Camelia, the Perl 6 bug

IRC log for #parrot, 2009-04-20

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:09 AndyA joined #parrot
00:10 kid51 joined #parrot
00:34 Topic for #parrotis now Parrot 1.0 Released | http://parrot.org | 334 RTs left!
00:39 LylePerl There is a Rakudo build problem on vista
00:39 LylePerl It won't build against parrot r38030 and later
00:39 tetragon joined #parrot
00:41 LylePerl The change in r38030 is that \src\pmc\sharedref.pmc and src\pmc\ref.pmc are removed
00:41 LylePerl Is this a parrot or Rakudo issue?
00:42 LylePerl Rakudo build dies with mingw32-make: *** [perl6_s1.pbc] Error -1073741819
00:52 kid51 Today is my lucky day.  I got a response to a bug ticket I filed at rt.cpan.org in March 2005!
00:52 cotto kid51, which ticket is that?
00:54 kid51 https://rt.cpan.org/Ticket/Display.html?id=12401
00:58 LylePerl kid51: I know mark takes on the role of maintainer for modules that have long bug queues
01:00 kid51 Is that Mark Stosberg?
01:00 cotto nie
01:00 cotto *nice
01:02 kid51 Funny, I probably haven't done anything CGI-related since submitting that ticket.
01:02 kid51 I wasn't even employed in IT at that point.
01:02 kid51 But it was the heyday of the Phalanx project!
01:03 kid51 cotto++ for examining and closing lots of RTs
01:06 Whiteknight cotto++ # agreed
01:06 dalek parrot: r38216 | coke++ | trunk/t/benchmark/benchmarks.t:
01:06 dalek parrot: [t] don't test removed file.
01:06 rg cotto++ # what they said :)
01:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38216/
01:07 kid51 I once said that the way to deal with the backlog of RTs was to lock Allison, chromatic and particle in a room and not let them out until the backlog was cleared.
01:08 kid51 Now, we'll have to lock cotto in there as well ;-)
01:08 kid51 No good deed goes unpunished!
01:08 purl no good deed goes unpunished is what self-righteous morons say when their thoroughly retarded plans backfire horribly
01:08 kid51 purl, "no good deed goes unpunished" is sometimes used ironically.
01:08 purl OK, kid51.
01:11 wayland No good deed goes unpunished!
01:11 purl no good deed goes unpunished is what self-righteous morons say when their thoroughly retarded plans backfire horribly
01:11 wayland Phooey
01:11 wayland I just wanted to see if things had changed :)
01:12 kid51 No, because I failed to capitalize the 'no'
01:13 wayland mo good deed goes unpunished!
01:13 wayland no good deed goes unpunished!
01:13 purl somebody said no good deed goes unpunished was what self-righteous morons say when their thoroughly retarded plans backfire horribly
01:13 rg that's not it. there seem to be some things hardwired into purl
01:13 kid51 purl, forget no good deed goes unpunished
01:13 purl kid51: I forgot no good deed goes unpunished
01:13 rg no good deed goes unpunished
01:14 rg no good deed goes unpunished!
01:14 rg interesting
01:14 kid51 Ah, the bot is speechless!
01:14 wayland No good deed goes unpunished!!!!!
01:14 wayland purl, kid51 will be happy today!
01:14 purl wayland: i'm not following you...
01:14 cotto no good purl goes unpunished
01:14 wayland phooey.  It was worth a go :)
01:15 cotto which is vacuously true in this case ;)
01:15 wayland I must not have that aura of authority :)
01:15 kid51 There is a web site where you can study purl's rules.  But I haven't always been able to apply them successfully.
01:17 bacek_ joined #parrot
01:19 kid51 The syntax for non-private messaging someone is what?  purl msg user  This is a message for you ????
01:19 Tene kid51: yes
01:19 Tene purl: msg kid51 this is a message
01:19 purl Message for kid51 stored.
01:19 kid51 And a private message is the same, only prefixed by /q   -- correct?
01:20 kid51 Got it, thx.
01:20 Tene kid51: yes, that should work
01:27 cognominal joined #parrot
01:33 dalek parrot: r38217 | whiteknight++ | trunk/src/pmc/fixedpmcarray.pmc:
01:33 dalek parrot: switch an instance of PMC_size(), which is specific to the FixedPMCArray PMC, to a call to VTABLE_elements instead. This way the sort method can be better inherited by subclasses (although it still doesn't work, but now it fails for a different reason. See TT #218 for details and progress
01:33 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38217/
01:51 Coke (#*@&$@#()$&@#$
01:52 Coke I am trying to run fulltest. people keep breaking 'codetest' while I do this.
01:53 Coke so I fix one thing, update, rerun... boom.
01:55 kid51 purl msg Infinoid Tested your patch in http://nopaste.snit.ch/16321 on Darwin/PPC:  'make' and 'make test' all pass.  See http://smolder.plusthree.com/app/pu​blic_projects/report_details/20177
01:55 purl Message for infinoid stored.
01:55 shorten kid51's url is at http://xrl.us/bepknf
01:56 kid51 Coke:  I use make fulltest_all for that
01:56 dalek parrot: r38218 | coke++ | trunk/config (3 files):
01:56 dalek parrot: [cage] Remove trailing whitespace.
01:56 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38218/
01:56 Coke kid51: that doesn't help.
01:56 kid51 y not?
01:56 Coke as then the failures are hidden in the backscroll.
01:56 Coke It might help me ||ize my own fixes, but not when I'm getting failures introduced by 'svn up'
01:58 Infinoid kid51++, thanks for the ppc test
01:58 nopaste "kid51" at 70.85.31.226 pasted "How I log make fulltest_all" (41 lines) at http://nopaste.snit.ch/16324
01:58 * Coke tries again.
01:58 Coke "unable to complete fulltest_all doesn't tell you anything, does it?
01:59 Coke I thought it ignored the results of all the various failures.
01:59 Coke yup.
01:59 dalek parrot: r38219 | coke++ | trunk/config/auto/perldoc.pm:
01:59 dalek parrot: [cage] no hard tabs, please.
01:59 dalek parrot: pass t/codingstd/perlcritic.t
01:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38219/
02:02 kid51 Well, I'm trying make fulltest_all now in a screen session.  I'll let you know the outcome.
02:02 Infinoid Coke: Looks like the last few codetest issues were my fault... sorry about that.
02:03 Coke kid51: it will pass. it always passes. =-)
02:05 rhr joined #parrot
02:05 Coke I wonder if anyone is running 'fulltest' anymore. Guessing not, or they would have found the shared_ref issue already.
02:06 kid51 In which core or variant?
02:06 Tene purl: msg jonathan Is there a ticket yet for 'is export' not working in modules not in a top-level namespace?  that is, module Foo::Bar { sub lol is export { ... } }
02:06 purl Message for jonathan stored.
02:07 Coke kid51: it was the benchmarks test.
02:07 kid51 my fulltest_all hasn't gotten that far yet
02:07 Coke I just fixed it earlier.
02:08 Coke ok. when you run that, are you actually going through the log looking for failures?
02:08 Coke that being 'fulltest_all'
02:12 kid51 Just now, am trying this while the process is still running:
02:12 kid51 egrep -B1 '^Files|running-make-test' 20090419.38216.parrot.test.txt
02:13 kid51 now doing benchmark; everything up to this point passed on Linux i386
02:19 allison joined #parrot
02:23 clunker3 joined #parrot
02:28 kid51 Yes, fulltest_all passed at r38216
02:35 Coke kid51: right. of course it does.
02:35 Coke since it ignores ALL failures...
02:36 Coke (now tons of failures in the example tests...
02:38 Tene purl: msg jonathan Nevermind about that bug... it was a spelling error on my part.
02:38 purl Message for jonathan stored.
