Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 13 tickets, merge outstanding branches, help NotFound clean up PBC tests
Set by moderator on 20 August 2010.
dukeleto cotto_work: perhaps git for-each-ref would help you 00:00
00:01 Paul_the_Greek left
dukeleto cotto_work: take a look at rails.wincent.com/wiki/Migrating_S...ies_to_Git and github.com/nothingmuch/git-svn-aban...n-fix-refs 00:01
00:01 Psyche^ joined
dukeleto cotto_work: or you can just parse the output of "git branch" 00:01
cotto_work: and do a git log on each branch 00:02
cotto_work Great! I'll play with those when I get home.
dukeleto cotto_work: you would want something like "git log -p master and then git log -p master...branch"
cotto_work git++ 00:03
dukeleto the ... makes a "commitish"
cotto_work rebase ftw
00:03 masak left
dukeleto goes for a run 00:03
cotto_work: good luck
purl You'll need it.
cotto_work purl, go play in traffic
purl wanders off to dent some cars.
cotto_work dukeleto: don't ;)
00:06 Patterner left, Psyche^ is now known as Patterner
dalek rrot: r48647 | jkeenan++ | branches/tt677_toolsdirs/t (3 files):
Correct construction of paths to programs inside corresponding test files.
 00:07
00:14 davidfetter left
cotto_work goes home 00:15
00:18 nwellnhof left 00:24 SingAlong left 00:25 SingAlong joined 00:32 kid51 is now known as kid51_at_dinner
dalek rrot: r48648 | jkeenan++ | branches/tt677_toolsdirs/config/gen/makefiles/root.in:
Use 'make' variable $DEV_TOOLS_DIR throughout in place of tools/dev.
 00:40
00:55 dduncan joined 00:56 dduncan left, Patterner left 01:00 Paul_the_Greek joined
Paul_the_Greek Evening, kids. 01:00
whiteknight hello Paul_the_Greek 01:01
Paul_the_Greek Hey there whiteknight.
Coke HEEEYYYYY YOOOOOU GUUUUYS! 01:05
Paul_the_Greek Coke!
01:06 Psyche^ joined, Psyche^ is now known as Patterner
Coke www.youtube.com/watch?v=KC2lrcyNtx8 (hey you guys) 01:09
msg plobsing Nope, I can't duplicate it either, thanks, closed. 01:10
purl Message for plobsing stored.
Coke plobsing++ 01:11
01:20 Paul_the_Greek left
dalek rtcl-nqp: 1d99ca5 | coke++ | / (3 files):
Add do-nothing stub for [encoding]
 01:20
rtcl-nqp: 1f33ddf | coke++ | t/tcl_var_subst.t:
untodo passing test.
rtcl-nqp: a2ef165 | coke++ | src/Partcl/Actions.pm:
make $::foo always refer to the top level HLL ns.
rtcl-nqp: 799f0e7 | coke++ | src/Partcl/commands/file.pm:
Avoid walking off the ends of strings and arrays.
rtcl-nqp: f79d43a | coke++ | / (3 files):
implement "in" and "ni" operators.
TT #1459 closed by coke++: missing dep on parrot_config 01:24
TT #1459: trac.parrot.org/parrot/ticket/1459
01:27 kid51_at_dinner is now known as kid51
cotto comcast-- 01:37
If I wanted random newsletters, I'd sign myself up for them. 01:38
01:39 whiteknight left 01:45 rurban_ joined 01:49 rurban left, rurban_ is now known as rurban
Coke digs the new ceelo video. 02:15
andy++'s protests aside. ;) 02:16
kid51 cotto Can you take a look at trac.parrot.org/parrot/ticket/865 ? It may now be closable. 02:25
cotto sure
02:26 SingAlong left 02:36 janus left 02:42 janus joined
dalek TT #1725 closed by jkeenan++: tools/build/headerizer.pl: documentation is confusing 02:48
TT #1725: trac.parrot.org/parrot/ticket/1725
TT #1255 closed by jkeenan++: fix PARROT_EXPORT visibility=default for gcc other than 4.x
TT #1255: trac.parrot.org/parrot/ticket/1255
cotto kid51, would it make sense to make that configure step fatal? 02:52
dalek rtcl-nqp: a5b7f74 | coke++ | / (2 files):
implement basic 2-arg rename.
 02:53
rtcl-nqp: 70d0cf1 | coke++ | src/Partcl/commands/namespace.pm:
Provide a few variants on getting the ns (actual Namespace, array of elements, or a canonical tcl string.)
02:53 tetragon left
moderator Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 20 tickets (18 to go), merge outstanding branches, help NotFound clean up PBC tests
cotto apart from that, that change addresses the ticket nicely. kid51++ 02:54
kid51 cotto: I think our stance has always been that if a user wants to make a particular configure step(s) fatal, user can do so with --fatal-step option. 03:10
But, by default, configure tries to go as far as it can and then prints out results for user to decide.
kid51 must sleep 03:11
purl Sleep is for the weak.
