|
Parrot 6.1.0 "Black-collared Lovebird" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 18 February 2014. |
|||
|
00:46
kid51 joined
01:07
rurban1 joined
01:38
Zyro joined,
Chirag_ joined
03:05
Zyro joined,
Chirag_ joined
03:40
rurban1 joined
04:52
rurban1 joined
05:45
Chirag__ joined
07:12
rurban1 joined
07:15
Zyro joined,
Chirag_ joined
07:21
FROGGS joined
07:43
Chirag__ joined
08:45
rurban1 joined
09:07
denis_boyun joined
09:52
wagle joined
10:01
wagle joined
10:40
rurban1 joined
11:19
wagle joined
13:07
denis_boyun joined
14:57
rurban1 joined
15:46
rurban1 joined
16:08
rurban1 joined
16:09
kshannon joined
16:25
mtk joined
16:29
rurban1 joined
|
|||
| rurban | looking into bison 3 now. repro with perl Configure.pl --maintainer | 17:09 | |
|
17:49
denisboyun joined
17:58
FROGGS joined
|
|||
| rurban | arglist: | STRINGC ADV_ARROW var /* State 385 conflicts: 1 shift/reduce */ | 18:14 | |
| 263 arglist: STRINGC . ADV_ARROW var | |||
| 326 stringc: STRINGC . | |||
| ADV_ARROW shift, and go to state 446 | |||
| ADV_ARROW [reduce using rule 326 (stringc)] | |||
| $default reduce using rule 326 (stringc) | |||
| FROGGS | O.o what kind of magic is this? | 18:15 | |
| rurban | bison | ||
| See www.tldp.org/HOWTO/Lex-YACC-HOWTO-7.html#ss7.2 | 18:16 | ||
| FROGGS | ahh, thanks | 18:17 | |
| rurban | And the imcc flags (optims and debug) are never passed to the imcc via the PMC | ||
| So it's a bit hard to optimize and debug it | |||
|
18:19
TonyC_ joined
18:24
TonyC joined
18:31
TonyC_ joined
18:56
sa1 joined
19:02
Zyro joined
19:29
denis_boyun__ joined
|
|||
| Coke | post to tpf-gsoc aboutimproving performance of method signatures - candidate is reaching out for pre-mentorship | 19:37 | |
| FROGGS | yeah, seen it too :o) | 19:38 | |
|
19:42
Chirag_ joined
|
|||
| rurban | Ok, I'll start digging... | 19:42 | |
| Coke | rurban++ | 19:43 | |
| rurban | I am just re-adding the imcc optimizations and flags back in. I needed imcc yydebug | ||
| bison 3 looks good with Util's fix, but there's one shift-reduce conflict which disturbs me | 19:44 | ||
| do we have a #gsoc channel on magnet? | 19:45 | ||
| FROGGS | #soc-help | 19:49 | |
| rurban | thanks | 19:50 | |
| FROGGS | you're welcome | ||
| Chirag_ | hey! | 19:51 | |
| Coke | Ho. | 19:54 | |
| rurban | hey, it's 1:30 am your time, right? | 19:55 | |
| I just wrote you a short letter about your questions. I'll need to dig into the history also to found to good points and backtraces to compare | 19:56 | ||
| Chirag | yup | ||
| rurban | and the relevant discussion in the parrot-dev mailinglist | ||
| lists.parrot.org/pipermail/parrot-dev/ | |||
| Chirag | registered already :D | ||
| FROGGS | Chirag++ | 19:57 | |
| rurban | the biggest problem is that ffi-like signatures are parsed as string at run-time "S->P" and not at compile-time | ||
| Chirag | but before it wasnt the case? | 19:58 | |
| rurban | they should be preprocessed and the method pointers probably cached | ||
| nope, before it was pretty good | |||
| before method calls needed to create 2 objects at run-time, now only one. That was an improvement | 19:59 | ||
| method calls are created automatically with the PMC compiler, which is pure perl5, so pretty easy to understand (if you know perl5) | 20:00 | ||
| Chirag | ok.. | ||
| rurban | how's the weather in Goa? I guess better than here | ||
| Chirag | definitely :D | ||
| its still frozen? | 20:01 | ||
| rurban | Not in Texas, just foggy | 20:02 | |
| Chirag | cool! :D | ||
| rurban | I'm working 8:30 to 5pm usually, but can be reached at home also on irc, if I'm not at the movies. I watch a lot of Hindi movies also | 20:04 | |
| Chirag | oh really! which ones have u watched? | ||
| rurban | So easiest is communication in your evening | ||
| About Goa only the bad sex comedies recently | 20:05 | ||
| www.rinema.com/ReiniUrban/ | |||
| Satyagraha was horrible, Ramleela fantastic | 20:06 | ||
| Dedh Ishqiya mediocre | |||
| My favorites were Paan Singh Tomar and Gangs of Wasseypur lately | 20:07 | ||
| Chirag | I could suggest you better movies! | ||
| rurban | Yes, please | ||
| Chirag | My favorite is Rang De Basanti | ||
| Lagey Raho Munnabhai | 20:08 | ||
| rurban | Oh, I heard about that. Missed it | ||
| Chirag | Bollywood's not coming up with good movies of late | ||
| rurban | My favorite was Mani Ratnam, but then he moved back to Chennai and got worse | 20:09 | |
| Chirag | Guru? | 20:10 | |
| rurban | Anurag Kashyap is pretty interesting but he wrote Yuva | 20:11 | |
| Chirag | havent seen | ||
| rurban | I forgot about Guru, I guess It didn't impress me | ||
| Yuva is Amores Perros in India | |||
| Chirag | hmm.. | 20:12 | |
| I have started watching movies that have unique storylines | |||
| rurban | 3 parallel, pretty predictable stories, with a new style of camera | ||
