Camelia, the Perl 6 bug

IRC log for #parrot, 2011-01-11

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 nwellnhof cotto: a missing generic platform function will be detected really early.
00:00 cotto_work ok.  That's what I care about.
00:01 dalek parrot/nwellnhof/platform_src: dc42962 | nwellnhof++ | config/gen/makefiles/root.in:
00:01 dalek parrot/nwellnhof/platform_src: Merge branch 'master' into nwellnhof/platform_src
00:01 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/dc42962b4a
00:03 nwellnhof kid51: i think we should skip that OS.readdir test for now. non-ascii filenames have never really worked with parrot. but i'm working on that.
00:04 dalek parrot: 8ff0b70 | cotto++ | t/tools/pbc_dump.t:
00:04 dalek parrot: Merge branch 'gci_pbc_dump2' of https://github.com/fernandobrito/parrot into fernandobrito-gci_pbc_dump2
00:04 dalek parrot: review: https://github.com/parrot/parrot/commit/8ff0b705c5
00:06 dalek parrot: 04c21ee | cotto++ | / (2 files):
00:06 dalek parrot: Merge branch 'master' of https://github.com/nol888/parrot into nol888-master
00:07 dalek parrot: review: https://github.com/parrot/parrot/commit/04c21ee323
00:07 cotto_work seen mat221
00:07 aloha Sorry, I haven't seen mat221.
00:07 cotto_work seen matt221
00:07 aloha matt221 was last seen in #parrot 20 hours 46 mins ago saying "cotto: thanks!".
00:07 kid51 nwellnhof: Is there a TT open on that issue yet?
00:08 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#2182) fulltest) at 2_11_0-928-g24280fc - Ubuntu 10.10 i386 (g++-4.5)
00:09 nwellnhof kid51: #1898 and #1836
00:12 dalek parrot: 48cb7c7 | (Matt Rajca)++ | t/pmc/complex.t:
00:12 dalek parrot: Increased coverage for Complex PMC
00:12 dalek parrot: review: https://github.com/parrot/parrot/commit/48cb7c76e6
00:12 dalek parrot: aaed459 | cotto++ | t/pmc/complex.t:
00:12 dalek parrot: [t] use a really incorrect PMC to misinitialize a Complex
00:12 dalek parrot: review: https://github.com/parrot/parrot/commit/aaed459065
00:14 dalek parrot: 05b2f96 | jkeenan++ | t/dynpmc/os.t:
00:14 dalek parrot: TODO test of non-ASCII filenames due to inconsistent results.
00:14 dalek parrot: review: https://github.com/parrot/parrot/commit/05b2f968fe
00:14 dalek parrot: 0cdd5b4 | jkeenan++ | / (4 files):
00:14 dalek parrot: Merge branch 'master' of git@github.com:parrot/parrot
00:14 dalek parrot: review: https://github.com/parrot/parrot/commit/0cdd5b4123
00:17 plobsing joined #parrot
00:22 janus left #parrot
00:22 janus joined #parrot
00:27 cotto_work kapace_: go ahead and nuke pack and unpack from pfbcs in your branch and I'll pull/merge.  If they're needed, we can revert.
00:27 kapace_ ok
00:30 cotto_work lmk when ready
00:31 dalek parrot: 3f4a9c6 | jkeenan++ | / (2 files):
00:31 dalek parrot: On Darwin, use installed dir, not build dir, for libparrot_soname.
00:31 dalek parrot:
00:31 dalek parrot: Apply patch submitted in http://trac.parrot.org/parrot/ticket/1890.
00:31 dalek parrot: This should also fix http://trac.parrot.org/parrot/ticket/344.
00:31 dalek parrot: review: https://github.com/parrot/parrot/commit/3f4a9c6f02
00:31 bacek_at_work ~~
00:32 cotto_work bacek_at_work: ohai.  any reason to keep pack/unpack in pfbcs?
00:32 kapace_ cotto_work, ok just building..
00:32 bacek_at_work kapace_, what about become parrot's commiter? Just send CLA and that's it.
00:32 bacek_at_work cotto_work, it was in original pdd. I would like to keep it. Just to have explicit call from NQP.
00:33 kapace_ bacek_at_work, ok I might do it... where do I get CLA from?
00:33 bacek_at_work aloha, cla?
00:33 aloha bacek_at_work: cla is http://www.parrot.org/files/parrot_cla.pdf
00:33 cotto_work bacek_at_work: it's inherited from pfs
00:33 kapace_ bacek_at_work, 404
00:33 cotto_work essentially the same implementation
00:33 bacek_at_work cotto_work, ah. Sorry, misunderstood. Nuke it. We need only Packfile.pack and .unpack. Not for segments.
00:34 cotto_work aloha: cla?
00:34 aloha cotto_work: cla is http://www.parrot.org/files/parrot_cla.pdf
00:34 bacek_at_work cotto_work, 404 for cla.
00:34 cotto_work indeed
00:34 cotto_work aloha: forget cla
00:34 aloha cotto_work: I forgot about cla.
00:35 * kapace_ brb, make testing...
00:35 cotto_work aloha: cla is http://www.parrot.org/sites/www.​parrot.org/files/parrot_cla.pdf
00:35 aloha cotto_work: Okay.
00:37 cotto_work bacek_at_work: so packfilesegment.pack and unpack can go away too?
00:43 dalek parrot: 940a3a1 | jkeenan++ | / (2 files):
00:43 dalek parrot: Merge branch 'master' of git@github.com:parrot/parrot
00:43 dalek parrot: review: https://github.com/parrot/parrot/commit/940a3a1639
00:43 dalek parrot: 2e35095 | jkeenan++ | lib/Parrot/Install.pm:
00:43 dalek parrot: Why use an 'eval' if you're not going to look at $@?  This eliminates a 'leaving eval via next' warning.
00:43 dalek parrot: review: https://github.com/parrot/parrot/commit/2e35095313
00:50 janus left #parrot
00:50 janus joined #parrot
00:51 dalek TT #1890 closed by jkeenan++: installed parrot requires build dir.
00:51 dalek TT #1890: http://trac.parrot.org/parrot/ticket/1890
00:51 dalek TT #344 closed by jkeenan++: Can't create working installed parrot for macports
00:51 dalek TT #344: http://trac.parrot.org/parrot/ticket/344
00:51 dalek parrot: 083132a | mikehh++ | t/tools/mk_language_shell.t:
00:51 dalek parrot: fix codetest failure - trailing whitespace
00:51 dalek parrot: review: https://github.com/parrot/parrot/commit/083132ada8
00:52 fbrito left #parrot
00:54 kapace_ cotto_work, ok, should be good, all in the same pull https://github.com/parrot/parrot/pull/120
00:55 kid51 is now known as kid51_at_dinner
00:58 cotto_work kapace_: ok.  I'll merge once everything passes.
01:00 plobsing that time approaches. fulltesting a fix for t/dynpmc/os.t
01:03 cotto_work kapace_: done
01:03 cotto_work thanks!
01:03 dalek parrot: 9005a63 | kapace++ | t/pmc/packfilebytecodesegment.t:
01:03 dalek parrot: Update pfbs tests for removal of pack methods
01:03 dalek parrot: review: https://github.com/parrot/parrot/commit/9005a636fd
01:03 dalek parrot: 0d31432 | cotto++ | / (2 files):
01:03 dalek parrot: Merge branch 'kapace/test_pbcs' of https://github.com/kapace/parrot into kapace-kapace/test_pbcs
01:03 dalek parrot: review: https://github.com/parrot/parrot/commit/0d31432651
01:03 kapace_ :)
01:03 dalek parrot: f89f985 | cotto++ | / (2 files):
01:03 dalek parrot: Merge branch 'kapace-kapace/test_pbcs'
01:03 dalek parrot: review: https://github.com/parrot/parrot/commit/f89f985096
01:04 cotto_work kapace++ because you missed some karma there
01:04 kapace_ hmm, i don't like the sound of this: http://www.google-melange.com/gci/​task/show/google/gci2010/parrot_pe​rl_foundations/t129452250650#c1001
01:05 cotto_work kapace_: which task is that?
01:05 kapace_ that was the last one I did, last minute
01:05 cotto_work bit ops coverage?
01:06 kapace_ increase pfbs coverage
01:06 mikehh t/src/checkdepend.t .... # couldn't find extra_nci_thunks.str, included from src/extra_nci_thunks.c
01:06 mikehh # couldn't find /usr/include/malloc.h, included from src/gc/malloc.c
01:06 mikehh # couldn't find glut_nci_thunks.str, included from src/glut_nci_thunks.c
01:06 mikehh t/src/checkdepend.t .... ok
01:06 cotto_work wacky.  I'm pretty sure I approved that one.
