Camelia, the Perl 6 bug

IRC log for #parrot, 2010-01-13

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 cconstantine joined #parrot
00:02 payload joined #parrot
00:07 nopaste joined #parrot
00:09 hercynium joined #parrot
00:22 zak__ joined #parrot
00:32 nopaste joined #parrot
00:42 nopaste joined #parrot
00:52 abqar joined #parrot
00:57 nopaste joined #parrot
01:02 zak_ joined #parrot
01:05 plobsing joined #parrot
01:05 Zak joined #parrot
01:08 zak__ joined #parrot
01:11 plobsing hi #parrot
01:11 plobsing sorry for missing #ps
01:12 cotto_work it was a very quiet meeting
01:14 plobsing I'm looking to merge pmc_freeze_cleanup in the next couple of hours unless anyone has any objections
01:17 kid51 plobsing:  Can you comment on whether merging in that branch will exacerbate the problems reported in http://trac.parrot.org/parrot/ticket/1393?
01:18 zak_ joined #parrot
01:20 kid51 plobsing:  Also, there was some discussion of that in #ps today.  chromatic alluded to a question about threads.
01:21 plobsing kid51: It's hard to predict if this issue will be exacerbated or alleviated by the branch. what happens is that the image sizes change, changing when collections happen revealing and/or hidding extant GC-related problems
01:21 plobsing kid51: the threads issue I already understand. simply put GC+Threads+COW=FAIL
01:23 kid51 I don't understand what image sizes are?  Also, what do you mean by collections?
01:23 plobsing GC sweep/compaction passes
01:23 plobsing image sizes => number of bytes a frozen PMC takes
01:24 kid51 Am smoking branch now.
01:24 plobsing arch?
01:24 purl well, arch is at language.perl.com/ppt/src/arch/ or A version control system at http://www.regexps.com/
01:27 cotto_work Having seen the freeze/thaw code, I object to *not* merging within the next couple hours.
01:27 nopaste joined #parrot
01:28 kid51 cotto_work:  Are you referring to the *current* freeze/thaw code?
01:29 cotto_work yes
01:31 zak__ joined #parrot
01:32 plobsing wrt revealing/hiding GC bugs, I've been considering adding another test$x target to the makefile to run the gcdebug core on all tests
01:34 kid51 plobsing:  http://smolder.plusthree.com/app/pu​blic_projects/report_details/31638  branch PASSes with one TODO passing
01:35 kid51 plobsing:  But isn't that what we do now?
01:35 kid51 'make test' entails '--gc-debug' by default.
01:35 plobsing nope we run with various cores with a "--gc-debug" option
01:35 plobsing I want to run with -R gcdebug
01:35 plobsing which is somehow different and seems to more reliably catch gc bugs
01:35 kid51 What does that mean?
01:35 purl That boy needs therapy.
01:36 plobsing -R gcdebug is the slow core but running a full mark/compact each instruction
01:36 kid51 purl:  Probably true -- but I have a choice between spending $150/hour on a shrink or being poor when I'm old.
01:36 plobsing not sure what if anything --gcdebug
01:36 plobsing does
01:37 plobsing anyways, with that smoke succeeding, I'll proceed with the merge
01:38 kid51 plobsing:  Yes.  Your branch was created before the 1393 problem emerged, so it won't show the problem by itself.  (But trunk after merging will probably still show problem.)
01:41 cotto_work plobsing, aren't you supposed to sync from trunk before merging?
01:41 plobsing I did. I ran svn up. Isn't that sync from trunk?
01:41 plobsing my svn-foo is a little on the weak side
01:42 kid51 No, it's not.  However, I'm unsure why cotto_work is recommending that.
01:42 cotto_work I was thinking more like svn merge https://svn.parrot.org/parrot/trunk .
01:42 cotto_work I thought it was sop
01:43 kid51 My two cents:  In this situation, I would do a fresh checkout of trunk.
01:44 kid51 Then call:  svn merge https://svn.parrot.org/parrot/trunk@ 43030 https://svn.parrot.org/parrot/b​ranches/pmc_freeze_cleanup@HEAD .
01:44 plobsing why aren't I supposed to do something along the lines of (cd trunk; svn up; svn merge -r A:B ../branch_name) ?
01:44 kid51 Examine output for conflicts.
01:44 kid51 Resolve conflicts.
01:44 kid51 configure, build, test
01:45 kid51 plobsing:  Suppose your merge leaves you with a lot of files in conflict.
