Camelia, the Perl 6 bug

IRC log for #parrot, 2011-03-31

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:18 Tene whiteknight: I should be able to be a backup gsoc mentor, I think.  Can you tell me what's involved there?
00:27 whiteknight Tene: awesome. we dont have complete lists yet
00:27 davidfetter left #parrot
00:28 theory left #parrot
00:29 Tene whiteknight: I suppose that reliability is important for mentor volunteers.  I haven't been very reliable in the past.  I can still aid gsoc students without being listed as a mentor.
00:29 whiteknight Tene: Don't worry about it, We have a robust community. Every little bit helps
00:31 whiteknight You're reliable enough :)
00:31 Tene whiteknight: 'k, just let me know what would be most appropriate for me to do there, and whether a potentially-unreliable mentor is better than one fewer mentor.
00:31 Tene 'k
00:31 Tene I'll probably fill out that form tonight, then.  I'm still dealing with the tail end of moving to a new apartment.
00:32 whiteknight We've been looking at new apartments ourselves. Probably going to be dealing with that in the next few months
00:33 kid51_at_dinner is now known as kid51
00:33 Tene whiteknight: the major lesson for me from this move is that I wish I had realized beforehand that I could pay someone else to move my furniture for me.
00:34 whiteknight We promised that to ourselves last time
00:34 whiteknight otherwise you call in too many favors, and ruin your back
00:34 whiteknight apts are not designed for moving large furniture into and out of
00:34 Tene srsly, not worth it
00:35 cotto I like helping friends move on occasion.
00:35 Tene I'm too used to being completely broke; I'm still learning that I'm capable of paying for things now that I have a reasonable income.
00:43 dalek parrot/opsc_llvm: 2fa2a5f | bacek++ | / (85 files):
00:43 dalek parrot/opsc_llvm: Merge branch 'master' into opsc_llvm
00:43 dalek parrot/opsc_llvm:
00:43 dalek parrot/opsc_llvm: Conflicts:
00:43 dalek parrot/opsc_llvm: config/auto/llvm.pm
00:43 dalek parrot/opsc_llvm: t/steps/auto/llvm-01.t
00:43 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/2fa2a5f4ec
00:57 woosley joined #parrot
01:02 theory joined #parrot
01:09 whiteknight Tene: How do you pronounce "Tene"?
01:10 whiteknight In my head it always sounds like "Teenie", but I hear pmichaud pronounce it "Tinay"
01:10 whiteknight or, I heard him say it that way once
01:10 cotto does nqp support MAIN these days?
01:11 whiteknight cotto: not the last time I checked
01:11 cotto hacks to the future!
01:21 whiteknight left #parrot
01:24 bacek_at_work looks like whiteknight is slightly scared about future and decided to leave :)
01:30 ShaneC left #parrot
01:30 dalek parrot/opsc_llvm: aff726f | jkeenan++ | t/steps/auto/llvm-01.t:
01:30 dalek parrot/opsc_llvm: Undeleted git-conflict marker caused Perl syntax error; delete it.
01:30 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/aff726f504
01:31 dmalcolm left #parrot
01:41 bubaflub left #parrot
01:46 ShaneC joined #parrot
01:53 kid51 bacek_at_work: Currently I'm testing the opsc_llvm branch on my iBook, where LLVM has never been installed.
01:54 kid51 bacek_at_work:  I noticed that during 'make', the build is running lots of commands which suggest that LLVM libraries are being used/built.
01:54 kid51 But why would we do that if LLVM is not found on a box?
01:56 bubaflub joined #parrot
01:56 nopaste "kid51" at 192.168.1.3 pasted "opsc_llvm branch on box with no LLVM: why is 'make' running these commands?" (12 lines) at http://nopaste.snit.ch/39093
02:05 benabik joined #parrot
02:07 dalek parrot: 5792dce | jkeenan++ | Configure.pl:
02:07 dalek parrot: Add entry for '--cxx' to help.
02:07 dalek parrot: review: https://github.com/parrot/parrot/commit/5792dcec8b
02:12 bubaflub left #parrot
02:26 sorear I tend toards "tener" -r
02:27 sorear using Spanish phonology
02:29 cotto nqp is fairly nice once I get over the hump of not remembering anything about it
02:37 nopaste "kid51" at 192.168.1.3 pasted "opsc_llvm branch: patch for makefile template: no need to build llvm if it's not there" (13 lines) at http://nopaste.snit.ch/39094
02:38 kid51 msg bacek http://nopaste.snit.ch/39094 is for your branch
02:38 aloha OK. I'll deliver the message.
02:40 dalek parrot: f1e7b7e | jkeenan++ | config/gen/makefiles/root.in:
02:40 dalek parrot: Adjust makefile template for better indentation in 'all' target in Makefile.