03:11 kid51 left
moderator Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 20 tickets (17 to go), merge outstanding branches, help NotFound clean up PBC tests 03:24
dalek TT #865 closed by cotto++: extra .pmc files in src/pmc cause silent pbc compatibility breaks 03:39
TT #865: trac.parrot.org/parrot/ticket/865
cotto it's quiet again 04:33
clearly everyone's so busy and engrossed hacking on some Parrot-related thing that they're ignoring irc. 04:34
I'd better not provide further distraction.
plobsing damn, too late 04:38
tcurtis isn't doing anything substantial.
dalek rrot: r48649 | plobsing++ | trunk (2 files):
convert pbc_to_exe over to Getopt;Obj to facilitate adding new options
 04:51
purl I don't know how to convert pbc_to_exe over to Getopt;Obj to facilitate adding new options.
cotto too late? 05:01
plobsing too late to avoid distraction 05:02
dalek p-rx: e71d569 | bacek++ | / (2 files):
Use @a.push form instead of pir::push
 05:06
05:09 plobsing left 05:18 tcurtis left 06:04 uniejo joined
dalek thub-trac: 6ce3f7a | ++ | github/github.py:
add validity check to hashes wikification syntax
 06:08
thub-trac: e430ed6 | ++ | / (2 files):
use git log output to populate the db, add sparse documentation
06:13 theory left 06:24 fperrad joined 06:29 aloha left, bacek left 06:31 aloha joined 06:40 bacek joined
dalek nxed: r602 | NotFound++ | trunk/winxedst1.winxed:
initial implementation of new keyed in stage 1
 06:58
07:55 tadzik joined
dalek TT #1446 closed by NotFound++: Exception PMC 'clone' does not work 08:41
TT #1446: trac.parrot.org/parrot/ticket/1446
rrot: r48650 | NotFound++ | trunk/t/pmc/exception.t:
delete referece to closed ticket TT #1446
 08:45
moritz fwiw, in the Perl 6 test suite I find references to closed tickets still helpful 08:46
two reasons: 1) the tickets often contain more explanation than the test itself and 2) if a test regresses, I know which ticket to re-open
08:51 nwellnhof joined
sorear I wonder how that will fare when there are more implementations than just Rakudo running around 08:57
moritz we just link to other bugtrackers too 08:58
09:00 tadzik left
NotFound moritz: a ticket thet just says 'implement this' is not very useful 09:09
moritz NotFound: agreed
09:13 tadzik joined
dalek rrot: r48651 | NotFound++ | trunk/t/pmc (2 files):
some improvement in exception/exceptionhandler tests
 09:19
09:25 aloha left 09:27 bacek left, aloha joined
dalek TT #1712 closed by NotFound++: Packfile tests incorrectly try to read non-native packfiles. 09:32
TT #1712: trac.parrot.org/parrot/ticket/1712
09:38 bacek joined 09:43 lucian joined 09:45 rurban_ joined 09:49 rurban left, rurban_ is now known as rurban
dalek kudo: 3a339ee | ++ | src/pmc/objectref_pmc.template:
Switch objectref_pmc.template to auto_attrs.

Fixes a nasty memory leak that would leak at least 3 scalars per block. Now a simple 'while 1 { }' loop seems not to leak anymore.
Signed-off-by: Moritz Lenz <moritz@faui2k3.org>		 10:21
10:22 masak joined 10:30 tadzik left 10:31 bacek left, aloha left 10:37 aloha joined, aloha left 10:38 aloha joined, aloha left 10:48 bacek joined, bacek left 11:04 bacek joined, bacek left 11:05 aloha joined
dalek TT #1756 created by NotFound++: distutils inst_lang fails on non-array 11:13
11:13 aloha left
dalek TT #1756: trac.parrot.org/parrot/ticket/1756 11:13
11:21 bacek joined 11:22 bacek left 11:23 bacek joined, bacek left 11:26 smash joined
smash hello everyone 11:26
11:33 aloha joined, aloha left 11:35 aloha joined, aloha left 11:39 bacek joined 11:40 bacek left 11:43 fperrad_ joined 11:45 fperrad left, fperrad_ is now known as fperrad, aloha joined 11:46 aloha left 11:53 fperrad left, whiteknight joined 11:54 fperrad joined 12:03 aloha joined
mikehh smash: ping 12:05
dalek rrot: r48652 | nwellnhof++ | trunk/src/pmc/capture.pmc:
Make sure custom_mark flag of captures gets set in pmc_set
 12:06
12:07 cotto_work left, aloha left 12:11 rblackwe left, rblackwe joined 12:16 aloha joined, aloha left 12:18 khairul joined 12:24 aloha joined 12:32 aloha left 12:38 aloha joined
whiteknight good morning, #parrot 12:43
masak morning, wk 12:45
whiteknight hello masak, how are you today?
masak I'm a bit tired, but otherwise fine. how are you? 12:46
whiteknight same. Exactly the same
smash mikehh: pong
masak I thought the chai I made yesterday was weak, but my heart rythm and my failed attempts to go to sleep at 2:30 suggest otherwise... :)
whiteknight at least you got good chai. All I got was a crying baby at 2:30 12:47
chai > baby
masak it wasn't particularly good chai.
it tasted mostly of milk.
whiteknight okay, maybe baby > chai
masak at least the non-crying kind. 12:48
nwellnhof masak: i made a bit of progress on #1746 12:49
masak nwellnhof: aye; I saw the updates to the ticket.
nwellnhof: I've made some progress here as well.
nwellnhof i got a reproducible segfault with the gcdebug runcore 12:50
masak soudns very good.