01:45 kid51 It's helpful to have one sandbox where you can sort that all out, ...
01:45 cotto_work btw, you forgot to "fix" compilers/pirc/src/bcgen.h after running headerizer
01:45 kid51 ... while retaining another sandbox for ongoing commits to trunk.
01:45 kid51 bbiaf
01:46 plobsing cotto_work: what is compilers/pirc/src/bcgen.h and why do I have to fix it ? doc ref ?
01:47 cotto_work for some reason having the declaration for emit_pbc_key breaks pirc
01:48 cotto_work also, I tools/dev/mk_native_pbc needs to be run and the changes committed
01:48 plobsing I already had that lined up. just commited
01:48 cotto_work there it is
01:49 cotto_work running fulltest now
01:49 cotto_work btw, this makes me happy: http://www.flickr.com/photo​s/davidcoallier/2207776375/
01:51 plobsing cotto_work: how do I reproduce the pirc breakage?
01:52 cotto_work try to build it
01:52 cotto_work it'll asplode
01:53 bacek joined #parrot
01:53 nopaste "cotto_work" at 131.107.0.80 pasted "apply this in reverse to fix pirc" (16 lines) at http://nopaste.snit.ch/19255
01:53 cotto_work plobsing: ^
01:53 cotto_work hio bacek_at_work
01:53 dalek parrot: r43434 | plobsing++ | failed to fetch changeset:
01:53 dalek parrot: merge branch pmc_freeze_cleanup
01:53 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43434/
01:53 cotto_work also, bacek
01:53 dalek parrot: r43435 | plobsing++ | trunk/t/native_pbc (4 files):
01:53 dalek parrot: update native_pbc to new PBC format
01:53 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43435/
02:01 cotto_work plobsing, is there somewhere (parrot-dev post, wiki page) where you've summarized the changes that were implemented by that branch?
02:02 plobsing no. where should I put that?
02:03 cotto_work the list or a wiki page would be fine, whichever is easier for you
02:09 dalek parrot: r43436 | plobsing++ | trunk/src/pmc/parrotthread.pmc:
02:09 dalek parrot: dissable GC when using threads - they don't play nice
02:09 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43436/
02:09 dalek parrot: r43437 | plobsing++ | trunk/compilers/pirc/src/bcgen.h:
02:09 dalek parrot: fix pirc build after headerizer
02:09 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43437/
02:10 cotto_work It looks like some tests in t/examples/pod.t are causing fulltest to fail.  Someone should probably fix those.
02:12 nopaste joined #parrot
02:16 plobsing_ joined #parrot
02:16 zak_ joined #parrot
02:18 Zak joined #parrot
02:27 kid51 Well here's some good news ...
02:28 kid51 I just did 'svn up' on trunk after plobsing's merges.
02:28 kid51 Am currently at r43437.
02:28 kid51 I got the same smoke results in trunk as I did in the branch ...
02:29 kid51 i.e., PASS with t/pmc/eval.t TODO passed:   12
02:29 zak_ joined #parrot
02:29 kid51 ... which means that the failure being discussed in TT #1393 did NOT occur!
02:29 kid51 The dog did not bark!
02:30 plobsing_ so the changes hide the bug? yay! sort of
02:34 dalek tracwiki: v1 | plobsing++ | pmc_freeze_cleanup
02:34 dalek tracwiki: create page to describe changes in pmc_freeze_cleanup branch
02:34 dalek tracwiki: http://trac.parrot.org/parrot/wiki/pmc_f​reeze_cleanup?version=1&action=diff
02:34 dalek tracwiki: v2 | plobsing++ | pmc_freeze_cleanup
02:34 dalek tracwiki: http://trac.parrot.org/parrot/wiki/pmc_f​reeze_cleanup?version=2&action=diff
02:40 nopaste "kid51" at 70.85.31.226 pasted "t/examples/pod.t: failures at r43437" (59 lines) at http://nopaste.snit.ch/19256
02:40 JimmyZ joined #parrot
02:51 plobsing_ same failures exist in r43433. errors not a result of pmc_freeze_cleanup merge
02:54 Coke the pod example failures may be a result of my earlier fix to properly delimit those examples.