02:40 dalek parrot: review: https://github.com/parrot/parrot/commit/f1e7b7e814
02:41 dalek parrot/opsc_llvm: f04d23a | jkeenan++ | config/gen/makefiles/root.in:
02:41 dalek parrot/opsc_llvm: Same commit as in master.
02:41 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/f04d23a88a
02:41 benabik left #parrot
02:44 benabik joined #parrot
02:47 kid51 left #parrot
03:34 hudnix left #parrot
03:41 soh_cah_toa left #parrot
03:49 bacek_at_work msg kid51 nopaste looks allright. I'll commit it tonight
03:49 aloha OK. I'll deliver the message.
04:06 shell joined #parrot
04:16 cotto what does <xdigit> match?
04:17 cotto I thought it was equivalent to 0-9a-fA-F
04:19 bacek_at_work cotto, I think so
04:20 nopaste "cotto" at 192.168.1.3 pasted "xdigit?" (12 lines) at http://nopaste.snit.ch/39095
04:21 cotto bacek_at_work, am I missing something there?
04:21 bacek_at_work "0x"
04:22 bacek_at_work hmmm
04:22 bacek_at_work it's "rule"
04:23 cotto ok
04:29 bacek_at_work yes, it works with "token"
04:30 benabik rule adds <.ws>
04:30 benabik So that's equiv to token { '0x' <.ws> <xdigit>+ }
04:30 benabik And ws has <!ww> in it, so won't match between x and A.
04:30 cotto ah
04:31 benabik I think you want it to be a token, as bacek said
05:05 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#13513) fulltest) at 3_2_0-108-gf1e7b7e - Ubuntu 10.10 i386 (g++-4.5)
05:08 mikehh what is the purpose of the --cxx in Configure.pl
05:10 mikehh it was taken out then put back in - I can't see what difference it makes
05:10 cotto it's wanted for llvm
05:10 benabik mikehh: It's to specify a C++ compiler, which will be needed for LLVM
05:12 dalek parrot: b48aa06 | petdance++ | src/spf_render.c:
05:12 dalek parrot: Cleaned up canonicalize_exponent:
05:12 dalek parrot: * Removed unused interp arg
05:12 dalek parrot: * consted SpfInfo pointer arg
05:12 dalek parrot: * Reformatted POD docs
05:12 dalek parrot: * Normalized variable declarations.
05:12 dalek parrot: review: https://github.com/parrot/parrot/commit/b48aa06e45
05:12 sorear mikehh: it was taken out because it wasn't being used at the time
05:13 sorear mikehh: there was a period of time when Parrot had no C++ components
05:13 sorear in hindsight it is obvious that this condition would not last
05:13 mikehh I already use --cc=g++ --link=g++ --ld=g++
05:14 bacek left #parrot
05:14 mikehh so is it just for LLVM
05:14 sorear mikehh: now? yes
05:15 sorear but when ICU comes back, we'll need it for that too
05:16 mikehh sorear: comes back? as opposed to using libraries?
05:16 sorear mikehh: ?
05:17 sorear for any library with a C++-specific API, we need --cxx
05:17 mikehh sorear: ICU comes back? as opposed to using libicu or something?
05:18 sorear mikehh: yeah - talk to jnthn if you want
05:19 * sorear wonders what effect 6model will have on jnthn's political power
05:20 mikehh sorear:  hey I thought ICU was pretty standard , or is it licencing issues
05:23 mikehh or is it a windoze problem?
05:25 sorear mikehh: icu is not very common, so most users don't have it
05:25 sorear mikehh: there are also some issues on windows that make it very hard to install
05:28 mikehh sorear: I mostly use Ubuntu/Kubuntu, so I generally install it when I set it up - new beta in a couple of days ;-}
05:31 theory left #parrot
05:48 shell left #parrot
06:35 bacek joined #parrot
07:12 wagle left #parrot
07:25 lucian joined #parrot
07:26 lucian left #parrot
07:47 perlite left #parrot
07:47 perlite joined #parrot
08:34 wagle joined #parrot
08:36 benabik left #parrot
08:41 clunker3__ joined #parrot
09:11 cosimo left #parrot
09:26 bacek left #parrot
09:39 ShaneC left #parrot
09:42 Hunger joined #parrot
09:56 contingencyplan left #parrot
09:58 woosley left #parrot
10:03 lucian joined #parrot
10:36 prank left #parrot
10:37 prank joined #parrot
11:06 cotto left #parrot
11:09 shell joined #parrot
11:09 shell left #parrot
11:10 shell joined #parrot
11:19 wagle left #parrot
11:40 jsut joined #parrot
11:45 jsut_ left #parrot
11:47 kid51 joined #parrot
11:49 Eduardow left #parrot
11:51 woosley joined #parrot
11:54 bacek joined #parrot
11:55 Patterner left #parrot
11:55 Psyche^ joined #parrot
11:55 Psyche^ is now known as Patterner
11:55 lucian_ joined #parrot
11:59 lucian_ left #parrot
11:59 wagle joined #parrot
12:08 hudnix joined #parrot
12:12 dalek parrot/opsc_llvm: 403668a | bacek++ | runtime/parrot/library/LLVM/Value.pm:
12:12 dalek parrot/opsc_llvm: Rename few generated methods for consistency with Builder.
