|
weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today Set by moderator on 20 February 2011. |
|||
|
13:32
[Coke] left
13:39
[Coke] joined
|
|||
| sorear | pre-report: | 15:27 | |
| DID: | |||
| * a bit more futzing with spectests | |||
| * a few spectest-motivated changes | |||
| * polymorphic ++ using .succ | 15:28 | ||
| * designed a mechanism for eval (which can't be used this month unless a v2.5 is cut) | 15:29 | ||
| * added minor stuff: [+], repeat/while, 1.5e2, unless{}, loop{}, m//, rx//, <&foo> | 15:30 | ||
| * disabled the generation of unsafe code in the CLR backend by default (5-10% performance penalty, but now if niecza segfaults on mono 2.11.$hash I can report Mono bugs since it's obviously not my fault) | 15:32 | ||
| WILL DO: | |||
| * small TODO items as tuits permit | |||
| * cut v3 | |||
| BLOCKERS: | 15:33 | ||
| * competing time demands | |||
| EOR | |||
| diakopter | pre-report: | 16:29 | |
| DID: | |||
| * made p6eval perlesque a ton faster by making it save the toplevel assembly to disk and process invoke (needs more investigation b/c I don't understand why that occurs) | |||
| * started a decent-sized subset of p6 grammars in perlesque, ala nqp | |||
| WILL DO: | |||
| * continue working on bootstrapping the grammars subset and regex literals, perhaps ~~ matching to literals | |||
| BLOCKERS: | |||
| * none that everyone else doesn't also have | |||
| EOR | |||
| moritz_ | sorear++ | ||
| diakopter++ | |||
| jnthn | Pre-report (got some friends arriving from UK this evening, and will need to feed and look after them, so may miss #phasers) | 16:41 | |
| * Mostly needed thinking time this week, on future serialization needs, problems with the current way the type check cache is structured, static lexpad bits, etc | |||
| * Some answers are forming, I hope to code some of them up pretty soon | |||
| * Did start to sketch out some changes to get a real lexical setting in NQP | |||
| * Plan to finish that up this week, and also do static lexpads | |||
| * Submitted some talks for NLPW, sorted out travel for YAPC::Russia | |||
| Only blockers are my brain needing a higher clock speed, and time. :-) | |||
| EOR | |||
| diakopter | jnthn++ | 16:48 | |
| tadzik | I'll probably miss phasers too, so here's a short-report: | 16:56 | |
| * Fixed modules.perl6.org, TiMBuS++ | |||
| * closed a few bugs on RT, wrote tests for another few, initiated some kind of a bug-fixing party on sunday, huge success | 16:57 | ||
| Plans: thought about making neutro working and testable (Pls maybe?) | 16:58 | ||
| Blockers: Semester started, distracted | |||
| EOR | |||
| moritz_ | I don't know if I'll make it to the meeting, but I'll try | 17:45 | |
| what I did | |||
| * ported more methods in HLL::Grammar from PIR to nqp | |||
| * sent Hague grant application (the one we discussed at YAPC::EU, about exception handling) | |||
| * in rakudo, I improved the error message for "type check on assignment" failures | 17:46 | ||
| * improved error mesasages for ternary parse failures, which tadzik++ had started to do | 17:47 | ||
| * closed a few tickets, wrote a few spectests, improved some spectests | |||
| What I will do: | |||
| * look for more LHFs | 17:48 | ||
| * maybe more nqp hacking | |||
| blocking on: | |||
| tuits, toddler, (and in the case of the book) motivation | |||
| EOR | |||
| tadzik++ for bug fixing etc. | |||
| Util | Pre-report: Nil; EOR | 18:45 | |
|
18:53
masak joined,
colomon joined
|
|||
| masak | preport: blogged some. will blog more, specifically the p5 post. excited. very busy. hoping for more hacking tuits in the cracks between things in the schedule. .eor | 18:58 | |
|
18:59
mberends joined
|
|||
| jnthn | Perl 5 po...oh. :) | 19:00 | |
| diakopter | \\ | ||
| masak | / | ||
| Util | \\ | 19:01 | |
| jnthn | \\/ | ||
| mberends | |o/ | ||
| diakopter | colomon & mberends: the rest of us pre-reported... | ||
| colomon | oh no. it's just us left? | 19:02 | |
| diakopter | (er, colomon & mberends & others) :) | ||
| mberends reads | |||
| colomon | colomon's report: I'm about 20% of the way to duplicating smash's benchmarking environment here. as far as I know, mostly a matter of finding the tuits. | 19:03 | |
| otherwise, er, have accomplished very very little. :( .EOR | 19:04 | ||
| mberends | report: un-procrastinating my NLPW talk. That's all :) | 19:06 | |
| *deprocrastinating | 19:07 | ||