01:06 dalek parrot: adbfc90 | plobsing++ | src/string/api.c:
01:06 dalek parrot: catch failed encoding of platform string and fallback on binary strings
01:06 dalek parrot: review: https://github.com/parrot/parrot/commit/adbfc90af6
01:07 plobsing NotFound: adbfc90 should fix your LANG=C issues
01:07 kapace_ i'm sure google'l figure it out later, not too worried (doesn't make much difference anyways)
01:07 cotto_work You'll still get some money out of the deal.
01:08 bacek_at_work cotto_work, yes, PFS.pack/unpack can go
01:09 cotto_work pfs doesn't do a whole lot
01:10 cotto_work done
01:10 dalek parrot: 90418f5 | cotto++ | src/pmc/packfilesegment.pmc:
01:10 dalek parrot: [pmc] remove unneeded pack/unpack from PackfileSegment
01:10 dalek parrot: review: https://github.com/parrot/parrot/commit/90418f5c8c
01:11 bacek_at_work cotto_work, you forgot about other PF*S pmcs.
01:12 bacek_at_work eg PFBS
01:18 cotto_work ah.  brain slowing.  will remove shortly
01:19 cotto_work looks like there aren't any others left apart from packfile
01:19 cotto_work pfbcs was removed by kapace++
01:20 whiteknight joined #parrot
01:31 Kristaba left #parrot
01:32 whiteknight good evening, #parrot
01:34 cotto_work hi whiteknight
01:40 whiteknight hello cotto_work, how are you this evening?
01:40 dalek parrot/nwellnhof/unicode_filenames: cfd716e | nwellnhof++ | compilers/imcc/ (5 files):
01:40 dalek parrot/nwellnhof/unicode_filenames: [imcc] Undo initial approach to slurp input files at once
01:40 dalek parrot/nwellnhof/unicode_filenames:
01:40 dalek parrot/nwellnhof/unicode_filenames: It's much easier to define our own YY_INPUT
01:40 dalek parrot/nwellnhof/unicode_filenames: review: https://github.com/parrot/parrot/commit/cfd716ef85
01:40 cotto_work done with gci and ready to rock something else
01:41 cotto_work nom, most likely
01:41 whiteknight nom definitely needs some love
01:41 cotto_work It's getting quite a bit.  jnthn has been committing a lot lately.
01:42 whiteknight yeah, it's good to see him back in the saddle. It always worries me when such prolific committers get sick or otherwise busy
01:48 whiteknight It's good seeing bacek producing so much good work too. I love waking up to two dozen of his commits
01:49 Coke cotto_work: http://www.parrot.org/sites/www.​parrot.org/files/parrot_cla.pdf !? that URL is borked.
01:50 Coke ... but, perhaps our drupal config just sucks.
01:51 Coke ah well.
01:58 fbrito joined #parrot
01:59 dukeleto ~~
01:59 dukeleto Coke: we had a drupal upgrade that messed some stuff of. I thought it was fixed.
02:00 dukeleto Coke: which URL links to that ?
02:03 gbacon joined #parrot
02:12 dalek parrot: 54364ee | plobsing++ | src/string/api.c:
02:12 dalek parrot: use proper exception handling mechanism in stead of abusing embed API
02:12 dalek parrot:
02:12 dalek parrot: now platform strings work with exception handlers in place
02:12 dalek parrot: review: https://github.com/parrot/parrot/commit/54364eee36
02:13 plobsing good thing I second guessed my fix.
02:19 Coke dukeleto: foundation, docs.
02:22 cotto ~~
02:26 nwellnhof plobsing: i thought about adding validating functions to the string code, so you don't have to catch exceptions.
02:26 plobsing nwellnhof: that shouldn't be a hot path. if it is, your machine is configured wrong.
02:27 nwellnhof yes, it probably isn't worth the trouble.
02:28 dalek parrot/nwellnhof/unicode_filenames: ee89c8c | nwellnhof++ | / (7 files):
02:28 dalek parrot/nwellnhof/unicode_filenames: Use platform_cstring functions for filenames
02:28 dalek parrot/nwellnhof/unicode_filenames: review: https://github.com/parrot/parrot/commit/ee89c8c779
02:32 nopaste "nwellnhof" at 192.168.1.3 pasted "imcc and unicode filenames" (7 lines) at http://nopaste.snit.ch/27676
02:33 plobsing yay unicode
02:34 nwellnhof the funny thing is that we can't delete etc. unicode files until we rewrite the OS dynpmc.
02:35 nwellnhof and we can't use perl for tests because it doesn't support cross-platform unicode filenames.
02:39 plobsing I'd like to say "it shouldn't be too hard", but my recent experiences say different.
02:39 nwellnhof it's not that hard, but i like to do it in a clean way and not with all that #ifdefs.
02:40 nwellnhof that's why i started the platform_src branch
02:45 plobsing while the strategy employed by config/gen/platform and now src/platform is very frequently a good approach, I don't think it is universally applicable to platform portability issues.
02:45 plobsing checkout PARROT_HAS_BSD_STAT_EXTN for where I had to use #ifdef because *nearly* every platform extends the baseline functionality in exactly the same way
02:46 plobsing having a separate, identical file fore every system which implemented the bsd stat extensions would have resulted in much duplication
02:51 nwellnhof the stat code sould also be moved to src/platform. having the ifdefs there isn't that bad.
02:52 mikehh The header in the .gitignore file mentions being generated by tool/dev/mk_manifest_and_skip.pl, surely thisis a holdover from the svn days
02:52 nwellnhof ah, it is in pltform already
02:53 mikehh in fact Parrot::Manifest still has the code there but I don't think it is used anymore
02:53 plobsing I do in general agree with the notion that #ifdefs should be avoided where possible. ifdef soup is not fun.
02:54 nwellnhof ifdefs in the platform-specific code are ok. but they shouldn't be needed in the rest of parrot.
02:54 cotto mikehh, are you up for a codectomy?
02:55 kid51_at_dinner is now known as kid51
02:55 mikehh cotto: I'm listening :-}
02:56 * kid51 whispers "html_cleanup merge"
02:56 mikehh kid51: yeah - will work on it now
02:56 plobsing do it!
02:56 * cotto whispers "what kid51 said"
02:57 kid51 unless you want me to do it while you do something else
02:58 kid51 hmm, *I* am working on something else
03:03 whiteknight left #parrot
03:05 ppant joined #parrot
03:06 dalek parrot/remove_auto_jit: 7ec45de | jkeenan++ | / (15 files):
03:06 dalek parrot/remove_auto_jit: Remove config step auto::jit and all its sequelae.
03:06 dalek parrot/remove_auto_jit: review: https://github.com/parrot/parrot/commit/7ec45de643
03:13 dalek Heuristic branch merge: pushed 62 commits to parrot/html_cleanup by mikehh
03:25 dalek parrot: 4352c13 | mikehh++ | / (62 files):
03:25 dalek parrot: [html_cleanup] merge in origin/master
03:25 dalek parrot: review: https://github.com/parrot/parrot/commit/4352c132a4
03:26 * kid51 hears a vast sigh of relief blowing westward across the Atlantic
03:29 fbrito Have you guys ever heard bout code_swarm?
03:29 fbrito http://www.vimeo.com/1093745
03:35 plobsing fbrito: I'd seen it before, but never thought of using it on parrot. now I'm curious
03:36 fbrito if you want I can take a look on it and see how it works
03:36 dalek parrot: 7d70de1 | jkeenan++ | t/pmc/filehandle.t:
03:36 dalek parrot: Remove duplicate 'use' statement.
03:36 dalek parrot: review: https://github.com/parrot/parrot/commit/7d70de192c
03:36 dalek parrot: aef023f | jkeenan++ | / (34 files):
03:36 dalek parrot: Merge branch 'master' of git@github.com:parrot/parrot
03:36 dalek parrot: review: https://github.com/parrot/parrot/commit/aef023f53c
03:40 fbrito plobsing: I am going to bed now, but I can take a look on code_swarm tomorrow to see how it works
03:40 cotto 'night fbrito
03:40 cotto dukeleto, ping
03:42 Coke so, I want to time all the makefile steps. I cannot think of a cleaner way than prepending every rule with $(TIMER) and then having that be conditionally =time
03:42 Coke any ideas?
03:43 kid51 left #parrot
03:46 fbrito left #parrot
03:51 cotto Coke, it really does seem like there'd be some better way to do that.