12:12 dalek parrot/opsc_llvm: review: https://github.com/parrot/parrot/commit/403668a27b
12:13 bubaflub joined #parrot
12:23 kid51 left #parrot
12:34 ambs joined #parrot
12:34 whiteknight joined #parrot
12:36 jsut_ joined #parrot
12:40 jsut left #parrot
12:51 whiteknight good morning, #parrot
12:53 lucian whiteknight: good day
12:53 whiteknight hello lucian, how are you doing?
12:53 lucian had my last meeting with my supervisor
12:53 lucian before i switch over to a new one
12:54 whiteknight oh cool. Does a new supervisor bring new and cool research opportunities?
12:54 lucian not really, the old one is just moving to germany
12:54 lucian and it's a BSc anyway, i'll finish in 1-2 months
12:56 whiteknight oh, okay
12:58 lucian bah, i have a lecture i'd rather not go to
12:58 lucian but i probably should
12:58 whiteknight which lecture?
12:58 lucian advanced computer graphics
12:58 lucian i usually like it, but the lecturer isn't very good, and today's subject is boring
12:58 whiteknight oh, fun
12:59 lucian i got possibly the most wide-and-thin spread set of modules this year
12:59 lucian graphics, games, computer language theory, real-time programming, mobile apps UI design
13:00 * lucian has to go. bye!
13:00 lucian left #parrot
13:07 clunker3__ is now known as clunker3
13:11 whiteknight I wish I had taken some of those kinds of classes
13:12 whiteknight My school didn't offer anything like a good language theory class, which is something that I had always wanted to take
13:12 whiteknight Technically they did offer it, but only every few years when demand was high
13:17 whiteknight msg cotto Found this page today on the interwebs. Interesting read about concurrency. A little bit tangential to the M0, but might help to inform some future decisions: http://www.cs.nmsu.edu/ALP/2011/03/concurrent-prog​ramming-constructs-and-first-class-logic-engines/
13:17 aloha OK. I'll deliver the message.
13:29 shell hello,whiteknight
13:29 whiteknight hello shell
13:29 whiteknight how are you today?
13:29 shell i've been trying winxed
13:30 shell and i have problem building it
13:30 whiteknight really? What's the problem?
13:30 shell it said "could not found config.pbc"
13:30 shell i was "making"
13:30 whiteknight did you install Parrot?
13:30 shell yes
13:31 whiteknight and it still says it can't find config.pbc?
13:31 shell right
13:31 whiteknight actually, was it looking for config.pbc or config.fpmc?
13:32 shell ok,i manually find the config.pbc
13:33 shell how can i tell it to winxed
13:34 whiteknight I don't know. Winxed should just be able to find it
13:34 whiteknight this is preventing you from building it at all?
13:34 shell yes
13:34 shell not
13:35 shell sorry
13:35 whiteknight ...?
13:35 shell it can produce some pir
13:35 whiteknight so you have winxed, but you can't use it?
13:36 shell right
13:36 whiteknight weird. You might need to ask NotFound what the problem is. He made Wixned
13:36 whiteknight Let's try something different. Get plumage, and use that to install winxed
13:37 whiteknight plumage is the parrot package manager, it will automatically build, test, and install winxed for you
13:37 whiteknight https://gist.github.com/894484
13:37 shell thanks
13:38 whiteknight shell: Have you drafted a proposal yet?
13:41 shell i've just go through the winxed and antlr
13:41 whiteknight okay
13:42 shell i'll do that soon
13:42 shell and i brought a proxy to access git and irc
13:42 whiteknight oh, nice
13:43 mtk joined #parrot
13:43 whiteknight Applications are due April 8th, so you have time
13:43 whiteknight but if you make drafts, we can help you with them
13:43 shell since my school doesn't allow me to use socket to access foreign sites
13:43 shell thanks
13:43 whiteknight shell: Is that going to be a problem during the summer
13:43 whiteknight ?
13:43 shell not at all
13:44 whiteknight okay
13:44 shell when would you be online
13:44 shell since we are in different timezone
13:44 whiteknight I don't know who your mentor would be
13:45 whiteknight we have developers all over the world
13:45 shell i got it
13:46 Coke I assume that winxed build problems would stem from installed parrot not being in path.
13:46 whiteknight Coke: that was my first thought
13:46 whiteknight I only asked if Parrot was installed, not whether it was installed "properly"
13:47 mtk left #parrot
13:47 shell i wonder how it found the library
13:47 mtk joined #parrot
13:47 shell if it does not find it,i'll do it manually
13:48 mtk left #parrot
13:48 mtk joined #parrot
13:50 whiteknight shell: did you get plumage?
