Parrot 5.3.0 "W00tstock Parrot" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 30 April 2013.
00:06 particle joined
dalek kudo/nom: d2af402 | pmichaud++ | tools/build/NQP_REVISION:
Bump NQP_REVISION to get ?-quantifier update and is-prime fixes.
00:15
01:22 woosley joined 03:17 woosley left 03:23 kurahaupo joined 04:25 Psyche^ joined
dalek p: 3d5e761 | pmichaud++ | src/vm/parrot/stage0/ (9 files):
Update parrot bootstrap with ?-quantifier changes.
06:17
06:37 Mike-PerlRecruiter_ joined
dalek p: 1b11be1 | pmichaud++ | src/QRegex/Cursor.nqp:
Convert NQPRegex.ACCEPTS to be multimethods based on concrete-ness.
06:53
07:08 xcombelle joined 08:18 bouncy joined 08:21 Timbus joined 08:23 donaldh joined 08:30 woosley joined 08:37 eternaleye joined
dalek p/rak-jvm-support: 26f19d6 | (Timo Paulssen)++ | src/vm/jvm/ (2 files):
implement unlink, cwd, rename, improve rmdir.
10:22
p/rak-jvm-support: 6e192ab | (Timo Paulssen)++ | src/vm/jvm/ (2 files):
copy, link and symlink. fix rename.
10:29
p: 72df616 | (Timo Paulssen)++ | README (2 files):
let github render the readme prettily
10:30
10:41 brrt joined 11:48 woosley left 12:21 brrt joined 12:59 brrt left 13:28 woosley joined 13:29 woosley left 13:58 contingencyplan joined 14:07 rurban1 joined 15:34 brrt joined, rurban1 joined
brrt hey parrot 15:58
any news on the gsoc front?
16:12 benabik joined
benabik ~~ 16:17
Util brrt: In the #parrotsketch weekly meeting yesterday, sa1 showed us his proposal and recieved comments.
benabik Someone should +o slavorg 16:18
WE don't have a msgbot here anymore, do we?
Util Over the last week or so, I have seen several potential GSoC candidates poke their nose int the channel and discuss ideas, but I was not tracking the number. 16:19
At least one mentor has registered; probably more, but I was only seeing the ones that dukeleto did not get first.
benabik Perhaps I should mentor this year. I'm getting paid for research this summer so I didn't register as a student. 16:20
Util benabik: irclog.perlgeek.de is down (and being rebuilt, I think I overheard). It handled logging. Perhaps it also hosted the bot? 16:21
benabik Util: IIRC bacek hosted aloha and lost access to the machine it was running on. It's been gone for a while.
Util benabik: mentoring would be great! Have you seen github.com/parrot/parrot/wiki/Summ...Task-Ideas ? Is there anything there you could mentor? Or perhaps you have a Task idea yourself? 16:23
benabik Considering I was the originator of PACT, I could probably mentor those. :-D 16:24
Although I'll note the disassembly half of PACT is ~99% done.
Util My Linux meeting tonight has some college students. I will make sure they are away of GSoC, and ask them to carry the info to the on-campus LUG.
benabik I mentioned it to people at school, but most of them were finding real internships. 16:25
(Or getting an RA contract like me.)
benabik added himself to the PACT and 6model projects as mentor. 16:27
I think I need to register on melange?
Util benabik: PACT looks like a good task. Would you add yourself as "Possible Mentors" on the Wiki page?
benabik Or is that not required yet?
Util: Ahead of you. :-) 16:28
Util benabik: I expect registration is required. In fact, go ahead and request to be a Mentor with Parrot Foundation.
It is not an obligation; it just links you to PaFo after we "approve" you. 16:29
benabik had forgotten how well melange worked. 16:33
Util: Do you have a handy link to the Parrot org page? Getting errors on their search. 16:34
Util benabik: www.google-melange.com/gsoc/org/goo...013/parrot 16:35
benabik Util++
Util Also, I see now that I could have "invited" you via email. Let me know if you need that avenue. 16:36
benabik Right, no aloha. I wonder how irritated my school would be if I ran an IRCbot on my desktop.
