Camelia, the Perl 6 bug

IRC log for #parrot, 2010-12-03

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 nopaste "bluescreen" at 192.168.1.3 pasted "implementation of Parrot_api_get_compiler" (14 lines) at http://nopaste.snit.ch/26394
00:00 bluescreen that's a preview of the function
00:01 donaldh left #parrot
00:03 whiteknight I thought there was something like that already. give me a few minutes to look
00:04 bluescreen sure..
00:07 TypeNameHere_____ joined #parrot
00:12 whiteknight hmm... I can't find the function I am looking for
00:13 whiteknight somewhere the PIR compreg is created, and I can't find it
00:13 whiteknight src/embed.c:Parrot_compile_string looks similar, but not the same
00:13 whiteknight and there is some compreg stuff in src/interp/inter_misc.c
00:14 whiteknight that Parrot_api_get_compiler function looks good
00:15 whiteknight Actually, create Parrot_get_compiler in inter_misc.c, that would be best
00:23 bluescreen parse_utils.c
00:23 whiteknight oh, let me look there in a minute (building now)
00:24 bluescreen whitenknight: then 2nd question...
00:24 bluescreen should we expose a Parrot_api_compreg
00:24 bluescreen that way we would allow our embbeders to use their own c compilers
00:24 bluescreen but... we have to expose couple more strctures
00:25 whiteknight what other structures? We should be able to return a compiler PMC
00:25 bluescreen sorry, maybe we don't need to export anything else with PMC we're fine..
00:25 whiteknight ok
00:26 whiteknight yes, a Parrot_api_get_compiler would be good
00:26 whiteknight Parrot_api_set_compiler too
00:26 bluescreen ok
00:26 whiteknight (and we will use Parrot_api_set_compiler in src/main.c to add in the PIR and PASM compilers to libparrot)
00:26 bluescreen that will require from embbeders to know the internals of the interp..
00:27 bluescreen and they have to return a sub pmc
00:27 TypeNameHere_____ left #parrot
00:27 bluescreen look at compilers/example/japh.c
00:27 TypeNameHere_____ joined #parrot
00:29 rfw joined #parrot
00:29 rfw hi again whiteknight
00:29 whiteknight where is parse_utils.c?
00:29 whiteknight hello rfw
00:29 rfw came to claim another task for parrot :D
00:29 rfw the y-combinator one
00:29 whiteknight rfw: really? awesome
00:30 rfw http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129130819823
00:30 whiteknight Y-Combinator is slightly crazy, so don't worry if it doesn't make sense
00:30 whiteknight I still don't always understand it
00:30 rfw lol
00:30 bluescreen compilers/imcc/parse_utils.c
00:31 whiteknight ah, I was in the wrong folder
00:32 whiteknight yeah, there it is. That's where it compregs the PIR and PASM compiler
00:32 bluescreen yeah
00:33 whiteknight oh damnit. Parrot_compreg takes a raw C function pointer
00:33 whiteknight that is unacceptable
00:33 bluescreen yeah
00:33 bluescreen why... sorry my complete ignorance
00:33 theory joined #parrot
00:36 whiteknight ...I changed the error-handling stuff, and now miniparrot doesn't link
00:36 whiteknight and I can't figure out why not
00:38 whiteknight oh, nevermind. realclean fixed it
00:40 whiteknight builds! Testing
00:41 bluescreen then... missing functions "Parrot_api_load_language", "Parrot_api_get_compiler", "Parrot_api_compile_file", "Parrot_api_set_compiler" do you agree?
00:42 whiteknight I don't think we need Parrot_api_compile_file, right? I mean, we already give them the compiler object
00:42 whiteknight and different compiler objects may have different interfaces, so Parrot_api_compile_file won't always know how to compile
00:42 whiteknight I think we need to just pass them the compiler PMC and let them use it
00:43 whiteknight what we do need to do is make a Parrot_api_call_method, which should basically be a copy of Parrot_ext_call
00:43 whiteknight and once we have that, we can do any methods we want
00:45 bluescreen mmm ok...
00:45 bluescreen so its their responsibility to call ".compile"
00:45 whiteknight yes
00:45 whiteknight the PIR and PASM compiler PMCs don't have ".compile"
00:45 bluescreen yeah...
00:46 theory left #parrot
00:46 bluescreen so we return an NCI PMC
00:46 sorear why don't we standardize the compilation interface?
00:46 dmalcolm left #parrot
00:46 dukeleto yes please!
00:46 sorear the IMCC PMC can support VTABLE_invoke for compatibility
00:46 sorear s/can/must/
00:46 sorear but we don't need a dep cycle to add a .compile method
00:47 whiteknight sorear: it's not a matter of just adding a .compile method. We would need to lay out exactly what we want a compiler object to do
00:47 whiteknight if you want to put together a proposal, that would go a long way
00:48 rfw what does error:imcc:The opcode 'sub_i_p_ic' (sub<3>) was not found. mean
00:48 rfw rather
00:48 rfw what is sub_i_p_ic
00:50 bluescreen rfw, thats an pir opcode, that takes 3 parameters
00:50 dukeleto rfw: you are trying to subtrac a PMC and an integer constant
00:50 rfw ah
00:50 dukeleto rfw: and store it in an int. There is no opcode for that.
00:50 dukeleto rfw: i = int, p = pmc, ic = integer constant
00:50 rfw yeah had a feeling it was something silly
00:50 dukeleto rfw: it happens a lot to me too :)
00:50 rfw didn't realize it was quite that silly
00:52 rfw hm can i define the return type of a function in winxed?
00:52 rfw or cast from pmc to num
00:52 sorear dukeleto: Why isn't there a sub_i_p_ic opcode?
00:54 bluescreen whiteknight: STRING * const sc       = CONST_STRING(interp, "PJt"); what is the calling convention
00:54 jhelwig left #parrot
00:54 bluescreen what does PJt means
00:55 sorear whiteknight: we already have one of those.  I 100% support "baby steps" towards IMCC supporting PDD-31 compreg object API
00:55 sorear bluescreen: look in nci.dat there's a comment header
00:55 whiteknight There are two "dialects" of function signatures. NCI signatures (like what you have there) are stupid
00:55 jhelwig joined #parrot
00:55 whiteknight P is PMC, J is interpreter, t is some kind of integer I think
00:55 sorear extra_thunks.nci
00:56 whiteknight the first item is the return value, so PMC * func(Interp *interp, ...)
00:56 sorear # t - character string (0-terminated)
00:56 rfw how do i cast a pmc to int?
00:56 sorear VTABLE_get_integer_native(INTERP, pmc)
00:56 whiteknight rfw: INTVAL i =VTABLE_get_integer(interp, pmc)
00:56 rfw ah
00:56 whiteknight oh, get_integer_native maybe
00:56 rfw interp?
00:57 rfw what's the interp parameter
00:57 sorear struct Parrot_Interpreter *interp
00:57 sorear reference to all global data
00:57 rfw oh
00:59 bluescreen so... having the compiler registered, then we don't need the Parrot_api_call_method... as the compreg is pointing to the compile method only
01:00 whiteknight no, the compreg points to the compiler object
01:00 whiteknight that object may have several methods .compile, .parse, etc
01:01 rfw oh
01:01 rfw er
01:01 rfw wrong window
01:02 whiteknight rfw: the interp is the thing that runs stuff in parrot
01:02 sorear see PDD-31 for the details on compiler objects and all the methods they support
01:02 sorear except for IMCC, which doesn't support PDD-31 at all
01:02 whiteknight rfw: interp is the first argument to most functions. Just make sure to pass it around everywhere
01:03 sorear and returns a NCI function, not an object at all
01:03 whiteknight sorear: Right. IMCC is bad
01:03 whiteknight and everything it does is terrible
01:03 whiteknight and anybody who likes it is wrong
01:04 sorear ...did I say something wrong?
01:04 whiteknight no
01:05 whiteknight unless you like IMCC, then you are wrong
01:05 whiteknight otherwise, you're cool
01:05 rfw uh oh, infinite recursion
01:06 dalek parrot: 6d5595e | (Gerd Pokorra)++ | README:
01:06 dalek parrot: change to Fedora package name
01:06 dalek parrot: review: https://github.com/parrot/parrot/commit/6d5595e04c
01:08 rfw hm i keep getting infinite loops when implementing a y-combinator
01:08 rfw probably doing something really wrong
01:10 whiteknight rfw: at least it's recursing!
01:10 rfw heh
01:10 whiteknight you have your code? nopaste it
01:10 whiteknight nopaste?
01:10 whiteknight aloha: nopaste?
01:11 whiteknight aloha nopaste?
01:11 rfw in a bit
01:11 whiteknight DAMNIT
01:11 whiteknight aloha nopaste
01:11 whiteknight aloha: nopaste
01:12 dukeleto http://nopaste.snit.ch/
01:14 nopaste "rfw" at 192.168.1.3 pasted "y-combinator in winxed" (38 lines) at http://nopaste.snit.ch/26395
01:15 whiteknight rfw: very nice!
01:15 rfw whiteknight: also very broken!
01:15 whiteknight Post the error message you are getting?
01:15 rfw just infinite recursion
01:15 rfw maximum recursion depth exceeded
01:16 whiteknight try it with a lower number than 6?
01:16 whiteknight start at the bottom. Try factoral(0), factorial(1), factorial(2)
01:16 whiteknight I dont know what the recursion depth is set at
01:18 rfw i think it's something broken in my implenetation
01:19 rfw ah whoops there is
01:19 rfw silly me
01:19 dalek parrot/embed_api2: 2483c68 | Whiteknight++ | / (16 files):
01:19 dalek parrot/embed_api2: implement the new error-handling mechanism. Instead of a string, we pass out the raw exception object for the embedder to query. Add some helper API functions. Parrot_x_exit should now only be called when actually exiting. For most operations, we should call the new Parrot_x_jump_out instead
01:19 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/2483c68859
01:20 whiteknight I don't even have winxed installed here right now
01:20 whiteknight I install/uninstall parrot so often
01:22 theory joined #parrot
01:23 whiteknight bluescreen: check out that commit! two tests still fail, but I think they are stupid
01:24 fbrito joined #parrot
01:25 rfw hm whiteknight
01:25 rfw it works fine in JS
01:26 rfw i think winxed is broken
01:27 whiteknight okay, that's a result
01:27 rfw http://pastebin.com/vT6M4aaX
01:28 rfw er
01:28 rfw that's the JS implementation
01:28 whiteknight the task was only to see if it would work, not to make it work
01:28 rfw haha
01:28 rfw http://pastebin.com/GtjWrdei there's the winxed implementation
01:29 rfw which is the same with a few types thrown around
01:29 whiteknight I suspect winxed is not handling lexicals correctly then
01:30 rfw my formatting makes it look like a sideways-M combinator more than anything
01:33 rfw whiteknight: so, is this a pass?
01:34 whiteknight rfw: yes. Let me mark it
01:34 rfw \o/
01:34 rfw oh also
01:34 rfw something i noticed
01:34 whiteknight rfw: change the task to NeedsReview
01:35 rfw oh
01:35 rfw question, though: if you declare a function in the global scope, can you not use the name as a variable in another function scope?