13:50 whiteknight plumage makes everything easier
13:50 shell i am pulling
13:50 whiteknight awesome
13:52 shell what the hell.when i am installing plumage,parrot also said that it could find some pbc
13:53 shell and i found it by myself
13:53 clunker9_ joined #parrot
13:53 whiteknight shell: Sounds like your parrot install is broken
13:53 whiteknight shell: What OS are you on?
13:53 shell it must be
13:54 shell windows
13:54 whiteknight make sure your librarys are on your PATH
13:55 shell i did that
13:55 bubaflub shell: any spaces in your path?
13:55 bubaflub i know we've had problems with that and windows before
13:56 shell i think it is.
13:56 clunker3 left #parrot
13:57 mtk left #parrot
13:58 mtk joined #parrot
13:58 bubaflub shell: try putting parrot in somewhere like C:\Parrot just to be sure
13:58 Coke can you run "parrot -?" from the command line?
13:59 mtk left #parrot
13:59 mtk joined #parrot
14:07 NotFound winxed bootstrap prooduces pir using stage 0 compiler without using parrot. The presence of generated pir doesn't mean it was able to use parrot.
14:08 NotFound That diagnostic message probably comes from pbc_to_exe
14:09 shell it still could not find the libaary,i fail to build plumage
14:11 whiteknight shell: Okay, you have a problem with Parrot
14:11 whiteknight somehow it isn't being installed or handled correctly
14:11 NotFound shell: How do you installed parrot?
14:11 shell it is a msi file
14:12 shell it installed automatically
14:14 NotFound shell: execute this: parrot_config libdir
14:20 shell it said "no such key"
14:22 NotFound Strange.
14:22 moritz shell: what does  parrot_config VERSION   say?
14:23 shell 3.2.0
14:28 shell ok,it ran sucessfully
14:28 shell but it is still a problem of parrot installer
14:28 clunker9_ is now known as clunker3
14:44 theory joined #parrot
14:56 shell well it tell me "could find lreadline"
14:58 cotto joined #parrot
14:58 shell how to get the readline library
15:00 bubaflub shell: on windows? i'm not sure.
15:07 whiteknight left #parrot
15:23 mj41 joined #parrot
15:25 dmalcolm joined #parrot
15:28 Eduardow joined #parrot
15:32 dalek winxed: r911 | NotFound++ | trunk/winxedst1.winxed:
15:32 dalek winxed: fix compile time concatenation of literal strings of mixed quoting types in
15:32 dalek winxed: stage 1
15:32 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=911
15:37 shell left #parrot
15:40 dalek parrot/m0-spec: c9574a5 | cotto++ | m0asm.pm:
15:40 dalek parrot/m0-spec: add a quarter-done prototype m0 assembler
15:40 dalek parrot/m0-spec: review: https://github.com/parrot/parrot/commit/c9574a5c73
15:42 mj41 left #parrot
15:49 plobsing left #parrot
15:53 dalek winxed: r912 | NotFound++ | trunk/ (3 files):
15:53 dalek winxed: rearrange string literal tests and examples
15:53 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=912
15:59 cotto_work ~~
16:00 contingencyplan joined #parrot
16:01 atrodo cotto_work> pong
16:02 woosley left #parrot
16:02 cotto_work atrodo, are you interested in working on an M0 assembler, disassembler or interpreter based on the current spec?
16:02 cotto_work well, disassembler or interpreter ;)
16:02 atrodo I've been thinking about that
16:03 atrodo I noticed that too ;)
16:05 atrodo I was also thinking through the differences between the spec and my prototype, since obviously I'd like to get the two unified
16:07 atrodo unfortunately, I haven't had time today yet to figure out what that would take
16:07 cotto_work atrodo: your're in the uk, right?
16:08 atrodo Close, Ohio
16:08 atrodo I'm curious why you thought uk
16:08 cotto_work I thought there was a bigger timezone mismatch.
16:14 wknight-phone joined #parrot
16:20 davidfetter joined #parrot
16:44 gg411 joined #parrot
16:45 davidfetter left #parrot
16:45 gg411 left #parrot
16:58 ShaneC joined #parrot
17:05 dodathome joined #parrot
17:07 ShaneC left #parrot
17:22 mj41 joined #parrot
17:22 wknight-phone left #parrot
17:23 plobsing joined #parrot
17:29 dalek parrot/m0-spec: 4721ce7 | cotto++ | hello.0bc:
17:29 dalek parrot/m0-spec: add a possibly-valid example m0 file to test the assembler on
17:30 dalek parrot/m0-spec: review: https://github.com/parrot/parrot/commit/4721ce7d1f
17:30 dalek parrot/m0-spec: d90c4c1 | cotto++ | m0asm.pm:
17:30 dalek parrot/m0-spec: minor fixes for hex->bits conversion
17:30 dalek parrot/m0-spec: review: https://github.com/parrot/parrot/commit/d90c4c11ce
17:41 ShaneC joined #parrot
17:43 wknight-phone joined #parrot
17:44 ligne joined #parrot
17:51 wknight-phone left #parrot
17:53 dalek winxed: r913 | NotFound++ | trunk/winxedst1.winxed:
17:53 dalek winxed: use a stringbuilder instead of hand-crafted code for argument substitution in
17:53 dalek winxed: predefs
17:53 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=913
17:55 plobsing left #parrot
17:58 dalek winxed: r914 | NotFound++ | trunk/pir/winxed_compiler.pir:
17:58 dalek winxed: update installable compiler
17:58 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=914
17:59 ligne left #parrot
18:00 cotto_work How easy is winxed to pick up?
