Camelia, the Perl 6 bug

IRC log for #parrot, 2010-01-22

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 TiMBuS joined #parrot
00:37 kurahaupo joined #parrot
00:37 kurahaupo1 joined #parrot
00:54 sri_ joined #parrot
00:54 abqar joined #parrot
00:56 sri_ joined #parrot
01:06 sri_ joined #parrot
01:18 mikehh joined #parrot
01:19 cotto_work http://www.archive.org/details/Visu​alDebuggingWithCustomPrettyPrinters
01:26 sri_ joined #parrot
01:28 sri joined #parrot
01:30 cotto_work It's Python-centric but looks pretty shiny.
01:43 cotto_work It might be fun to write some pretty printers for PMCs and other Parrot structs.
01:48 cotto_work a more textual introduction: http://tromey.com/blog/?p=524
02:04 sri joined #parrot
02:24 mtk joined #parrot
02:24 mtk left #parrot
02:37 eternaleye joined #parrot
02:46 cconstantine joined #parrot
02:50 kurahaupo joined #parrot
03:04 davidfetter joined #parrot
03:11 kid51 joined #parrot
03:30 purl joined #parrot
03:34 cotto_w0rk joined #parrot
03:49 cognominal joined #parrot
04:05 cconstantine_ joined #parrot
04:05 abqar joined #parrot
04:18 sri joined #parrot
04:20 sri joined #parrot
04:29 sri left #parrot
04:37 sri joined #parrot
04:48 sri joined #parrot
04:59 kurahaupo joined #parrot
05:03 cspencer joined #parrot
05:05 cspencer i'm attempting to compile parrot under OS X and having some issues...anyone out there able to help?
05:11 cspencer left #parrot
05:43 cotto Some of us can help and some can try.  Let's see into which category I fall.
05:46 cotto seen kid51
05:46 purl kid51 was last seen on purl 2 hours, 32 minutes and 36 seconds ago, saying: <private message>
06:36 particle1 joined #parrot
06:56 dukeleto 'ello
06:58 davidfetter all yuor parrotz are belong to us
06:58 dukeleto davidfetter: funny seeing you here
06:58 davidfetter heh
06:59 uniejo joined #parrot
07:15 davidfetter joined #parrot
07:17 patspam joined #parrot
07:29 dalek tapir: 616aec8 | dukeleto++ | t/harness.pir:
07:29 dalek tapir: Alter the format of the runtime a bit
07:29 dalek tapir: review: http://github.com/leto/tapir/commit/61​6aec88144a31c08ed98ef71357adaf0d964c0d
07:29 dalek tapir: 0d642b9 | dukeleto++ |  (3 files):
07:29 dalek tapir: Implement "Bail out!" parsing
07:29 dalek tapir: review: http://github.com/leto/tapir/commit/0d​642b9666c6b7b7252261b2dd34b60b62445770
07:38 dalek parrot: r43532 | chromatic++ | trunk (3 files):
07:38 dalek parrot: [IMCC] Forbade use of relative macro labels in PIR (in IMCC anyway).  This
07:38 dalek parrot: fixes the segfault reported by Will Coleda in TT #902.  His test case is now a
07:38 dalek parrot: test.
07:38 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43532/
07:39 dukeleto chromatic: where does the magic constant of 12 come from?
07:42 fperrad joined #parrot
07:42 dukeleto fperrad: howdy
07:43 chromatic I think I know.  I didn't like it much myself when I saw it.
07:46 dalek TT #902 closed by chromatic++: segfault with macro
07:46 chromatic Alright, this should make IMCC a little bit clearer, though something something whistling in the wind something something.
07:49 * dukeleto whistles a hypnotic tune from the Elder Ones
07:51 dukeleto has anybody played around with NekoVM?
07:52 dalek tapir: 0a7c40f | dukeleto++ | t/02-parse_tapstream.t:
07:52 dalek tapir: Add tests to start a Bug Hunt
07:52 dalek tapir: There is a bug when printing out how many failing tests occured:
07:52 dalek tapir: (kadath)(~/git/plparrot master )$ parrot t/harness.pir t/test.sql
07:52 dalek tapir: t/test.sql .. failed 2/3 tests
07:52 dalek tapir: FAILED 1 test(s) in 1/1 files (7.5512 seconds)
07:52 dalek tapir: review: http://github.com/leto/tapir/commit/0a​7c40fb74fd5a7cca456f10b293484b936a8cab
07:54 dalek parrot: r43533 | chromatic++ | trunk/compilers/imcc (2 files):
07:54 dalek parrot: [IMCC] Refactored a magic constant into intention-revealing code and commented
07:54 dalek parrot: another magic constant.  It's IMCC, so it's a minor change, but it's a minor
07:54 dalek parrot: change for the better.  Credit goes to dukeleto for the suggestion.
07:54 purl dalek: that doesn't look right
07:54 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43533/
08:00 * chromatic can't wait for one_make
08:02 cotto If all configure-generated files need to be build-generated, it'll be a while.
