www.parrot.org/ | 291 RTs left | Next release: 2009-07-21
Set by moderator on 14 July 2009.
bacek_at_work Whiteknight: I'm waiting for TT#759 resolution. 00:11
If decision willnot be made by Saturday I'll add few useless test files. 00:12
00:12 kid51 joined
kid51 seen Coke 00:12
purl Coke was last seen on #parrot 4 hours, 25 minutes and 42 seconds ago, saying: allison: thanks. that'll help me get an osx distro of padre cut. =-)
00:13 mikehh joined
dalek cnum-dynpmcs: r109 | darbelo++ | trunk/aux/decTest/ (3 files):
Add the start of this tool's future build infrastructure.
 00:25
00:30 TiMBuS joined 00:36 mokurai joined 00:43 mikehh_ joined 01:32 mokurai joined
dalek cnum-dynpmcs: r110 | darbelo++ | trunk/aux/decTest/cfg/Makefile.in:
Some makefile touchups for aux/decTest.
 01:50
01:58 darbelo left 02:09 athomason joined 02:20 sekimura joined 02:22 Zak joined 02:33 allison joined 02:34 kid51 joined 02:36 janus joined 03:10 mokurai joined 03:13 Andy joined 03:21 donaldh joined 03:32 Andy joined 03:33 Theory joined
Andy am I here? 03:36
03:37 petdance joined 03:47 particle1 joined 03:51 GeJ joined 04:21 iblechbot joined 04:22 skids joined 04:37 MikHel joined
MikHel Hello! 04:37
MikHel is a Parrot newbie having troubles....
How do I debug a grammar and its actions? 04:38
I cannot get a working AST... The error messages lack context :(
"Method 'ast' not found for invocant of class 'Undef'" is definitely not telling me anything I can work with :( 04:39
Anybody awake? :) 04:45
Util Awake, but working in Parrot down below the PCT/AST level. 04:50
MikHel: Can you nopaste your failing code? 04:51
as in: nopaste.snit.ch/ 04:54
MikHel Util. Sure. Bit long. Two files. 04:56
nopaste "MikHell" at 67.165.213.109 pasted "Failing code. First the grammar.pg" (80 lines) at nopaste.snit.ch/17285 04:57
"MikHell" at 67.165.213.109 pasted "Failing code. Second the actions.pm" (144 lines) at nopaste.snit.ch/17286 04:58
MikHel Util. There you go. 04:59
The input I tried is "a=3;"
bacek_at_work MikHel: $() form is deprecated. Use .ast call directly. 05:00
MikHel: rule assignment is culpit.
You are using $(<assignment>) 05:01
sorry. <expression>
It should be <simple_expression>
MikHel bacek_at_work: Hmm lemme look to see if I can understand what you are saying :) 05:04
bacek_at_work: Hmmm OK I see what you are saying. It is indeed a mistake. Corrected it, but did not help. Lemme check again. 05:06
bacek_at_work MikHel: rule assignment { <identifier> '=' <simple_expression> ';' }
dalek rdinal: 38f6a07 | (Ted Reed)++ | (2 files):
Add new build system.
bacek_at_work so in Actions.pm you should use <identifier> and <simple_expression>
treed I'd be interested in hearing if that build system works for people, BTW. 05:07
MikHel bacek_at_work: Got it corrected. Thanks :) 05:09
Util bacek++ # for PCT knowledge
MikHel bacek_at_work: Is there some way to run the thing in a way that would show me where it fails when it does?
bacek_at_work: I also must admit I am a total novice, so most of what I do so far is cut and paste from a tutorial... not understanding what you said about the $() form at all :( ... What should I be doing? 05:11
treed Yay, no more changes to stage. I can get back to implementing new shit. 05:31
dalek rdinal: 224a460 | (Ted Reed)++ | src/classes/Array.pir:
Remove debug says from Array's * operator. :-/
 05:34
rdinal: 24e44a2 | (Ted Reed)++ | .gitignore:
Add .DS_Store and *.swp to .gitignore, and fix test.pir to Test.pir
rdinal: 5a8ce29 | (Ted Reed)++ | t/ (3 files):
Fix some test files to generate sensible TAP output, and to test what they're actually testing.
Util MikHel: If you are using the tutorial from parrotblog.org, I think you will find that this one is more up-to-date: 05:35
en.wikibooks.org/wiki/Parrot_Virtua...troduction
05:37 Zak joined
Util sleeps... 05:37
purl sleeps are being interr!!!
MikHel Util: Thanks. 05:45
cotto hi 05:51
purl what's up, cotto.
MikHel bacek_at_work: Where do I find info on the $() form and .ast ?