01:35 rfw since winxed puked a bit when i declared factorial_y as a function outside of main
01:36 cotto_work We have two of the top three ranking gci students hacking on our stuff.  Nice.
01:36 bluescreen left #parrot
01:36 whiteknight rfw: that's a good question. I don't know the answer
01:36 whiteknight I'll have to ask NotFound (he's Winxed's creator)
01:36 rfw cotto_work: first place guy just writes a bucketload of asm functions for x264 :(
01:36 whiteknight rfw: I'm going to use your work to open a ticket for Winxed
01:37 rfw ah
01:37 rfw cool
01:37 whiteknight rfw: we're coming up with new tasks every day. Keep checking back with us, and we will keep you stocked with good tasks
01:37 rfw will do :)
01:38 whiteknight awesome
01:38 whiteknight rfw++
01:38 cotto_work rfw: well, one of them does ;]
01:38 rfw cotto_work: heh
01:39 whiteknight msg NotFound a student wrote the Y-Combinator in Winxed and it broke. PRobably a bug in Winxed not handling lexicals correctly. I'll open a ticket for you, and can try to help fix it
01:39 cotto_work rfw: is there anything you especially like hacking on?  I'm trolling through parrot's tickets tonight looking for gci tasks.
01:40 rfw trolling tickets?
01:40 rfw that's a funny way to phrase it
01:40 rfw but uh, not really
01:48 kid51 joined #parrot
01:48 kid51 msg fbrito If you ever run out of GCI tasks, we have several pertaining to translation of our web pages into Spanish
01:49 kid51 (hmm, no/slow response from aloha)
01:52 whiteknight yes, aloha is being very disappointing tonight
01:55 cotto_work aloha: aloha?
01:56 fbrito kid51: spanish? :s
01:57 fbrito what about brazilian portuguese? :D
01:57 kid51 fbrito I got the impression you were a native Spanish speaker
01:57 kid51 Was that wrong?
01:57 fbrito Yes :P. I am from Brazil :)
01:57 kid51 Which city?
01:58 fbrito in the easternmost city in the Americas
01:58 fbrito http://en.wikipedia.org/wiki/Jo%C3%A3o_Pessoa
01:58 kid51 Recife?  Belo Horizonte?
01:59 fbrito Really close to Recife
01:59 fbrito (120km)
02:00 cotto_work Nice.  You're right on the coast.
02:00 kid51 Ah, somehow I thought you were from Buenos Aires.
02:01 cotto_work kid51: you're thinking of darbelo
02:01 kid51 In any case, on our Google Code In wiki page there are several tasks for translation of our web pages, etc., in various languages.
02:01 cotto_work the originator of our proud tradition of naming branches *_massacre
02:01 kid51 I don't know if Portuguese was listed there, but, hey, it wouldn't hurt!
02:02 kid51 cotto_work:  Yes, I know darbelo's from there.  I thought we had picked up a second Argentine Parrot hacker.
02:02 kid51 Instead, we picked up our first Brazilian!
02:02 cotto_work well, he picked us
02:03 kid51 In any case, we have at least one European Portuguese contributor who could double-check any translations fbrito did.
02:03 cotto_work either way, awesome
02:03 kid51 namely, ambs
02:04 mikehh_ joined #parrot
02:08 dalek TT #1879 created by whiteknight++: Deprecate Parrot_compreg
02:08 dalek TT #1879: http://trac.parrot.org/parrot/ticket/1879
02:08 mikehh left #parrot
02:11 fbrito rfw: congrats :D. you are now in the 2nd/1st place :D
02:11 rfw fbrito: i was there before too ^^
02:12 rfw well
02:12 rfw before before
02:17 fbrito rfw: have you tried those tasks related to ganeti-webmgr?
02:18 rfw ah no
02:18 rfw not yet
02:18 rfw i'm running windows so it's a bit of a pain
02:18 rfw and i'm not going to run VMs in my VMs lol
02:18 fbrito there are sooo many of them, and really easy ones (basic things with python for web)
02:18 fbrito but setting the environment is REALLY hard... realy
02:18 rfw yeah
02:18 fbrito really*
02:18 rfw i'm just making a "how to install dragonflybsd" screencast right now
02:18 fbrito I had to recompile my kernel like 3 times to get xen working
02:19 rfw haha yeah
02:25 fbrito wow, Parrot tasks are getting harder and harder :)
02:26 whiteknight fbrito: sorry! We used up all the easy ones that we could think about
02:26 whiteknight we need to come up with more tasks
02:26 fbrito no! its a good thing
02:26 whiteknight it's your fault. If you guys weren't so good, we wouldn't offer harder tasks
02:27 fbrito this contest is about students helping open source, and not the other way around :P
02:28 cotto_work fbrito, your next gci task is to write a PIR program to demonstrate P=NP by decrypting an SSL connection on the fly
02:28 dalek parrot/embed_api2: 6e591fa | Whiteknight++ | src/ (2 files):
02:28 dalek parrot/embed_api2: fix some remaining issues where we were still caling Parrot_io_eprintf in die_from_exception. Now all error message/backtrace stuff is done through the API. However, the formatting is off and a few tests are barfing because they are trying to regex match the error message format.
02:28 dalek parrot/embed_api2: review: https://github.com/parrot/parrot/commit/6e591fa600
02:28 whiteknight I'll add more tasks tomorrow. Right now, bed.
02:28 whiteknight goodnight
02:29 fbrito good night :)
02:29 whiteknight left #parrot
02:30 fbrito cotto_work: ahhaha, ok. how many points?
02:36 bacek_at_work fbrito, http://www.getacoder.com/projects​/solve%20p%20vs%20np_132036.html
02:36 bacek_at_work 500-1000 bucks apparently :)
02:37 bacek_at_work And this one just arrived in my Google.Reader http://i.imgur.com/ixZ9o.jpg :)
02:41 fbrito ahahahhahahahaha
02:42 fbrito "problem?" ahha, I really like "troll science" (http://knowyourmeme.com/memes​/troll-sciencetroll-physics)
02:47 mikehh_ is now known as mikehh
02:47 mikehh opbots, names
02:55 fbrito we should be able to claim another task while we have a task marked as NeedsReview :(
02:56 dalek parrot/tt532_headerizer_refactor: 277b983 | jkeenan++ | / (3 files):
02:56 dalek parrot/tt532_headerizer_refactor: Refactor some code out of Object.pm into new Functions.pm subroutine  handle_split_declaration().  Write tests for all execution paths of that subroutine.
02:56 dalek parrot/tt532_headerizer_refactor: review: https://github.com/parrot/parrot/commit/277b983086
03:04 dalek parrot/tt532_headerizer_refactor: 92f3471 | jkeenan++ | lib/Parrot/Headerizer/Functions.pm:
03:04 dalek parrot/tt532_headerizer_refactor: Eliminate an implicit, but unreachable branch in handle_split_declaration().
03:04 dalek parrot/tt532_headerizer_refactor: review: https://github.com/parrot/parrot/commit/92f34715d5
03:08 plobsing joined #parrot
03:10 kid51_ joined #parrot
03:13 kid51 left #parrot
03:23 plobsing left #parrot
03:24 fbrito I created a account on Parrot Trac in order to edit a wiki page but it seems that I don't have permissions to do that
03:25 fbrito an account*
03:26 cotto dukeleto, ^
03:29 fbrito There is a task about updating a wiki page... But don't worry... it is not really important. I still have to wait until tomorrow to get a work on another task reviewed.
03:29 nopaste "kid51_" at 192.168.1.3 pasted "embed_api2 branch: Build failure in src/embed/api.c, observed on Darwin/PPC at 6e591fa6" (328 lines) at http://nopaste.snit.ch/26400
03:50 plobsing joined #parrot
04:03 Coke fbrito: I can grant perms, momentito.
04:03 fbrito No, wait :P
04:04 Coke uid?
04:04 Coke hokay.
04:04 fbrito I am locked to another task :~
04:05 fbrito Anyway, I think that this wiki task is overrated.  There are like just 3 or 4 lines that need to be updated
04:08 kid51_ left #parrot
04:13 fbrito In my opinion it should be marked as "easy" :P
04:18 dukeleto fbrito: you need anything?
04:19 fbrito dukeleto: no, thank you. I am waiting for review on my task on another project
04:19 cotto dukeleto, can you give trac users wiki editing bits?
04:22 fbrito I started working on this Parrot task: http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129130043418 but according to GCI rules I am not supposed to work on a task that I have not claimed, and since I can't claim tasks while I still have a task on NeedReview, I am doing nothing :(
04:25 cotto fbrito, which task is awaiting approval?
04:25 fbrito GNOME task :P Submitting a patch to add text/x-prolog mimetype to shared-mime-info project
04:26 rfw oh
04:26 rfw hi fbrito
04:26 cotto ah.  nothing we can do about that
04:26 fbrito So GEdit can identify if a *.pl file is Perl or Prolog
04:26 cotto it's definitely no good that you can get blocked like that
04:27 fbrito I will start working on the Parrot task and if anyone claim it, I just share my work with him and done
04:28 cotto wfm
04:28 cotto have you been waiting for the gnome folks for long?
04:29 fbrito no, no :P
04:29 fbrito but I think there is only 1 guy responsible for the gnome tasks there, and he is sleeping right now
04:31 cotto Really?  I'd expect them to have half a dozen people involved.
04:34 fbrito Hm, true. I think the guy I contacted is only responsible for the Gnome/Gedit tasks
04:39 fbrito rfw: may I ask something? on which task are you working right now? :)
04:39 rfw Create a brief video, similar to a rails-cast or one of the many other well put together how-to videos common on the internet today, detailing the entire installation process of DragonFlyBSD.
04:39 rfw taking some time off coding
04:40 fbrito ahha, nice
04:40 rfw have fun with your gnome task
04:40 rfw haven't played with it in a while
04:40 fbrito I gave up all the screencasts tasks because my english sucks :(
04:41 rfw heh yeah, some tasks are rather unfairly biased
04:41 rfw especially those apertium ones
04:42 fbrito ahahah. translating to czench, slovenian, greek, ukrainian, danish, faroese, friulian, slovak...
04:43 rfw yeah :(
04:43 rfw oh wow
04:43 fbrito the only documentation related task that I did was writing this page: http://docs.limesurvey.org/tiki-in​dex.php?page=Creating,+Editing+and​+Translating+LimeSurvey+Wiki+Pages
04:43 rfw you've done a fair few tasks for parrot
04:44 fbrito yes. I am in love with parrot :)
04:44 rfw hehe yeah
04:44 rfw i'm loving parrot too
04:44 fbrito but not even close to the kind of tasks that the 1st guy did
04:44 rfw hah yeah
04:45 fbrito he has ALL (100%) the same task :D
04:45 rfw i'm still waiting for someone from google to show up in #x264dev
04:45 cotto is that on freenode?