08:03 cotto The digest PMCs would be a good next step.
08:03 chromatic I'd settle now for better dependency tracking to avoid unnecessary rebuilding.
08:04 cotto It wouldn't be a bad idea to nail down an end goal upon the meeting of which the branch could be merged.
08:06 chromatic Coke seemed to see a near merge point, even if the branch doesn't meet all of its goals yet.
08:09 cotto I'm all for it.
08:11 chromatic Two failing tests for TT #389.
08:18 mikehh_ joined #parrot
08:28 barney joined #parrot
08:30 chromatic ... and that may do it.
08:43 dalek parrot: r43534 | chromatic++ | branches/tt389_fix (2 files):
08:43 dalek parrot: [PCT] Made PCT and PGE use objects and classes to look up methods, rather than
08:43 dalek parrot: poking in the namespaces themselves.  This fixes the test failures in
08:43 dalek parrot: t/compilers/pct/complete_workflow.t.
08:43 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43534/
08:43 dalek parrot: r43535 | cotto++ | branches/one_make/tools/dev/checkdepend.pl:
08:43 dalek parrot: [checkdepend] minor simplification, add a couple quote-munching regexes
08:43 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43535/
08:58 JimmyZ joined #parrot
09:00 JimmyZ joined #parrot
09:13 JimmyZ joined #parrot
09:16 JimmyZ joined #parrot
09:31 chromatic msg Tene Down to the P6Object failure in the TT #389 branch.  r43536 is unpleasant; if you have a better solution, there's the place to fix it.
09:31 purl Message for tene stored.
09:32 dalek parrot: r43536 | chromatic++ | branches/tt389_fix/compilers/pge/PGE/Exp.pir:
09:32 dalek parrot: [PGE] Modified PGE::Exp to look up rules in the current grammar from the
09:32 dalek parrot: appropriate class, not the attached namespace directly.  This is not efficient,
09:32 dalek parrot: but it works.
09:32 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43536/
09:37 cognominal joined #parrot
10:00 mj41 joined #parrot
10:04 jsut_ joined #parrot
10:36 payload joined #parrot
11:23 uniejo joined #parrot
11:24 mj41 joined #parrot
11:30 clinton joined #parrot
12:07 bluescreen joined #parrot
12:20 preflex joined #parrot
12:23 cconstantine joined #parrot
12:29 payload joined #parrot
12:39 LaVolta joined #parrot
12:40 LaVolta err...may be there's the right place to talk about a build error?
12:42 LaVolta sorry...i think someone already issued an ticket, sorry for the disturbing
12:43 ruoso joined #parrot
13:01 LaVolta joined #parrot
13:14 sri joined #parrot
13:48 tetragon joined #parrot
13:54 payload joined #parrot
13:59 smash joined #parrot
13:59 smash hello everyone
14:09 JimmyZ joined #parrot
14:10 cognominal joined #parrot
14:24 riffraff joined #parrot
14:49 dalek TT #1411 reopened by doughera++: [BUG] configure problem on the arm architecture
15:21 bluescreen joined #parrot
15:21 payload joined #parrot
15:28 bubaflub joined #parrot
15:44 solarion joined #parrot
15:45 Psyche^ joined #parrot
15:46 davidfetter joined #parrot
15:59 theory joined #parrot
16:01 theory_ joined #parrot
16:08 darbelo NotFound++ # "And you need a computer."
16:19 Coke that was a little over the top. =-)
16:19 Coke msg kid51 - re: the XXX; no worries until we're ready to merge back to trunk.
16:19 purl Message for kid51 stored.
16:33 cspencer joined #parrot
16:36 cspencer left #parrot
17:03 japhb What's the current idiom for creating and filling a raw buffer with 8-bit data, ready to send off to a C lib?  I want to generate texture data for some new OpenGL example code, which means big buffers and lots of accesses to fill them, so I need an efficient method.
17:05 cotto_work joined #parrot
17:18 cotto_w0rk joined #parrot
17:20 jan joined #parrot
17:20 cotto_working joined #parrot
17:22 theory_ joined #parrot
17:23 darbelo japhb: c level or PIR level?
17:27 japhb darbelo, PIR level preferred.  If I have to write C, I will, but only to write some API, PMC, or class that I can then access from PIR.  :-)
17:29 japhb I wonder if the parrot-linear-algebra project has something efficient I can use ...
17:29 darbelo Hmm. We might have a way to hadle that, but I can't think of any.
17:30 japhb darbelo, ah, OK.
17:30 japhb afk for a while but will backlog
17:30 darbelo japhb: pla uses custom PMCs
17:36 ruoso joined #parrot
17:55 dukeleto 'ello
18:01 davidfetter oi ruoso
18:01 ruoso oi davidfetter
18:01 davidfetter e aí, novidades?