06:04 mokurai left 06:09 uniejo joined
MikHel How do I make my parser "ignore" new lines and only do something when it has a complete TOP object? 06:11
06:12 Zak joined 06:29 flh joined
flh MikHel, when you say "only do something when it has a complete TOP object", are you talking about interactive sessions (where you type some code, and return evaluates it)? 06:32
you might have a look at: git.berlios.de/cgi-bin/gitweb.cgi?p...mpiler.pir 06:38
I've subclassed HLLCompiler, replacing the "readline_interactive" method so that it doesn't only read one line at a time, but waits until it sees ";;" (it also cares about comments and strings) 06:40
MikHel flh: Yeah that's what I was wondering :) 06:41
flh: But though it will work, I do not think it is quite the perfect approach.... 06:42
But then, I am just starting :) Imperfect will do for now.
flh: To me it seems that the perfect way would be that the TOP expect only one statement, or function definition or class definition at a time. (one of the possible top object) and return when it has found it and it is executed. 06:44
Then the compiler outside would loop back to the TOP to find the next object... 06:45
06:51 barney joined
dalek rdinal: 7f387b6 | (Ted Reed)++ | Rakefile:
Fixes to the Rakefile. Make it generate Test.pir when needed, and fix some mistakes with the test harness.
 06:53
rdinal: d547aa0 | (Ted Reed)++ | (2 files):
Correct array.uniq+block's behavior to match the actual implementation in Ruby SVN. Also add a test for same.
rdinal: efff91b | (Ted Reed)++ | src/classes/Array.pir:
On second thought, have the value? check occur only if a block has been passed to Array.uniq.
 06:58
07:13 mikehh joined
dalek rdinal: c757e57 | (Ted Reed)++ | Rakefile:
Add overall statistics to the test:all rake target.
 07:15
07:20 donaldh joined 07:25 mikehh joined 07:54 AndyA joined
bacek_at_work waves 08:00
cotto: around?
MikHel: $( $<expression> ); is deprecated form of $<expression>.ast 08:02
08:06 chromatic joined 08:29 mikehh joined
mikehh codetest FAIL, ALL others PASS (pre/post config, smolder, fulltest) at r40111 - Ubuntu 9.04 amd64 08:44
nopaste "mikehh" at 90.209.69.171 pasted "codetest failures from make fulltest at r40111" (33 lines) at nopaste.snit.ch/17288
08:50 TiMBuS joined 09:43 clinton joined 09:46 bacek joined
bacek o hai 09:47
mikehh: trac.parrot.org/parrot/ticket/759 09:50
09:50 MoC joined
bacek I'm waiting for resolution. Or just dump few useless test files. 09:50
.oO( Or I can do it right now. )
 09:51
mikehh bacek: I seem to remember that - I think I reported the original codetest failure - IO-handles or something like that 09:58
bacek mikehh: indeed
mikehh As far as I know kid51 was looking at it 09:59
dalek rrot: r40112 | bacek++ | trunk (2 files):
[cage][t] Add tests for ArrayIterator.
 10:24
10:34 payload joined 10:41 payload1 joined
dalek rrot: r40113 | bacek++ | trunk/t/pmc/stringiterator.t:
[cage][t] Add tests for StringIterator
 10:45
10:56 bacek_ joined
dalek rrot: r40114 | bacek++ | trunk/t/pmc/orderedhashiterator.t:
[cage][t] Add tests for OrderedHashIterator
 10:58
rrot: r40115 | bacek++ | trunk/t/pmc (2 files):
[cage] Add pop_eh to *Iterator tests
rrot: r40116 | bacek++ | trunk/t/pmc/stringiterator.t:
[cage][t] One more pop_en in stringiteratort. bacek-- for commiting before saving file.
 11:15
rrot: r40117 | bacek++ | trunk/t/pmc/hashiteratorkey.t:
[cage][t] Add dummy test for HashIteratorKey
bacek_ mikehh: make codetests should be happy now :) 11:17
11:20 donaldh joined
spinclad (note to all: purl's 'probably' and 'probably. ' (note trailing space after period) are Dead.) 11:21
(10 days late, i know. i'm way backlogged.)
moritz anyway, thanks spinclad++ 11:22
anything that makes purl less annoying is greatly appreciated
bacek_ probably 11:23
bacek :)
moritz probably? really? are you certain? why not?