nwellnhof but then i hit a wall :(
masak nwellnhof++
moritz should teach the IRC logs to treat 4-digit ticket numbers as track tickets
just as it does now for TT#1746
masak moritz: the correct algorithm will be a sliding-window thing checking year and month against the ticket number :)
moritz masak: or just be channel sensitive, and default to RT on #perl6, and to parrot on #parrot 12:51
nwellnhof i think i have to leave it to someone more knowledgeable
masak moritz: that sounds like a good idea, yes. 12:52
nwellnhof moritz: i can also stick with TT#... 12:53
moritz: but it seems that #... is also used a lot
moritz nwellnhof: lots of people here only use #... - which is a sign that the technology should adapt to the people, not the other way round 12:54
opbots: trust nwellnhof
slavorg Ok
slavorgn Ok
mikehh smash: was just wondering about PaFo voting, thought I was a member, but appears not
nwellnhof moritz: do you handle git commit hashes in the irclogs? 12:55
moritz: i imagine that's tricky
smash mikehh: i can verify if you're in the members list, just confirm me your name
moritz nwellnhof: no; too many git projects out there
mikehh mikehh - Michael Hind 12:56
nwellnhof moritz: that's one advantage of svn
moritz if we invent a convention for it (like pg:a1b2c3 for "parrot git" or so) it's fairly easy to do
nwellnhof moritz: maybe github has a commit hash redirection service 12:57
moritz interesting idea 12:58
smash mikehh: you are not a foundation member, you can apply if you want (and meet the requirements) 12:59
moritz if mikehh needs a nomination, I can contribute one :-)
mikehh smash: yeah - thought I was joined up after last years meeting, and AFAIK I meet the requirements 13:01
whiteknight I hereby nominate mikehh
mikehh smash: been committing for a year, last Release Manager 13:02
moritz also nominates mikehh
mikehh whiteknight: thanks
and moritz 13:03
smash mikehh: ok, i'll process your apply, thks moritz++ and whiteknight++ for nominations
mikehh: i'll get back to you as soon as i wrap everything up 13:04
mikehh smash: I'll be around here for a while if you need anything
smash mikehh: ok, thks 13:05
mikehh: am i guessing correctly that you already supplied a signed CLA right ? 13:06
mikehh smash: correct 13:08
smash mikehh: ok 13:10
mikehh smash: about a year or so ago 13:11
dalek nxed: r603 | NotFound++ | trunk/setup.json:
one more little step towards install
13:16 uniejo left
nopaste "NotFound" at 192.168.1.3 pasted "load_language wined works" (37 lines) at nopaste.snit.ch/23099 13:19
masak whiteknight: re trac.parrot.org/parrot/ticket/1746 -- I took the 2010.02 release of Rakudo and massaged it so that it could accept the patch that causes recent Rakudos to segfault. 13:26
whiteknight: so what you said yesterday about Parrot releases being more stable is not only not true in this case, but very not true.
I don't think I can go back further than 2010.02 -- before that there's only the old development branch of Rakudo, and I fear the differences are too big. 13:27
but this segfault has been in Parrot for many months.
Coke (configure, fata-step) I think our stance is that "no one wants to own configure so we don't talk about it" =-) - I think expecting a user to use "fatal-step" is too much of a burden on the user. I personally would like it if it a failed config step would abort the entire process. I cannot tell you how many times I've hit ^C fruitlessly while config is running to try to /make/ it stop. ;) ... but I don't want to own configure, so I will now l
Free Karma? someone rerun the faces script and update the wiki... 13:28
13:28 fperrad_ joined
Coke mikehh: you don't have commit bits!? 13:28
ah, you don't have member bitrs. 13:29
yah, that process is... suboptimal.
smash Coke: we were handling his PaFo membership
NotFound plays 'Give a little bit'
Coke I thought smash tried to fix that in the months leading up to voting.
smash! You're here! 13:30
smash++ mikeh++
mikehh++ #even
13:31 fperrad left
smash Coke: i processed all the requests that i got 13:31
13:31 fperrad_ is now known as fperrad
Coke I'm sure. I wasn't saying you screwed up, I was just thinking out loud that I thought you worked on it. 13:33
smash Coke: hehe 13:34
Coke sorry, I know it's frustrating when you volunteer and someone says "YOU SCREWED UP", so just watned to be clear I am not saying that. ;) 13:35
smash Coke: ah, no worries.. and pleas do can tell me that i screwed up when if i do something wrong 13:36
Coke hulk?
purl Hulk SMASH!!!
smash s/pleas do can tell/please do tell/
Coke wonders if he still has some issues of the hulk lying about.
Coke used to collect comics over 20 years ago and still has most of them in plastic in boxes. 13:37
... there's a guy at work here who does not say "you are ell", but "earl" for URLs. it's like nails on chalk. 13:38
13:40 fperrad_ joined
pmichaud (rakudo should only target parrot releases...) are we still suffering from that delusion that this makes any sort of difference? 13:41
moritz well yes, dangerous branch merges are usually held off until after a release
whiteknight pmichaud: only in the sense that targetting HEAD leads to the least stability. You can't complain about Parrot's stability if you target Parrot HEAD
pmichaud rakudo releases never target HEAD 13:42
whiteknight but rakudo development does track it relatively closely
pmichaud we don't have a choice.
moritz pmichaud: or do you means "target *supported* releases"?
pmichaud well, we *do* have a choice, we could stall our development speed by about 2 weeks.
whiteknight I don't really care in any case. I'm not against anybody tracking Parrot HEAD
pmichaud Then why bring it up as a (false) reaction to a legitimate bug? 13:43
moritz whiteknight: ironically we do have to track HEAD to be able to complain about parrot bugs in time
whiteknight I'm against the combination of tracking HEAD *and* complaining about stability. You can do one or the other, you can't do both
pmichaud we can't complain about the stability of the releases? wtf?
whiteknight pmichaud: no releases. Forget the word releases. I'm not talking about releases
13:44 fperrad left, fperrad_ is now known as fperrad
whiteknight Releases should always be perfectly stable, and if they are not there should be complaints 13:44
masak in this case, the bug has been in Parrot for more than half a year. so in this case, tracking HEAD or not is moot. 13:45
I'm now complaining that releases from February and on have not been stable.
pmichaud If masak supplies his patch to the release of Rakudo that targets Parrot 2.6.0, and that exhibits a segfault, are we not allowed to "complain about stability"?
masak sorry I didn't discover it sooner :)
moritz it's been a very stable bugs for many months :-) 13:46
masak in a way, this is a kind of ideal segfault. it occurs very regularly. more eyeballs should be able to dig it out and eliminate it.
