Camelia, the Perl 6 bug

IRC log for #parrot, 2011-01-18

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 NotFound_b That trans.t thing needs review. At least its comments looks wrong if you look at what the pow opcode does.
00:02 cotto_work NotFound_b: which trans.t thing?
00:02 kid51 NotFound_b: The tests are not failing for me now -- which, of course, doesn't mean they are *good* tests.  Perhaps we're at the point where we should open a TT.
00:03 cotto_work It's nice that *some* of our releases are boring.
00:03 NotFound_b cotto_work: the pow overflow tests
00:03 * kid51 gets ready to do laundry
00:04 NotFound_b kid51: thet are not failing now just because its skipping is more consistent.
00:13 dukeleto we should recruit this dude: http://ruby-php.org/
00:15 cotto_work dukeleto: wtg.  He sounds like someone who'd be interested.
00:20 whiteknight joined #parrot
00:21 whiteknight good evening #parrot
00:22 kid51 NotFound_b: True
00:22 kid51 linux/i386 make fulltest PASS
00:27 cotto_work hio whiteknight
00:27 whiteknight heidi-ho neighboroony
00:29 nwellnhof left #parrot
00:29 cotto_work whiteknight: I think you're going to like the release announcement.  Barring insanity (which I really wish we could do), it should be waiting for you when you wake up.
00:31 whiteknight awesome. I look forward to it
00:32 whiteknight a new era of Parrot is coming. I think the 3.0 months are going to be awesome
00:39 kid51 darwin/ppc: make test PASS
00:42 cotto_work Keep the good news coming!
00:42 cotto_work also bad news if you've got it
00:44 kid51 linux/i386: configured with optimize and g++ all around: PASS
00:44 kid51 make test PASS, that is
00:45 dukeleto kid51: glad to hear that you agree about Parrot::Revision
00:47 kid51 It may have been a good idea ... but it's too close to release and he didn't revise the corresponding test
00:47 cotto_work yup
00:47 * kid51 to laundry
00:47 cotto_work I'll be glad to see it go in after the release.
00:48 kid51 I want to look at it more closely first.
00:48 cotto_work kid51: it's in a branch now
00:48 kid51 ... and I want it tested somehow.
00:48 kid51 afk
00:48 cotto_work whiteknight: next time you move an op from experimental.ops, please run make bootstrap-ops.
00:49 whiteknight ok
00:49 dukeleto it does O(N) work to figure out something about git, where git would tell it in O(1) time.
00:49 * dukeleto goes
00:49 whiteknight we should document those steps somewhere
00:50 cotto_work whiteknight: the question is where.  I made a note at the top of each .ops file but that was only because I couldn't think of something better.
00:51 whiteknight actually it doesnt matter where. i wouldnt see it if it were tattood on my face
00:52 cotto_work A bad but effective idea would be to add it in a comment to the beginning of each op.
00:52 whiteknight i thought i did make bootstrap-ops
00:53 whiteknight had to for the previous release
00:53 cotto_work Running bootstrap-ops now results in finalize moving to a different place in the op ordering.
00:53 whiteknight hmm
00:54 cotto_work either way, I'll do it to tonight
00:54 whiteknight cotto++
00:54 cotto_work whiteknight: did you ever get a functional laptop?
00:55 whiteknight nope
00:56 cotto_work You're being pretty productive for not having a laptop.
01:12 NotFound_b The Float cmp_p_p equality test fails for me in --optimize g++ build.
01:13 whiteknight so dont do it
01:14 NotFound_b A bit I don't understand: the comment says cmp_p_p but it does cmp_num, not cmp.
01:14 whiteknight problem solved
01:15 PerlJam left #parrot
01:15 PerlJam joined #parrot
01:16 davidfetter left #parrot
01:16 pmichaud left #parrot
01:17 dmalcolm left #parrot
01:20 pmichaud joined #parrot
01:22 cotto_work NotFound_b: trying to repro
01:23 mikehh I got the same problem with an g++ --optimized build, ok without --optimize
01:26 cotto_work NotFound_b: t/pmc/float.t?
01:26 NotFound_b cotto_work: yeah
01:27 NotFound_b The String get_number gives a result that differs aprox. 3e-15 from the expected vaue
01:29 cotto_work same result with cmp instead of cmp_num
01:29 cotto_work NotFound_b: do you think there should be an epsilon in there?
01:31 NotFound_b cotto_work: I don't think so. The pir compiler should use the same function to get the float value than the string PMC, thus the result should be equal.
01:42 KatrinaTheLamia left #parrot
01:45 cotto_work NotFound_b: is this something we can fix before the release or do we need to selectively TODO it?
01:46 NotFound_b cotto_work: not sure, I not even know were the problem lies. It can even be a compiler bug.
01:46 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#3504) fulltest) at 2_11_0-1176-gea2a249 - Ubuntu 10.10 amd64 (g++-4.5)
01:47 mikehh I find it strange that it passes without --optimize nut fdails with
01:47 mikehh but
01:49 cotto_work NotFound_b: it sounds like we should selectively TODO the test.
01:49 mikehh gah - had a Microsoft 1000 wireless desktop, now have a Microsoft 5000 wireless desktop - the keys are differently placed
01:49 NotFound_b cotto_work: and then maybe we get a todo passing report from any other platform
01:50 mikehh you can check the compiler and if optimize is used
01:50 cotto_work mikehh: exactly
01:52 mikehh anyway, doing a test run on amd64, g++ --optimize
01:52 mikehh everything ok without it
01:52 cotto_work mikehh: it's x86 too
01:53 mikehh I think that is where I found the problem
01:53 NotFound_b I'm using i386
01:53 mikehh so, let's see what happens on 64 bit
01:53 KatrinaTheLamia joined #parrot
01:55 mikehh I did comment on it yesterday, so should be in the #parrot logs
01:57 cosimo left #parrot
01:58 mikehh t/pmc/float.t - Failed test:  102 with g++ configured with --optimize (passes without and with gcc) on both Ubuntu 10.04 i386 and Ubuntu 10.10 i386 (g++ 4.4.3 and 4.5.1)
02:01 mikehh that was 2011-01-16 - 19:30 (http://irclog.perlgeek.de/)
02:03 NotFound_b I have a possible fix, but is cargo cult programming.
02:03 cotto_work NotFound_b: is this a todo or an actual fix?
02:03 NotFound_b Fix
02:03 cotto_work TODoing this is really ugly
02:04 cotto_work NotFound_b: can you nopaste the fix?
02:04 whiteknight left #parrot
02:04 cosimo joined #parrot
02:05 nopaste "NotFound" at 192.168.1.3 pasted "Bizarre fix for Float cmp problem" (16 lines) at http://nopaste.snit.ch/28063
02:05 kid51 mikehh: Recently I have begun testing with --optimize --cc=g++ etc ... but I'm getting PASS.  As I think you mentioned, this may be a g++ version problem.
02:06 cotto_work NotFound_b: that fixes it?  *sigh*
02:06 NotFound_b cotto_work: passes for me in i386
02:06 cotto_work kid51: which gcc version?
02:06 cotto_work NotFound_b: which gcc version?
02:06 cotto_work mikehh: you too
02:06 NotFound_b 4.4.5
02:07 cotto_work same here
02:07 cotto_work if that's what "gcc (Ubuntu/Linaro 4.4.4-14ubuntu5) 4.4.5" means
02:08 mikehh I tested with g++ 4.4.3 and 4.5.1, kid51 passed with an earlier version
02:08 cotto_work mikehh: same result for both?
02:09 mikehh yes - see above (from the logs)
02:09 cotto_work NotFound_b: that fixes it for me
02:09 kid51 On Linux, I have:  g++ --version
02:09 kid51 g++ (Debian 4.3.2-1.1) 4.3.2
02:09 kid51 Copyright (C) 2008
02:09 cotto_work mikehh: moritz's irclogs let you link to a specific line (for future reference)
02:09 NotFound_b We got the Cargo! Joe bless us!
02:10 cotto_work fbrito++ for adding the tests that detected that bit of inanity
02:11 cotto_work I'll commit it.