Coke some people like schizoid bots.
bacek should stop drinking... 11:24
moritz Coke: do you?
purl it has been said that do you is it compulsory? ;)
bacek I see "moritz" where it should be "purl"...
moritz bacek: I don't know if that's related to your changes, but I get again some packfile errors 11:36
bacek moritz: o shi...
moritz: What errors? 11:37
purl i think errors is buggy? funny
moritz smolder.plusthree.com/app/public_pr...ails/25022
t/pmc/packfile.t (Wstat: 256 Tests: 10 Failed: 0) Non-zero exit status: 1 Parse errors: Bad plan. You planned 34 tests but ran 10. 11:38
and so on, for 6 packfile tests or so
perl t/harness t/pmc/packfile.t
t/pmc/packfile.t .. 1/34 Unknown PMC type to thaw -100
current instr.: '_pbc' pc 35 (t/pmc/testlib/packfile_common.pir:26)
bacek moritz: hmm... Is is after make realclean/Configure.pl? 11:39
moritz bacek: aye 11:40
bacek moritz: ouch.... 11:42
purl hmmm... ouch is www2.epscylonb.com/image001.gif
bacek Time to bump PBC_REVISION and rebuild native PBCs...
rakudo: say "bacek--" 11:43
polyglotbot OUTPUT[bacek--ā¤] 11:44
bacek moritz++ 11:46
dalek rrot: r40118 | bacek++ | trunk (5 files):
Bump PBC_REVISION and rebuild native PBCs after merging
 11:49
bacek moritz: can you retest at r40118?
moritz bacek: smoke running already 11:50
bacek moritz: thanks
11:56 ruoso joined
bacek O! 11:57
robots: trust bacek_at_work 11:58
hmm.. Wrong spell
bacek cast memoize on himself 11:59
opbots, trust bacek_at_work 12:00
clunker9__ But I do not trust you bacek
slavorg Ok
bacek Aha!
opbots, trust cotto 12:02
slavorg But I already trust cotto
clunker9__ But I do not trust you bacek
12:02 bacek joined
bacek moritz: any luck with smoke? 12:08
12:10 mikehh joined
Coke opbots, trust bacek_at_work 12:17
slavorg But I already trust bacek_at_work
clunker9__ But I do not trust you Coke
Coke so, who owns the new packfile pmc tests?
They are failing on all three core platforms now. 12:18
dalek TT #840 created by coke++: t/op/io.t fails on win32 12:19
bacek Coke: I own them... 12:21
Coke then #823 is probably for you. 12:22
bacek Coke: can you retest them on latest parrot? 12:23
I've rebuild "native pbc" recently
(It should be "rebuilt" if I correctly understand this barbarian language :) 12:24
Coke bacek;that was as of 2 minutes ago.
they've been failing for weeks on linux.
bacek Oh...
Weeks???
purl Weeks is good enough to make money today
Coke you seem to have re-introduced the tests that were explicitly disabled by allison pre-1.0 here. (or at least in the same vein.) 12:25
the ones that are todo'd "pending a robust test strategy" - those also required the pbcs to be made up to date often.
bacek;yes. weeks. I mentioned it casually here, it never got fixed, I opened the ticket a week or so ago. 12:26
I rarely run 'make test' in parrot, concentrate more on partcl.
bacek oookey...
Coke that's with r40118 on all three failing platforms. 12:27
bacek spotted "constant" in backtrace. 12:28
Coke (the failures on feather predated the landing of the hash branch.)
I can try to figure out a bisect if you're unable to duplicate the failures.
bacek Coke: I've tested branch carefully... A lot of GC errors that I focused on was "inspired" by this branch. But looks like I've missed some of them. 12:30
Coke this is on trunk, and the failures at least predate the most recent mergeback from a branch. 12:31
bacek "feather" is...? 12:33
(In terms of platform)
Coke feather is the shared linux devel box. 12:34
juerd?
purl juerd is probably root or at juerd.nl/ or mailto:juerd@juerd.nl
Coke he can give you an account if you like.
bacek i386/x86_64?
Coke $ uname -a
Linux feather 2.6.18-6-xen-686 #1 SMP Sun Feb 10 22:43:13 UTC 2008 i686 GNU/Linux
bacek Bah!
Coke ? 12:35
bacek bacek@icering:~/src/parrot$ uname -a
Linux icering 2.6.26-1-686 #1 SMP Fri Mar 13 18:08:45 UTC 2009 i686 GNU/Linux
Coke what's the -xen -?
bacek It's my main "dev box"
Coke I meant on feather. does -xen just mean a virtual machine?
bacek "xen" is for vitualisation
nopaste "coke" at 72.228.52.192 pasted "t/pmc/packfilerawsegment.t" (52 lines) at nopaste.snit.ch/17289 12:37
Coke let me know if there's anything else I can provide. 12:39
bacek Coke: checking... 12:40
It;s not "branch failure" btw. keys_revamp branch just exposed few GC related bugs in parrot's internals... 12:42
Coke ... I cannot stress enough, this is happening IN TRUNK. 12:43
this has nothing to do with a branch.
and these failures were occuring IN TRUNK before you merged that branch back. 12:44
so the branch is completely blameless.
mikehh I got only a codetest failure at r40111 but lots of failures at r40117 12:47
nopaste "bacek" at 122.110.4.63 pasted "Patch for Coke" (13 lines) at nopaste.snit.ch/17290 12:48
bacek Coke: can you check build with nopaste? 12:49
Coke bacek: with that patch, I can't even build parrot. 12:50
mikehh IOW all tests but codetest PASSed at r40111 but smolder, most coretests FAIL at r40117 codetest PASS, example_tests FAIL
Coke (should I realclean and try again?)
bacek oh...
