Camelia, the Perl 6 bug

IRC log for #parrot, 2010-11-15

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 dngor left #parrot
00:02 dngor joined #parrot
00:10 ash_ left #parrot
00:12 dalek left #parrot
00:17 dukeleto tcurtis: are you doing a trial release?
00:18 GeJ running `fulltest` to make sure I didn't break anything. In the meantime...
00:18 nopaste "GeJ" at 192.168.1.3 pasted "RFC : Ignore ext/ subdir in Parrot::Test::Pod (see TT #1272)" (22 lines) at http://nopaste.snit.ch/25616
00:18 GeJ could someone tell me what he thinks about this patch?
00:18 tcurtis dukeleto: I have done one.
00:23 GeJ Oh yeah! Heisenbug!
00:24 tcurtis On a related note, how do I gain the appropriate access to the FTP server?
00:25 dukeleto tcurtis: i think you need to give your public ssh key to somebody important. I don't think it is me.
00:25 dukeleto tcurtis: Coke, allison or particle can probably do it
00:29 p6eval left #parrot
00:35 kid51 joined #parrot
00:36 GeJ If I ignore that Heisenbug in t/pmc/threads.t the nopasted patch makes codingstd PASS. Any objection?
00:37 kid51 Gej:  I don't see the need for this line:  +        my $IN_EXT_SUBDIR_RX = qr{^ext/};
00:37 GeJ to prevent recompiling the regexp for every file.
00:37 kid51 Why can't we just directly say:  if ($file =~ m/^ext/)
00:38 dukeleto GeJ: use qr{} directly in the if-statement
00:38 kid51 Then: if ($file =~ m/ext/o)
00:38 kid51 Or, as duke just suggested
00:39 kid51 $IN_EXT_SUBDIR_RX is only used once.  I don't see that it is any more clear than /^ext/
00:39 kid51 You might also want to include some inline comments re purpose of this exclusion (similar to following 'unless' blocks)
00:40 p6eval joined #parrot
00:40 kid51 (basic idea of patch is, however, sound)
00:47 kid51 make fulltest PASS on Linux/i386 at af8991f
00:51 kid51 make test PASS on Darwin/PPC
00:53 nopaste "GeJ" at 192.168.1.3 pasted "RFC : Ignore ext/ subdir in Parrot::Test::Pod (see TT #1272) -- take #2" (26 lines) at http://nopaste.snit.ch/25617
00:55 kid51 Gej:  If you are using the /o extension, you don't need the 'qr'
00:58 kid51 m{^qr/}o     should suffice
00:58 kid51 wrong
00:58 kid51 m{^ext/}o   should suffice
00:58 kid51 Gej:  On a different note, I attended the New York City BSD Conference this weekend.
01:12 kid51 left #parrot
01:22 p6eval left #parrot
01:28 GeJ pushed. Thanks.
01:30 dngor left #parrot
01:30 dngor joined #parrot
01:30 dukeleto kid51: qr has been around since 5.6, how long has /o been around?
01:32 p6eval joined #parrot
01:34 GeJ I've ran the generation of .pod_examinable* 10 times with and without the /o and it *seemed* to be faster with it. However I didn't include it in the commit per Jim's suggestion.
01:34 GeJ the difference is barely noticeable anyway (if ever).
01:35 GeJ s/ever/at all/ ?
01:35 GeJ boy my English stinks.
01:36 dngor_ joined #parrot
01:39 dngor left #parrot
01:57 ash_ joined #parrot
01:58 dukeleto GeJ: no worries, your english is pretty good :)
01:59 * dukeleto only speaks bad English
02:01 dukeleto https://github.com/blog/744-today-s-outage
02:02 dukeleto Classic "tests got run on production" bug. Classic.
02:03 dukeleto Worse, however, is that we may have lost some data from between the last good database backup and the time of the deletion. Newly created users and repositories are being restored, but pull request state changes and similar might be gone.