02:11 cotto_work NotFound_b: where'd you find that fix?
02:11 kid51 With the patch, t/pmc/float.t continues to pass for me on linux/i386 configured with cc=g++ (4.3.2)
02:12 cotto_work kid51: thanks
02:13 cotto_work It's times like this that I'm grateful for ccache, or at least I would be if I'd bothered to installed it on this machine.
02:13 NotFound_b cotto_work: just by remembering ugly ways of avoiding old g++ bugs years ago.
02:13 cotto_work wheee
02:14 kid51 So I think this is a g++ version thing ... which means it's way above my paygrade
02:14 NotFound_b One time I even created a non const global var with zero value to make == ZERO work in optimized bulds.
02:15 cotto_work notfound++
02:15 kid51 is now known as kid51_at_dinner
02:16 dalek parrot/gcc-float-fix: a69afcc | cotto++ | src/pmc/float.pmc:
02:16 dalek parrot/gcc-float-fix: fix an odd bug that resulted in some lost precision in the Float PMC
02:16 NotFound_b Usually people fails to recognize that kinds of bugs, they fall to the standard answer "never compare floats for exact equality"
02:16 dalek parrot/gcc-float-fix: review: https://github.com/parrot/parrot/commit/a69afcc939
02:17 mikehh from http://irclog.perlgeek.de/parrot/2011-01-16:
02:17 mikehh 16:12 mikehh I am getting a failure in t/pmc/float.t (g++ --optimize) - Failed test:  102 - comparison ops: cmp_p_p: equality (passes g++ without --optimize and gcc - Ubuntu 10.10 i386)
02:17 mikehh 16:45 kid51  mikehh: Unable to reproduce that on an older Debian, with all other arguments same as yours
02:17 mikehh 16:47 mikehh kid51: looking at it - it works with g++ (no --optimize) and gcc with or without --optimize
02:17 mikehh 16:49        I am using gcc/g++ 4.5.1
02:17 cotto_work NotFound_b: I want to test with msvc before I throw it into trunk.
02:17 mikehh ...
02:17 mikehh 19:30 mikehh t/pmc/float.t - Failed test:  102 with g++ configured with --optimize (passes without and with gcc) on both Ubuntu 10.04 i386 and Ubuntu 10.10 i386 (g++ 4.4.3 and 4.5.1)
02:18 NotFound_b cotto_work: ok
02:18 cotto_work It'll probably be fine, but I want to be really sure.
02:19 NotFound_b cotto_work: it's all yours, I must sleep now.
02:19 cotto_work NotFound_b: 'night
02:22 mikehh cotto_work: there are a bunch of files that should be in .gitignore - they are removed by make realclean - let me nopaste
02:22 * cotto_work wonders if that was really NotFound's evil twin
02:22 NotFound_b left #parrot
02:23 cotto_work mikehh: do you mind filing a ticket on this Float bug?  I want something more than irclogs to keep a record of it.
02:23 nopaste "mikehh" at 192.168.1.3 pasted "files to go in .gitignore" (19 lines) at http://nopaste.snit.ch/28066
02:23 cotto_work affected gcc versions, the fix and the gist of the conversation would be fine
02:25 dalek parrot: a69afcc | cotto++ | src/pmc/float.pmc:
02:25 dalek parrot: fix an odd bug that resulted in some lost precision in the Float PMC
02:25 dalek parrot: review: https://github.com/parrot/parrot/commit/a69afcc939
02:28 cotto_work mikehh: fixed
02:28 dalek parrot: 23156aa | cotto++ | .gitignore:
02:28 dalek parrot: add some generated files to .gitignore
02:28 dalek parrot: review: https://github.com/parrot/parrot/commit/23156aa71c
02:33 KatrinaTheLamia left #parrot
02:42 mikehh cotto_work: just as a matter of interest - I ran the tests before the patch and it passes (t/pmc/float.t) 64 bit g++ --optimize
02:44 mikehh but mind you it uses a different float type (not the co-processor i387 as does 32 bit)
02:45 mikehh anyway testing with the patches now
02:48 cotto_work mikehh: same gcc version?
02:50 mikehh cotto_work: yes
02:50 mikehh cotto_work: MANIFEST.SKIP needs to be re-generated
02:50 cotto_work mikehh: in the future (and especially before a release), please be sure that someone like whiteknight, dukeleto or me notices failure reports.  Don't feel bad if you need to repeat yourself or send a msg.
02:51 mikehh I done think I got the next release
02:51 cotto_work done
02:51 dalek parrot: eff8e03 | cotto++ | MANIFEST.SKIP:
02:51 dalek parrot: regenerate MANIFEST.SKIP; mikehh++ for noticing
02:51 dalek parrot: review: https://github.com/parrot/parrot/commit/eff8e03640
02:54 NotFound left #parrot
02:58 NotFound joined #parrot
03:07 fbrito joined #parrot
03:19 cotto_work left #parrot
03:20 contingencyplan joined #parrot
03:21 cotto_work joined #parrot
03:28 vmspb left #parrot
03:28 kid51_at_dinner How TF did whiteknight manage to take a ticket at the instant I created it?
03:29 kid51_at_dinner is now known as kid51
03:30 kid51 Has he got some hook set up to automatically assign all tickets to him?
03:30 dalek TT #1976 created by jkeenan++: manifest-related files still reference Subversion
03:30 dalek TT #1976: http://trac.parrot.org/parrot/ticket/1976
03:31 bacek_at_work kid51, "component owner"?
03:31 kid51 What is that?
03:32 kid51 Does that mean that if I designate a component on a brand-new ticket, that ticket can automatically be assigned to someone designated as the 'owner' of that component?
03:33 kid51 Hmm, I guess so.
03:34 kid51 And, much to my surprise, *I* am designated the component owner of coding_standards.
03:34 KatrinaTheLamia joined #parrot
03:34 kid51 And, you, dear bacek, are the component owner for pirate!
03:35 bacek_at_work kid51, no idea ;)
03:35 kid51 I'm not sure I like this.  About half of the "component owners" are not currently active committers.
03:36 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#3527) fulltest) at 2_11_0-1179-geff8e03 - Ubuntu 10.10 amd64 (g++-4.5 with --optimize)
03:37 kid51 make release!
03:37 kid51 tag it and ship it!
03:38 * kid51 must sleep
03:38 kid51 left #parrot
03:47 cotto ~~
03:51 mikehh hey cotto
03:51 mikehh tests ok for me - up to fulltest on 64 bit, with or without --optimize
03:52 cotto good news there
03:54 mikehh only had that one problem with 32 bit, but hopefully that is fixed now
03:54 cotto mikehh, it is as far as my testing indicates
03:54 cotto and it didn't make anything worse
03:55 mikehh I will test later, it's approaching 4am for me, think I need some sleep
04:41 cotto mikehh, good idea
05:49 rurban_ joined #parrot
05:52 rurban left #parrot
05:52 rurban_ is now known as rurban
06:16 davidfetter joined #parrot
06:25 davidfetter left #parrot
06:26 davidfetter joined #parrot
06:34 Patterner left #parrot
06:35 Psyche^ joined #parrot
06:35 Psyche^ is now known as Patterner
06:38 Tene left #parrot
06:43 Tene joined #parrot
06:50 simcop2387 left #parrot
07:11 dukeleto ~~
07:14 cotto hi dukeleto
07:15 cotto dukeleto, git log ..origin/master doesn't show any changes, even though I have a couple unpushed commits.
07:17 simcop2387 joined #parrot
07:20 dukeleto cotto: hmm
07:20 dukeleto cotto: what about git log origin/master.. ?
07:22 dukeleto cotto: does git log --graph --all show you anything ?
07:23 cotto dukeleto, that does it
07:23 cotto origin/master..
07:23 cotto I'll fix that now
07:24 theory left #parrot
07:24 moritz good morning
07:24 cotto hi moritz
07:29 * cotto wonders why his laptop makes a noise when gcc compiles a file
07:30 moritz hiccup? :-)
07:31 cotto It's a distinct sounds, almost like a click.  It shouldn't be hdd access because I have an ssd.
