Camelia, the Perl 6 bug

IRC log for #parrot, 2009-12-20

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:11 mikehh pre/post-config, make corevm/makecoretest, smoke (#31108) PASS - fulltest FAIL at at r43168 - Ubuntu 9.10 i386 (g++ with --optimize)
00:11 mikehh t/examples/pir.t - Failed test:  2 in examples_tests
00:11 mikehh all other tests PASS
00:26 dalek winxed: r292 | julian.notfound++ | trunk/winxedst1.winxed:
00:26 dalek winxed: implement for classic style and allow break and continue in for in, in stage 1
00:26 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=292
00:40 dalek parrot: r43169 | jkeenan++ | branches/tt473_remove_memcpy_aligned (8 files):
00:40 dalek parrot: Schedule files for deletion from repository.  Regenerate MANIFEST so that 'make manifest_tests' will pass.
00:40 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43169/
01:09 cotto Is there any sense in using a ManagedStruct or UnmanagedStruct for PackFiles?  There's a TODO from Infinoid added at r35191 that mentions the idea.
01:14 dalek winxed: r293 | julian.notfound++ | trunk/examples/xlibtest.winxed:
01:14 dalek winxed: bug fix in example
01:14 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=293
01:29 dalek parrot: r43170 | mikehh++ | trunk/examples/pir/quine_ord.pir:
01:29 dalek parrot: fix t/examples/pir.t - test 2 by modifying examples/pir/quine_ord.pir
01:29 dalek parrot: (run ../../parrot quine_ord.pir --print-code-as-data and paste output over previous data)
01:29 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43170/
01:31 kid51 mikehh:  How the hell did you know to change exactly those lines?
01:34 Whiteknight cotto: I don't know, weren't there going to be dedicated PMC types for that?
01:35 mikehh kid51: there is a little documentation in the file (at the end) and a reference which I read through
01:35 kid51 Thanks!
01:38 mikehh I generated the output from the file and compared it with kdiff3 and found that change and re-reading the documentation worked out what to do
01:39 mikehh It took me a little while :-}
01:40 kid51 That file was so big it made me dizzy!
01:40 dalek TT #1384 closed by jkeenan++: t/examples/pir.t: test #2 (quine_ord) failing
01:46 * kid51 listens to the ultimate christmas music:  The Little Drummer Boy as sung by Marlene Dietrich!
01:47 cotto Whiteknight, there are already dedicated PMCs for those.  I'm not sure what the TODO is about.
01:47 dalek winxed: r294 | julian.notfound++ | trunk/winxedst1.winxed:
01:47 dalek winxed: operator += in stage 1
01:47 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=294
02:01 dalek parrot: r43171 | cotto++ | trunk/docs/pdds/pdd13_bytecode.pod:
02:01 dalek parrot: [pdd] minor fixes to pdd13
02:01 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43171/
02:03 kid51 dukeleto:  Thanks for the smoke test
03:42 preflex joined #parrot
04:46 JimmyZ joined #parrot
04:46 * JimmyZ couldn't open http://icanhaz.com/parrotroadmap
04:58 bacek joined #parrot
05:48 dalek parrot: r43172 | bacek++ | branches/context_unify3_simple/MANIFEST.generated:
05:48 dalek parrot: Install pmc_callcontext.h instead of pmc_callsignature.h. Spotted by pmichaud++
05:48 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43172/
05:48 dalek parrot: r43173 | bacek++ | branches/context_unify3_simple/src/call/context.c:
05:48 dalek parrot: Remove trailing spaces.
05:48 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43173/
05:56 cotto bacek, do you know what's in the Packfile debug segment?
05:57 cotto Is that what maps bytecode offsets to PIR/PASM line numbers?
05:59 bacek cotto, sorry, no idea.
05:59 cotto I suspect that answer will be the general consensus. ;)
06:00 bacek cotto :)
06:01 cotto pbc_dump isn't very smart about printing debug segments.
06:39 bacek purl, ( 4198893760 - 3578293848 ) / 4198893760 * 100
06:39 purl 14.7800813136077
06:39 bacek ~15% speed improvements. Not so bad
06:44 cotto It's a welcome improvement.
07:04 JimmyZ joined #parrot
07:13 JimmyZ joined #parrot
07:22 * JimmyZ doesn't know when he can get the commit bit.
07:48 cotto I proposed that you get a bit at the last #ps but didn't get any confirmation.  If you like I can ask what you could do better.
07:56 * cotto sleeps
07:58 JimmyZ cotto: thanks. I had sent a cla again.
08:03 * cotto really wants to know what the garbage at the end of a packfile debug segment is in pbc_dump's output.