Why?
purl Left field.
moritz bacek: still fails
bacek Coke: no.
nopaste "coke" at 72.228.52.192 pasted "bacek" (25 lines) at nopaste.snit.ch/17291
bacek Coke: it's some weird bugs... 12:51
12:54 awhitworth joined
bacek ok. It's some very strange bug... I have to sleep on it (or Whiteknight can have a look) 12:55
Ho! 12:56
whiteknight ?
what can I look at?
bacek whiteknight: welcome! :)
whiteknight: nopaste.snit.ch/17291
Coke that's misleading. 12:58
that's with a patch.
TT #823 is the problem.
bacek (And a lot of other nopaste) Just read irclogs
Coke trac.parrot.org/parrot/ticket/823
bacek must sleep 12:59
purl $bacek->sleep(8 * 3600);
Coke ~~
bacek "Method 'wave' not found for class Undef" :) 13:00
Coke doesn't supposes there's anyone from singapore here. 13:01
Coke grumbles at various countries who do not post ALL their printed tax forms as PDFs. 13:03
mikehh t/codingstd/pbc_compat.t now fails 13:06
maifest_tests fail - # Failed test 'No need to regenerate MANIFEST' 13:07
s/maifest/manifest/ 13:09
13:11 ruoso joined 13:13 ruoso joined
mikehh at r40111 about 6 hours ago there was a codetest failure (not all PMC have tests) All other tests passed. 13:13
bacek put these tests together and since then we have had multiple failures
13:14 ruoso joined 13:16 ruoso joined 13:18 ruoso joined 13:20 ruoso joined 13:22 ruoso joined 13:23 slavorg joined
Infinoid `op slavorg 13:24
whiteknight we need some kind of dual-bot solution 13:25
Infinoid yeah. I ran one for a while, and am now set up to do so again
(remember slavorgn?)
whiteknight oh yeah 13:26
mj41 bacek: r40111 broke "packfile" ... tt.ro.vutbr.cz/report/pr-Parrot/do?...un-6395=on 13:30
13:32 slavorgn joined
Infinoid slavorgn, trust slavorg 13:32
slavorgn Ok
moritz slavorgn: trust me 13:34
slavorgn But I don't trust you there, moritz
Infinoid opbots, trust moritz
slavorg But I already trust moritz
clunker9__ But I do not trust you Infinoid
slavorgn Ok
Infinoid can I kick clunker9__? pretty please?
moritz clunker9__: trust Infinoid 13:35
opbuts, trust
opbuts, trust Infinoid
damn, ESC-. doesn't work in irssi :-)
Infinoid: if slavorgn has a reasonable trust list, why not? 13:36
Infinoid moritz: I don't have access to slavorg's list, but I can add everyone in here
moritz Infinoid: fine by me 13:37
13:37 MoC joined
Coke I hate irc. 13:40
Infinoid IRC hates us, too. 13:41
moritz Coke: the protocol? or the people? or the habits? or your client? or everything?
Coke the technology. we shouldn't have to act like a bunch of squatters. 13:43
PerlJam trundles off to find some bunnies and sunshine for Coke 13:44
Infinoid you think IRC is bad? look at real estate law sometime. 13:45
Coke Infinoid: I deal at least tangentially with non-US tax law. I'm set. =-) 13:46
13:47 whoppix joined
PerlJam real estate law is cool. One of my great aunts paid the property taxes on her mother's property after she died. then when her father died all of the siblings started squabbling over how to divvy up their property. Turns out that there was a law in the state at that time that read if you pay the property taxes on a property for more than 7 years, you are the de facto owner of the property. 13:47
My great aunt's siblings learned the hard way that they should have helped their mother and father out in their old age 13:48
Infinoid heh
MikHel is back. Hi bacek! 13:51
13:56 szabgab joined 14:04 skids joined 14:11 mokurai joined
Coke cd ../san 14:26
ww 14:27
particle1 heh, i finally got padre installed, only to find that typing 'padre' doesn't work, i need to run 'sudo padre'. 14:34
moritz why's that?
szabgab particle, why ?
purl well, particle, why is the file still open in r33938?
szabgab check ~/.padre maybe it is owned by root for some reason 14:35
particle Could not copy the My plugin
szabgab that could be the problem , the ownership of ~/.padre
particle yep, that's it indeed 14:36
jdv79 does padre know pmc an pir?