07:34 Topic for #parrot is now Parrot 2.11.0 Released | http://parrot.org | Log: irclog.perlgeek.de/parrot/today | Goals: write GettingStartedWithPlumage, review html documentation, test HLLs, review deprecations | RELEASE IN PROGRESS - no commits to master
07:43 dukeleto cotto: the difference between ..origin/master and origin/master.. is (A-B) vs (B-A), where A and B are sets
07:43 dukeleto cotto: so if origin is ahead, one of those will show a difference, but the other won't
07:44 dukeleto cotto: when local master is ahead, the other comittish shows the difference
07:44 dukeleto cotto: makes sense?
07:44 cotto yup
07:44 * cotto has encountered a yak
07:44 dukeleto cotto: perhaps the docs should just say "git diff origin/master"
07:45 dukeleto cotto: that doesn't care about set differences, if that commands returns something, then local master and origin/master are not in sync, IF "git fetch" has been run recently
07:45 cotto dukeleto, wfm
07:45 dukeleto cotto: what flavor yak did you find?
07:46 cotto a stupid one that smells fairly hairy
07:46 cotto trouble with tools/dev/mk_native_pbc
07:46 dukeleto cotto: i saw your float pmc commit, that is quite interesting
07:46 dukeleto cotto: what kind of trouble?
07:46 cotto dukeleto, that was notfound's
07:46 cotto but yes
07:46 dukeleto cotto: do you need somebody to generate native pbc's again?
07:46 cotto src/nci/libffi.c:274: error: ‘ffi_type_parrot_numval’ undeclared (first use in this function)
07:47 cotto breaks with Configure.pl "--floatval=long double"
07:48 cotto dukeleto, any idea what to do about that?
07:48 dukeleto cotto: that --floatval option is NEVER tested, as far as I can see
07:49 dukeleto cotto: i don't know that i consider that config arg 'supported'
07:49 dukeleto cotto: it would be nice to, but we need to say that, and then test against it
07:49 cotto dukeleto, it's run by mk_native_pbc
07:52 dukeleto cotto: huh
07:52 dukeleto cotto:  so every time mk_native_pbc is run, it gives Configure.pl those flags?
07:52 cotto that seems to be the case
07:52 dukeleto cotto: but that is never triggered in the test suite?
07:53 cotto guess not
07:54 dukeleto cotto: blarg
07:54 cotto I'm not sure I like that script.
07:54 cotto dukeleto, are you on a 64-bit machine?
07:54 dukeleto cotto: i can be
07:55 * dukeleto is on the train right now on a 32bit netbook, but have ssh access to various 64 bit boxen
07:55 dukeleto cotto: perhaps you need to give some --with-extra-nci-thunk junk to the script ?
07:56 cotto blargh
07:57 cotto dukeleto, can you see if tools/dev/mk_native_pbc works on a 64-bit box?
07:57 dukeleto cotto: will check when i get home. what exactly do you want me to do?
07:58 dukeleto cotto: you gonna be up for a bit?
07:58 cotto just configure Parrot normally and run sh tools/dev/mk_native_pbc
07:58 cotto dukeleto, probably for at least an hour or so
07:58 * dukeleto will be back soon
07:58 cotto I want to kick the release out before bed.
07:58 cotto eta?
08:03 gg411 joined #parrot
08:03 AndChat| joined #parrot
08:09 gg411 left #parrot
08:21 dukeleto cotto: back on a 64 bit machine
08:21 * dukeleto updates parrot.git and builds a fresh parrot
08:23 cotto good times
08:23 cotto thanks
08:25 dukeleto cotto: it seems to be running
08:26 cotto dukeleto, it takes a couple cycles to fail on my box
08:26 cotto lmk if it completes
08:27 dukeleto cotto: will do, it is warming up my desk as we speak
08:27 cotto it'll do that
08:28 dukeleto cotto: failed
08:28 cotto gah
08:28 cotto h8h8h8
08:28 dukeleto cotto: https://gist.github.com/784135
08:28 dukeleto cotto: yeah, this is no fun. Did something in ffi change since the last release?
08:28 cotto same thing I see, more of less
08:29 cotto If only git exposed some kind of log that I could examine.
08:29 dukeleto yeah, they should add that
08:29 cotto looks like essentially nothing changed since the last release
08:30 particle left #parrot
08:30 dukeleto cotto: src/nci/libffi.c:38:4: error: #error "unhandled NUMVAL_SIZE value"
08:30 dukeleto cotto: is that the first actual error?
08:31 dukeleto cotto: there is nothing wrong with post-poning the release a day to figure this out. it is much better than rushing and getting something wrong
08:31 cotto I didn't keep track of the full output
08:32 cotto I'll see if I can get something to work.
08:32 cotto in the meantime, this: http://www.jwz.org/blog/2011/01/what-if​-i-didnt-want-to-buy-the-potion-indeed/
08:34 cotto fulltest passes
08:36 dukeleto cotto: which files got modified when you ran it?
08:36 dukeleto i am seeing (integer|number|string)_4.pbc
08:36 cotto a bunch in t/native_pbc
08:36 cotto _1.pbc
08:38 dukeleto cotto: mine only changed 3 files
08:38 cotto 6 here
08:38 dukeleto fulltest bombs out for me with the same src/nci/libffi.c:38:4: error: #error "unhandled NUMVAL_SIZE value"
08:39 cotto same here
08:40 dukeleto cotto: gerd just replied to you on parrot-dev
08:40 * dukeleto is tired and have epsilon/2 tuits left tonight
08:41 cotto dukeleto, no worries.
08:42 cotto dukeleto, I don't see any response
08:44 cotto dukeleto, I may have a fix.
08:52 cotto and it's fixed
08:54 cotto dukeleto, I'm still not seeing anything on parrot-dev.
08:59 cotto I guess you're not conscious.  'night
09:25 fbrito left #parrot
09:26 dalek parrot: 974663a | cotto++ | docs/pdds/ (2 files):
09:26 dalek parrot: [pdd] kick pdd10 out of draft
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/974663a43b
09:26 dalek parrot: 68d24d0 | cotto++ | NEWS:
09:26 dalek parrot: update and polish NEWS for 3.0.0
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/68d24d0f52
09:26 dalek parrot: 59ec665 | cotto++ | CREDITS:
09:26 dalek parrot: update CREDITS to be less untrue
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/59ec66537e
09:26 dalek parrot: 994be38 | cotto++ | / (7 files):
09:26 dalek parrot: various updates to prepare for 3.0.0
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/994be38f77
09:26 dalek parrot: 3f1765e | cotto++ | MANIFEST:
09:26 dalek parrot: manifix after kicking pdd10 out of draft
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/3f1765ee55
09:26 dalek parrot: 1fc3a3f | cotto++ | / (4 files):
09:26 dalek parrot: bootstrap ops to 3.0.0
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/1fc3a3fe0f
09:26 dalek parrot: e97ed5a | cotto++ | PBC_COMPAT:
09:26 dalek parrot: bump PBC_COMPAT for 3.0.0
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/e97ed5a050
09:26 dalek parrot: add0c97 | cotto++ | PBC_COMPAT:
09:26 dalek parrot: fix PBC_COMPAT formatting
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/add0c97f2d
09:26 dalek parrot: 01cefe5 | cotto++ | src/nci/libffi.c:
09:26 dalek parrot: fix mk_native_pbc
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/01cefe5d76
09:26 dalek parrot: 8c3d3b4 | cotto++ | t/native_pbc/ (6 files):
09:26 dalek parrot: update native pbc files
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/8c3d3b4066
09:26 dalek parrot: 502d035 | cotto++ | t/manifest/01-basic.t:
09:26 dalek parrot: skip the correct number of tests in a release tarball
09:26 dalek parrot: review: https://github.com/parrot/parrot/commit/502d035ad7
09:38 treed left #parrot
09:38 treed joined #parrot
09:40 treed left #parrot
09:40 treed joined #parrot
09:40 treed left #parrot
09:40 treed joined #parrot
09:40 treed left #parrot
10:03 dalek website: cotto++ | Parrot 3.0.0 "Beef Stew" Released!