02:03 dukeleto Our plan moving forward:
02:03 dukeleto 1. Completely isolate the test environment from the production environment, i.e. make production hosts unreachable from the testing VM.
02:04 dukeleto good idea ;)
02:11 seanstickle left #parrot
02:13 whiteknight very good idea
02:33 dngor_ is now known as dngor
02:33 rurban_ joined #parrot
02:36 rurban left #parrot
02:36 rurban_ is now known as rurban
02:39 treed joined #parrot
02:39 treed left #parrot
02:50 whiteknight left #parrot
02:52 ash_ left #parrot
02:53 Andy joined #parrot
02:54 dukeleto sure is quiet in here.
02:54 Andy I could delete the repo again.
02:54 dukeleto are people getting the hang of git?
02:54 Andy Or github could crash for me instead.
02:54 dukeleto Andy: that could be fun.
02:55 Andy I'm waiting for the first time someone deletes a SHA1 that someone else has already.
03:01 seanstickle joined #parrot
03:03 dalek joined #parrot
03:04 * dukeleto needs to institute the no-force pushes on master policy
03:05 sorear dalek seems to have crashed secondary to the github fault
03:05 sorear the CREDITS parser fell into an infinite loop
03:16 * dukeleto puts on his lollerskates
03:29 seanstickle left #parrot
03:37 dukeleto nopaste?
03:37 dukeleto aloha, nopaste?
03:37 aloha dukeleto: nopaste is is http://nopaste.snit.ch (works with the script in $_PARROT/tools/dev/nopaste.pl)
03:48 cotto sorear, ...
03:48 cotto how...
03:48 sorear cotto: it cuts off lines until it gets one with -----
03:48 * cotto borrows dukeleto's lollerskates
03:48 sorear but the file was empty after the github database wipe
03:49 sorear Perl, DWIM to a fault, will (with a warning) let you chop off an empty line from a list with no items
03:49 sorear instead of throwing an exception, which would have been caught by the main loop :/
03:50 dukeleto Learning Is Fun!
03:50 cotto btw, I have more free time on the 5th of December than I thought.  We can bump the meeting back to 3:30 if nobody's got their heart set on 5:00.
03:51 cotto That still works for everyone who's filed in the doodle poll and will give mikehh a chance to participate for the first while.
03:58 cotto tcurtis, ping
03:58 tcurtis cotto: pong
03:59 cotto tcurtis, I heard you like ftp servers
04:00 cotto if you get me your public ssh key, I can give you access to the ftp server
04:03 cotto If you don't have one, they're pretty easy to generate.
04:07 cotto You can nopaste it safely.  The only thing people can use it for is to give you access to their machines.
04:07 nopaste "tcurtis" at 192.168.1.3 pasted "cotto, here you are" (1 line) at http://nopaste.snit.ch/25620
04:09 cotto tcurtis, there should be a username@host at the end of that.
04:11 tcurtis Ah, right. Is it problematic if it's a local hostname?
04:11 cotto nope
04:11 nopaste "tcurtis" at 192.168.1.3 pasted "cotto, this time" (1 line) at http://nopaste.snit.ch/25621
04:12 cotto try sshing to parrot@ftp-osl.osuosl.org
04:16 tcurtis It works! Now I just need to get the editor privileges on parrot.org and figure out what I need to update docs.parrot.org.
04:20 cotto I can't help you there.  For docs.parrot.org you need access to vm that hosts *.parrot.org.  Ask in #osuosl on freenode.  Alternately, you can have someone else do it.  It's not much work.
04:20 bluescreen left #parrot
04:27 plobsing joined #parrot
04:29 plobsing ~_~
04:30 cotto hi plobsing
04:31 cotto bacek_at_work, ping
04:33 dukeleto cotto: good evening
04:34 cotto hi dukeleto.  Are you back on normal time?
04:51 dukeleto cotto: for some definition of normal.
04:51 dukeleto cotto: wazzup?
