Camelia, the Perl 6 bug

IRC log for #parrot, 2011-04-17

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:18 cotto bacek, do you know of anything that should be added to that file?
00:34 woosley joined #parrot
00:40 shell joined #parrot
01:36 Coke left #parrot
01:36 Coke joined #parrot
01:46 shell left #parrot
02:03 kid51_ left #parrot
02:03 Coke left #parrot
02:03 Coke joined #parrot
02:17 Coke left #parrot
02:18 Coke joined #parrot
02:24 Coke_ joined #parrot
02:24 Coke left #parrot
02:31 Coke_ left #parrot
02:32 Coke joined #parrot
02:39 Coke left #parrot
02:39 Coke joined #parrot
02:46 Coke left #parrot
02:46 Coke joined #parrot
02:46 petdance joined #parrot
02:52 Coke_ joined #parrot
02:52 Coke left #parrot
03:04 Coke_ left #parrot
03:04 Coke joined #parrot
03:08 JimmyZ left #parrot
03:10 soh_cah_toa left #parrot
03:13 theory joined #parrot
03:18 JimmyZ joined #parrot
03:20 Coke left #parrot
03:20 Coke joined #parrot
03:23 dalek parrot: 48f56a3 | petdance++ | lib/Parrot/Pmc2c/Method.pm:
03:23 dalek parrot: Annotate non-SHIM PMC arguments in generated PMC C code
03:23 dalek parrot: review: https://github.com/parrot/parrot/commit/48f56a3545
03:30 cotto dukeleto, ping
03:32 hudnix left #parrot
03:35 Coke left #parrot
03:35 Coke joined #parrot
03:40 ShaneC1 joined #parrot
03:40 ShaneC left #parrot
03:45 contingencyplan left #parrot
03:49 Coke left #parrot
03:49 Coke joined #parrot
04:03 JimmyZ left #parrot
04:06 Coke left #parrot
04:06 Coke joined #parrot
04:19 Coke left #parrot
04:19 Coke joined #parrot
04:26 Coke left #parrot
04:26 Coke joined #parrot
04:44 petdance left #parrot
05:03 Eduardow left #parrot
05:27 dalek tracwiki: v1 | cotto++ | ParrotTalks
05:27 dalek tracwiki: http://trac.parrot.org/parrot/wiki/P​arrotTalks?version=1&action=diff
05:28 benabik I wonder if I can make my way to NC in June...
05:31 cotto benabik, it's always fun to have more Parrot hackers around.
05:32 benabik cotto: It just occurred to me that the main thing I'll be doing over the summer is hacking on Parrot, so heading down there isn't really going to take away from anything else.
05:33 benabik cotto: (Hacking on Parrot, hopefully...)
05:33 cotto benabik, all the better.
05:34 cotto Even if nobody else shows up, I'm sure dukeleto, kid51 and I will spend some quality hacking time together.
05:37 benabik cotto: I'll look into it.  Flights are a bit steep (ROC isn't the cheapest to fly out of), but I might be able to work out something.
05:42 * benabik should probably also ask the wife.  :-D
05:47 cotto that's one advantage of being single
05:49 benabik Aye, but I'm willing to put up with it. ;-)
05:51 cotto not at all saying it's a bad trade-off
06:08 woosley left #parrot
06:09 woosley joined #parrot
06:10 benabik cotto: Are there reg fees?  The site seems a little haphazard.
06:19 woosley left #parrot
06:20 cotto worst header ever
06:21 cotto yapc is intended as an affordable high-quality conference, but there is a fee of some kind
06:23 cotto Once you sign up, it'll be on the "edit profile" page, just where you'd expect
06:23 benabik Mail got eaten by spamassassin.
06:24 cotto special
06:26 benabik Ah, the info I was looking for needed a login.  Whee.  $90 isn't too bad.  Hopefully will be there.  :-)
06:33 cotto great
06:36 benabik It's amusing that it'll be cheaper to rent a car and drive there than to fly.  Although the 12 hour drive is less amusing.