10:03 dalek website: http://www.parrot.org/news/2011/parrot-3.0.0
10:09 gerd joined #parrot
10:13 Topic for #parrot is now Parrot 3.0.0 Released | http://parrot.org | Log: irclog.perlgeek.de/parrot/today | Goals: write GettingStartedWithPlumage, review html documentation, test HLLs, review deprecations | merge at will
10:14 mikehh cotto++
10:17 cotto aloha, clock?
10:17 aloha cotto: LAX: Tue, 02:17 PST / CHI: Tue, 04:17 CST / NYC: Tue, 05:17 EST / UTC: Tue, 10:17 UTC / LON: Tue, 10:17 GMT / BER: Tue, 11:17 CET / TOK: Tue, 19:17 JST / SYD: Tue, 21:17 EST
10:18 cotto good morning, mikehh
10:19 mikehh cotto: well done on the release, looks like you sorted out a lot of problems there
10:20 cotto I'm very ready for bed.
10:21 mikehh cotto: sleep well, cu in #ps
10:22 cotto 'night
10:22 dalek parrot: df5403b | cotto++ | docs/project/release_manager_guide.pod:
10:22 dalek parrot: don't insult the release manager after at the end of the process
10:22 dalek parrot: review: https://github.com/parrot/parrot/commit/df5403b412
10:23 gerd cotto++; nice to have the release so early on the day
10:23 cotto That commit message is actually a pretty good indicator of my mental state.
10:24 cotto definitely time to limit the damage.
10:24 * cotto out
10:37 gerd left #parrot
11:07 vmspb joined #parrot
11:25 Coke cotto++ # ooh, even docs.parrot.org is updated.
11:49 Coke msg dukeleto any idea where smolder is installed? is it a box I have access to?
11:49 aloha OK. I'll deliver the message.
11:55 whiteknight joined #parrot
11:56 contingencyplan left #parrot
12:01 dalek parrot: aad4e61 | Coke++ | docs/project/release_manager_guide.pod:
12:01 dalek parrot: minor nit on docs.parrot.org release.
12:01 dalek parrot: review: https://github.com/parrot/parrot/commit/aad4e61a9a
12:09 cogno joined #parrot
12:12 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#3612) fulltest) at 3_0_0-1-gdf5403b - Ubuntu 10.10 amd64 (g++-4.5 with --optimize)
12:22 mikehh rakudo (388eed7) - builds on parrot (3_0_0-1-gdf5403b)- make test, make spectest_smolder[(#3626), roast (d96889b)] PASS - Ubuntu 10.10 amd64 (g++-4.5 with --optimize)
12:24 Coke t/tools/mk_language_shell.t fails.
12:28 mikehh Coke: not for me
12:34 mikehh Coke: btw I am getting three couldn't find messages in t/src/checkdepend.t  - checked # couldn't find /usr/include/malloc.h, included from src/gc/malloc.c - should be ok
12:36 mikehh - #  include "/usr/include/malloc.h" on line 854 of src/gc/malloc.c -> /usr/include/malloc.h exists
12:37 Coke mikehh: it's got an explicit skip for darwin.
12:38 cogno left #parrot
12:38 Coke if you invert the conditional, you can probably duplicate the error.
12:38 Coke mikehh: those have been complaining since nwell* changed things a few weeks ago.
12:41 mikehh Coke: yeah, was just wondering, try and follow up on messages that don't cause the test to fail
12:45 cogno joined #parrot
13:17 cogno left #parrot
13:24 cogno joined #parrot
13:29 gg411 joined #parrot
13:30 atrodo cotto++ # Release announcement
13:33 AndChat| left #parrot
13:34 gg411 left #parrot
13:39 whiteknight release announcement has a lot of extra snark this time around. cotto++
13:40 davidfetter left #parrot
13:40 whiteknight left #parrot
13:42 davidfetter joined #parrot
13:47 cogno left #parrot
13:47 Coke I can't build rakudo after installing 3.0.0
13:48 Coke Reading configuration information from parrot_config ...
13:48 Coke PARROT VM: Could not initialize new interpreterPackFile_Header_validate: This Parrot cannot read bytecode files with version 10.0.
13:48 Coke it's because I installed 3.0.0 into the same location where I had originally installed 2.11.0-devel
13:49 Coke the installed parrot_config gets confused. :(
13:49 mikehh Coke: it built fine for me with Ubuntu 10.10 amd64
13:49 rurban_ joined #parrot
13:50 mikehh I usually remove the installed stuff before building parrot, or use a different --prefix on Configure
13:50 cogno joined #parrot
13:50 Coke right. it's fine if you start with a clean install directory.
13:50 Coke but the fact that you need to do that is a bug, IMO.
13:50 atrodo Coke++ I agree
13:51 rurban left #parrot
13:51 rurban_ is now known as rurban
13:52 mikehh Coke: I also agree on that, have had quite a few problems if I did not remove it
13:52 fperrad joined #parrot
13:53 * mikehh need to go out, no milk left :-}
13:55 Coke ticket opened.
13:57 ambs joined #parrot
13:59 dalek TT #1977 created by coke++: parrot requires fresh install directory
13:59 dalek TT #1977: http://trac.parrot.org/parrot/ticket/1977
14:00 cogno left #parrot
14:11 plobsing left #parrot
14:19 particle joined #parrot
14:34 plobsing joined #parrot
14:36 mtk joined #parrot
15:04 mberends left #parrot
15:16 mberends joined #parrot
15:26 plobsing left #parrot
15:27 sorear left #parrot
15:30 p6eval left #parrot
15:31 dalek left #parrot
15:33 sorear joined #parrot
15:33 TimToady left #parrot
15:33 TimToady joined #parrot
15:36 dalek joined #parrot
15:40 p6eval joined #parrot
15:41 cogno joined #parrot
15:42 whiteknight joined #parrot
15:44 whiteknight good morning, #parrot
15:45 hudnix left #parrot
15:46 hudnix joined #parrot
15:52 fbrito joined #parrot
15:56 cogno left #parrot
15:57 theory joined #parrot
16:08 cogno joined #parrot
16:12 dalek TT #1978 created by mikehh++: possible problem with g++ --optimize (t/pmc/float.t failure)
16:12 dalek TT #1978: http://trac.parrot.org/parrot/ticket/1978
16:15 fbrito1 joined #parrot
16:16 fbrito left #parrot
16:34 plobsing joined #parrot
16:40 dukeleto ~~
16:41 whiteknight good morning, dukeleto
16:41 dukeleto Coke: smolder is on the parrotvm box at OSUOSL. That is all I know.
16:45 Coke dukeleto: danke.
16:46 Coke ah, found it.
16:46 dukeleto Coke: please document what you find
16:46 dukeleto Coke: i am still having a problem getting nqp-rx and tre-optimization to submit smolder reports
16:47 dukeleto Coke: so if you find any useful error logs, let me know. I want to see them.
16:48 Coke found the sqlite db, that's about it.
16:48 dukeleto cotto++ # awesome job on the release, the announcement, and fixing mk_native_pbc
16:50 Coke dukeleto: when did you last get an error?
16:52 Coke dukeleto: I see no log entries on the last 2 days for report/8 (nqp-rx)
16:54 dukeleto Coke: that is the bug. Smolder returns 200 and the report never shows up. Very frustrating.
16:54 dukeleto Coke: what about in tree-optimization
16:56 dukeleto Coke: i see you had the pleasure of opening the first github issue on my Util.git repo. Congrats ;)
16:57 dukeleto Coke: i agree, and would gladly accept a patch
16:57 Coke dukeleto: found the log. looks like something is dying trying to email directors@parrot.org
16:58 Coke (patch) well, the easiest thing locally is for me to hardcode it too. ;)
16:58 dukeleto Coke: sad face!
16:59 dukeleto Coke: when i was writing that stuff i wasn't taking into account others using it, but it wouldn't be hard to document it a bit and have a doc about "how to setup a smoker"
16:59 Coke can you try to do another upload for nqp-rx for me?
