Camelia, the Perl 6 bug

IRC log for #parrot, 2013-04-17

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
01:23 Psyche^ joined #parrot
02:56 preflex joined #parrot
03:44 dalek rakudo/nom: 4109be8 | coke++ | INSTALL.txt:
03:44 dalek rakudo/nom: RT #76828 - document memory build requirements.
03:44 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4109be8905
04:15 dalek parrot: 8874a86 | util++ | / (7 files):
04:15 dalek parrot: Prepare for the 5.3.0 release
04:15 dalek parrot: review: https://github.com/parrot/parrot/commit/8874a86fb1
04:39 Util Here come the drums.
04:40 pmichaud Util++
04:48 Hunger joined #parrot
04:49 myhrlin_ W00tstock
04:49 myhrlin_ nice
04:51 Topic for #parrot is now Parrot 5.3.0 "W00tstock Parrot" | http://parrot.org/ | Log: http://irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
07:50 kurahaupo joined #parrot
08:41 yawnt joined #parrot
08:41 yawnt hai
08:46 sorear Heya.
09:17 dalek rakudo/nom: 24532ac | (Brent Laabs)++ | src/core/IO.pm:
09:17 dalek rakudo/nom: inch IO::Path closer to the recent spec changes (path, volume, Str)
09:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/24532aca13
09:17 dalek rakudo/nom: 02b7524 | moritz++ | src/core/IO.pm:
09:17 dalek rakudo/nom: Merge pull request #126 from labster/nom
09:17 dalek rakudo/nom:
09:17 dalek rakudo/nom: inch IO::Path closer to the recent spec changes (path, volume, Str)
09:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/02b7524a58
11:02 PacoAir joined #parrot
11:08 saikiran_a joined #parrot
11:21 jlaire joined #parrot
11:26 PacoAir joined #parrot
11:28 xcombelle joined #parrot
12:44 dalek rakudo/nom: c946d0f | (Tobias Leich)++ | src/Perl6/ (3 files):
12:44 dalek rakudo/nom: bring runtime importation closer to spec
12:44 dalek rakudo/nom:
12:44 dalek rakudo/nom: This enables:
12:44 dalek rakudo/nom:     require PATH <imports>
12:44 dalek rakudo/nom:     require $varibale <imports>, and
12:44 dalek rakudo/nom:     require ModuleName:file(PATH) <imports>.
12:44 dalek rakudo/nom: So a given path or variable is always treated as a filename,
12:44 dalek rakudo/nom: and a module name can be specified to ease importing.
12:44 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c946d0f7f7
13:07 darbelo joined #parrot
13:17 dalek rakudo/nom: c39b887 | moritz++ | src/core/Mu.pm:
13:17 dalek rakudo/nom: remove reporting of variable name from uninit warnings
13:17 dalek rakudo/nom:
13:17 dalek rakudo/nom: they produced too many false positives
13:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c39b887388
13:17 dalek rakudo/nom: f3c3feb | moritz++ | t/spectest.data:
13:17 dalek rakudo/nom: track deletion of test file
13:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f3c3feb3e9
13:24 darbelo joined #parrot
13:52 saikiran_a joined #parrot
14:03 contingencyplan joined #parrot
14:10 awwaiid joined #parrot
14:29 Mike-PerlRecruiter_ joined #parrot
14:42 bouncy joined #parrot
16:06 darbelo joined #parrot
16:18 saikiran_a joined #parrot
16:33 saikiran_a joined #parrot
17:37 * tangentstorm wonders what it would take to really achieve cross-language interop.
17:42 tangentstorm http://equi4.com/minotaur/minotaur.html was another project with these ambitions from around the dawn of parrot.
17:43 tangentstorm It uses forth to allow perl, python, and tcl to integrate.
17:43 bluescreen joined #parrot
18:02 darbelo joined #parrot
18:09 alester joined #parrot
18:22 Mike-PerlRecruiter_ joined #parrot
18:30 rurban tangentstorm: maybe parrot
18:31 tangentstorm haha.. parrot isn't an answer to the problem yet, unfortunately. i'm asking what would have to be done to parrot to get it to that point.
