Parrot 3.10.0 "Apple Pi" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 16 November 2011.
whiteknight the fact that we have green threads now for dispatching said events in a sane way is a great thing 00:02
Tene whiteknight: I'd go with src/dynpmc 00:11
whiteknight thanks Tene. That's what I was thinking
I probably won't get to it today, but when I am ready to play that's where I'll do it 00:12
bacek_at_work minusm.com 00:18
looks interesting
whiteknight wow, does look interesting 00:20
cotto whiteknight: I'd start by writing some non-functional code to see how well the interface would fit into a dynpmc, but a dynpmc doesn't sound like a bad place for something like that. 00:22
whiteknight cotto: I thnk it's going to be a dynpmc, the question was whether to do it in the parrot repo or elsewhere 00:23
cotto whiteknight: ok. We can call it experimental, so I'm fine with it living in the repo. 00:24
bacek_at_work: nice. Too bad that I get a "500" when trying to sign up.
bacek_at_work cotto, hmm. I successfully signed up...
cotto bacek_at_work: it seems like the kind of thing that'd get overwhelmed pretty quickly by excited nerds. 00:25
bacek_at_work they were referenced in HN :)
cotto yes 00:26
bacek_at_work cotto, can you quickly test kill_constants branch? 00:33
00:34 bluescreen joined
whiteknight bacek_at_work: I will give it a shot 00:35
bacek_at_work whiteknight, thanks
whiteknight bacek_at_work: I'm so happy to see you, I'll do anything you need :) 00:36
bacek_at_work whiteknight, I was quite busy with RL. And still is. But I can squeeze few hours per week 00:38
whiteknight bacek_at_work: we're always happy when you have time, but don't squeeze! We don't want anybody to be uncomfortable
bacek_at_work :)
dukeleto bacek_at_work: welcome back, meatbag 00:41
bacek_at_work dukeleto, no, you!
I'm magical drinking robot!
dukeleto bacek_at_work: what timezone are you in these days? 00:46
dalek p: ec2f2a8 | diakopter++ | src/QRegex/NFA.nqp:
skip non-declarative code when building NFAs
 00:52
dukeleto seems like a lot of activity in nqp-land lately 00:56
bacek_at_work dukeleto, still Sydney 00:58
whiteknight bacek_at_work: kill_constants passes all tests for me 00:59
I'm looking forward to this branch
bacek_at_work whiteknight, arch/mem?
whiteknight bacek: linux x86_64
4GB mem, un-nice'd 01:00
bacek_at_work dukeleto, looks like jnthn and pmichaud are finally implementing LTM.
whiteknight, thanks
dalek sella/gh-pages: 1e159fd | Whiteknight++ | libraries/commandline.md:
Add in some stub docs for CommandLine.
 01:16
dukeleto Now With More 10% CommandLine! 01:19
02:07 kid51 joined
cotto ~~ 02:14
alvis Evening #parrot. Would someone take a look at ayardley.github.com/ and my repo at git@github.com:ayardley/ayardley.github.com.git 02:18
Is this what we want? And, do we want the docs/*.pod in the repo as well? (But, of course, in git@github.com:parrot/parrot.github.com.git) 02:19
dalek p: 9dd7eab | diakopter++ | src/QRegex/ (2 files):
make the NFA work with () captures
 02:20
p: f5ad380 | diakopter++ | src/QRegex/NFA.nqp:
make NFA work with rxtype subcapture (= captures)
 02:31
sella: b0b6644 | Whiteknight++ | rosella/data/templates/test_ns.winxed.template:
Make sure to call load functions in tests from the winxed/namespace template
 02:36
whiteknight alvis: I'm heading to bed now, But I'll look at it tomorrow 02:37
alvis++
alvis whiteknight: Ok, thanks.
whiteknight: Good night
cotto dukeleto++ 02:59
03:24 benabik joined 03:31 rurban joined 03:50 bluescreen joined 04:59 jsut_ joined 05:06 schmooster joined
dukeleto cotto: what did I do now? 06:45
alvis: looks good and yes.
07:21 rfw joined
cotto dukeleto, PaFo stuff 07:25
08:28 lucian joined 08:53 mj41 joined 09:08 lucian joined 09:40 nine joined 09:41 masak joined
masak was there no release announcement email this month? :/ 09:42
bacek masak, yeah. Looks like whiteknight forgot to send it... 09:50
msg whiteknight Looks like you forgot to send release announcement. 09:59
aloha OK. I'll deliver the message.
dalek p: 5523be6 | jnthn++ | src/QRegex/NFA.nqp:
Make sure we terminate LTM on imperative pastnode cases.
 10:42
kudo/nom: 81c23d2 | jnthn++ | src/Perl6/Actions.pm:
Mark :my declarations as declarative for the purposes of LTM.
 11:15
kudo/nom: 2f44b37 | jnthn++ | tools/build/NQP_REVISION:
Bump to an NQP with various LTM improvements by diakopter++.
p: 5a3575c | jnthn++ | src/QRegex/P6Regex/Actions.nqp:
Get ~ goal matching syntax working again; Cursor.FAILGOAL is missing yet, so the error sucks, but otherwise it seems to do the right thing.
 11:32
p: 0b22858 | jnthn++ | src/QRegex/Cursor.nqp:
Basic FAILGOAL implementation so we at least get a bit better error.
 11:41
kudo/nom: 4456cdd | jnthn++ | tools/build/NQP_REVISION:
Bump to NQP revision with goal matching syntax implemented.
 11:46
11:56 cosimo joined
dalek rrot/kill_constants: 4bbe707 | bacek++ | src/pmc/key.pmc:
Don't use constant Keys. It's not needed anymore.
 12:09
rrot/kill_constants: 0a752a7 | bacek++ | / (3 files):
Made HLL entries first class GC citizens.