02:55 zak__ joined #parrot
02:57 nopaste joined #parrot
03:13 bacek joined #parrot
03:14 dalek parrot: r43438 | plobsing++ | trunk/docs/pmc/subs.pod:
03:14 dalek parrot: fix docs/pmc/subs.pod wrt t/examples/pod.t
03:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43438/
03:14 dalek parrot: r43439 | plobsing++ | trunk/docs/user/pir/intro.pod:
03:14 dalek parrot: fix docs/user/pir/intro.pod wrt t/examples/pod.t ('=begin PIR' becomes 'begin PIR_FRAGMENT')
03:14 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43439/
03:14 plobsing_ tunk fulltest passes on linux x86_64
03:18 cotto happy times
03:20 cognominal joined #parrot
03:27 nopaste joined #parrot
03:39 zak_ joined #parrot
03:40 Coke sorry about that.
03:43 Zak joined #parrot
03:46 * kid51 must sleep
03:46 purl $kid51->sleep(8 * 3600);
03:47 plobsing_ dalek slow?
03:48 dalek parrot: r43440 | plobsing++ | trunk (3 files):
03:48 dalek parrot: add -G flag to t/harness and testG makefile target to run tests with '-runcore gcdebug'
03:48 dalek parrot: running these tests currently gives 8 failures on linux x86_64
03:48 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43440/
03:57 nopaste joined #parrot
03:58 zak_ joined #parrot
03:59 plobsing joined #parrot
04:05 zak__ joined #parrot
04:22 nopaste joined #parrot
04:32 bacek joined #parrot
04:42 nopaste joined #parrot
04:57 nopaste joined #parrot
05:05 zak_ joined #parrot
05:08 abqar joined #parrot
05:17 nopaste joined #parrot
06:26 chromatic joined #parrot
06:34 JimmyZ joined #parrot
06:51 cotto_w0rk joined #parrot
06:54 abqar_ joined #parrot
06:54 leto joined #parrot
06:57 bacek joined #parrot
06:57 Maddingue joined #parrot
06:58 baest joined #parrot
07:24 uniejo joined #parrot
07:25 mikehh joined #parrot
07:40 jsut_ joined #parrot
07:42 uniejo_ joined #parrot
07:44 uniejo joined #parrot
07:55 iblechbot joined #parrot
08:02 davidfetter joined #parrot
08:13 bacek joined #parrot
08:38 fperrad joined #parrot
08:56 aninhumer joined #parrot
09:11 bacek joined #parrot
09:34 dalek joined #parrot
09:36 AndyA joined #parrot
09:45 bacek o hai
09:48 dduncan left #parrot
11:31 dalek joined #parrot
11:46 payload joined #parrot
11:56 kid51 joined #parrot
12:07 bluescreen joined #parrot
12:09 kid51 make fulltest:  Darwin/PPC: r43437:  PASS
12:23 bluescreen joined #parrot
12:43 dalek parrot: r43441 | jkeenan++ | failed to fetch changeset:
12:43 dalek parrot: Merge tt473_remove_memcpy_aligned branch into trunk.  Applying patch submitted by Andy Dougherty in �http://trac.parrot.org/parrot/ticket/473 to remove unused configuration probes.
12:43 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43441/
12:43 dalek parrot: r43442 | jkeenan++ | branches/tt473_remove_memcpy_aligned:
12:43 dalek parrot: Branch has been merged into trunk and is no longer needed at HEAD.
12:43 purl i already had it that way, dalek.
12:43 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43442/
12:56 aninhumer Editing chXX_hlls.pod, not sure what this is trying to say: "If the PMCs were written properly to satisfy this interface
12:56 aninhumer all the necessary information from those PMCs. Operate on the PMCs at the
12:56 aninhumer VTable level, and we can safely ignore the implementation details of them."
12:57 aninhumer Also, trivial, but what's the canon capitalisation of vtable?
12:59 clinton joined #parrot
12:59 dalek parrot: r43443 | jkeenan++ | tags/tt473_remove_memcpy_aligned-43162:
12:59 dalek parrot: Branch to which tag corresponded has been merged into trunk; tag may be removed.