06:36 * benabik is not sure why he's still up.
06:37 cotto you should probably do something about that
06:42 fperrad joined #parrot
07:01 theory left #parrot
07:32 contingencyplan joined #parrot
08:03 cotto anyone familiar with Open Source Bridge?
08:19 tadzik this one in which you see the players' hands?
08:37 plobsing left #parrot
08:41 plobsing joined #parrot
08:52 cotto plobsing, ping (with short timeout)
09:06 contingencyplan left #parrot
09:08 mj41 joined #parrot
09:13 dalek parrot/packfile_wrap: 37ce94f | cotto++ | docs/project/hacking_tips.pod:
09:13 dalek parrot/packfile_wrap: add some tips on reducing unhelpful randomness
09:13 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/37ce94f10e
09:13 dalek parrot/packfile_wrap: 48f56a3 | petdance++ | lib/Parrot/Pmc2c/Method.pm:
09:13 dalek parrot/packfile_wrap: Annotate non-SHIM PMC arguments in generated PMC C code
09:13 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/48f56a3545
09:13 dalek parrot/packfile_wrap: 6f1f07b | bacek++ | / (5 files):
09:13 dalek parrot/packfile_wrap: Merge branch 'master' into packfile_wrap
09:13 dalek parrot/packfile_wrap: review: https://github.com/parrot/parrot/commit/6f1f07bd99
09:42 woosley joined #parrot
09:56 zby_home joined #parrot
10:49 whiteknight joined #parrot
10:55 whiteknight good morning, #parrot
11:04 jrtayloriv left #parrot
11:07 moritz good morning whiteknight
11:18 whiteknight hello moritz
11:51 bacek aloha, whiteknight
11:51 whiteknight good morning bacek
11:51 bacek it's kind of late evening here :)
11:52 whiteknight good late evening, backe
11:52 whiteknight bacek
11:52 bacek :)
11:53 bacek whiteknight, looks like I fixed "all" problems of maser in packfile_wrap banch
11:53 whiteknight bacek: what's the story with the packfile_wrap branch? It
11:53 bacek branch
11:53 whiteknight okay, so can we merge it?
11:53 whiteknight I'm working on unmerging imcc_compreg_pmc from master now, and it's very messy
11:53 bacek At least I couldn't reproduce rakudo's segfault anymore.
11:53 whiteknight is it really much slower?
11:54 bacek Nope. About same as master.
11:54 whiteknight okay, that's what I was hoping
11:54 whiteknight let's merge it
11:54 bacek Let's _review_ it first.
11:54 whiteknight damn the torpedos! Full speed ahead!
11:54 whiteknight okay. I'll review it
11:54 whiteknight :)
11:54 bacek I've made all this changes in kind of harry.
11:55 bacek My confidence level is still quite high, but I want more eyes on it
11:55 bacek "hurry"
11:55 Patterner left #parrot
11:56 whiteknight okay. I'll spend today looking at it and reviewing/testing. We can merge tonight (tomorrow morning for you)
11:56 whiteknight We'll do all the testing while you sleep :)
11:59 bacek Sounds like a good plan :)
12:00 whiteknight bacek++
12:00 * bacek doesn't sleep. He recharges.
12:00 Psyche^ joined #parrot
12:01 Psyche^ is now known as Patterner
12:03 arnsholt left #parrot
12:06 whiteknight I'm going to build and test the branch with ms2 first, since that's the default. When all tests Pass I'm moving over to gms
12:41 arnsholt joined #parrot
12:46 jaffa4 joined #parrot
12:47 jaffa4 hi
13:06 arnsholt left #parrot
13:21 arnsholt joined #parrot
13:27 arnsholt left #parrot
13:41 petdance joined #parrot
14:01 ambs joined #parrot
14:08 arnsholt joined #parrot
14:08 hudnix joined #parrot
14:23 dukeleto ~~
14:27 dukeleto jaffa4: hola
14:27 jaffa4 hola que tal?