02:42 janus joined #parrot
03:02 particle1 joined #parrot
03:06 dukeleto joined #parrot
03:46 dalek parrot: r38220 | allison++ | trunk/docs/book/ch09_pasm.pod:
03:46 dalek parrot: [book] HTML output can render an entity within a code example, but LaTeX
03:46 dalek parrot: can't. Also, Pod parsers other than Pod::Simple can't parse entities in
03:46 dalek parrot: code examples. Removing it, since it's not really necessary to use the
03:46 dalek parrot: symbol for 'pi' in a code example comment. See TT #556.
03:46 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38220/
04:05 dukeleto joined #parrot
04:07 davidfetter joined #parrot
04:59 Infinoid he: FYI, kid51++ tested your asmfun.S patch on darwin and it didn't cause any problems... but we're getting too close to our 1.1.0 release (Tuesday) and just to be safe, I don't want to destabilize anything.  I've got it in my local patch stack, I'll commit after the release.  Are there any other issues on your list?
05:21 Infinoid Do cygwin/strawberry/activestate perl all have a Win32 module?  (I'm looking at https://trac.parrot.org/parrot/​attachment/ticket/471/svn.diff)
05:21 shorten Infinoid's url is at http://xrl.us/bepk7d
05:25 davidfetter patch stack?
05:25 * davidfetter tries to make a bad joke about registers, but fails
05:51 dalek parrot: r38221 | fperrad++ | trunk/NEWS:
05:51 dalek parrot: [release] more NEWS
05:51 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38221/
06:00 uniejo joined #parrot
06:00 nopaste "szabgab" at 212.251.241.27 pasted "building pipp fails r38208 of parrot" (13 lines) at http://nopaste.snit.ch/16325
06:02 masak joined #parrot
06:05 dalek parrot: r38222 | fperrad++ | trunk/docs/project/release_manager_guide.pod:
06:05 dalek parrot: [release] remove useless note about Source0 in parrot.spec.
06:05 dalek parrot: (done by variable substitution %{version})
06:05 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38222/
06:08 szabgab cardinal Configure.pl has --gen-parrot option but the relevant script is missing from build/ directory
06:09 Tene szabgab: feel free to steal what's appropriate from rakudo or elsewhere and post a patch on github
06:10 Tene I'll apply any patches I get there, and grant commit privs to anyone who asks.
06:15 szabgab thanks
06:17 szabgab so I'd be glad to get a commit bit (szabgab on github)
06:18 pmichaud the create_language.pl script in the parrot repo will generate an generic build/gen_parrot.pl that can be used
06:18 pmichaud (tools/dev/create_language.pl)
06:18 Tene szabgab: lemme see if I can remember how to do that... 'sec :)
06:19 szabgab I have to go now anyway... later
06:19 Tene szabgab: added
06:36 iblechbot joined #parrot
07:07 mikehh joined #parrot
07:09 baest joined #parrot
07:18 helozjisky joined #parrot
07:20 krunen_ joined #parrot
07:27 mikehh when I run make fulltest_all I log it - make fulltest_all 2>&1 | tee fulltest_all.38209.log - load the log into Kate and search for Result:
07:27 szabgab joined #parrot
07:27 raiph joined #parrot
07:47 HG` joined #parrot
07:48 Eevee joined #parrot
07:56 cotto allison, ping
07:57 allison hey cotto
07:57 cotto In PIR, should it be possible to give a named register and a label the same name?
07:58 cotto Current IMCC allows it, but it doesn't strike me as a good idea.
07:58 cotto and can produce some misleading error messages
07:59 wayland_ joined #parrot
07:59 allison they're different contexts, so it should be possible, but the misleading error messages mean that IMCC isn't keeping them as separate as it should
07:59 cotto ok.
08:00 allison we had a similar problem with string constants in method/sub calls, and eventually decided to just make them require quotes
08:00 dalek parrot: r38223 | cotto++ | trunk/tools/dev/pbc_to_exe.pir:
08:00 dalek parrot: [pbc_to_exe] avoid using a named register with the same name as a label (no functional changes)
08:00 dalek parrot: This doesn't cause any problems, but it should.
08:00 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38223/
08:00 allison in this case, it's probably an op that accepts either a label or a named variable?
08:02 cotto The problem is that labels and named variables are stored in the same SymHash by imcc.
08:02 allison yes, that's just not right
08:03 cotto The easy fix is to detect duplicates, which I have a patch for.
08:03 cotto The correct fix is pirc. ;)
08:04 allison aye, but the easy fix might be a good patch until we get to pirc
08:04 allison it should at least warn, even if it doesn't die
08:04 cotto The patch causes some new test failures (where labels and named vars collide), so I'll hold off until after the release.
08:04 dalek rakudo: 71033ee | pmichaud++ | docs/spectest-progress.csv:
08:04 dalek rakudo: spectest-progress.csv update: 372 files, 10381 passing, 0 failing
08:04 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​1033eef9fcedd7283843994009c3691f2e5ce8b
08:04 shorten dalek's url is at http://xrl.us/bepmd2
08:05 cotto Thanks.
08:05 allison technically, if it removes old features it has to wait until after 1.4
08:05 allison (if test are failing, that's a good sign it needs a deprecation cycle)
08:05 cotto *sigh*
08:06 allison :)
08:06 allison but, the tests can be fixed now anyway :)
08:06 cotto yes
08:06 cotto and there are only a handful
08:16 GeJ msg Coke Just made a 'fulltest'. Went pretty well except for t/examples/pod.t (most of the errors seem to be in ch09_pasm.pod)
08:16 purl Message for coke stored.
08:19 dalek rakudo: 11352de | jnthn++ | build/gen_setting_pm.pl:
08:19 dalek rakudo: Fix gen_setting_pm.pl changes yesterday to also work on Win32.
08:19 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​1352def7454f9fecbb63a10b49f58544370e93e
08:19 shorten dalek's url is at http://xrl.us/bepmd8
08:19 dalek rakudo: 039f6d5 | jnthn++ | :
08:19 dalek rakudo: Merge branch 'master' of git@github.com:/rakudo/rakudo
08:19 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​39f6d5ee448955a422395b1e08cd316f58ade97
08:19 shorten dalek's url is at http://xrl.us/bepmea
08:23 Ademan joined #parrot
08:30 particle2 joined #parrot
08:31 dalek parrot: r38224 | cotto++ | trunk/t/compilers/imcc/syn/pcc.t:
08:31 dalek parrot: [t] remove an unused named variable
08:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38224/
08:34 dalek parrot: r38225 | cotto++ | trunk/t/library/streams.t:
08:34 dalek parrot: [t] remove an unused named variable
08:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38225/
08:36 cotto bsdz++ #catching accidentally passing tests
08:38 dalek parrot: r38226 | cotto++ | trunk/t/pmc/class.t:
08:38 dalek parrot: [t] remove another unused named variable that's later overwritten
08:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38226/
08:53 rblasch joined #parrot
09:28 helozjisky joined #parrot
09:29 Infinoid joined #parrot
09:29 cognominal joined #parrot
10:00 he joined #parrot
11:12 * bacek nicely messed up with textarea in Trac :-/
11:36 TiMBuS joined #parrot
11:48 kid51 joined #parrot
11:50 dalek parrot: r38227 | bacek++ | trunk/src/pmc/fixedpmcarray.pmc:
11:50 dalek parrot: Use VTABLE method instead of direct accessing attributes in FixedPMCArray.sort. Closes TT#218.
11:50 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38227/
11:55 dalek rakudo: cb2b28b | jnthn++ | src/ (2 files):
11:55 dalek rakudo: First cut of traits on variables; impelment of trait verb so 'my $x of Int' and similar now work.