04:51 * dukeleto is writing a blog post for GCI
05:17 dukeleto Very interesting wiki: http://wiki.pentaho.com/display/​BEEKEEPER/3.+The+Beekeeper+Model . The Parrot community has a lot to learn from that page.
05:20 * cotto loves link dumps
05:26 * simon_ loves analogies
05:27 simon_ also, by abusive extension of that analogy, RMS must be an anti-capitalist vegan? :D
05:29 dukeleto great article about parrot i haven't read before: http://www.wellho.net/solutions/perl-p​arrot-perl-s-new-virtual-machine.html
05:29 dukeleto old, but a nice overview for the time.
05:31 simon_ also, that type of honey with parts of the beehive wax still in it is like Ubuntu: it's essentially a product, but it smells of beehive (open source).
05:31 simon_ and when you eat it, you get chunks of wax in your mouth (quirks) and tell yourself it's a good thing.
05:32 dukeleto does TPF carry any legal responsibility for Parrot?
05:33 cxreg aloha: you're working on GC, right?
05:33 cxreg wait, are you a bot?
05:33 cxreg lol
05:34 cxreg /who fail :)  looks like i want bacek
05:36 dukeleto cxreg: lollerskates
05:36 * dukeleto hopes that bacek_at_work is working on a bot that can fix bugs in the GC
05:37 arnsholt joined #parrot
05:38 cxreg dukeleto: shows how closely i follow this channel
05:40 bacek_at_work ~~
05:40 cotto hi bacek_at_work
05:40 bacek_at_work cotto, hio
05:40 arnsholt_ left #parrot
05:47 dukeleto can someone take a look at this and see if it looks sane? http://socghop.appspot.com/gci/org/show​/google/gci2010/parrot_perl_foundations
05:51 cotto bacek_at_work, how likely is it that you can make a non-empty subset of the Parrot developer summit?
05:51 cotto s/make/attend/
05:51 bacek_at_work cotto, where is proposed schedule?
05:52 cotto http://www.doodle.com/q4nwdt9i5k929n2s
05:54 dukeleto bacek_at_work: it would be nice to hear from you there
05:55 bacek_at_work cotto, added. I prefer 7am-8:30am (just before departing to $dayjob)
05:57 bacek_at_work ookey. Second me is me.
05:57 cotto bacek_at_work, you can edit
05:57 cotto "Edit an entry"
05:57 * bacek_at_work had some schizophrenia moment.
05:58 cotto Sweet!  Dual baceks!
06:02 cxreg bacek_at_work: i was wondering if your gc work had any intention of moing the m&s bits out of the gc-ables to a lookaside
06:03 bacek_at_work cxreg, erm... Which bits?
06:03 cxreg the "mark" in particular, anything that causes something to be written to as part of the process
06:04 bacek_at_work cxreg, not really. Why?
06:04 cxreg copy on write fork
06:05 bacek_at_work COW is _slower_ in modern, mutlithreaded world
06:05 cxreg its not always about speed
06:05 dngor 5005 threads or bust!
06:05 dngor and/or bust.  more like coredump, actually.
06:05 bacek_at_work cxreg, we did remove COWed strings for speed purpose also
06:06 bacek_at_work anyway, decommute time
06:06 cxreg bacek_at_work: i dont mean anything implemented in parrot, i mean what the OS already does for you
06:06 cxreg it'd be nice to not immediately double real-ram when forking a parrot-based program
06:06 bacek_at_work cxreg, hmm. Interesting idea.
06:07 bacek_at_work But it will require redesign of whole "PObj" stuff.
06:09 cxreg i'll have to read up on that
06:09 theory left #parrot
06:22 preflex left #parrot
06:24 sorear more importantly
06:24 sorear each of our PMC headers fits into a cache line
06:25 preflex joined #parrot
06:25 sorear if we set a bit on each one, then we'll force the CPU to read and write a cache line for eahc PMC
06:25 sorear if we keep mark bits out of line, then we still have to read each PMC, but not write them, for half the memory cycles
06:25 sorear the mark bitmap is much smaller and can be resident entirely in L2
06:26 sorear this is, of course, theoretical; actual profiling is needed to make decisions
06:36 * jjore thought it was pretty standard design to do marks in look-aside.