12:59 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43443/
13:05 ruoso joined #parrot
13:22 cconstantine joined #parrot
13:35 jsut joined #parrot
13:47 payload joined #parrot
13:56 payload joined #parrot
14:06 dalek lua: 8b826f5 | fperrad++ | dynext/pmc/lua (3 files):
14:06 dalek lua: update PMC freeze/thaw after merge branch pmc_freeze_cleanup
14:06 dalek lua: see http://trac.parrot.org/parrot/changeset/43434
14:06 dalek lua: review: http://github.com/fperrad/lua/commit/8b​826f598cdd68dab065b08c9c7e9eeb20b12f46
14:09 PerlJam joined #parrot
14:19 zak_ joined #parrot
14:19 PerlJam good morning #parrot
14:32 davidfetter oh hai
14:36 dalek parrot: r43444 | mikehh++ | trunk/docs/user/pir/intro.pod:
14:36 dalek parrot: fix examples_tests TODO pass - docs/user/pir/intro.pod
14:36 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43444/
14:52 plobsing joined #parrot
14:52 zak__ joined #parrot
14:55 Zak joined #parrot
15:02 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#31657), fulltest) at r43444 - Ubuntu 9.10 amd64 (gcc with --optimize)
15:02 mikehh t/pmc/eval.t - TODO passed:   12 in make coretest, make test and in all cores in fulltest
15:02 fperrad_ joined #parrot
15:10 cognominal joined #parrot
15:10 Psyche^ joined #parrot
15:23 iblechbot joined #parrot
15:40 payload joined #parrot
15:40 bubaflub joined #parrot
16:02 Zak joined #parrot
16:17 zak_ joined #parrot
16:17 dalek tracwiki: v1 | Freddy1++ | ZendStudioIde
16:17 dalek tracwiki: http://trac.parrot.org/parrot/wiki/Ze​ndStudioIde?version=1&action=diff
16:36 cotto unusual spam there
16:40 zak__ joined #parrot
16:44 Zak joined #parrot
16:45 zak_ joined #parrot
16:46 cotto baleeted
16:47 plobsing joined #parrot
16:51 cotto (fwiw, it was a partial copy/paste from http://forums.zend.com/view​topic.php?f=59&t=2134)
17:10 zak__ joined #parrot
17:19 zak_ joined #parrot
17:22 theory joined #parrot
17:25 zak__ joined #parrot
17:30 mtk joined #parrot
17:33 jsut_ joined #parrot
17:36 chromatic joined #parrot
17:51 dalek parrot: r43445 | mikehh++ | trunk/t/pmc/eval.t:
17:51 dalek parrot: fix t/pmc/eval.t - test 12 TODO pass
17:51 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43445/
17:53 cotto_work joined #parrot
17:54 zak_ joined #parrot
17:56 jsut joined #parrot
17:59 zak_ joined #parrot
18:02 zak__ joined #parrot
18:17 davidfetter joined #parrot
18:19 zak__ joined #parrot
18:23 theory joined #parrot
18:25 coke joined #parrot
18:25 coke feather is out of ptys.
18:38 plobsing joined #parrot
18:55 payload joined #parrot
19:03 joeri joined #parrot
19:05 aninhumer Is it suggested to submit a separate "[PATCH]" ticket, even if a bug ticket already exists? Or just attach the patch?
19:06 coke just attach.
19:06 coke (woot)
19:07 coke feather?
19:07 purl feather is http://feather.perl6.nl/ or a perl6 community development server or run by juerd.
19:07 coke juerd?
19:07 purl i heard juerd was root or at http://juerd.nl/ or mailto:juerd@juerd.nl
19:09 mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#31660), fulltest) at r43445 - Ubuntu 9.10 amd64 (g++ with --optimize)
19:10 AndyA joined #parrot
19:31 jsut_ joined #parrot
19:35 bluescreen joined #parrot
19:38 japhb joined #parrot
19:42 ash_ joined #parrot
20:00 TimToady phone
20:00 ash_ so... i have a question, what does parrot do if you have a .lex '$name', $S0 ? that's not an error, and it compiles, it just silently fails to create the lexical variable
20:01 ash_ i guess my question is, why?
20:01 chromatic IMCC takes an odd approach to error handling.
20:03 ash_ so... thats an imcc problem? just curious, i was banging my head against the wall trying to figure something out and it turned out i was just being dumb and passing the wrong thing to .lex, but that seemed like it should of failed before i did a find_lex
20:06 chromatic I'm pretty sure it's an IMCC problem, and I'm certain it's an easy fix to add a better error there.
20:06 chromatic If you had a few test cases, it'd be almost trivial.
20:07 ash_ alright, i'll write up some test cases for you
20:07 chromatic Thanks!
20:12 Psyche^ joined #parrot
20:16 ash_ so, since this is a parser error, how would i write a test for it? i'd need to have the file contents in a string that is going to run the imcc and expect an error to be raised, but i am not sure how to accomplish that
20:16 ash_ or if you want i can just give you an example file that i thought should raise an error that doesn't
20:17 chromatic Some of the files in t/ in Parrot use a function called pir_error_output_is().
