Camelia, the Perl 6 bug

IRC log for #parrot, 2008-08-01

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:05 teknomunk__ joined #parrot
00:09 AndyA joined #parrot
00:32 wknight8111 2**5
00:32 purl 32
00:32 wknight8111 1**6
00:32 purl 1
00:32 wknight8111 0**7
00:32 purl 0
00:32 wknight8111 0**(-1)
00:32 wknight8111 0**-1
00:32 wknight8111 0**1
00:32 purl 0
00:32 wknight8111 5**-1
00:32 purl 0.2
00:32 wknight8111 purl doesn't like being segfaulted
00:32 purl wknight8111: sorry...
01:13 Infinoid -1**256
01:13 purl -1
02:03 bacek joined #parrot
02:08 tetragon joined #parrot
02:08 tetragon So, I've had a bit of a platform change happen over the past month
02:09 tetragon I'm now on Intel OS X 10.5 (64-bit)
02:11 Lorn joined #parrot
02:23 Infinoid nice!
02:25 Theory joined #parrot
02:26 tetragon I'm still waiting for some of my backups to transfer over, and I have quite a bit to recompile
02:27 jdv79 joined #parrot
02:27 tetragon The Apple techs who worked on my iBook were unable to put my box back together again without snapping something in the process
02:27 jdv79 kid51: around?
02:27 tetragon And this happened multiple times in the month
02:29 kid51 jdv79 wazzup?
02:30 jdv79 kid51: what ever happened to that school idea?  OT.  you mentioned it a year ago at a semny thing...
02:31 kid51 school idea?
02:31 Infinoid tetragon: so they gave you an extra 32 bits instead? :)
02:33 tetragon Infinoid: Yep, although I went and paid a bit to go to a MacBook Pro instead of a MacBook.  I can hardly stop playing with the auto-dimming screen and the illuminated keyboard
02:34 Infinoid hehe
02:38 * kid51 must restart computer for software update; bye
02:42 Infinoid I must say, I'm quite impressed with the "mylyn" extension to eclipse
02:42 Infinoid one of the biggest problems I've had so far with PDD13 is the fact that there are a *lot* of PMCs in src/pmc/, and a *lot* of tests in t/pmc/
02:43 Infinoid I end up spending a lot of time paging through the list to find the file I want
02:43 Infinoid mylyn hides the stuff I'm not interested in, quite nicely.
03:22 coumbes joined #parrot
03:23 Andy joined #parrot
03:50 teknomunk__ joined #parrot
04:27 rurban_ joined #parrot
04:28 Infinoid jonathan: PDD13's definition of the PackfileDirectory PMC has methods to list and fetch segments, and a method that inserts/updates segments, but no way to remove one.  is that intentional?
04:45 seneca joined #parrot
05:15 Theory joined #parrot
05:17 dalek r29918 | infinoid++ | pdd13pbc:
05:17 dalek : [PDD13]
05:17 dalek : * Add a get_pmc_keyed() function to PackfileDirectory, which dispatches to
05:17 dalek :   .get_pmc_keyed_str().
05:17 dalek : * Fix some erroneous test descriptions.
05:17 dalek : * Add a couple of TODO tests for the things in PackfileDirectory I haven't
05:17 dalek :   implemented yet.
05:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29918
05:42 iblechbot joined #parrot
05:43 seneca joined #parrot
05:45 Psyche^ joined #parrot
05:54 Ademan joined #parrot
06:29 uniejo joined #parrot
06:52 silug joined #parrot
06:58 jan joined #parrot
07:03 Ademan joined #parrot
07:40 cosimo joined #parrot
07:50 bacek joined #parrot
07:54 * moritz wonders why autounfudge is b0rked
07:56 nopaste joined #parrot
08:02 alvar joined #parrot
08:09 masak joined #parrot
08:31 TonyC joined #parrot
09:37 barney joined #parrot
09:45 dalek r29919 | kjs++ | trunk:
09:45 dalek : [pirc/new] small refactorings and add a few comments.
09:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29919
09:50 kj joined #parrot
11:00 ambs joined #parrot
11:16 bacek joined #parrot
11:25 moritz any rakudo hacker or tester around?