18:06 NotFound cotto_work: for me very easy, but I'm a special case ;)
18:06 cotto_work You're very special.
18:06 NotFound cotto_work: with a minimal familiarity with javascript syntax and pir knowledge, it should be easy.
18:07 NotFound For a lot of usages is just pir on steroids.
18:11 zby_home left #parrot
18:12 lucian joined #parrot
18:16 plobsing joined #parrot
18:29 benabik joined #parrot
18:30 dalek winxed: r915 | NotFound++ | trunk/README:
18:30 dalek winxed: Very delayed README upadte
18:30 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=915
18:35 ambs left #parrot
18:36 ambs joined #parrot
18:41 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#13559) fulltest) at 3_2_0-109-gb48aa06 - Ubuntu 10.10 i386 (gcc-4.5 --optimize --gc-gms)
18:43 mj41 left #parrot
18:52 mikehh winxed (r015) on parrot  3_2_0-109-gb48aa06 - make all/test/test1/test2 ok - fly and opengl still work
18:52 mikehh bah that should be r915
18:54 NotFound I'll be surprised if opengl worked on r15
18:54 mikehh NotFound: :-}
18:57 atrodo cotto_work> ping
19:18 hercynium joined #parrot
19:19 plobsing who should I light a fire under to get better thread-safety?
19:21 benabik plobsing: Wouldn't thread-safety involve parrot having threads?
19:22 plobsing benabik: not necessarily
19:22 plobsing being thread-friendly and exposing threads are 2 different levels
19:22 plobsing the first is usually required before the second
19:23 benabik Ah.
19:23 plobsing my situation is that I'd like to provide ØMQ bindings for parrot, but it calls the free-message-struct callback from different threads (it manages its own IO thread pools)
19:26 plobsing we have "asynchronous callbacks", which would reasonably be threadsafe, but the scheduler operations aren't thread-safe
19:26 lucian thread-safety is pretty critical, i'd agree
19:42 ambs left #parrot
19:50 bacek Good morning, humans
19:51 dalek parrot/jit_prototype: 0052c45 | bacek++ | compilers/opsc/src/Ops/JIT.pm:
19:51 dalek parrot/jit_prototype: Remove accidentally committed _dumper.
19:51 dalek parrot/jit_prototype: review: https://github.com/parrot/parrot/commit/0052c453e7
19:54 bacek plobsing, ping
19:54 tadzik good evening bacek
19:54 bacek tadzik, aloha
19:54 plobsing bacek: pong
19:54 bacek plobsing, what is "the best way" to free resources allocated in nci-called function?
19:55 bacek e.g. wrap "malloc" and "free" using nci
19:55 plobsing bacek: which resources are being allocated? if the native code does the allocation, free it using native code
19:56 plobsing I wouldn't wrap malloc if you can get away with StructView's allocation (and auto-cleanup)
19:56 bacek plobsing, do we have example how to use StructView?
19:56 plobsing t/pmc/structview.t has some basic examples
19:57 bacek simple case: "char *foo()" returns new allocated c-string. How to call "free" on it?
19:57 plobsing parrot-zeromq also has a number of usages
19:58 plobsing bacek: if you don't control the allocation, you almost certainly need to wrap free(). there's really no way around that.
19:59 bacek plobsing, yes, I don't mind to "wrap free". I just don't want to write "C" for it.
20:00 plobsing I suspect you are using "t" for the signature on the foo() wrapper. That won't work. "t" doesn't handle resource lifecycle very flexibly, which is why it is deprecated.
20:00 plobsing you need to use "p" in stead
20:01 bacek yes, "t" looses original string.
20:01 bacek "p" will return (Un)ManagedStruct
20:01 bacek There is no way (afaik) to attach custom .destroy to UMS
20:02 plobsing which you can (a) convert into a PtrObj with an appropriate destructor (b) call free() on it explicitly when you're finished
20:03 cotto_work atrodo: pong
20:03 bacek plobsing, any samples/examples how to do it?
20:04 plobsing bacek: how to do which/
20:04 plobsing ?
20:04 bacek convert to PtrObj
20:05 plobsing not yet, SFAIK. you can do it with PtrObj.init_pmc().