*and
14:36 Theory joined
szabgab we used to have that bug that a test created it 14:36
I'll look at the tests again 14:37
particle chmod -R parrot:parrot ~/.padre # works
szabgab jdv79, there was a pugin that was doing a naive highlighting of those
I think that was the Parrot plauing
for obvious reasons :-)
but I have not touched it for a while
particle is now one step closer to trac.parrot.org/parrot/wiki/Parrot...lAppliance 14:38
jdv79 the word appliance makes it sound like a refrigerator or an oven but only virtually i guess 14:40
PerlJam it is like those things.
it's furniture.
NotFound I'd like virtual refrigerator full of virtual beers
PerlJam (furniture that does stuff)
jdv79 like those japanese toliets that do way too much stuff? 14:41
particle well, this is a vm you should be able to put on a memory stick, boot up with vmware player, and develop parrot, rakudo, pipp, partcl, mod_parrot, perl6-examples, november, etc, as well as have local copies of the bugs from the respective bug trackers, while editing in padre, vim, or emacs. 14:42
jdv79 i use nano
PerlJam particle: so, the PVA is for language developers?
particle PerlJam: for anyone interested in parrot 14:43
jdv79 (just kidding)
particle /kick jdv79
PerlJam particle: it's just an "integrated parrot environment" so they can to anything parroty then/
jdv79 cool idea though. 14:44
PerlJam ?
s/to/do/
particle PerlJam: something like that.
purl hmmm... something like that. is that true?
particle i use virtual machines for all my development
since they're portable
for this one, i've added prophet and sd, which allows local mirroring of bug trackers 14:45
jdv79 particle: vmware?
purl vmware is falling apart
PerlJam purl: forget vmware
particle allowing me to, for example, update parrot tickets on a plane, and sync with trac when i land
purl PerlJam: I forgot vmware
particle jdv79: yes, vmware 14:46
jdv79 be cool if we had a standard set of images for testing mostly but for dev too...
is that not possible for some reason?
kinda like Alias's PITA idea in spirit 14:47
PerlJam particle: what's missing from the PVA? How many more steps do you need to go?
particle i haven't done packages, just source 14:49
and that may be all i do for rakudo and parrot atm
basically, i'd like to make a script that creates the vm from scratch
check in the script somewhere
then regenerate a pva using the script, and put that on ftp.parrot.org 14:50
then folks that want to help can modify the generator script
oh, and i need to write instructions on cloning the bug queues 14:53
15:20 donaldh joined
moritz particle: can't you distribute the pre-cloned bug trackeers, and write instructions on updating? 15:22
particle moritz: the user needs to enter his credentials to do the pull, so the push updates the ticket appropriately 15:29
moritz particle: good point 15:30
particle i'm debating putting minicpan on the pva 15:32
nah, i'll leave that out 15:33
moritz 1.1GB
(at least that's the size of my checkout)
particle yep, i have one on my laptop, and one on my desktop
i share them with all my VMs
moritz very handy when you're offline 15:34
I haven't got the minicpan web server working though
(haven't tried to hard... some dependencies wouldn't install)
15:38 Theory joined
particle didn't even know about a web server 15:38
moritz CPAN::Mini::Webserver - Search and browse Mini CPAN 15:39
15:40 ilia joined
particle hrmm... my backpan mirror is close to done 15:40
14.1GB atm 15:41
15:53 ruoso joined 15:55 ruoso joined 15:57 ruoso joined
Coke news.bbc.co.uk/2/hi/technology/8148730.stm //LOL memory 15:58
16:05 davidfetter joined 16:06 eiro_ joined 16:07 davidfetter joined 16:23 darbelo joined 16:47 solarion joined
jdv79 i saw code in parrot last night that pokes at the system stack for gc purposes. is that used and is that necessary? 16:52
16:58 Psyche^ joined 17:00 whiteknight joined
dalek l: e606b0b | fperrad++ | (2 files):
use the opcode 'iter' instead of a instanciation of 'Iterator'
 17:09
whiteknight good afternoon #parrot 17:13
dalek rrot: r40119 | fperrad++ | trunk/tools/dev (2 files):
[Languages] use the opcode 'iter' instead of a instantiation of 'Iterator'
 17:16
17:24 donaldh joined
treed Can anyone around answer a question about multiple dispatch? (PDD 27 mentions it, but doesn't quite explain.) 17:24
17:26 chromatic joined
pmichaud treed: I might be able to answer it. 17:26
treed PDD27 mentions that the iter from a MultiSub returns thing in "best match" order. 17:27
But doesn't go into exactly how the determination of "best" is made.