04:45 rfw and say "look you're not allowed to add asm tasks anymore"
04:45 rfw cotto: yes
04:46 * cotto joins and and expectantly prepares some popcorn
04:46 cotto not that it's likely
04:46 rfw heh
04:47 rfw it's just mostly jumpyshoes talking about asm
04:47 fbrito ahha
04:50 fbrito what about the playtest tasks? have you take a look on them?
04:50 rfw oh hell no
04:50 rfw i can't play games for peanuts :p
04:51 fbrito they usually request previous experience in the game
04:52 rfw yeah, tactical games aren't my kettle of fish
04:52 rfw i mostly (read: only) play shmups, since i'm relatively good at them and i don't lose self-esteem after playing them online
04:53 fbrito :P
04:53 fbrito wow, I can't find those tasks on the list anymore :o. how is that possible? there were almost 10 from The Battle for Wesnoth
04:53 rfw i guess people took them all
04:54 fbrito long time ago I used to play a lot (starcraft, CS, warcraft, mmo rpgs, ...) but when I started to code I got so addicted that I couldn't stop :P
04:55 rfw hah
04:55 davidfetter joined #parrot
04:56 cotto allison, ping
05:09 cotto dukeleto, ping
05:27 fbrito left #parrot
05:27 fbrito joined #parrot
05:39 dalek tracwiki: v31 | cotto++ | CottoTasklist
05:39 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Cot​toTasklist?version=31&amp;action=diff
06:22 plobsing ping rfw
06:22 rfw pong
06:22 rfw what's up?
06:23 plobsing I've debugged your ycombinator.
06:23 rfw ah
06:23 rfw what was wrong with it?
06:23 rfw it runs fine javascript though ._.
06:23 plobsing it is a problem with winxed lexicals. you can work around by having the variable closed over as outer coming from a local variable (as opposed to an argument)
06:24 rfw ah
06:24 plobsing the problem is that the 2 closed-over variables in the combinator (outer and func) get aliased to the same lexical.
06:24 rfw wasn't aware of that
06:24 plobsing not your fault.
06:24 rfw still a bug in winxed though :)
06:25 fbrito I have interest in http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129123759504 but I still think it is overrated. I think 2 points are a lot when all you have to do is change 2 or 3 lines on a wiki page.
06:26 fbrito are you guys allowed to edit this?
06:30 cotto I can if melange isn't broken.
06:30 cotto a successful sign-in results in me not being signed in
06:31 cotto perhaps it has something to do with the url change
06:31 fbrito which URL are you trying? google-melange.com or the one with appspot?
06:31 fbrito there seems to be http://socghop.appspot.com/ and  http://www.google-melange.com/
06:32 fbrito and they don't share the same session
06:32 rfw (and they both suck as much as each other)
06:32 cotto appspot seems to work
06:33 fbrito rfw: ahahhaah! I totally agree with you
06:33 cotto google-melange doesn't
06:33 fbrito rfw: on the first days the list all tasks page was REALLY slow
06:33 rfw fbrito: i was there as soon as it started
06:34 fbrito rfw: and even now we have some silly bugs like tasks with 0 points :o
06:34 rfw i got a crapload of 500 errors
06:35 rfw fbrito: i also hate the fact they used jquery ui everywhere
06:35 rfw i wanted properly paginating pages ._.
06:35 rfw not some javascript mess
06:35 cotto I like jquery, but only if it's used properly.
06:35 rfw yeah
06:35 fbrito why? did you want to write a crawling bot? :P
06:35 rfw you don't use jquery for the whole goddamn interface
06:35 rfw fbrito: no, because it takes a few minutes to load all the tasks
06:36 davidfetter .o(pquery)
06:36 rfw fbrito: also the backend is json data anyway
06:36 cotto but I can complain with one hand and read cool research papers with the other and see which goes to sleep having no regrets about how it spent its day
06:37 fbrito rfw: yes, true. like when you open "list all tasks" and filter by "parrot", you still have to wait some seconds while it loads everything
06:37 rfw fbrito: it's sorta "oh look there's some parrot tasks WHY IS IT FILLING UP WITH OTHER TASKS"
06:37 fbrito ahhaha! it is even worse in the rank page
06:38 fbrito "yes... i am the 2nd! HEY, WAIT!"
06:38 rfw lolol
06:38 rfw i wonder if poor artem shmelev on 0 points at the bottom is going to get a t-shirt
06:38 fbrito and than you can see that it is loaded in alphabetical order
06:39 fbrito then*
06:39 fbrito like you. your name starts with T, so it always loaded after mine
06:39 rfw :p
06:40 fbrito rfw: I sent an email to gci-discuss list and they said that 0 points task was a bug :P
06:40 rfw does he even get 1 point
06:40 rfw lol
06:40 fbrito poor guy :D
06:41 rfw Under no circumstances shall the submission of an Entry into the Contest, the awarding of a prize, or anything in these Official Rules be construed as an offer or contract of employment with either Google or any participating Open Source Organization.
06:41 rfw aw :(
06:42 dukeleto rfw: don't read too much into that ;)
06:42 dukeleto cotto: pong, but distracted
06:43 cotto dukeleto, ooc how much time do you spend looking for people building Parrot on unusual platforms and how often do you try to recruit them?
06:44 fbrito rfw: have you seen those numbers? http://pastie.org/1340596 they are from yesterday
06:44 cotto rfw, we'll hire you for an annual salary of twice what dukeleto makes.
06:44 cotto aloha, 2*0
06:44 rfw ahaha i was just going to ask if dukeleto even made a salary :p
06:45 rfw fbrito: hoping i can still stay in the top 10
06:45 rfw trip to mountain view == awesome
06:45 fbrito it would be nice if they had offered a scholarship on a american university
06:45 cotto (for his Parrot work.  I'm sure his dayjob pays nicely.)
06:45 fbrito MIT costs like 55k/year
06:45 rfw yeah
06:45 rfw hm there's another brazilian in here
06:46 * rfw pokes Kovensky
06:46 fbrito really?
06:46 rfw yah
06:46 rfw Kovensky is
06:46 rfw i don't remember where he's from though
06:46 fbrito his name sounds like russian, ahhaa
06:46 sorear most US unis have radically different rates for locals (same state or in some cases city) and non-locals
06:46 cotto Oh noes.  You're taking over!
06:46 rfw sorear: aren't californian unis free for californians
06:47 rfw i guess the poor californians get the hell taxed out of them :p
06:47 fbrito I am going to write a test to go to a uni here in my city on monday :o
06:48 sorear rfw: they were for my parents' generation :(
06:48 rfw sorear: dang :(
06:49 fbrito cotto: were you able to edit that wiki task?
06:49 fbrito rfw: have you ever been to USA?
06:49 rfw fbrito: my university entrance restuls arrive on 23rd of january ._.
06:49 rfw andn o
06:49 rfw no*
06:50 fbrito me neither. do you know if we will be able to choose the ticket dates?
06:50 stilgar joined #parrot
06:50 rfw i'd rather not think of that yet
06:50 rfw we have to win first, you know
06:50 cotto fbrito, which task?
06:51 rfw fbrito: still got quite a few weeks to go :(
06:51 NotFound plobsing: ping
06:52 fbrito this task: http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129123759504 I heard another student saying that parrot tasks were too overrated, and I agree in some cases. in that task we only have to edit 3 or 4 lines in the wiki page. I think it should be an "easy" (1 point) task
06:52 plobsing NotFound: pong
06:52 fbrito rfw: yes, you are right
06:52 NotFound plobsing: what's the problem you've found?
06:52 stilgar cotto: i try to recruit anyone who seems interested in parrot, regardless of their platform
06:53 plobsing closure-within-closure can cause name conflicts in lexical name selection
06:53 plobsing and it wasn't me, it was rfw++ and his y-combinator
06:53 * stilgar is  dukeleto's phone
06:54 NotFound Ah, yes, I've seen the issue now.
06:54 NotFound rfw: good work
06:54 rfw \o/
06:54 rfw thanks
06:55 cotto fbrito, update
06:55 cotto d
06:55 fbrito rfw: may I ask what 'rfw' means?
06:55 NotFound plobsing: you're right, the problem is the simplistic scheme used to assign internal names to arguments.
06:55 rfw fbrito: abbreviation of my full nickname which is rather silly
06:57 fbrito I don't know why but I always type rwf and then I have to fix it :(
06:57 rfw r<tab>?
06:57 sorear opbots trust rfw
06:57 slavorgn Ok
06:57 slavorg Ok
06:57 NotFound rfw++ whitheknight++ plobsing++
06:57 sorear whitheknight-- whiteknight++
06:57 stilgar cotto: why do you ask?
06:58 rfw \o/
06:58 rfw thanks
06:58 cotto stilgar, there's a ticket for porting to more platforms that strikes me as unnecessary
06:58 fbrito do you guys know what gives great 1 point tasks? translating wikipedia project page to other languages!
06:59 fbrito I've seen quite a lot of projects doing that
06:59 rfw fbrito: http://www.google-melange.com/gci/student_ran​king/show_details/google/gci2010/rdaniel93_1 FILLING GODDAMN WIKI INFOBOXES
06:59 fbrito AHAHHAHAHAHA, I KNOW!
07:00 fbrito but have you tried to fill them?
07:00 fbrito I tried for like 30 minutes and then I gave up
07:00 rfw no
07:00 rfw i've kept away from 1 point tasks for now
07:00 NotFound rfw: you asked how to cast to int in winxed? Use the cast psuedo-function: i = int (whatever);
07:01 rfw NotFound: ah
07:01 rfw thanks
07:01 fbrito (rfw: I feel that we are flooding the channel with off topic matters :P)
07:01 rfw fbrito: yeah i do that everywhere i go
07:01 NotFound I should add that to the quick documentation section on the home page.
07:02 rfw offtopic discussion follows me like the plague
07:02 fbrito do you guys know who could provide me editing bits to the wiki?
07:05 stilgar cotto: which platforms?
07:05 cotto stilgar, the ticket doesn't specify.  That's part of why I don't like it.
07:06 cotto tt 609
07:06 stilgar fbrito: cotto should be able to give you an edit bit
07:06 cotto stilgar, nope
07:06 NotFound If someone wants to make me happy, pick the task about chunked encoding. I know it doesn't looks so cool as the Y-combinators and the like, but will be highly useful.
07:10 fbrito the task that I had most fun doing was writing unit tests to nummatrix2d on PLA. it was the hole day googling for LaPlace, Sarrus and so on :P
07:10 fbrito I mean examples, not unit tests
07:15 fbrito so, should I wait until tomorrow to ask who can give me permission to edit wiki pages?
07:15 stilgar fbrito: i am going to make some challenging tasks with you in mind :)
07:16 fbrito and when I find it too hard I will pass it to rfw :D hahaha
07:17 fbrito rfw: how are you going with your screencast?
07:23 fbrito Ok, I am going to bed. Please, if someone find who has the permission to give editing bit on wiki pages, tell him that I need it to a GCI task. My account on Trac is "fbrito" :)
07:27 cotto fbrito, did you respond to your welcome email?
07:29 fbrito respond?