20:05 atrodo cotto_work> If I can get my vision for my prototype and m0-spec close enough, I would like to make it a base for the m0-spec interpreter
20:06 cotto_work atrodo: I'm game
20:06 atrodo cotto_work> I'm drawing out the differences and my thoughts now
20:06 cotto_work excellent
20:06 plobsing bacek: destructors are set/get using the destroy_func() method
20:06 bacek plobsing, ok. Thanks.
20:14 ligne joined #parrot
20:15 mj41 joined #parrot
20:17 bubaflub plobsing: can i pick your brain about some NCI stuff?
20:20 plobsing sure
20:22 bubaflub plobsing: ok, here's what i'm thinking of doing to get bindings to GMP
20:22 bubaflub 1) create a definition file for ncidef2pir which maps function names to signatures and return types
20:23 bubaflub 2) create a PMC which wraps a StructView to represent a GMP integer type
20:23 bubaflub 3) write some convenience functions for that PMC wrapper class for easy initialization and stringifcation
20:24 benabik Quick Q: Does anyone know if POST is used for anything other than PIR generation?  I know there was a tree-optimizer GSoC before, but don't know what trees it works on.
20:24 bubaflub 4) run the ncidef2pir script on my definition file to generate a boat load of PIR
20:24 plobsing I would support this strategy. I've done exactly that for parrot-zeromq
20:25 bubaflub so is this the bare minimum i need to call GMP functions?
20:26 plobsing bubaflub: the bare-minimum is just the raw bindings. an object layer over top is nice, but not strictly necessary.
20:26 plobsing raw bindings should probably be done in a day at the most
20:26 bubaflub plobsing: ok, yeah.  i can write a script to scrape the headers and the docs to generate it
20:27 bubaflub plobsing: so some basic examples - mpz_init() returns void and takes a mpz_t - a GMP integer
20:27 bubaflub so the signature is vp
20:27 plobsing benabik: IIUC, POST is fairly specialized to Parrot. it is likely only suitable for generating PIR and PBC. PAST on the otherhand, is more high-level and may be amenable to other usages
20:28 bubaflub and i make an appropriate StructView so parrot knows how much space to allocate
20:28 bubaflub and then pass that StructView into the generated function for mpz_init
20:29 rohit_nsit08 joined #parrot
20:29 plobsing bubaflub: StructView is the representation, not the instance. You use it to allocate instances (which come out as PtrObj instances with destructors that manage cleanup automatically).
20:29 rohit_nsit08 hello #parrot
20:29 plobsing hi rohit_nsit08
20:29 rohit_nsit08 plobsing: hi
20:29 rohit_nsit08 whiteknight: hi
20:29 dukeleto benabik: POST can be use for other things
20:30 rohit_nsit08 dukeleto: hi
20:30 dukeleto rohit_nsit08: howdy
20:30 bubaflub plobsing: ok, so when i setup the StructVIew i'm just telling Parrot how this C thing is structured, the actual instance is a PtrObj
20:30 bubaflub and then i pass that PtrObj into the PIR generated function
20:30 dalek parrot/jit_prototype: ba48854 | bacek++ | / (3 files):
20:30 dalek parrot/jit_prototype: Handle some variable declarations
20:30 dalek parrot/jit_prototype: review: https://github.com/parrot/parrot/commit/ba48854890
20:30 bubaflub and it loads the library and calls the function
20:30 benabik I'm digging around a bit for the bytecode generation GSoC task.  Trying to figure out if tree-walking, smarter nodes or a re-write (ew) is better.
20:31 rohit_nsit08 dukeleto: perfect :-) , has student started submitting their proposals on google?
20:31 dukeleto rohit_nsit08: not sure, been busy and haven't looked yet :)
20:31 dukeleto benabik: what do you *want* to work on?
20:31 benabik dukeleto: Something useful.  :-D
20:31 rohit_nsit08 dukeleto: ok, i am making some final additions, and planning to submit tomorrow.
20:32 dukeleto bacek: how does POST interact with the JIT prototype that you are hacking on
20:32 rohit_nsit08 dukeleto: or shall i wait ?
20:32 bacek dukeleto, no interaction at all. jit_prototype works on PBC
20:32 benabik dukeleto: I'm planning on submitting two proposals, at least.
20:33 lucian unless you're doing tracing, i don't see what interactions there would be
20:34 plobsing bubaflub: that sounds about right
20:35 bubaflub plobsing: great, then the PMC itself just wraps this PtrObj and makes things pretty
20:35 dukeleto benabik: awesome!
20:36 dukeleto bacek: just wondering
20:36 bubaflub plobsing: i can override addition and multiplication and things like that, right?
20:36 dukeleto i don't think anybody has a gsoc proposal relating to Cardinal yet
20:37 plobsing bubaflub: yes, parrot supports HLL overloading of many vtables, including math ops (those are the easiest overrides)
20:37 dukeleto it might be somewhat blocked on 6model, tho
20:37 Tene dukeleto: It's not clear to me that Cardinal is in an appropriate place for gsoc work.
