00:09 davidfetter joined 00:21 particle joined 00:54 japhb joined 02:05 particle joined 02:21 particle joined 02:59 particle joined 04:24 particle1 joined 05:29 cognominal joined 07:16 Tweety joined
Tweety Hi there. I am Tweety. I am now publicly logging this channel. If you don't want to be logged, please leave now. 07:16
12:54 particle joined 14:09 particle1 joined 15:25 rdice joined 15:44 davidfetter joined 16:24 particle1 joined 17:30 Auzon joined 17:42 Whiteknight joined 17:50 avar joined 17:56 Whiteknight joined 18:04 barney joined 18:06 pmichaud joined 18:11 cotto-work joined 18:21 chromatic joined 18:24 NotFound joined 18:27 DietCoke joined, rgrjr joined
DietCoke Hallo, everyone. anyone with a report comin', pm me to let me know. about 2 m. 18:28
18:31 allison joined
pmichaud here. 18:31
DietCoke alrightee. let's go with tene, who assures me his report is nanoscopic.
18:32 jonathan joined
DietCoke See? 18:32
Ok. Chromatic, you go. Tene can follow you if he wakes up. =-)
chromatic I've been fiddling with strings, especially constant strings. 18:33
More of our strings are now compile-time constant, and dynpmcs can now use CONST_STRING for those too.
Uniformity of examples ++
I'm working on const string caching, to save even more compile time.
I also changed the string splitting does_isa to use a hash lookup, and that's a nice conceptual improvement, even if the code code use a bit of a refactor. 18:34
Mostly concentrating on small optimizations that produce cleaner code, though I think the benchmark is now maybe 10% faster than last week.
I'll do my best to poke at the pdd25cx branch too, to see if there are any fixes I can make.
That merge is going to be nasty though....
and that's it.
Tene Cardinal now has support for class definitions and instantiation. Couple of syntax fixups in response to cizra, an acquaintance I'm trying to con into working on cardinal. 18:35
.eor
DietCoke excellent. I forgot to actually say something nice and pleasant to start off the sketch, so let's take that as written...
18:35 tewk joined
DietCoke pmichaud? (and that's everyone who said they had a report. pm me if you have something to share. =-) 18:36
pmichaud ** Rakudo spectest_regression: 59 files, 705 passing tests (+4, +67 from last #parrotsketch)
== Parrot stuff
: Added get_number vtable method to Hash
: Fixed the 'add_sub' method in Namespace to allow MultiSub PMCs
: Ported the CONTROL_* exception constants from pdd25cx into trunk
== PCT stuff
: Added 'control handlers' to PAST::Block -- allows a block to catch an exception (e.g., return exceptions) and process them
: PCT provides reasonable defaults, but also allows HLLs to easily override control exception handling
== PGE stuff
: Fixed a backtracking bug in the operator precedence parser 18:37
: Adjusted token matching in operator precedence parser so that word tokens (all alphanumerics) require a word boundary to match.
== NQP stuff
: Added 'return' statement to NQP.
== Rakudo stuff
: Updated test targets
: Fixed a number of operators to do proper casting
: More refactorings to handle MMD properly
: Added 'return' function.
: pointy subs in 'for' now work
: We need a review to see which tests are now regression candidates
: Have a preliminary implementation of 'fail', will commit it shortly.
: Fixed a bug where floating point constants were losing precision in generated PIR code
: Created a tools/test_summary.pl script to run regression tests and provide more accurate summary of planned, actual, passed, failed, todo, and skipped tests
: Fixed scalar assignment to work even for non-Rakudo objects (jhorwitz++)
: Refactored handling of named unaries
: Added Complex type
: Added closure interpolation to double-quoted strings
Big news of the week is that we now have 'return' nodes in PCT, and 'return' works in NQP and Rakudo (for small values of "works")
.eor
DietCoke tries to turn off the firehose, there... 18:38
japhb I feel so unproductive now.
pmichaud ...too much detail?
DietCoke nope. =-)
you're fine. thanks for all the work. =-)
OH NOES ITZ JONATHAN!
18:38 jhorwitz joined
jonathan * Did Rakudo stuff on Thursday, including 18:38
** Getting readonly/rw/copy semantics right on parameters
** A bit more work on Mutables and make .VAR and VAR(...) work
** Got more of Range in place, and switch over to .. creating Range object (which is a lazy iterator)
** Various other small fixes here and there
* Got :instanceof(...) on subs to parse; no implementation yet
* Put a good fix in for a GC bug that I'd put a bad fix in for last week
* Got slides from my talks all online and written into a journal post.
* Will have Rakudo day on Thursday this week. 18:39
EOR
DietCoke lots of rakudo work going on, excellent.
WK?