08:40 pdcawley joined #parrot
09:22 fperrad joined #parrot
09:22 fperrad_ joined #parrot
09:26 iblechbot joined #parrot
09:51 dalek parrot: r43174 | fperrad++ | trunk/runtime/parrot/library/distutils.pir:
09:51 dalek parrot: [distutils] tweak spec rpm
09:51 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43174/
09:55 fperrad morning
09:59 JimmyZ fperrad: morning
10:13 Zak joined #parrot
10:37 mikehh context_unify3_simple branch - All tests PASS - (pre/post-config, make corevm/make coretest, test, fulltest at r43173 - Ubuntu 9.10 i386 (g++, --optimize)
10:38 dalek winxed: r295 | julian.notfound++ | trunk/winxedst1.winxed:
10:38 dalek winxed: call objects in stage 1
10:38 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=295
10:48 dalek winxed: r296 | julian.notfound++ | trunk/examples/xlibtest.winxed:
10:48 dalek winxed: some clean and refactor in example xlibtest
10:48 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=296
11:07 dalek winxed: r297 | julian.notfound++ | trunk/winxedst1.winxed:
11:07 dalek winxed: allow replacing statements in optimize step in stage 1
11:07 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=297
11:13 bacek joined #parrot
11:13 bacek o hai
11:21 moritz o hai
11:21 moritz there seem to be some dependencies broken in context_unify3_simple
11:21 moritz a make -j 3  fails
11:21 moritz ordinary make succeeds
11:28 bacek moritz, erm... It's probably broken in trunk already.
11:28 moritz bacek: last I tried it worked in trunk (but that doesn't mean too much)
11:29 bacek moritz, try it after realclean. "make docs" target is definitely broken with -j3
11:29 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#31121), fulltest) at r43174 - Ubuntu 9.10 i386 (gcc with --optimize)
11:29 bacek -j2 works pretty well
11:29 bacek mikehh++ # testing-testing-testing
11:32 mikehh I did a make -j corevm, make coretest TEST_JOBS=40 then make -j world and make test TEST_JOBS=40 - then I did a make -k fulltest TEST_JOBS=40
11:34 mikehh I did not try make -j but let me clean and see
11:35 bacek mikehh, "make corevm" should work with any -j. But "all" includes "docs" which depends on generated files but there is not dependencies in Makefile
11:36 fperrad context_unify3_simple is OK with mingw, see http://smolder.plusthree.com/ap​p/projects/report_details/31122
11:38 bacek fperrad, thanks!
11:38 mikehh since we had some problems with tests which passes make test but failed with make corevm/make coretest I have modified me testing procedure to test that first
11:38 mikehh we really need to fix those docs dependencies
11:42 mikehh anyway - need to reboot into amd64 for more testing - bbiab
11:48 mikehh BTW I just did a make realclean - config, make -j world, make -j test TEST_JOBS=40 in trunk and it worked ok
11:49 mikehh let me clean up and will be back in a minute
12:04 mikehh joined #parrot
12:04 joeri joined #parrot
12:07 bacek interesting. "make -j4" works on my box perfectly. (on context_unify3_simple branch)
12:23 bacek msg pmichaud nqp-rx just works on context_unify3_simple branch
12:23 purl Message for pmichaud stored.
12:23 mikehh I did a make -j world, make test TEST_JOBS=40 and it went ok (on amd64)
12:24 bacek mikehh, it works for me too. Looks like make dependencies were fixed recently
12:26 bacek mikehh, if context_unify3_simple passed on amd64 I can merge it back to trunk in minutes.
12:27 mikehh context_unify3_simple branch - All tests PASS (pre/post config, test, fulltest) at r43174 - Ubuntu 9.10 amd64 (g++ with --optimize)
12:27 bacek oookeeey
12:27 mikehh so it passes all tests for me on i386 and amd64
12:27 bacek Merge party!
12:27 bacek It also passed on win32/mingw
12:31 bacek r43175
12:32 bacek Everybody panic!
12:32 purl i guess everybody panic is http://web.archive.org/web/20070216080911/​http://www.autofrog.com/~chrisp/panic.gif
12:34 dalek parrot: r43175 | bacek++ | trunk (58 files):
12:34 dalek parrot: Merge context_unify3_simple branch back to trunk.
12:34 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43175/
12:37 bacek TapTinder is green so far
12:41 nopaste "bacek" at 122.110.241.182 pasted "Patch for build rakudo after contex_unify3_merge." (42 lines) at http://nopaste.snit.ch/19150
12:51 JimmyZ joined #parrot
12:55 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#31127), fulltest) at r43175 - Ubuntu 9.10 amd64 (gcc with --optimize)
12:55 JimmyZ oh hai
12:56 mikehh hello JimmyZ
12:58 JimmyZ mikehh: can you open http://icanhaz.com/parrotroadmap?