I'm hopeful, at least.
as further encouragement, recall that I'm blocking on this for the book "Using Perl 6". the patch I can't apply to Rakudo enables things I want to write about in the book. 13:47
here's the ticket again: trac.parrot.org/parrot/ticket/1746 13:48
kthx
pmichaud masak: is that patch still an issue on release #32 of Rakudo? 13:51
i.e., does it still apply cleanly and exhibit the segfault? 13:52
masak last I checked against Rakudo HEAD, it did.
I think that might have been around or after the release.
masak checks now
Coke pmichaud: OOC, will R** include the tylercurtis fix?
I think complaining about stability in a general sense is fine.
Bug reports are awesome, please keep them coming.
13:53 mikehh_ joined
Coke (because parrot is not stable, at releases or anywhere.) 13:53
masak I'm complaining about stability in a specific sense. :P
pmichaud yeah, I'm a little distributed by any attitude of "you should only report bugs for releases".
*disturbed
masak and that didn't matter in this case.
whiteknight I never had that attitude
pmichaud (wow, that was a weird typo)
masak it affects several recent releases.
pmichaud 13:43 <whiteknight> I'm against the combination of tracking HEAD *and* complaining about stability. You can do one or the other, you can't do both 13:54
whiteknight there's a difference between reporting bugs and complaining that parrot is lousy and bug-riddled
pmichaud evidence please.
Please show me where someone said "Parrot is lousy and bug-riddled"
Coke There is certainly some snark that exists on #perl6 about parrot. but it often results in reportable (and therefore fixable) bugs. 13:55
whiteknight pmichaud: I'm not trying to start any arguments here. There is a distinct difference between a "complaint" and a "bug report"
Coke pmichaud: if you review from timtoady matching parrot, you'll find something close, anyway. ;)
13:55 mikehh left, mikehh_ is now known as mikehh
whiteknight I'm very careful to use one term when I mean the one concept and not the other 13:55
pmichaud whiteknight: what complaint are you referring to, then?
mikehh opbots, names
pmichaud if Rakudo is incorrectly complaining about Parrot, I'd like to put a stop to it. 13:56
Coke: TimToady/parrot -- I think most of that snark has to do with Parrot design issues, not with bug-ridden-ness
whiteknight pmichaud: the exact instance is irrelevant. That doesn't mean I should be accused of having some kind of disturbing attitude when I was very careful about the words I used
Coke pmichaud: you can see where it would be hard to tell. 13:57
pmichaud Coke: certainly.
Coke so, can we all just hug and make up now?
pmichaud Coke: but since whiteknight is now claiming precision, I'd like to see some.
Coke except TimToady, he's too mean. Þ
moritz should I send hugme in? :-)
masak :)
pmichaud moritz: likely.
Coke in here? why don't you try him out in #perl first. Þ 13:58
moritz Coke: :-)
pmichaud I'm just tired of hearing "Rakudo should only be targeting Parrot releases", because it is now and always has been a delusion on our parts to think that it matters in any significant way.
I mean, we file a bug report, and the response is essentially "don't complain about things that don't have to do with releases" 13:59
whiteknight pmichaud: who ever said that? I've never seen that said
pmichaud (I grant this response is coming from only one person.)
whiteknight bug reports have always been appreciated
pmichaud whiteknight: did you not just say that someone is complaining about stability? 14:00
Okay, okay, never mind. I'll drop it.
whiteknight pmichaud: again, difference between "bug report" and a "complaint". That said, I did see something yesterday that I did legitimately take as an insulting complaint about stability
Coke whiteknight: instead of saying "i said what I said and I meant what I meant", perhaps you could consider that someone was easily able to read what you said the wrong way and apologize. 14:01
whiteknight Coke: that may be, and I'm going to lengths to try and clarify
Coke Perhaps instead of clarifying, we can all just start over. 14:02
So... pmichaud... what's up, dude?
how's rakudo working out for you?
pmichaud Coke: sorry, I don't have a response I can make at the moment. 1/2 :-) 14:03
masak pmichaud: confirmed; patching Rakudo HEAD and running 'use A; use B' causes the segfault.
Coke I don't know what "1/2" means in this context. 14:04
pmichaud masak: I was somewhat interested in patching the #32 release, so we could tie it to a parrot release. But it's not important now, other than I'll look at your bug report shortly and see if I can figure out what's going on. That looks like a really weird GC issue.
masak whiteknight: I sincerely hope that it wasn't something I said about stability. Parrot is far, far better in this department than in, say, s008.
er, *2008
14:04 Andy joined
masak pmichaud: i"ve already tied it to the February and May parrot releases. I'd be very surprised if the problem somehow fixed itself in time for a release, and then re-appeared. 14:05
Coke ah, here's our peacemaker. except when he's being a jerk. Andy, everyone!
Andy You all suck.
purl gang rape costs extra
Coke hugs andy, and was just kidding about the jerk thing.
Andy Kiss my ass, Coleda. 14:06
Let's fight!
masak you're the worst peacemaker ever. :)
Coke is this your first time here? then you must fight.
Coke thinks we need our own infobot.
Andy I could try to be a pacemaker, but keep me away from microwave ovens.