11:55 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​b2b28b87b953b671979fd245a1cefb04782eb4b
11:55 shorten dalek's url is at http://xrl.us/bepmn9
12:06 nopaste "kid51" at 68.237.3.168 pasted "make fulltest_all results on Darwin/PPC at 38219 (excerpts)" (608 lines) at http://nopaste.snit.ch/16326
12:09 kid51 Of the failing tests reported in that paste, I got the same failures on Linux/i386 for t/examples/pod.t and t/examples/streams.t at r38216.
12:10 kid51 So, 'make fulltest_all' *does* report un-TODOed test failures.
12:21 Coke .
12:22 Infinoid ;
12:22 pmichaud drat, missed kid51
12:22 Coke Infinoid: that declares my presence to purl without saying anything committal.
12:22 Coke and look, I have a msg.
12:23 Infinoid Coke: same here, but no msg :(
12:25 Coke ah. the pasm test for ch09 is failing because chapter 9 has horribly broken pir. =-)
12:25 Coke s/pir/pasm/
12:26 * Coke wonders how much bigger he can make the IF YOU EDIT THIS FILE, FIX THE TEST message for the streams test.
12:27 ruoso joined #parrot
12:30 * Coke throws fperrad under the bus on that one.
12:30 bacek purl: msg Infinoid good morning :)
12:30 purl Message for infinoid stored.
12:30 Coke msg bacek dobra otra
12:30 purl Message for bacek stored.
12:31 bacek Coke: it's night in AU :)
12:31 Coke um...
12:31 Coke dobra vecher?
12:31 * Coke guesses based on the little bulgarian he remembers.
12:32 * Coke thinks that's close.
12:32 * Coke finds "'dobry 'vecher"
12:32 * Coke declares victory.
12:33 Coke . o O (ya ne govo'ryu po russki)
12:33 bacek Coke: it's can be fixed :)
12:35 Coke I had some familiarly with cyrillic after poking at bulgarian. I just need a language parter, i would think.
12:36 Coke I have no desire to add another language to the list I've dabbled with, though. be nice to add something I can claim fluency in. =-)
12:36 bacek Coke: добро пожаловать в мир русского языка :)
12:37 Coke my client renders that as a bunch of \u's. =-)
12:37 * Coke digs.
12:39 bacek Coke: it's weird... irssi in non utf8 console?
12:40 Infinoid in my case, it looks like greek stuff, irssi handles it fine but my terminal doesn't have the right font for it
12:42 Coke bacek;yes. normally it would render, but I'm running in screen, which seems to break a lot of that for me.
12:43 Infinoid So https://trac.parrot.org/parrot/changeset/38030 broke rakudo.  Was Ref deprecated?
12:43 Infinoid ah, it was
12:53 bacek Look like Lyle forgot to run "make clean" in rakudo between makes.
12:57 * Coke wonders if he should submit a PLATFORMS update that lists failures in make examples_tests.
12:57 * Coke guesses yes.
12:59 iblechbot joined #parrot
13:01 dalek parrot: r38228 | coke++ | trunk/PLATFORMS:
13:01 dalek parrot: [docs] update osx/intel results.
13:01 dalek parrot: All platforms are probably having the failures in 'make fulltest',
13:01 dalek parrot: not just this one.
13:01 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38228/
13:01 bacek We need access to interp->op_lib for emitting PBC from PCT. Any ideas about best way of implementing it?
13:02 rdice joined #parrot
13:04 Infinoid Does anyone have a realname for darbelo?  I'm updating CREDITS for the recent patches I've applied.
13:06 Infinoid Found it.
13:07 gryphon joined #parrot
13:11 pmichaud bacek: (Lyle forgetting "make clean")  -- we're getting the same error under Darwin here, and I guarantee we've been running make realclean between makes.
13:11 pmichaud (on both parrot and rakudo)
13:12 pmichaud also, the problem we saw under Darwin wasn't tied to 38030 -- I saw failures in earlier commits.  So, I don't know.
13:12 pmichaud we'll want/need some more Darwin and Vista tests, I guess.
13:12 LylePerl I ran make realclean before each rebuild
13:14 LylePerl I've updated the ticket with all the different versions of parrot I built, and which ones would build Rakudo
13:14 LylePerl On the ones that wouldn't the error was always exactly the same
13:17 Infinoid LylePerl!
13:17 Infinoid I just read http://perl.bristolbath.org/blog/lyle/200​9/04/first-perl-6-experiences-part-2.html
13:18 shorten Infinoid's url is at http://xrl.us/bepmu4
13:18 Infinoid You made me sound like a hero... but really, you were the extremely patient one.
13:18 LylePerl Infinoid: I had no idea what I was doing. It was all down to you really
13:19 Infinoid not everyone has the stomach for that gdb stuff :)
13:19 LylePerl I just hope one day I understand it :)
13:20 Infinoid speaking of gdb... I'd really love to get some more info on that "[perl6_s1.pbc] Error -1073741819" failure
13:20 LylePerl Me too, I can't build the latest Rakudo.
13:20 pmichaud on darwin it was getting an invalid address
13:20 Infinoid hmm, bad pointer?
13:20 pmichaud as part of loading perl6_group
13:22 Infinoid LylePerl: Are you still building with mingw?  Do you still have gdb installed?
13:22 LylePerl Would you be able to talk me through it later? I can't right now, I've got to get things ready for my class. But I should be available from 8pm BST (it's 2:22pm now)
13:23 Infinoid Sounds good.
13:23 LylePerl :)
13:24 LylePerl Yes, Mingw and gdb
13:28 pmichaud LylePerl++ # excellent blog article
13:29 moritz where?
13:30 pmichaud http://perl.bristolbath.org/blog/lyle/200​9/04/first-perl-6-experiences-part-2.html
13:30 shorten pmichaud's url is at http://xrl.us/bepmu4
13:32 rakudohudson joined #parrot
13:32 moritz indeed, LylePerl++
13:33 megazooch joined #parrot
13:34 rg joined #parrot
13:35 rg joined #parrot
13:35 masak LylePerl++
13:43 dalek rakudo: 07bfffc | pmichaud++ | :
13:43 dalek rakudo: Merge branch 'master' of git@github.com:rakudo/rakudo
13:43 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​7bfffc46439a67b48d1f1986de1d0dc52fa5926
13:43 shorten dalek's url is at http://xrl.us/bepmxv
13:43 dalek rakudo: 480902e | pmichaud++ | src/ (2 files):
13:43 dalek rakudo: Updates to enable .hash to work on Match objects.
13:43 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/4​80902edc93fd79bcfe354038c2cce078b9cb868
13:43 LylePerl Thanks everyone. Now that I've done most of my guide, my next blog post will aim to get more people from B&BPM to help with development
13:43 shorten dalek's url is at http://xrl.us/bepmxx
13:44 pmichaud Speaking of which... we're still looking for a name for Rakudo's April release :-)
13:47 masak pmichaud: did you make the post you talked about?
13:48 dalek parrot: r38229 | Infinoid++ | trunk/CREDITS:
13:48 dalek parrot: Add entries to CREDITS for darbelo++, he++, LylePerl++.
13:48 dalek parrot: Add a little more detail to the entries for rg++, bacek++.
13:48 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38229/
13:51 PacoLinux joined #parrot
13:54 rg infinoid: thanks. however i don't think anyone committed the fp jit workaround yet. pending review by tewk (who seems to be mia)
13:55 rg since you're at it ... i think i'll also have a PLATFORMS update for you. let me update and get you a diff
14:01 Infinoid rg: You submitted it, you get credit for that :)
14:07 nopaste "rg" at 62.216.214.73 pasted "PLATFORMS update for FreeBSD and Solaris" (44 lines) at http://nopaste.snit.ch/16327
14:07 rafl joined #parrot
14:14 dalek parrot: r38230 | Infinoid++ | trunk/PLATFORMS:
14:14 dalek parrot: Apply PLATFORMS update from rg++.
14:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38230/
14:31 PacoLinux joined #parrot
14:35 Coke pmichaud: is the request for a hold on the parrot release still valid?
14:36 Coke (given that rakudo seems to depend on non-release versions of parrot, I would hesitate to delay the release in general.)
14:36 davidfetter joined #parrot
14:39 pmichaud Coke: I'm not sure what to do there.
14:39 pmichaud In theory, I'd like the release versions of Rakudo to be able to build against the appropriate release versions of Parrot
14:39 pmichaud thus, while the git-checkout version of Rakudo might not build against 1.1, it seems that the April 23 release of Rakudo should be able to build against the April 21 release of Parrot.
14:40 pmichaud (that's why Rakudo releases are timed so closely after Parrot releases)
14:40 pmichaud so, if the Parrot release comes out and is completely unable to build Rakudo, it's not possible for me to create a Rakudo release against that Parrot release.
14:41 Coke tcl's "solution" there would be to continue to target the old release.
14:41 Coke perhaps branching if necessary. Not a great plan.
14:46 Andy joined #parrot
14:51 szabgab joined #parrot
14:52 pmichaud well, there's no way that Rakudo's Apr release would be able to target the Parrot Mar release.
14:52 pmichaud Rakudo currently relies on various Parrot improvements.
14:56 Coke I think trying to keep the HLL development generally linear, but with release tracking against parrot is going to be hard without branching.
14:57 Coke not that I'd recommend branching, per se.
14:57 Coke Hopefully someone can bisect to find the error before it's an issue. =-)
14:57 Infinoid They did, it was the Ref deprecation
14:57 pmichaud For Vista it appears to be the Ref deprecation, yes.
14:57 pmichaud For Darwin it wasn't the Ref deprecation.
14:58 Infinoid uh oh
14:58 pmichaud at least not on the machine I was able to test with.
14:58 Infinoid Some backtraces would help, if possible
14:58 pmichaud yes, I don't have ready access to the darwin machine here.
14:59 Infinoid No problem, I'll work with LylePerl on the vista side when he has a moment
15:00 pmichaud Coke: With Rakudo I'm just trying to follow the "recommended plan" on a one-month timescale instead of a 6-month timescale.
15:00 iblechbot joined #parrot
15:00 pmichaud but if the idea is that HLL's should only ever build to a release... that's not quite good for hll developers yet.
15:02 Coke pmichaud: was this darwin/x86 ?
15:03 Coke I have that and can run something for you.
15:06 rafl joined #parrot
15:09 Coke trying with rakudo-latest...
15:11 mikehh just looking at the results of make fulltest_all - failures in make testj and make examples_tests - the rest PASS
15:11 mikehh Ubuntu Intrepid i386 at r38226
15:12 mikehh testj as per TT#530 - passes if I apply rg's patch
15:13 mikehh I think I saw the examples_tests failures reported earlier
15:15 mikehh t/examples/pod.t - Failed tests:  4, 9-10, 80-83, 85, 87, 89-96, 98, 100-101, 103, 105-106, 108-110, 112, 117-119, 121, 153
15:15 Coke mikehh: one of them has a ticket. you can probably open a trac for the t/examples/pod.t
15:15 Coke (the streams.t has a ticket)
15:17 mikehh the streams.t reads an additional 2 lines at the beginning
15:17 jaffa joined #parrot
15:18 jaffa hi
15:18 jaffa Is there a support for structs in Parrot?
15:18 pmichaud Coke: yes, darwin/x86.
15:18 Theory joined #parrot
15:19 Coke pmichaud: failure in 'make test' ?
15:20 rg jaffa: no, parrot isn't that high level
15:20 Coke when you say structs, do you mean c-level structs? something /like/ structs?
15:20 jaffa yes
15:20 Coke "which"
15:21 Coke pmichaud: 'make test' passes here.
15:21 mikehh Coke: ok will check further and then open a ticket - hopefully with a patch or something
15:21 jaffa c level structs
15:22 jaffa Coke, clevel structs
15:22 Coke jaffa: see perldoc src/pmc/*struct*.pmc
15:22 rg jaffa: you could maybe use objects and attributes
15:23 Coke rg: or he can use the PMCs that appear to already wrap the c-level structs?
15:23 particle2 parrot can interact with c functions via the native call interface (nci) and c structs with the managedstruct and unmanagedstruct pmc
15:24 jaffa ok
15:24 rg coke: right, i guess it depends a lot on what you're trying to do.
15:24 jaffa I implemented a small language
15:24 jaffa I just wonder how I could make structs
15:25 jaffa I ran into difficulties , I remember
15:27 rg jaffa: can you point us at some code that illustrates your problem?
15:27 jaffa I use s1, s2 and n1, n2 etc
15:28 jaffa no, I cannot
15:28 jaffa I made a small compiler
15:28 * Infinoid tries to reproduce the rakudo fail on mingw/winxp
15:28 jaffa I wanted to extend it so it could handle structs as well
15:28 jaffa and I used pasm
15:29 jaffa but I could not see how I could declare a variable based on a struct
15:29 rg you mean you're trying to write a c compiler?
15:29 jaffa not c
15:30 jaffa just a compiler which uses some elements of C
15:30 Infinoid Hmm.  On mingw, the "try re-running Configure.pl with the '--gen-parrot' option" lines aren't emitted properly, it just says "'parrot' is not recognized as an internal or external command" and barfs there.
15:31 Coke Infinoid: where were the failures occurring?
15:31 Coke I'm, what, 10 hours into running 'make spectest' and things seem fine so far.
15:31 Infinoid Coke: The vista issue looks like a build failure... http://rt.perl.org/rt3/Tic​ket/Display.html?id=64694
15:31 Infinoid Apparently the darwin failure was something else.  I don't know any details
15:32 Coke jaffa; if you're trying to develop something C-like that isn't actually C, I would just do the structs as a PMC and not try to provide direct access to them.
15:32 Coke build failure of parrot or rakudo?
15:32 jaffa ok
15:32 Infinoid rakudo
15:32 pmichaud Coke:  failure was in build.
15:32 jaffa I did not want direct access to them
15:32 Coke pmichaud: ok. there is no failure in the build for me.
15:32 Coke (on darwin)
15:33 jaffa 'just as a language feature
15:33 pmichaud If 'make test' passes, I'm willing to accept that the problem was local to the machine.
15:33 Coke I can provide more details if you need them.
15:33 jaffa so there is pmc for that ok
15:33 Coke jaffa: you can roll your own PMC, or use a PIR object, or...
15:33 Coke seems like a Hash would be "close enough".
15:33 Infinoid pmichaud: Should it make any difference whether I use --parrot-config=../parrot or let rakudo make its own checkout?
15:33 pmichaud Infinoid: it might, depending on your parrot.
15:33 Coke perhaps a PIR level subclass of hash that restricted the valid set of keys?
15:33 * rg would probably go for a pir object. but then coke has more experience implementing a language ;)
15:34 Infinoid my parrot is svn latest.  (well, git latest)
15:34 estrabd joined #parrot
15:34 Coke pmichaud: I'm did a 'git pull', removed parrot/ , and then ran perl Configure.pl --gen-parrot
15:34 pmichaud Infinoid: that sounds okay.
15:34 pmichaud Coke: yes, you did the correct sequence.
15:34 Coke s/I'm/I/
15:34 jaffa What is the speed of parrot nowadays?
15:34 pmichaud I'm willing to accept that there may have been something odd on the machine itself.
15:34 jaffa has anyone made a test?
15:34 Infinoid 316512 feathers per minute
15:35 pmichaud I'm only getting 280,000 feathers per minute :-|
15:35 Infinoid fsvo "feather"
15:36 pmichaud Given that you've been able to make it work on Darwin, I'm quite willing to retract the hold.
15:38 Coke if you have some kind of smoke test, I can setup a cron job to give you daily reports.