11:25 moritz (except me, that is ;)
11:25 masak I don't have commit access, but I consider myself a tester. do I count?
11:25 moritz yes ;)
11:25 masak woot
11:26 masak anyway, I'm around :)
11:26 moritz for me 'make spectest_regression' works nearly fine (just 3 failed tests)
11:26 masak checking...
11:26 moritz whereas both tools/autounfudge.pl and tools/test_summary.pl have problems
11:26 moritz for example test_summary:
11:26 moritz t/spec/S02-builtin_data_types/num.rakudo...............  47Statement not terminated properly at line 5, near ".8;\n\n#if s"
11:26 moritz masak: is that the same on your box?
11:27 masak hold on
11:28 * moritz holds on
11:28 * masak makes
11:30 * moritz finally releases his breath
11:30 masak good. makes you live longer.
11:32 masak my spectest_regression has now passed S02-builtin_data_types/
11:32 moritz yes, it was an EOL problem
11:32 masak EDL?
11:32 masak btw, what is num.rakudo?
11:32 moritz jonathan++ for finding that
11:32 masak ah, EOL
11:33 moritz the fudged version of num.t
11:33 masak I don't get that output, I just get num
11:33 moritz did you try tools/test_summary.pl?
11:34 masak uh... no
11:34 masak still spectest_regressing
11:34 moritz that's where my error message came from
11:34 masak I see.
11:34 masak will try
11:34 moritz spectest_regression has only a failure in S03-operators/arith
11:35 masak I was just going to say that :)
11:35 moritz which I guess was over-agressive unfudging or something ;)
11:35 masak 3 failing tests
11:35 moritz aye
11:35 moritz Failed tests:  130-132
11:39 kid51 joined #parrot
11:46 masak moritz: maybe it's no longer an issue, but I didn't get a 'statement not terminated properly' error running num.rakudo in tools/test_summary.pl
11:47 moritz masak: so did it run (mostly) fine?
11:49 masak aye
11:49 moritz which means that I probably have a local fuckup. Bad.
11:49 masak (local fuckups)++
11:50 masak eh, (local fuckups)--
11:50 moritz (local fuckups)-- # at least in terms of karma
11:50 masak I've been ++ing to much lately
11:50 masak it's a reflex
11:51 moritz masak++ # karma reflexes ;)
11:51 masak :)
11:51 moritz karma masak
11:51 purl masak has karma of 14
11:51 masak it's a start. :)
11:51 masak karma moritz
11:51 purl moritz has karma of 119
11:51 masak whoa.
11:52 moritz half of it is for commiting other people's patches ;)
11:52 moritz the other half is for adding files to spectest_regression that Auzon++ fudged
11:52 masak ...both halves deservedly, I'd add.
11:52 moritz and the third half is actually earned ;)
11:52 masak hehe.
11:52 masak it's all earned
11:52 masak karma never lies
11:52 purl never lies has neutral karma
11:52 moritz lol
11:52 masak purl: you are very stupid
11:52 purl ...but purl is a tool or a good source of vitamin ass or marnanel's mom or the state of western Transylvania on a thursday or the death of me....
11:53 masak it won't be long until racism against IRC bots becomes rampant. and no-one will speak in their favor.
11:54 moritz except themselves, of course
11:54 masak they could flood the channels, nobody in power will care
11:55 masak actually, it'd probably be a good deed to put together by-laws for bot writers. I opine that purl breaks a few such unwritten laws by being too aggressively helpful at times.
11:58 masak "a bot should not speak unless spoken to"
12:00 moritz which outlaws our nice svn bots
12:00 masak uh, yes...