PerlJam Seems like a good fundraiser for YAPC/TPF/whatever (fight your "favorite" perl personality :)
Coke trying to make purl forget that fact in private, no love. ah well. 14:07
14:07 bacek joined
Coke bacek_at_work: speaking of infobots... 14:08
I don't mind a chatty infobot as long as he's not contaminated by #perl.
moritz +1 to having our own 14:09
szbalint well volunteered! :)
moritz damn 14:10
14:11 fperrad_ joined
Coke moritz: this is why I try to keep my mouth shut in here. ;) 14:11
moritz Coke: well, as long as the volunteering is not tied to a time frame... :-) 14:12
14:15 fperrad left, fperrad_ is now known as fperrad
TiMBuS you can use my perl 6 irc bot framework if you want :) 14:17
(but yeah dont, i keep refactoring it) 14:19
bacek aloha, humans 14:21
aloha, perl?
aloha bacek: No clue. Sorry.
bacek aloha, perl is great
aloha bacek: Okay.
bacek aloha, perl?
purl hmmm... aloha, perl is great
aloha bacek: perl is great
bacek stupid girl
Coke, I have infobot plugin in aloha :) 14:22
14:27 plobsing joined
pmichaud (R** and tylercurtis fix) I was planning to use the release version of Rakudo, which wouldn't include the fix. 14:27
Coke hokay. there's always next time. no worries. 14:28
(yay, R*** already wins!)
whiteknight good morning, bacek 14:29
bacek whiteknight, it's 0:30am here :) 14:30
whiteknight technically: morning
bacek definitely
14:31 fperrad left, fperrad joined
Coke aloha? 14:32
aloha: coke?
purl somebody said coke was Will Coleda <mailto:will@coleda.com>
aloha Coke: Sorry, I don't know.
Coke aloha, coke is Will Coleda <mailto:will@coleda.com>
aloha Coke: Okay.
Coke aloha: coke?
purl rumour has it coke is Will Coleda <mailto:will@coleda.com>
aloha Coke: coke is Will Coleda <mailto:will@coleda.com>
purl i already had it that way, aloha.
Coke what else does purl do that we can't live without. msg?
whiteknight aloha: i already had it that way, aloha is <reply>Coke? 14:33
moritz aloha: msg Coke maybe it even works
purl Message for coke stored.
moritz aloha: purl is really stupid
aloha moritz: Okay.
Coke aloha: no, purl is just misguided.
purl okay, Coke.
aloha Coke: Okay. 14:34
Coke moritz: looks like purl took the msg, not aloha.
if we can get that working, I'm happy to boot purl. 14:35
moritz Coke: that's why I wrote "purl is really stupid"
Coke hurm. we'd also lose purl-karma. 14:37
we could also ask hachi about options.
moritz Coke: In the long run I plan to extract karma from the public IRC logs 14:38
whiteknight how does purl store it's karma info? DB? flat file?
14:42 aloha left, aloha joined
Coke so, to channel chromatic for a minute; would it be helpful to reduce that perl6-with-patch chunk to a PIR-only snippet? 14:45
(that certainly seemed to help track down the auto_attrs issue in rakudo yesterday.) 14:46
14:46 tcurtis joined
moritz I'm sure it would; I'm not sure how easy or hard it would be 14:46
Coke I basically ran --target==pir and then started ripping things out until the leak went away. 14:47
(so, still had the ops & the setting.) 14:48
14:49 aloha left
pmichaud I wouldn't be surprised if it's a runloop issue. 14:49
14:49 aloha joined
bacek aloha, tell Coke it should work 14:49
aloha bacek: OK. I'll deliver the message.
masak the patch is made to the setting, so there are good chances that it might be detached from the Rakudo source altogether. 14:50
it also hasn't been properly minimized.
pmichaud I'm a little concerned about that class-within-a-loadinit
bacek aloha, karma Coke
aloha bacek: Coke has karma of 0.
bacek Coke++
aloha, karma Coke 14:51
aloha bacek: Coke has karma of 1.
pmichaud oh, and there's a role in there as well.
hmmmmm
that feels like it's likely to cause problems
masak: have you tried it using a named class/role instead of the anonymous ones? 14:52
masak no. I also think those anon things are the ultimate cause of the problem.
masak tries
moritz masak: fwiw, that patch doesn't apply cleanly on my rakudo copy 14:53
14:53 theory joined
masak moritz: I have a local branch which I keep rebasing. so I'm not actually applying a patch here. 14:53
pmichaud + method WHAT { $enumeration-object } 14:54
that looks Evil.
masak oh?
it's spec.
pmichaud since $enumeration-object is not in fact a type object.
masak it's an undefined object just like type objects are.
pmichaud even with the .new? 14:55
masak hm.
pmichaud it looks to me like it's a full-fledged instance of the class, not the class's type object
masak (class {}) doesn't yield a type object? 14:56
pmichaud yes, but you do (class {}).new
masak oh wait. there's a .new there.
why did I add that again?
that doesn't seem necessary. 14:57
pmichaud beyond that, I'm not sure why you'd need to declare .WHAT in the first place. 14:58
the objects will already know their .WHAT 14:59
masak I remember there being some reason. 15:01
I'm now unexpectedly called away, but I'll run the patch with named class/role while offline.
o/
15:02 masak left 15:04 theory left 15:08 theory joined
dalek thub-trac: 22fa9a4 | ++ | github/github.py:
make tooltips work
 15:21
15:40 theory left
Coke pmichaud: hey, you implemented partcl-nqp's version of variables. Got a sec to discuss namespace vars? 15:42
pmichaud sure 15:43
but just a short second -- need to head off to lunch
15:43 cotto_work joined
cotto_work good morning 15:44
purl For you maybe.
Coke so, you can do: namespace eval foo { variable bar} , and then you have $::foo::bar (or would if that syntax did anything yet.)
cotto_work . o 0 (We should probably kill the bots before they turn on us. Let's start with purl.) 15:45
Coke for storage, i was thinking of adding a %GLOBALS whenever a namespace was created, like the correspoding one in the HLL tcl root.
pmichaud Coke: seems reasonable
Coke In the old system, i was putting namespace vars in as globals and lexicals, and I have no idea why. :( 15:46
pmichaud well, in Rakudo we put "our" variables in as both package and lexical bindings
Coke Probably to simplify differentiating between variable-in-ns and variable-in-proc.
moritz right
pmichaud my $x = 5; { say $x; { our $x = 7; say $x; } } 15:47
the innermost $x needs to know that it's bound to the package $x, not the outer lexical.