12:58 JimmyZ mikehh: I can't open it here.
12:59 bacek JimmyZ, looks like some problem with google apps.
13:00 mikehh JimmyZ: yes I can open it
13:00 JimmyZ now I open it.
13:01 JimmyZ but It is really difficult
13:21 kid51 joined #parrot
13:39 iblechbot joined #parrot
13:43 nopaste "kid51" at 70.85.31.226 pasted "make corevm/coretest: failures in t/src/library.t (r43175 Linux/i386)." (47 lines) at http://nopaste.snit.ch/19151
13:55 kid51 Those failures in t/src/library.t are occurring during regular 'make test' as well.
14:07 mikehh t/src/library.t ?
14:07 kid51 Yes
14:08 mikehh don't have it
14:08 kid51 There's something strange going on here.  I did *not* get those failures in my smoke test on Darwin/PPC.
14:09 kid51 Yeah, where did this file come from?
14:09 patspam joined #parrot
14:09 mikehh also t/src is not tested in coretest
14:15 nopaste "kid51" at 71.246.106.84 pasted "I wonder where this file came from" (170 lines) at http://nopaste.snit.ch/19152
14:25 mikehh kid51: dunno, it's not in my source tree or in trac
14:32 mikehh afaik: it was removed from the tree months ago
14:35 mikehh in fact I am not sure it was there - it was in branch pdd_install or something like that
14:40 payload joined #parrot
14:46 dalek winxed: r298 | julian.notfound++ | trunk/winxedst1.winxed:
14:46 dalek winxed: handle predefined functions, define several of them and some related changes in
14:46 dalek winxed: stage 1
14:46 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=298
14:50 mikehh kid51: searching trac I got 4 references to the file from about a year ago - TT #128 might have got it for you
14:51 preflex joined #parrot
15:02 kid51 mikehh:  again, you nailed it!  I was reviewing old tickets and decided to try out the patch in TT #128.  So I put it there!
15:08 Psyche^ joined #parrot
15:15 dalek winxed: r299 | julian.notfound++ | trunk/winxedst1.winxed:
15:15 dalek winxed: fix and refactor operators * and / in stage 1
15:15 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=299
15:21 dalek partcl-nqp: c18a576 | coke++ |  (4 files):
15:21 dalek partcl-nqp: Add [lappend] - pass all but one test.
15:21 dalek partcl-nqp: Add some helper methods to make it easier to get a list value from a PMC.
15:21 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/c18a576981d1de0577ee3eeef42169d218dd6669
15:21 dalek partcl-nqp: 4016034 | coke++ | src/Partcl/commands/main.pm:
15:21 dalek partcl-nqp: switch [join] to use getList()
15:21 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/40160346c7ce399d7f35155e23c8daba8d4a23ef
15:21 dalek partcl-nqp: e5b0dbf | coke++ | .gitignore:
15:21 dalek partcl-nqp: Ignore more generated files
15:21 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/e5b0dbfb9eabe002e4f1921c26874500b606995f
15:34 KatrinaTheLamia joined #parrot
15:45 payload joined #parrot
16:04 cognominal joined #parrot
16:31 dalek winxed: r300 | julian.notfound++ | trunk/winxedst1.winxed:
16:31 dalek winxed: fix and refactor comparator ops and emit correct .param types in stage 1
16:31 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=300
16:35 dalek parrot: r43176 | fperrad++ | trunk/compilers/pirc/src (2 files):
16:35 dalek parrot: [pirc] fix build after merge context_unify3_simple branch
16:35 dalek parrot: see r43175
16:35 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43176/
16:55 dalek winxed: r301 | julian.notfound++ | trunk/winxedst1.winxed:
16:55 dalek winxed: handle sources with utf8 encoded string literals in stage 1
16:55 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=301
16:55 dalek winxed: r302 | julian.notfound++ | trunk/Makefile:
16:55 dalek winxed: pass all stage 0 tests in make test1
16:55 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=302
16:55 Whiteknight joined #parrot
17:06 Whiteknight good morning #parrot
17:14 Zak joined #parrot
17:16 dalek matrixy: d3bf9ad | Whiteknight++ | src/parser/actions.pm:
17:16 dalek matrixy: small change to the parser to allow printing multiple values on assignment without ;
17:16 dalek matrixy: review: http://github.com/Whiteknight/matrixy/commi​t/d3bf9adc4104a37ce2aeca6022325b81a0e27807
17:27 iblechbot joined #parrot
17:33 dalek winxed: r303 | julian.notfound++ | trunk/winxedst1.winxed:
17:33 dalek winxed: initial support for parameter modifiers in stage 1
17:33 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=303
17:55 cognominal joined #parrot
19:02 ash_ joined #parrot
19:04 particle joined #parrot
19:14 ash_ is there a way to list all the attributes on parrot object?