18:02 davidfetter 'sup dukeleto
18:04 ruoso davidfetter, all work and no play makes ruoso a dull boy ;)
18:04 davidfetter play more :)
18:10 clinton joined #parrot
18:13 dukeleto davidfetter: hola
18:13 purl hola, dukeleto.
18:13 Andy joined #parrot
18:14 davidfetter ¿que paso?
18:15 darbelo I don't know, but it turned one of your question marks upside down ;)
18:16 clinton ¡NO!
18:16 dukeleto davidfetter: i worked on Tapir a bit last night. I was close to adding it to PL/Parrot, but then I found a bug and started a Bug Hunt
18:17 davidfetter heh
18:18 davidfetter oh, that reminds me. are we pulling from theory's repo for pgtap, or...?
18:18 davidfetter we == pl/parrot
18:19 theory davidfetter: Wouldn't generally recommend pulling from the repo, as things can break on non-8.4 builds
18:19 theory though at the moment it works fine on 8.0-8.5.
18:19 theory But I generally do dev on 8.4 and only fix it for earlier versions when preparing a release.
18:19 davidfetter k
18:20 * davidfetter pretty uninterested in anything < 8.4
18:20 theory yeah
18:20 theory then it should be fine.
18:20 davidfetter and could be persuaded that being interested in >= 9.0 is the thing
18:21 theory heh
18:22 dukeleto for now, pgTAP is in the PL/Parrot repo
18:23 dukeleto theory: have you thought more about splitting pgTAP into two pieces?
18:23 dukeleto theory: PL/Parrot only needs Test::More-ish things in pgTAP, not the schema stuff, which is VERY pg version-specific
18:23 dukeleto and I meant to say "for now, we will keep a copy of pgTAP in the PL/Parrot repo"
18:24 dukeleto but I am open to git submodules or somesuch
18:24 theory dukeleto: no, I don't have the tuits to do it.
18:24 * theory is way way overcommitted at the moment
18:25 * davidfetter sets theory's overcommit setting to -17 to avoid the OOM killer
18:25 davidfetter oh, wait. you don't use linux much, do you?
18:26 davidfetter nm
18:26 theory heh
18:26 dukeleto OH SNAP
18:26 dukeleto theory: then I am going to cannabalize our copy of pgTAP and rip out the schema stuff
18:27 dukeleto theory: so we will not be as version-specific
18:30 theory dukeleto: As you wish.
18:30 theory If you have a way to break it up and want to contribute it back, that'd be great.
18:38 plobsing joined #parrot
18:41 dukeleto theory: I am good at breaking things.
18:44 riffraff joined #parrot
18:44 * theory is too
18:53 mikehh joined #parrot
19:00 mikehh joined #parrot
19:05 mikehh joined #parrot
19:08 dalek rakudo/master: d872a6f | pmichaud++ | build/PARROT_REVISION:
19:08 dalek rakudo/master: Bump PARROT_REVISION to Parrot 2.0.0 release.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/d​872a6fdf735dcc7a2ae9e8fde133c2f5a39a53f
19:08 dalek rakudo/master: 98f43b6 | pmichaud++ | docs/ChangeLog:
19:08 dalek rakudo/master: Update ChangeLog for release.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/9​8f43b6f2a5eca2c42054319e5974a519d72a4f2
19:08 dalek rakudo/master: c548b61 | pmichaud++ | docs/spectest-progress.csv:
19:08 dalek rakudo/master: Update spectest-progress.csv .
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/c​548b61fc214b30677b8cedbca6d0db8fecfadf0
19:08 dalek rakudo/master: e64677c | pmichaud++ | docs/announce/2010-01:
19:08 dalek rakudo/master: Add draft announcement for release #25.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/e​64677c95799aaf11559ce283f620740b36fd6fa
19:08 dalek rakudo/master: 5c70e10 | pmichaud++ | docs/announce/2010-01:
19:08 dalek rakudo/master: Add notice about future change to release identifier format.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/5​c70e10b5a3aa95eed0ea74ab89e7c9c443d1f57
19:08 dalek rakudo/master: 3c7289b | pmichaud++ | docs/release_guide.pod:
19:08 dalek rakudo/master: Small improvement to guidelines for failing spectests at time of release.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/3​c7289b49261e2c0e15da01b76f3ae8eb8089f70
19:08 dalek rakudo/master: 4da4e4d | pmichaud++ | docs/release_guide.pod:
19:08 dalek rakudo/master: Update release dates and names.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/4​da4e4d7bd69f0b72d846aa8e273fd95ac0d3ad2
19:08 dalek rakudo/master: 6fc7264 | pmichaud++ | README:
19:08 dalek rakudo/master: Update copyright date.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/6​fc72646bb29f25d0d8ca2fe5ab74883839a1057
19:08 dalek rakudo/master: a609d77 | pmichaud++ | docs/announce/2010-01:
19:08 dalek rakudo/master: Last announcement update before release.