15:38 Infinoid this mingw build will take a while, I'll let you know
15:38 Coke (can just tack it onto my parrot tests)
15:39 pmichaud Coke: I have a script I regularly run to produce the daily status updates -- I might be able to adapt it a  bit.
15:39 Coke might be nice to have a new project setup for rakudo at the smolder site.
15:40 pmichaud yes, I agree.
15:41 Infinoid my build barfed on a Makefile double-quoting issue... g++ -o perl6_group.dll [snip a bunch of objects and libs] -shared ""C:/Documents and Settings/mark.glines/parrot-trunk/libparrot.dll""
15:44 nopaste "coke" at 65.91.151.195 pasted "perl6 spec test failures on darwin." (10 lines) at http://nopaste.snit.ch/16328
15:45 rg infinoid: that sounds familiar. i think there is a ticket that says parrot can not be built in a dir containing spaces
15:45 Infinoid parrot builds fine, it just can't build rakudo in a dir containing spaces
15:45 pmichaud well, that makefile is generated by parrot.
15:46 Infinoid actually, the Makefile is fine
15:46 Infinoid I removed the $PATHQUOTEs from tools/build/dynpmc.pl:116, and now it's gettting farther
15:46 Infinoid now I think I need to do the same for the ops script
15:47 pmichaud I think those scripts are deprecated, fwiw.
15:47 pmichaud Rakudo needs to evolve beyond them (and we're still looking for the exact mechanism to do that)
15:48 Infinoid ok.  well, I'll patch them anyway, even if they're going away
15:51 Infinoid weird.  in RT #64694, the build log shows parrot running an empty file to generate src/gen_setting.pir, but that's not what happened here
15:51 Infinoid I have a successful build.
15:53 pmichaud Infinoid: what platform?
15:53 Infinoid mingw/winxp
15:53 Coke Infinoid: perhaps a missing dep?
15:54 Coke (so an earlier failure didn't prevent something that depended on it from running?)
15:54 nopaste "Infinoid" at 75.140.33.106 pasted "Portion of my build log (for comparison against RT #64694)" (5 lines) at http://nopaste.snit.ch/16329
15:55 Infinoid it's possible.  looks like the dynpmc.pl and dynoplibs.pl scripts ignore some error returns in child processes, too, though that might have been my imagination
15:56 Infinoid Actually, I think you're right, it must be the aftermath of some other previous failure
15:58 Infinoid erm, that nopaste wasn't very useful.  Let me make a fresh build
15:58 jaffa Coke, thanks
16:03 dalek parrot: r38231 | Infinoid++ | trunk/config/gen/makefiles (2 files):
16:03 dalek parrot: [config] Fix a build issue on mingw when building out of dirs containing spaces.
16:04 dalek parrot: $LIBPARROT already contains quotes, the extra $PATHQUOTEs negate the quoting.
16:04 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38231/
16:14 dalek rakudo: 5919b0c | jnthn++ | src/builtins/guts.pir:
16:14 dalek rakudo: Fix does at the point of declaration. (We're kinda cheating because we should really do it at compile time. This may result in shoes loudly being banged on tables.)
16:14 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/5​919b0ce29ab47f8d056f8e3973bae9041aeb511
16:14 shorten dalek's url is at http://xrl.us/bepnot
16:18 nopaste "Infinoid" at 75.140.33.106 pasted "full rakudo build log on mingw32-winxp (for reference)" (89 lines) at http://nopaste.snit.ch/16330
16:19 clunker3 joined #parrot
16:30 dalek parrot: r38232 | pmichaud++ | trunk/NEWS:
16:30 dalek parrot: More NEWS update.
16:30 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38232/
16:40 acajou joined #parrot
16:43 PacoLinux_ joined #parrot
17:16 dalek parrot: r38233 | jonathan++ | trunk/src/pmc/fixedpmcarray.pmc:
17:16 dalek parrot: [core] Revert r38227 because it's wrong wrong wrong. The address of an array should not change when you push an element onto the array, and the addresses of two empty arrays should not be the same!
17:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38233/
17:19 particle- jonathan: updating tt#218?
17:20 jonathan particle-: re-opened and commented on it just a few seconds ago :-)
17:20 jonathan and pmichaud is writing a test case
17:20 particle- so i see :)
17:20 dalek parrot: r38234 | pmichaud++ | trunk/t/pmc/resizablepmcarray.t:
17:20 dalek parrot: Tests for RPA and commits 38233 (revert of 38277).
17:20 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38234/
17:20 particle- whee!
17:20 pmichaud three tests, actually.
17:20 jonathan nice
17:21 jonathan Hate to just rip out people's work, but it was *very* broken and release is tomorrow.
17:27 particle- that's entirely appropriate
17:33 bsdz joined #parrot
17:51 Infinoid "This process will run forever, or until it is killed, whichever is sooner."
17:53 darbelo joined #parrot
18:02 bsdz Infinoid: hi, you wouldn't know of any way i can look inside my pmc data structure whilst in a gdb/vs debug session? perhaps there's an included function
18:02 pmichaud bsdz: normally "print" does that.
18:03 bsdz pmichaud: i'll give that go.
18:03 Infinoid bsdz: the details of what you want to print will vary a lot depending on what kind of pmc it is
18:04 bsdz tbh it's a rpa. i know the info i want is in the data element. is there a way i can freeze it from C then i could thaw it in pir?
18:04 barney joined #parrot
18:04 Infinoid hmm, maybe the yaml dumper will do what you want
18:05 bsdz does it have a corresponding C function?
18:06 Infinoid or maybe just RPA.get_repr()
18:06 bsdz oh hold i'll try that
18:06 Infinoid assuming all your pmcs stringify properly, that will give you something, at least.
18:08 bsdz something like this? (*((*rpa).vtable)).get_repr()
18:10 Infinoid oh, from C?  try VTABLE_get_repr(interp, rpa)
18:10 HG` joined #parrot
18:15 dalek pipp: af39365 | (Gabor Szabo)++ | README:
18:15 dalek pipp: improve instruction on how to build pipp
18:15 dalek pipp: Signed-off-by: Bernhard Schmalhofer <Bernhard.Schmalhofer@gmx.de>
18:15 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/af39365bd2152b3797093a8e5c4a69943c77ae86
18:15 shorten dalek's url is at http://xrl.us/bepoa3
18:15 dalek pipp: 9951e7c | (Gabor Szabo)++ | Configure.pl:
18:15 dalek pipp: warn about missing fields from parrot_configure
18:15 dalek pipp: Signed-off-by: Bernhard Schmalhofer <Bernhard.Schmalhofer@gmx.de>
18:15 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/9951e7ccc14eca4618eae5a2736330666385ee8b
18:15 shorten dalek's url is at http://xrl.us/bepoa5
18:15 dalek pipp: 123a4b1 | (Gabor Szabo)++ | Configure.pl:
18:15 dalek pipp: eliminating further warnings from Configure.pl
18:15 dalek pipp: Signed-off-by: Bernhard Schmalhofer <Bernhard.Schmalhofer@gmx.de>
18:15 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/123a4b1370a02bfdf562efb7a58ea95d12de0e73
18:15 shorten dalek's url is at http://xrl.us/bepoa7
18:15 dalek pipp: f72b269 | (Gabor Szabo)++ | Configure.pl:
18:15 dalek pipp: more warning on missing parrot_config field
18:16 dalek pipp: Signed-off-by: Bernhard Schmalhofer <Bernhard.Schmalhofer@gmx.de>
18:16 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/f72b2695108c2e192493b455ad867d261082c6e6
18:16 shorten dalek's url is at http://xrl.us/bepoa9
18:16 bsdz Infinoid: thanks. i eventually found Parrot_PMC_get_repr
18:18 Infinoid ah, you're using the extend interface.  ok
18:19 bsdz seems to work. my empty rpa has a strstart = "[  ]" :)
18:22 Infinoid cool
18:24 acajou left #parrot
18:30 Coke just had a failing smolder report on linux i386, r38234
18:30 Coke http://smolder.plusthree.com/app/public_pr​ojects/report_details/20232#first_failure
18:30 shorten Coke's url is at http://xrl.us/bepodu
18:32 rg if i'm allowed a wild guess, that looks like a conflict with an older installed parrot
18:37 Coke That shouldn't be happening anymore. :|
18:37 Coke not necessarily because I know someone fixed it. more youthful optimisim.