06:37 jjore It's one of the things standard to anyone who likes forking processes anyway.
06:39 tcurtis left #parrot
06:42 Andy left #parrot
06:43 jjore for giggles, I added that feature to ruby-1.8.6 ages ago at http://diotalevi.isa-geek.net/~jo​sh/090410/0001-COW-nice-gc.patch just to avoid having my pre-forked procs go completely unshared merely because of GC.
06:43 sorear did it get accepted?
06:44 jjore though real users would go use something like REE (but I was enjoying writing it for myself)
06:44 sorear I beleive the Parrot GC is what is called "legacy code"
06:45 jjore it... consumed years of people's lives. Legacy? Bummer. Ok, how about plugging in something like boehm to avoid having to actually do all the work?
06:45 plobsing everything can be considered legacy code if you look into the future far enough
06:46 cotto we had a branch for using the boehm gc a while back.  istr that performance suffered
06:47 jjore shrug. I just figured you know, save the dev time for later? but er, whatever.
06:47 bacek jjore, I did plug boehm gc. It's slower than current one.
06:48 sorear jjore: it's legacy in the sense that nobody here likes it but -G inf and -G boehm don't work well enough
06:48 sorear what's your connection to the GC?  that nick is familiar
06:48 jjore I haven't touched the parrot gc at all.
06:49 jjore maybe you used something from http://search.cpan.org/~jjore/ ?
06:53 NotFound sorear: -G inf works nice, you just need a machine with infinite memory to fully profit from it.
06:55 bacek left #parrot
07:04 dukeleto sorear: why do you call the GC legacy?
07:04 dukeleto sorear: it has changed a lot recently
07:05 dukeleto sorear: what is your definition of "legacy" ?
07:05 dukeleto sorear: I call the IMCC legacy, but I wouldn't lump the GC into that category
07:08 sorear dukeleto: has gc_massacre been merged yet?
07:09 sorear I was under the impression nothing had changed in the trunk gc for years
07:12 dukeleto sorear: you are quite wrong about that. a bunch of gc branches have been merged in the last few months
07:12 dukeleto sorear: and it is "master" now :)
07:25 bacek joined #parrot
07:50 simcop2387_ joined #parrot
07:52 simcop2387 left #parrot
07:52 simcop2387_ is now known as simcop2387
08:16 ascent joined #parrot
08:21 dukeleto http://leto.net/perl/2010/11/parrot-foundat​ion-the-perl-foundation-google-code-in.html
08:26 TiMBuS left #parrot
08:37 bacek left #parrot
08:41 fperrad joined #parrot
08:50 contingencyplan left #parrot
09:05 bacek joined #parrot
09:13 TiMBuS joined #parrot
09:38 bacek seen nwellnhofer
09:38 aloha Sorry, I haven't seen nwellnhofer.