14:29 dukeleto jaffa4: enjoying a lazy sunday morning
14:30 jaffa4 that is somthing
14:30 dukeleto cotto: pong
14:30 dukeleto jaffa4: what brings you to #parrot ?
14:31 jaffa4 chance
14:31 dukeleto msg benabik yes, you should go to YAPC::NA this year. There will be some kind of hackathon
14:31 aloha OK. I'll deliver the message.
14:31 jaffa4 I added one line for debuggin parrot
14:32 jaffa4 and the result is crash
14:32 dukeleto jaffa4: fun!
14:32 jaffa4 sound like fun?
14:32 dukeleto jaffa4: people in here have a warped sense of fun :)
14:33 jaffa4 some people like challenges
14:33 dukeleto jaffa4: i think it is safe to say that parrot hackers like a good challenge
14:33 dukeleto jaffa4: what was the one line debug statement?
14:34 jaffa4 puts(full_name);
14:34 dukeleto jaffa4: can you gist/nopaste the diff?
14:35 jaffa4 library.c  line 680, insert that line.
14:35 dukeleto jaffa4: if you need help debugging something about parrot, asking in here and on the parrot-dev mailing list are the best options
14:35 dukeleto jaffa4: which version of parrot?
14:36 dukeleto jaffa4: also, what are you trying to debug?
14:36 dukeleto jaffa4: is something not working as you would expect?
14:36 jaffa4 3.2.0 commit 2c7060dd68e4c2badd44a140597a0bb7d82e440d
14:36 jaffa4 Yes
14:37 jaffa4 I am trying to execute Perl6 pbc
14:37 jaffa4 and I get "load_language" couldn't find a compiler module for the language 'perl6'
14:37 jaffa4 by the way
14:37 dukeleto jaffa4: ah
14:37 jaffa4 What would you use for debugging?
14:38 dukeleto jaffa4: sounds like you didn't install perl6, only compiled it. does that sound right?
14:38 dukeleto jaffa4: rakudo perl6 needs to be installed for some things to work correctly, like library loading
14:38 jaffa4 per6 installed
14:38 jaffa4 working
14:38 jaffa4 I just want to run pbc
14:39 jaffa4 using parrot
14:39 dukeleto aloha nopaste?
14:39 aloha dukeleto: nopaste is is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
14:39 dukeleto jaffa4: can you nopaste exactly what commands you are running and the error messages?
14:39 dukeleto jaffa4: that will help us help you
14:39 jaffa4 http://nopaste.snit.ch/40730
14:43 dukeleto jaffa4: have you asked the people in #perl6 on freenode? They may be able to help as well
14:43 jaffa4 yes
14:43 jaffa4 asked
15:03 jaffa4 ok
15:03 jaffa4 I can debug it now
15:09 Coke left #parrot
15:09 Coke joined #parrot
15:11 jaffa4 hi Cokr
15:11 jaffa4 hi Coke
15:19 dukeleto jaffa4: did you make progress?
15:19 jaffa4 yes
15:19 dukeleto jaffa4: good to hear
15:19 dukeleto jaffa4: what was the issue?
15:19 jaffa4 path is not set
15:20 jaffa4 or I do not know how to set the oath
15:20 jaffa4 -L
15:20 jaffa4 -I has not effect
15:20 dukeleto jaffa4: aha!
15:20 jaffa4 This is what I expected somewhere
15:21 jaffa4 The difference is I know now what the path is
15:21 jaffa4 so I managed to test the speed of small Perl 6 program
15:34 benabik ~
15:36 benabik dukeleto: I'm going to try to get to YAPC::NA.  Possibly depends on finding a friend or maybe someone from Buffalo.pm to share travel/room expenses with.
15:38 theory joined #parrot
15:42 dukeleto benabik: sounds awesome. I promise you will have fun.
