perl6lexpad.pmc:
23:03 dalek rakudo/nom: Add a write barrier.
23:03 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5440fdefe3
23:06 whiteknight sleeping? slacker
23:09 cotto_work seriously
23:11 bubaflub joined #parrot
23:51 nopaste joined #parrot
23:58 cotto ~~

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

Parrot | source cross referenced