09:39 bacek left #parrot
09:39 bacek joined #parrot
10:06 dalek parrot: 34f29c2 | fperrad++ | / (2 files):
10:06 dalek parrot: [hello] clean & ignore more files
10:06 dalek parrot: review: https://github.com/parrot/parrot/commit/​34f29c2d02a1eb109d14e3cf78a7e6605edfddf0
10:22 plobsing left #parrot
10:34 rurban_ joined #parrot
10:37 rurban left #parrot
10:37 rurban_ is now known as rurban
10:51 dalek parrot: a8192ac | fperrad++ | config/gen/platform/win32/misc.h:
10:51 dalek parrot: [win32] check WINVER, see TT#1846
10:51 dalek parrot: review: https://github.com/parrot/parrot/commit/​a8192ac19b7f338117a1ac1c355219cb4a12ba5c
11:05 dalek parrot: 8bc48ad | bacek++ | .gitignore:
11:05 dalek parrot: Fix .gitignore
11:05 dalek parrot: review: https://github.com/parrot/parrot/commit/​8bc48ad4b852e3af5ba13fd7a7dbcd2abafc604e
12:30 bluescreen joined #parrot
12:56 mlb joined #parrot
13:10 silug left #parrot
13:27 whiteknight joined #parrot
13:27 whiteknight good morning, #parrot
13:42 GeJ Well, good morning indeed. :/
13:46 elmex left #parrot
13:51 bluescreen left #parrot
13:51 bluescreen joined #parrot
13:54 elmex joined #parrot
14:01 whiteknight GeJ: you seem less than excited about it
14:04 bluescreen left #parrot
14:15 bluescreen joined #parrot
14:33 dalek TT #1847 created by coke++: runtime/parrot/library/URI/Escape.pbc
14:34 dalek TT #1847: http://trac.parrot.org/parrot/ticket/1847
14:41 dalek parrot: f29109b | Coke++ | ports/macports/Portfile:
14:41 dalek parrot: pull in latest changes from
14:41 dalek parrot: http://trac.macports.org/browser/​trunk/dports/lang/parrot/Portfile
14:41 dalek parrot: (our macport users are now only 1 supported release behind)
14:41 dalek parrot: review: https://github.com/parrot/parrot/commit/​f29109bf805acfa02811e4455f5d7be868c84338
14:49 Andy joined #parrot
14:49 dalek TT #1415 closed by coke++: parrot macport for 2.0.0
14:49 dalek TT #1415: http://trac.parrot.org/parrot/ticket/1415
14:51 PacoLinux joined #parrot
14:53 dalek parrot: f0f8663 | fperrad++ | .gitignore:
14:53 dalek parrot: Clean & fix .gitignore
14:53 dalek parrot: review: https://github.com/parrot/parrot/commit/​f0f8663095b3a33c5b1e327af7a03c67e4d21d17
14:53 GeJ whiteknight: given that it's 1:49 AM here and I'm still writing an import proggy between a craptastic undocumented proprietary 4D application and a moronically designed database... yeah, not really thrilled about that "morning" thingy. :)
14:54 GeJ I'm too old to pull an all-nighter.
15:01 whiteknight GeJ: sounds....shitty. Good luck with all that
15:04 GeJ thanks.
15:05 dalek TT #1847 closed by fperrad++: runtime/parrot/library/URI/Escape.pbc
15:05 dalek TT #1847: http://trac.parrot.org/parrot/ticket/1847
15:16 silug joined #parrot
15:18 Andy dukeleto: Is Code-In really a contest?
15:19 whiteknight Andy: in a sense. The students are competing with each other for tasks
15:20 Andy ok
15:34 whiteknight I can't configure Parrot on a system that does not have git installed
15:34 whiteknight config fails at auto::sha1 and auto::git_describe
15:41 contingencyplan joined #parrot
16:05 theory joined #parrot
16:07 hercynium joined #parrot
16:14 atrodo left #parrot
16:15 dukeleto 'ello
16:16 atrodo joined #parrot
16:16 dukeleto whiteknight: hmmm, I should fix that
16:16 dukeleto Andy: yes, Code-In is really a contest
16:16 Andy ok
16:16 Andy it's all posted
16:17 whiteknight dukeleto: nevermind, I can build. It generates the makefile
16:17 whiteknight it just spits out some really ugly error messages when a git client can't be found
16:18 dukeleto whiteknight: ok, i will try to clean that up
16:18 whiteknight I'm on Win64 right now. I'm using msysgit to do git stuff, but I'm in the normal windows cmd prompt to build it
16:18 whiteknight so there is a .git directory, but no git executable visible in PATH
16:19 whiteknight (I know this is unnecessarily complicated)
16:23 dukeleto whiteknight: i see. it should be trivial to not puke out error messages
16:24 bluescreen left #parrot
16:25 cotto_work ~hello~
16:25 * dukeleto jumps into a cup of coffee
16:33 whiteknight kid51: ping
16:35 bluescreen joined #parrot
16:37 dalek parrot: c414af3 | unknown++ | config/gen/platform/win32/misc.h:
16:37 dalek parrot: fix build on windows. I don't think there is such a thing as 'w32api.h'. It would be 'win32api.h' or, what I used, 'windows.h'
16:37 dalek parrot: review: https://github.com/parrot/parrot/commit/​c414af3416d60d67e99c75d184b1c8185765df69
16:48 whiteknight damnit. unknown is stealin my karmas
16:51 jan left #parrot
16:52 dmalcolm joined #parrot
17:17 bluescreen where is headerizer? and is it automagically run by "make" ?
