Parrot 4.3.0 "In Which..." | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 18 April 2012.
whiteknight dukeleto: ping 00:02
mayank: ping
dukeleto whiteknight: pong 00:03
whiteknight: i just saw the duplicate email
whiteknight yeah, I'm replying now
my general thought is: it would be good/great for us. If it's better than that for them, that's fine by me 00:04
or, whatever his preference is
dalek rrot/remove_sub_flags: 2ab8cb0 | Whiteknight++ | compilers/imcc/main.c:
Don't execute immediate subs twice. Only execute them the instant that they are compiled, not again later.
 00:11
00:12 aloha joined
whiteknight We've got 99 branches on github, apparently 00:17
dukeleto whiteknight: one more to make it a century 00:19
whiteknight++ # gsoc duplicate email
whiteknight I think I want to merge pcc_ops
I've got 99 branches, but soon pcc_ops won't be one 00:46
hmm. I think I've already merged a replacement for this 00:56
ha, anybody mind if I merge my branch to delete the Eval PMC? 01:05
actually, nevermind. It doesn't pass tests 01:08
dalek rrot/eval_pmc: 8749bb3 | Whiteknight++ | / (13 files):
Merge branch 'master' into eval_pmc
 01:20
rrot/eval_pmc: 68f0666 | Whiteknight++ | tools/dev/nci_thunk_gen.pir:
One fix to nci_thunk_gen.pir to get the build moving again. PackfileView doesn't have an invoke vtable like Eval did
rrot/eval_pmc: 53273f8 | Whiteknight++ | / (3 files):
Fix data_json and several tests so that the test run executes without hanging (some failures)
rrot/eval_pmc: 24c35b6 | Whiteknight++ | / (317 files):
Merge branch 'master' into eval_pmc
03:28 benabik joined 03:36 nbezzala joined 03:46 schmooster joined 03:47 davidfetter joined
_mayank whiteknight: pong 04:09
04:21 dalek joined, pmichaud_ joined, Util_ joined
dalek rrot: a71d1fd | petdance++ | src/ (3 files):
consting args
 04:42
07:10 nbezzala_ joined 07:23 lucian__ joined 07:40 fperrad joined 08:18 lucian joined 08:43 fperrad joined
dalek rrot: 613e1f9 | fperrad++ | t/pmc/timer.t:
[t] fix shebang
 09:15
10:07 brrt joined 10:08 lucian joined
lucian got disconnected yesterday, on the train 10:19
lucian waves hello
i've noticed the Jaesop proposal is marked as duplicate 10:25
brrt hello 10:29
_mayank lucian: Hi, I submitted the Jaesop proposal . What is meant by "marked as duplicate"? 10:41
lucian _mayank: i don't really know
it could be either to signal mistakes, like the same person mistakenly submitting the same proposal twice 10:42
or it could mean there's another very similar proposal
neither appears to be the case here
_mayank Actually, I did submit another proposal to another organisation.
lucian hmm. i don't think that could be it 10:43
_mayank Is it related to that? I got a mail from the mentor from the other organisation, asking for my preference out of the 2 projects.
It was cc'ed to Andrew also
lucian it wouldn't be the same proposal, obviously
_mayank: i see. have you expressed a preference?
_mayank Not yet, I need to. I wanted to have a word with Andrew before I do. Waiting for him. 10:45
brrt if history is an indication, he'll be here in an hour or so 10:46
lucian _mayank: btw, i also expressed an interest to mentor for this project 10:49
12:01 benabik joined 12:07 brrt left 12:13 bluescreen joined 12:16 whiteknight joined
whiteknight good morning, #parrot 12:25
_mayank good morning whiteknight 12:26
benabik o/ #parrot, whiteknight 12:27
_mayank whiteknight: I saw your ping from yesterday, was it about the conflicts? 12:29
whiteknight _mayank: yes, the conflicts
hello benabik
13:12 PacoAir joined 13:30 bluescreen joined 13:34 rich joined, rich left 13:50 hercynium joined 14:20 contingencyplan joined
dalek p: e5e2dea | coke++ | VERSION:
bump VERSION to 2012.04
 14:25
kudo/nom: 0c5d318 | coke++ | tools/build/NQP_REVISION:
update to latest NQP
 14:29
kudo/nom: b27b216 | coke++ | src/core/tai-utc.pm:
ooh, new leap second
 14:32
14:47 dmalcolm joined 14:58 Justin joined
Justin good day everyone 14:59
benabik o/ Justin 15:01
whiteknight hello Justin 15:09
dalek CT: e7aed76 | benabik++ | / (3 files):
Notes on REPLs