15:42 dukeleto jaffa4: what are your results?
15:42 jaffa4 very bad
15:43 jaffa4 a loop from 1..10000 that contains $p++
15:43 jaffa4 ran for 50 seconds
15:43 jaffa4 for 1..10000 { $p++}
15:44 benabik jaffa4: In Rakudo?
15:44 jaffa4 rakudo
15:44 woosley left #parrot
15:45 jaffa4 without compilation times
15:45 jaffa4 just running pbc
15:45 benabik That does seem poor, I got 4.9 sec for `time perl6 -e 'my $p; $p++ for 1..10000; say $p'`
15:47 jaffa4 sorry
15:47 jaffa4 100000
15:47 jaffa4 I missed one digit
15:48 jrtayloriv joined #parrot
15:48 benabik Huh.  I couldn't get perl6 to output a pbc.  Wierd.
15:50 ambs left #parrot
15:51 jaffa4 I do not know how weird is that
15:51 jaffa4 perl6 outputs pir
15:51 jaffa4 then I convert pir into pbc using parrot
15:51 jaffa4 thatI run pbc using parrot
15:52 benabik jaffa4: 50 seconds then seems above average.  I just got 1m29.757s
15:52 Coke left #parrot
15:52 Coke joined #parrot
15:52 benabik (Although I'm just running via -e so that includes compilation, etc.)
15:53 jaffa4 compilation 1 second
15:53 benabik jaffa4: Compiling using perl6 -o ?
15:54 jaffa4 yes
15:55 benabik That does nothing for me.  That's the weird part.
15:56 jaffa4 perl6 --target=pir  -oD:\m\p6\pirout.pir  D:\m\p6\e.p6?
15:56 benabik And ruby accomplishes the same task in 0.039 seconds.  Bleh.
15:56 benabik Ah.  Forgot the --target=pi
15:56 benabik *pir
16:04 davidfetter left #parrot
16:04 jaffa4 not a oroblem of parrot
16:04 jaffa4 similar program written in pir
16:04 jaffa4 runs in 0.05 seconds
16:05 benabik It does $p 100_000 times in only 30s, so the dispach and execution of postfix:<++> is a big part of the problem, although that loop could likely be tightened somewhere.
16:06 jaffa4 here is pir program
16:06 jaffa4 that does it more efficiently
16:06 jaffa4 http://nopaste.snit.ch/40731
16:07 benabik jaffa4: Do you know enough NQP to try it there?  I'm curious where the inefficiency is, exactly.  I'm betting some of it is because Rakudo using Integer PMCs instead of Integer registers.
16:08 jaffa4 What is NOP?
16:08 benabik Ah.  Well, I'll try after lunch.
16:08 jaffa4 I think
16:08 benabik NQP = Not Quite Perl
16:09 benabik It's the language much of Rakudo is written in.  (installed by parrot as parrot-nqp)
16:09 jnthn__ I'm curious how Rakudo should know to use integer registers in that program. :)
16:09 jaffa4 by program analysis
16:09 jnthn__ *sigh*
16:09 benabik jnthn__: True enough, but the PIR code he wrote uses the registers, so I'm wondering if the PMCs are that much slower.
16:10 jnthn__ benabik: It's *much* more than that, that makes it slow.
16:10 jaffa4 I guess that problem is that variables are looked up by their name
16:11 jnthn__ jaffa4: There's a *lot* of reasons. I talked about a lot of them in http://www.jnthn.net/papers/2011-nlpw-rakudo.pdf
16:12 jaffa4 ok. noone said rakudo was written for performance
16:12 jnthn__ That hasn't been a focus so far.
16:12 jnthn__ It's becoming one now that it's got a lot of features in place.
16:12 jaffa4 yes
16:12 jaffa4 it is clear
16:12 jaffa4 I just wanted to have first hand experience
16:13 benabik jaffa4: Rakudo has nothing resembling an optimization phase, AFAICT.  Static analysis probably can't get down to integer registers, but might be able to only perform the variable lookup once.
16:13 benabik (Might.)
16:13 benabik jnthn++ # Awesome and informative presentations
16:15 jnthn__ In NQP I played with an optimization that can reduce various lexicals to local register accesses. Full-blown Perl 6 is much more complex than NQP so it's harder to apply the optimization...needs more analysis.
16:16 jnthn__ I'm still hopeful it's do-able though. :)
16:16 benabik jnthn__: Shouldn't that be possible with tree analysis?
16:16 benabik Or does full Rakudo skip PAST somewhere?
16:18 jnthn__ No, Rakudo certainly makes PAST.
16:19 benabik I'd think tree-analysis on the PAST would make the job similar between NQP and Rakudo.  (Admittedly, I haven't tried.)
16:20 petdance left #parrot
16:21 jnthn__ There's probably plenty of interesting work there, yes.
16:21 jnthn__ At the moment I'm worrying more about optimizing our building blocks.
16:22 benabik Fair enough, I don't actually want to distract you from the awesomeness of 6model
16:22 jnthn__ :)
16:23 jnthn__ Yes, my tuit supply for that seems to have improved. Hoping to get lots of progress in the next couple of weeks. Also started doing more docs...
16:23 jaffa4 where is 6model?
16:24 benabik Oooo, documentation
16:24 jnthn__ jaffa4: The Parrot implementation of 6model lives in the nqp repo at the moment.
16:25 jnthn__ benabik: Yes. Want to make examples of how to build different kinds of object system with 6model. :)
16:26 jaffa4 gradual typing
16:26 jaffa4 WHat is it?
16:26 benabik jnthn__: So you can say RTFM when people say "But [insert language here]'s object system won't work in 6model"?
16:26 jnthn__ benabik: ;)
16:27 jnthn__ jaffa4: It's where instead of either being fully dynamic or fully static, instead programs can choose to add more type information. Then the compiler can use it to produce better programs and give more safety.
16:27 cotto dukeleto, do you have some notes on what your yapc::na parrot talk will cover?  It'd be nice to minimize overlap.
16:28 jaffa4 What more type information?
16:29 benabik jaffa4: Things like "my Int $p" instead of "my $p"
16:29 cotto whiteknight, ping
16:29 whiteknight pong
16:29 jaffa4 oh I know what you mean
16:30 jnthn__ benabik: Yes, that.
16:30 cotto whiteknight, kid51 had a couple fairly minor patches for tickets he wanted to see in 3.3.  Are we frozen yet?
16:30 whiteknight not frozen yet, no. go ahead
16:30 cotto ok
16:31 whiteknight I'm probably going to merge bacek's packfile_wrap branch tonight. That will mark the start of feature freeze
16:31 whiteknight bugs and fixes of course are always welcome
16:31 cotto awesome
16:32 cotto bacek++ did some good work there
16:46 dalek parrot: 370a418 | cotto++ | / (4 files):
16:46 dalek parrot: Introduce --with-llvm option to explicitly link to LLVM
16:46 dalek parrot:
16:46 dalek parrot: fixes tt #2092, patch courtesy of kid51++
16:46 dalek parrot: review: https://github.com/parrot/parrot/commit/370a41843c
16:46 dalek parrot: 73edd49 | cotto++ | t/op/spawnw.t:
16:46 dalek parrot: Eliminate use of 'env'
16:46 dalek parrot:
16:46 dalek parrot: fixes tt #2095, patch courtesy of kid51++
16:46 dalek parrot: review: https://github.com/parrot/parrot/commit/73edd49575
16:50 dalek TT #2095 closed by cotto++: t/op/spawn.t: failure due to Perl 5 problem
16:50 dalek TT #2095: http://trac.parrot.org/parrot/ticket/2095
17:01 luben whiteknight, I have tested packfile_wrap branch, (--gc=gmc --optimize=-O3). It compiles rakudo, its tests do not segfault, I got test fail only in t/spec/S02-builtin_data_types/instants-and-durations.t but I think this failure is not related
17:01 jaffa4 What is  IGLOBALS_LIB_PATH?
17:05 dalek TT #2092 closed by cotto++: config/auto/llvm.pm:  Offer explicit optout
17:05 dalek TT #2092: http://trac.parrot.org/parrot/ticket/2092
17:12 cotto jaffa4, that's an array of arrays of paths where parrot will look for different kinds of libraries
17:13 cotto see enum_lib_paths in include/parrot/library.h for the list of what gets stored in that array
17:15 tadzik luben: that fail happens
17:15 tadzik luben: if you retry it, you'll probably get a PASS
17:16 tadzik my guess is just that Rakudo can't instantiate time object fast enough, and they don't match :)
17:40 mj41 left #parrot
17:54 rohit_nsit08 joined #parrot
17:54 rohit_nsit08 hello #parrot
17:57 cotto I can't believe how many interesting-looking presentations OS Bridge has.
17:57 cotto hio rohit_nsit08
18:00 contingencyplan joined #parrot
18:02 cotto rohit_nsit08, have you started hacking yet?
18:04 cotto also, as someone who's continuing to realize how deep js can get, I'd highly recommend http://www.w2lessons.com/2011/0​4/you-dont-know-javascript.html
18:09 SHODAN joined #parrot
18:12 kid51 joined #parrot
18:12 rohit_nsit08 cotto: hi, sorry was away from system. yup i am studying winxed's stage 1 compiler. and some PIR also
18:14 rohit_nsit08 cotto:i am focusing on implementing prototype based objects right now
18:16 whiteknight luben: Thanks. the instants test is not related
18:31 ambs joined #parrot
18:57 preflex left #parrot
18:58 Eduardow joined #parrot
18:59 contingencyplan left #parrot
19:00 preflex joined #parrot
19:01 contingencyplan joined #parrot
19:34 S_Arms joined #parrot
19:38 kid51 Over the last few days, I've been running 'make test' and 'make fulltest' in both master and the packfile_wrap branch.  I have generally gotten the same, good results in either branch on both Linux/i386 and Darwin/PPC.
19:38 whiteknight awesome
19:38 whiteknight I'm testing packfile_wrap now with rakudo
19:38 kid51 Which makes me hopeful that if we merge packfile_wrap into master, we will continue to get good results.
19:39 kid51 Sadly, neither of my boxes has enough free space to test rakudo effectively.
19:39 kid51 So I haven't tried to build Rakudo in 8 months.
19:39 tadzik I tried it with packfile_wrap few days ago, it went okay
19:39 tadzik I mean, the spectests passed
19:50 pjcj left #parrot
19:56 Tene kid51: if I gave you ssh access to an x86_64 box with plenty of space, would you have any use for it?
19:58 Andy_ joined #parrot
19:59 theory left #parrot
20:15 soh_cah_toa joined #parrot
20:27 kid51 Tene:  Well, the only thing that I would possibly want to do that I can't do now is to test Rakudo on top of parrot or parrot branches.
20:27 kid51 So, yes, that would be useful ... though perhaps more for Rakudo than for me ;-)
20:28 * kid51 would also welcome offers for access to boxes on other OSes, particularly FreeBSD or OpenBSD.
20:31 pjcj joined #parrot
20:34 dalek nqp/ctmo: 1d1ea2f | jonathan++ | src/metamodel/how/NQP (4 files):
20:34 dalek nqp/ctmo: Get packages meta-objects to set a hash in their .WHO (can promote it to some more full-on Stash type later, probably, but this is fine for now.) (Foo.WHO)<$x> := 42 style stuff now works fine.
20:34 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/1d1ea2fda4
20:34 dalek nqp/ctmo: 06f82a8 | jonathan++ | src/metamodel/knowhow_bootstrapper.c:
20:34 dalek nqp/ctmo: Should have a stash for KnowHOWs too.
20:34 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/06f82a86fc
20:34 dalek nqp/ctmo: 9869549 | jonathan++ | src/NQP/ (2 files):
20:34 dalek nqp/ctmo: s/$*PKGMETA/$*PACKAGE/ - it's the compile time $?PACKAGE, after all. Install $?PACKAGE. Set up a GLOBALish, though nothing is done with it just yet.
20:34 dalek nqp/ctmo: review: https://github.com/perl6/nqp/commit/9869549033
20:48 Tene kid51: Would you have interest in a VM server, to run other OSs on?  I'm not interested in installing other OSs myself, but I could give you a few VMs on my server to to use.
20:52 kid51 Tene: Well, installing other OSes on VMs would probably be more work than I could take on right now.
20:53 Tene kid51: 'k.  feel free to let me know if that changes.  :)
20:54 kid51 But the x86_64 ... that would be Linux? which distro?
20:55 Tene kid51: yeah, linux.  Probably fedora or CentOS.  I'm replacing my old underpowered i386 home server as soon as I get around to installing an OS on the new box.
20:59 SHODAN left #parrot
21:00 dafrito left #parrot
21:18 dafrito joined #parrot
21:27 gbens joined #parrot
21:31 fperrad left #parrot
21:33 ambs left #parrot
21:40 gbens hello
21:50 Andy_ left #parrot
21:59 kid51 nopaste?
22:00 kid51 aloha, nopaste is http://nopaste.snit.ch/
22:00 aloha kid51: ... but nopaste is is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl) ...
22:00 kid51 nopaste
22:01 benabik aloha: nopaste?
22:01 aloha benabik: nopaste is is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
22:01 kid51 So aloha insists on being followed by a colon.
22:01 benabik aloha, nopaste?
22:01 aloha benabik: nopaste is is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
22:01 benabik aloha nopaste?
22:01 aloha benabik: nopaste is is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
22:01 benabik kid51: Looks like aloha just insists on being addressed.
22:03 kid51 Yes, but when I privmsg aloha with that wording, she says she has no idea.
22:03 bacek good morning, humans
22:03 nopaste "gbens" at 192.168.1.3 pasted "prove -v t/dynoplibs/debug." (35 lines) at http://nopaste.snit.ch/40741
22:05 nopaste "kid51" at 192.168.1.3 pasted "t/dynoplibs/debug.t test #5 -- the one we're trying to debug now on darwin/intel" (11 lines) at http://nopaste.snit.ch/40742
22:09 kid51 gbens: Yes, let's get that nopaste into a Trac ticket.
22:09 kid51 Have you set up a Trac account for yourself, yet?
22:09 gbens yes
22:11 kid51 So you should now be able to go to trac.parrot.org, log in (or refresh), and hit the New Ticket button
22:12 petdance joined #parrot
22:12 gbens I've hit the new ticket button
22:13 benabik kid51: t/dynoplibs/debug.t passes for me on Darwin/intel.
22:13 kid51 benabik: Hold on.  We're trying to figure out why it passes on some boxes but not others.
22:14 kid51 gbens:  For Summary, enter: t/dynoplibs/debug.t: fails one test on some Darwin
22:14 benabik kid51: Let me know if I can help.  I'll be distracted but will keep an eye out.
22:14 kid51 Thx.  We're training gbens in how to file a TT as well.
22:15 gbens kid51: ok, I've put the summary
22:15 kid51 For Description:  Ordinary text, you just type.  Code blocks start with {{{ at the left margin, code starting on the next line, then code block closes with }}} at the right margin.
22:16 kid51 So you say something like:  In my Smolder reports from Mac OS X 10.whatever, this test consistently fails for me.
22:16 kid51 Then, take the code that you just pasted and plug it in between the triple-brace lines.
22:16 kid51 Lemme know when you've got that.
22:21 gbens kid51: ok
22:21 jaffa4 left #parrot
22:21 kid51 So now you see that there are various panes and such near the bottom of the page.