17:18 cotto_work bluescreen: you need to run make headerizer manually
17:19 bluescreen where is it?
17:19 cotto_work if you've just changed one file, it's quicker to run it directly via perl tools/dev/headerizer.pl src/foo.o
17:19 cotto_work make headerizer will also work
17:19 bluescreen ok
17:20 cotto_work (and yes, you pass the name of the object file, not the c file)
17:20 bluescreen why's thaT?
17:22 cotto_work That's what the magic requires.
17:23 cotto_work Andy wrote it.  I'm sure he had a sensible reason.
17:23 Andy ?
17:23 Andy No, it is not run by make.
17:23 cotto_work It is if you use make headerizer
17:24 cotto_work but no, not by default
17:24 Andy Well, yes.
17:25 bluescreen why does it run over the object file instead of the c file? because of the macro expansion?
17:42 whiteknight because it is typically called by the makefile, and the makefile has a list of .o files not .c files
17:44 dalek parrot: 2d9999d | fperrad++ | config/gen/platform/win32/misc.h:
17:44 dalek parrot: [win32] w32api.h is only with MinGW
17:44 dalek parrot: review: https://github.com/parrot/parrot/commit/​2d9999d291ea2aed23833f50e8cd42ebdd0b22cd
17:45 whiteknight ah, that's the detail I was missing
17:46 whiteknight I couldn't figure out why w32api.h was in there in the first place. But it's an MinGW thing
17:46 whiteknight fperrad++
17:48 fperrad whiteknight, have you seen that you commit c414af3 as 'unknown' ?
17:49 cotto_work It's probably a temporary problem while github restores their db.
17:51 dukeleto cotto_work: not so sure. he is commiting from a machine without a properly setup ~/.gitconfig
17:51 cotto_work also a possibility
17:57 whiteknight fperrad: yeah, I didn't have my git config set up on this box. Thanks
17:59 whiteknight LAPACK came out with a standard C API. That will make my integration work with PLA much easier when the time comes
18:03 arnsholt left #parrot
18:33 rurban_ joined #parrot
18:37 plobsing joined #parrot
18:37 rurban left #parrot
18:37 rurban_ is now known as rurban
18:41 mlb left #parrot
18:48 dalek winxed: r681 | NotFound++ | trunk/winxedst0.cpp:
18:48 dalek winxed: rename efirst and esecond members to lexpr and rexpr in stage 0 for consistency
18:48 dalek winxed: with stage 1
18:48 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=681
18:54 arnsholt joined #parrot
19:06 hercynium_ joined #parrot
19:09 dalek winxed: r682 | NotFound++ | trunk/t/harness:
19:09 dalek winxed: make harness a bit more idiomatic
19:09 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=682
19:12 hercynium left #parrot
19:12 hercynium_ is now known as hercynium
19:34 seanstickle joined #parrot
19:34 theory left #parrot
19:35 theory joined #parrot
19:35 seanstickle left #parrot
19:35 seanstickle joined #parrot
19:36 nomad joined #parrot
19:36 seanstickle left #parrot
19:45 hercynium left #parrot
19:46 hercynium joined #parrot
20:14 whiteknight I want everybody to submit a smolder report before the release goes out
20:15 cotto_work deal
20:17 NotFound Aye, sir!
