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.
00:02 benabik joined
benabik ~~ 00:16
kid51 threads branch make fulltest PASS: smolder.parrot.org/app/projects/rep...ails/27691 00:23
dalek rrot/eval_pmc: b616087 | Whiteknight++ | / (8 files):
Fix Test::More to use the PIR compreg correctly. Fix most of coretest
00:33
kid51 whiteknight: ping 00:44
whiteknight pong
kid51 You may list me as backup mentor for the ModParrot GSOC project 00:45
If I need to click on any buttons at google-melange to make that official, just let me know. 00:46
whiteknight thanks. I'll write you down 00:54
kid51 I actually found a button to click there, which has added me to the Possible Mentors list for that project.
whiteknight google only has a concept of a single mentor per project. "backup mentor" is something Parrot invented 00:55
so I'm going to do something high-tech, like write your name down on a piece of paper 00:58
Actually, I spend so much time in front of the keyboard that my handwriting is probably too embarrassingly bad for that 00:59
kid51 Well, if you manage to do that, then the next time I see you I'll give you lessons in high-tech calendaring, i.e., a pocket yearly calendar of the kind I've been using since the 70s 01:00
And if you want to send me links to things that I should read as background to that project, please do. 01:01
I want to learn *something* via this.
whiteknight sure thing.
dalek rrot/eval_pmc: 843d395 | Whiteknight++ | compilers/pge/PGE/Exp.pir:
Fix p5rx.t, which was in an infinite loop of exception handlers and blocking the test suite from running.
01:11
01:43 brambles joined
dalek rrot/eval_pmc: d61cabb | Whiteknight++ | / (3 files):
Fix HLLCompiler to use PackfileView. This fixes several tests, though it does add in the weirdness that the .evalpmc() method doesn't do anything with Eval.
02:06
rrot/eval_pmc: d80d94a | Whiteknight++ | MANIFEST:
Remove a deleted test file from MANIFEST
04:51 jashwanth joined 05:10 alvis joined 05:31 particle1 joined 05:34 jjore joined 06:21 nbrown joined 07:19 fperrad joined 07:20 preflex_ joined
dalek kudo/nom: 082dd1a | masak++ | docs/release_guide.pod:
[docs/release_guide.pod] Frankfurt!

  \\o/
07:46
08:17 lucian joined
dalek p: 1e988f1 | jonathan++ | src/6model/serialization.c:
When chasing the outer chain serializing closures, we may encounter a NULL lexinfo pointer. Don't follow it.
09:12
09:38 fperrad joined
dalek kudo/nom: 8658d94 | pmichaud++ | src/core/Any.pm:
Implement hash autoviv. Array autoviv comes next.
10:05
kudo/nom: ff875ef | pmichaud++ | / (2 files):
Array autoviv, add a passing test file.
10:38
kudo/nom: bff4e5d | pmichaud++ | t/spectest.data:
Add another passing spectest file.
11:43
11:49 kid51 joined
dalek kudo/nom: e59619c | jonathan++ | src/Perl6/Metamodel/MROBasedMethodDispatch.pm:
Eliminate use of %h.update(...) in the meta-objects.
12:03
12:06 mdupont joined 12:12 alvis joined 12:30 bluescreen joined
dalek p: 77ee7b5 | jonathan++ | / (2 files):
Add OwnedHash, which will track hashes that need to trigger the SC write barrier if modified.
12:43
p: 6f8ec7c | jonathan++ | src/how/NQPClassHOW.pm:
Eliminate use of %hash.update(...); that code ain't a hot path with serialization, the method ain't in Perl 6, and (most pressingly) the Parrot implementation doesn't handle hash subclasses properly.
p: 362983f | jonathan++ | src/stage0/ (6 files):
Update the bootstrap.
p: 0adfeb8 | jonathan++ | src/ (7 files):
Some more .update elimination; bootstrap update also.
12:43 hercynium joined
p: db9d01c | jonathan++ | src/pmc/ownedhash.pmc:
Don't trigger SC write barrier if owner is null; also add mark().
p: 0a519ef | jonathan++ | src/6model/serialization. (2 files):
Have deserialization create OwnedHash and sets the owner.
p: ba5c237 | jonathan++ | src/6model/serialization.c:
Serializing an OwnedHash just does exactly the same as serializing a Hash.
p: 2be3071 | jonathan++ | src/ops/nqp.ops:
Add ops for temporarily enabling/disabling the SCWB.
p: 0f16cba | jonathan++ | src/6model/serialization.c:
Don't try and serialize file handles.
kudo/nom: c9905df | jonathan++ | tools/build/NQP_REVISION:
Bump NQP_REVISION.
12:44
kudo/nom: 7f62119 | jonathan++ | src/Perl6/ModuleLoader.pm:
Disable SC write barrier while loading CORE.setting.
12:46 whiteknight joined 12:55 PacoAir joined
whiteknight good morning, #parrot 13:12
nine Good morning, whiteknight 13:27
whiteknight hello nine 13:28
nine: I'm starting to line up some projects I want to work on when threads get merged
I really want to put together a streams implementation like what node.js has 13:29
nine You actually want to use this stuff? Getting nervous :)
whiteknight yes I want it!
I want to start working on an eventing system like POE, and I want to start putting together an async networking library 13:30
I really need to start focusing on those greenthreads for windows 13:31
nine Yeah, that'd be awesome
whiteknight I got a simple build environment set up on my winxp vm, which is better than nothing.
benabik Known issue? "Eliminate use of %hash.update(...); that code ain't a hot path with serialization, the method ain't in Perl 6, and (most pressingly) the Parrot implementation doesn't handle hash subclasses properly." 13:32
whiteknight I know that our subclasses of built-in types is flawed. I wonder exactly which flaw they find to be fatal in that case
13:32 PacoAir_ joined
whiteknight damn, we really need 6model 13:32
13:38 jashwanth joined
whiteknight If I can get that Eval PMC deprecation wrapped up this weekend, I think I can knock green threads on windows out next week 13:39
nine \\o/
whiteknight The remove_sub_flags stuff isn't high priority, since we probably wouldn't want to even consider a merge before 4.4, so I can move on to 6model after that 13:40
14:23 contingencyplan joined
dalek kudo/nom: 51bfca4 | jonathan++ | src/binder/types.c:
Fix regression due to some hashes now being OwnedHash. pmichaud++ for noticing.
14:28
dngor whiteknight: I'd rather see something like Reflex. Eventiness baked into the OO. 14:47
whiteknight dngor: I'm not familiar with Reflex. Link?
dngor github.com/rcaputo/reflex/tree/master/eg ... search.cpan.org/dist/Reflex/ 14:48
dalek kudo/nom: aa2eb5b | masak++ | src/core/Exception.pm:
[core/Exception] output nested error