12:00 masak that's not what I meant :)
12:01 moritz I know ;)
12:01 masak good point, though
12:01 masak there's discussion-speech, such as that of purl, and info-speech such as that of our nice svn bots
12:01 moritz I quite like it to show people when what they say differs from what they mean ;)
12:02 moritz (but people don't always love me for that ,)
12:02 masak moritz: in that respect, you have much in common with a compiler :)
12:02 moritz perhaps that's why they fascinate me ;)
12:03 masak me too :)
12:04 masak esp. GHC, which actually has very informative and readable errors
12:07 * moritz loves spin (a promela interpreter/compiler/verification tool) which actually gives you counter examples that violate your assertions
12:08 moritz otoh, promela (being just a regular language, not turing complete) sucks
12:17 ruoso joined #parrot
12:28 rurban_ joined #parrot
12:39 Infinoid msg jonathan PDD13's definition of the PackfileDirectory PMC has methods to list and fetch segments, and a method that inserts/updates segments, but no way to remove one.  is that intentional?
12:39 purl Message for jonathan stored.
13:01 seneca joined #parrot
13:29 jhorwitz joined #parrot
13:48 gmansi joined #parrot
13:57 gryphon__ joined #parrot
14:05 Andy joined #parrot
14:06 barney joined #parrot
14:12 Lorn joined #parrot
14:37 jdv79_ joined #parrot
14:57 ambs seen coke
14:57 purl coke was last seen on #parrot 19 days, 9 hours, 53 minutes and 11 seconds ago, saying: yes?  [Jul 13 05:04:35 2008]
14:57 ambs seen dietcoke
14:57 purl dietcoke was last seen on #parrot 1 days, 1 hours, 5 minutes and 28 seconds ago, saying: later. =-)  [Jul 31 13:52:24 2008]
15:00 DietCoke joined #parrot
15:00 DietCoke ambs: yes?
15:01 ambs DietCoke, I was just preparing to call you 'Spammer' :)
15:01 ambs but now, I feel I should call you, something like 'omnipresent'
15:01 ambs :D
15:04 DietCoke I figured you'd appreciate having all the coke-interactions be front-loaded this quarter.
15:13 davidfetter joined #parrot
15:16 rhr joined #parrot
15:32 ambs_ joined #parrot
15:34 Whiteknight joined #parrot
15:47 DietCoke left #parrot
15:49 tewk x86_64 jit status?
16:04 alvar joined #parrot
16:05 Theory joined #parrot
16:17 iblechbot joined #parrot
16:52 jan joined #parrot
17:06 Ivatar joined #parrot
17:14 dalek r29920 | chromatic++ | trunk:
17:14 dalek : [t] Ensured cleanup of temporary directories created during testing.
17:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29920
17:14 tewk joined #parrot
17:16 dalek r29921 | chromatic++ | trunk:
17:16 dalek : [t] Ensured that temporary files created during testing get cleaned up.
17:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29921
18:21 peepsalot joined #parrot
18:37 gryphon__ joined #parrot
19:29 Senak1 joined #parrot
19:35 apeiron joined #parrot
19:36 Senak1 left #parrot
19:52 jhorwitz joined #parrot
20:05 jhorwitz hm.  if i have Foo::bar, is there any way in rakudo to tell if it's a method or a sub?
20:10 Whiteknight try { $foo() } catch { say "not a sub!"; }
20:10 Whiteknight :)
20:11 Whiteknight I don't even remember P6's exception handling syntax
20:11 jhorwitz i obviously didn't specify the rules of the game.  :)
20:11 jhorwitz no calling the sub.  but i think i have a workaround anyway.
20:12 Tene sub foo { die "fail in a sub" }
20:13 jhorwitz fyi, i'm trying to implement method handlers in mod_perl6.  i have to know if i'm calling a method or regular sub at runtime.
20:14 jhorwitz my workaround is to call .WHAT() on the handler name, which will reveal if it's a class.  unfortunately rakudo bombs out when the handler name is a regular namespace.  i'll open a ticket.
20:17 Tene Yeah, Parrot's core classes don't implement WHAT, HOW, etc.
20:17 Tene Rather a pain at times.
20:17 jhorwitz indeed
20:17 jhorwitz but if you do, RANDOMSTRING.WHAT(), it runs fine.
20:17 jhorwitz as long as it's not a namespace.  :-P
20:18 Tene Or a ResizablePMCArray.
20:18 purl a resizablepmcarray is probably what we use nowadays
20:18 particle um... so check for a namespace first?
20:18 jhorwitz can i do that in rakudo?