18:37 rg it's hard to tell just from a smolder report
18:38 Coke crap. I never did get partcl working with an installed parrot. :|
18:38 Coke if someone has tuits, that would be a lovely project before tomorrow. =-)
18:44 Coke pmichaud: ping
18:45 Coke msg pmichaud can you verify if haud: ping [14:45] [@Coke(+i)
18:45 purl Message for pmichaud stored.
18:45 Coke er.
18:45 Coke msg pmichaud whoops. Can you verify if http://rt.perl.org/rt3/Tic​ket/Display.html?id=47970 is still an issue for you?
18:45 purl Message for pmichaud stored.
18:45 Infinoid darbelo++ # thanks for testing, I'll close #529
18:59 dalek parrot: r38235 | fperrad++ | trunk (3 files):
18:59 dalek parrot: [codingstd] partial revert of r38115,
18:59 dalek parrot: 2 files in examples/streams are handled as exception
18:59 dalek parrot: by t/codingstd/svn_id.t
18:59 dalek parrot: See TT #554.
18:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38235/
19:10 Coke fperrad++
19:14 japhb joined #parrot
19:14 cotto bacek++ for finding old tickets
19:15 japhb joined #parrot
19:24 amoc joined #parrot
19:29 slavorgn joined #parrot
19:30 LylePerl Hi
19:31 LylePerl Back a bit later than I'd hoped
19:31 Infinoid howdy
19:31 particle- tpf gsoc accepted projects list here: http://socghop.appspot.com/org​/home/google/gsoc2009/dukeleto
19:31 shorten particle-'s url is at http://xrl.us/bepomj
19:32 Infinoid LylePerl: I posted a reply to your RT #64694, but basically, what would be really useful is to get a nopaste (or whatever) of your full rakudo build log
19:32 Infinoid I think there's something else that went wrong, earlier than the stuff you pasted
19:33 helozjisky joined #parrot
19:33 LylePerl ok
19:38 fperrad joined #parrot
19:39 * cotto realizes that he'll need to finish updating the PMC docs for darbelo
19:39 Infinoid opbots, trust fperrad
19:39 slavorg Ok
19:39 clunker3 But I do not trust you Infinoid
19:39 slavorgn Ok
19:40 Coke clunker3, die!
19:40 Infinoid clunker3, I do not trust you either.
19:40 * darbelo only now notices that Christoph Otto contracts to cotto.
19:41 cotto yup
19:43 darbelo I guess we can start bonding now :)
19:44 * cotto gets the superglue
19:44 particle1 joined #parrot
19:45 Infinoid darbelo?
19:45 purl darbelo is Daniel Arbelo Arrocha <mailto:dany.arbelo@gmail.com>
19:45 Infinoid darbelo is also into bonding
19:45 purl okay, Infinoid.
19:46 darbelo darbelo is also arbelo@gmail.com
19:46 purl okay, darbelo.
19:46 Infinoid darbelo?
19:46 purl darbelo is, like, Daniel Arbelo Arrocha <mailto:dany.arbelo@gmail.com> or into bonding or mailto:arbelo@gmail.com
19:46 Infinoid beautiful.
19:46 purl beautiful is a love thing
19:46 Infinoid me?
19:46 purl you are Mark Glines <mailto:mark@glines.org> or likes shiny things
19:47 LylePerl infinoid: Just building the latest parrot, then I'll try and build Rakudo and give you all the details
19:47 Infinoid LylePerl++
19:47 cotto darbelo, how familiar are you with the current way PMCs work?
19:52 darbelo I've read the docs and played with some PMCs, mostly easy stuff.
19:55 cotto Much of the current documentation needs to be updated.  I'm currently working on docs/pmc.pod, although there are a couple others too.
19:56 cotto The big change is that the UnionVal shouldn't be used and ATTRs should.
19:58 cotto Hmmm.  It appears at first glance that pmc_reuse leaks whatever's in the old pmc's PMC_data.
19:59 rg should pir and pasm registers be the same or does parrot assign whatever it wants?
20:00 Infinoid they're dynamically assigned in pir
20:02 Infinoid pir has a register allocator, and there are lots of hits for that in the documentation, but I'm not sure where it's actually implemented.
20:03 barney cotto++ for converting PhpResource to ATTRs
20:03 Infinoid ah, compilers/imcc/reg_alloc.c
20:04 rg infinoid: thanks. if they're not the same, that's all i need to know to tell that -t output won't exactly match the .pir file.
20:04 dalek pipp: 0e5bea9 | (Bernhard Schmalhofer)++ | src/pmc/phpresource.pmc:
20:04 dalek pipp: Convert PhpResource to using ATTRs, Courtesy of Christoph Otto
20:04 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/0e5bea9b2204ebc786ec9dcde78a229a7c6e845a
20:04 shorten dalek's url is at http://xrl.us/bepork
20:04 dalek pipp: 83422c3 | (Bernhard Schmalhofer)++ | :
20:04 dalek pipp: Merge branch 'master' of git@github.com:bschmalhofer/pipp
20:04 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/83422c38f4abcfc7d907b1f594e8bfba2d31b60f
20:04 shorten dalek's url is at http://xrl.us/beporn
20:05 rg would you happen to know if the register values shown in -t output is before or after the statement in the same line is executed?
20:06 nopaste "LylePerl" at 78.86.207.240 pasted "Vista Rakudo build" (93 lines) at http://nopaste.snit.ch/16333
20:06 rg looks a lot like before to me
20:07 Infinoid the register values are before, yes.
20:07 Infinoid LylePerl: That looks like a broken dynpmc.pl.  one moment
20:08 Infinoid which perl are you using?
20:09 Infinoid (just wondering how many similarities between our platforms I can rely on)
20:09 Infinoid from the path, I'm guessing maybe activestate?  (mine is strawberry)
20:12 LylePerl ActivePerl 5.10
20:12 dalek parrot: r38236 | fperrad++ | trunk/t/codingstd/c_function_docs.t:
20:12 dalek parrot: [t] allows this test to run on Windows
20:12 dalek parrot: Previously (since r37817) fails with :
20:12 dalek parrot: Trailing \ in regex m/C:/fperrad/Parrot/trunk\/ at t/codingstd/c_function_docs.t line 37.
20:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38236/
20:13 Coke I wonder if that fixes the 5.8 vs. 5.10 discrepancy also.
20:13 Infinoid ok, thanks.  I'm at a loss as to why your dynpmc.pl would act so completely different from mine
20:15 LylePerl what rakudo are you on, I checked out a version from last week so I could build on the older parrot
20:15 Infinoid oh, wait.  no, that's doing the same thing as mine, I was just looking at the wrong section
20:15 LylePerl I'll update my rakudo
20:15 Infinoid this was rakudo HEAD as of a few hours ago
20:19 nopaste "LylePerl" at 78.86.207.240 pasted "Vista Rakudo build latest" (93 lines) at http://nopaste.snit.ch/16334
20:19 Infinoid LylePerl: Ready for some gdb-fu?
20:20 LylePerl yeah! :)
20:20 Infinoid Judging from that make error on the last line, parrot is returning 0xc0000005
20:20 Infinoid which is a little odd.