18:31 rurban but this forth attempt surely looks convincing.
18:32 rurban parrot is at that point already
18:32 rurban it's just that nobody wants to use it, because its not complete and too slow
18:34 tangentstorm Okay, I have somewhere around 4-5 bodies of code that I'm trying to integrate, which use in java, python, javascript/coffeescript, pascal... And I'll probably want to merge in some stuff in perls 5 and 6, c, c++, and retro soon (retro being a forth dialect)
18:34 tangentstorm Can parrot do that?
18:35 rurban java, pascal not, the rest propably
18:35 rurban parrot is just the toolkit for external cross-langs
18:35 tangentstorm I'm skeptical.
18:36 tangentstorm Especially about python on parrot, having attempted the task myself in the past...
18:37 tangentstorm And also I asked about in here the other day.
18:37 rurban python = pynie, https://github.com/parrot/ecmascript is a bit old.
18:37 rurban it depends how much you need of it
18:40 * tangentstorm looks through the pynie source
18:40 tangentstorm Hm. Pirate actually does some of the stuff that pynie doesn't do yet.
18:40 tangentstorm Like list comprehensions.
18:41 tangentstorm It's cool that that objects are actually written in PIR though.
18:45 rurban This looks like a good overview of both: http://intertwingly.net/blog/1870.html
18:46 tangentstorm "Michal Wallace has given me commit access to Pirate,"
18:46 tangentstorm <- Michal Wallace :)
18:47 tangentstorm Sam came through and did a bunch of refactorings, and contributed some PMCs written in C.
18:48 tangentstorm Pynie seems to be a reboot by allison and pmichaud based on the newer parrot compiler tools.
18:48 Coke which are, no doubt, already out of date.
18:48 Coke (snark based on several ground up rewrites of partcl)
18:48 rurban yeah, long long time ago
18:49 tangentstorm Dan Sugalski also made a version early on. He had a bet with GvR that parrot would run faster than python. His approach was based on translating the python bytecode.
18:50 Coke Yes, and he ate a pie for his hubris. ;)
18:50 rurban :)
18:50 tangentstorm Pirate uses the AST provided by python.
18:50 tangentstorm yeah
18:50 Coke I was never able to partcl anywhere NEAR tcl speeds.
18:51 rurban wonder how he came up with his benchmark projections
18:51 tangentstorm The basic idea was that register machines in general should be faster than stack machines.
18:52 * tangentstorm looks at partcl
18:52 Coke Much hay was made of that, yes.
18:52 Coke tangentstorm: I have not touched it seriously in years now, apologies for any dust you find.
18:52 tangentstorm It's okay, pirate is still in a CVS repository. :)
18:55 tangentstorm Does anyone have an example of writing code in PAST somewhere?
18:55 rurban register machines yes. cps hmmm...
18:56 * tangentstorm thinks PAST isn't what he thought it was.
18:56 darbelo joined #parrot
18:56 allison tangentstorm: yes, Pynie and Pirate are both out of date
18:56 tangentstorm <- author of pirate :)
18:57 allison ah, is this a new pirate?
18:57 tangentstorm no it's an old pirate
18:57 allison well, feel free to kidnap any bits of Pynie that are useful in a reboot :)
18:58 tangentstorm somebody made a pirate for the PIR language but it was also the name of a python-to-parrot compiler
18:58 tangentstorm http://pirate.he.sabren.com/viewvc.cgi/pirate/
18:59 tangentstorm it was also the name of the original lua->parrot compiler by klaas-jan stol, but he renamed his
18:59 allison maybe it's time for a new name :)
19:00 allison there's also Puffin https://github.com/lucian1900/puffin
19:00 allison which is probably the most modern out of all of them
19:00 tangentstorm using winxed.. interesting.
19:06 tangentstorm hmm... looks like there's another pascal compiler called PAPAgei listed on http://trac.parrot.org/parr​ot/wiki/AbandonedLanguages but google's not being helpful with the source... anyone know where it is?