20:18 Tene purl: forget resizablepmcarray
20:18 purl Tene: I forgot resizablepmcarray
20:19 jonathan (half paying attention) .WHAT will end up being a compiler macro rather than a real method anyway, so we'll be able to make it do the Right Thing on non-Perl 6 things later.
20:19 Tene jhorwitz: you could .HLL_map Namespace.
20:19 jhorwitz Tene: can't use PIR here.
20:19 Tene jhorwitz: do it in rakudo first.
20:20 jhorwitz jonathan: i trust that it will all work right later.  :)
20:20 particle should rakudo see parrot namespaces as packages?
20:20 particle * perl 6 packages
20:20 jhorwitz ignore parrot namespaces.  run this....
20:20 jhorwitz nopaste?
20:20 purl rumour has it nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl
20:20 jonathan Probably...
20:21 jonathan Wowz, I haz two messages
20:21 nopaste "jhorwitz" at 96.245.16.64 pasted "namespace & WHAT" (3 lines) at http://nopaste.snit.ch/13718
20:22 jonathan jhorwitz: I'm not sure we've really paid a load of attention to namespaces yet.
20:22 jonathan erm
20:22 jonathan to modules
20:22 jhorwitz i just traced it -- akin to what Tene said, parrot tries to call typeof on a namespace, which returns PMCNULL
20:22 jonathan They're in the ROADMAP.
20:23 particle Foo is a module
20:23 particle it perl6land
20:23 jonathan Ano.
20:23 jonathan *yes!
20:23 particle it should implement .WHAT
20:23 * jonathan shold speak the correct language
20:23 jonathan particle: I agree it *should*, I'm just saying I don't think anyone has worked on modules to make sure they do work yet.
20:23 particle right
20:24 particle does grammar implement .WHAT?
20:24 jonathan particle: No - we parse it as a normal method call and have it just as a normal method at the moment.
20:24 jonathan But S12 says it should be a macro.
20:24 jonathan As for .HOW and so forth.
20:24 particle i see
20:25 jonathan So we're kinda cheating at the moment.
20:25 particle src/classes/Code.pir has a WHAT method
20:25 jonathan jhorwitz: Let me read backscroll and see if I can work out what you're needoing to do.
20:26 particle jonathan: he needs module Foo; print Foo.WHAT(); to work
20:26 jhorwitz jonathan: particle hit the nail on the head.  after tracing, i see the root cause is that typeof is returning PMCNULL for a namespace.
20:26 jonathan OK.
20:27 jhorwitz and getprop on the PMCNULL, well, you know what happens.
20:27 particle an earth-shattering kaboom.
20:27 jhorwitz big badaboom
20:27 jonathan OK
20:27 jonathan un momentito...
20:28 jhorwitz so many languages....
20:28 jonathan I'll try and fix that for you now.
20:28 jonathan jhorwitz: Well, I live in Slovakia now, and lived in Spain for 6 months.
20:28 rurban_ joined #parrot
20:28 jonathan So I speak ok Spanish (enough to get by and do the day to day trips to the bank, buying train tickets, food, beer, etc) and basic Slovak.
20:29 jhorwitz Na zdravie!
20:30 jonathan :-)
20:30 jhorwitz speaking of that, i have a happy hour in 30...
20:30 jonathan Wow. It's been a quiet day for Parrot commits today...
20:32 jhorwitz LOL when i put it in a try block, it bombs out with "Lexical '$!' not found"
20:33 * jhorwitz can't win today
20:33 Tene nice
20:33 jhorwitz Tene: OBTW, mod_lolcode was a hit at OSCON, at least in the lightning talks
20:34 Tene Yeah?  Wish I could have been there.
20:34 Tene What sort of response?
20:34 particle clapping and laughter
20:35 particle and a few people threw kittens
20:35 Tene nice
20:38 jhorwitz better than the response at the talk.  *crickets*
20:40 jonathan Argh. Even with a patch that is meant to do the Right Thing, it's because get_class in NameSpace is overridden.
20:40 jhorwitz jonathan: headed out for some beers.  msg me if you get that working and i'll owe YOU a few.  :)
20:40 teknomunk joined #parrot
20:41 jonathan HA! Win.
