Parrot 4.1.0 "Black-headed Parrot" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 28 February 2012.
whiteknight hello dukeleto 00:23
we should find out tomorrow whether we are in GSOC
01:28 kid51 left, kid51 joined 03:33 benabik joined
benabik ~~ 03:34
msg whiteknight "shotgun cleanup" is such a wonderful phrase. I still like "refactoring with fire" 03:40
aloha OK. I'll deliver the message.
benabik msg whiteknight The Parrot_pfw_ API sounds sane to me. 03:41
aloha OK. I'll deliver the message.
04:36 woosley joined 06:17 awwaiid joined 07:52 mj41 joined 08:10 johbar_work joined
dalek kudo/nom: 8c7747b | moritz++ | docs/ChangeLog:
[ChangeLog]: note that scoped multis now die
08:33
08:47 bacek joined
bacek ~~ 08:50
08:52 woosley left 09:18 lucian joined 09:33 preflex_ joined 12:04 not_gerd joined
not_gerd hello, #parrot 12:05
bacek: what do you think about using a single, shared call context per thread for passing information down the call graph instead of re-using the parent one? 12:06
Coke is glad to see some technical discussion on list. 12:07
even if it seems to not be happening in the same room yet.
12:19 mj41 joined 13:04 plobsing_ joined
dalek p: 07ce402 | (Gerhard R)++ | src/6model/base64. (2 files):
Make base64.[ch] compile with new error levels
13:14
p: ffd778a | moritz++ | src/6model/base64. (2 files):
Merge pull request #31 from gerdr/gerdr/cleanup

Make base64.[ch] compile with new error levels
allison not_gerd: Parrot's CallContexts are just the same as old-fashioned call stack frames 13:23
not_gerd: there is some data that you need to store for each call that you make
not_gerd: it's a small amount of data, but it's also unique to each call, so every call needs one 13:24
ttbot Parrot d296f64f i386-linux-thread-multi make error tt.taptinder.org/cmdinfo/72826 13:40
Parrot d923903b i386-freebsd-64int make error tt.taptinder.org/cmdinfo/72837 13:43
13:45 preflex_ joined
not_gerd allison: what I'm getting at is that the caller only needs to pass the information necessary to create a CallContext to the callee, but that can be done without actually allocating a new one (more) 13:52
allison: the long term goal could be to get rid of call contextx altogether and instead us reserved registers, ie the caller puts the necessary information into a thread-local location, and the callee moves it to a set of reserved registers 13:54
allison There was a time that Parrot passed arguments in reserved registers
but, it was problematic, for several reasons 13:55
partly, it burned through registers
partly, it triggered register spilling (which is expensive) with far greater frequency 13:56
partly, it meant we had to allocate registers sooner
partly, it meant that even simply asynchronous calls (not even full concurrency) were unsafe 13:57
partly, it made continuations very expensive 13:58
because taking a continuation meant snapshotting all active registers so they could be restored later
(and restoring them later had a nasty habit of trashing active registers that were still needed) 13:59
dalek p: fdec156 | moritz++ | tools/build/PARROT_REVISION:
bump parrot revision

just to get some better testing of newer birds, no specific reason
14:05
kudo/nom: 40b20ca | moritz++ | tools/build/NQP_REVISION:
bump NQP revision to get NQP build fixes by not_gerd++
14:07
14:20 fperrad joined
dalek kudo/nom: 16bf0f4 | moritz++ | src/ops/perl6.ops:
unbreak the build.
14:27
14:33 not_gerd left 14:43 PacoAir joined 15:18 Psyche^ joined 15:28 fperrad joined 15:31 benabik_ joined 15:46 Psyche^ joined 15:59 davidfetter joined 16:46 contingencyplan joined
Coke sounds like the recent CFLAGS changes have broken dyncall. 17:29
moritz is about to open an issue 17:32
Coke moritz++ 17:33
I'm pretty sure those changes are just to make porting easier, so if they break something, reverting them should be OK.
but if it's not a rush, we should ping alester.
17:45 fperrad_ joined 17:54 whiteknight joined
whiteknight Parrot is accepted into GSOC 2012 17:55
tadzik \\o/
whiteknight hello tadzik
I just got the email.
tadzik 'ello whiteknight
I can see you're in a different timezone. Here it says that orgs will be announced in an hour and 4 minutes :) 17:56
whiteknight They are probably reviewing and accepting them on a rolling basis 17:57
tadzik no, it must be timezones
aloha (parrot/parrot) Issues opened : 735 (adding -Werror=strict-prototypes to CFLAGS breaks nqp build on some platforms) by moritz : github.com/parrot/parrot/issues/735 17:58
18:03 not_gerd joined
aloha (parrot/parrot) Issues opened : 736 ("perl Configure.pl --gen-parrot" fails) by parthm : github.com/parrot/parrot/issues/736 18:08
18:14 parth joined
parth Hello. I am trying to build a language shell created with create_language.pl. 18:21
whiteknight parth: yes, I just saw the ticket 18:22
what kind of a language are you working on?