20:20 Infinoid all right, gdb C:\temp\rakudo\parrot\parrot.exe
20:20 Infinoid run -o perl6_s1.pbc perl6.pir
20:22 nopaste "LylePerl" at 78.86.207.240 pasted "gdb rakudo parrot" (10 lines) at http://nopaste.snit.ch/16335
20:22 LylePerl segfault
20:22 purl well don't DO that, then.
20:22 Infinoid bt
20:23 cotto Infinoid++ #pun
20:23 * Infinoid feels like he's missed something :)
20:24 cotto Infinoid Judging from that make error on the last line, parrot is returning 0xc0000005
20:24 cotto Infinoid which is a little odd.
20:24 Infinoid oh, hah
20:25 LylePerl ?
20:25 Infinoid as opposed to even
20:25 Infinoid the pun wasn't intended :)
20:26 Infinoid cotto: would you mind looking at LylePerl's crash?
20:27 Infinoid http://nopaste.snit.ch/16335 looks like the interp->vtables array may have been allocated with fewer than 83 entries.  And his bisect says it started at r38030, where you removed Ref and SharedRef
20:29 LylePerl Anything else I can try here to get more info from gdb?
20:29 Infinoid absolutely.  as a sanity check, see what you get from "print interp->vtables" and "print interp->vtables[type]'
20:31 cotto Infinoid, np.  I'll look at it later today.
20:31 nopaste "LylePerl" at 78.86.207.240 pasted "gdb vtables prints rakudo parrot" (4 lines) at http://nopaste.snit.ch/16336
20:31 cotto Is that after a make realclean?
20:32 LylePerl It was a make realclean, then I tried to build rakudo, then I tried gdb
20:32 Infinoid I think it's actually a fresh checkout built by the rakudo makefiles, cotto
20:32 LylePerl Shall I realclean, then straight to gdb?
20:33 LylePerl Yes I grabbed the latest with git pull
20:33 cotto Was it a rakudo realclean or a Parrot realclean?
20:33 LylePerl both
20:33 cotto ok
20:33 NotFound PARROT_ASSERT(type < (int)((sizeof interp->vtables) / (sizeof interp->vtables[0])));
20:33 NotFound Adding this assertion before that lines fails
20:34 Infinoid isn't sizeof interp->vtables just sizeof(a pointer)?
20:34 LylePerl If you want I can start from scratch with a whole new folder?
20:34 NotFound Sorry, <=
20:34 NotFound Also fails
20:35 Infinoid NotFound: interp->vtables is a VTABLE **, so sizeof() is probably just 4 or 8, depending on your platform
20:35 NotFound Forget it, is not an array
20:35 purl NotFound, I didn't have anything matching it, is not an array
20:35 Infinoid LylePerl: it sounds like you've already done that, several times
20:35 Infinoid LylePerl: print interp->vtables[type]->_namespace
20:36 Infinoid see if that returns data, or just crashes.
20:36 LylePerl (gdb) print interp->vtables[type]->_namespace $3 = (PMC *) 0x0
20:36 cotto LylePerl, if you haven't already done that, yes, but there's no point if you're just doing something you've done (several times) before.
20:38 LylePerl I'll do it again anyway, can't harm
20:38 Infinoid ok.  so that line will try to call ((PMC*)0)->vtable->set_pmc_keyed_str, which explains the segfault
20:39 NotFound PARROT_ASSERT(type < PARROT_MAX_CLASSES); <-- This one passes
20:39 Infinoid NotFound++
20:40 Infinoid so that doesn't explain why Ref or SharedRef have anything do do with this... maybe we're just lucky.
20:41 NotFound PARROT_ASSERT(interp->vtables[type]->_namespace); <--- Pass
20:41 LylePerl Maybe they are a red herring?
20:41 Infinoid probably.  maybe it changed a data structure or buffer size enough to exceed some boundary
20:42 Infinoid NotFound: It doesn't seem to occur anywhere other than LylePerl's (vista, activestate, mingw) box so far.
20:42 Infinoid I can't reproduce it on winxp (strawberry, mingw)
20:43 NotFound This function is called from dynpmc. I think it lacks a runtime check, better than an assertion.
20:43 Infinoid Sure, we can crash prettier.  I'm a little concerned that we crash at all though :)
20:43 LylePerl I could try ActiveState, xp, mingw on another machine
20:44 LylePerl Or strawberry, mingw, vista on another
20:45 Infinoid if all else fails, we might have to do that.  At the moment, I'm trying to figure out where _namespace is assigned, though, because that shouldn't be null
20:45 NotFound I CAN HAZ BACKTRACE?
20:45 Infinoid LylePerl: Yeah, the output of "bt" wouldn't hurt
20:45 * Coke sees the mentor list approach the GIFT.
20:45 Coke GIFT?
20:45 purl giFT is a daemon the can handle multiple file-sharing protocols loaded via plug-ins. http://www.giftproject.org or German for "Poison" or Dutch for "Poison" or Norwegian for "poison"
20:46 dalek parrot: r38237 | allison++ | branches/pcc_rewiring/src/multidispatch.c:
20:46 dalek parrot: [pcc] Use a resizable Integer array, instead of trying to calculate it
20:46 dalek parrot: based on the length of a string that isn't the same length as the array
20:46 dalek parrot: needs to be.
20:46 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38237/
20:46 Infinoid GNU Interactive Flow-control/Throttling?
20:46 nopaste "LylePerl" at 78.86.207.240 pasted "gdb bt rakudo parrot" (25 lines) at http://nopaste.snit.ch/16337
20:47 Coke Infinoid: penny arcade reference.
20:47 Coke the Greater Internet ****** Theory.
20:48 Coke (it's not that bad, but some people need to relax.)
20:48 Infinoid oh, heh
20:49 Infinoid the only reason it's not that bad is because of incomplete anonymity
20:51 Infinoid what platform, NotFound?
20:51 cotto LylePerl, good news.  The rakudo build segfaults on my system too.
20:52 Infinoid oh, oops, I got lost in scrollback for a sec
20:53 LylePerl That one I was trying from scratch in a new folder just worked :s
20:54 ascent joined #parrot
20:54 LylePerl *Lyle* wants to shoot himself
20:56 LylePerl I don't understand... Is this a realclean issue then??
20:56 Infinoid You could just remove rakudo and parrot entirely, and start over from a fresh checkout to be sure.
20:57 Infinoid (I call that "really really realclean")
20:57 cotto afk
20:57 LylePerl Infinoid: That's what I just did and it worked
20:58 Infinoid ok.  yeah, it's possible that something didn't get cleaned properly
20:58 LylePerl All the other ones I did was realclean parrot, realclean rakudo, build parrot, build rakudo
20:59 LylePerl I feel so stupid, sorry I should have done that sooner. I trusted the realclean
20:59 NotFound Reporting realclean failures is not bad
20:59 moritz realclean usually works for me, but sometimes I just realclean parrot and forget to clean rakudo as well
21:00 Infinoid does rakudo's realclean chain and call the parrot realclean?
21:00 moritz don't think so
21:00 Infinoid I've never used the auto checkout stuff, I use --parrot-config= instead
21:01 LylePerl So the r38029 to r38030 issue was probably just a realclean issue
21:02 Infinoid well, this is good news.  maybe we can have a non-broken 1.1 release tomorrow :)
21:02 LylePerl Maybe realclean was leaving in .o versions of those 2 files and they were being picked up in the build?
