|
weekly Rakudo status meetings with phase transitions: Tue 19:00 UTC | IR clogs at irclog.perlgeek.de/phasers/today Set by moderator on 25 June 2010. |
|||
|
04:20
ash_ joined
04:21
ash_ joined
06:37
eternaleye joined
06:49
eternaleye joined
09:24
masak joined
16:55
ash_ joined
18:15
mberends joined
18:31
takadonet joined
18:51
colomon joined
|
|||
| mberends | ENOMORITZ | 19:01 | |
| pmichaud | ah | ||
| mberends | ENOMASAK | ||
| ENOJNTHN | |||
| pmichaud | "Captain! The phasers are empty!" | ||
| [Coke] | ~~ | ||
| *pew pew* | 19:02 | ||
| sorear | \\o | ||
| mberends | lots of progress this last week, anyway :) | ||
| pmichaud | really? I didn't do anything. :-) | ||
| mberends | heh | ||
| pmichaud | do we want me to post what I did since last #phasers? Or should I just say "look at the rakudo commit log"? | 19:05 | |
| mberends | for me, there are enough records in the logs and the other places where postings occur. #phasers should be more about the future than the past. | 19:07 | |
| [Coke] | I am looking forward to a smooth release of the compiler next month. =-) | 19:09 | |
| pmichaud | okay | ||
| future, then | |||
| mberends | it's definitely getting smoother every single day :) | ||
| pmichaud | right now I'm working on a refactor of the series operator, because the current one makes me feel bad | ||
| I'm a good way along in fixing autoviv of hashes and arrays -- most of the commits over the weekend were prep for that. I expect that in the next day or so. | 19:10 | ||
| after autoviv, my next tasks are closures, subrule backtracking, and repl | |||
| I don't have an eta on those, other than "before Wed of next week". | 19:11 | ||
| No significant blockers (other than time) at this point. | |||
| EOR | |||
| mberends | those are probably big jobs. how is that working out with your private life? | ||
| sorear | closures and repl we understand well enough to duct-tape over if they don't get dealt with | ||
| pmichaud | it's fine, so far. | ||
| mberends | :) | 19:12 | |
| pmichaud | Paula has been really good about making sure private life things don't interfere with rakudo progress for the next week or two | ||
| PerlJam | paula++ | 19:13 | |
| pmichaud | the next part that might get dicey around here will be this next weekend | ||
| ...but we'll deal with that this weekend. :) | 19:14 | ||
| mberends | ok. it seems to be going fairly well under the circumstances. paula++ indeed. | ||
| pmichaud | speaking of which, I have to run errands right now, will bbiah | ||
| oh, before I go, I could use any further reviews/comments on my slides -- especially theones from my rakudo star presentation. | 19:15 | ||
| I'm planning to turn those slides into a blog post, and possibly a video | |||
| mberends will look | 19:16 | ||
| jnthn gave me a plan for making a dynpmc to implement select(). I haven't worked on it yet, but hope to have it done well before R*. With that, we would have some value of async socket I/O. | |||
| moritz_ back | 19:18 | ||
| sorear | why a dynpmc? | ||
| why not just use the NCI? | |||
| (and by select I'm hoping you mean poll) | 19:19 | ||
| mberends | erm, I think it was necessary to implement part of it in C. it was the select(2) function we were looking at. Perhaps with a zero timeout it becomes similar to poll. | 19:23 | |
| also, the dympmc approach would hopefully lead to the code making its way into Parrot core. | 19:25 | ||
| reading the poll(2) manual there's considerable overlap in functionality. It could go either way during implementation. | 19:26 | ||
| moritz_ | mberends: if you want to know what kind of functions people want, talk to _sri on #perl6 | 19:27 | |
| mberends | yes, I've read his offer(s) of firstborn etc ;) | 19:28 | |
| sorear | _sri is lying | 19:29 | |
| mberends | initially I'll be happy enough just to have anything that works for a proof of concept. | ||
| sorear | I gave him async sockets :from<perl5> | ||
| no firstborns for me | |||
| mberends | yeah, that's ironic | 19:30 | |
| sorear | mberends: poll is basically select, but without the stupid interface | ||
| mberends | oh :-) | ||
| sorear | it doesn't require macros and doesn't limit file descriptors to <1024 | ||
| the functionality is almost exactly the same | 19:31 | ||
| mberends | makes sense. the interface would get a P6 makeover anyway :) | ||
| but some people think async means threads and callbacks, which are beyond me to implement :( | 19:32 | ||
| sorear | you can implement threads and callbacks on top of select/poll/epoll/kqueue | 19:33 | |
| see search.cpan.org/perldoc?Coro::EV and related modules for a demonstration of principles | 19:34 | ||
| mberends | POE basically did that on top of P5 select(), that was the pattern we followed. | ||
| yes, I think some call it co-operative multitasking | 19:35 | ||
| sorear | you're involved with POE? | ||
| mberends | no, but I think I should have been | ||
| I did write some code like that a while back. POE looks like an idiosyncratic style, that did put me off a little. | 19:36 | ||
| I think it may work to use poll() for implementation, but to present it as a cleaned up select() to help P5->P6 migration of modules. | 19:40 | ||
|
19:45
ash_ joined
|
|||
| Tene | It really makes me sad to see people talking about things that I put on my todo list ages ago, and never got around to. | 20:20 | |
| My todo list is looking more and more like a graveyard of burned-out ideas, lately. | 20:21 | ||
| Oh, this is #phasers, not #perl6. | |||
| colomon | no worries | 20:22 | |
| jnthn | o/ | 20:23 | |
| Sorry I missed #phasers | |||
| After work event. | |||
| colomon | \\o | 20:24 | |
| mberends | \\o | 20:27 | |
|
23:05
pmichaud joined
23:10
PerlJam joined
|
|||