19:08 dalek rakudo/master: review: http://github.com/rakudo/rakudo/commit/a​609d778df4790801ff24bf45d3abe9770f62a66
19:09 chromatic joined #parrot
19:15 dukeleto wheeee!
19:19 chromatic Why are my edits to pages on parrot.org not sticking?
19:20 chromatic "This content has been modified by another user, changes cannot be saved."
19:21 chromatic Whee.  First Drupal doesn't understand HTML (line breaks in tags?  ALLOWED!) and then it fails to do what a publishing system should let you do (publish things).
19:21 chromatic The edit button is a tease.
19:23 chromatic Oh.  It's aggressive locking, even if I'm the one holding the lock.  Brilliant.  http://drupal.org/node/394694
19:23 Tene But if it does that, how can you edit while you edit?
19:24 chromatic You follow its intended workflow to the letter, to comply with its locking strategy.
19:24 chromatic DO NOT use the back button on your browser.  It is a trap.  Do not close the editing window, you hopeless fool!
19:25 chromatic Above all, never EVER make typos.
19:29 ash_ joined #parrot
19:29 Coke YO, I HEAR YOU LIKE EDITING HTML SO I PUT AN EDITOR IN YOUR EDITOR SO YOU CAN EDIT WHILE YOU EDIT!
19:29 Coke </meme>
19:31 chromatic Hm, and somehow I forced the change through.  Somehow.
19:33 Tene you forced an edit on an unwilling, unconsenting wiki?!
19:33 mikehh joined #parrot
19:34 chromatic Motivated by a comment on the release announcement: http://use.perl.org/~chromatic/journal/40114
19:35 chromatic Oh joy, and here too: http://news.ycombinator.com/item?id=1070094
19:41 Coke he donated money? i wonder where to.
19:42 chromatic TPF, back in the early days I'm sure.
19:42 Coke the people comment is essentially true. I'm not convinced about the design myopia, but I can see how it's readable that way.
19:43 Coke (specifically regarding perl6. I think we have design issues, but I don't think that's one of them.)
19:43 chromatic I don't care about that part so much.
19:44 chromatic The part that bothers me is "I didn't do my research, I don't know that there's a Python 3 implementation under way, and therefore no one should take the project seriously."
19:45 chromatic And it's easy to predict the response.  "Oh, it's just a partial implementation of Python 3.  I didn't look at the code and I don't know the Python 3 specification, but I'm comfortable predicting that it'll fail, and therefore no one should take you seriously."
19:48 Tene chromatic: they don't want a python "under way", they want it "done". :)
19:48 Tene to be fair, it's only allison working on it lately, and she's not very available.  Last commit on pynie was 5 weeks ago.
19:50 chromatic I think rather Ben wants to be critical and seen as brilliant rather than informed.
19:51 Tene Yeah, probably. :)
19:51 Coke msg moritz I see I have sudo on timtowdi. what's the pkg manager to use there?
19:51 purl Message for moritz stored.
19:52 chromatic By the way, I think I figured out the problem in the P6object test.
19:52 chromatic Methods get looked up in different places for C-based PMCs than PIR classes.
19:52 darbelo It's not called PyObject
19:53 Tene Ah.
19:53 chromatic r43398 (Git 3e420e7a025) made that change.
19:53 chromatic The find_method VTABLE is different for the default PMC and the Class PMC.
19:53 chromatic I think the right solution is to unify the two.
19:53 Tene darbelo: the plan is for pynie to not use p6object, but reimplement its own object system.
19:53 chromatic See also find_method_direct_1() in src/oo.c.
20:15 * PerlJam just read the btillyism
20:16 * darbelo googles btillyism
20:16 PerlJam I don't see what his problem is, the announcement says Parrot is "aimed at running all dynamic languages"  it doesn't say that Parrot currently runs all dynamic languages
20:17 darbelo All dynamic languages == Python
20:18 darbelo We don't run python, therefore we are irelevant.
20:19 Coke that whole history with sam was painful.
20:20 Tene That was before my time.
20:21 chromatic The project was doomed then.
20:22 chromatic I remember trying to figure out how to release a new version.
20:24 chromatic Did I call his logical fallacy or what?  http://use.perl.org/comments​.pl?sid=44429&amp;cid=71555
20:25 leto joined #parrot
20:25 Tene so what's the story with sam?
20:26 chromatic Sam had a lot of trouble supporting Python semantics.
20:26 chromatic He made some suggestions, and I thought they were mostly reasonable.
20:26 chromatic Leo didn't.
20:26 chromatic Dan let them fight it out.
20:26 darbelo Ouch.
20:27 chromatic Sam didn't want to fight.  (I don't know if Leo did either, but the spoken language barrier hurt when discussions grew heated.)
20:27 chromatic Sam decided it wasn't worth the effort.
20:33 chromatic I think this is also the time when you couldn't run much beyond simple PASM without triggering lots of segfaults.
20:37 ash_ is pynie under active development still?