03:55 dalek parrot: f63d76e | mikehh++ | docs/index/ (2 files):
03:55 dalek parrot: change some case sensitive references - (the commit for this seemed to have got lost somewhere)
03:55 dalek parrot: review: https://github.com/parrot/parrot/commit/f63d76e657
03:56 kurahaupo1 joined #parrot
04:02 Coke cotto: yah, i was hoping for a make flag.
04:02 Coke basically want to find things to target to reduce build time.
04:03 plobsing you could use strace/dtrace and try to work out what is being done when
04:03 plobsing but that could be a lot of work
04:04 Coke meh. for now, just picking random .pl things and throwing NYTprof at them.
04:05 Coke ooh, that might help.
04:13 ppant left #parrot
04:20 cotto Coke++ for trying to profile the build
04:26 cotto mikehh++ for getting html_cleanup merged.  Nice work.
04:35 mikehh cotto: still need to clean up a lot of the cruft left, will probably create a branh html_cleanup_2 for that, but tomorrow
04:36 cotto what kind of cruft?
04:36 cotto either way, the only complaint I have is that you didn't use "massacre" in the branch name.
04:38 mikehh there are a lot of leftovers from the previous incarnation in lib/Parrot/Docs we don't need but are referenced in places eg make chm or something
04:39 cotto Ripping out old unused stuff is fun.
04:39 plobsing msg fbrito if you are still interested in codeswarm, there are some good instructions at http://sams-rants.blogspot.com/2010/02/co​deswarm-01-on-os-x-106-with-git-repo.html
04:39 aloha OK. I'll deliver the message.
04:39 mikehh cotto: aint completed the mass-acre yet
04:39 cotto codeswarm for Parrot would be interesting
04:40 mikehh or should it be mass-a-cre
04:40 Coke in mikehh's defense, I named that branch.
04:40 Coke mikehh++ for getting something merged back to trunk!
04:40 Coke (esp. something usable! ;)
04:41 cotto very mikehh++
04:41 cotto no longer need mortals fear to look into how our html docs are generated
04:43 mikehh AFAIK I have clecked the links in make html, some of you might want to check further, also the refs in the index files are now taken directly from the pod so some changes might be needed there
04:44 mikehh or should I say the names in the index/header files
04:44 cotto That should definitely be done before the release.
04:48 Coke we can do a make html test run and put it on docs under TESTME or something.
04:49 Coke s/we/you/ ;)
04:49 Coke ah, think I can drop extract_delimited from c2str.pl
04:54 Coke is t/pmc/exception-old.t failing for anyone on master?
04:55 plobsing not here
04:55 plobsing what are you getting?
04:56 ppant joined #parrot
04:56 nopaste "coke" at 192.168.1.3 pasted "boom" (6 lines) at http://nopaste.snit.ch/27677
04:56 Coke I do have a local mod.
05:14 dalek TT #1945 created by dukeleto++: A parrot with all external libraries disabled fails t/src/atomic.t
05:14 dalek TT #1945: http://trac.parrot.org/parrot/ticket/1945
05:14 dalek TT #1946 created by coke++: Smolder parrot feed incorrect URLS.
05:14 dalek TT #1946: http://trac.parrot.org/parrot/ticket/1946
05:16 plobsing msg fbrito you might also want to have a look at gource, which I find more understandable (gource.googlecode.com)
05:16 aloha OK. I'll deliver the message.
05:24 Coke urk. hard to verify that our generated all_cstring.str is unchanged.
05:25 nopaste "coke" at 192.168.1.3 pasted "this is faster, but causes t/pmc/exception-old.t to fail. Any clues?" (23 lines) at http://nopaste.snit.ch/27678
05:38 kurahaupo1 left #parrot
05:44 theory left #parrot
05:44 theory joined #parrot
05:45 kurahaupo_mobi joined #parrot
05:45 kurahaupo left #parrot
05:46 rurban_ joined #parrot
05:46 Coke weird - my change is apparently impacting Exception's CONST_STRING for Hash, and nothing else. no other tests fail.
05:48 rurban left #parrot
05:48 rurban_ is now known as rurban
05:48 * cotto attempts to repro
05:49 Coke AHAH!
05:49 Coke that is the one case in the REPO where we use CONST_STRING('') instead of CONST_STRING("")
05:49 cotto sneaky us
05:51 cotto yup.  one place
05:52 Coke yay, one more incredibly incremental build speedup.
05:53 cotto how incredibly incremental?
05:53 Coke momentito. retesting to make sure it all works.
05:53 kurahaupo_mobi left #parrot
05:55 kurahaupo joined #parrot
05:59 Coke ok. with my fix, c2str for scalar PMC takes ~88ms - without it, takes 113ms - basically replacing Text::Balanced with a regexp. (avoids the startup cost of loading the module)
05:59 Coke aloha, 113-88
05:59 aloha Coke: 25
05:59 Coke so, 25ms times # of .str files. yay, hardly worth it!
06:03 dalek parrot: 633cd46 | Coke++ | tools/build/c2str.pl:
06:03 dalek parrot: Remove unused 'use lib'
06:03 dalek parrot: (took 2.55/35.4ms in one invocation)
06:03 dalek parrot: review: https://github.com/parrot/parrot/commit/633cd461e6
06:03 dalek parrot: e4c1406 | Coke++ | src/pmc/exception.pmc:
06:03 dalek parrot: use " on CONST_STRING, not '
06:03 dalek parrot: review: https://github.com/parrot/parrot/commit/e4c14060f8
06:03 dalek parrot: 4380854 | Coke++ | tools/build/c2str.pl:
06:03 dalek parrot: avoid Text::Balanced; use an RE instead.
06:03 dalek parrot: - side effect: forces " delim on CONST_STRING
06:03 dalek parrot: - shaves about 1/40s of each c2str.pl invocation.
06:03 dalek parrot: review: https://github.com/parrot/parrot/commit/4380854f2a
06:04 dukeleto ~~
06:04 cotto hio dukeleto
06:05 dukeleto cotto: howdy
06:05 cotto 300*25
06:05 aloha 7500
06:06 cotto 125*25
06:06 aloha 3125
06:06 cotto Coke, a 3s shorter build is a shorter build
06:08 sorear 700 0.1% improvements is a 50% improvement
06:08 sorear a good day for chromatic finds 20 0.1% improvements; don't expect anything higher of yourselfg
06:10 contingencyplan left #parrot
06:12 dukeleto Coke++ # shorter builds
06:19 gbacon left #parrot
06:22 Coke looks like pmc2c.pl has options we don't use.
06:25 Coke --no-lines?
06:30 bacek_at_work Coke, --no-lines is quite useful option. And it's definetely used
06:30 bacek_at_work It blocks putting #line directives into generated files
06:31 Coke bacek_at_work: where is it used?
06:31 Coke not in any build rule...
06:32 bacek_at_work Configure.pl --no-line-directives
06:35 Coke danke.
06:35 dalek parrot: fc57a47 | Coke++ | tools/build/pmc2c.pl:
06:35 dalek parrot: Add comment pointing to common way to enable this.
06:35 dalek parrot: review: https://github.com/parrot/parrot/commit/fc57a4731b
06:42 kurahaupo left #parrot
06:43 nwellnhof left #parrot
06:53 NotFound msg plobsing Looks like you haven't applied your setlocale patch, see http://smolder.parrot.org/app​/projects/report_details/2196
06:54 aloha OK. I'll deliver the message.
07:34 dalek parrot/leto/embed_grant: 5a5526c | dukeleto++ | t/src/embed.t:
07:34 dalek parrot/leto/embed_grant: [t] Parrot_PMC_add_float
07:34 dalek parrot/leto/embed_grant: review: https://github.com/parrot/parrot/commit/5a5526ca68
07:48 cotto seen tcurtis
07:48 aloha tcurtis was last seen in #parrot 6 days 11 hours ago joining the channel.
07:55 theory left #parrot
08:28 bacek aloha, humans
08:29 bacek mj41++ # moving taptinder forward
08:29 bacek seen moritz
08:29 aloha moritz was last seen in #parrot 13 hours 20 mins ago saying "array sounds find".
08:29 fperrad joined #parrot
08:29 bacek moritz, yeah... I can reproduce your failures in rakudo. Don't quite understand why they are happen.
08:32 bacek msg moritz, yeah... I can reproduce your failures in rakudo. Don't quite understand why they are happen.
08:32 aloha OK. I'll deliver the message.
08:40 * moritz neither
08:43 bacek moritz, bisecting now
08:58 bacek moritz, sigh... It's way too coarse commit to find out real issue.