20:41 jonathan module Foo;
20:41 jonathan say Foo.WHAT
20:41 jonathan Module
20:42 jhorwitz jonathan++
20:42 jhorwitz jonathan++
20:42 jhorwitz jonathan++
20:43 jonathan jhorwitz: Running tests, then will ci.
20:43 jonathan Hope this helps. :-)
20:43 jhorwitz thanks!
20:44 jonathan jhorwitz: Does mod_parrot run on Win32 yet?
20:44 jhorwitz particle needs to finish up the configure branch, and i need to test on a threaded MPM on unix first.
20:44 jonathan OK.
20:45 jhorwitz then we close our eyes, hold our breath, and try win32.
20:45 jonathan I don't even have an Apache install here, but it would be fun to play with at some point.
20:46 jonathan Anyways, enjoy the beers!
20:46 * jonathan is able to drink beer again today, after a couple of weeks without thanks to medication.
20:47 jhorwitz :)
20:48 ambs :)
20:57 Whiteknight joined #parrot
21:07 dalek r29922 | jonathan++ | trunk:
21:07 dalek : [rakudo] Make .WHAT called on a NameSpace return a Module proto-object. Added to unblock mod_perl; probably will want a re-visit when we look at modules more comprehensively.
21:07 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=29922
21:08 jonathan purl message Infinoid Do you have a use case for removing a segment?
21:08 purl Message for infinoid stored.
21:09 Infinoid jonathan: I know nothing about .pbc file usage, but I figure if you can replace it, why not.  it seems like a hole in an otherwise complete CRUD style interface
21:10 jonathan Infinoid: I do know a bit about PBC file usage, and couldn't think of a use case. OTOH, it's probably not hard to implement, and someone may think of a usage.
21:11 jonathan So, for completeness sake, let's have it.
21:11 Infinoid heh, here's a use case.  binutils has a "strip" command for removing debugging sections... seems like that would be useful here too
21:11 jonathan Aha.
21:11 jonathan Yes, that would certainly be a use case.
21:11 jonathan OK, it's in.
21:11 jonathan If you would, please update the PDD too.
21:11 Infinoid you want to modify the pdd, or should I?  (I'd have to look up the right vtable name and all of that first)
21:12 Infinoid ok, will do.
21:12 Infinoid thanks!
21:13 jonathan If you're not sure how to update it, just mail me...I'll also review the commit when it gets sent to p2 anyway.
21:17 * jonathan afk for the evening...relaxing
21:18 Infinoid happy weekend, jonathan :)
21:19 jonathan Thanks, you too. :)
21:26 davidfetter joined #parrot
22:13 kid51 joined #parrot
22:20 teknomunk_ joined #parrot
22:24 gmansi joined #parrot
22:26 teknomunk__ joined #parrot
22:47 justin joined #parrot
22:48 Andy joined #parrot
23:09 dngor joined #parrot
23:09 workbench joined #parrot
23:11 Casan joined #parrot
23:11 davidfetter joined #parrot
23:11 Casan left #parrot
23:11 Casan joined #parrot
23:12 Andy joined #parrot
23:27 seneca joined #parrot
23:30 Limbic_Region joined #parrot
23:44 cotto_work is there any reason why GET_ATTR_globals is rewritten to GETATTR_Exporter_globals instead of GET_ATTR_Exporter_globals?
23:46 Ademan joined #parrot
23:46 cotto_work and if there isn't, could my alter ego submit a patch preserving the underscore now, before most of the PMCs have been converted?
23:47 Infinoid svn blame?
23:47 purl i heard svn blame was just like p4 annotate, only better
23:47 * Infinoid has no idea, sorry
23:48 cotto_work you're in good company, then ;)
23:48 cotto_work I'll probably just submit a patch and see what happens.
23:48 Infinoid cool.
23:49 * cotto_work looks forward to being able to submit a patch from work
23:55 cotto_work Infinoid++ #I hadn't seen svn blame before.  Nice tool, that.
23:55 Infinoid oh, cool

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

Parrot | source cross referenced