| jnthn hands mberends some antiprocrastinationals. | 19:08 | ||
| Any other reports? And questions/discussion items? | 19:09 | ||
| colomon | I'd love to know how to help with nqp. | ||
| but then, I don't know that I have the tuits. | |||
| jnthn | colomon: The issues tracker has some ideas, though some are EHARD | 19:11 | |
| Seeing if Rakudo's grammar and actions compile with nqp would be interesting. | 19:12 | ||
| colomon | what do you need to get started? git lone nqp and go? | ||
| *clone | |||
| diakopter | someone should reply to dwhipp on #perl6 | 19:14 | |
| (not I since I wouldn't know what to say) | |||
| jnthn | colomon: Should be enough | 19:15 | |
| colomon | jnthn++ | 19:18 | |
| jnthn | colomon: If it Just Works, awesome. | 19:19 | |
| And it's one thing I know I don't ahve to worry over. ;) | |||
| My immediate nqp tasks are quite hard to delegate because I didn't work out how to explain what to do to myself yet. :) | 19:20 | ||
| A very self-contained task is to try porting the new multi-dispatch cache. There's a ticket about it. It'll need C skills. | 19:21 | ||
| But since it's porting something already existing, just on another VM, it may be fairly achievable. | |||
| pmichaud | good afternon, #phasers | ||
| jnthn | pmichaud! \\o/ | ||
| colomon | \\o/ | ||
| mberends | \\=/ | 19:22 | |
| pmichaud | I have a very short report... let me know when to post it :) | ||
| jnthn | pmichaud: now is fine :) | ||
| pmichaud | Early part of last week was spent on holiday with my wife. We had a good time. (more) | 19:23 | |
| jnthn | I gotta run to the shop...just realized I'm missing an ingredient for making noms for @guest | ||
| jnthn will backlog | 19:24 | ||
| pmichaud | Upon returning, I learned that $otherjob is largely finishing after this week, meaning that +@jobs will be effectively zero. :-) | ||
| so, this week I'm finishing up $otherjob for good, and should return to nqp/perl6 hacking middle of next week (if not sooner, depends on how long it takes to finish final $otherjob tasks :-) | |||
| masak | \\o/ | 19:25 | |
| pmichaud | I'm planning to write a longish article to respond to the "perl 6 critic" thread on p6u | ||
| basically saying "here's why we think we'll be able to get back a ton of speed and memory" | 19:26 | ||
| I'll probably ask for comments on the draft before posting it officially | |||
| eor for me | |||
| colomon | pmichaud++ | 19:31 | |
| is there a handy link to the perl 6 critic thread, or do we not really want to look? | |||
| pmichaud | oh, it's a fine thread | 19:32 | |
| I think all of the comments are extremely well thought out -- but I think it deserves a response that outlines why I think there's more room for improvement than would seem apparent at first | 19:33 | ||
| looking for thread link | |||
| groups.google.com/group/perl.perl6....11436bbd9# | |||
| the main thing that makes Rakudo startup slow at the moment is that we don't have good serialization (more) | 19:34 | ||
| thus we have to build at init-time things that really should've been frozen/built when Rakudo itself was built | |||
| this also hugely increases the memory footprint | |||
| because we have to store subs to do the initialization instead of simply storing the results of the initialization | |||
| the work that jnthn++ has been doing with 6model starts to also look at the serialization issues | 19:35 | ||
| colomon | pmichaud++ | 19:37 | |
| jnthn | back | 19:41 | |
| pmichaud: yay! | |||
| pmichaud: Yes, I've started to make some preps in that area. | 19:42 | ||
| pmichaud: Think I know how to prepare the ground for going all the way, without having to go all the way yet. :) | |||
|
20:02
cognominal joined
20:21
mberends left
20:26
tylercurtis joined
20:41
pmichaud left,
Util left
20:49
cognominal left
21:00
pmichaud joined,
Util joined
21:05
diakopter left,
moritz_ left,
pmichaud left,
Util left,
tylercurtis left,
colomon left,
PerlJam left,
TimToady left,
eternaleye left,
masak left,
[Coke] left,
Tene left,
[particle] left,
tadzik left,
sorear left,
jnthn left
21:06
moritz_ joined,
Util joined,
pmichaud joined,
tylercurtis joined,
colomon joined,
masak joined,
[Coke] joined,
eternaleye joined,
jnthn joined,
Tene joined,
[particle] joined,
sorear joined,
PerlJam joined,
tadzik joined,
TimToady joined
21:07
diakopter joined
21:49
pmichaud left,
Util left
21:50
Util joined,
pmichaud joined
22:22
masak left
|
|||