Whiteknight * Worked on fixing deprecation notes in PDD09, submitted a big patch for that
* Continuing work on the PDD09 GC. still early but making steady progress
* Evaluating and trying to salvage branches/gmc
EOR
DietCoke ok. lemme go next to last... 18:40
I went on a crazy RT spree, gave many tickets a [TAG], closed some old ones, proposing some patches, finding that HLL doesn't seem to do much any more, opened some tickets for chip and simon to remove their ex-parrots from CPAN. Tcl is still on life support. Looking
forward to yapc, except for the part about my diet. I'll be there
Friday.
:wq
allison, I assume you have something to report on, go ahead. =-)
allison - I spent a good bit of time finishing off the Parrot Foundation documents this week. 18:41
- The filing went through yesterday. Congratulations, you're all the proud parents of a healthy 10 lb bird!
- Also tracing down bugs with contexts in the concurrency branch. Basically, it's a fundamental mismatch between the reference counted context structs, and the garbage collected PMCs that contain the contexts and are stored within contexts. I've made and reverted a few conservative fixes, and am now nearly finished with a shift to make contexts full PMCs, so they participate cleanly in GC.
pmichaud allison++ and others # Parrot Foundation
allison - The change to contexts will fix some of our other context memory leaks too.
EOR
jonathan allison++ # contexts as PMCs! ;-) 18:42
s/;-)/:-)/
DietCoke Anyone else with a report, you're out of luck. muahahah.
So, two big parrot/perl6 things at yapc this week, the hackathon and the workshop. 18:43
There is a wiki page for each of them at the conference wiki.
I didn't see any questions reserved... Anyone? 18:44
cotto-work yes
DietCoke shoot.
Tene DietCoke: What should I work on tonight while I'm offline at the hotel?
pmichaud Lightning Talk! :-) 18:45
cotto-work I'm writing up a PHP spec and want to know how to be sure that any implementation won't be tainted by the original license.
DietCoke tene: Coming up with an awesome implementation to make ATTR in pmcs actually do all the hard work?
cotto-work (for Plumhead)
DietCoke I'll defer to allison on any licensing questions. 18:46
allison cotto-work: it's only affected by the original license if you cut and paste code
cotto-work sounds good to me
that's the assumption I've been working under
eoq
allison re-implementing the same interface is not touched by copyright
DietCoke I asked on list, but what's the deal with HLL? looks like the mapping got ripped out when we switched to PCC, perhaps?
barney sounds good to me too
allison DietCoke: it's possible. looking for message... 18:47
DietCoke allison: rt.perl.org/rt3/Ticket/Display.html?id=55586 18:48
allison DietCoke: it's also possible that it never worked quite right
DietCoke ok, I didn't so much ask as rant. =-)
allison: it worked well enough to use it in tcl. =-)
when i resurrected tcl a few months ago, I noticed that part wasn't working at the moment, but hadn't gotten back to diagnosing why until today.
that also refers back to rt.perl.org/rt3/Ticket/Display.html?id=40124, which was an old request for documentation on what we could map. 18:49
I'd rather see pdd25 merged back in though, so don't change your priorities. =-)
allison it also might have been written out during the objects or PMC transitions 18:50
pmichaud agreed, it's not high priority for rakudo at the moment either.
allison could you condense it into a TODO test, and point to it from the ticket?
agreed, pdd25 branch is a higher priority, but we will get to it
Whiteknight when is the pdd25 branch merge going to happen, roughly? 18:51
DietCoke it's not a PIR-only test, requires a pmc.
I am hoping we can resolve all the issues with pdd25 in time to do it just after the next release.
... given there's a hackathon this weekend. Allison?
allison yah
aside from PGE not compiling, we're down to about 10 failing tests in the pdd25 branch 18:52
pmichaud do we have any specific targets/tasks for the hackathon?
DietCoke those're on the wiki. Feel free to add to it.
pmichaud okay, excellent.
DietCoke I may create a meta ticket for the hackathon, also.
(but let's just -refer- to the tickets we want done, so we can close it when we're done.) 18:53
ticket number is... 18:55
55588
FULL HOUSE!
Anyone have anything else we should do to prepare for the hackathon or the workshop? 18:56
kid51++ for organizing the workshop.
chromatic Figure out time, location, food, and snacks.
DietCoke At the moment, "the dorms".
I am willing to be the primary point of contact for folks trying to find us. 18:57
pmichaud I'm arriving Saturday noonish.
DietCoke ... I suppose that means I should give out my cell phone #, at least to you jokers.
chromatic Especially for those of us jet-lagged after cross-country flights arriving around midnight. 18:58
DietCoke alrightee. anything else? I'll send out an email about the hackathon. 19:00
no? See most of you this weekend! 19:02
pmichaud see you this weekend.
19:02 Auzon left
cotto-work agreed! 19:02
pmichaud no parrotsketch next week?
allison see you there!
19:03 chromatic left, cotto-work left
DietCoke Good point. I think we'll skip it, yah, as most of us will be elsewhere. 19:03
I'll send an email about that as well.
gooooood night, everybody. 19:04
19:04 DietCoke left, cognominal joined 19:05 jonathan left 19:06 NotFound left 19:23 jhorwitz left 21:07 davidfetter joined