Coke hokay. I can wire up the %GLOBAL creation to [namespace eval] and then make [variable] work and then make $foo::bar work.
in tcl, vars aren't really lexical like that, so it's easier.
pmichaud right.
moritz Coke: how far is partcl-nqp behind the old partcl? 15:48
Coke moritz: <cry>
the old one was able to run most of the spec tests.
the new one is extensible! but not extended.
so it doesn't even run the partcl-only tests that were running before.
moritz I know that feeling all too well
pmichaud afk, lunch 15:49
Coke hurm. I think I can simplify $foo::bar by putting wiring up [namespace eval] to take the GLOBALS, if any, and make them lexical for that block. then I can just keep looking in the same place when getting their vals. 15:51
16:04 pyrimidine joined 16:06 khairul left
cotto_work How do I delete any local-only commits in git? 16:18
moritz git reset --hard origin/master 16:19
cotto_work moritz++
moritz (but it's not "hard" gone... git reflog will still show you the commits for a while)
szbalint yeah, until some gc run right? 16:20
moritz right
git gc
or git repack -A -d # or so
mikehh All tests PASS (pre/post-config, make corevm/make coretest, test, fulltest) at r48652 - Ubuntu 10.04 amd64 (g++ with --optimize) 16:24
cotto_work allisonrandal.com/2010/08/20/ubuntu-ta-intro/ 16:27
szbalint ooh 16:28
moritz doesn't find a date for that blog entry
except in the uRL 16:29
TiMBuS oh boy! allison can now fix all of mark shuttleworths bad ideas :P 16:30
dalek rrot: r48653 | gerd++ | trunk/lib/Parrot/Configure/Compiler.pm:
add a method to replace entries like "@key@" only for shebang at the first line of a file
 16:34
Andy wait, what? 16:46
purl yeah, seriously
Andy Allison got a real job?
dukeleto Andy: seems that way 16:48
dalek rrot: r48654 | gerd++ | trunk (3 files):
set the "@key@" shebang value in the "tools/dev/mk_language_shell.pl.in"
 17:08
rrot: r48655 | plobsing++ | trunk/tools/dev/pbc_to_exe.pir:
add --runcore option to pbc_to_exe in order to optionally use the slow core
17:09 robin-gvx joined
dalek TT #1653 closed by plobsing++: pbc_to_exe loses file information. 17:10
TT #1653: trac.parrot.org/parrot/ticket/1653
moderator Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 20 tickets (16 to go), merge outstanding branches, help NotFound clean up PBC tests 17:10
17:14 brianwisti joined 17:15 plobsing left 17:23 allison joined 17:25 nwellnhof left 17:39 fperrad_ joined 17:41 fperrad left, fperrad_ is now known as fperrad 17:42 ruoso joined 17:45 rurban_ joined 17:49 rurban left, rurban_ is now known as rurban 17:50 rurban left 17:56 ruoso left
Coke wonders if he would like his job more if he was paid to hack on parrot related issues. 19:00
atrodo Coke> I'd assume all of us would like our job more if we could hack on things that interest us all day 19:01
cotto_work doesn't
moritz btw trac.parrot.org/parrot/ticket/1752 should be pretty easy for somebody who actually understands some parrot guts 19:02
brianwisti I'm not sure it would interest me as much if I would get fired for not doing it. 19:06
But still, it would be nice if Parrot was in some vague way work-related.
atrodo brianwisti> I think about it in optimal terms. If I get paid to hack on things that interest me, then by definition, as long as I'm hacking, I'm fulfilling my jobs. 19:10
However, I'm still looking for a jobs with that description
szbalint what I need at a job is freedom and to do something interesting. Interesting in itself is not enough 19:14
Coke moritz: if the function docs for the function it calls are to be believed, there's your answer. 19:15
19:27 hercynium joined 19:31 lucian_ joined 19:32 mikehh_ joined 19:34 lucian left, mikehh left 19:39 nwellnhof joined
TimToady phoan in faiv 19:54
Coke meh 19:55
dalek rrot: r48656 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
[distutils] fix TT #1756
 19:56
20:09 whiteknight left 20:10 fperrad left 20:16 jan left 20:18 eternaleye_ joined
NotFound fperrad++ 20:20
20:20 eternaleye_ left, eternaleye_ joined
cotto_work NotFound: did you intend to close that ticket? 20:23
20:25 eternaleye_ left, jan joined 20:27 eternaleye_ joined 20:29 eternaleye_ left 20:30 eternaleye_ joined
dalek rrot: r48657 | NotFound++ | trunk/tools/dev/pbc_to_exe.pir:
add a cast to pbc_to_exe generated code for c++ builds
 20:30
20:32 eternaleye_ left
NotFound cotto_work: if fperrad doesn't, I'll do it. 20:32
cotto_work is not a fan of the name Parrot_Run_core_t 20:33
NotFound If someone wants easy karma, go for it.
cotto_work done
moderator Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 20 tickets (15 to go), merge outstanding branches, help NotFound clean up PBC tests 20:34
dalek TT #1756 closed by cotto++: distutils inst_lang fails on non-array 20:34
TT #1756: trac.parrot.org/parrot/ticket/1756
NotFound You can drop that call for help from the topic, it was for the last week ;) 20:35
20:36 eternaleye_ joined
moderator Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 20 tickets (15 to go), merge outstanding branches
cotto_work I'm not sure if we discussed merging outstanding branches this week, but it wouldn't hurt to do some merging or pruning. 20:37
NotFound Agreed
20:38 eternaleye left 20:40 eternaleye_ is now known as eternaleye 20:46 plobsing joined 20:47 lucian joined 20:51 lucian_ left 21:01 robin-gvx left 21:10 whiteknight joined 21:12 robin-gvx joined
GeJ Bonjour everyone. 21:19
cotto_work nakkaya.com/2010/08/24/a-micro-manu...nted-in-c/ 21:21
21:22 lucian_ joined
whiteknight hello GeJ 21:23
GeJ Good evening whiteknight 21:24
21:24 lucian left
GeJ clock? 21:24
purl GeJ: LAX: Wed 2:24pm PDT / CHI: Wed 4:24pm CDT / NYC: Wed 5:24pm EDT / LON: Wed 10:24pm BST / BER: Wed 11:24pm CEST / IND: Thu 2:54am IST / TOK: Thu 6:24am JST / SYD: Thu 7:24am EST /
GeJ s/evening/afternoon/
cotto_work You can't be accurate for everyone. 21:26
21:26 Paul_the_Greek joined
Paul_the_Greek Good everning, one and all. 21:27
whiteknight hello Paul_the_Greek 21:28
Paul_the_Greek Hey ho whiteknight
whiteknight did you just call me a ho?!?!