16:59 dukeleto ** POST http://smolder.parrot.org/app/​projects/process_add_report/9 ==> 200 OK
16:59 dukeleto Coke: i just submitted a tree-optimization report
16:59 cogno left #parrot
17:00 Coke dukeleto: once more?
17:00 dukeleto Coke: i am not sure if the bug in nqp-rx is on my side or smolder, but tree-optimization's smolder stuff is using setup.pir, so it *should* work
17:00 dukeleto Coke: en route
17:01 Coke did it hit?
17:01 Coke (just saw the log timestamp change)
17:01 dukeleto Coke: yep, just sent it
17:01 dukeleto Coke: it gave me the same 200 junk
17:02 Coke I think it might be the email config. let me see if I can find the per project config.
17:02 Coke ... there's a web GUI admin? haz you the password?
17:03 Coke I imagine it's much easier to config that way than rooting about on the command line
17:04 Themeruta joined #parrot
17:05 whiteknight I would really like to see plobsing's branch get merged before anything else
17:05 whiteknight that's a really imporant branch
17:05 NotFound is now known as NotFound_idle
17:06 Themeruta is now known as NotFound
17:06 dukeleto Coke: yes
17:08 dukeleto plobsing: ping!
17:08 dukeleto plobsing: we are going to merge your branch for you, if you don't do it soon ;)
17:08 dukeleto whiteknight: i agree
17:08 dukeleto whiteknight: i want that branch to have the smallest chance of annoying conflicts
17:10 whiteknight I am not going to merge either of my branches until he does his, but I am going to update them now to current master
17:13 tadzik o/
17:13 dukeleto tadzik: wazzup
17:13 dukeleto tadzik: you gonna be able to hack on deprecations as data anytime soon?
17:13 tadzik dukeleto: oh, loads of Uni work
17:13 tadzik dukeleto: yeah, I scheduled that on fri-sat
17:14 dukeleto tadzik: awesome!
17:14 dukeleto tadzik: good luck with school. what classes are you taking? Any good math classes?
17:15 tadzik dukeleto: discrete mathematics and analysis. I'm having a DM exam tomorrow
17:15 tadzik Not a final yet, mainly from recursive equations
17:16 dukeleto tadzik: oh boy, sounds like fun :) are you swimming in Hilbert space?
17:17 tadzik coming soon this week: physics, fundamentals of electronics and fundamentals of digital techniques (or how is it named. Asynchronous FSMs this time)
17:17 tadzik dukeleto: never heard of the guy :)
17:18 * dukeleto can't tell if he is being trolled
17:20 dalek parrot/imcc_cleanups: 592b8ce | Whiteknight++ | compilers/imcc/ (2 files):
17:20 dalek parrot/imcc_cleanups: remove old vestige of IMCC -w handling. We do all that in the frontend now
17:20 dalek parrot/imcc_cleanups: review: https://github.com/parrot/parrot/commit/592b8ce31d
17:20 dalek parrot/imcc_cleanups: b347b0c | Whiteknight++ | compilers/imcc/Rules.in:
17:20 dalek parrot/imcc_cleanups: add rules for checkdepend, though it still fails and I can't figure out why
17:20 dalek parrot/imcc_cleanups: review: https://github.com/parrot/parrot/commit/b347b0c9c1
17:22 * dukeleto can see a bumper sticker on the horizon : "The Parrot Embed API does it in the frontend"
17:23 whiteknight ::facepalm::
17:24 dukeleto http://smolder.parrot.org/app​/projects/report_details/3683 <-- first smoke report from tree-optimization !!!
17:24 tadzik dukeleto: no really. Just wikipedied him, doesn't look familiar
17:24 tadzik oh, so it works?
17:25 dukeleto tadzik: it works! Coke++ figured out that the parrot-autobot user wasn't made a part of the new projects
17:25 dukeleto tadzik: http://mathworld.wolfram.com/HilbertSpace.html
17:25 tadzik cool :) My commit was valuable then
17:26 dukeleto tadzik: yes, it was!
17:26 tadzik awesome :)
17:27 kthakore Hello on this page http://www.parrot.org/news/2011/parrot-3.0.0
17:27 kthakore Can some one update  docs/pdd/pdd10_embedding.pod
17:27 kthakore to  docs/pdds/pdd10_embedding.pod
17:27 kthakore ?
17:28 dukeleto kthakore: nice catch
17:28 dukeleto can someone do that?
17:29 kthakore I could if I had access
17:29 kthakore but cotto was the one to add it
17:29 kthakore any way I have to get back to work
17:30 dukeleto kthakore: we will find someone to fix it. Thanks for letting us know. kthakore++
17:31 dukeleto bacek_at_work: interesting new failure in tree-optimization http://smolder.parrot.org/app​/projects/report_details/3683 on parrot master
17:32 sjn_ is now known as sjn
17:35 dalek TT #1975 closed by doughera++: pbc_dump and pbc_merge may crash on illegal options
17:35 dalek TT #1975: http://trac.parrot.org/parrot/ticket/1975
17:42 tadzik http://isparrotfastyet.com/ -- whoa, what's that dropdown?
17:42 dmalcolm joined #parrot
17:43 cotto Let's see some merging!
17:47 dukeleto MERGE MERGE MERGE
17:48 Coke (dropdown) that's /good/, right?
17:49 dukeleto if it is correct, yes, it is good
17:49 dalek parrot/imcc_cleanups: 0e3282c | Whiteknight++ | / (37 files):
17:49 dalek parrot/imcc_cleanups: Merge branch 'master' into imcc_cleanups
17:49 dalek parrot/imcc_cleanups: review: https://github.com/parrot/parrot/commit/0e3282cc09
17:51 tadzik Use of uninitialized value within %config in substitution iterator at Configure.pl line 179.
17:51 tadzik (3.0.0)
17:52 whiteknight tadzik: any idea which value?
17:53 Kristaba joined #parrot
17:57 tadzik whiteknight: no idea, let me look into the file
17:58 dukeleto tadzik: i see that error too
17:58 dukeleto tadzik: warning, rather
17:58 dukeleto whiteknight: it is looking for some config key that doesn't exist anymore, probably
17:58 dukeleto tadzik: that has been happening since before 3.0.0 as well
18:00 tadzik hmm, haven't noticed
18:00 tadzik oh wait, that's Rakudo's problem
18:02 jimmy joined #parrot
18:02 jimmy good morning
18:04 dalek Heuristic branch merge: pushed 47 commits to parrot/exception_backtraces by Whiteknight
18:04 jimmy is now known as JimmyZ
18:04 whiteknight good morning JimmyZ
18:04 tadzik good morning
18:05 JimmyZ good morning whiteknight
18:07 whiteknight I'm going to try updating pbc-backref to master now, to see how cleanly that goes
18:07 dukeleto whiteknight++
18:08 JimmyZ (anyone who makes parrot better)++
18:09 cotto_work ~~
18:10 dukeleto cotto_work: mornin'
18:10 bluescreen joined #parrot
18:14 contingencyplan joined #parrot
18:15 cotto_work hi JimmyZ
18:16 whiteknight okay, pbc-backref updates to master with no apparent issues
18:17 whiteknight builds without problem and all coretests pass
18:17 cotto_work excellent
18:19 cogno joined #parrot
18:19 dukeleto release the hounds!
18:19 whiteknight I don't want to do it without plobsing. It's his baby
18:21 cotto_work no reason not to let him do the honors
18:21 JimmyZ hello cotto_work
18:21 JimmyZ it's 2:21am here :)
18:22 cotto_work JimmyZ: that's about when I got to bed.
18:22 JimmyZ cotto_work: I will go to bed soon. hehe
18:24 cotto_work JimmyZ: I'd like to get that pull request dealt with.  What are you suggesting that WINVER be #defined to?
18:24 cotto_work https://github.com/parrot/parrot/pull/121
18:26 JimmyZ cotto_work: if I'm on windows 98, then defining WINVER 0x0500 is wrong
18:26 dukeleto #ps soon?