20:17 chromatic Otherwise, ping dukeleto and ask him about error catching in PIR.
20:18 ash_ okay, i'll look for that one and use it
20:22 ash_ chromatic: so you know, i looked in the test helper, there is a pir_error_output_is but its not used much, pir_error_output_like is used far more, i am going to use pir_error_output_like
20:22 chromatic Sure, that's fine.
20:27 ash_ so, should .lex box the value into a pmc? or should it raise an error? the design doc 20 which cover's .lex doesn't cover that, and the docs on opcode's show that the store_lex op takes a pmc, store_lex raises an error if anything other than a pmc is passed to it
20:28 cotto_work It's nice to see there some other Parrot hackers who read LtU.
20:29 bacek joined #parrot
20:29 chromatic My plan is to make IMCC throw a syntax error when using anything other than a PMC register with .lex.
20:31 ash_ okay, i'll try to come up with examples that should fail, i'll just look for a generic error message for now
20:36 chromatic I can change the text of the error message if you don't think it's brilliant.
20:36 ash_ sure
20:43 dalek TT #1409 created by Aninhumer++: [PATCH] Copyedit + Fixes to submissions.pod
20:45 redbrain that paper Leto just sent around is pretty good!
20:53 GeJ Good morning everyone
20:57 aninhumer I'm editing docs/user/pir/intro.pod, if it was made at the time of 0.8.1 I imagine "make install" is supposed to work now? (It did for me)
20:58 cotto_work yes
21:03 cotto_work btw, thanks.  Documentations updates are most welcome.
21:05 davidfetter joined #parrot
21:14 payload joined #parrot
21:15 davidfetter joined #parrot
21:20 plobsing joined #parrot
21:21 ewilhelm left #parrot
21:25 hudnix joined #parrot
21:28 payload joined #parrot
21:32 ash_ chromatic: i have some basic tests http://gist.github.com/276594 it adds 6 tests, 3 of them already pass, just to verify store_lex only excepts a pmc as its argument, anything else raises a parser error, .lex's tests fail though
21:33 chromatic Sounds reasonable.
21:33 ash_ the gist is a svn diff
21:37 ash_ chromatic: anything else you'd need from me? want me to put a ticket in trac or something?
21:52 chromatic If I don't get to it by this time tomorrow, please do.
21:53 ash_ sure, i'll try to remind you, if i don't forget
21:53 chromatic I wrote it on my task list.
22:01 theory joined #parrot
22:05 aninhumer I'm trying to work out how the scheduler works, but as far as I can see, the code attribute of my Tasks is being reset to null?
22:05 aninhumer http://gist.github.com/276617
22:06 aninhumer This produces: Null PMC access in invoke()
22:06 aninhumer current instr.: 'testcaller' pc 62 (conctest.pir:39)
22:10 chromatic Look out, here comes bacek!
22:17 dalek parrot: r43446 | bacek++ | branches/gc_encapsulate/src/gc (2 files):
22:17 dalek parrot: Reimplement allocation of attributes in GC MS
22:17 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43446/
22:17 dalek parrot: r43447 | bacek++ | branches/gc_encapsulate/src/gc/gc_private.h:
22:17 dalek parrot: Fix Variable_Size_Pool *compact declaration.
22:17 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43447/
22:17 dalek parrot: r43448 | bacek++ | branches/gc_encapsulate/src/gc/gc_ms.c:
22:17 dalek parrot: Implement GC MS gc_allocate_buffer(_with_pointers)
22:17 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43448/
22:17 dalek parrot: r43449 | bacek++ | branches/gc_encapsulate/src/gc (2 files):
22:17 dalek parrot: Implement GC MS mark/sweep block/unblock
22:17 dalek parrot: review: http://trac.parrot.org/parrot/changeset/43449/
22:27 cognominal joined #parrot
22:31 mikehh joined #parrot
22:47 nopaste joined #parrot
22:57 cognominal joined #parrot
23:03 Limbic_Region joined #parrot
23:08 davidfetter joined #parrot
23:14 bacek_at_work o hai
23:24 GeJ G'Day bacek.
23:24 cconstantine joined #parrot
23:24 bacek_at_work G'Day GeJ
23:34 kid51 joined #parrot
23:54 bacek joined #parrot

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

Parrot | source cross referenced