20:19 whiteknight :)
20:20 whiteknight the top 15 most recent reports are all on either windows or solaris. My linux peeps need to represent
20:20 NotFound Doing ubuntu i386 and debian amd64
20:21 seanstickle joined #parrot
20:22 NotFound Solaris? The oracle has talked.
20:23 atrodo oracle--
20:25 NotFound There is joke in spain: oracle<=>elcaro (el caro == the expensive)
20:25 atrodo That is amazing and accurate
20:26 NotFound Usually not so accurate, because people that said it haven't paid the licenses X-)
20:27 NotFound Disclaimer: I deny everything!
20:32 whiteknight fperrad: ping
20:39 zby_home joined #parrot
20:39 whiteknight msg fperrad http://smolder.parrot.org/app​/projects/report_details/1085 <-- Is there something special about this report? With the release tomorrow I want to know if we have bugs to fix. Most other reports from that platform are clean
20:39 aloha OK. I'll deliver the message.
20:41 particle left #parrot
20:44 particle joined #parrot
20:55 whiteknight atrodo: T3H FA1LZ
20:57 whiteknight atrodo: do you have any local mods or anything? I'm trying to figure out why t/pmc/threads.t would fail on your system
20:59 jan joined #parrot
21:06 GeJ Oh, I haz the t/pmc/threads.t heisenbug!
21:06 GeJ Well, I have one. Not sure if it's the same.
21:07 whiteknight great
21:09 GeJ ok, not the same. Mine coredumps saying that keyed_int vtable not implemented in ExceptionHandler... or something along those lines.
21:09 GeJ As I said, Heisenbug. I can run make fulltest 10 times and not see the thing explode.
21:11 dukeleto GeJ: the test probably has a race condition in it
21:12 GeJ from the expected output, I'd say it's probably L366: "CLONE_CODE | CLONE_GLOBALS"
21:12 GeJ (if memory serves)
21:12 simcop2387 left #parrot
21:12 simcop2387_ joined #parrot
21:12 simcop2387_ is now known as simcop2387
21:12 whiteknight we should skip that test for the release if it's a problem
21:12 GeJ but I slept 30 minutes last night, so feel free to dismiss everything I might say in the next 12 hours.
21:13 * whiteknight thinks we should skip at least one threads test for each release, until there are none left
21:13 whiteknight the threading implementation is almost as bad as the tests that attempt to exercise it
21:14 moritz what happened to the gsoc branch?
21:14 * GeJ thinks that since everybody and their broother think that FreeBSD's thread library is weird, I second whiteknight's proposal.
21:15 moritz GeJ: even my wife agrees :-)
21:15 whiteknight moritz: Added greenthreads, I think, but was only really working on linux
21:16 whiteknight I need to go through that code again eventually and see if there is anything in there that can be salvaged
21:17 whiteknight I suspect the greenthreads implementation can be used with minor modification. I don't know if we will be able to use the non-blocking IO parts of it though
21:18 GeJ moritz: I lack the technical knowledge to have a reasonable opinion on the subject. I'll just say that from a user point of view, I'm having less and less problems with threaded programs as I upgrade my system.
21:19 GeJ that doesn't mean that I'm not willing to trust your wife's opinion.
21:19 moritz GeJ: I've never even installed FreeBSD. Neither did my wife :-)
21:19 moritz she just didn't want to contradict when I asked her :-)
21:20 dngor I just installed FBSD on my old G4 powerbook.
21:20 whiteknight I won't use FreeBSD until it starts working in VirtualBox
21:20 whiteknight I don't have any hardware that I'm willing to do the install on
21:20 nomad left #parrot
21:23 perlite_ joined #parrot
21:25 seanstickle left #parrot
21:26 perlite left #parrot
21:26 simcop2387 left #parrot
21:26 perlite_ is now known as perlite
21:28 simcop2387 joined #parrot
21:28 GeJ okay, back to work.