20:37 dukeleto ash_: from what I know, people are still interested in hacking on it, but development has been slow recently
20:37 darbelo ash_: Yeah, but it move slowly.
20:38 dukeleto ash_: allison is very interested in pynie, but has classes to deal with
20:38 Tene ash_: I've been very interested in working on it for quite a while, but haven't had the time.
20:39 Tene the last time I worked on it, I stopped because of a lack of :vtable('invoke') override, and never started again.
20:39 Tene Looks like we have that now, though.
20:40 PerlJam So Ben may inadvertently spur pynie development  :)
20:41 chromatic How can he?  It doesn't exist!
20:42 ash_ PerlJam: i feel like there is a cowboy joke in there somewhere
20:43 PerlJam ash_: only if we're in Blazing Saddles
20:56 ash_ anyone know how cardinal is doing?
20:57 Tene cardinal stalled on implementing a new object system
20:57 mikehh joined #parrot
20:57 Tene treed ran into a lot of trouble with it, because of the inability to subclass Class.
20:59 chromatic Are there tests for that somewhere?  I think I saw a ticket.
21:00 joeri joined #parrot
21:00 darbelo Tene: He also got a job IIRC. Which limited his time to hack on it.
21:00 Coke JOBS!?
21:00 purl jobs are overrated
21:00 Tene Yes, that's right.
21:00 * Coke ponders giving up on partcl and focusing on pynie instead.
21:01 Coke (in an homage to alice's restaurant.)
21:01 Tene how's lua doing these days?  does it run on 2.0.0 ?
21:01 jsut joined #parrot
21:01 Coke . o O (allison's restaurant?)
21:01 chromatic She did that as a song at one conference once.
21:01 PerlJam allison++  That would have been neat to see
21:02 Coke NO SINGING!
21:02 Coke (one day, all these languages will be yours...)
21:03 chromatic http://www.radwin.org/michael/blog​/2003/07/perl_lightning_talks.html
21:04 PerlJam chromatic: where's the video/audio?  ;)
21:04 chromatic I don't know if any exists.  allison might still have the lyrics somewhere.
21:05 PerlJam Coke: are unimplemented languages like credit card debt that we should put most resources into each language one at a time?
21:06 Coke PerlJam: it is certainly frustrating to have N very incomplete languges.
21:06 chromatic That assumes developer time, motivation, interest, and skill is fungible.
21:06 Coke I'm willing to funge mine a bit.
21:06 PerlJam Coke++
21:06 Coke But realize not everyone is.
21:06 chromatic As am I, but I don't even assume that on so-called "professional" projects.
21:07 Coke but *I* rea...
21:07 Coke (that was not imperitive!)
21:07 chromatic It wasn't even categorical.
21:07 Coke step one stopping me from contributing to pynie is that it's in hg!
21:08 Coke (once hg is on timtowtdi.org, I can at least check it out.)
21:09 chromatic Is there no git-hg bridge?
21:09 treed Yeah. I got a job, and fulltime work combined with the fact that I live several hundred miles from my girlfriend (thereby spending every other weekend travelling) severely limits my time to work on cardinal.
21:09 treed Plus the last time I looked at it, I got stumped, and wanted some help with it.
21:10 treed But it's been so long that I don't remember exactly what the issue was.
21:10 PerlJam chromatic: I thought there was, but I've never been that interested in hg to know for sure.
21:11 chromatic I used it for Drizzle.  It works.  It's not a lot different from Git, but Git was faster.  This was a couple of years ago.
21:11 dalek parrot: r43537 | mikehh++ | trunk/t/op/trans_old.t:
21:11 dalek parrot: set svn properties on t/op/trans_old.t
21:11 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43537/
21:14 Tene treed: the issue is that you couldn't subclass Class
21:14 treed Tene: that's a different issue
21:14 Tene oh?
21:14 treed I was working around that and had a different issue.
21:15 dukeleto chromatic: there is a git-hg bridge
21:15 treed I created this elaborate scheme of doing cardinal;Class has-a parrot;Class
21:15 treed Which makes my head hurt at times.
21:15 treed But it was kinda working.
21:15 treed And then didn't work for a particular case.
21:15 treed Integer I think.
21:15 treed In the class declaration function.
21:16 treed 3-4 classes get made, and then it gets to one and fails
21:17 treed I never actually tried to subclass parrot;Class
21:17 treed I just kinda assumed from the beginning that it wouldn't work.
21:18 treed Probably because I had so many issues in the past with subclassing.
21:18 treed Then you said "Why aren't you just subclassing?" and then you tried it and it didn't work.
21:20 treed Man I wish we were using git natively.
21:21 dukeleto treed++
21:21 treed oh, wait
21:21 treed That was a mischan.
21:21 treed I meant using git natively at work.
21:21 dukeleto treed: no, it wasn't ;)
21:21 treed But the statment still stands.
21:21 PerlJam heh
21:21 treed We use SVN here, and I use git-svn.