19:15 ash_ a PBC i mean
19:15 ash_ errr pmc not pbc
20:11 bacek joined #parrot
20:47 dalek partcl-nqp: 36df97f | coke++ |  (2 files):
20:47 dalek partcl-nqp: implement [lassign] - all tests pass.
20:47 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/36df97fd6fbbd67a6dd7bd2d133034a5343a2b43
20:53 Coke ash_: moment.
20:54 Coke ash_: hurm. I do not see anything obvious in src/ops/object.ops
20:55 Coke seems like a reasonable thing to want; I'd ask on the parrot-users list.
20:56 ash_ Coke: its cool, i think i have it figured out
20:59 Coke sweet.
21:00 ash_ Coke: if i wanted to create an instance of my own object (well really a rakudo one) how would i do that? do i need to get_hll_global ['Perl6'; 'NameOfClass'] ?
21:06 Coke no, you want some variant of 'new'
21:06 Coke new 'NameOfClass' is probably sufficient.
21:06 bacek joined #parrot
21:06 Coke I know pmichaud had a need for 'root_new' at one point.
21:08 ash_ Coke: k, i'll try that
21:15 ash_ Coke: can you think of a reason why doing Foo.HOW in rakudo and doing foo.'HOW'() in parrot would be different?
21:15 ash_ where foo is Foo
21:15 purl foo Foo is probably a little RABBIT
21:18 ash_ I have a perl 6 role defined in rakudo and when i do Foo.HOW it returns a RoleHOW, but in parrot it returns itself
21:18 ash_
21:18 ash_ so Foo.'HOW'() == Foo
21:22 dukeleto ash_: example code?
21:22 purl hmmm... example code is perl
21:22 dukeleto purl, forget example code
21:22 purl dukeleto: I forgot example code
21:24 Coke my guess without seeing code is that you're using an instance instead of the class or vice versa.
21:32 ash_ Coke's right, thats what my mistake was
21:41 dukeleto Coke++
22:02 Coke yay, 8 years of parrot good for something.
22:08 jan joined #parrot
22:12 iblechbot joined #parrot
22:26 theory joined #parrot
22:42 cotto joined #parrot
22:58 * Coke wonders if someone broken nqp-rx. :P
22:58 Coke *broke
23:03 NotFound Nice, I got the first parrot segfault with winxed.
23:08 NotFound Solved by closing output files before exiting. Looks like flushing output during global destruction doesn't work well.
23:13 dalek winxed: r304 | julian.notfound++ | trunk/winxedst1.winxed:
23:13 dalek winxed: refactor modifiers, some cleaning, and closing output before exit in stage 1
23:13 dalek winxed: review: http://code.google.com/p/w​inxed/source/detail?r=304
23:15 Coke yup. someone broke partcl-nqp; downgrading parrot in the meantime.
23:16 PerlJam Coke: broke it how?
23:17 Coke PerlJam: see the lsit.
23:18 Coke bacek's last mergeback appears to have borked nqp-rx.
23:18 Coke -> afk
23:19 dalek partcl-nqp: 5a6184d | coke++ | src/Partcl/commands/main.pm:
23:19 dalek partcl-nqp: add [linsert] - all tests but one pass.
23:19 dalek partcl-nqp: review: http://github.com/partcl/partcl-nqp/commit​/5a6184db1cc0110f0bed3acf4a3abdda690e2ab6
23:25 xenoterracide joined #parrot
23:33 ash_ how does box work?
23:36 NotFound ash_: creating a pmc of the appropiate type and asigning a value to it.
23:43 ash_ in a HLL how do you associate types with it though? so they get turned into the right thing
23:46 NotFound Look at the hll_map tests in t/pmc/parrotinterpreter.t for example
23:47 ash_ kk, thanks
23:51 ash_ NotFound: so, hll_map is how you'd associate them, right? what if they are not sublcasses of each other?
23:52 NotFound I think that there is no problem, as long at it provides the appropiate vtable functions.
23:56 ash_ so, here's the problem i am running into in the ng-branch of rakudo, maybe i am going about this wrong but at compile time a certain method is needed to compose role's into other roles, but at runtime you can use the same method to get an array of all the methods in the role, the problem is its currently returning a ResizablePMCArray, but at runtime we need it to return an actual perl6 type
23:56 ash_ could hll_map take care of that if you mapped them after the compile phase is finished?
23:57 eternaleye joined #parrot

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

Parrot | source cross referenced