07:29 dukeleto fbrito: stupid wiki
07:29 dukeleto fbrito: i will try to give you a bit
07:29 dukeleto fbrito: we had to do that because our wiki kept getting spammed
07:30 fbrito cotto: I think so, because to create tickets we have to verify the email, and I have already created a ticket
07:30 cotto I'm not seeing a response to a test account I just set up, but our trac site might not send out email instantly
07:30 dukeleto fbrito: have you ever used PostgreSQL ?
07:31 * cotto sleeps
07:31 fbrito dukeleto: I have used MySQL and SQLite on "production", but never PostgreSQL
07:32 fbrito but I have some basic knowledge of databases (queries, joins, relationship...)
07:32 dukeleto fbrito: you should have a wiki edit bit now
07:32 dukeleto fbrito: well then. I am going to make a PL/Parrot task for you. http://pl.parrot.org . And a good one ...
07:34 fbrito :o
07:34 fbrito seems interesting
07:35 NotFound If you are looking for ideas for new tasks: add support for winxed to PL/Parrot
07:40 fbrito I should be already on bed by this time (04:40am :D). Good "night" guys! I am going to finish the wiki task right after I wake up.
07:43 fbrito left #parrot
07:50 theory left #parrot
07:57 dukeleto NotFound: that could be interesting, and is nontrivial, to say the least :)
08:10 NotFound dukeleto: some are seying that the tasks are too easy, they aren't? So let's add a few challenging ones.
08:17 lucian joined #parrot
08:22 dukeleto NotFound: That will motivate me to make some very challenging tasks. But now I sleep.
08:22 dukeleto NotFound: PL/Winxed might make a nice task...
08:22 dukeleto NotFound: if you can give me PBC of Winxed, it could work
08:23 dukeleto NotFound: but i must sleep. Think of some hard tasks and we will talk tomorrow
08:33 fperrad joined #parrot
09:32 rfw left #parrot
09:37 contingencyplan left #parrot
10:25 preflex left #parrot
10:29 preflex joined #parrot
11:10 preflex left #parrot
11:14 preflex joined #parrot
11:54 kid51 joined #parrot
11:55 kid51 aloha is present but continues to be unresponsive
12:16 preflex left #parrot
12:19 preflex joined #parrot
12:36 zby__ left #parrot
13:16 kid51 left #parrot
13:29 bacek joined #parrot
13:43 Patterner left #parrot
13:43 whiteknight joined #parrot
13:44 whiteknight good morning, #parrot
13:53 Psyche^ joined #parrot
13:53 Psyche^ is now known as Patterner
14:21 smash joined #parrot
14:21 smash hello everyone
14:21 whiteknight good morning smash
14:35 atrodo good morning
14:38 whiteknight good morning atrodo
14:44 * whiteknight wonders if we could move plumage from gitorious to the parrot org on github
14:44 whiteknight or, at least fork it there
14:48 atrodo i'd imagine it'd be pretty easy
14:48 whiteknight seen japhb
14:48 whiteknight aloha help
14:48 aloha whiteknight: Ask me for help about: msg, convert, status, vars, karma, auth, seen, maths, translate, infobot, clock, loader (say 'help <modulename>').
14:49 whiteknight aloha seen japhb
14:49 fbrito joined #parrot
14:53 contingencyplan joined #parrot
14:56 whiteknight stupid bot
14:57 moritz seen japhb
15:02 NotFound whiteknight: the Y-combinator task has been a great idea.
15:08 dalek tracwiki: v19 | fbrito++ | NewParrotDeveloperGuide
15:08 dalek tracwiki: Updating from SVN to Git
15:08 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=19&amp;action=diff
15:08 dalek tracwiki: v20 | fbrito++ | NewParrotDeveloperGuide
15:08 dalek tracwiki: Link to git_workflow.pod
15:08 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=20&amp;action=diff
15:20 fbrito left #parrot
15:21 fbrito joined #parrot
15:24 dalek tracwiki: v21 | fbrito++ | NewParrotDeveloperGuide
15:24 dalek tracwiki: Link .pod files to GitHub source
15:24 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=21&amp;action=diff
15:24 dalek tracwiki: v22 | fbrito++ | NewParrotDeveloperGuide
15:24 dalek tracwiki: Link to tests.pod
15:24 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=22&amp;action=diff
15:24 dalek tracwiki: v23 | fbrito++ | NewParrotDeveloperGuide
15:24 dalek tracwiki: There is already a sentence starting with "once again" later
15:24 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=23&amp;action=diff
15:38 fbrito Is anyone here actually using "diffstat" or can I remove it from the wiki?
15:40 zby_ joined #parrot
15:50 whiteknight what is diffstat?
15:51 Coke it's nice to have, but not a requirement for new developers.
15:51 Coke kill it.
15:51 Coke whiteknight: tool to show you how  many lines changed in a given diff.
15:52 Coke sample: http://freshmeat.net/screenshots/c3/ef/c3ef27b4​c948dac9c6f15f6abea35395_medium.png?1240088976
15:52 whiteknight Coke++
15:54 dip joined #parrot
15:56 lucian allison: ping
15:59 dalek tracwiki: v24 | fbrito++ | NewParrotDeveloperGuide
15:59 dalek tracwiki: Remove diffstat and old ticket system
15:59 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=24&amp;action=diff
15:59 dalek tracwiki: v25 | fbrito++ | NewParrotDeveloperGuide
15:59 dalek tracwiki: Fix header level
15:59 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=25&amp;action=diff
16:01 fbrito and can anyone open those linked files on the wiki? (names.t, patch example, etc)
16:01 fbrito documentations on docs/ are so great that I think sooner or latter this wiki page will be just links pointing to them :P
16:02 dmalcolm joined #parrot
16:05 jan left #parrot
16:07 allison lucian: pong
16:08 lucian allison: i've sort of disappeared from pynie development lately :)
16:08 lucian I've been looking at it lately
16:08 lucian i'm wondering if using http://codespeak.net/pypy/​dist/pypy/doc/parser.html might be a better approach
16:09 allison so have I :)
16:09 allison well, actually going even further
16:09 allison and just making parrot a backend for pypy
16:09 lucian yeah, it's bitrotten so far it didn't even build. i fixed that, but it still doesn't eval
16:10 lucian allison: ah, but that would generate a python interpreter
16:10 lucian with a PASM jit, or something like that
16:10 lucian i think that's suboptimal, at best
16:10 allison that depends
16:11 allison right now, our compiler tools infrastructure is way too heavyweight
16:11 lucian allison: and weird. very, very weird
16:11 allison so we're not losing too much by switching to a different heavyweight one
16:11 allison and, we gain the advantage of  an array of work by people other than Parrot contributors
16:11 lucian allison: i don't think the problem with PyPy is size, but redundancy
16:12 allison parrot is heading toward being much smaller
16:12 lucian PyPy deals with a lot of issues that don't exist in parrot, even a smaller one
16:12 allison and, it should be possible to trim away most of the fat, so parrot is much leaner and lighter than CPython as a backend
16:12 lucian you can't reasonably compile the full python language to x86 or jvm bytecode
16:13 lucian but the full python language could be compiled to some parrot bytecode
16:13 lucian so there isn't really a need for an interpreter running on top of parrot
16:13 lucian it'd just be wasteful
16:13 lucian allison: but i do agree pynie should be written in python entirely
16:13 allison those would be my two best options
16:14 allison either strip pynie down so it's just the tiniest possible translator to Parrot bytecode
16:14 lucian or get a full interpreter running
16:14 allison or throw it away and make it a backend for pypy
16:14 allison yup
16:14 allison what we have right now is really a waste
16:14 lucian a pypy backend isn't hard conceptually, just time-consuming
16:15 lucian and it'd probably be slow and without a chance of interop with other languages on parrot
16:15 allison in some ways, a pypy backend is just a special case of the first idea
16:15 allison afaict, the first step would be implementing an RPython translator to Parrot bytecode
16:15 lucian allison: so a PyPy backend would mean two major things:1
16:15 lucian 1) RPython on parrot
16:15 lucian and 2) parrot object space
16:15 lucian yeah
16:16 allison both of which could be a good first step to a more extensive python implementation
16:16 Andy joined #parrot
16:16 allison so, the two possibilities aren't necessarily mutually exclusive
16:16 whiteknight fbrito: try to open those files again
16:17 lucian allison: yeah. RPython on parrot could at the very least bootstrap pynie
16:17 lucian allison: but if you compare CPython bytecode and parrot bytecode, there's not much difference
16:18 allison lucian: btw, if we start over, we should scrap the current code, and contribute the new code to the PSF from the start
16:18 lucian looking at PyPy's parser, it should be easier to replace the compiler bit and instead transform the AST to PAST or something
16:18 davidfetter left #parrot
16:18 fbrito whiteknight: still can't open them
16:18 lucian allison: would they be interested?
16:19 allison lucian: they were already interested
16:19 lucian then what's missing is the pynie object space
16:19 lucian allison: i see
16:19 allison lucian: but there was a bit of complexity in the earlier ownership
16:19 whiteknight fbrito: try it again (there are a lot of permissions here, I don't know which you need to view that file)
16:19 allison lucian: scrapping and starting over gets rid of all that
16:20 lucian allison: i see
16:20 fbrito whiteknight: nothing
16:21 whiteknight fbrito: ah, I just tried myself. Those files may have been deleted
16:21 whiteknight delete the links
16:21 whiteknight (oh, and you have WIKI_ADMIN privileges now, so don't break nothing)
16:22 lucian allison: so writing a PyPy backend will certainly get an almost entirely compatible interpreter
16:22 whiteknight allison, lucian: We've been talking about a very similar approach to a new JavaScript interpreter. Don't use NQP/PCT, write the compiler directly in JavaScript and bootstrap
16:22 lucian whiteknight: yes, we have :)
16:22 fbrito whiteknight: ok, I will try to update the links. thank you :)
16:22 whiteknight In fact, there are some JavaScript parsers written in JavaScript we've been looking at that we might just make codebackends for
16:23 whiteknight for languages sufficiently capable, this really starts to become a more attractive option
16:23 lucian allison: but a PyPy backend would be doubly interpreted, once parrot interpreter bytecode and twice pypy interpreting python
16:23 lucian s/parrot interpreter/parrot interpreting/
16:23 whiteknight I've heard several complaints from Pythonish people who don't want to write a Python interpreter in Perl6
16:24 whiteknight lucian: must PyPy generate an interpreter too? Can't it just parse and generate an AST?