21:21 * dukeleto just switched to git at $work and is very happy
21:21 treed But local branches aren't as fun when you git svn rebase and the commit tree screws up.
21:22 treed because the id changes
21:23 Tene treed: never ever change the history of anything pushed to svn.
21:23 treed Tene: I don't.
21:23 treed But in the case of: git svn rebase, make new branch, hackhackhack, merge branch back to master and dcommit
21:23 treed then that branch is useless.
21:24 treed Well, I guess you can still merge or something?
21:24 treed But you can't just -d it anymore
21:24 treed because it's not an ancestor
21:25 dukeleto git-svn is a swiss-army-ICBM
21:25 theory joined #parrot
21:26 treed I'd also like the ability to push a local branch and say "Check my branch here"
21:27 ash__ joined #parrot
21:28 ash___ joined #parrot
21:28 treed I tend to work on master, and then make a branch to stash things if I don't want to push it right away.
21:29 hercynium joined #parrot
21:29 ash_ joined #parrot
21:34 jsut_ joined #parrot
21:35 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#31840), fulltest) at r43537 - Ubuntu 9.10 i386 (g++ with --optimize)
21:43 kgilmer joined #parrot
21:44 kgilmer hi!  can anyone point me to how i can cross compile parrot 2.0 for arm?
21:44 Coke kgilmer: we don't support cross compilation out of the box yet, I don't think.
21:44 Coke (but I know someone has cross compiled parrot before by hand.)
21:45 Coke dukeleto?
21:45 kgilmer hi Coke, ok... is it possible to pass custom toolchain to configure.pl ?
21:46 japhb joined #parrot
21:46 darbelo Sort of, you could try to hijack the hints file.
21:47 darbelo But you have to keep in mind that parrot is used pretty extensively by it's own build process.
21:47 kgilmer hmm, ok thanks darbelo.  sounds like java build :0
21:48 kgilmer you need java to build java...
21:48 kgilmer makes things hard.
21:48 kgilmer sorry if this is a n00b question.  is cross compiling on the roadmap for parrot?
21:48 bacek joined #parrot
21:50 darbelo I'm not sure. Right now our roadmap is more focused on vm features.
21:51 kgilmer ok darbelo, thx.  so if I want parrot on ARM now I'll need to build natively on an arm machine.
21:51 darbelo Pretty much. Yeah.
21:51 kgilmer w/ gcc 4.4.1
21:51 kgilmer ok
21:51 kgilmer thx for your help darbelo and Coke !
21:51 kgilmer Parrot has come a long way, good work!
21:52 darbelo Well, we *do* support older gcc.
21:52 kgilmer ah ic.  the PLATFORMS file specifies that version I think but good to know it's not strict.
21:52 darbelo I use 3.3.5 but that's on OpenBSD i386.
21:53 darbelo PLATFORMS is ussually restricted to what testers have access.
21:54 kgilmer ic makes sense darbelo
21:54 darbelo If it's there it works. If it isn't, it might work too, but nobody told us.
21:54 kgilmer i have been wanting to integrate parrot into openembedded for awhile now, but will need to be able to pass toochain to the build before that will work.
21:55 kgilmer seems like a serious effort though.
21:55 kgilmer due to the self referential aspect of the build
21:56 darbelo There might be a way around that.
21:56 Coke try to build 'corevm' first.
21:57 Coke (that should avoid at least some of the stuff that requires parrot.)
21:57 kgilmer oe as concept of "native" package that can be used in staging to build a target package.
21:57 kgilmer so i imagine i'd have to build parrot-native, stage that, and then use that to build parrot for my target
21:57 kgilmer everything in oe is internally hosted, including the toolchain.
21:58 kgilmer ok Coke thanks for the suggestion.  corevm only requires gcc, perl, linker etc?
21:59 clinton joined #parrot
22:03 ash_ joined #parrot
22:05 mikehh joined #parrot
22:06 Coke think so.
22:07 cotto_working icu may be a hard requirement too
22:07 darbelo I haven't built without it in a while. But we supported it in the past.
22:08 chromatic It's not a hard req yet.
22:09 kgilmer well i updated my oe build script for parrot native...will see in a bit if it works.
22:10 kgilmer looks like oe has icu 3.6 native package.
22:11 kgilmer is 3.6 ok?
22:11 dukeleto Coke: yes?
22:11 dukeleto kgilmer: what are you trying to do?
22:12 kgilmer dukeleto, my aim is to write an openembedded 'recipe' such that i can run parrot on arm machines like the beagelboard and bug.
22:12 chromatic Alright, jsut_.  Let's hear your brilliant project management strategy.  This should be good.
22:13 * Coke thinks he's missing a send.
22:14 chromatic http://news.ycombinator.com/item?id=1070798
22:15 chromatic C'mon Ben.  We're waiting.
22:16 kgilmer this channel is fun.
22:16 darbelo fperrad: ping
22:16 ash__ joined #parrot
22:16 davidfetter what pie?
22:16 purl pie are square!