09:01 moritz bacek: renaming method 'backtrace' in src/Perl6/Compiler.pir should give you PIR level stacktraces
09:02 moritz maybe that will make it easier to debug
09:02 bacek moritz, "renaming"?
09:02 moritz bacek: or deleting
09:02 bacek ah, ok. Thanks. Will try
09:03 * moritz would like to make it a command line flag
09:03 moritz how can I install a method at run time?
09:03 moritz or a method alias
09:07 bacek moritz, intersting. Generated PIR is exactly same for master and nqp_pct. At least for (1..5).perl.say
09:08 moritz bacek: was that an example that failed?
09:08 bacek moritz, one of them
09:09 moritz bacek: also notice that it calls into the compiled setting, so you'd have to compare their PIR too
09:09 bacek "1Nil()..Nil()5"
09:09 moritz src/gen/core.pir or so
09:09 bacek it will be... hard.
09:09 bacek there is way to many generated ids inside
09:10 moritz write a perl script that normalizes them :-)
09:11 bacek moritz, actually no. I didn't rebuild perl6. Just install master and run same executable.
09:11 bacek parrot's master
09:11 bacek which became even more interesting
09:16 dalek Heuristic branch merge: pushed 39 commits to parrot/nqp_pct by bacek
09:23 bacek perl6: say "hi"
09:23 p6eval pugs, rakudo ca62f2: OUTPUT«hi␤»
09:23 bacek perl6: (1,2).join('').say
09:23 p6eval pugs, rakudo ca62f2: OUTPUT«12␤»
09:24 sorear uh
09:24 sorear moritz: how many p6evals are there?
09:25 moritz sorear: one for freenode, one for irc.perl.org
09:26 moritz evalbot control restart
09:26 p6eval left #parrot
09:26 p6eval joined #parrot
09:26 moritz often when we change the source code, we only bother to restart the freenode instance
09:26 moritz perl6: say 1
09:26 p6eval pugs, rakudo ca62f2, niecza v1-123-g04bce2a: OUTPUT«1␤»
09:27 bacek perl6: ("a", ("^" if Bool::False), "b").join("").say;
09:27 p6eval pugs: OUTPUT«*** ␤    Unexpected "if"␤    expecting operator or ")"␤    at /tmp/5srL95ggIV line 1, column 12␤»
09:27 p6eval ..rakudo ca62f2, niecza v1-123-g04bce2a: OUTPUT«ab␤»
09:27 bacek moritz, ("a", ("^" if Bool::False), "b").join("").say; gives me "aNil()b" on nqp_pct branch. Interesting.
09:33 bacek perl6: ("a", Nil, "b").join("").say
09:33 p6eval rakudo ca62f2, niecza v1-123-g04bce2a: OUTPUT«ab␤»
09:33 p6eval ..pugs: OUTPUT«*** No such subroutine: "&Nil"␤    at /tmp/HZItLKutQa line 1, column 1 - line 2, column 1␤»
09:33 bacek interesting...
09:35 sorear bacek: does it give the same pir?
09:35 bacek sorear, exactly same
09:35 bacek rakudo: say join("a", Nil, "b")
09:35 p6eval rakudo ca62f2: OUTPUT«b␤»
09:35 bacek rakudo: say join("", "a", Nil, "b")
09:35 p6eval rakudo ca62f2: OUTPUT«ab␤»
09:36 moritz rakudo: say Nil.Str
09:36 p6eval rakudo ca62f2: OUTPUT«␤»
09:36 moritz rakudo: say Nil.Stringy
09:36 p6eval rakudo ca62f2: OUTPUT«Method 'Stringy' not found for invocant of class ''␤  in main program body at line 22:/tmp/9Z4IREN53O␤»
09:36 moritz rakudo: say ~Nil
09:36 bacek ~/src/rakudo (master)$ ./perl6 -e 'say join("", "a", Nil, "b")'
09:36 bacek ab
09:36 bacek ~/src/rakudo (master)$ ./perl6 -e '("a", Nil, "b").join("").say'
09:36 bacek aNil()b
09:36 p6eval rakudo ca62f2: OUTPUT«Nil()␤»
09:36 bacek it's what I've got on nqp_pct branch...
09:37 moritz bacek: it's probably .Str vs get_string vtable
09:37 sorear bacek: did you rebuild rakudo's setting after switching branches?
09:37 bacek moritz, yes... But isn't C<join> just export version of List.join?
09:37 bacek sorear, nope
09:38 moritz the sub has the separator as first arg
09:38 moritz the method has the list as invocant
09:39 bacek sub join($j, *@list) { @list.join($j) }?
09:47 moritz yes
09:47 bacek than why result is different?
09:50 dalek parrot: fb65249 | bacek++ | DEPRECATED.pod:
09:50 dalek parrot: Deprecate "constant" PObjs.
09:50 dalek parrot: review: https://github.com/parrot/parrot/commit/fb65249e44
09:53 dalek parrot: 93731ed | bacek++ | DEPRECATED.pod:
09:53 dalek parrot: Add Parrot_new_constant_* functions and PObj_constant_FLAG into deprecation.
09:53 dalek parrot: review: https://github.com/parrot/parrot/commit/93731ede61
10:26 dalek TT #1947 created by cotto++: excessively noisy output from Configure.pl
10:26 dalek TT #1947: http://trac.parrot.org/parrot/ticket/1947
10:33 dalek parrot/kill_constant: 7213e11 | bacek++ | / (2 files):
10:33 dalek parrot/kill_constant: Remove Parrot_pmc_new_constant* functions. Build is broken for now
10:33 dalek parrot/kill_constant: review: https://github.com/parrot/parrot/commit/7213e11955
10:33 dalek parrot/kill_constant: fb5514e | bacek++ | / (8 files):
10:33 dalek parrot/kill_constant: Remove usage of Parrot_pmc_new_constant. Register few PMCs where appropriate
10:33 dalek parrot/kill_constant: review: https://github.com/parrot/parrot/commit/fb5514eb74
10:36 cotto Yay.  more killing of unnecessary code!
10:36 cotto I'm sleeping now.  'night.
11:35 prabuinet joined #parrot
11:36 prabuinet hi there anybody know how to run a PIR file in utf8?
11:45 TiMBuS joined #parrot
11:46 ppant left #parrot
12:20 moritz prabuinet: PIR has a syntax for escaping non-ASCII characters in string literals
12:20 moritz which you should use
12:20 moritz then you don't need any special options
12:22 prabuinet moritz: what to do if my variables should be in someother language?
12:26 Coke if the variables are exposed to the user (as opposed to, say, ".local" variables in a PIR sub), then they are always referred to as strings in PIR anyway.
13:00 gbacon joined #parrot
13:42 cosimo left #parrot
13:45 rurban_ joined #parrot
13:48 rurban left #parrot
13:48 rurban_ is now known as rurban
14:01 whiteknight joined #parrot
14:08 whiteknight good morning, #parrot
14:15 plobsing left #parrot
14:20 cosimo joined #parrot
14:31 plobsing joined #parrot
14:32 prabuinet left #parrot
14:41 tadzik cosimo: ping
14:41 cosimo tadzik: hi
14:43 gbacon left #parrot
14:45 tadzik cosimo: mind updating your projects to the new infrastructure?
14:45 tadzik see http://ttjjss.wordpress.com/2011/01/08/perl-​6-module-ecosystem-%E2%80%93-news-and-ideas/
14:46 tadzik oh, that's #parrot. Pardon my offtpic
14:46 cosimo tadzik: yeah, why not
14:46 cosimo i'll put it in my todo list
14:48 tadzik thx
15:03 cosimo left #parrot
15:05 dalek TT #1948 created by whiteknight++: Configure with --no-line-directives still produces .c files filled with ...
15:05 dalek TT #1948: http://trac.parrot.org/parrot/ticket/1948
15:07 Patterner left #parrot
15:08 Psyche^ joined #parrot
15:08 Psyche^ is now known as Patterner
15:15 bacek left #parrot
15:17 contingencyplan joined #parrot
15:18 dalek Heuristic branch merge: pushed 40 commits to parrot/exception_backtraces by Whiteknight
15:21 whiteknight msg plobsing take a look at the work in the new exception_backtrace branch. It does multi-level backtraces now following rethrow points like you requested. TT #1283 and a forthcoming mail to the list will have more details
15:21 aloha OK. I'll deliver the message.