Properly store and mark created HLL entries.
rrot/kill_constants: 29f6873 | bacek++ | src/multidispatch.c:
Made multidispatch guts non-constant.

We already marking 'const' Subs. Which should also mark all related PMCs as well.
rrot/kill_constants: 84fbdb5 | bacek++ | src/pmc/ (2 files):
Don't clone self to constant PMC.

It's not needed, unfinished, inconsistent and ugly.		 12:10
rrot/kill_constants: 3a4fad3 | bacek++ | / (2 files):
Don't create constant ParrotLibrary.

They are stored in interp->iglobals and will be properly marked anyway.
rrot/kill_constants: 5e1b6cf | bacek++ | / (2 files):
Remove Parrot_pmc_new_constant* functions.

Say goodbay to old "poor man GenGC hack"
12:12 Psyche^ joined
dalek rrot/kill_constants: ecfb375 | bacek++ | / (47 files):
Merge branch 'master' into kill_constants

Conflicts:
  \tinclude/parrot/pmc.h		 12:12
13:12 bluescreen joined 13:32 whiteknight joined
whiteknight good morning, #parrot 13:44
Coke morning. did you get the message about the release email? 13:45
(was that for you? Now I cannot remember.)
whiteknight yeah, it was for me 13:56
and yes, i did forget
At this point, unless somebody else wants to send them out sooner, I'll send out emails tonight 14:19
14:25 mtk joined 14:45 jsut joined
dalek p: 3071d43 | jnthn++ | src/QRegex/P6Regex/Actions.nqp:
Handle nibbler case of named assertions.
 15:07
15:14 bluescreen joined 15:45 bluescreen joined
dalek p: 7fc7a08 | jnthn++ | src/QRegex/Cursor.nqp:
Fix implementation of <before ...>.
 15:46
kudo/nom: a0e51ac | jnthn++ | tools/build/NQP_REVISION:
Get <before ...> fixes. Seems we can now parse with JSON::Tiny::Grammar again.
 15:53
15:56 bluescreen joined 16:02 davidfetter joined, whiteknight joined
whiteknight msg dukeleto I was on the google-melange website and I don't seem to have an option for creating or editing (or even viewing) tasks. Do you have such an interface? Am I missing a permission or something? Thanks 16:45
aloha OK. I'll deliver the message.
16:54 jeffreykegler joined 16:59 jeffreykegler left 17:23 AzureSto_ joined 17:48 lucian joined
cotto ~~ 18:12
whiteknight hello cotto
cotto good morning, whiteknight 18:13
18:20 fperrad joined
dukeleto ~~ 18:59
18:59 Kovensky joined
tadzik ~~ 19:01
whiteknight hello dukeleto, tadzik 19:15
dukeleto whiteknight: i saw your msg, not sure yet 19:25
whiteknight dukeleto: do you have an interface for adding/modifying tickets? 19:26
dukeleto whiteknight: i haven't even tried it yet, sadly. Trying to juggle too many things right now
whiteknight gotcha
that's the thing, I don't know if I'm supposed to have it and don't, or if they just haven't opened it up yet 19:27
dukeleto whiteknight: i thought i made you backup admin 19:28
whiteknight that's what you said. My dashboard when I log in is very spartan 19:29
dalek p: 87bb1c3 | jnthn++ | src/ (2 files):
In a token foo { <bar>+ }, the captures would not be stored. This was due to a bad interaction between :r and the cstack; the bstack contains the current cstack index we're at, but we don't grow the bstack when :r is on. We thus lost the count of captures, popping the entire cstack. This adds an extra bstack frame when :r is on to keep such counts, fixing the issue.
 19:42
19:51 NotFound joined 20:21 mj41 joined 20:39 ambs joined
dalek sella: 78e1923 | Whiteknight++ | src/event/Event.winxed:
Use an object factory for creating payloads, so you can substitute in a new type without needing to subclass Event