I guess, in particular, is it just the most specific match? (Taking into account parent classes and whatnot.)
whiteknight treed: it compares the manhattan distance between the signatures and sorts them in order of smallest distance 17:28
treed Immediately, I just want to know whether calling a multisub on two CardinalIntegers will take the :multi('CardinalInteger','CardinalInteger') over :multi(_,_)
dalek rdinal: 60a96ea | (Ted Reed)++ | src/classes/Integer.pir:
Add an infix:< to CardinalInteger.
pmichaud treed: yes.
treed ^ that in particular
Okay, cool.
whiteknight treed, yes it will
treed I figured it would, but the docs didn't specify so I figured I'd ask. 17:29
Thanks, guys.
whiteknight no problem
(we should really update the docs)
pmichaud all that multi(_,_) does is (a) signify the sub as a multi, and (b) require at least two arguments
whiteknight but when it takes the difference is signatures, it counts the specific types as being "closer" then the more general types 17:30
treed And it really only acts as a fallback version in the event nothing else matches?
whiteknight that is, multi(_,_) only matches if nothing better does
right
treed Gotcha.
I'd figured that would be the case. (At least, that's how I would have done it.) 17:31
whiteknight it's basically the worse of the acceptable matches
treed "Nothing else here, giving you the backup. Hope it works."
pmichaud well, there can be backups to the backup
for example, :multi(_) and :multi(_,_) will result in different arguments 17:32
sorry, will result in different invocations
treed And :multi(_) will only be called if there's only the one argument?
whiteknight correct
pmichaud and if nothing else more specific matches :-)
whiteknight and if you stand on one leg and should "kalukaleh!" 17:33
treed (Assuming that :(_,_) will act as the back up for anything with 2 or more)
whiteknight s/should/shout/
dalek rdinal: 632831b | (Ted Reed)++ | README:
Update the README to reflect the recent active development.
 17:39
17:40 Theory joined
cotto bacek_at_work, ping 17:45
17:55 szabgab joined
treed I can't recall if :method subs require an extra parameter slot at the beginning for self. 17:58
whiteknight nope, it's automatic
treed In the :multi signature? 17:59
(Meant to have that in there to begin with)
whiteknight oh, that is a good question
I don't know how the invocant object is handled in a multi signature
treed Seems like I've done it before, but I can't remember where, to look it up. 18:00
particle cd parrot && ack \\:multi examples 18:15
heck, might as well check docs/ and t/ while you're at it 18:16
treed 2084:.sub "print" :multi(_, int, int, string) :method
2091:.sub "print" :multi(_, string) :method
The _ certainly indicates something of the sort there.
docs doesn't have anything that specifically mentions it 18:18
18:18 Andy joined
treed although the PDD on multiple dispatch mentions :invocant as an attribute of a .param 18:18
particle _ means any
treed But that seems only to be about marking which params are relevant. 18:19
Rather than which represents the invocant.
whiteknight I don't know if :invocant is implemented yet
treed Tene says that you do need a _ to represent self in method multi sigs. 18:21
And I found where I'd done it before and that agrees. 18:22
whiteknight Tene's pretty bright, you should listen to him 18:24
Infinoid is blinding, but you shouldn't listen to him 18:25
Infinoid also sunburns easily.
Coke :invocant isn't implemented, btw.
18:29 flh joined
treed Interesting. 18:32
I'm trying to debug a test failure, so I run the test again with -t 1, and it goes into a loop declaring something about double free.
(infinite loop, it would seem)
But that only happens when I'm trying to trace.
dalek cnum-dynpmcs: r111 | darbelo++ | trunk/cfg/ (2 files):
More Makefile changes. Moved some operations out of the pmc makefile, and

This should fix the weirdness we were experiencing with BSD make. Removing the src/pmc Makefile is left for later.		 18:36
18:36 allison joined
allison I've got 81 failing tests on Ubuntu karmic, all in t/pmc/packfile.t and t/pmc/packfile*.t 18:38
NotFound treed: tracing has some problems, because in order to print things it must execute code, that can fail.
allison: I think is the same in all platforms.
allison NotFound: that's good to know
NotFound allison: TT #823
allison ah, yup, my intrepid build this morning has the same problem 18:39
NotFound Debian amd64 here, the same.
Coke those have been failing on feather for some time, btw. =-) 18:40
NotFound It must be an epidemy ;) 18:41
dalek rrot: r40120 | allison++ | trunk/config/init/hints.pm:
[install] Provide configuration details for GnuFreeBSD, which uses the

to Petr Salinger for reporting the issue, testing and revising the patch.		 18:55
allison mmm... that's GNU/kFreeBSD 18:56
why can't they use memorable names, like animals or tribes?