15:22 fbrito joined #parrot
15:26 bacek joined #parrot
15:30 plobsing left #parrot
15:31 whiteknight msg cotto 3.0 is about a week away. We have a few branches that might be mergable within the next few days. I suspect we need lots of testing since this is a supported release. When do you want to call a feature freeze?
15:31 aloha OK. I'll deliver the message.
15:32 whiteknight msg mikehh the html_cleanup branch was merged but not deleted from github. Is it still necessary?
15:32 aloha OK. I'll deliver the message.
15:32 tadzik fbrito: is the GCI rank available?
15:36 mtk joined #parrot
15:36 whiteknight I think what we really need is a BackTrace PMC type that can encapsulate information about previous backtraces, and context/sub information, and provide both a complete backtrace string on demand or OO-based access to individual parts of it
15:39 fbrito left #parrot
15:40 moritz +1
15:40 moritz I've been planning to spec such a thing for Perl 6 too
15:41 whiteknight If I produce something for Parrot, it should be possible to subclass it from PIR/Rakudo
15:41 rurban Oi, python!
15:41 plobsing joined #parrot
15:48 Patterner left #parrot
15:49 Psyche^ joined #parrot
15:49 Psyche^ is now known as Patterner
15:49 fbrito joined #parrot
15:54 dalek tracwiki: v3 | whiteknight++ | ExceptionRefactor
15:54 dalek tracwiki: +break this up into a commentary section and a list of specific tasks. Start listing tasks.
15:54 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Exce​ptionRefactor?version=3&amp;action=diff
15:54 dalek TT #1854 closed by whiteknight++: Re-write of the embedded API
15:54 dalek TT #1854: http://trac.parrot.org/parrot/ticket/1854
15:59 theory joined #parrot
16:00 fbrito left #parrot
16:00 whiteknight msg cotto can you take a look at TT #1406? I would like to reject it because it's too much work for too little gain. Let me know if you have any opinions before I drop the hammer
16:00 aloha OK. I'll deliver the message.
16:07 whiteknight msg NotFound I don't think there are any remaining issues in TT #804. The patches were applied several hundred revisions ago and there are no reported problems. Can you close that ticket?
16:07 aloha OK. I'll deliver the message.
16:11 dalek TT #1780 closed by whiteknight++: Using StringHandle for STDERR can cause segfaults
16:11 dalek TT #1780: http://trac.parrot.org/parrot/ticket/1780
16:11 dalek TT #1455 closed by whiteknight++: Exception 'birthtime' attribute not accessible
16:11 dalek TT #1455: http://trac.parrot.org/parrot/ticket/1455
16:11 dalek TT #944 closed by whiteknight++: Exception handling in top-level main function is broken.
16:11 dalek TT #944: http://trac.parrot.org/parrot/ticket/944
16:11 fbrito joined #parrot
16:16 gbacon joined #parrot
16:45 plobsing whiteknight: exception_backtraces is a step in the right direction.
16:46 plobsing I see rethrow as 'exception handler defer', so intermediary information is not so necessary IMHO
16:47 plobsing also, not having that info in the backtrace array (as described on parrot-dev) means HLLs which implement their own backtrace formatter don't benefit from this.
16:49 p joined #parrot
16:50 p parrot cannot run a pir file with utf8 encoding?
16:51 plobsing p: it can and frequently does so for me
16:51 p in windows?
16:51 plobsing ah there's the rub
16:52 plobsing what error message do you get?
16:52 p error:imcc:syntax error, unexpected $endInvalid character in ASCII string
16:52 p imcc syntax error
16:53 plobsing in what kind of token are you using non-ascii characters (eg: string literal, local variable name, etc)?
16:53 p i've just give two lines
16:53 whiteknight plobsing: yes, it's a step in the right direction. I think of exception_backtrace as being more of a temporary fix than the final solution
16:53 p .sub main
16:53 p .end
16:54 whiteknight plobsing: I think the real solution is a proper BackTrace PMC type that can handle all the necessary information in an encapsulated and subclassable way
16:54 plobsing that error message indicates to me that you probably have 2 errors: 1 is an unexpected end-of-file, and a non-ascii character near that EOF
16:55 plobsing whiteknight: I don't care about intermediary catch info. if I have the original frame, all intervening scopes are already visible
16:56 plobsing p: the only way to get the error message you've described is if you have a non-ascii encoded character somewhere in your file. the code you've pasted does not contain such characters
16:57 plobsing utf-8 == ascii if ord(c) < 127
16:59 p i'm editing with notepad, saving the same file in ansi format and running with parrot , then not getting any error.
16:59 p if changed it to utf-8 or any other format , i'm getting error.
17:00 plobsing p: can you fire up a hex editor and look for characters greater than 0x80?
17:00 plobsing the file you've described should be identical in ascii or utf8
17:04 fbrito left #parrot
17:13 plobsing p: is it possible your editor is adding a BOM (byte-order-mark)?
17:21 p you are correct
17:21 p thanks
17:21 p it's BOM
17:22 plobsing I wonder if it is possible to make lex/yacc deal with BOMs correctly
17:25 p plobsing: identifiers in pir cannot contain foreign language characters?
17:28 plobsing it isn't in the spec anywhere, but I suspect the parser wouldn't like it
17:29 whiteknight left #parrot
17:29 plobsing p: why do you want foreign language identifiers in PIR variable names?
17:29 plobsing s/variable/register/
17:31 p i'm writing a compiler which converts a programming-language to PIR in which variables are declared in foreignlang
17:31 plobsing yep, the lexer defines identifiers to be [a-zA-Z_@][a-zA-Z0-9_@]*
17:32 plobsing p: you shouldn't try to map HLLs directly to PIR. it always ends in frustration.
17:32 p so what is the right way to do?
17:32 p directly to pbc?
17:33 plobsing use $Pxx registers. possibly add comments to indicate the mapping.
17:34 plobsing compiling directly to PBC might be a long-term objective, but there are some caveats involved. PBC is not static, but changes to better serve the needs of users.
17:34 p ok
17:34 plobsing p: look at Winxed (winxed.googlecode.com) for probably the closest thing to direct mapping from an HLL to PIR that is advisable
17:35 p ok
17:36 plobsing OOC, what language are you trying to compile to parrot? if it is custom, do you have a description of it somewhere?
17:39 p no new ideas, just for learning.
17:39 p i'm trying to write a lang similar to c-like language , with just symbols changed to my local language.
17:40 p i'm trying to write a c-like language , with just symbols changed to my local language.
17:41 plobsing sounds like close (close.googlecode.com) with a twist
17:41 Kristaba joined #parrot
17:42 nwellnhof joined #parrot
17:45 p plobsing: do you know any such lang's using antlr ?
17:46 plobsing p: I do not think there are any comiplers that target parrot implemented with ANTLR
17:47 plobsing the usual advice is to use Perl 6 grammars (either through PGE, NQP, or Rakudo), however there are other compilers doing their own thing (eg: Winxed)
17:47 plobsing and of course, I'm implementing another compiler framework based on PEGs (OMeta)
17:48 particle antlr does not have a gode generation target for PIR
17:48 particle s/gode/code/
17:49 p i tried using pct/pge/nqp, but since i'm new to perl, i'm not able to understand a the new version of regular expression...etc
17:49 particle read synopsis 5, it's enlightening: http://perlcabal.org/syn/S05.html
17:50 moritz p: maybe you're interested in the chapters on Regexes and Grammars in the "Using Perl 6" book
17:50 particle that'll help, too!
17:50 moritz p: https://github.com/perl6/book/downloads
17:54 dukeleto ~~
17:54 moritz what do you want to smartmatch today?
17:55 plobsing moritz: is that the new Perl 6 tagline?
17:55 moritz plobsing: it's an appropriate response to  ~~ :-)
17:56 p i've a perl6 question, can we change the grammar of the perl6 at runtime..?
17:56 extreme joined #parrot
17:56 dukeleto moritz: not a bad idea for a Tshirt ...
17:56 dukeleto p: i think you might want #perl6 on freenode
17:56 PerlJam p: philosophically, or in an actual implementation? :)
17:56 dukeleto p: but the answer is yes
17:56 moritz p: yes (though currently only a bit)
17:58 p moritz: what do you mean by a bit?
17:58 moritz p: rakudo implements custom operators, which are small grammar changes
18:01 cotto ~~
18:04 * jnthn is doing some groundwork that will hopefully get us slangs in a little while.
18:04 jnthn Which allow bigger changes.
18:05 tadzik nice
18:08 dalek TT #1406 closed by cotto++: get_results opcode used to fetch exception object
18:08 dalek TT #1406: http://trac.parrot.org/parrot/ticket/1406
18:10 dukeleto jnthn: i love slang. It was my first language...