Util karma Util
hmmm. I have a machine to host it, if someone has the source code. 16:37
benabik Util: github.com/bacek/aloha
I think bacek put a copy of the old DB up somewhere, but I don't know where. 16:38
Util benabik++ :)
I can't get to it today, but I put it on my short list. 16:39
benabik Util++++++++++++
:-D
I haven't pulled Parrot/NQP/Rakudo in a month. Oy. Classes will do that.
Actually, I should probably be writing a paper while the kid is asleep instead of sitting on IRC. Oh well.
arnsholt I should probably be writing an article instead of sitting on IRC. What can ya do, eh? =) 16:41
16:45 rurban1 joined
dalek p/rak-jvm-support: 66ea32d | pmichaud++ | src/ (4 files):
"Temporarily" convert ?-quantified captures to use **0..1 instead.

This is in preparation for bringing the ?-quantifier in line with S05, such that ? is an "item quantifier" instead of a "list quantifier".
16:48
p/rak-jvm-support: 54a9a07 | pmichaud++ | / (2 files):
Start adding tests for quantified captures.
p/rak-jvm-support: afbc772 | pmichaud++ | t/qregex/ (2 files):
Fix up incomplete quantified captures commit 54a9a07.
p/rak-jvm-support: cde4d6d | jnthn++ | src/vm/parrot/6model/serialization.c:
Fix a version check thinko.
p/rak-jvm-support: 16fa719 | util++ | src/vm/parrot/ops/nqp_bigint.ops:
Fix bigint_is_prime via intermediate int result (RT #117731)
p/rak-jvm-support: b48b5c3 | pmichaud++ | / (3 files):
P6regex: Switch ?-quantifier to be an item quantifier instead of a list quantifier.

In other words, using ? on a capturing expression no longer produces a list of captures; it's a single capture (or Nil/nonexistent).
p/rak-jvm-support: 3d5e761 | pmichaud++ | src/vm/parrot/stage0/ (9 files):
Update parrot bootstrap with ?-quantifier changes.
p/rak-jvm-support: 1b11be1 | pmichaud++ | src/QRegex/Cursor.nqp:
Convert NQPRegex.ACCEPTS to be multimethods based on concrete-ness.
p/rak-jvm-support: 72df616 | (Timo Paulssen)++ | README (2 files):
let github render the readme prettily
p/rak-jvm-support: 0f1be31 | jnthn++ | / (22 files):
Merge branch 'master' into rak-jvm-support
p/rak-jvm-support: 11d1840 | jnthn++ | src/vm/jvm/stage0/ (10 files):
Update bootstrap after merging from master.
16:56 rurban1 joined 16:57 mtk joined 17:06 mtk joined
dalek Heuristic branch merge: pushed 17 commits to rakudo/jvm-support by jnthn 17:10
tangentstorm I would like to mentor an M0 conversion for GSoC 17:14
Util tangentstorm: That sounds ambitious for a college student who is (probably) just starting on the Parrot source. Then again, I am not aware of how far along the M0 work-in-progress branches are. 17:20
tangentstorm: Would you please write up a proposed task for this, and put it on github.com/parrot/parrot/wiki/Summ...Task-Ideas , for review? 17:21
tangentstorm Hrm. You may be right about the difficulty level. 17:22
M0 seems small enough...To me the real work is mapping all the existing opcodes to M0 17:23
benabik A functioning M0 interpreter would be a nice pre-requisite.
tangentstorm I spend a lot of time mentoring through reddit's #learnprogramming channel on freenode... I don't actually have a student yet but I'm asking around 17:25
benabik: I think writing a new interpreter is the wrong approach. 17:29
benabik tangentstorm: You need a new interpreter to read a new bytecode.