16:24 lucian whiteknight: there is a bytecode compiler in PyPy, we could write a new code generator for that
16:25 lucian but i'm not sure that could cover all corner cases
16:25 lucian we might never get more than a language very similar to python
16:25 lucian whiteknight: http://codespeak.net/pypy/​dist/pypy/doc/parser.html
16:26 whiteknight lucian: once you bootstrap the first time, you can use your own compiler to write successively-better parsers
16:26 lucian whiteknight: btw, pypy is mainly two things: 1) a framework for writing generating VMs with JIT and GC from interpreters written in RPython and 2) a python interpreter written in RPython
16:26 whiteknight so PyPy might not give a complete Python, but then you use that subset to bootstrap a proper Python
16:27 lucian whiteknight: that's not the issue. i'm not certain parrot bytecode can express the whole of python
16:27 whiteknight lucian: whatever you have a problem with can be added
16:27 lucian whiteknight: right, that sounds good
16:27 lucian i guess i'm used to thinking about VMs as static targets
16:27 whiteknight lucian: Parrot? static? Give me a few minutes so I can laugh until I start crying
16:28 lucian whiteknight: yes, yes i know
16:28 lucian half of it seems to get deprecated every release :)
16:28 whiteknight only the bad half
16:29 whiteknight Oh, I wish we could rip out half of Parrot
16:29 whiteknight but I think Coke would go Super-Saiyan and murder us all
16:29 lucian allison: whiteknight: anyway, a first implementation might be PyPy's parser + generating PAST/PIR/PBC
16:29 PerlJam whiteknight: you can rip out as much as you want ...  you just have to fork and change the name   :-)
16:30 allison whiteknight: there's always the "do-over" option
16:30 whiteknight allison: no, I'm mostly kidding. i don't think half of it needs to go
16:31 allison whiteknight: indeed
16:31 whiteknight once we finally rip IMCC out, I think we'll be a MUCH better product
16:32 allison whiteknight: that will certainly help
16:32 whiteknight I don't know how much of the codebase is IMCC. It might be 50%
16:33 whiteknight allison: I've actually been wanting to pick your brain about that "concurrency-friendly CPS" ticket you opened a while back.  Do you have any more ideas about that?
16:33 smash whiteknight: poor IMCC :)
16:34 whiteknight (I was about to say something about ripping out threads too, and segued)
16:34 lucian whiteknight: is there a consensus on what sort of concurrency parrot should have?
16:35 whiteknight lucian: no, which is why I would like to explore this idea allison had
16:35 whiteknight We have an implementation of green threads in a branch that we could probably get pulled into trunk if that's what everybody wanted
16:35 lucian whiteknight: if you want to support all languages, you'll need an implementation of pthreads
16:35 whiteknight I suspect we will want to go a different way, but the option is open
16:35 lucian as evil as pthreads is
16:36 whiteknight lucian: at least, something that looks like pthreads
16:36 lucian whiteknight: i'm not sure that's even possible
16:36 lucian whiteknight: are you thinking green threads + non-blocking io to similate pthreads + blocking io?
16:37 lucian s/similate/simulate/
16:37 whiteknight lucian: I'm thinking Greenthreads as a stop-gap measure to simulate concurrency until we come up with something better
16:38 whiteknight We are going to want a system where we can have multple Parrot interpreters, each running on a separate OS thread, and being able to communicate with each other
16:38 whiteknight that approach becomes a heavy-weight pthread-like situation, but we are also going to want things that are much lighter
16:38 whiteknight greenthreads do seem to be that "lighter" alternative, especially if we get non-blocking IO working with them
16:39 lucian whiteknight: i don't know if you've seen eventlet http://wiki.secondlife.com/wiki/Eventlet
16:39 lucian whiteknight: and green threads are in fact much ligher on the OS's scheduler
16:40 lucian whiteknight: erlang processes and stackless threads can run in the millions, while pthreads kill the machine in the (very) low thousands
16:40 lucian python stackless that is
16:40 lucian whiteknight: there's also the other, slighty more low level approach, STM
16:41 lucian clojure has an awesome concurrency story
16:41 lucian but it makes some tradeoffs that I don't think are possible for parrot
16:41 PerlJam would that parrot could gain an awesome story too
16:42 lucian so an STM is nice to have, but without efficient immutable data structures, it's much less useful
16:42 whiteknight lucian: I've seen eventlets in the past, yes. I should probably re-read them
16:42 whiteknight and STM isn't really a concurrency solution, just a memory sharing solution
16:43 allison whiteknight: what would you like to know?
16:43 lucian whiteknight: it solves contention without locks
16:43 whiteknight I think that erlang processes and stackless threads could be simulated pretty easily with greenthreads multiplexed onto a small number of OS threads
16:43 allison whiteknight: it was the topic of my PhD thesis, but I had to set that aside for a while
16:43 lucian whiteknight: that's exactly what erlang does, in fact
16:44 whiteknight allison: anything! The ticket was extremely vague. I'm just wondering if you have any specific details that I could latch onto
16:44 allison I'll send you the thesis proposal as the start
16:44 whiteknight allison: you have a draft I could borrot? I'm an extremely good proof-reader :)
16:44 whiteknight nice! allison++
16:45 allison whiteknight: I never extended it into a more complete paper, but I could (probably should)
16:45 lucian that reminds me, I have a dissertation project proposal to write for next week ...
16:46 * lucian wishes he would find less things interesting
16:46 whiteknight lucian: what kind of dissertation are you proposing?
16:46 lucian whiteknight: mobile application SDK+IDE that is itself a mobile application
16:47 lucian basically developing mobile applications directly from your mobile device, without a PC
16:48 allison whiteknight: email sent
16:48 lucian whiteknight: and it involves generating JavaScript from something like http://education.mit.edu/drupal/openblocks
16:49 dalek tracwiki: v26 | fbrito++ | NewParrotDeveloperGuide
16:49 dalek tracwiki: Remove old (probably deleted) attachments
16:49 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=26&amp;action=diff
16:49 Coke whiteknight: why would I be upset that you ripped out half of parrot?
16:50 allison Coke: I suspect he meant from a code instability perspective
16:51 Coke ... you can't break partcl any more than it's been broken, unless you try REALLY hard. (having rakudo as a bulwark in front of partcl helps. ;)
16:53 whiteknight allison++ # Thanks! Reading intently
16:54 * allison steps away for a few minutes, brb
16:55 * Coke reads http://lwn.net/SubscriberLi​nk/417952/bf6a55b67170ff0e/ and hopes that that sort of thing isn't happening in the perl community. *sigh*
17:00 whiteknight Coke: That is extremely upsetting
17:00 moritz Coke: I've asked some women who attend Perl conferences, and they've told me that the Perl community is one of the friendliest towards women
17:01 moritz and I haven't heard anything like that from a YAPC
17:01 moritz (which might be connected to Larry brining his family, and others too)
17:05 lucian left #parrot
17:05 dalek tracwiki: v27 | fbrito++ | NewParrotDeveloperGuide
17:05 dalek tracwiki: Add "feedback" section
17:05 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=27&amp;action=diff
17:06 lucian joined #parrot
17:07 whiteknight Parrot doesn't really have a community code of conduct like they mention here
17:13 plobsing left #parrot
17:19 theory joined #parrot
17:19 allison Coke: I've attended hundreds of conferences and never been treated with anything less than total respect
17:19 jan joined #parrot
17:21 allison moritz: Larry setting an example for the whole Perl community does help
17:24 allison whiteknight: a code of conduct isn't a bad idea for every project, even Parrot (where we pretty much have an implicit code of conduct anyway)
17:25 whiteknight allison: right. I would hate to be caught in a position where we were having problems and didn't have a code written up already. But then again, the situation is such that we haven't felt like it's a priority yet
17:25 cotto_work ~~
17:25 whiteknight which is good and bad. Obviously we have a great community, but we don't want to be oblivious to things
17:25 * cotto_work prepares for epic backscolling
17:26 TimToady some of our code of conduct is written into the very first paragraph of perl6.org
17:27 PerlJam spokesbug ... spunky ... world-famous ... dedicated ... ah, there it is:  be nice   :-)
17:28 TimToady actually "know how to be nice"--you don't always have to be nice :)
17:30 * Coke reads the ubuntu code of conduct. that looks nice.
17:30 TimToady we are not nice to people who don't know how to be nice :)
17:30 Coke TimToady++
17:30 Coke TimToady: Thanks for not being a jerk!
17:30 allison Coke: yes, I've always been a big fan of Ubuntu's
17:32 dukeleto 'ello
17:32 dukeleto fbrito: looks like you are really improving our NewParrotDeveloperGuide, awesome!
17:32 PerlJam William Ury has some good ideas too:  http://www.ted.com/talks/william_ury.html
17:32 PerlJam :)
17:33 PerlJam I particularly like his opening story in that talk
17:33 allison Coke: did the article mention http://openrespect.org/?
17:33 dukeleto Is anyone planning on going to PyCon in Atlanta? There is a Virtual Machine Summit this year.
17:33 dukeleto actually, it is next year, but you know what I mean
17:33 dukeleto http://us.pycon.org/2011/
17:34 fbrito dukeleto: yes, but there are so much covered in docs/ that I don't know what else could I write there
17:34 dukeleto "This is a networking and discussion group for those involved in the implementation of Virtual Machines for dynamic languages, and is not Python specific. Invitees include developers from the CPython, JRuby, IronRuby, IronPython, PyPy, Jython, Parrot, LLVM, Lua, Smalltalk and JavaScript VM projects."
17:34 dukeleto fbrito: it just needs to be a good intro that points people in the right directions
17:34 allison dukeleto: I am, thought I didn't submit a talk
17:35 allison dukeleto: I've been to the VM summit every year since it started and found it valuable
17:35 dukeleto allison: I am playing with the idea of going, but figuring out logistics. If you find it valuable every year, that says a lot.
17:36 allison dukeleto: it partly depends on your interest in Python
17:36 allison while they say it's not Python specific... it actually is
17:37 allison so if you go, be prepared to listen and learn about a new way of doing things
17:38 allison and for the fact that no one there will be interested in Perl 6 at all (and will probably make jokes about it)
17:38 allison don't be defensive, be open
17:39 allison it's a great cross-cultural experience
17:39 dukeleto allison: that is actually the kind of environment I want to be in. I want to learn about what other VM's are doing, specifically PyPy
17:39 dukeleto allison: i think we have a lot to learn from Rubinius and PyPy
17:39 allison dukeleto: definitely
17:40 preflex left #parrot
17:40 TimToady some of which will be carping that their languages don't give them enough type annotations to optimize well  :)
17:40 dukeleto allison: I am not sure if I would go just for the summit or PyCon as well.
17:42 dukeleto TimToady: what are you focusing on with respect to Perl 6 these days?
17:42 dukeleto TimToady: anything in particular ? Or just making sure the washing machine doesn't stop?
17:43 dip left #parrot
17:45 dukeleto hmmm. PyCon costs $300 (early bird) and $350 for the normal rate.
17:45 preflex joined #parrot
17:45 TimToady mostly just making sure people don't block on lack of understanding
17:46 TimToady and a lot of actually playing with Perl 6 to see how it feels
17:46 dukeleto TimToady: I am very interested in working with other parrot hackers to port some of jnthn's 6model stuff to Parrot. Parrot's current meta object model is blocking the implementation of Ruby on Parrot and other HLL's
17:46 TimToady indeed, that feels quite convergent to me
17:46 dukeleto TimToady: I think it is awesome that Parrot can learn from Perl 6 as well as the other way around
17:47 dukeleto TimToady: convergence is happening, and that means stuff is getting righter
17:51 * TimToady is getting migraine auras that make it difficult to read anything; better wander off for now...