18:11 jnthn :)
18:12 p left #parrot
18:12 * jnthn finds it cute that *s*ub-*lang*uage => slang :)
18:12 jnthn Suspect that may be backtomology rather than etymology though.
18:13 plobsing left #parrot
18:13 dukeleto jnthn: I like how the Germans say slang : Umgangsprache = "sidewalk language"
18:13 dukeleto die Umgangsprache # in case my old german teachers are looking
18:14 jnthn Nice :)
18:14 tadzik heh "Die GNU Autotools" :)
18:15 dukeleto tadzik: yes, the German word "damit" also makes for fun mistranslations
18:15 moritz dukeleto: that's not how I'd translate "Umgang"
18:15 dukeleto moritz: perhaps I was mistaught. How would you translate it?
18:15 moritz Umgang = contact
18:16 tadzik „Ich besischtige die sehenswurdigkeiten” is my German as its best ;)
18:16 moritz or the people you have contact with
18:17 dukeleto moritz: you seem to be totally correct. My teacher lied to me.
18:18 jnthn No no, it's quite simple. It's the, um, gang speak. :)
18:18 moritz lol
18:18 moritz dukeleto: maybe he confused it with Umweg = detour  or something similar
18:19 dukeleto moritz: i think i took literally what he was saying metaphorically.
18:19 * dukeleto feels the rays of enlightenment
18:20 dukeleto moritz: what are you hacking on these days?
18:20 moritz dukeleto: yesterday I hacked Devel::Trace into rakudo
18:20 dukeleto moritz: i have been meaning to backport your rakudo build stuff into create_language/mk_language_shell in parrot's tools/dev
18:20 dukeleto moritz: $cow.holy!
18:22 moritz rakudo: use Devel::Trace; say rand; say (1..5).pick;
18:22 p6eval rakudo ca62f2: OUTPUT«use Devel::Trace␤say rand␤0.621661508546339␤say (1..5).pick␤1..5␤5␤»
18:22 moritz it's lexically scoped
18:22 moritz and it was just like 15 lines or so
18:24 moritz I wonder why the 1..5 is printed separately
18:24 moritz apart from that, masak's p6 coding contest
18:34 cotto left #parrot
18:48 plobsing joined #parrot
18:51 cotto_work ~~
18:55 dukeleto cotto_work: wilkommen!
18:59 cotto_work hallo, dukeleto
19:08 kurahaupo joined #parrot
19:14 whiteknight joined #parrot
19:20 cotto_work mikehh: ping
19:31 whiteknight good afternoon, #parrot
19:34 cotto_work good morning, whiteknight
19:35 mtk left #parrot
19:37 kurahaupo left #parrot
19:37 kurahaupo joined #parrot
19:38 mtk joined #parrot
19:41 tadzik good evening cotto_work, whiteknight
19:42 whiteknight hello cotto_work, tadzik
19:47 nwellnhof left #parrot
19:49 atrodo dukeleto> You liking linode?
19:50 Yuki`N joined #parrot
19:51 cotto_work Yuki`N: ohai
19:51 Yuki`N Hello thar.
19:51 cotto_work nice to see you back
19:52 Yuki`N Mmm.
19:52 Yuki`N Doing course selections for senior year actually.
19:53 sorear oh goody, they didn't all leave after the end :>
19:55 dukeleto atrodo: yeps
19:55 dukeleto Yuki`N: you are still here! awesomesauce
19:55 atrodo dukeleto> Me too.  Real nice setup they have going
19:55 cotto_work Yuki`N: you should try to hack on Parrot as an independent study project.
19:56 dukeleto Yuki`N: we can write a syllabus for you ;)
19:56 atrodo Ah independent study.  Best two semesters of my college career.
19:56 Yuki`N Hah.
19:56 Yuki`N I really hope Google doesn't screw all us AP-testers over.
19:57 dukeleto Yuki`N: why do you say that?
19:58 Yuki`N The consent form says "on or about may 1."
19:58 Yuki`N That week is AP exam week.
19:59 * cotto_work doesn't see a connection.  Are you talking about Summer of Code?
19:59 Yuki`N The "trip to cali" for the GCI winners.
19:59 cotto_work ah
19:59 dukeleto Yuki`N: if you can't attend at a certain time, i am sure you can visit another time. Don't worry
20:00 dukeleto Yuki`N: they *want* you to go there :)
20:00 mikehh cotto_work: pong
20:00 cotto_work Yuki`N: the rules say "certain blackout dates apply".  That implies that you have some choice.
20:00 cotto_work mikehh: what all needs review from the html_cleanup merge?  istr something about links possibly being wonky.
20:01 dukeleto did html_cleanup finally get merged?
20:01 tadzik si
20:02 Yuki`N I'm not sure if it's a preplanned thing or whether or not the date is negotiable lol.
20:02 mikehh cotto_work: not really, I checked the navigation as much as I could and fixed any problems I found, I think I got most
20:02 Yuki`N CompSci A is actually may 3rd the tuesday, us history is may 6th the friday, and I have another AP the next week.
20:02 Yuki`N The first 2 weeks of may really are no good.
20:02 cotto_work dukeleto: it did
20:03 dalek TT #804 closed by NotFound++: Avoid entering inner runloop when catching a exception thrown from C with ...
20:03 dalek TT #804: http://trac.parrot.org/parrot/ticket/804
20:03 mikehh cotto_work: the possible problems occur in that it takes the Name of the link to pages from the page title that is generated from thy pod
20:04 mtk left #parrot
20:04 mikehh cotto_work: and this might need fixing in the documentation of the page
20:05 cotto_work mikehh: ok
20:07 mikehh in the old system there was (in the Parrot::Docs::Section files bothe the title and the page, now we only have the page and have to pick up the title from that page
20:08 Yuki`N left #parrot
20:09 mikehh for example the first page in index.html was Introduction to Parrot now it picks up Parrot Primer
20:09 Yuki`N joined #parrot
20:10 NotFound plobsing: What is setlocale used for?
20:10 cotto_work #ps in 20
20:12 mikehh whiteknight: the branch is no longer necessary - feel free to delete it (I tried but I had a problem there - probvably from the wrong place)
20:12 kurahaupo left #parrot
20:12 kurahaupo joined #parrot
20:12 acrussell joined #parrot
20:12 Coke we can put a copy of the the new "make html" out on docs.parrot.org in an unadvertized URL for testing if folks want.
20:18 dukeleto Coke: sounds nice
20:18 cotto_work +1
20:19 NotFound +1
20:19 mtk joined #parrot
20:19 dalek TT #1878 closed by cotto++: Increase minimum perl 5 version to 5.10.0
20:19 dalek TT #1878: http://trac.parrot.org/parrot/ticket/1878
20:20 dalek parrot: 06edc1c | plobsing++ | / (2 files):
20:20 dalek parrot: only be locale-specific about character type (LC_CTYPE, eg: UTF8)
20:20 dalek parrot:
20:20 dalek parrot: respecting locale globally breaks tests in many non-english locales by changin
20:20 dalek parrot: printf, errno messages, etc
20:20 dalek parrot: review: https://github.com/parrot/parrot/commit/06edc1c9f3
20:21 plobsing NotFound: setlocale can be used to set and query the locale. if used with an empty string, it will use the locale specified by the environment. for back-compat, all programs start with "C" for all locales
20:21 cotto_work #ps in 9
20:22 NotFound plobsing: I know, but what feature are we using now?
20:22 plobsing LC_CTYPE (which controls character type only)
20:23 NotFound And for what? Collating order in strcmp?
20:24 plobsing nope, just for determining character type. the only libc functions it controls are the 'multi-byte string' functions, which we don't use.
20:26 NotFound LC_CTYPE - Character classification and case conversion --> This may have unexpected side effects in our codebase.
20:27 NotFound For example, isalpha is used in several places.
20:27 kid51 joined #parrot
20:28 NotFound From the ma page: "In  some locales, there may be additional characters for which  isalpha() is true"
20:28 whiteknight the OS is a constant. We need to handle whatever inputs the OS throws at us. If we have to reimplement our own isalpha, we do that
20:31 hercynium joined #parrot
20:31 kurahaupo1 joined #parrot
20:31 cotto_work #ps in now
20:32 kid51 left #parrot
20:32 plobsing we use isalpha() in 2 places, both of which are not well equipped to deal with unicode regardless of localization
20:32 NotFound If we use the setlocale just for look up the user encoding, doing it one time and then revert to "C" locale may be the easiser solution.