21:02 Coke I would imagine rakudo's realclean could cheat and just rm -rf parrot/
21:03 moritz Coke: that's a bad idea, I might have local modifications in there
21:03 rg and it would waste a lot of bandwidth and time to do a full checkout
21:04 moritz maybe there's a compromise
21:04 moritz that just wastes disk space ;-)
21:04 Coke I did say it was a cheat. =-)
21:05 Theory joined #parrot
21:05 moritz checkout out parrot to parrot-svn, and then rm -rf parrot, and svn copy to parrot/ again
21:06 Infinoid if parrot supported separate source and build dirs, that would be a very clean solution
21:08 NotFound I'v tried today to build with parrot installed in a user directory and it works well
21:09 NotFound perl Configure.pl --prefix=/myhome/insparrot ; make ; make install-dev
21:11 NotFound Then you just add /myhome/insparrot/bin to the PATH and configure and build rakudo
21:12 ascent joined #parrot
21:15 ascent joined #parrot
21:16 LylePerl Infinoid and others, sorry I wasted your time
21:16 LylePerl I'm doing a few more builds to see if I can get the problem again and find out what the real cause was
21:16 Infinoid LylePerl, no worries.  Thanks for the successful test report :)
21:18 ascent joined #parrot
21:20 * rg just found what to occupy his spare time with: lots of rakudo spectest failures on freebsd :(
21:21 * moritz got 4 IO related failures
21:41 donaldh joined #parrot
21:41 bsdz hi, is there anyone who might be able to help with a pmc internals question?
21:44 he joined #parrot
21:48 Infinoid bsdz: yes.  Was that the question? :)
21:49 moritz (in general, just ask, instead of asking to ask)
21:52 bsdz cool, when i run find_method on my class it doesn't find anything. find_method does a hash search on ->methods but inspect shows all my methods under ->namespace. is that right
21:55 nopaste "bsdz" at 78.105.78.157 pasted "methods or namespace?" (108 lines) at http://nopaste.snit.ch/16339
21:56 nopaste "he" at 158.38.152.119 pasted "Add NetBSD/arm 4.0 info to PLATFORMS, ref smoke report 20223" (12 lines) at http://nopaste.snit.ch/16340
21:57 particle1 joined #parrot
21:59 contingencyplan joined #parrot
21:59 Infinoid he++ # probably the first testing we've gotten on ARM so far this year!
21:59 particle1 hey, nice! which arm rev?
21:59 he heh.  StrongARM, sa-110
22:00 he Note, this is with parrot 1.0.0, so no jit.
22:00 particle- nice. one step closer to working on a BUG
22:00 particle- there's no jit for arm in any parrot release yet
22:01 Infinoid though jit for arm would be a beautiful thing
22:01 he There appears to be some code for it, but it's not enabled.
22:02 Infinoid I know enough arm asm to be able to imagine some pretty awesome possibilities
22:08 dalek rakudo: 2703c20 | (Moritz Lenz)++ | src/ (2 files):
22:08 dalek rakudo: rename s/readline/get/g, and remove prefix:<=>
22:08 dalek rakudo: That's not fully correct, .get() still does unwanted context dwimmery, but
22:08 dalek rakudo: it does make three of the four failing spectests pass again.
22:08 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​703c2075e5b990f5a32fcd1e217a242d58f074c
22:08 shorten dalek's url is at http://xrl.us/beppe2
22:09 rg he: i highly doubt jit would still work on arm. it's not enabled by default, although it seems to have worked at some point.
22:12 dalek rakudo: ae5e785 | (Moritz Lenz)++ | src/setting/IO.pm:
22:12 dalek rakudo: implement limit to IO.lines
22:12 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​e5e78590857e3351e9191aa31f3f53e1bd08c96
22:12 shorten dalek's url is at http://xrl.us/beppfg
22:12 Infinoid I'm sure we've made some api changes in the meantime
22:13 particle- does llvm jit work on arm? if so, we may have a prototype by the end of the summer
22:15 Whiteknight joined #parrot
22:15 dalek parrot: r38238 | Infinoid++ | trunk/PLATFORMS:
22:15 dalek parrot: PLATFORMS update for netbsd4.0-arm-gcc-4.1.2 from he++.
22:15 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38238/
22:16 rg particle: not according to their release notes :(
22:16 Infinoid presumably they work fine in 32 bit EABI mode
22:18 dalek parrot: r38239 | fperrad++ | trunk/config (2 files):
22:18 dalek parrot: [doc] revert patch from TT #555 (he)
22:18 dalek parrot: perldoc detection broken on Windows XP with Strawberry Perl 5.10.0
22:18 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38239/
22:20 Infinoid uh oh
22:23 Infinoid fperrad: good catch.  I've got strawberry to test with here... but I'll wait until after the release regardless of whether I find a fix
22:25 fperrad infinoid, i want build/upload a CHM with release 1.1.0
22:25 Infinoid chm?
22:25 purl it has been said that chm is but compiled, indexed, html.
22:25 Infinoid ah, nice
22:28 fperrad available on http://parrotwin32.sourceforge.net/
22:28 fperrad built on Windows, but readable on [m]any platform
22:31 bsdz i find my chm manual like a bible :)
22:33 fperrad Coke, t/codingstd/c_function_docs.t has many failures(86/358)
22:33 fperrad do you have a plan before the release ?
22:52 tetragon joined #parrot
22:54 dalek rakudo: 49e00b4 | git++ | src/classes/Object.pir:
22:54 dalek rakudo: Added "get_number" method to Object.pir to allow numification/intification of Perl 6 classes.
22:54 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
22:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/4​9e00b4073895e86580a277a1594b767c775a3ba
22:54 dalek rakudo: cea34fd | (Moritz Lenz)++ | t/spectest.data:
22:54 dalek rakudo: we pass S13-type-casting/methods.t
22:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​ea34fd87f5d302bbe3a7fc78c08b9cbe9caf436
22:54 shorten dalek's url is at http://xrl.us/beppku
22:54 shorten dalek's url is at http://xrl.us/beppkw
23:01 kid51 joined #parrot
23:05 donaldh left #parrot
23:25 Topic for #parrotis now Parrot 1.0 Released | http://parrot.org | 333 RTs left!
23:31 rg hehe. nice :)
23:31 nopaste "LylePerl" at 78.86.207.240 pasted "I feel like I'm taking crazy pills :(" (114 lines) at http://nopaste.snit.ch/16341
23:32 LylePerl infinoid: ping?
23:38 Limbic_Region joined #parrot
23:43 darbelo LylePerl: Have you tried a "svn status" to check for local modifications?
23:45 darbelo or the git equivalent.
23:46 LylePerl I think I've narrowed it down to a realclean problem... one sec I've got another nopaste...
23:46 cotto LylePerl, is this still the problem that's keeping current rakudo from building?
23:47 LylePerl cotto: If I started from scratch with a completely new folder it builds
23:48 LylePerl but cleaning up my old ones it doesn't
23:49 nopaste "LylePerl" at 78.86.207.240 pasted "realclean misses libparrot.dll and the src/ops folder" (53 lines) at http://nopaste.snit.ch/16342
23:50 LylePerl my last 2 nopastes tell the story. realclean isn't doing it's job, at least not on Vista
23:50 cotto Hmmm.  I'll see if that happens for me too.
23:51 cotto That substantially narrows down the source of the brokenness.
23:51 LylePerl darbelo: the diff -qr would pickup local modifications in the svn
23:51 rg lyleperl: btw i've just commented on the ticket you showed me (Null PMC access ...). I can't verify the problem after all and rakudo builds fine on freebsd and openbsd.
23:53 cotto LylePerl++ #taking the crazy pills
23:54 LylePerl rg: you mean #64694? Null PMC access is someone else I think...
23:55 rg i mean #63922
23:56 LylePerl rg: That wasn't me :/
23:57 rg ah right, that was he. sorry, misread the log there.
23:58 LylePerl rg: no worries
23:58 Infinoid LylePerl: pong
23:58 dalek parrot: r38240 | allison++ | branches/pcc_rewiring (2 files):
23:58 dalek parrot: [pcc] Convert PCCMETHODs over to new calling convention internals.
23:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/38240/
23:59 Whiteknight allison++ #excellent work!

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

Parrot | source cross referenced