Infinoid Most of those are probably trademarked. :( 18:57
allison branch out into different languages?
Kasuku
Infinoid nice
allison (african gray parrot in swahili)
18:58 whiteknight joined
dalek rdinal: 74ff9b6 | (Ted Reed)++ | (4 files):
Add tests for the things I've implemented lately where they didn't already exist.
 19:05
NotFound I have a program called YAYAPN: Yet Another 'Yet Another' Program Name 19:07
dalek rdinal: e355309 | (Ted Reed)++ | t/ (2 files):
A few more tests.
 19:11
rdinal: f09fc6d | (Ted Reed)++ | Rakefile:
Minor change to debug output of Rakefile.
rdinal: f6dbd6a | (Ted Reed)++ | src/classes/Array.pir:
Remove leftover debug say.
rdinal: 4cdc142 | (Ted Reed)++ | src/classes/Hash.pir:
Add a Hash.new that takes a default object.
19:12 Theory joined
dalek rrot: r40121 | NotFound++ | trunk/PBC_COMPAT:
[cage] fix PBC_COMPAT formatting
 19:16
19:20 donaldh joined
allison NotFound: (YAYAPN) excellent :) 19:24
pmichaud treed: in :multi(_,_) used with method, the first item in the :multi corresponds to the invocant 19:28
chromatic jdv79, it's necessary because a PMC may be live but only reachable from the system stack. 19:29
Tene allison: did you ask me to write something up about NCI and callbacks? 19:31
NotFound chromatic: talking about that, don't we must make all PMC * local vars volatile, to prevent being stored in registers? 19:32
allison Tene: that sounds familiar
Tene: that is, writing up what you figured out as you were working on them last week
dalek rrot: r40122 | allison++ | trunk/ports/ubuntu (7 files):
[ubuntu] Store packaging details as they went into Ubuntu Karmic (9.10).
 19:33
Tene and where should I put it?
allison: is parrot 1.4 getting into the next ubuntu release for sure?
allison Tene: how about docs/user/pir? 19:34
Tene Okay. I'll try to work on that tonight. Thanks.
chromatic NotFound, the setjmp trick should suffice to anchor them in a scannable place.
allison Tene: there's a good chance it will get into 1.4, especially if I can get the packages out quickly
Tene: I mean that 1.4 will get into 9.10
Tene :) 19:35
allison Tene: Ubuntu generally prefers to treat it as a sync from Debian, so getting into Debian is the first step
NotFound allison: Are you learning Swahili in order to talk more fluently with the Ubuntu people?
allison Tene: fingers crossed that we can do both nearly simultaneously 19:36
NotFound: heh :) I actually learned it as a kid in Kenya
NotFound: but, hey, you never know what might come in handy
NotFound: speaking of names, I was actually a little offended when I first heard the name "Ubuntu", kind of "what right do they have to that name?" 19:37
NotFound: then I found out the founder was South African. :)
19:38 Andy joined
NotFound allison: a friend of mine make this joke: 'Ubuntu is an ancient afrikan words that means: "I'm too lazy to install Debian" ' X-) 19:39
allison NotFound: :)
NotFound allison: Can you take a look at TT #747 ? 19:46
allison NotFound: sure
NotFound: looks good 19:48
NotFound allison: I was wondering about marking it as experimental, just to be on the safe side. 19:49
allison NotFound: nah, those constants probably should have been exposed from PASM a while ago 19:50
NotFound Ok, I'll close the ticket, then. 19:51
allison NotFound: and, that's a part of the system that isn't slated for revision, so the names won't be changing
19:53 flh joined
dalek TT #747 closed by NotFound++: Constants in library.h are not available as PASM constants 19:56
19:56 RobertLJ1 joined
dalek rrot: r40123 | NotFound++ | trunk/t/compilers/imcc/syn/file.t:
[cage] use constant in libpaths instead of magic value
 20:07
20:40 Theory_ joined 20:48 dalek joined
mikehh still getting a bunch of pmc/packfile errors nearly all come up with Unknown PMC type to thaw -100 20:52
current instr.: '_pbc' pc 35 (t/pmc/testlib/packfile_common.pir:26)
in fact all the pmc/packfile errors first come up with that 20:54
Unknown PMC type to thaw -100 20:55
followed by
current instr.: '_pbc' pc 35 (t/pmc/testlib/packfile_common.pir:26)
then they differ but that seems to me to be the error
currently at r40123 20:57
last ok test at r40111 20:58
moritz yes, that's well known :(
20:59 Theory_ joined
mikehh well there was a codetest failure - which bacek tried to fix resulting in the subsequent errors 20:59
21:00 Theory__ joined
jdv79 chromatic: would that change if the whole L1 thing happens? 21:01
NotFound What codetest failure? I thinked I had fixed all.