18:26 whiteknight dukeleto: 2 hours, I think
18:27 dalek parrot: 248434e | cotto++ | src/pmc/float.pmc:
18:27 dalek parrot: [pmc] add tt reference to NotFound++'s Float PMC fix
18:27 dalek parrot: review: https://github.com/parrot/parrot/commit/248434efa3
18:27 whiteknight 20:30 UTC, right?
18:28 cotto_work 2 hours
18:28 cotto_work JimmyZ: what should it be defined to?
18:30 JimmyZ cotto_work: and, WINVER was nerver used by parrot, _MEMORYSTATUSEX was already copied from winbase.h to parrot's win32/sysmen.c
18:31 dalek winxed: r731 | NotFound++ | trunk/winxedst1.winxed:
18:31 dalek winxed: add target named parameter to the compile method on the compreg'ed object
18:31 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=731
18:34 JimmyZ IIRC,  if WINVER < Windows2000 was introduced by fperrad++, but parrot did'nt define WINVER dynmically, and then is defined 0x400 by windows.h. that's why it's failing
18:35 JimmyZ err, 0x0400
18:40 JimmyZ I think the better way is check windows version, and then define WINVER dynmically. :), or remove these lines, which is bad, _MEMORYSTATUSEX doesn't work when < windows2000
18:42 cotto_work JimmyZ: can you suggest a patch to check the windows version more intelligently?
18:42 dalek winxed: r732 | NotFound++ | trunk/winxed_installed.winxed:
18:42 dalek winxed: initial support for -c -e options in installable compiler
18:42 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=732
18:42 dalek winxed: r733 | NotFound++ | trunk/pir/winxed_ (2 files):
18:42 dalek winxed: update installable files
18:42 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=733
18:44 dalek parrot: 10b4386 | plobsing++ | / (12 files):
18:44 dalek parrot: Merge branch 'pbc-backref'
18:44 dalek parrot: review: https://github.com/parrot/parrot/commit/10b43866f5
18:44 tadzik \o/
18:44 JimmyZ cotto_work: I don't know how to do it, but I can give it a try.
18:45 cotto_work JimmyZ: ok.  Unfortunately I can't help much beyond testing.
18:45 cotto_work thanks
18:45 JimmyZ you are welcome :)
18:45 * JimmyZ must sleep
18:45 cotto_work 'night
18:46 JimmyZ good night
18:46 JimmyZ left #parrot
18:48 whiteknight plobsing++ indeed
18:48 cotto_work plobsing++
18:49 fbrito1 left #parrot
18:50 cogno left #parrot
18:51 bluescreen left #parrot
18:56 dalek parrot: c327bca | cotto++ | lib/Parrot/Install.pm:
18:56 dalek parrot: [install] update the libparrot symlink when installing, fixing TT #1960
18:56 dalek parrot: review: https://github.com/parrot/parrot/commit/c327bca662
18:59 dalek TT #1960 closed by cotto++: make install doesn't update libparrot.so symlink.
18:59 dalek TT #1960: http://trac.parrot.org/parrot/ticket/1960
18:59 dalek TT #1977 closed by cotto++: parrot requires fresh install directory
18:59 dalek TT #1977: http://trac.parrot.org/parrot/ticket/1977
19:00 whiteknight Maybe for future releases we should prepare and release a "release candidate" a few days prior to the actual release
19:00 whiteknight that way we can shake out bugs that only appear in the release before we actually cut it
19:04 bluescreen joined #parrot
19:05 cotto_work whiteknight: that's not a bad idea, provided that creating the rc doesn't involve hours of testing in the same way the release does
19:06 cotto_work The fixes didn't end up being too difficult, but it was a bit disheartening that they were needed at all.
19:10 NotFound We can create a branch for the release, apply only fixes to it, and after realese done and tagged merge it to master.
19:11 cogno joined #parrot
19:12 whiteknight cotto_work: The benefit of the release candidate is that we can do the process early, do all the testing in a branch, get users and HLLs to test with it, and then we basically have a release ready-made when we want to cut it
19:12 whiteknight so you still do all the testing necessary, but you do it earlier
19:13 whiteknight We can cut a release branch pretty early, and people can be breaking master so long as the release branch is stable
19:15 dalek parrot: f99c01f | Whiteknight++ | / (14 files):
19:15 dalek parrot: Merge branch 'master' of github.com:parrot/parrot
19:15 dalek parrot: review: https://github.com/parrot/parrot/commit/f99c01f849
19:15 dalek TT #1978 closed by cotto++: possible problem with g++ --optimize (t/pmc/float.t failure)
19:15 dalek TT #1978: http://trac.parrot.org/parrot/ticket/1978
19:16 whiteknight when I merge imcc_cleanups everybody is going to have to reconfigure
19:16 whiteknight barring test failures, that's about 10 minutes from now
19:16 cotto_work whiteknight: I also kicked pdd10 out of draft.
19:16 whiteknight I saw that. Most awesome
19:18 hercynium joined #parrot
19:21 dukeleto plobsing++ # pbc-backref merge
19:23 whiteknight what does the output of t/src/checkdepent.t mean?
19:23 cotto_work it means that the makefile dependencies need fixing
19:24 whiteknight right. it says "got 'include/pmc/pmc_callcontext.h(2)' expected 'include/pmc/pmc_callcontext.h(1)'"
19:24 whiteknight so what's that error, and how do I fix it?
19:24 cotto_work I guess that means it's included twice.
19:24 whiteknight great
19:25 cotto_work whiteknight: is this in a branch?
19:25 whiteknight yeah, imcc_cleanups
19:28 cogno left #parrot
19:28 whiteknight I can't wait to get this branch merged so I can start on the next stage of the project
19:31 whiteknight ...and here it comes
19:31 dalek parrot: 7e33ed2 | Whiteknight++ | / (20 files):
19:31 dalek parrot: Merge branch 'imcc_cleanups'
19:31 dalek parrot: review: https://github.com/parrot/parrot/commit/7e33ed26d9
19:31 dalek parrot: a0d48ce | Whiteknight++ | compilers/imcc/Rules.in:
19:31 dalek parrot: fix dependencies in the imcc makefile
19:31 dalek parrot: review: https://github.com/parrot/parrot/commit/a0d48ceef6
19:31 KaeseEs joined #parrot
19:35 Kapace hahaha! cool "beef stew" reference to homestarrunner :D
19:35 Kapace well done, good work everyone!
19:36 cotto_work git++
19:36 cotto_work there's no way this would be so smooth with svn
19:41 Coke jimmyz: if you're on win98,  well, why?
19:43 ambs_ joined #parrot
19:43 ambs left #parrot
19:43 ambs_ is now known as ambs
19:47 dukeleto whiteknight++ && imcc--
19:47 dukeleto karma imcc
19:47 aloha imcc has karma of -7.
19:47 tadzik poor imcc
19:48 dalek winxed: r734 | NotFound++ | trunk/winxedst0.cpp:
19:48 dalek winxed: minimal support for named parameters in stage 0
19:48 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=734
19:49 tcurtis joined #parrot
19:51 cotto_work tcurtis: ping
19:51 tcurtis cotto_work: pong
19:51 cotto_work tcurtis: there's a test failure in tree-optimization (which has been moved under the parrot org).  Do you have the tuits to fix it?
19:56 tcurtis cotto_work: possibly. I likely won't have time to look into it much tonight, but immobilizing ankle injuries do free up one's schedule nicely. :)
19:57 KaeseEs what doc should i start looking at first if i want to dig into the gc but haven't done parrot work before?
19:58 cotto_work tcurtis: I'm not glad that you're injured, but I'll accept the benefits.
20:03 whiteknight KaeseEs: we have a docs/pdds/pdd09* document that describes the GC architecture
20:04 whiteknight but it's a little bit out of date since bacek has been doing so much work there
20:04 whiteknight bacek has two branches, one of which contains a pretty detailed description of the new algorithm he wants to implement
20:05 tcurtis cotto_work: It's minor, but makes walking/standing/going to class inconvenient.
20:06 whiteknight KaeseEs: https://github.com/parrot/parrot/bl​ob/generational_gc/src/gc/gc_gms.c
20:07 whiteknight KaeseEs: Unfortunately, it's one of those kinds of systems that you really need to just start digging into. We can answer questions that you have
20:07 Eclesia joined #parrot
20:07 Eclesia good evening
20:07 dukeleto KaeseEs: firstly, welcome to #parrot !