22:17 davidfetter chromatic, i'm unfamiliar with the pie incident. when was it?
22:17 chromatic Dan bet Guido that Parrot would run Python faster than CPython by OSCON 2004, I think.
22:17 ash__ joined #parrot
22:17 davidfetter um, did he win?
22:17 chromatic Nope.
22:18 chromatic http://www.mail-archive.com/pe​rl6-all@perl.org/msg40418.html
22:18 darbelo davidfetter: google pie-thon
22:18 davidfetter k
22:18 dukeleto kgilmer: are you doing real-timey stuff?
22:19 dukeleto kgilmer: i am interested in that stuff. do you have a repo somewhere?
22:19 dukeleto kgilmer: the RTEMS guys have cross-compiled Parrot, but they have hand-crafted magic
22:19 dukeleto rtems?
22:19 dukeleto RTEMS?
22:19 dukeleto darn
22:20 kgilmer dukeleto, no real time stuff.  i am working locally against oe-dev git repo.
22:20 kgilmer if i make some non-trivial progress i'd be happy to share what i've got
22:20 chromatic No jsut_.  Pity.  I was looking forward to his brilliance descending down from Mt. Sinai to correct all of our flaws in 2004 retroactively.  Guess his time machine's not working.
22:20 dukeleto purl, rtems is http://rtems.org/ or Real Time Embedded OS
22:20 purl OK, dukeleto.
22:21 dukeleto kgilmer: sounds great.
22:22 darbelo msg fperrad lua's t/standalone.t fails (for me) if there's no 'native' lua binary. Is that expected or should the test be skipped if no lua is available?
22:22 purl Message for fperrad stored.
22:23 ash_ kgilmer: i'd be interested in your successes, i was going to get a gumstix to play with some to see if i could use rakudo perl 6 (which is running on parrot) for scripting on it, just to see the performance
22:24 chromatic ash_, particle had talked to some people on a project like gumstix, but I haven't heard anything about it since last summer.
22:24 dukeleto ash_: gumstix?
22:24 darbelo http://www.gumstix.com/
22:24 ash_ dukeleto: its a processor on a chip, embedded system, uses an arm processor
22:25 * dukeleto likes
22:25 darbelo ... and is the size of a gumstick.
22:25 * dukeleto no likes flash on front page, tho
22:25 kgilmer sounds interesting ash_ :)  well if parrot is in OE then it should be pretty easy to build for gumstix as well.
22:26 dukeleto kgilmer: i am not very familiar with OE
22:26 ash_ i also am sorta trying to get parrot to compile as a static library for the iphone, its just xcode's being annoying when i do the make process so i am kinda hold off for now on that, but i thought it might be cool to write an iphone app in rakudo using NCI to call the interface stuff
22:27 kgilmer well i am not an expert, but i have written some stuff, mainly java recipes.
22:27 kgilmer i work for bug labs, and i would like to get parrot on bug so we can have all these cool languages to play with :)
22:28 chromatic kgilmer, were you the one who talked to particle about this?
22:28 kgilmer it's probable chromatic
22:28 kgilmer a year or so back
22:28 chromatic That sounds right.  Before YAPC?  At YAPC?
22:28 PerlJam ash_: remind me on Monday.  I may have a gumstix around here I could send you.  Though it would be about 4-5 years old if so
22:28 kgilmer i was looking to talk to someone at SCaLE last year but no luck
22:29 dukeleto kgilmer: i have heard of bug labs. cool stuff!
22:29 kgilmer thx dukeleto
22:29 chromatic I'm definitely interested in seeing a Parrot port there too.
22:29 kgilmer well now with native arm working it should be easier
22:30 dukeleto kgilmer: let us know how we can help
22:30 ash_ PerlJam: wow, that would be cool. I am still in college and i am working on some student based robotics projects and we have gumstix in our lab, just being used on a robot right now, so i'd rather not mess with theirs too much :P
22:31 kgilmer ok will do dukeleto .  looks like i'll resume on monday.
22:31 PerlJam ash_: we got ours to develop a cheap data collection platform for some instrumentation, but no one has done anything with it in a few years.  They're just sitting on a shelf somewhere (I'd need to find them even)
22:36 ash_ thats cool, PerlJam, i am working on ION Robotics lawnmower competition, but that robot's running linux on an embedded atom processor
22:43 plobsing hi #parrot
22:43 plobsing anyone interested in testing libjit_framebuilder2?
22:43 cotto_working hi plobsing
22:44 ash_ sure, i'd try it
22:44 cotto_working plobsing, I'll give it a shot
22:45 cotto_working what do you want tested?
22:45 plobsing first test without libjit installed. make sure nothing breaks.
22:45 plobsing after, if you're feeling adventurous, install libjit and reconfigure (it should detect automatically)
22:46 ash_ which version of libjit is needed?