19:06 tangentstorm porcupine only implements a few bits and pieces.
19:11 rurban targetting pascal looks real trivial though. looks like a two days job
19:12 tangentstorm hmm.. well there's ISO pascal and then there's the pascal people actually use, which is delphi/object pascal.
19:12 tangentstorm which has generics, classes, units, iterators, lambdas, etc...
19:12 rurban And the Borland had a much better library
19:17 tangentstorm Yeah but the library is all written in pascal thanks to http://freepascal.org/
19:32 Coke tangentstorm: if I ever restart partcl AGAIN, I will consider using winxed.
19:32 Coke but my current effort is using rakudo's nqp.
19:33 Coke (but then I'm targetting nqp, not parrot. but it's definitely more pleasant for me to write nqp that it was to write PIR)
19:33 Coke *than
19:33 darbelo joined #parrot
19:57 awwaiid joined #parrot
20:10 gtodd Coke: is winxed errm "completed"  i.e. not actively developed but not needing development
20:12 Coke gtodd: It's not my project. I can't answer that question. (it's an exterior project that parrot snapshotted for internal use when it was usable.)
20:12 Coke (much like parrot-nqp)
20:13 gtodd ah ok so it's "inside" parrot
20:13 gtodd hmmm winxed.org
20:13 gtodd heh
20:55 perlite joined #parrot
21:12 darbelo joined #parrot
21:52 dalek nqp/qrpa_experiment: d0af2bb | (Timo Paulssen)++ | src/6model/serialization.c:
21:52 dalek nqp/qrpa_experiment: add serialisation for qrpa
21:52 dalek nqp/qrpa_experiment: review: https://github.com/perl6/nqp/commit/d0af2bb408
21:52 dalek nqp/qrpa_experiment: af36e92 | (Timo Paulssen)++ | src/ (12 files):
21:52 dalek nqp/qrpa_experiment: search&replace nqp::list -> nqp::qlist; change .push to nqp::push etc
21:52 dalek nqp/qrpa_experiment: review: https://github.com/perl6/nqp/commit/af36e9244e
21:54 Alxandr joined #parrot
21:57 Alxandr Hello. I've been considering trying to develop a language for the Parrot VM, and I'm just wondering, what languages can said language be developed in? (like C/python/java/whatever)
21:58 masak you could do it in nqp.
21:58 Tene Alxandr: As long as you're producing Parrot bytecode, you can produce it in any language you like; however, there aren't many facilities for producing parrot bytecode from non-parrot languages.  The best you'll get conveniently from another language is generating an intermediary language, like PIR, which has some problems.
21:59 Tene Most language development on Parrot is done from nqp or winxed, I believe.
21:59 Tene I don't think there are any others these days.
22:01 Alxandr Cause, I tried the tutorial found somwhere on the parrot-site (can't recall where, it starts by calling a perl-script that generates the "sceleton" for your compiler), and I had problem getting even basic stuff working
22:01 Alxandr And I have no experience with perl, so I had no idea what the errors I was getting was either.
22:02 Alxandr And it was sort of ... discurraging when I couldn't even get a simple AST-parser to work -.-
22:07 Alxandr But in other words, there are no bindings to generate PAST in non parrot-languages like C?
22:08 kurahaupo joined #parrot
22:11 Alxandr winxed seems to be written in C++ though
22:13 dalek nqp/qrpa_experiment: 2102ba9 | (Tobias Leich)++ | src/ (5 files):
22:13 dalek nqp/qrpa_experiment: more list -> qlist
22:13 dalek nqp/qrpa_experiment: review: https://github.com/perl6/nqp/commit/2102ba9545
22:55 dalek nqp/qrpa_experiment: 491eec9 | (Timo Paulssen)++ | src/ (6 files):
22:55 dalek nqp/qrpa_experiment: so many search and replace. probably too many.
22:55 dalek nqp/qrpa_experiment: review: https://github.com/perl6/nqp/commit/491eec9193

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

Parrot | source cross referenced