 20:41
sella: 5ac444f | Whiteknight++ | src/core/ObjectFactory.winxed:
small, quick cleanups and modernizations for ObjectFactory
20:53 bacek_ joined
dalek sella: bbaac88 | Whiteknight++ | / (2 files):
Rearrange the way we get payloads during event publish to streamline it.
 21:02
sella: 15f1394 | Whiteknight++ | src/ (2 files):
Rearrange some global accessor functions to be functions instead of coroutines. It's almost the same amount of time to execute, and doesn't have the overhead of a coroutine
sella: 8a40d9a | Whiteknight++ | src/ (2 files):
Replace a few more global accessor coroutines
21:23 claudemiro joined
claudemiro hi 21:23
whiteknight hello claudemiro 21:24
claudemiro anybody knows why parrot does not have a pow instruction?
whiteknight parrot does have a pow instruction. It might not be available by default
claudemiro whiteknight, do you know why? 21:25
Coke it's available in one of the dynamic opcode libraries.
whiteknight at least, I thought we had a pow op. I can't find it right now
Coke not all opcodes are available on initial load. You need to load the opcode lib to use it.
whiteknight yeah, it's in the transcendental opcode set
Coke it's in trans.ops in the dynopli 21:26
*b
whiteknight claudemiro: it's not a popular op, so we don't include it by default
claudemiro: in your PIR file, include the text .loadlib 'trans_ops'
Coke historically, I don't think we have a good design decision backing what is dynamic and what is "core".
claudemiro yes I can see now 21:27
21:27 perlite_ joined
whiteknight Coke: no, we certainly do not. I don't think all ops need to be core, but the dividing line is not sanely drawn 21:27
benabik Splitting up the core opcode might help keep PBC files useful between versions.
*opcodes
or *oplib
whiteknight is going home now. Will be back online tonight 21:31
dalek p: 2c0f695 | jnthn++ | src/QAST/Compiler.nqp:
Fix fail in setting the cursor variable.
 21:33
p: f599448 | jnthn++ | src/QRegex/P6Regex/Actions.nqp:
Avoid name leakage that led to action methods get invoked at weird times.
 21:34
kudo/nom: 472ec45 | jnthn++ | tools/build/NQP_REVISION:
Get NQP revision with fixes that make JSON::Tiny mostly work.
 21:40
kudo/nom: c732039 | jnthn++ | t/spectest.data:
Turn on a couple of tests.
 21:55
22:04 lucian joined
dalek nxed: c6e7008 | NotFound++ | winxedst1.winxed:
experimental optimization: switch with all cases strings of length 1
 22:06
nxed: 553e2d5 | NotFound++ | winxedst1.winxed:
class FinalExpr to prevent some mistakes
nxed: abd0d40 | NotFound++ | winxedst1.winxed:
extract the function subid part in IdentifierExpr to other class
nxed: 0942ce4 | NotFound++ | winxedst1.winxed:
oops, forgot to delete a commentend out line
 22:23
kudo/nom: cf8c0e1 | jnthn++ | src/core/Str.pm:
Make substr handle *-1 style third arg.
 22:24
kudo/nom: 5d5940c | jnthn++ | t/spectest.data:
Turn on non-capturing.t.
22:30 rfw joined 22:40 donaldh joined, zby_home joined 22:42 masak left 22:50 bluescreen joined
dalek p: 879321b | jnthn++ | src/QRegex/P6Regex/ (2 files):
First stab and % and %%.
 23:07
23:24 whiteknight joined
whiteknight good evening, #parrot 23:28
benabik o/ whiteknight
dalek kudo/nom: 43be3c4 | jnthn++ | tools/build/NQP_REVISION:
Get % and %% support in regexes.
 23:29
whiteknight hello benabik

Powered by Cro and the Raku® Programming Language.

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

Thank you!