20:37 Tene I would not recommend work on cardinal before 6model lands, and it's not clear that 6model work is appropriate for gsoc
20:37 Tene dukeleto: I'm mostly done dealing with moving to a new apartment, so I'll have time to start on cardinal 6model again this weekend.
20:38 bubaflub plobsing: great, as well as comparison functions and what not.  hmmmm what else would go in convenience functions?
20:38 cotto_work seen whiteknight
20:38 clunker3 Whiteknight was last seen on #parrot 2 years, 9 days, 1 hour, 54 minutes and 44 seconds ago, saying: now the real test is to see if I can use it for my own needs
20:38 aloha whiteknight was last seen in #parrot 6 hours 27 mins ago saying "somehow it isn't being installed or handled correctly".
20:38 clunker3 whiteknight was last seen on #parrot 6 hours, 27 minutes and 22 seconds ago, saying: somehow it isn't being installed or handled correctly
20:38 plobsing bubaflub: I recomend you take a look at parrot-zeromq. It does exactly what you are proposing, but for libzmq.
20:38 clunker3 was kicked by plobsing: clunker3
20:38 cotto_work plobsing++
20:39 cotto_work ban might have been better
20:41 bubaflub plobsing: great.  i'm looking at that and docs/book/pir/ch07_objects.pod.  seems perfect for me.
20:43 atrodo i thought moritz got the owner to stop clunker3 from joining
20:43 cotto_work atrodo: he was going to.  He might not have caught him yet.
20:43 atrodo must have misunderstood
20:45 plobsing bubaflub: I would avoid PIR if I could. Winxed makes static class definitions simple and syntactically beautiful.
20:45 plobsing and you can define vtable overrides just as easily
20:45 bubaflub plobsing: hmmm, i'm definitely more familiar with PIR but i suppose as long as it compiles down to PBC that's fine
20:45 bubaflub plobsing: if i'm going the winxed route should i also be using rosella for my tests?
20:46 rohit_nsit08 left #parrot
20:46 plobsing you should be able to use rosella from any language
20:46 plobsing PIR included
20:46 plobsing winxed included
20:46 bubaflub of course.
20:46 dodathome left #parrot
20:47 bubaflub i guess my question is if rosella is easier or simplier than PIR's Test::More
20:47 plobsing I cannot speak to its suitability or ease of use, not having used it.
20:48 ambs joined #parrot
20:48 bubaflub ok, i'll ask whiteknight to sell me on it
20:55 dalek winxed: r916 | NotFound++ | trunk/winxedst0.cpp:
20:55 dalek winxed: fix a few hard tabs and improve indentation of generated code in stage 0
20:55 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=916
20:58 fperrad left #parrot
20:58 lucian jnthn_: ping
21:03 lucian 6model: ping
21:03 cotto_work 6model is not that advanced.  It won't respond.
21:03 plobsing it isn't sentient... yet
21:03 Tene lucian: you have questions?
21:04 Tene lucian: any chance you could be around to chat about it in... 5 hours?
21:04 lucian sure
21:04 Tene lucian: I'm at work right now, but I should be home and available by then.  I don't know wha ttime zone you're in.
21:05 lucian wait, 5h is too much
21:05 lucian brb
21:06 lucian cotto_work: plobsing: it was my not-so-subtle ping to whoever knows about it
21:08 lucian Tene: it's 10pm here right now. 3am is a bit too late
21:08 Tene lucian: Ah, what time do you wake up?  I can go much later.
21:08 Tene 14:00 here right now.
21:09 ShaneC left #parrot
21:09 ShaneC joined #parrot
21:09 lucian Tene: well, i have lectures. i should have plenty of time after 11-12
21:11 Tene lucian: Here's a better idea.  Can you email me and jnthn your questions, and I'll write up a response when I get home?
21:11 lucian Tene: sure. i was going to explain roughly how python's object system works, if you don't know already
21:11 Tene That would be great.
21:12 dalek parrot/jit_prototype: 2b43930 | bacek++ | compilers/opsc/src/Ops/JIT.pm:
21:12 dalek parrot/jit_prototype: More register/constants handling.
21:12 dalek parrot/jit_prototype: review: https://github.com/parrot/parrot/commit/2b43930523
21:12 dalek parrot/jit_prototype: ac86bda | bacek++ | compilers/opsc/src/Ops/JIT.pm:
21:12 dalek parrot/jit_prototype: Vivify declared variables.
21:12 dalek parrot/jit_prototype: review: https://github.com/parrot/parrot/commit/ac86bda280
21:12 dalek parrot/jit_prototype: 13ac7bc | bacek++ | t/jit/jitted.ops:
21:12 dalek parrot/jit_prototype: Update test ops.
21:12 dalek parrot/jit_prototype: review: https://github.com/parrot/parrot/commit/13ac7bc81c
21:14 ligne left #parrot
21:15 Tene I've never had notable success waking up with enough time before work to do anything.