17:53 dukeleto TimToady: feel better! Take a walk and sniff some fresh air.
17:54 dukeleto So our GCI students have been complaining that our tasks are too easy. Time to take out the big guns.
17:57 dukeleto whiteknight: ping
17:58 whiteknight dukeleto: pong
17:58 TypeNameHere_____ left #parrot
17:59 jnthn dukeleto: I already started that port. It's in the nom branch of nqp-rx :)
17:59 jnthn dukeleto: Contributions *very* welcome. :)
17:59 dukeleto whiteknight: very nice write-up on the embed_api2 branch
18:00 dukeleto jnthn: can you tell me a bit more about the nom branch?
18:00 lucian sorry, catching up to the backlog
18:00 dukeleto whiteknight: question: Can we get rid of the nonesense where creating a parrot interp requires passing in any previously created interps?
18:00 lucian rubinius and pypy are actually nothing alike
18:00 dukeleto whiteknight: that gives me the heeby-jeebies WRT security
18:01 dukeleto lucian: that doesn't mean we don't have a lot to learn from them
18:01 * lucian has been reading a lot, perhaps at the expense of his uni
18:01 jnthn dukeleto: At the moment it contains a port of knowhow (pure prototype) and a class implementation (NQPClassHOW) written in NQP. Current state is trying to transition NQP to use that NQPClassHOW implementation. It also has a port of some of the reprs.
18:01 dukeleto lucian: Rubinius is effectively using LLVM right now. We aren't. PyPy has a very finely-tuned GC (their 5th). Ours still has some pointy edges.
18:01 lucian dukeleto: of course. but there's the general opinion that they're very similar, but for different languages
18:01 jnthn dukeleto: Note how I'm writing NQP's object model in...NQP. :)
18:02 lucian dukeleto: well, rubinius is a ruby VM written in C++ with llvm for jit
18:02 dukeleto lucian: i don't have that opinion. I just know I want parrot to learn a lot from both of them.
18:02 dukeleto lucian: yes, i know :)
18:02 jnthn dukeleto: So the core really does not offer much at all.
18:02 lucian dukeleto: good :) i probably read too many confused blog posts
18:02 whiteknight dukeleto: as I understand, you can create multiple interpreters without them needing to reference each other
18:02 jnthn Other than a little case to build everything on.
18:02 dukeleto lucian: the core Rubinius guys live in Portland, and I went to talk about it a few days ago.
18:02 jnthn s/case/base/
18:03 dukeleto lucian: they hang out in #rubinius on freenode, and are very friendly
18:03 dukeleto whiteknight: that is wrong
18:03 lucian dukeleto: yep, and they do great work
18:03 dukeleto whiteknight: i.e. untrue
18:03 whiteknight dukeleto: Okay, we can address that, though probably not in this branch
18:03 whiteknight dukeleto: create a ticket and assign it to me?
18:03 Coke allison: I didn't see that link, but might have.
18:03 lucian dukeleto: i like how they work around llvm's bad performance as an actual vm
18:03 dukeleto whiteknight: Parrot_new requires being passed in the 1st interp every created, or HORRIBLE THINGS WILL HAPPEN
18:04 lucian i'm much more familiar with PyPy, though
18:04 dukeleto s/every/ever/
18:04 dukeleto whiteknight: perhaps your branch isn't the correct place to fix it
18:04 dukeleto whiteknight: i think it is just sharing constants across interps, but still, from a security viewpoint, it is dicey.
18:05 dukeleto whiteknight: in PL/Parrot i have to create trusted and untrusted interps, but i have to tell each about the other, and I don't like that.
18:05 whiteknight dukeleto: yes, that is bad. We will fix that eventually
18:06 whiteknight first step is to come up with a failing test/example, and then research why it fails
18:08 dukeleto whiteknight: http://trac.parrot.org/parrot/ticket/1880
18:09 cotto_work allison: can you put the paper on the wiki?
18:09 dukeleto whiteknight: the documentation says that horrible, horrible things will happen, and i remember seeing something bad, but can't remember what it was
18:09 dalek TT #1880 created by dukeleto++: Parrot_new should not need to be passed previously-created interps
18:09 dalek TT #1880: http://trac.parrot.org/parrot/ticket/1880
18:11 cotto_work or whiteknight if allison has taken off
18:11 allison cotto_work: which paper? my PhD thesis proposal?
18:12 cotto_work yes
18:12 cotto_work That was what I pinged you about yesterday evening.
18:13 Coke re #1880 - it really requires the first interpreter and not just any ole interpreter?
18:13 lucian dukeleto: speaking of GC/JIT in parrot, the PyPy folks are more productive because they write RPython and not C
18:13 allison cotto_work: will skim through it quickly, but should be fine
18:13 Coke (I could see requiring "a".)
18:13 lucian dukeleto: i've been thinking whether anything similar would be possible for parrot
18:13 cotto_work allison: thanks!
18:14 fbrito left #parrot
18:17 whiteknight lucian: yes, being able to write something better than C has it's advantages
18:17 lucian whiteknight: even simple things, like not caring about memory when implementing parrot (except when implementing the GC of course)
18:18 whiteknight if our post-Lorito JIT brings with it an AOT compiler, it becomes very possible to do things like that ourselves
18:19 whiteknight we write Parrot in a high-level language, compile to PBC, optimize, compile to machinecode, and we have our binary
18:19 fbrito joined #parrot
18:20 cotto_work dukeleto: it looks like PyCon will be a good place to steal some ideas and hang out with some different kinds of hackers.  I'll have to see if I can avoid not going.
18:25 lucian whiteknight: i was thinking more of something like Vala
18:27 rfw joined #parrot
18:29 dukeleto Coke: i am not sure if it is only the 1st, or any
18:30 dukeleto lucian: i looked into Vala. It looks cool, but i heard debugging errors in it is a pain, since you are then debugging errors in generated C
18:31 rfw left #parrot
18:38 rfw joined #parrot
18:39 lucian dukeleto: sort of, yeah
18:40 lucian dukeleto: i was thinking of something more system-like
18:40 lucian even D would fit the bill
18:40 lucian but D brings a lot of un-parrot things with it
18:41 lucian something like SafeC + a few extra features would probably be enough
18:47 fbrito GCI mentors: I think I have finished my task. Please take a look when you have some time: http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129123759504
18:48 rfw fbrito: just woke up and you've already finished something :o
18:50 nwellnhof joined #parrot
18:51 smash left #parrot
18:54 lucian_ joined #parrot
18:56 lucian left #parrot
18:57 dalek tracwiki: v28 | cotto++ | NewParrotDeveloperGuide
18:57 dalek tracwiki: specify minimum perl version
18:57 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=28&amp;action=diff
19:00 fbrito rfw: it is still my wiki task from yesterday :P
19:01 lucian_ whiteknight: and I just realised that the lorito interpreter could be written in RPython and get a free JIT from PyPy
19:01 rfw fbrito: trying to do my screencast task :(
19:01 rfw it doesn't help that i keep going "the installer here lets you blehgabldgdahhfds"
19:02 cotto_work lucian_: the Lorito interp can be written in all kinds of things
19:02 cotto_work that's part of its charm
19:03 cotto_work doing it in RPython and getting a free JIT sounds cool.  That might even be a good place to do the initial filling out of the spec once the spec is ready.
19:06 whiteknight dukeleto: ping
19:09 lucian_ cotto_work: yeah, i assumed that as well
19:10 cotto_work I hope that no implementation will have a monopoly on Lorito interpreters.
19:10 Kristaba joined #parrot
19:11 fbrito (great. now I have to wash my mom's car and do grocery shopping. see you guys later)
19:12 Coke wow, usually we have to go pick up the kids from school or something.
19:23 atrodo cotto_work> That would make it nice.  That way parrot becomes a specification and not a program
19:28 cotto_work atrodo: yes
19:28 cotto_work atrodo: did you see that I'm meeting with allison and chromatic to get a brain dump on Lorito?  Now's the time to add questions to the wiki page.
19:29 atrodo cotto_work> I heard that, and I'm still trying to think of questions to little success
19:29 whiteknight cotto_work: I just assigned TT #61 to you after adding in info from the GCI student. I don't think it's a rush issue.
19:30 cotto_work whiteknight: good idea
19:30 cotto_work I was going to make gci tasks out of the tickets last night but ran out of steam before I finished looking through the ticket queue.
19:31 cotto_work It's mentally draining.
19:32 whiteknight yes it is
19:32 whiteknight I created a report on Trac to list just the open RFCs and I was trying to weed out those
19:32 Coke any kind of queue management is, yah. My condoloence.s
19:33 whiteknight breaking up into small lists is the only way to deal with it
19:33 cotto_work There's a handful that are tagged gci though.  When I have the tuits I (or someone) can make them into proper tasks.
19:33 whiteknight okay
19:34 cotto_work I can see much more clearly the value of having a small ticket queue now.
19:35 whiteknight MUCH more clearly
19:38 fbrito (oh, tomorrow is opendataday! http://www.opendataday.org/)
19:46 dukeleto whiteknight: pong
19:54 whiteknight dukeleto: the links on the /foundation page on parrot.org are broken now after the upgrade
19:55 whiteknight er, /foundation/legal
19:57 dukeleto whiteknight: bugger
19:57 whiteknight :)
19:57 dukeleto whiteknight: http://parrot.org/foundation/legal works for me
19:57 dukeleto whiteknight: what do you see?
20:00 whiteknight click one of the link
20:00 whiteknight I get "http://www.parrot.org/sites/www.parrot.org/f​iles/files/first_organizational_meeting.pdf"
20:02 Coke often things change for logged in/not users.
20:02 Coke (which in turn is usually a server side cache issue)
20:03 dukeleto whiteknight: i see now
20:04 * dukeleto just sent a support ticket to OSUOSL
20:05 whiteknight oh, that's not something we can fix?
20:05 whiteknight is that page editable?
20:05 dukeleto i have no clue. But the upgrade broke it, so i emailed them.
20:06 * dukeleto has never fiddled with parrot.org content
20:14 perlite_ joined #parrot
20:16 allison dukeleto: OSUOSL doesn't work on our Drupal site at all, we do that
20:17 dukeleto allison: sure. But something broke when they upgraded our site
20:17 dukeleto allison: so i am asking them for help. They are free to say "we don't care", but they are usually pretty helpful
20:17 allison dukeleto: they are very helpful
20:17 dukeleto allison: we had an old version of drupal that had various security issues, so they did an upgrade
20:18 dukeleto allison: yep, those OSUOSL peeps are awesome.
20:18 perlite left #parrot
20:18 perlite_ is now known as perlite
20:18 dukeleto I've never edited parrot.org in any way, so i don't know where the line between OSUOSL and us is, exactly.
20:18 sorear whiteknight++ whiteknight++ whiteknight++ whiteknight++ whiteknight++
20:18 Coke if you can do it through the web gui, it's us.\
20:19 sorear # embed_api2 report is puire awesome
20:19 Coke we have command line access so we can probably do anything, but I find it easier to ask them for help beyond that point.