tangentstorm: Yes, in theory, we could do a C->M0 conversion leaving the .pbc format intact but you still need something that executes the underlying M0. 17:30
tangentstorm well ... not necessarily wrong by itself. i think there should be multiple implementations (especially in javascript) with some small changes to break the c dependency.
benabik: the idea is that each of the current opcodes would be reimplemented in terms of the simpler ones 17:31
so i think parrot itself should serve as the m0 interpreter.
benabik tangentstorm: I'm very aware, yes. But the simpler ops need to be executed by an interpreter. Which we don't have.
tangentstorm each new op would just be added to parrot
benabik I suppose you could implement parrot ops that do M0 ops that we then implement other parrot ops in terms of. But that way lies madness and much much slowdown.
tangentstorm on demand, as the old ops were ported. 17:32
benabik I'm willing to be proven wrong on that one.
tangentstorm why is that madness? it seems to me that's the only way to do it without breaking all the existing tools.
at some point, each old op has to be translated to the new ops anyway, so why not approach it as a refactoring problem? 17:33
that way parrot continues to run at each step along the way.
benabik Every translated op would need to start a nested interpreter, which has been a cause of much pain in the past. 17:34
tangentstorm hmm
benabik I suppose you could expand them in place, but our current bytecode manipulation bits would handle that very poorly.
If you want to approach it that way, I suggest having an M0 interpreter that knows how to call parrot ops, rather than adding M0 ops to the parrot interpreter. 17:37
Then when you convert the PBC to M0, you turn unconverted ops into the 'call old parrot ops' bit. 17:38
tangentstorm that makes sense 17:40
like the outer interpreter in forth.
dalek rrot: 0ae1d2b | util++ | / (5 files):
[GH #967] Add readlink() and Parrot_file_readlink(), with tests.
17:56
18:05 chekkaa left 18:37 Mike-PerlRecruiter_ joined 18:52 brrt joined 19:21 rurban1 joined
dalek kudo/nom: 5edb135 | pmichaud++ | src/core/Cursor.pm:
Failed regex matches should return Nil.
19:21
p/rak-jvm-support: 7d9f0b9 | jnthn++ | src/vm/jvm/ (2 files):
Case consistency with Parrot backend.
19:48
p/rak-jvm-support: 1e2629a | jnthn++ | src/vm/jvm/QAST/Compiler.nqp:
Add infrastructure for getting op result type.
p/rak-jvm-support: 907a86c | jnthn++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/CallFrame.java:
Make 24-module.t less busted on JVM.

This also helps towards the Rakudo work.
Coke wonders if we need to have a chat somewhere not on channel about incoming proposals. 19:53
dalek kudo/jvm-support: e14ef54 | jnthn++ | src/ (2 files):
Get hold of QAST compiler in more portable way.
kudo/jvm-support: 623d45e | jnthn++ | src/vm/jvm/runtime/org/perl6/rakudo/Ops.java:
Implement p6bool op for JVM.
kudo/jvm-support: f5cd6d9 | jnthn++ | src/Perl6/Actions.nqp:
Correct atkey that should be atpos.

Also remove an unrequired "my"; it's the default for subs.
kudo/jvm-support: 11c7841 | jnthn++ | src/Perl6/World.nqp:
Replace a join with nqp::join.
kudo/jvm-support: 80f3017 | jnthn++ | src/Perl6/Actions.nqp:
Make sure $*DISPATCHER is declared before use.
20:19
20:29 rurban1 joined 20:40 brrt joined
dalek kudo/nom: 8025743 | pmichaud++ | lib/Test.pm:
Switch Test.pm's &diag function to accept Mu arguments. We may
20:55
21:45 donaldh joined 21:49 kurahaupo joined 22:16 rurban1 joined 23:30 WillWorkForBytes joined 23:33 rurban1 joined