|
Parrot 4.2.0 "Ornithopter" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 21 March 2012. |
|||
|
00:15
kid51 joined
|
|||
| dalek | rrot: ba6af98 | jkeenan++ | config/gen/makefiles/root.in: Merge branch 'kid51/makefile_variables' |
00:35 | |
|
00:54
workbench joined
|
|||
| dalek | sella: 382f1d6 | Whiteknight++ | s (7 files): [Net] Add Net to the build and fix errors and warnings. Add more request handlers to the file protocol |
01:48 | |
| sella: 228bd32 | Whiteknight++ | src/ (3 files): [Core] play around with the format of reported backtraces in try_report. Aesthetic change, I'm just trying it out |
|||
| sella: fc15ef6 | Whiteknight++ | src/core/Parrot.winxed: [Core] Add an ExceptionFormatter type and use it to implement custom backtrace formatting |
|||
| sella: 5c36917 | Whiteknight++ | s (6 files): [Net] Several fixes so a simpel ad hoc test of file:// urls works |
|||
|
01:57
workbench joined
|
|||
| dukeleto | ~~ | 02:19 | |
|
04:00
benabik joined
04:54
Psyche^ joined
05:37
parthm joined
|
|||
| nine | *yawn* good morning, #parrot | 06:56 | |
| tadzik | hello #parrot | 08:14 | |
| nine | o/ | ||
|
09:28
JimmyZ joined
09:32
mj41 joined
|
|||
| moritz | \\o | 10:57 | |
| patshaughnessy.net/2012/3/23/why-yo...-ruby-2-0# ruby's GC optimized for copy-on-write/fork | |||
|
11:45
whiteknight joined
|
|||
| whiteknight | good morning, #parrot | 11:51 | |
|
12:00
taryk joined
12:07
taryk joined
12:32
kid51 joined
12:48
mj41 joined
13:03
parthm joined
|
|||
| parthm | Hello. I generated a language shell using mk_language_shell.pl. When I try to compile it using "parrot setup.pir" the compilation fails. I am using parrot 4.2.0. pastebin.com/YyQRdpRc | 13:24 | |
| Whats the recommended way to use parrot, create_language.pl, mk_language_shell.pl or to use pir directly? | 13:25 | ||
| whiteknight | mk_language_shell.pl is the newer one, I think | 13:27 | |
| parthm | whiteknight: thanks. i will try to generate the lang shell using 4.1.0 and see if that builds. unfortunately, i haven't reached a point where i understand the error messages well enough. | 13:28 | |
| i see a similar error with 4.1.0. | 13:32 | ||
| dalek | sella: d17b551 | Whiteknight++ | src/unstable/net/ (2 files): [Net] Add a Request.set_method method, so we can make other requests besides GET |
13:38 | |
| sella: 9210a94 | Whiteknight++ | s (11 files): [Net] Initial implementation of the Http protocol. A simple HTTP get works in an ad hoc test |
|||
| nine | moritz: very interesting article | 13:46 | |
|
14:14
autark joined
14:46
PacoAir joined
14:54
benabik joined
15:46
Psyche^ joined
|
|||
| nine | Who'd have thought? Parrot's Test::More is not yet thread safe! | 16:28 | |
|
16:58
sri joined
17:01
whiteknight joined
|
|||
| nine | whiteknight: seems like the threads branch is 2.42 % slower than current master with signficance 8.9σ according to moritz' mandelbrot-color.pl | 17:39 | |
| whiteknight | okay, that's not entirely unexpected | 17:40 | |
| that's in a normal single-threaded case? | |||
| nine | I wonder how much of that is due to the orig_interp pointer in PMC | ||
| whiteknight: yes | 17:41 | ||
| moritz | how many interpreters are there, usually? | 18:18 | |
| nine | moritz: one per thread, why? | ||
| moritz | nine: because if you don't need 2**64 of them, it might be cheaper to keep a global array of interpreters, and sneak some extra bits into some other header, instead of making a full new pointer | 18:19 | |
| nine | moritz: in truth I only need a single flag to differ between our own PMCs and those belonging to a foreign interp and only during forwarding a call by proxy. So yes, this pointer can be optimized away. But for now its darn convenient for debugging this stuff. | 18:21 | |
| moritz | nine: ok | 18:22 | |
|
19:06
lucian joined
19:15
GodFather joined
|
|||
| nine | Ironic: ResizablePMCArray supports delete while PMCList doesnt | 19:51 | |
| dalek | sella: 0d45ae4 | Whiteknight++ | / (11 files): [Net] Add additional HTTP status codes. Cleanup Protocol.Http to break it up into logical blocks. Start implementing some POST logic for Requests. Remove Message.Factory and be more simple about the way we create requests and responses. Change the protocol interface to take a user_agent and not keep any permanent state. Change the Protocol.Factory to return existing instances, since they are stateless |
19:57 | |
|
20:13
perlite_ joined
|
|||
| nine | msg whiteknight chameneos.pir got 3.5 times faster by using the wait op which I just fixed for threads :) It's still slow as hell but there's plenty of optimization opportunity left | 20:16 | |
| aloha | OK. I'll deliver the message. | ||
|
20:31
contingencyplan joined
21:03
PacoAir_ joined
22:46
mj41 joined
23:01
davidfetter joined
23:12
benabik joined
|
|||
| benabik | Good evening, #parrot | 23:13 | |