20:07 whiteknight hello Eclesia
20:07 dukeleto Eclesia: howdy
20:08 KaeseEs thanks whiteknight, dukeleto
20:08 cotto_work KaeseEs: also be aware that the algorithm described has not actually been implemented.
20:08 KaeseEs ah-ha.
20:08 whiteknight KaeseEs: I know that system pretty well. What kind of work do you want to do in it? Any particular questions you have to get you moving in the right direction?
20:09 davidfetter left #parrot
20:09 Eclesia are there syntaxe colorations tools who handle actions and grammar files ?
20:10 whiteknight Eclesia: None that I know of for the Perl6 variant
20:10 dalek parrot: 2a46de6 | NotFound++ | / (2 files):
20:10 dalek parrot: fix c++ build and add a few decorations
20:10 dalek parrot: review: https://github.com/parrot/parrot/commit/2a46de6c63
20:10 whiteknight Some syntax highlighters may handle Flex and Bison, but it's not common
20:10 dukeleto Eclesia: good question. I think a perl6 syntax color highlighter exists, which will mostly work for the actions and grammar files
20:11 Eclesia dukeleto: that would be great, you remember it's name ?
20:12 davidfetter joined #parrot
20:12 dukeleto Eclesia: i am attempting to find it
20:12 * Eclesia applaud for parrot 3.0 release (while I think about it)
20:13 dukeleto Eclesia: http://perlgeek.de/blog-en/perl​-6/syntax-highlighter.writeback is a bit old but could be useful
20:13 mtk left #parrot
20:13 kid51 joined #parrot
20:14 Eclesia dukeleto: thanks, I'll have to find a way to use that now ^^
20:14 KaeseEs whiteknight: at this stage i'm mostly just looking to get an overview of the system and see what sort of tasks i might be able to do (hopefully my classes this semester should leave me a decent amount of time to work on stuff)
20:14 fperrad left #parrot
20:14 dukeleto Eclesia: http://lastofthecarelessmen.blogspot.com​/2009/11/perl-6-syntax-highlighting.html
20:14 whiteknight KaeseEs: Okay. And you're specifically interested in GC?
20:15 dukeleto Eclesia: those are mostly for producing HTML output, but they are a start
20:15 dukeleto KaeseEs: what kind of programming language background do you have?
20:16 dukeleto Eclesia: http://search.cpan.org/dist/Padre-Plugin-Perl6/
20:17 dukeleto Eclesia: that is probably what you want, but it is for Padre, a pretty awesome GUI IDE
20:17 dalek nqp-rx/nom: 88e53e7 | tadzik++ | build/Makefile.in:
20:17 dalek nqp-rx/nom: Fix a whitespace-o in Makefile.in
20:17 dalek nqp-rx/nom: review: https://github.com/perl6/nqp-rx/commit/88e53e7efe
20:18 KaeseEs dukeleto: junior in cse at rpi, written a fair amount of C (mostly in the embedded space), done a little bit of sysprog, use perl 5 as my goto apps language.  GC seemed like an interesting and challanging task, although if folks think i wouldn't be useful there i could always hack at something else :)
20:19 whiteknight KaeseEs: no, GC is definitely interesting and challenging. Plus, we do need it done
20:19 whiteknight I just want to make sure it's something you really do want to work on before I start scaring you with details
20:19 Eclesia dukeleto: never heard about this ide before, I'm not surprised to see that it's a perl IDE lol.
20:21 Eclesia KaeseEs: maybe you could code the java language on parrot vm. to ease those who which to migrate (like me) :D
20:22 KaeseEs whiteknight: also i've heard that most of the hard bits of of implementing a modernish gc system (generational/tricolor/etc) boil down to graph problems, and staring that in the face will help me decide whether to get my concentration nest yeat in graph theory or hardware :V
20:22 KaeseEs s/nest yeat/next year/ # crumbs in keyboard, fml
20:23 mtk joined #parrot
20:24 bacek aloha, humans
20:24 whiteknight KaeseEs: Ah, so a life decision. GC is only really a graph in name only. That is, memory objects do form a graph, but you don't do anything fun with it like tracing edges, or applying a colorization, or anything
20:24 tadzik hey bacek
20:24 whiteknight basically you have a loop, and you loop over all the pointers in an array
20:24 bacek tadzik, hi
20:24 kid51 #parrotsketch in 6 minutes
20:24 whiteknight KaeseEs: if you want a better application of graphs, the MMD system or the register allocator might be better for you
20:25 dalek parrot: 096e9c6 | NotFound++ | t/src/embed/api.t:
20:25 dalek parrot: fix an embed test for c++ build and add linedirective to it
20:25 dalek parrot: review: https://github.com/parrot/parrot/commit/096e9c63f5
20:26 KaeseEs i'll look into that too.
20:28 whiteknight KaeseEs: actually, the generational algorithm is pretty graph-able. It basically involves sorting nodes from one large graph into various subgraphs depending on how often they are touched
20:28 whiteknight (as a radical oversimplification)
20:28 chromatic joined #parrot
20:28 Coke KaeseEs: you're at RPI? I'm class of '93.
20:29 KaeseEs didn't mjd graduate around then too?
20:29 Coke (and am probably about 30m south of you right now)
20:29 Coke we're about the same vintage, but didn't share classes.
20:29 Coke syntax highlighter... vim?
20:29 mikehh #ps time
20:30 Coke https://github.com/petdance/vim-perl
20:30 KaeseEs ah-ha.  i was going to ask him if he wanted to maybe give a talk in paul lalli's class about hop-type stuff, but apparently it isn't being offered any more :(
20:31 Coke is paul still around? used to see him on lily all the time, but not in a while.
20:32 KaeseEs i think so.  would have to check.  when i took his class his day job was doing some sort of bank software in 5.8.x on solaris i think in the vicinity of clifton park, so it's plausible that he hasn't gotten too far away :)
20:34 dalek parrot: 34dca63 | NotFound++ | t/src/embed/api.t:
20:34 dalek parrot: main lacked return on an embed test
20:34 dalek parrot: review: https://github.com/parrot/parrot/commit/34dca6333c
20:45 mtk left #parrot
20:46 mtk joined #parrot
20:51 bacek plobsing++ # pbc-backref!
20:52 dalek parrot: 09c60a0 | cotto++ | docs/project/release_manager_guide.pod:
20:52 dalek parrot: add whiteknight as the release manager for 3.3
20:52 dalek parrot: review: https://github.com/parrot/parrot/commit/09c60a0fd3
20:54 dalek TT #1978 reopened by doughera++: possible problem with g++ --optimize (t/pmc/float.t failure)
20:54 dalek TT #1978: http://trac.parrot.org/parrot/ticket/1978
20:54 mtk left #parrot
20:55 jdhore joined #parrot
20:55 jdhore left #parrot
21:03 fbrito joined #parrot
21:06 plobsing left #parrot
21:09 NotFound Why trac doesn't allow to add a comment if someone had added another while writing it? Look like a silly restriction to me.
21:10 cotto_work NotFound: because trac--
21:14 Eclesia left #parrot
21:21 jnthn Every time you think trac--, remember: "We could be using Microsoft's Team Foundation System" ;-)
21:21 * jnthn has a comparatively high opinion of Trac compared to that junk heap.
21:22 cotto_work jnthn: that's a helpful perspective.
21:22 NotFound jnthn: we couldn't, no one is going to pay the licenses ;)
21:23 jnthn No tool is perfect. Some are just more imperfect than others. :)
21:24 jnthn Question. Is Parrot's fixed size allocator API?
21:24 whiteknight jnthn: extending API maybe
21:24 whiteknight we don't really have a defined extending API other than "anything with PARROT_EXPORT is available for extenders"
21:24 jnthn If I want to allocate object bodies, which will tend to fall into various common sizes, is it a better choice than just mallocing?