moritz fwiw it seems that cloning parrot fails
parth hi whiteknight: i am exploring the possibility of creating an experimental smalltalk variant. 18:23
whiteknight parth: oh, awesome
moritz it might help to use the git:// URL instead of the https URL
parth i tried using --parrot-config option. that seems to fail too pastebin.com/K4KUBvKV is this expected?
whiteknight oh, okay, do you have Parrot installed? 18:24
parth i haven't installed it. i just built the downloaded tarball. 18:25
whiteknight Can you try to "make install" it?
I think create_language.pl is looking for the installed binary, not the local one
parth mortiz: https seems to work ok from the command prompt using git pastebin.com/HzYb3SPG
whiteknight: yup. i was hoping it would pick up the one built. is this by design? 18:26
whiteknight parth: I don't think so. Looks like a bug to me
parth: We can fix the script, but if you install parrot you shouldn't have to wait 18:27
parth whiteknight: i will go ahead and file it. yes. will try with install.
whiteknight: thanks for verifying this. 18:28
whiteknight thanks. Good luck! Let us know if you have any questions
Coke whiteknight: i am tentatively interested in being a mentor. 18:31
whiteknight Coke++ 18:32
Coke: And I hope we can pair you up with somebody awesome
aloha (parrot/parrot) Issues opened : 737 ("perl Configure.pl --parrot-config <path>" looks for parrot-nqp in /usr/local/bin) by parthm : github.com/parrot/parrot/issues/737 18:33
Coke either way I hope to do a better job personally than I did last year. 18:35
cotto ~~
Coke wait, high school students can apply? 18:36
I thought that was GCI.
I thought you had to be college aged for GSOC.
18:37 Psyche^ joined
whiteknight Coke: I'll have to go back and look at the requirements, I thought that highschool kids could apply 18:38
is github crapping out for anybody else? I'm having trouble loading pages 18:39
moritz yes 18:44
whiteknight okay, good. The network here at work is terrible, so if it was just me I would have to go yell at somebody 18:45
19:06 Psyche^ joined 19:08 johbar_work joined
whiteknight Coke: must be at least 18. Otherwise can be in (or recently graduated from) highschool or college 19:10
Coke TPF didn't make it into GSOC this year.
that's... odd.
whiteknight lolwat? That stinks
cotto That's most surprising.
whiteknight did they apply?
Coke aye. 19:11
whiteknight If Rakudo has any project ideas, I suspect we can offer those under our name
Coke we'll find out after the irc meeting next week.
whiteknight: aye.
(the more projects the better.) 19:12
whiteknight yes, and the bigger the supply of mentors, the better
last year we had just a few too few mentors, and at least one project couldn't be taken because of that
not that all available mentors are interested in all available proposals 19:13
moritz I'm happy to mentor any Perl 6 stuff under the umbrella of PaFo, though I'll understand it if you want to reserve your slots for stuff that's more directly related to parrot 19:15
dalek p: 99f34ab | (Gerhard R)++ | 3rdparty/dyncall/GNUmakefile:
Temporary fix for dyncall build failure

See github.com/parrot/parrot/issues/735
19:26
whiteknight moritz++ 19:32
...and so it begins. The shitstorm of open-source developers who criticize everything google does during GSOC because some people don't know how to graciously accept charity
sorear whiteknight: are you saying that in a way that includes the reactions on #parrot and #perl6? 19:34
whiteknight sorear: no, on the gsoc mailing list
sorear ok. 19:35
Coke bunchawhiners, summathem.
whiteknight Every day during GSOC and GCI those lists get bullshit complainy emails. "Your form isn't intuitive", "the wording on your website is weird", "I wish the data would fill in automatically", "Can you make an aesthetic change?"
Coke granted, email makes them sound worse, but they should try to compensate for that. 19:36
sorear I find that generally the people who complain most annoyingly are people I don't care to listen to in the first place
whiteknight it's bugging me that TPF didn't get in
sorear (mlehmann being the main exception) 19:37
whiteknight unless their application was woefully underwhelming or something
"We're perl. Maybe you've heard of us. We're kind of a big deal"
sorear (probably some of you will disagree with me about mlehmann being worth listening to)
whiteknight sorear: I'm not sure I know who that is
sorear JSON::XS, AnyEvent, Coro, rxvt-unicode, libevent 19:38
whiteknight oh, okay 19:39
Coke I only know mlehmann from his responses on [perl #108470] 19:42
awwaiid whiteknight, also libev (foundation of node.js), unicode-rxvt 19:45
sorear I meant libev, yes 19:47
benabik o/ 19:55
whiteknight benabik! 20:06
benabik \\o/ whiteknight!