21:01 Theory___ joined 21:05 Theory___ joined 21:06 Theory_ joined 21:07 Theory__ joined
mikehh NotFound: the codetest failure was # Failed test 'there are test files in t/pmc for all PMC files' arrayiterator, hashiteratorkey, orderedhashiterator and stringiterator at r40111 21:14
NotFound Ah, yes, forget those. 21:16
chromatic jdv79, it could; we'd have much easier time analyzing lifespan and liveness without worrying about C calling conventions and storage locations.
mikehh when bacek fixed them and I reran the tests at r40117 the pmc/packfile failures were occurring 21:17
dalek rrot: r40124 | allison++ | trunk/examples/tutorial (13 files):
[examples] Documentation fixes for the PIR tutorial examples.
 21:37
21:40 Limbic_Region joined
mikehh NotFound: make test and coretests all fail the same tests nbut also manifest_tests FAIL with # Failed test 'No need to regenerate MANIFEST' - could this be significant? 21:44
NotFound Looks like someone forgot to update MANIFEST. One moment... 21:46
dalek rrot: r40125 | NotFound++ | trunk (5 files):
[cage] update MANIFEST and set svn properties
 21:57
22:03 Theory_ joined
pmichaud here's a related item on the binding issue 22:04
oops, wrong chan
22:06 bacek joined
bacek morning. good morning 22:07
dalek rdinal: 71e0e5e | (Ted Reed)++ | Rakefile:
Make some additions to the statistics reported by the test harness.
rdinal: fe2f4e3 | (Ted Reed)++ | Test.rb:
Fix some of the Test module to output correctly.
rdinal: 9832133 | (Ted Reed)++ | (26 files):
Massive housecleaning on the test files. All plans now reflect the actual number of tests. Anything that crashes is commented out and marked with a skip. Anything that fails is marked with an appropriate todo. "rake test:all" now runs cleanly.
rrot: r40126 | bacek++ | trunk/t/native_pbc (4 files):
[cage] Rebuild native PBCs again. Should fix TT#823
bacek seen fperrad 22:08
purl fperrad was last seen on #parrot 30 days, 1 hours, 23 minutes and 35 seconds ago, saying: Whiteknight, when I encoured the same problem, I send a email (with my IP) to osuosl support [Jun 16 20:40:01 2009]
22:09 Theory joined 22:11 Theory__ joined
NotFound Looks like the packfile test failures are a problem of handling exceptions not working well. 22:23
bacek sigh... 22:30
moritz POST? 22:31
purl POST is, like, Parrot Opcode Syntax Tree or for changing data and uncacheable
moritz thanks purl
NotFound Well, with the native pbc updated it doesn't throw, and then doesn't fail. 22:32
Not all of them, I mean.
mikehh that looks much more better 22:33
bacek NotFound: are they still failing??? 22:34
NotFound The one that fails is TODOed, so all pass now. 22:35
mikehh make test PASSes - running fulltest
22:37 rg1 joined
dalek rdinal: f1b5f9b | (Ted Reed)++ | t/range/each.t:
Additional test cleaning.
 22:46
22:52 rg joined 22:53 bacek joined
bacek me again 22:54
mikehh examples_tests FAIL - All Others PASS (pre/post config, smolder, fulltest) at r40126 - Ubuntu 9.04 amd64
NotFound t/examples/namespace.t fails, looks like if depends on the ordering of a hash 22:55
dalek rrot: r40127 | bacek++ | trunk/t/pmc/hash.t:
[cage] Add tests for Hash.elements
bacek NotFound: it is...
purl Oh no it isn't!
mikehh t/examples/namespace.t FAILs
I have had it pass on and off through the last few revisions 22:56
NotFound If it depends on the ordering of some that we don't grant the order, no wonder. 22:57
Looks like the real problems are solved :) 23:01
bacek Which one?
23:03 payload joined
NotFound The namespace isa hash and we don't grant order in hashes, isn't it? 23:03
mikehh Well all the test failures we were having today 23:04
NotFound Too late for me, I'm mixing questions X-) 23:06
mikehh Well I started with parrot at about 8am this morning - it is now just after midnight here 23:09
bacek Ok... I can't find why Lua is failing. 23:13
23:17 magnachef joined
bacek Time for $dayjob. 23:20
23:20 donaldh joined 23:21 Theory joined
mikehh rakudo faills 1 test/2 subtests - t/spec/S12-attributes/instance.rakudo - Failed tests: 121-122 23:26
23:41 elmex joined 23:43 elmex joined 23:56 patspam joined 23:58 Whiteknight joined
dalek rdinal: 4d319ce | (Ted Reed)++ | Rakefile:
Fix typo in Rakefile.
 23:58

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!