20:20 dukeleto i just heard back from OSUOSL and they say the problem is on their end. They are looking into it.
20:20 dukeleto Those guys rock. They replied 14 minutes after I sent them an email.
20:22 allison dukeleto: basically, they do the os-level admin (the command-line stuff) and we do everything else
20:22 dukeleto allison: see above :)
20:22 allison dukeleto: if you don't have full admin permissions on parrot.org yet, let us know
20:22 allison dukeleto: (yup, saw above, and figured clarification was good for-the-record)
20:22 dukeleto allison: yes, i appreciate that.
20:23 dukeleto allison: I am concerned that some of our docs are not linked to from our site.
20:23 dukeleto allison: such as docs/embed.pod and many others
20:23 dukeleto how does http://docs.parrot.org/parrot/latest/html/ get updated ?
20:24 allison dukeleto: that's the whole purpose of docs.parrot.org
20:24 allison dukeleto: it's part of the release process
20:26 Kristaba Hi everyone
20:26 dukeleto Kristaba: howdy!
20:27 dukeleto allison: is the file at http://docs.parrot.org/parrot/latest/html/ actually under version control ? Is it generated from "make html" or something?
20:27 dukeleto allison: it is not clear from our release docs
20:27 allison dukeleto: generated by "make html"
20:29 allison dukeleto: item 10.f. in the release document
20:30 dukeleto allison: which file is index.html generated from? I still can't seem to find that
20:30 whiteknight sorear: thanks! But I didn't think it was worth all that karma!
20:30 allison dukeleto: it's not generated from a file, it's generated by the modules
20:31 dukeleto allison: please explain what you mean "generated by the modules" ?
20:31 dukeleto allison: all i want to do is add a few links to index.html . How do I do that? It is definitely not obvious.
20:32 dukeleto Kristaba: how is GCI treating you?
20:33 allison dukeleto: see lib/Parrot/Docs/Section/Parrot.pm
20:33 dukeleto allison: thanks. I've never looked at those modules.
20:34 fbrito can anyone please take a look on my GCI wiki task? (http://www.google-melange.com/gci/task/show/googl​e/gci2010/parrot_perl_foundations/t129123759504). I would like to claim another Parrot task but this one is blocking me :)
20:34 allison dukeleto: it's a simple declarative structure, add an item with a title and file path to a given group, and it'll do everything else for you (convert the pod to html, set up the site links, etc)
20:35 dukeleto allison: sweet!
20:35 dukeleto fbrito: on it
20:36 dukeleto fbrito: approved.
20:37 dukeleto fbrito: what are you working on next? I still need to make some challenging tasks for you.
20:37 fbrito http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129130043418
20:37 whiteknight yes, we need more challenging tasks
20:37 whiteknight I will try to make some more tonight, but I am at work late today
20:37 dukeleto fbrito: that looks really fun!
20:38 dukeleto whiteknight: you have been kicking my ass in the task-making department. I need to catch up.
20:38 fbrito :D
20:38 whiteknight dukeleto: You said you would manage the event if other people picked up the slack and didn't leave you with all the work
20:38 fbrito allison: look! it looks that now you have a portuguese wikipedia article on your name: http://pt.wikipedia.org/wiki/Allison_Randal
20:38 whiteknight this is just part of the deal
20:38 hercynium joined #parrot
20:39 dukeleto whiteknight: sounds good to me :) You are doing a great job mentoring. I have lots of task ideas, but life has been getting in the way.
20:39 whiteknight dukeleto: if you send me a list of task-idea oneliners I will put them in as tasks
20:39 whiteknight I'm actually slowly draining out my pool of ideas, so I need fresh blood
20:40 whiteknight and BRAINS
20:40 plobsing joined #parrot
20:40 dukeleto MMMMMMMM, BRAAAAAAAAAAAAAAAAAAINNNNSSS
20:41 dukeleto has anybody implemented LOLZOMBIE yet?
20:43 whiteknight LOLZOMBIE?
20:44 allison fbrito: cool! :)
20:45 allison fbrito: I like the sound of "Arquiteta Chefe do Ubuntu"
20:46 fbrito yes. and it sounds really important :P
20:47 fbrito (wow. portuguese wikipedia has no article to Perl6 and doesn't even mention it on Perl page)
20:47 whiteknight fbrito: You need to be our Arquiteta Chefe do Perl6 article
20:47 allison fbrito: or really flamboyant :)
20:48 fbrito whiteknight: ahahhaha, NICE! oh, wait a second... Arquiteta is for girls :(
20:48 allison fbrito: like I need a big red dress and 6" heels
20:49 * allison gets ideas for next halloween...
20:52 Kristaba dukeleto: (sorry for the delay) No problem, my task is about the PLA, so I ask help to whiteknight for now :)
20:55 whiteknight fbrito: sorry, I don't know portugese
20:56 fbrito whiteknight: no problem :P
21:01 dukeleto Sample LOLZOMBIE program: IF (I CAN HAZ BRAINS) FEAST() ELSE WANDER()
21:02 whiteknight dukeleto: nice syntax. What does it do?
21:02 Coke dukeleto: see also the branch for making it easier to change the html docs.
21:02 Coke (which I never quite made to a state that made it pushable to master.)
21:02 lucian_ dukeleto: halloween DSL for lolcode?
21:02 fbrito left #parrot
21:03 fbrito joined #parrot
21:03 allison fbrito: hmm... spanish does gender agreement jefe/jefa too
21:03 lucian_ allison: i'm pretty sure all latin languages do (i'm romanian)
21:04 lucian_ is now known as lucian
21:04 allison fbrito: portuguese doesn't do chefe/chefa?
21:05 allison fbrito: as in "arquiteta chefa"?
21:05 fbrito allison: no, it doesn't :)
21:05 allison fbrito: interesting (language geek here, asks curious questions)
21:05 dukeleto lucian: not only for halloween :)
21:05 fbrito allison: some words like "chefe" or "presidente" (president) doesn't do gender agreement
21:06 davidfetter joined #parrot
21:06 allison fbrito: makes sense
21:07 lucian allison: romanian does do something like that, though
21:07 lucian it's not consisdent, however. doctor doesn't vary with gender
21:07 fbrito I, particularly, find portuguese really hard to learn
21:08 fbrito but not as hard as german :)
21:09 fbrito there are some details, like: in english you only have 1 gender (the house, the dog, the cat). in portuguese you have 2 (masculine/feminine), and in german you have 3 (masc/fem/neuter) :(
21:10 rfw ohi Kovensky
21:10 allison fbrito: better than Swahili, it has dozens of genders
21:10 fbrito really?! ohh
21:10 allison fbrito: trying to remember how to pluralize all of them is a real headache
21:10 rfw how do you have more than 2 genders
21:11 lucian rfw: plenty of languages have three
21:11 rfw oh i meant 3
21:11 lucian neutral tends to be for things that don't reproduce
21:11 fbrito and in german you have the 4 cases (genitive, accusative, nominative and dative)
21:11 rfw but how do you have dozens of genders
21:11 lucian fbrito: that's also quite standard
21:11 allison rfw: they go off into all kinds of semantic details like "tree-like" or "cow-like"
21:11 rfw masculine, feminine, neuter, notquitemasculinebutslighylfeminine?
21:11 rfw oh
21:11 lucian rfw: i don't know. how do you only have one gender in a language? (and some do)
21:11 lucian allison: haha
21:12 allison rfw: but, they play exactly the same role in the language as the masculine/feminine distinction in other languages
21:12 Kovensky <@fbrito> there are some details, like: in english you only have 1 gender (the house, the dog, the cat). in portuguese you have 2 (masculine/feminine) <-- only because everything decays to masculine by default
21:12 lucian allison: that sounds like fun
21:12 rfw ohi Kovensky
21:12 allison lucian: it is fun
21:12 * rfw decays Kovensky to masculine
21:13 allison lucian: and once you learn the rules, it's pretty straightforward. Like, a noun with an "m" prefix always pluralizes with a "wa" prefix
21:13 lucian Kovensky: and it gets worse, in Hungarian there's no gender at all, you have to specify "male" or "female"
21:13 allison lucian: "mtu" person, "watu" people
21:13 Kovensky lol
21:13 Kovensky hungarian sounds funny ._.
21:13 lucian Kovensky: english still has it good
21:14 Kovensky tho I've only heard hungarian as sung by Dalriada
21:14 lucian Kovensky: it's funny to hear a Hungarian speak Romanian, as it often happens. they tend to get genders all wrong
21:14 Kovensky lol
21:14 lucian like: He's the chief architect(female) of Parrot. (speaking of allison)
21:15 Kovensky brazilians tend to get english sentence order wrong :/
21:15 whiteknight I have to sign off. Be on later
21:15 rfw bye whiteknight
21:15 lucian Kovensky: that's a bit funny too
21:15 whiteknight left #parrot
21:15 Kovensky and most of them memorize the meanings of words instead of learning them so they take EVERY word at face value
21:15 Kovensky and mostly only learn one or two meanings so they're completely lost at parsing semantics
21:15 rfw Kovensky: how is your english so good ._.
21:15 lucian Kovensky: meh, i guess that's a general rule for people not yet proficient in a certain language
21:16 Kovensky rfw: no idea =p
21:16 rfw Kovensky: you still have that accent though :(
21:16 Kovensky lucian: http://randomgraphs.blogspot.com/​2009/06/feynman-on-education.html
21:16 fbrito Kovensky: did you go to an english school or learned all by yourself?
21:17 Kovensky rfw: which is russian for some reason, at least according to Emess
21:17 rfw LOL
21:17 Kovensky fbrito: by myself
21:17 fbrito wow :D. really nice
21:17 rfw Kovensky: well, half of everyone online says i'm british
21:17 rfw so ._.
21:17 Kovensky fbrito: I attended classes on CCAA ~6 months
21:17 rfw inb4newzealandisacolonyofbritain
21:17 Kovensky +for
21:17 Kovensky but they weren't that much helpful
21:17 rfw -much
21:17 Kovensky rfw: at least you don't speak sheepese ;)
21:18 rfw baa
21:18 rfw time to make another retake at this screencast again :(
21:18 dukeleto rfw: i am excited to watch it soon :)
21:18 rfw dukeleto: i just don't want to mess it up five times over :(
21:18 rfw it's also very boring
21:19 dukeleto rfw: put some good music in the background
21:19 dukeleto rfw: i have seen lots of screencasts that do it. It really helps
21:19 rfw i could put that in later
21:19 rfw but my music collection isn't very uh
21:19 dukeleto rfw: perhaps some speed metal/
21:19 rfw it's very "unique"
21:19 rfw Kovensky knows
21:20 Kovensky I can send you some Moonsorrow
21:20 Kovensky :P
21:20 rfw wtf is moonsorrow
21:20 rfw is there a sunjoy
21:20 Kovensky some finns
21:20 Kovensky but I guess their song would be too "slow"
21:21 Kovensky so put Demetori
21:21 rfw maybe i should just put some instrumental tōhō in the background
21:21 dukeleto rfw: instrumental is good for a screencast
21:21 rfw yes, tōhō
21:21 Kovensky why are you using macrons
21:21 rfw i dunno
21:21 Kovensky stop trying to hide your hobbies :P
21:21 rfw shut up ;___;
21:21 rfw Kovensky: demetori is kinda fast though
21:22 rfw and loud
21:22 Kovensky isn't fast good
21:22 rfw is it?