21:15 lucian i see. same here
21:15 lucian btw, my proposal for who's interested https://gist.github.com/891481
21:16 Tene I go to sleep around 08:00 UTC
21:16 cotto_work lucian: you might as well loop in parrot-dev
21:16 cotto_work I'm interested in that discussion and I'm sure I'm not alone.
21:16 lucian cotto_work: i see. good idae
21:17 benabik Is there a machine readable table of opcode name to number?
21:18 Tene lucian: email EVERYONE
21:18 lucian Tene: yep
21:18 Tene I'm sure we coudl find a spambot network somewhere to distribute your GSoC proposal to the world. :)
21:18 dukeleto benabik: yes. trying to remember where it lives
21:18 dukeleto benabik: it gets autogenerated somewhere
21:18 Tene Yes, I'll read that when I get home.
21:19 Tene afk $job
21:20 benabik dukeleto: The only thing I can see is include/parrot/oplib/ops.h
21:21 plobsing benabik: every oplib has an op_info_t array
21:21 plobsing likewise every bytecode segment (for the mapping)
21:21 plobsing if you wanted something dynamic
21:23 plobsing there's a hash of those names somewhere too
21:23 plobsing I was considering converting those to a trie at some point
21:26 dukeleto benabik: src/ops/core_ops.c and friends
21:26 dukeleto benabik: starting at line 78
21:27 benabik dukeleto: Alrighty then.  Thanks.
21:30 ambs left #parrot
21:34 eternaleye left #parrot
21:37 bubaflub hola dukeleto
21:40 soh_cah_toa joined #parrot
21:41 bubaflub left #parrot
21:44 soh_cah_toa hey guys. there has been some debate as to whether the gsoc parrot debugger project should be kept in a separate repo or in a branch off parrot/master. what do you think?
21:45 plobsing split out if possible. I'm in the thin-core camp.
21:45 soh_cah_toa i think a branch makes the most sense. isn't that the prupose of a branch/merge after all? plus, the debugger is packaged as part of parrot and a separate repo would be hard
21:45 soh_cah_toa *purpose
21:52 mj41 left #parrot
21:54 lucian soh_cah_toa: i think it might be nice if there was an instrumentation api for parrot, and a debugger written as an application that uses it
21:54 soh_cah_toa lucian: i plan on doing that. it's in my proposal
21:55 lucian right. then perhaps the instrumentation api would be in parrot, and the debugger separate
21:55 lucian i don't know if it's a good idea, or whether parrot wants to ship a debugger
21:56 soh_cah_toa oh wait, i see what you mean now
21:56 soh_cah_toa wouldn't is make more sense to include instrumentation in the debugger instead of an api since the only practical use for it would be in a debugger?
21:56 soh_cah_toa it*
21:57 lucian soh_cah_toa: i don't know. perhaps other things can be done with it
21:58 soh_cah_toa lucian: sure. the finished product will definitely use code from parrot-instrument though. we'll see what else we can do w/ it once i play around w/ parrot-instrument a little more
21:58 lucian soh_cah_toa: if i were you i'd branch/fork
21:58 lucian separate it later if possible, or not
21:59 soh_cah_toa lucian: i agree. everyone else seems to agree as well. whiteknight was the only who felt differently
22:01 donaldh joined #parrot
22:01 plobsing left #parrot
22:02 soh_cah_toa lucian: when would you recommend working in the documentation into my project? i had it at the end but dukeleto made a good point that it might not get done then. he recommended doing it during design. what do you think?
22:02 lucian soh_cah_toa: https://gist.github.com/891481
22:03 lucian i have mine in bits
22:05 soh_cah_toa yeah
22:07 bacek left #parrot
22:12 soh_cah_toa lucian: have you done gsoc before?
22:12 lucian yeah
22:12 soh_cah_toa lucian: w/ the parrot foundation or another organization?
22:12 lucian with Sugar Labs
22:12 eternaleye joined #parrot
22:13 soh_cah_toa lucian: never heard of them. what project was it for?
22:13 lucian Sugar :)
22:13 lucian ever heard of OLPC?
22:13 lucian soh_cah_toa: http://sugarlabs.org
22:13 soh_cah_toa yeah - one laptop per child
22:14 lucian Sugar is the UI for it
22:15 soh_cah_toa yeah i'm looking at it now. it's pretty neat
22:20 plobsing joined #parrot
22:29 dukeleto ~~
22:31 * lucian waves at dukeleto
22:57 soh_cah_toa left #parrot
23:01 whiteknight joined #parrot
23:04 whiteknight good afternoon, #parrot
23:05 hercynium left #parrot
23:05 donaldh left #parrot
23:06 bubaflub joined #parrot
23:07 * dukeleto waves back
23:10 bubaflub ~~
23:12 kurahaupo joined #parrot
23:55 dmalcolm left #parrot
23:58 cosimo joined #parrot

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced