Camelia, the Perl 6 bug

IRC log for #parrot, 2012-08-03

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:12 lucian joined #parrot
00:20 whiteknight joined #parrot
00:25 whiteknight good evening, #parrot
00:27 benabik joined #parrot
00:48 whiteknight I think I've got my linux/i386 vm up and running now
00:48 whiteknight so that's one platform off my list
00:49 whiteknight the fact that my mouse doesn't work makes using it unnecessarily difficult
00:50 benabik Mice are for the week!
00:51 benabik *Weak
00:52 whiteknight it's not a matter of the mouse just not working. It doesn't take my input, but it will freak out when my palms rub against it, jump the mouse around, click things, and activate gestures that I don't know how to turn off
00:53 whiteknight and sometimes it does this when I swear my palm has not rubbed it in any way
00:53 benabik In the VM or just anytime?
00:53 whiteknight the whole laptop. At all times
00:54 benabik That's less than ideal.
00:54 whiteknight just barely, slightly, less than ideal
00:57 whiteknight If I go through all this hassle, and I don't see a test failure, I'm going to be annoyed mightily
00:58 whiteknight heh, I do see it!
00:58 whiteknight yay for test failures!
01:12 kid51 joined #parrot
01:20 dalek parrot/whiteknight/io_cleanup1: dd4ce1b | Whiteknight++ | / (3 files):
01:20 dalek parrot/whiteknight/io_cleanup1: Add in a new Parrot_io_buffer_size legacy routine to provide a fake, backwards-compatible behavior for the FileHandle.buffer_size method, which is worthless in the new system but which does appear in some ill-conceived tests.
01:20 dalek parrot/whiteknight/io_cleanup1: review: https://github.com/parrot/parrot/commit/dd4ce1b7d4
01:23 whiteknight msg kid51 the linux/i386 failures on whiteknight/io_cleanup1 should be fixed now. Uninitialized variable. In fact, I think that test should have been failing on more systems too
01:23 aloha OK. I'll deliver the message.
01:24 kid51 Hmm.  Just now I was building on that platform and on that branch but with all g++.  'make' failed.
01:25 kid51 I will paste that build failure, then I will re-test with all gcc.
01:25 alester_ joined #parrot
01:27 nopaste "kid51" at 70.85.31.226 pasted "io_cleanup1 branch on linux/i386: build failure with all g++" (608 lines) at http://nopaste.snit.ch/153650
01:27 kid51 That may not be the most recent commit, however.
01:28 whiteknight I'm testing g++ on linux/x86_64 and linux/x86 right now
01:29 whiteknight kid51: I've been trolling through craigslist looking for reasonably priced, working macs with PPC
01:29 whiteknight they are hard to find like unicorn
01:31 kid51 linux/i386: io_cleanup1 branch:  all-gcc build:  t/pmc/filehandle.t test #14 now passes, but test #28 still fails
01:31 whiteknight ah, I think I found the g++ build failure
01:32 whiteknight it's related to my last round of readline fixes
01:32 nopaste "kid51" at 70.85.31.226 pasted "io_cleanup1 branch on linux, all gcc build: down to 1 failing test in t/pmc/filehandle.t" (57 lines) at http://nopaste.snit.ch/153651
01:33 * kid51 whiles away his evenings refactoring 17-year-old non-encapsulated code in the Perl 5 core distribution.
01:36 whiteknight sounds like fun
01:38 kid51 whiteknight: read your last post to list; glad you identified the shortage of core developers as part of the problem.
01:38 kid51 That saved *me* from having to post about that.
01:38 benabik sounds like
01:38 dalek parrot/whiteknight/io_cleanup1: 1705fc4 | Whiteknight++ | src/pmc/socket.pmc:
01:38 dalek parrot/whiteknight/io_cleanup1: fix a readline-related build failure on g++. kid51++ for the catch
01:38 dalek parrot/whiteknight/io_cleanup1: review: https://github.com/parrot/parrot/commit/1705fc4609
01:38 benabik sounds like "fun" (FIFY)
01:39 whiteknight kid51: Lots of different devs have lots of individual reasons why they can't or won't hack on parrot anymore
01:39 whiteknight but it's all come to a bit of a head recently
01:39 kid51 Now testing 1705fc4 all g++ on linux
01:41 kid51 Okay, the branch built on all g++
01:41 whiteknight nice
01:41 whiteknight kid51++
01:41 kid51 And in all g++, prove t/pmc/filehandle.t PASS
01:42 kid51 'make test' PASS.  http://smolder.parrot.org/app/​projects/report_details/30374 Let me try an optimize as well.
01:42 whiteknight I'll have to see that result tomorrow morning. It's coming up on my bedtime now
01:43 benabik I used to think my parents were crazy for going to be so early.  Now I say "why did I stay up until 11?"
01:44 kid51 'cause you only had work and school; you didn't have: work and school and baby
01:44 whiteknight I try to get into work very early, so I can leave early and have more afternoon at home. That means I have to get to bed early so I can wake up early...
01:44 whiteknight anyway, Whiteknight Out!
01:46 kid51 whiteknight/io_cleanup1 branch: linux/i386: all-g++ build with optimize: PASS: http://smolder.parrot.org/app/​projects/report_details/30375
01:48 kid51 Damn, still getting a test failure with all-gcc build.
01:50 nopaste "kid51" at 70.85.31.226 pasted "io_cleanup1 branch: linux/i386: all-gcc build: still that same failure in t/pmc/filehandle.t" (12 lines) at http://nopaste.snit.ch/153661
01:51 kid51 whiteknight/io_cleanup1 branch: linux/i386: all-gcc build with optimize: FAIL: http://smolder.parrot.org/app/​projects/report_details/30376; http://nopaste.snit.ch/153661
01:51 kid51 t/pmc/filehandle.t linux/i386 fails test 28 on io_cleanup1 branch with all-gcc build no-optimize/optimize
02:04 nopaste "kid51" at 71.255.49.227 pasted "io_cleanup1 branch: darwin/ppc (gcc/g++g++): fails 1 test with write error" (12 lines) at http://nopaste.snit.ch/153671
02:05 kid51 So the io_cleanup1 branch, when built with gcc, is failing t/pmc/filehandle.t test #28 on both linux/i386 and darwin/ppc, and with a spectacular error in the darwin/ppc case
02:05 kid51 While earlier today, that file PASSed entirely on that branch on darwin/*i386*!
02:06 kid51 enough parrot for tonight
02:12 kid51 On Darwin/PPC, I can induce that "Write error: No space left on device" error simply with:  ./parrot t/pmc/filehandle_28.pir
04:54 benabik joined #parrot
05:56 fperrad joined #parrot
07:16 brrt joined #parrot
07:28 lucian joined #parrot
09:13 brrt joined #parrot
11:21 JimmyZ joined #parrot
11:43 mtk joined #parrot
12:09 dalek rakudo/nom: 1f662c9 | (Felix Herrmann)++ | src/core/Temporal.pm:
12:09 dalek rakudo/nom: change constructor of Date to use default values for year/month/day
12:09 dalek rakudo/nom:
12:09 dalek rakudo/nom: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
12:09 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1f662c9aa6
12:29 dalek nqp: 553ee86 | (Arne Skjærholt)++ | src/ (2 files):
12:29 dalek nqp: Expose nativecall write-barriering through an nqp::op.
12:29 dalek nqp: review: https://github.com/perl6/nqp/commit/553ee862f3
13:20 bluescreen joined #parrot
13:23 JimmyZ joined #parrot
13:26 * Coke doesn't think whiteknight should be trolling craiglist to buy hardware to support a platform that can never be better than 5 years obsolete.
13:26 Coke there's not point in running parrot great on something that few people are going to run parrot on.
13:28 * JimmyZ agrees
13:29 Coke that said, I also agree with the fact that sounds like threads isn't ready to merge and needs more eyes.
13:29 moritz I don't think mac powerpc support is very important. {Net,Free}BSD is
13:29 Coke but that isn't a reason to keep darwin/PPC support. Great it if works, but not something we should be spending tuits on.
13:30 JimmyZ well, I am +1 mark the failing nci tests to todo
13:30 Coke moritz: those are more important than powerpc, sure.
13:30 JimmyZ rakudo doesn't use parrot's nci
13:32 * JimmyZ wonders who uses parrot's nci system on *BSD
13:47 Coke "why do we even have that lever?"
13:48 * JimmyZ doesn't know where that lever was from
13:48 Coke a thing that would be helpful: a list of all the mismatches between rakudo and parrot and a plan to unify them.
13:49 Coke and/or document a valid reason why the mismatch should exist.
13:51 * Coke supposes he should try again to kill PASM.
13:52 * Coke wonders what happened to chromatic's recent appearance, branch, and disappearance.
13:52 * JimmyZ hopes kill pir and use m1 eventually too
13:52 * Coke wonders what happened to m0.
13:53 * Coke wonders if we need our too infrequent "what the hell are we doing, anyway" meeting.
13:54 * JimmyZ is looking forward to not_gerd's m0
13:54 moritz I'm pretty sure Parrot could benefit from a developer's summit
13:54 moritz as does Rakudo
13:55 atrodo Weren't those suppose to happen quarterly? Or have I just missed them?
13:55 Coke I think "slightly more regularly than we elect new officers", but that's not happening either.
13:58 Coke seen allison?
13:58 aloha allison was last seen in #parrot 16 days 18 hours ago saying "Coke: I'm really not sure what the problem is on ia64".
14:08 moritz is allison mentoring a gsoc student right now?
14:12 PacoAir joined #parrot
14:21 rurban msg whiteknight: I've purchased yesterday a mac powerbook 4 ppc32 with macos 10.5 for $75 from a fellow perl hacker for parrot smokes.
14:21 aloha OK. I'll deliver the message.
14:22 rurban The first mips32 smoke is there also: http://smolder.parrot.org/app/​projects/report_details/30372, sparc32 and ppc32 soon to follow
14:23 rurban Coke: p5p is having a summit before YAPC::Europe in a german town called Perl
14:25 rurban I've only found a Parrot, KY, USA near towns called London, Manchester and Thousandsticks
15:01 brrt joined #parrot
15:03 brrt ping whiteknight
15:04 brrt t
15:05 brrt tell whiteknight i might be able to arrange *bsd / solaris virtual machines
15:06 brrt it won't be free of charge but i've found somebody capable at least
15:12 dmalcolm joined #parrot
15:24 Coke That reminds me, I wonder what the PaFo finances are like at the moment.
15:27 brrt yeah i wonder about that a bit, too
15:28 brrt a coworker from $job can provide vms (he has his own hosting company)
15:28 brrt just not free as in beer
15:38 Coke I'm asking around as well. I'm thinking we don't really have a budget here.
15:39 Coke I have a super crappy *bsd box running at home that I used to be able to build parrot on. I'll see if I can resurrect it.
15:40 Coke think it might /barely/ be a pentium.
15:42 rurban brrt: I have a lot of virtual solaris and bsd machines as debian qemu/kvm images around
15:43 rurban just no sparc64 yet. A coworker promised a UltraSparc though. sun4u with sol10
15:44 Coke rurban++ #testbot
15:44 rurban I'll remove my private stuff from it and put them in my webserver for others also
15:45 rurban mips32 and sparc32 are already privatized and almost ready to publish
15:45 pmichaud the few times I tried to install *bsd, I didn't have much luck.  I'm just too unfamiliar with the toolsets, I suspect.
15:45 rurban pkg sucks a bit
15:46 pmichaud That said, there are several *bsd enthusiasts here in the Dallas area that have dedicated co-location boxes, and some solaris guys too; perhaps I could put out a call to the local unix group and see if any of them could help.
15:47 pmichaud I should wait a day or so to do that, though; many of them are off at the Texas Linux Fest this weekend.
15:47 rurban solaris on intel is really to setup though in a vm. there's even a free sol24-sparc32 on the web somewhere, but this has no compiler.
15:47 rurban really trivial
15:58 benabik joined #parrot
15:58 benabik ~~
16:00 rurban pmichaud: I"m also going to San Antonio tommorrow!
16:07 rurban kid51: on my mac g4 libpcre fails to load with the nci test. I'm using macports
16:41 * rurban off to San Antonio, bye
17:09 Coke ~~
17:09 Coke (belatedly, apparently!)
17:45 dngor joined #parrot
18:59 allison Coke: I'm around
18:59 allison moritz: no, I'm not mentoring this year
19:00 allison atrodo: If you mean face-to-face developer summits, then they're funding dependent (and we don't really have any)
19:00 allison atrodo: The virtual summits are supposed to happen just after each supported release.
19:00 allison atrodo: But, I have the sense that hasn't been maintained.
19:01 atrodo allison: those were what I meant
19:01 * moritz things non-virtual summits are needed at the moment
19:01 moritz *thinks
19:01 allison atrodo: all that's needed for those is one developer to say "hey, let's have one, how is this time for everyone"
19:02 allison moritz: they are helpful, but can be expensive
19:02 moritz I know
19:02 allison moritz: best if they're attached to an event where many people are already going
19:14 * Coke suggests trying to have one at yapc::eu.
19:17 moritz how many parrot devs are going there?
20:05 zby_home joined #parrot
20:20 Khisanth joined #parrot
20:27 Coke moritz: how many parrot devs are there?
20:28 moritz Coke: I have no idea :-)
20:36 sorear I suspect that generally, Parrot devs are more likely to be found at YAPC::NA
20:37 Coke historically that's true.
20:37 * Coke tries to get sorear to hack on parrot.
20:38 atrodo This year at YAPC::NA, there was not a lot of parrot people
20:39 Coke I blame my $dayjob for teasing me into thinking it'd be covered for too long.
20:49 sorear Coke: won't help, I'm a USian
20:54 Coke sorear: sorry, I don't understand the sequitor.
20:55 sorear Coke: me being a parrot dev would not fix "I suspect that ... at YAPC::NA"
20:56 sorear since I, too, will tend ot find NA easier than EU
20:57 bluescreen joined #parrot
21:01 Coke no, I just want you to hack on parrot so we have more parrot hackers.
21:01 Coke Conference and meeting attendance purely optional.
22:01 zby_home joined #parrot
22:25 ingy joined #parrot
22:37 mtk joined #parrot
23:43 benabik joined #parrot

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

Parrot | source cross referenced