Implementing a REPL framework has some tricky parts. Document them somewhere for further reference and consideration.		 15:11
benabik is currently writing PACT/docs/optimization.mkd
whiteknight nice 15:12
benabik++
benabik whiteknight: Did you look at the axe job I did to opcode trees? I'm hoping my brief explanations are enough to carry the basic ideas. 15:13
whiteknight benabik: I saw the commit, but didn't read it too closely 15:14
I probably need to go back and actually pay attention this time 15:15
benabik Well, if you get a chance. :-D
whiteknight if I get a chance? My free time is boundless
benabik I suppose it's not too big a deal at the moment, given that I'm the only person working on it, but I want to have bus > 1
whiteknight: Sarcasm doesn't travel well over the internet...
whiteknight :) 15:17
I will look at it. Don't worry about that.
PACT is very near and dear to my heart
Justin lol 15:24
Coke hey, everybody - if you're interested in mentoring a project for GSOC 2012 for parrot, it's not too late to sign up. 15:35
help avoid burnout, sign up today!
whiteknight We actually have several mentors signed up already, but not all of them are interested in mentoring all projects 15:36
matching mentors up to the right projects is the hard part 15:37
15:37 jashwanth joined 15:39 alvis joined
Justin aw :( no one wants to mentor me? 15:46
lol 15:47
PerlJam Justin: What's your project proposal? 15:48
15:48 davidfetter joined
Justin The security sandbox :) 15:48
15:51 dngor joined 16:35 Justin joined 17:01 GodFather joined 17:03 PacoAir joined
cotto ~~ 17:25
whiteknight hello cotto 17:29
cotto hi whiteknight 17:32
17:33 awwaiid joined 17:54 lucian__ joined
moritz hi everybody 19:43
how do I switch off lcoale-dependent behavior in parrot?
benabik moritz: Somehow I knew that was coming. Sadly, I don't know the answer.
moritz the code seems to be in src/platform/linux/encoding.c and not easily configurable 19:45
NotFound moritz: Configuring, or runtime? 19:46
moritz NotFound: runtime 19:47
NotFound moritz: try LANG=C
moritz NotFound: but then I don't get UTF-8 :/ 19:48
as Parrot_platform_encoding_ptr
maybe the correct thing would be to never access that at runtime
benabik If you want UTF8, you probably want to set LANG=UTF-8 instead of LANG=C. Or something like.
NotFound moritz: I thik so, relying in platform enconding is almost always wrong.
moritz it's not that easy
NotFound: problem is, the documentation doesn't say where it's used 19:49
relying on LANG is also wrong, because then subprocesses we spawn also has the modified LANG
and that's not something a VM should do to its host programs
maybe it's something as simple as manually setting the encoding of some file handle, but I just don't know 19:50
benabik You can get an object for stdin, I think... Perhaps you can set the encoding on it? 19:51
NotFound moritz: What is your current problem? 19:53
I've had problems with open "rp", in that case you just need to set the encoding in the handle. 19:54
benabik I think it's that our output functions translate to platform encoding. Which is probably vaguely sane.
moritz NotFound: the problem is that non-UTF-8 locales (or ones that are set up incorrectly) produce test failures like that: pastebin.com/FCj0dHU6 20:07
NotFound: which basically means that strings don't round-trip correctly from -e on the command line to STDOUT 20:08
benabik I suppose we could default to UTF8 instead of ASCII. :-/
(Although we might explicitly recognize C as ASCII.)
NotFound Defaulting to utf8 is not a general solution, that way you get failures with byte sequences not parseable as uft8 20:11
moritz well, if you default to ASCII, you get errors on bytes > 127 20:12
at least if you really mean ASCII
benabik Possible encoding errors in any case.
NotFound The assumption is that if you have a utf8 locale you pass utf8 strings to -e 20:13
And so on.
moritz yes, and I'd like to override that assumption 20:16
NotFound moritz: I don't understand that test. Does it expect that parrot must guess what encoding you want in the output? 20:24
moritz NotFound: no. Rakudo could pass it if rakudo simply always assumed UTF-8 20:29
so, how do I make rakudo always use UTF-8 instead of (sometimes?) what the locales say? 20:30
benabik You can set encoding on IO objects, right? So you should be able to get the stdio objects and set them to UTF-8. Less sure about dealing with ARGV.
NotFound benabik: command line arguments are not stdin 20:31
benabik NotFound: Yes, but I bet stdout is set to ASCII too.
NotFound: And that's why I said "less sure about argv".
Probably need to set both to get it to work right. 20:32
benabik could be wrong.
NotFound If you get a utf8 valid byte pair interpreted as latin-1 and then you recode to utf8, the result will be wrong. 20:33
benabik We may just want to add a --encoding parameter that overrides whatever we think the platform is. 20:43
NotFound I think that we should first document when and why we use the platform encoding. Otherwise adding more options can be just adding more confusion. 20:46
benabik True story. 20:47
21:27 Hunger joined 21:49 dip joined 22:09 whiteknight joined 22:43 kid51 joined
whiteknight good evening, #parrot 23:08
tadzik good night whiteknight 23:09
whiteknight hello tadzik, how are you? 23:11
tadzik pretty well. Survived the Killer Week on the university and now preparing for Perl 6 Hackathon in Oslo 23:12
it's 1:12 AM in here, and I'm leaving for the airport on 4:15 so I figured there's pretty much no point in going to bed :)
23:26 brambles joined 23:36 dmalcolm joined
dalek kudo/nom: 6fcae56 | coke++ | docs/announce/2012.04:
reluctantly add myself; remove duplicate masak
 23:44
kudo/nom: 96ab007 | coke++ | docs/release_guide.pod:
Add note about script requirement

  (match other note earlier in the doc)
kudo/nom: 73b1756 | coke++ | docs/ChangeLog:
update ChangeLog slightly
kudo/nom: 65ab8e3 | coke++ | docs/release_guide.pod:
nearly done planning...
kudo/nom: 12411d0 | coke++ | VERSION:
2012.04 bump VERSION
kudo/nom: 2e550a8 | coke++ | docs/release_guide.pod:
Update step number. Add note.
 23:52

Powered by Cro and the Raku® Programming Language.

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

Thank you!