Better than just "Error while creating error string"
15:12
kudo/nom: c40f488 | masak++ | src/core/Exception.pm:
[core/Exception] whitespace fix
15:21 brrt joined 15:45 dmalcolm joined 16:31 brrt1 joined
dalek rrot/threads: 7ef9990 | nine++ | / (2 files):
Don't do return foo() in void functions

Fixes compliation on stricter compilers like on Solaris
  Andy Dougherty++ for finding this and suggesting the patch
16:38
16:52 PacoAir joined
cotto ~~ 17:12
17:24 Justin joined
Justin good afternoon 17:24
whiteknight hello justin 17:29
brrt just out of curiosity, is there anyone who applied to mentor mod_parrot? 17:34
whiteknight brrt: we have mentors assigned for all accepted proposals 17:40
The googlers get really angry if we say who did and did not get accepted, so I'm really trying not to talk about specifics 17:41
brrt thats cool :-) 17:42
benabik Wow. The announcement is Monday? 17:43
brrt mentor deadline is today :-) 17:44
i was wondering if i could enthuse anyone to sign up 17:45
btw, how does google determine who can be a mentor for what project? 17:46
benabik I think that's up to the orgs. 17:47
brrt i know
but if i were to submit as a mentor for a project i know nothing about 17:48
what would happen?
benabik I think Google hopes that either the mentor or the org wouldn't allow it. :-/
brrt i'd hope sp
whiteknight brrt: Organizations accept and assign mentors 18:03
brrt: If you applied to be a mentor for Parrot, I would get an email, and I would respond "no, but thanks anyway" 18:04
brrt so, basically you manage the whole operation
brrt brb 18:05
benabik I think Carol would argue with that statement. :-D
whiteknight I'm running the Parrot organization, yes 18:07
Carol is the supreme leader
18:17 Justin joined
Coke also, you cannot be a student and a mentor in the same year, even cross-org. 18:22
18:22 nbrown joined
whiteknight that comes up every year. I don't know how Google enforces that 18:40
cotto if you were dead-set on being both a student and a mentor and knew orgs where you could do it, you could figure out a way. I don't know why you'd want to though. 18:44
Coke whiteknight: because you cannot be a student and a mentor in melange for the same year at the same time. 18:54
(you'd have to setup 2 logins, I suppose.)
cotto yeah 19:01
19:07 lucian__ joined 20:20 _mayank left 21:00 alester joined
dalek rrot: b5026c2 | petdance++ | t/codingstd/case_sensitive_files.t:
Add a test to check for conflicting file names in the repo
21:02
rrot: f8df6a7 | petdance++ | lib/Parrot/Distribution.pm:
Add get_all_files() method
21:04
alester this doesn't solve an existing problem, but prevents one in the future. We have one like this at work when a guy up in Marketing added a Foo.html file and there was already a foo.html. 21:06
dalek kudo/nom: 45ef4e2 | pmichaud++ | src/core/ (4 files):
Refactor autoviv methods to better handle type objects, use multidispatch.
21:42
22:26 plobsing joined 23:07 moritz joined