21:22 rfw for a screencast?
21:22 fbrito it depends on how fast you speak, i guess
21:22 Kovensky if you want slow you could use that one where they slew down some Justin Bieber song by 8x and it sounds like some ambient music
21:22 rfw ohgodno
21:22 fbrito ahahhahaha
21:22 rfw also >slew
21:23 rfw Kovensky: so do they play baby x8 slower in elevators now
21:23 rfw er 8x*
21:23 Kovensky they should
21:23 Kovensky it actually sounded rather decent
21:23 rfw "when you hear it you will shit bricks"
21:23 rfw Kovensky: do you have a link
21:24 rfw i don't particularly want to google "justin bieber"
21:24 Kovensky http://www.youtube.com/wat​ch?v=JI0jAW7s0YY&amp;hd=1 <-- just use this :D
21:24 rfw oh i found it
21:24 Kovensky rfw: ask kuu on rizon
21:24 * dukeleto hits his gavel for order in the court
21:24 rfw oh, back on topic
21:24 * Kovensky goes back to mla
21:25 * dukeleto writes some tasks to challenge these darn youngins
21:27 dalek tracwiki: v29 | fbrito++ | NewParrotDeveloperGuide
21:27 dalek tracwiki: Add parrot-users list
21:27 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=29&amp;action=diff
21:37 Kristaba Can someone can check few lines I wrote using the PMC system, to see if my code seems correct, please?
21:37 Kristaba http://pastebin.com/urz0Y32g
21:40 jan left #parrot
21:42 jan joined #parrot
21:42 Kristaba If it's right, it should chack if the PMC is an array, and if it is, check if the array size is less than 3. Then, it get float value (if it's possible of course) of the first and the second element of this array
21:43 dalek tracwiki: v30 | fbrito++ | NewParrotDeveloperGuide
21:43 dalek tracwiki: Add GitHub README link
21:43 dalek tracwiki: http://trac.parrot.org/parrot/wiki/NewParro​tDeveloperGuide?version=30&amp;action=diff
21:43 fbrito Kristaba: about line 10
21:45 fbrito I think it sounds better if you do INTVAL size = VTABLE_elements(interp, value) on line 9 and line 10 just "if (size <=2)"
21:45 Kristaba Yes, you're right ;)
21:45 fbrito http://pastebin.com/CSxdDNw1
21:45 fbrito it sounds more readable :D
21:46 Kristaba Of course, thank you!
21:46 fbrito but I am not sure if your whole code is going to work. on which task are working?
21:46 Kristaba This task : http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129104235563
21:48 fbrito Really nice :D
21:49 Kristaba The PMC concept seems very powerful, but a bit complicated when you start to code with, I think :p
21:51 Kristaba Yes, this task is very interested, but need to learn a lot of different things
21:53 cotto_work dukeleto: no luck on a group ACM subscription.  The closest thing they have is corporate subscriptions that start at $very_expensive and go up.
21:54 fbrito Kristaba: I am working on: http://www.google-melange.com/gci/task/show/goog​le/gci2010/parrot_perl_foundations/t129130043418
21:55 rfw left #parrot
21:55 fbrito I need to make sure that $m[1,1,3]; and $m{Key.new(1,1,3)} throw exceptions :P
21:56 Kristaba fbrito: Yeah, very nice too! :D
21:57 fbrito Kristaba: But I am still figuring out what functions and macros (and in which order) are called when I do those kind of things
21:59 Kristaba fbrito: Yes, it's a *really* GCI difficult task ;)
22:01 Andy left #parrot
22:01 dukeleto cotto_work: perhaps OSUOSL can help
22:02 dukeleto cotto_work: they have ties to the university, who already pays for access, probably. Perhaps something can be done.
22:03 Benabik joined #parrot
22:06 rfw joined #parrot
22:06 cotto_work dukeleto: excellent idea
22:08 cotto_work asking about it now
22:17 cotto_work nope
22:21 Benabik Is it my imagination, or is most of the PCT documentation out of date?
22:23 dukeleto Benabik: it is not your imagination
22:23 dukeleto Benabik: which docs are you talking about in particular?
22:24 Benabik dukeleto: PDD29 and the "Implementing Languages on Parrot" section of docs.parrot.org
22:24 Benabik dukeleto: If I'm looking at the code right, things are spelled HLL:: instead of PCT::HLL and nqp-rx has absorbed PGE.
22:24 dukeleto Benabik: what part of the docs are you finding to be out of date? We will fix it.
22:25 dukeleto Benabik: yes, it is a somewhat recent change. All of our docs didn't get updated.
22:25 dukeleto Benabik: nqp-rx is the successor to PGE, but many things are still based on PGE.
22:25 dukeleto aloha msg whiteknight i have a task template for parrot tasks, this should make your life easier: https://github.com/leto/gci/blob/​master/templates/parrot_task.html
22:26 dukeleto is aloha down?
22:26 dukeleto cotto_work: see the above link as well
22:26 cotto_work aloha: help
22:26 aloha cotto_work: Ask me for help about: msg, convert, status, vars, karma, auth, seen, maths, translate, infobot, clock, loader (say 'help <modulename>').
22:26 cotto_work msg aloha aloha
22:26 Benabik dukeleto: I'm not sure what parts exactly are out of date.  Just starting to dive in and noticing that mk_language_shell.pl doesn't reference PGE or the PCT namespace.
22:26 cotto_work bacek: aloha doesn't seem to be taking messages
22:27 Benabik dukeleto: Also, running mk_language_shell.pl results in a few $Revision$ tags.
22:28 dukeleto Benabik: I am sorry to tell you that mk_language_shell.pl is broken right now, and it is mostly my fault.
22:28 dukeleto Benabik: I helped transition Parrot to Git, and all our tools were migrated, except mk_language_shell and create_language
22:29 Benabik dukeleto: Good news is that it still seems to produce a working language.  :-D
22:29 dukeleto Benabik: i think, after recent work from moritz++, they can be fixed easily
22:29 dukeleto Benabik: yes, only the way that it checks for the version of parrot is not working correctly
22:33 Benabik dukeleto: I'm working on a presentation on Parrot for a compiler class.  I'll try to write up a couple notes on where the docs don't match reality.  And if I get the time, a couple patches to parrot.git/docs
22:33 Benabik dukeleto: Knowing that things are known to be out of sync helps keep me from running in circles.
22:34 cotto_work Benabik: If you see something you think is inconsistent between the docs and the code, please let us know.  There's a good chance you're correct.
22:34 dukeleto Benabik: we welcome any and all doc patches that you can send us
22:35 dukeleto Benabik: and we would love to see your presentation about Parrot, if that is possible :)
22:36 Benabik dukeleto: I'm sure there'll be a PDF on the web at the end.  Don't expect anything too detailed, I've only got 15 minutes to give an overview. :-)
22:36 dukeleto Benabik: what level class is this for?
22:36 dukeleto Benabik: an introduction from the perspective of someone on the outside is exactly what we need :)
22:37 dukeleto aloha clock?
22:37 aloha dukeleto: dukeleto: LAX: Fri, 14:37 PST / CHI: Fri, 16:37 CST / NYC: Fri, 17:37 EST / UTC: Fri, 22:37 UTC / LON: Fri, 22:37 GMT / BER: Fri, 23:37 CET / TOK: Sat, 07:37 JST / SYD: Sat, 09:37 EST
22:37 Benabik dukeleto: Grad compiler construction.  Supposed to be a bit of an overview and comparison.
22:37 dukeleto aloha msg aloha anybody home?
22:37 Benabik cotto_work: Since I don't know what the code does yet, I'm not sure where it doesn't match, although the PCT book and PDD seems to reference PGE a lot more than the code generated from mk_language_shell.pl does.
22:37 dukeleto aloha msg whiteknight have all tasks from the GCI parrot wiki page been transferred to melange?
22:38 dukeleto bacek_at_work: is aloha not working? She doesn't respond to message requests.
22:38 lucian hmm http://quokforge.org/projects/horizon/wiki
22:38 dukeleto Benabik: very cool. is your pres only on parrot? Or is parrot just one thing you talk about?
22:39 Benabik dukeleto: Intending to talk about creating a new HLL, so it'll touch on what Parrot is, NQP-rx, and the rest of the PCT.
22:39 cotto_work Benabik: I mean if you find that the docs say one thing but the programs you're trying to use act differently.  I don't expect that you'll know what most of the code does quite yet.
22:40 Benabik cotto_work: So far, looks like mostly naming differences, but same functionality.
22:41 dukeleto Benabik: nqp-rx can do regexen, PGE only supported them in a very limited way. You can mention that.
22:41 rfw left #parrot
22:43 Benabik cotto_work: Although I don't know how well the PIR scripts that test grammars with PGE and PCT::HLL work anymore. (docs/book/pct/ch0[34]*)
22:55 bacek_mobile joined #parrot
22:56 bacek_mobile Cotto, meh. I broke my box at work yesterday.
22:57 rfw joined #parrot
22:58 fperrad left #parrot
23:15 preflex left #parrot
23:17 preflex joined #parrot
23:34 cotto_work seen whiteknight
23:38 Benabik left #parrot
23:41 dalek parrot/leto/embed_grant: eaa40a9 | dukeleto++ | docs/embed.pod:
23:41 dalek parrot/leto/embed_grant: s/type signature/function signature/ and clarify what Pi means
23:41 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/eaa40a93e5
23:44 dalek parrot/leto/embed_grant: 9d74e5f | dukeleto++ | docs/embed.pod:
23:44 dalek parrot/leto/embed_grant: Clarify what the structure of a Parrot function signature describes
23:44 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/9d74e5f11b
23:51 nwellnhof left #parrot
23:52 dukeleto Coke: yes, just verified. Our docs say the 1st ever initialized parrot interp must be used to create all others
23:53 bacek_at_work aloha, seen bacek
23:53 aloha bacek_at_work: bacek was last seen in msg 1 days 13 hours ago <private message>.
23:53 bacek_at_work ok, aloha is fixed now.
23:55 sorear What is the deal with that anyway?
23:55 sorear Why can't I just create interpreters whenever I feel like it?
23:56 dalek parrot/leto/embed_grant: 69533d8 | dukeleto++ | docs/embed.pod:
23:56 dalek parrot/leto/embed_grant: Document interpreter flags in docs/embed.pod
23:56 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/69533d841b
23:57 dukeleto sorear: i am loading bytecode in many different interpreters. Some I trust. Others I don't.
23:59 cotto_work aloha msg aloha aloha
23:59 aloha cotto_work: OK. I'll deliver the message.
23:59 cotto_work msg bacek_at_work thanks
23:59 aloha OK. I'll deliver the message.

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

Parrot | source cross referenced