21:28 * GeJ &
21:29 * whiteknight is leaving work. later
21:29 whiteknight left #parrot
21:40 bluescreen left #parrot
21:40 theory left #parrot
21:44 Andy left #parrot
21:45 fperrad left #parrot
21:50 Andy joined #parrot
21:52 Andy left #parrot
21:55 Andy joined #parrot
21:59 dalek TT #1848 created by coke++: delete placeholder t/tools/parrot_debugger.t
21:59 dalek TT #1848: http://trac.parrot.org/parrot/ticket/1848
22:01 Tene I'm been considering trying to work on tt#757 sometime.  I'm still grumpy about it.
22:02 theory joined #parrot
22:05 bluescreen joined #parrot
22:12 bluescreen left #parrot
22:17 theory_ joined #parrot
22:17 theory left #parrot
22:17 theory_ is now known as theory
22:19 theory_ joined #parrot
22:19 theory left #parrot
22:19 theory_ is now known as theory
22:21 zby_home left #parrot
22:22 simcop2387 left #parrot
22:29 simcop2387 joined #parrot
22:30 dalek parrot: ae3500c | GeJ++ | MANIFEST.SKIP:
22:30 dalek parrot: Regenerate MANIFEST.SKIP
22:30 dalek parrot: review: https://github.com/parrot/parrot/commit/​ae3500cef7cb03b6210cfa61eb1b1d95871ca1a8
22:31 dngor_ joined #parrot
22:31 dngor left #parrot
22:37 donaldh joined #parrot
22:37 dngor_ is now known as dngor
22:37 theory left #parrot
22:40 atrodo aloha msg whiteknight I'm not sure what the deal was, but when I ran smoke again, everything passed.
22:40 aloha atrodo: OK. I'll deliver the message.
22:41 dalek winxed: r683 | NotFound++ | trunk/winxedst1.winxed:
22:41 dalek winxed: simplify a few string operations
22:41 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=683
22:57 whiteknight joined #parrot
23:03 * whiteknight has started his "testall" script. this should give us a few smoke reports to mull through
23:04 whiteknight anybody else seeing the heisenbug in the thread tests?
23:04 gert joined #parrot
23:06 whiteknight atrodo++ # Thanks for the re-test
23:06 dngor_ joined #parrot
23:06 dngor left #parrot
23:07 dngor_ is now known as dngor
23:08 cotto_work whiteknight: haven't we had a thread-related heisenbug for a long time?
23:14 whiteknight probably. All the more reason to crumple our current thread system and all it's tests up into a little ball and throw it in the trash can
23:14 GeJ whiteknight: didn't exactly smoke. But I ran a 3 or 4 fulltests this morning and all passed.
23:15 whiteknight GeJ: that's good too
23:15 whiteknight all tests are always appreciated
23:22 theory joined #parrot
23:32 * dukeleto smokes master
23:36 bluescreen joined #parrot
23:39 dukeleto PASS http://smolder.parrot.org/app​/projects/report_details/1098
23:39 * dukeleto smokes an optimized build
23:44 dngor_ joined #parrot
23:44 dalek winxed: r684 | NotFound++ | trunk/winxedst1.winxed:
23:44 dalek winxed: refactor Token class to a bunch of derived classes in stage 1
23:44 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=684
23:44 dngor left #parrot
23:44 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#1097) fulltest) at ae3500c - Ubuntu 10.10 i386 (g++-4.5 with --optimize)
23:45 dukeleto PASS (optimized) http://smolder.parrot.org/app​/projects/report_details/1099
23:48 dngor_ is now known as dngor
23:54 cotto_work anyone know what time tcurtis++ is planning on cutting the release?
23:58 mikehh ok how do you build rakudo with the latest parrot release - it reports Parrot revision RELEASE_2_9_1-261-g495feaf required (currently RELEASE_2_9_1-273-gae3500c)
23:58 mikehh AFAICS 273 is greater than 261

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

Parrot | source cross referenced