21:24 whiteknight jnthn: but it is a pretty stable part of the system, so you can use it. It is better than malloc
21:25 whiteknight it's not automatically managed. No GC for fixed size chunks. So you need to trac them yourself
21:25 jnthn whiteknight: Sure. These are things that'll hang off the ->data pointer of PMCs.
21:25 whiteknight okay. If you can keep them under control you can definitely use them
21:26 jnthn OK. Where to look for docs?
21:26 jnthn The appropriate .h file will do as docs. ;-)
21:27 whiteknight include/parrot/gc_api.h is probably it
21:27 whiteknight they are used in src/pmc.c, among other places. That's where the pmc->data structure comes from in the first place
21:28 whiteknight and in src/hash.c now, I think
21:29 jnthn Parrot_gc_allocate_fixed_size_storage looks the part :)
21:29 whiteknight yeah, I think that's it
21:29 whiteknight and there's a _free_ variant of it too
21:29 jnthn yes, just found that :)
21:30 * jnthn will be sure to call it ;)
21:32 dalek parrot: ba669a8 | Whiteknight++ | / (2 files):
21:32 dalek parrot: Merge remote branch 'origin/gci_orderedhash_coverage'
21:32 dalek parrot: review: https://github.com/parrot/parrot/commit/ba669a8198
21:32 dalek parrot: 14f54bc | Whiteknight++ | / (4 files):
21:32 dalek parrot: Merge branch 'master' of github.com:parrot/parrot
21:32 dalek parrot: review: https://github.com/parrot/parrot/commit/14f54bccdb
21:33 plobsing_ joined #parrot
21:37 cotto_work atrodo: ping
21:37 atrodo cotto_work> pong
21:37 cotto_work atrodo: what are you thinking about wrt M0 hashes and arrays?
21:38 atrodo cotto_work> In what regard?
21:38 cotto_work what the syntax will look like and how structs will work
21:39 atrodo Currently, they are just PMC/Objects, accessed as such
21:39 atrodo as for structs, i have load/store ops
21:39 kid51 left #parrot
21:43 Andy joined #parrot
21:48 cotto_work atrodo: do you have an way of defining structs or is the structure implicit?
21:49 rurban_ joined #parrot
21:50 atrodo cotto_work> It's more implicit, like how C handles the structure of a struct/union.  lasm has a way of defining it, but that's a convenience more than a definition.
21:51 rurban left #parrot
21:51 rurban_ is now known as rurban
21:57 whiteknight left #parrot
22:13 dukeleto atrodo: so where are you on your lorito roadmap?
22:14 bacek_at_work dukeleto, (tree-optimizations) I have no idea why it's failed. I'm simple "customer" of this code :)
22:16 ambs left #parrot
22:19 dukeleto bacek_at_work: you brake it you buy it!
22:20 bacek_at_work dukeleto, nope. You moved it from tcurstis in broken state!
22:21 dukeleto bacek_at_work: well, at least I tried :)
22:21 cotto_work anyone know of a good set of tools for managing screen sessions?
22:21 bacek_at_work dukeleto, at least I fixed one of broken tests :)
22:21 bacek_at_work cotto_work, man screen?
22:21 cotto_work bacek_at_work: not awesome enough
22:23 PacoLinux cotto_work: you can try http://tmux.sourceforge.net/
22:24 cotto_work I mean tools to manage sessions started with screen.  I see a handful of projects on github, but I'm curious if anyone here has a favorite.
22:33 Coke cotto_work: ?
22:38 dukeleto cotto_work: what do you mean?
22:38 dukeleto cotto_work: i just have one screen session per machine
22:39 cotto_work dukeleto: I usually have a large number.  One for Parrot, one for Rakudo, etc
22:39 dukeleto cotto_work: you can use CTRL-a-" to name a screen terminal and CTRL-a-A shows a list of all the names (i have a list of the terms at the bottom of my screen screen as well)
22:39 * dukeleto hopes those kb shortcuts are correct
22:40 cotto_work The names are helpful, but it'd also be nice if there were a tool to tell me which process were running under all sessions.  I occasionally get into the situation where a vim session has a file open, but I don't recall which screen session it's under.
22:40 cotto_work I started to write a script, but I thought I'd try the lazy approach first.
22:40 dukeleto cotto_work: oh, i used to feel that pain. Then I went back to a single screen session :)
22:41 cotto_work I'd get claustrophobic.
22:42 bacek_at_work cotto_work, I've got 6 virtual screens on my laptop. 4 of them have terminal with few tabs and gvim running :)
22:42 dukeleto cotto_work: https://github.com/leto/Util/blob​/master/config/.bashrc.aliases#L1
22:42 dukeleto cotto_work: you could easily fiddle with that alias to give it a named argument
22:42 dukeleto cotto_work: currently i do: screenssh machinename
22:42 dukeleto cotto_work: but you could to : screenssh screenname machinename
22:42 dukeleto cotto_work: or somesuch
22:43 cotto_work I need to dig through your scripts and steal the useful-looking bits.
22:43 dukeleto cotto_work: Coke has started opening up github issues for my Util repo :)
22:43 cotto_work win
22:44 cotto_work that's awesome
22:44 dukeleto cotto_work: https://github.com/leto/Ut​il/blob/master/bin/screen
22:44 dukeleto cotto_work: that is my screen wrapper
22:44 cotto_work I've only got my laptop at home, so I'm sure some of the stuff won't be as usefult o me
22:45 dukeleto cotto_work: https://github.com/leto/Uti​l/blob/master/bin/setup_env
22:45 dukeleto cotto_work: you could tweak that script to setup your work env :)
22:45 dukeleto cotto_work: i have that script sitting somewhere accessible online, and I wget and execute it and it sets up my work environment on new boxes
22:46 dukeleto cotto_work: it is a pretty useful thing to have. I wrote it when I had to setup the smokers
22:46 cotto_work shiny.  I'm sure I'd do something similar if I ended up having to work on new machines frequently
22:57 dukeleto any more branches going to land today? You know you want to...
22:58 vmspb left #parrot
22:59 cotto_work git is handling them like a dream
23:03 dukeleto indeed. git++
23:10 cotto_work msg whiteknight Make sure to add some NEWS items about the stuff you merged.
23:10 aloha OK. I'll deliver the message.
23:12 dukeleto looks like we have some failures for plain parrots on master: http://smolder.parrot.org/app​/projects/report_details/3740
23:13 dukeleto PROTIP: a plain parrot is what the smokers call parrots with every --without-foo Configure.pl flag set
23:14 shuja joined #parrot
23:14 shuja left #parrot
23:17 dalek parrot: 004f53d | mikehh++ | compilers/imcc/main.c:
23:17 dalek parrot: add ASSERT_ARGS
23:17 dalek parrot: review: https://github.com/parrot/parrot/commit/004f53d91c
23:17 dalek parrot: 8f579d3 | mikehh++ | t/src/embed/api.t:
23:17 dalek parrot: fix perlcritic failure - make bareword filehandle lexical
23:17 dalek parrot: review: https://github.com/parrot/parrot/commit/8f579d3d21
23:18 plobsing_ left #parrot
23:31 bluescreen left #parrot
23:37 dalek parrot: 92c952e | dukeleto++ | docs/project/release_manager_guide.pod:
23:37 dalek parrot: [doc] Add release dates for the rest of the year
23:37 dalek parrot: review: https://github.com/parrot/parrot/commit/92c952e938
23:40 dalek parrot: 131c6c5 | dukeleto++ | docs/project/release_manager_guide.pod:
23:40 dalek parrot: [doc] Sign up for 3.9 release manager
23:40 dalek parrot: review: https://github.com/parrot/parrot/commit/131c6c5a0d
23:44 dalek parrot: 9f8e1de | cotto++ | docs/project/release_manager_guide.pod:
23:44 dalek parrot: "volunteer" cotto for a couple releases
23:44 dalek parrot: review: https://github.com/parrot/parrot/commit/9f8e1de1ac
23:47 hercynium left #parrot
23:49 dukeleto cotto must be a glutton for punishment ;)
23:54 cotto_work It's a good interval.
23:55 plobsing_ joined #parrot

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

Parrot | source cross referenced