22:21 kid51 For component: Select 'testing'
22:21 kid51 For Platform: darwin
22:22 kid51 Then you can hit Preview to see if it displays okay.
22:22 kid51 Then you hit Create ticket
22:22 kid51 Don't worry too much about it; we can always fine-tune the description
22:23 gbens kid51: ok, ticket created
22:28 kid51 Yes.  I just added a comment with the code for the test that's failing.
22:28 kid51 And I re-ragged some lines in the description.
22:28 kid51 gbens: In that ticket, can you post the version of Mac OS X you're on?
22:29 kid51 IIRC one of the Parrot GSOC projects is to overhaul the Parrot debugger.
22:29 kid51 Assuming that project gets assigned, we can have the GSOC student look at that ticket.
22:30 kid51 benabik: Can you add a note to TT #2096 about the version of Mac on which you do *not* get that failure?
22:30 gbens kid51: the OS version is in the first line of the description.
22:30 gbens is there someplace else I should put it?
22:31 kid51 No, that's okay.
22:31 dalek TT #2096 created by gbens++: t/dynoplibs/debug.t: fails one test on some Darwin
22:31 dalek TT #2096: http://trac.parrot.org/parrot/ticket/2096
22:32 kid51 And that's an Intel mac, correct?
22:32 gbens Mac mini, Intel Core 2 duo
22:33 kid51 Thanks.
22:33 kid51 is now known as kid51_at_dinner
22:34 gbens left #parrot
22:35 benabik kid51_at_dinner: Comment added.
22:36 benabik Hm.  I guess he's at dinner.  He'll see it when he gets back.
22:48 marcel_r joined #parrot
23:08 mikehh joined #parrot
23:34 bacek_at_work ~~
23:37 kid51_at_dinner left #parrot
23:41 soh_cah_toa cotto: ping
23:43 cotto soh_cah_toa, pong
23:43 cotto working on the library paths bug?
23:43 soh_cah_toa cotto: i have a question about src/library.c
23:43 soh_cah_toa cotto: beat me too it
23:44 soh_cah_toa cotto: it looks like the . directory is pushed onto "paths" 3 times
23:44 cotto no need to ask to ask
23:44 soh_cah_toa cotto: i know i should just delete all 3 occurences of VTABLE_push_string but why is it like that? why push it 3 times?
23:45 cotto soh_cah_toa, that's sloppy reuse of a variable
23:45 cotto or distracting, at least
23:45 soh_cah_toa cotto: yes, verry
23:45 soh_cah_toa *very
23:45 cotto note that paths is a separate PMC each time it gets . pushed
23:46 soh_cah_toa cotto: alright, i'll remove it. i just wanted to make sure there was something special about it before i removed those lines
23:46 cotto If you wanted to make several different variables for each kind of path, I wouldn't complain.
23:48 kid51 joined #parrot
23:50 soh_cah_toa cotto: all i really have to do though is remove lines like "entry = CONST_STRING(interp, "./"); VTABLE_push_string(interp, paths, entry);", right?
23:50 lucian joined #parrot
23:51 cotto soh_cah_toa, that's about the extent of it, as far as coding goes
23:51 cotto it'll be important to run parrot's tests and see how Rakudo does with that change, though.
23:52 soh_cah_toa cotto: right
23:52 cotto I'd also hold off until after 3.3 gets cut.  Changing something like that right before the release is asking for it.
23:53 soh_cah_toa cotto: okay
23:53 soh_cah_toa cotto: which tests in particular should i run?
23:54 cotto I'd run all of them, but I'd focus on the dynop and dynpmc tests
23:54 cotto or anything having to do with library loading
23:54 cotto For bonus points, write a test that a file in the cwd doesn't get loaded.
23:55 * cotto needs to take off
23:55 cotto happy testing
23:55 soh_cah_toa cotto: okay, thanks

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

Parrot | source cross referenced