22:46 ash_ i am on OS X
22:47 plobsing I'm running 0.1.2 on x64, but IIRC I had it working on 0.1.0 my x86 vm. so anything "recent"
22:47 ash_ macports only has 0.0.6 so, i guess i'll have to manually install it
22:48 plobsing ash_: you could try with the old version. I'm not sure how stable/unstable their api/feature-set is.
22:49 cotto_working Configure detected libjit fine.  testing nao
22:49 dukeleto languages?
22:49 purl languages are tools.  just like php programmers.
22:49 dukeleto dammit
22:49 ash_ lol purl
22:49 dukeleto purl, why do you forget usefult things and remember nonsense?
22:49 purl no idea, dukeleto
22:49 dukeleto botsmack
22:50 ash_ plobsing: is 0.1.2 the most recent version?
22:51 chromatic I used 0.1.2 when I tried it a couple of months ago.  It seemed to work.
22:52 ash_ checking out the branch, i'll let you know when its done
22:53 plobsing ash_: as far as I know, thats the latest main branch version. libjit-linear-scan has 0.1.2.5, but I'm not sure if that's a fork or the new name for the project or what
22:53 cotto_working make test with libjit looks good
22:53 cotto_working testing without...
22:53 ash_ i found the tar for 0.1.2 on their website
22:53 * dukeleto begs for free embedded hardware to port Parrot to. cross your appendages
22:54 cotto_working lots of failures, but it smells like makefile stupidity.  retesting with reconfigure
22:54 ash_ i wonder if libjit builds on os x still.... lots of things dont like 10.6
22:55 cotto_working I can see why chromatic wants one_make.  The parallel build is broken-flavored atm.
22:55 ash_ libjit didn't build
22:55 plobsing ash_: nopaste of error?
22:56 chromatic Who are you begging, dukeleto?
22:56 ash_ http://nopaste.snit.ch/19331
22:56 ash_ plobsing: there's my make
22:56 ash_ configure did fine
22:56 ash_ gcc-4.2 (apple's)
22:57 cotto_working plobsing, make test w/o libjit looks good too
22:57 ash_ are you doing just make test? or is there a more comprehensive test suite?
22:57 plobsing cotto_working: what platform/version?
22:58 cotto_working Ubuntu Jaunty x64
22:59 plobsing ash_: make test is probably good because it runs t/pmc/nci.t which is the main nci test, but if you want to be comprehensive, make fulltest is always best
22:59 dduncan joined #parrot
22:59 cotto_working libjit 0.1.2
22:59 ash_ http://git.savannah.gnu.org/cgi​t/dotgnu-pnet/libjit.git/refs/ which is supposidly the repo for libjit only has tags for 0.0.6 0.0.4 and 0.0.2 :-/
23:01 clinton joined #parrot
23:03 plobsing ash_: http://download.savannah.gn​u.org/releases/dotgnu-pnet/ has 0.1.0 and 0.1.2
23:03 ash_ yeah, i found the 0.1.2, i'll try 0.1.0
23:04 ash_ all tests passed without libjit on os x
23:04 particle joined #parrot
23:04 darbelo plobsing: all test pass on OpenBSD without libjit.
23:05 ash_ libjit 0.1.0 fails to build too
23:06 ash_ make immediately dies with an error saying gen-apply-macosx.h is not a file
23:06 ash_ sorry, it seems libjit isn't working on os x, i guess i can email someone at libjit if they don't already know
23:07 ash_ do they have a bug tracker that you know of?
23:07 Infinoid joined #parrot
23:09 plobsing ash_: I'm not sure which of the sites (and associated bug-trackers) is the current one, but they do have a mailing list: dotgnu-libjit on lists.gnu.org
23:10 ash_ k, i'll email there in a bit
23:16 plobsing OK, here's where I think libjit framebuilder is at:
23:16 plobsing * doesn't break anything where libjit isn't installed
23:16 plobsing * works on linux x86/x64 where libjit is installed
23:17 plobsing do I need to test anything else?
23:18 chromatic Does it work on Mac OS X when installed?
23:19 chromatic ... I mean, if libjit worked on Mac OS X.
23:21 plobsing I don't see a reason why it wouldn't, but I haven't seen anyone able to get libjit working on mac
23:22 chromatic As long as the build still works on Mac OS X without libjit installed, I see no problem.
23:23 darbelo No breakage means no blockers. Merge it!
23:23 plobsing ash_: can you verify that libjit_framebuilder2 doesn't break the build on OSX when libjit isn't installed?
23:31 dduncan left #parrot
23:31 Whiteknight joined #parrot
23:34 riffraff joined #parrot
23:43 kid51 joined #parrot
23:46 * kid51 reads backscroll
23:47 kid51 plobsing Could you use a smoke report for libjit_framebuilder2 on Darwin/PPC?
23:47 kid51 darbelo ping
23:56 plobsing kid51: yes please
23:56 kid51 It will take 30 min or so.
23:58 darbelo joined #parrot
23:59 darbelo kid51: pong

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

Parrot | source cross referenced