20:36 dalek TT #1949 created by doughera++: Remove C++ style comments in pbc_to_exe.pir
20:36 dalek TT #1949: http://trac.parrot.org/parrot/ticket/1949
20:36 dalek TT #1950 created by doughera++: pbc_disassemble dumps core on illegal options
20:36 dalek TT #1950: http://trac.parrot.org/parrot/ticket/1950
20:44 plobsing NotFound: that can be arranged
20:45 NotFound About LC_MESSAGES, setting it makes fail a lot of tests that look for error strings. Not easy to fix that without extensive testing with a lot of locales.
20:46 whiteknight honestly, I hate all the tests that test for the exact strings in error message
20:46 whiteknight even error_like tests are evil
20:47 NotFound whiteknight: yes, but is not easy to catch them all.
20:49 whiteknight Andy: ping
20:49 cotto_work whiteknight: yes, but they're better than no tests.
20:49 whiteknight cotto_work: true, but it would be easy enough to go through all instances of pir_error_output_like and c_error_output_like, and family, and change them to not match on the text of error messages
20:50 whiteknight they're not necessarily better than no tests if they get in the way
20:52 plobsing if you don't want your c library error messages localized, why do you have the locale set?
20:53 dalek TT #1951 created by doughera++: t/dynpmc/os.t incorrectly uses system-specific error strings
20:53 dalek TT #1951: http://trac.parrot.org/parrot/ticket/1951
20:53 dalek TT #1952 created by doughera++: t/src/misc.t tests 1 and 2 fail under C89 compiler
20:53 dalek TT #1952: http://trac.parrot.org/parrot/ticket/1952
20:53 plobsing and we should just as easily be able to check against errno as we can against strings
20:53 plobsing checkign against errno is probably a better idea
20:54 NotFound plobsing: the locale system is not granular enough to say: "I want applications errors and screen text localized, but let "C" system errors unchanged"
20:55 NotFound plobsing: there is ticket since years to make errno values available in pir.
20:55 AzureStone left #parrot
21:01 AzureStone joined #parrot
21:06 plobsing left #parrot
21:07 dalek Heuristic branch merge: pushed 38 commits to parrot/kill_packfile_new_dummy by Whiteknight
21:08 mtk left #parrot
21:09 [hercynium] joined #parrot
21:10 [hercynium]_ joined #parrot
21:11 dalek parrot: 06b1fad | Whiteknight++ | DEPRECATED.pod:
21:11 dalek parrot: Deprecate functions in src/embed.c. TT #1910
21:11 dalek parrot: review: https://github.com/parrot/parrot/commit/06b1fad972
21:14 mtk joined #parrot
21:14 hercynium left #parrot
21:15 [hercynium] left #parrot
21:19 plobsing joined #parrot
21:20 nwellnhof joined #parrot
21:25 nopaste "plobsing" at 192.168.1.3 pasted "[PATCH] reset locale for NotFound" (81 lines) at http://nopaste.snit.ch/27681
21:25 plobsing NotFound: try ^^this^^ please
21:26 NotFound plobsing: going
21:28 whiteknight left #parrot
21:30 plobsing bacek: if you are changing the syntax of PIR, please eliminate bare label names as op-args. these lead to *very* misleading error messages.
21:30 bacek plobsing, e.g.?
21:30 bacek "goto foo_label"?
21:31 plobsing bacek: yes that. replace that with goto &foo_label except for sugared cases.
21:31 NotFound There is a bunch of opcodes that can take labels,.
21:31 plobsing because, right now, to support arbitrary ops, any not found identifier must be assumed to be a forward reference to a label
21:31 bacek plobsing, interesting idea. I can try to implement it in PIRATE. But not in IMCC.
21:32 plobsing which means 'set $P0, no_such_identifier' becomes 'set_p_ic'
21:32 plobsing *very* broken
21:34 jnthn oh, *that's* why it does that.
21:34 jnthn Yes, it's very annoying.
21:34 cotto_work plobsing: wow and yuck
21:35 plobsing I have a ticket for that somewhere...
21:36 NotFound Is the rot dynop fixed, or has the test been todoed or skiped?
21:37 cotto_work NotFound: I committed a fix that didn't make it worse.
21:38 cotto_work from Yuki`N iirc (which I may not)
21:38 NotFound cotto_work: pass for me now in amd64, then
21:38 cotto_work yup.  That was from Yuki`N++
21:38 cotto_work good news then
21:39 NotFound plobsing: all test pass both with mu usual locale and with "C", and TODO'ed "OS.readdir with ord >127" also pass on both.
21:39 * cotto_work needs to attend to $dayjob
21:40 plobsing committing...
21:40 plobsing PIR labels ticket is TT #1323
21:41 bacek left #parrot
21:43 dalek parrot: ae83646 | plobsing++ | / (3 files):
21:43 dalek parrot: reset locale after grabbing character encoding - fixes tests on internationalized systems
21:43 dalek parrot: review: https://github.com/parrot/parrot/commit/ae83646865
21:43 plobsing #ps goal met before end of #ps. win.
21:45 rurban_ joined #parrot
21:49 rurban left #parrot
21:49 rurban_ is now known as rurban
21:51 NotFound plobsing: any reason to keep that OS test todo'ed ?
21:52 plobsing I don't know why it was todoed in the first place
21:52 mtk left #parrot
21:53 NotFound I dunno, some people just don't like to see fail reports.
21:53 bacek joined #parrot
21:53 NotFound Even when things are really failing.
21:53 plobsing not wanting to see something doesn't mean you should stumble around with your eyes closed
21:54 NotFound Going to untodo...
21:57 Yuki`N That todo'd OS test passes for me.
21:58 plobsing Yuki`N: was it failing for you before?
21:59 Yuki`N I don't think so.
21:59 Yuki`N It even passes on win32.
21:59 dalek TT #1953 created by doughera++: platform/generic/env.c incorrectly frees string passed to putenv()
21:59 dalek TT #1953: http://trac.parrot.org/parrot/ticket/1953
22:02 dalek parrot/nwellnhof/platform_src: 33c545b | nwellnhof++ | src/platform/ (34 files):
22:02 dalek parrot/nwellnhof/platform_src: Make headerizer work on src/platform files
22:02 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/33c545b5b5
22:02 dalek parrot/nwellnhof/platform_src: 03127e6 | nwellnhof++ | / (33 files):
22:02 dalek parrot/nwellnhof/platform_src: Move platform IO code to src/platform
22:02 dalek parrot/nwellnhof/platform_src:
22:02 dalek parrot/nwellnhof/platform_src: And get rid of io_portable. It must have been broken for years.
22:02 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/03127e69d4
22:02 dalek parrot/nwellnhof/platform_src: 72f9128 | nwellnhof++ | / (10 files):
22:02 dalek parrot/nwellnhof/platform_src: Merge platform/*/pid.c into platform/*/exec.c
22:02 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/72f9128234
22:02 dalek parrot/nwellnhof/platform_src: 688366d | nwellnhof++ | / (7 files):
22:02 dalek parrot/nwellnhof/platform_src: Add Parrot_lstat_info_intval
22:02 dalek parrot/nwellnhof/platform_src:
22:02 dalek parrot/nwellnhof/platform_src: And use Parrot_stat_* functions in dynpmcs
22:02 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/688366d5e6
22:04 nwellnhof parrot/nwellnhof/platform_src: 122 files changed, 1182 insertions(+), 3591 deletions(-)
22:07 plobsing nwellnhof: please make sure to grab my recent setlocale fix
22:10 dalek parrot: d539bfb | NotFound++ | t/dynpmc/os.t:
22:10 dalek parrot: untodo OS test, TT #1836
22:10 dalek parrot: review: https://github.com/parrot/parrot/commit/d539bfbcda
22:12 [hercynium]_ is now known as hercynium
22:13 NotFound Yuki`N: It was passing on windows? What compiler?
22:16 Yuki`N NotFound, vs2010
22:18 NotFound Yuki`N: Can you try to unskip the tests skiped for windows in that file and report to TT #1836?
22:19 NotFound Minus the "Hardlinks to files not possible on Windows", of course.
22:20 NotFound Better said, minus a lot of them that doesn't make sense on windows.
22:26 dalek parrot: c638c31 | bacek++ | tools/dev/pbc_to_exe.pir:
22:26 dalek parrot: Remove c++-style commented out line. Closes #1949. doughera++
22:26 dalek parrot: review: https://github.com/parrot/parrot/commit/c638c312ef
22:28 Yuki`N NotFound, you know that Win7 and Vista-ish supports symlinks on NTFS.