:)
GeJ cotto_work: true. I usually go with 'morning' but there's always someone who needs to single himself with a "not here it's not." 21:29
.o0(I think there's a "out" missing somewhere)
cotto_work I like to go with "Good morning, and if you complain that it's not morning go do some Parrot hacking instead you slacker." 21:30
GeJ It's way too morning-y to write proper english anyway.
Paul_the_Greek You're a coding whore, whiteknight.
NotFound GeJ: Let's write C instead
Paul_the_Greek Oh wait, if you don't get paid then I guess the word doesn't apply. 21:31
21:31 robin-gvx left
GeJ NotFound: I'll pass thank you. :) 21:32
21:33 aloha left
Paul_the_Greek whiteknight: If we break apart the mark and sweep, doesn't the marking become obsolete if anything happens in between? 21:33
atrodo If it's marked as garbage, as in no way to get to it, I would guess that when the sweep runs it's still garbage since no one can get to it. But what do I know 21:34
21:35 bacek left
Paul_the_Greek I thought the mark phase marks the good stuff, not the garbage. 21:36
21:36 davidfetter joined 21:37 mikehh_ is now known as mikehh
mikehh opbots, names 21:37
21:37 aloha joined 21:44 lucian_ left 21:47 bacek joined 22:22 nwellnhof left 22:26 pyrimidine left 22:33 kid51 joined 22:34 ruoso joined 22:36 brianwisti left
dalek nxed: r604 | NotFound++ | trunk/winxedst1.winxed:
add compile_from_file method to WinxedHLL
 22:39
nxed: r605 | NotFound++ | trunk/winxed_installed.winxed:
initial version of driver for installed winxed - only for testing, not installed
kid51 seen Gerd 22:40
purl Gerd was last seen on #parrot 71 days, 10 hours, 17 minutes and 33 seconds ago, saying: I will start to copy the tar file to the ftp server. [Jun 15 12:22:48 2010]
aloha Sorry, I haven't seen Gerd.
cotto_work aloha: status 22:43
purl Since Thu Aug 5 20:55:17 2010, there have been 3258 modifications and 1927 questions. I have been awake for 20 days, 1 hour, 47 minutes, 58 seconds this session, and currently reference 855425 factoids. Addressing is in optional mode.
cotto_work botkick
22:48 davidfetter left
nopaste "kid51" at 192.168.1.3 pasted "[codingstd] tools/dev/mk_language_shell.pl.in causes failure in filenames.t." (11 lines) at nopaste.snit.ch/23119 22:49
"kid51" at 192.168.1.3 pasted "[codingstd] tools/dev/mk_language_shell.pl.in causes failure in filenames.t." (17 lines) at nopaste.snit.ch/23120 22:50
kid51 2nd paste is better
Paul_the_Greek seen chromatic 22:56
purl chromatic was last seen on #parrot 5 days, 3 hours, 4 minutes and 47 seconds ago, saying: Give me a few minutes and I can look at it. [Aug 20 19:51:22 2010]
aloha Sorry, I haven't seen chromatic.
Paul_the_Greek What the heck is aloha?
kid51 aloha?
It's a bot who doesn't respond to her own name ;-) 22:57
Paul_the_Greek And purl doesn't know either.
cotto_work it's bacek's bot
its purpose appears to make tab completion of greetings easier
Paul_the_Greek So aloha is what allows me to type part of a username and then hit tab? 22:58
cotto_work No. That's your irc client. aloha makes it easier to type aloha 22:59
kid51 doubts that
cotto_work s/aloha$/"aloha"/
22:59 davidfetter joined
Paul_the_Greek I was going to be scared if a bot could see my keystrokes, so I'm glad it can't. 22:59
Is chromatic away? 23:00
cotto_work He posted to parrot-dev earlier today
seen chromatic
purl chromatic was last seen on #parrot 5 days, 3 hours, 9 minutes and 37 seconds ago, saying: Give me a few minutes and I can look at it. [Aug 20 19:51:22 2010]
aloha Sorry, I haven't seen chromatic.
kid51 He's not here now ... but then again aloha is very new and doesn't know most of us 23:01
cotto_work seen cotto
purl cotto was last seen on #parrot 17 hours, 59 minutes and 17 seconds ago, saying: too late?
aloha Sorry, I haven't seen cotto.
cotto_work seen cotto_work
purl cotto_work was last seen on #parrot 0 seconds ago, saying: seen cotto_work
tcurtis seen cotto
purl cotto was last seen on #parrot 17 hours, 59 minutes and 25 seconds ago, saying: too late?
aloha cotto_work was last seen in #parrot 0 seconds ago saying "seen cotto_work".
Sorry, I haven't seen cotto.
kid51 seen kid51
purl kid51 was last seen on #parrot 0 seconds ago, saying: seen kid51
aloha kid51 was last seen in #parrot 0 seconds ago saying "seen kid51".
kid51 seen tcurtis
purl tcurtis was last seen on #parrot 12 seconds ago, saying: seen cotto
aloha tcurtis was last seen in #parrot 12 seconds ago saying "seen cotto".