| Chirag | Wow! | 20:13 | |
| U have more idea about Hindi movies than me! LOL | |||
| rurban | The latest Urdu movie was pretty unusual, forgot its name | ||
| I'm working as film-critic so I should know :) | |||
| Chirag | Oh nice :) | 20:14 | |
| I saw this movie recently - The man from earth | |||
| btw when I was going through Parrot's API last summer, I remember the implementation was in C++ | 20:17 | ||
| rurban | nope, only pure C. I just looks like C++ | ||
| Chirag | That was for providing OpenCV bindings | 20:18 | |
| openCL* | |||
| Coke is reminded that Chennai Express is in his netflix queue. | 20:20 | ||
| rurban | look in pmc/src/*.c at the generated C files, set a breakpoint there and look at the backtrace. we have enough testcases in t/pmc/ | ||
| Chirag | ok.. will do | ||
| rurban | I'll gonna find a good testcase for you. My current one, imccompiler is not good | ||
| nci is for "Native Call Interface", the ffi | 20:21 | ||
| this is the problem | |||
| Chirag | I will have to read up on that.. | 20:22 | |
| I have never worked on Perl before | |||
| Just Python and C .. and some C++ and a little Java | |||
| rurban | ok, the other compiler for our ops is written in nqp, which is a simplified perl6 | 20:23 | |
| Chirag | ok.. | 20:24 | |
| rurban | So our sigs are declared like "PiSSoIpSnSoIpSnPoIp" | ||
| Pointer, integer, String, .... | |||
| parsed via Parrot_pcc_fill_params_from_c_args(interp, _call_object, "PiSSoIpSnSoIpSnPoIp", ... | 20:25 | ||
| which parses the string "PiSSoIpSnSoIpSnPoIp" and fills the given args dependent on the type of the string, tyhe sig | |||
| But we can precompile all that into generated C code for the pmc, depending on the sig. | 20:26 | ||
| Chirag | Parrot's Documentation will help me understand I hope.. | ||
| rurban | A PMC is a generic parrot object. Almost everything is an object | 20:27 | |
| Chirag | like Python | ||
| rurban | yes, we have 3 native objects (int, num, string) and then PMC for the rest | 20:28 | |
| the methods for all pmc's are specified in src/pmc/*.pmc | |||
| and the pmc compiler emits src/pmc/*.c files and include/pmc/pmc_*.h headers | |||
| Chirag | ok.. I will look into it | 20:29 | |
| rurban | it's actually easier and more readable than python | ||
| Chirag | nice! | ||
| rurban | and it should be more efficient than python also | ||
| Chirag | yes python is slow | ||
| rurban | I'll let you sleep, because it's pretty late, see ya tomorrow | 20:30 | |
| Chirag | sure! | ||
| What time does everine come online? | |||
| rurban | european morning time | ||
| until US evening time | |||
| there are no asians here | 20:31 | ||
| but most of the europeans are passive nowadays, mostly US folks | |||
| Chirag | 2.30 pm IST - 6.00 am IST | ||
| roughly | 20:32 | ||
| rurban | yes, your afternoon and evening | ||
| Chirag | i ll idle on this channel and ping | ||
| rurban | the work could be done in 2-3 days for a very experienced parrot programmer, but for a newcomer it will need a lot of time to get into it. | 20:33 | |
| bye | |||
| Chirag | ok.. I will try my best | ||
| :) | |||
| bye! | |||
| rurban | no prob. failure is also no problem | ||
| dalek | rrot: 90e3128 | rurban++ | compilers/imcc/imc (3 files): [imcc] bison 3 fixes and more [gh #1031] Fix #1031 compat with bison 3: remove deprecated and unneeded YYID macro note one remaining shift/reduce conflict in arglist: STRINGC ADV_ARROW var at ADV_ARROW change deprecated %pure_parser to new %pure-parser regenerate compilers/imcc/imcparser.[ch] Note: repro with perl Configure.pl --maintainer |
21:08 | |
| rurban | I'm creating now the generated include/*.pasm files with vim and emacs read-only coda and proper modes | 21:33 | |
| dalek | rrot/rurban/h2inc-modes-gh1032: 0cad250 | rurban++ | / (4 files): wip: add imcc flags to the refactored API and use it (again) |
21:40 | |
| rrot/rurban/h2inc-modes-gh1032: 45d8e31 | rurban++ | / (2 files): [cage] Provide coda for generated include/*.pasm files #1032 |
|||
| rurban | oops, 0cad250 does not belong here. | 21:41 | |
| dalek | rrot/rurban/imccflags: ea57c49 | rurban++ | / (4 files): [imcc] add old imcc optim flags as Parrot_trace_flags |
21:44 | |
|
21:45
travis-ci joined
|
|||
| travis-ci | [travis-ci] parrot/parrot#947 (master - 90e3128 : Reini Urban): The build is still failing. | 21:45 | |
| [travis-ci] Change view : github.com/parrot/parrot/compare/d...e3128edec0 | |||
| [travis-ci] Build details : travis-ci.org/parrot/parrot/builds/19529920 | |||
|
21:45
travis-ci left
|
|||
| dalek | rrot/rurban/h2inc-modes-gh1032: 42184f2 | rurban++ | / (2 files): [cage] Provide coda for generated include/*.pasm files #1032 |
21:47 | |
|
21:48
awwaiid joined
22:35
rurban1 joined
23:35
kid51 joined
|
|||