22:29 NotFound Yuki`N: my windows knowledged is several years outdated.
22:30 Yuki`N So what am I trying to unskip?
22:30 NotFound I've fixed several problems using my old laptop with XP Home, go figure X-)
22:31 Yuki`N I also have some weird problems involving the configure script on Windows.
22:31 Yuki`N Sometimes steps will spontaneously fail, when they've passed before.
22:32 dalek TT #1949 closed by bacek++: Remove C++ style comments in pbc_to_exe.pir
22:32 dalek TT #1949: http://trac.parrot.org/parrot/ticket/1949
22:33 NotFound Yuki`N: this one: pir_output_is( <<'CODE', $entries2, 'Test OS.readdir with ord >127' );
22:33 dalek parrot/nwellnhof/platform_src: 770403b | nwellnhof++ | / (3 files):
22:33 dalek parrot/nwellnhof/platform_src: Make win32 tests pass
22:33 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/770403be47
22:34 NotFound Yuki`N: and the others in that block, if readdir is really working on windows now.
22:36 Yuki`N Readdir works, but the non-ascii filenames does not.
22:36 nwellnhof Yuki`N: i have the same problem with Configure.pl
22:37 NotFound Yuki`N: but that test was passing, wasn't it?
22:37 Yuki`N No, readdir never passed.
22:37 Yuki`N The todo'd can_write OS test passed.
22:37 Yuki`N The readdir tests were skipped.
22:38 nwellnhof we should move the windows code away from readdir to the native windows api
22:38 Yuki`N Yeah.
22:38 NotFound <Yuki`N> That todo'd OS test passes for me. <Yuki`N> It even passes on win32.
22:38 Yuki`N Todo'd, I said. readdir was skipped.
22:38 preflex left #parrot
22:38 Yuki`N Native WinAPI supports widechars.
22:39 NotFound Yuki`N: I'm not sure, the mix of skip and todo in that part of the file confuses me.
22:40 dalek parrot/nwellnhof/platform_src: 644658e | nwellnhof++ | / (54 files):
22:40 dalek parrot/nwellnhof/platform_src: Merge branch 'master' into nwellnhof/platform_src
22:40 dalek parrot/nwellnhof/platform_src: review: https://github.com/parrot/parrot/commit/644658e933
22:40 Yuki`N Yeah. :/
22:40 Yuki`N I'm getting test failures with t/dynoplibs/debug.t and t/library/streams.t
22:41 nwellnhof does dalek handle merges to branches now?
22:41 NotFound Yuki`N: yes, but we should change several file and directory functions to use the wide char interface, recoding when needed.
22:41 preflex joined #parrot
22:41 Yuki`N That is true.
22:42 NotFound And we are short of developers interested in low level windows details.
22:42 nwellnhof NotFound: that's one of the reasons i created the platform_src branch
22:43 nwellnhof now we have all platform-specific code in one place and don't do silly things like generating src/platform.c during config.
22:43 NotFound nwellnhof++
22:44 nwellnhof but we could really need some win32 hackers...
22:45 cotto_work yeah.  They're in short supply.
22:45 NotFound Windows developers are mostly interested in .net things these days, it seems.
22:46 Yuki`N .net is beautiful tbh.
22:46 Yuki`N At least I like it.
22:46 jnthn Well, also there's plenty of jobs in .Net development.
22:47 NotFound Compiling parrot with managed C++ may be doable?
22:47 jnthn There's a decided shortage of them where I live. Well, of good ones anyway.
22:47 jnthn NotFound: Maybe, but I don't know how much of a win it'd be. :)
22:47 jnthn NotFound: If any. :)
22:48 NotFound jnthn: me neither, is just an idea.
22:48 gbacon left #parrot
22:48 Yuki`N imho managed c++ is pretty stupid though.
22:49 NotFound Yuki`N: I don't care, if it can be a way to have a version regularly built and tested.
22:50 Yuki`N Parrot is C, anyway, isn't it?
22:50 NotFound Yuki`N: is mostly written in the common subset of C and C++
22:51 NotFound With conditional compling hiding the differences.
22:53 NotFound Yuki`N: to compile it as C++ in linux: perl Configure.pl --cc=g++ --cxx=g++ --link=g++ --ld=g++
22:54 Yuki`N Ah.
22:54 NotFound It helps to catch several kinds of errors easily
22:56 plobsing left #parrot
22:58 cotto_work mingw32 and msvc smolder reports submitted
22:59 NotFound cotto_work: unfortunately, smolder reports doesn't show what happens with skipped tests.
23:12 hercynium left #parrot
23:13 dalek nqp-rx/nom: fac4f86 | jonathan++ | src/ (2 files):
23:13 dalek nqp-rx/nom: Pass name along to meta-object so classes know how they're called.
23:13 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/fac4f86680
23:13 dalek nqp-rx/nom: cbd0747 | jonathan++ | src/stage0/ (3 files):
23:13 dalek nqp-rx/nom: Update bootstrap with recent changes.
23:13 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/cbd07472dc
23:13 dalek nqp-rx/nom: ab3a2a5 | jonathan++ | / (3 files):
23:13 dalek nqp-rx/nom: Add a barebones NQPMu and make it the default base class for anything declared with NQPClassHOW.
23:13 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/ab3a2a53be
23:20 dalek nqp-rx/nom: 5821a4c | jonathan++ | t/nqp/28-subclass.t:
23:20 dalek nqp-rx/nom: Correct a test (also previously corrected in 6model repo copy of the test suite).
23:20 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/5821a4cbe5
23:20 dalek nqp-rx/nom: 67bb846 | jonathan++ | src/NQP/ (2 files):
23:20 dalek nqp-rx/nom: Switch the class package declarator to use 6model. There are a couple of regressions out of this. One is of multis, which I don't care about since they're getting re-done Real Soon Now. The other is the :vtable trait handling, which I'll fix at some point, but is low priority. At this point, just grammars are left using P6metaclass.
23:20 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/67bb846358
23:21 plobsing joined #parrot
23:32 fperrad left #parrot
23:38 kid51 joined #parrot
23:39 dalek nqp-rx/nom: d357201 | jonathan++ | src/stage0/ (3 files):
23:39 dalek nqp-rx/nom: Update bootstrap with class declarator using 6model, so class_6m elimination can start.
23:39 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/d357201486
23:39 dalek nqp-rx/nom: 09ed288 | jonathan++ | src/ (7 files):
23:39 dalek nqp-rx/nom: Remove usage of class_6m.
23:39 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/09ed2886fe
23:48 kid51 cotto_work: ping
23:48 cotto_work jnthn: it's exciting to see nom start to fully use the new object model
23:48 cotto_work kid51: pong
23:48 kid51 cotto_work: http://trac.parrot.org/parrot/ticket/1925 about removing auto::jit ...
23:49 kid51 ... Do you (a) accept it in general; (b) want it done before 3.0?
23:49 jnthn cotto_work: Yes, it's going surprisingly well.
23:50 cotto_work That's a very internal change that I'd be comfortable having removed any time before the weekend.  There's no way anyone should be depending on that outside Parrot itself unless I'm misunderstanding it.
23:50 kid51 k
23:51 cotto_work I would prefer sooner than later, though.
23:53 dalek nqp-rx/nom: f0b00d8 | jonathan++ | src/NQP/ (2 files):
23:53 dalek nqp-rx/nom: Remove last remains of class_6m from the grammar/actions.
23:53 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/f0b00d807f
23:53 dalek nqp-rx/nom: 013ebe2 | jonathan++ | src/metamodel/how/NQPClassHOW.pm:
23:53 dalek nqp-rx/nom: Publish type cache (pretty much directly taken from 6model/common).
23:53 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/013ebe2308
23:53 dalek nqp-rx/nom: 80c8322 | jonathan++ | src/metamodel/how/NQPMu.pm:
23:53 dalek nqp-rx/nom: A cheating NQPMu.ACCEPTS. Need multi-dispatch done to do the non-cheating version that we have in nqpclr. It's enough to reclaim a regressed test file, for now.
23:53 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/80c8322b2b
23:53 cotto_work kid51: is there a branch for that ticket?
23:56 whiteknight joined #parrot
23:57 plobsing left #parrot
23:58 cotto_work kid51: making a branch now]
23:58 kid51 cotto_work:  Huh? I was just preparing to merge the remove_auto_jit branch
23:58 cotto_work silly me
23:58 kid51 You don't need to do anything.
23:59 cotto_work proceed
23:59 kid51 Am testing.
23:59 kid51 Should have results < 15 min.

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

Parrot | source cross referenced