Paul_the_Greek msg chromatic I have some ideas for speeding up allocation/free of PMC attributes. Is my best bet to wait until you merge gc_massacre?
purl Message for chromatic stored.
cotto_work hopes they don't create a feedback loop
kid51 So,, aloha only knows people seen since her activation
Paul_the_Greek: Why don't you post those ideas to list? 23:02
dalek rrot: r48658 | jkeenan++ | trunk/t/native_pbc (3 files):
(Once again ...) Run tools/dev/mk_packfile_pbc to update t/native_pbc files for Darwin/PPC.
cotto_work Paul_the_Greek: gc_massacre has been on the list to merge for at least a month.
mikehh kid51: ping 23:03
kid51 mikehh: pong
cotto_work I thought there was some agreement about getting it merged, but it remains unmerged.
Paul_the_Greek You read my mind. I was going to ask if there is a proposed merge date. 23:04
Sounds like I should wait for the merge.
cotto_work Applying pressure won't hurt.
NotFound I don't understamd the problem with the native_pbc files. They are no longer used in any test not skipped.
mikehh kid51: I am getting a codetest failure - t/codingstd/filenames.t - Multi-dot filename found in 1 files: tools/dev/mk_language_shell.pl.in 23:05
kid51 mikehh: nopaste.snit.ch/23120
mikehh kid51: never seen that before - don't even know what it means
kid51 I've seen it very infrequently. 23:06
I saw it just before leaving $job but couldn't fix it from there.
Paul_the_Greek I swear I saw a pile of double-dot files.
kid51 I emailed gerd, who did that commit.
Paul_the_Greek: Not in the repository, at any rate. 23:07
dalek TT #1750 closed by jkeenan++: t/pmc/packfile*.t: Failures in multiple files on Darwin/PPC; not ...
TT #1750: trac.parrot.org/parrot/ticket/1750
Paul_the_Greek Hmm...
kid51 mikehh: If you want to change its name's ending from '.pl.in' to '.in', that would probably work (with suitable adjustments in files referring to that file) 23:08
moderator Parrot 2.7.0 "Australian King" Released! | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | close 20 tickets (14 to go), merge outstanding branches 23:09
cotto_work Why do have that test? 23:09
Paul_the_Greek Ah, I remember. They were generated C files.
mikehh kid51: if gerd doesn't sort it out, I will
Paul_the_Greek Because multiple-dot file names are annoying? 23:10
At least we don't have multiple files that differ only by case.
sorear Paul_the_Greek: multi-dot files are forbidden by many filesystems
thankfully, the least dead such filesystem is FAT16 23:11
but codetest (and ISO C) have a long memory
Paul_the_Greek Good. Then we won't be tempted. :D
And we can run on VMS, too! 23:12
cotto_work Sure. Just submit a patch. 23:13
23:13 kid51 is now known as kid51_at_dinner
Paul_the_Greek Just a small patch or two and it should run on VMS. 23:15
mikehh anyone got access to VMS? 23:16
I mean we really, really should try to make parrot work everywhere or how can we take over the world.... 23:17
Paul_the_Greek I could probably find someone with a system at home.
I spent three years as a VMS developer at DEC.
szbalint I got VMS access @ uni. I've once written a write-once script for VMS. The experience still haunts me. 23:18
Paul_the_Greek If we want to take over the world we have to make it run on the ARM and other such little guys.
You wrote a DCL program? 23:19
dalek rrot: r48659 | gerd++ | trunk (2 files):
changed filename so that it pass the coding standards test "t/codingstd/filenames.t" (not more than one dot "."); Thanks to Jim for the advice.
mikehh yeah - I once did some work on MCP at Burroughs 23:20
szbalint Paul_the_Greek: yes.
Paul_the_Greek Then you might find this helpful: www.amazon.com/Writing-Real-Program...amp;sr=8-1
You may be amused by the name of one of the authors. 23:21
cotto_work I know that name from somewhere.
szbalint yeah, sounds familiar :) 23:22
cotto_work it's an easy one to remember
once you sound it out a couple times
Paul_the_Greek That's how I ended up in the book production business for $work.
Some people remember it as Soluopotsongana backward.
szbalint there are advantages and disadvantages for having a UUID as a surname :)
trust me, I know. 23:23
cotto_work pricey book 23:24
Paul_the_Greek If only I had some more copies, I could make some money on the side.
mikehh Paul_the_Greek: love the availability - 5 used from $129.95 23:25
Paul_the_Greek I do have the copyright back from Digital Press, so I could print it and sell it agaiin.
129. Crap. I never sold my extra copies for more than about $60.
$270? Hell, I should start selling advice on the Net. 23:26
I wrote the VMS code management and user installation facilities in DCL, so I became the world's expert at DCL. 23:27
For what it's worth.
Both have since been replaced, of course.
mikehh Paul_the_Greek: yes - but for the better? 23:29
Paul_the_Greek I suspect so. At least they are faster. 23:30
DCL had some pretty ancient interpreter guts, but no one would fund a rewrite.
It would have to be crock-for-crock compatible, of course. 23:31
mikehh if we had some new alpha chips running VMS or something, but bah 23:32
Paul_the_Greek You mean you don't want to install a VAX/11-780 in your house? 23:33
They have one in on display in the Brown CS department. I wonder if it runs? 23:34
mikehh I see that OpenVMS r8.4 was released in June 2010 23:39
that Open in OpenVMS has nothing to do with open source of course 23:40
Paul_the_Greek Yeah, I'm not sure what's open about it. They changed the name after I left. Some marketing pukes got at it. 23:41
Wait, a new version in June? Interesting. 23:42

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!