Unless I broke something, in which case...
not_benabik I wasn't here.
whiteknight benabik: start getting your project idea together. We're going full-throttle this summer
sleep and meal breaks are for the weak 20:07
benabik Well, I've got a plan already in place for not having any sleep.
Coke ah, for me that was this week!
20:10 mj41 joined, donaldh joined
benabik I'm going to have to be careful what I promise this summer. My thesis is probably going to run a little long and then who knows what'll happen with the kiddo. 20:11
benabik would rather aim a little low and overshoot than the other way around. 20:12
whiteknight okay, we'll recalibrate expectations later. Today is for celebrating! 20:13
benabik \\o/
20:16 preflex_ joined
whiteknight anyway, time to go. I'll be back online tonight 20:27
ttbot Parrot 8032bfa2 darwin-thread-multi-2level make error tt.taptinder.org/cmdinfo/73062 20:28
mj41 Hi. Ignore the last ttbot message. I'm trying to setup MacOS client. 20:42
cotto phew 20:43
20:48 not_gerd left
mj41 tt client machine macos is back tt.taptinder.org/buildstatus/parrot/master 21:09
dukeleto ~~ 21:18
blarg 21:21
benabik dukeleto: ?
dukeleto benabik: my irc proxies were fighting, better now :) 21:22
benabik dukeleto: You? IRC problems? Never. ;-) 21:24
davidfetter wonders what the other duke leto would have done about duelling proxies 21:30
dukeleto Yeah, i need to find something better than dircproxy, or change the default settings. 21:32
davidfetter for some reason, dircproxy reminds me of "team america: world police" 21:33
benabik One of these days I'm going to get up znc on my server. 21:39
Mostly because it supports push notifications to my phone.
21:46 lucian joined
cotto dukeleto: how'd dircproxy deficient? 22:35
22:38 luben_at_work joined
dukeleto cotto: the default behavior to is autoreconnect, which is useful, unless you leave an irc window open somewhere and then open another 22:41
cotto: then the irc clients fight to the death forever disconnecting each other
cotto: this happens when i leave my irc client open on my desktop at home and then try to get on irc on my laptop
cotto yeah. lta 22:42
22:43 whiteknight joined
whiteknight good evening, #parrot 22:59
benabik evening!
22:59 johbar_work joined
whiteknight hello benabik 23:08
dukeleto whiteknight: congrats on getting us into gsoc! 23:13
whiteknight++
whiteknight Thanks!
davidfetter who's all using parrot in production? 23:14
dukeleto davidfetter: what does "in production" mean" ;) 23:18
davidfetter: i use it to produce bytecode every day!
davidfetter dukeleto, i'm pretty sure you know what i mean 23:19
cotto I'm not aware of any production users.
benabik You can be one of the first! How exciting! 23:20
dukeleto davidfetter: yeah, maybe you can ask VMware if they want to use Parrot in production... 23:31
davidfetter: do you have suggestions about who are first prod users could be?
davidfetter: is it PL/Parrot, maybe?
davidfetter: or something else?
davidfetter dukeleto, i need an elevator pitch, and i don't have one 23:32
partly because i don't know enough about parrot
dukeleto davidfetter: that is the hard part. Parrot is an enabling technology, but I am not sure what kind of elevator pitch Parrot core would have 23:35
23:35 brambles joined
dukeleto davidfetter: an app built on parrot could have a good elevator pitch, though 23:35
davidfetter: i still have high hopes for parrot in the mobile and embedded world 23:36
davidfetter: "We make bytecode so you don't have to"
davidfetter i've got a perl6 pitch: grammars
benabik We'd have to be a bit more svelte for mobile.
dukeleto davidfetter: "We wrote 5 garbage collectors so you don't have to"
benabik: yeah, i have a plan for parrot "flavors"
benabik: parrot is already more svelte than perl5 23:37
benabik When trying to explain Parrot during my interviews, I explained it as bootstrapping. Doing all the hard parts for new languages: portability, libraries, GC.
davidfetter benabik, my mobile phone kicks the butt of the server systems i had running 10 years back. i think the "svelte" part's overrated
dukeleto davidfetter: a good point, indeed 23:41
davidfetter "it's not fatter than perl5" is damning with faint praise, and it's definitely not an elevator pitch 23:42
dukeleto davidfetter: indeed, again
davidfetter: "Parrot is a toolbox for language and DSL development" 23:43
whiteknight: feel free to put me down as backup admin/mentor for gsoc 23:44
whiteknight: if you haven't already :)