Camelia, the Perl 6 bug

IRC log for #parrot, 2009-03-07

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
00:19 Whiteknight joined #parrot
00:20 allison joined #parrot
00:23 tetragon joined #parrot
00:34 dalek parrot: r37147 | jkeenan++ | trunk (2 files):
00:34 dalek parrot: We should call '$conf->data->set( has_negative_zero => $has_neg_0 );' in only
00:34 dalek parrot: one location:  just before the close of runstep().  Either we'll assign the
00:34 dalek parrot: original value of $has_neg_0 -- 0 -- or we'll assign the return value of
00:34 dalek parrot: _evaluate_cc_run().  Eliminate two tests as consequence.
00:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37147/
00:36 rurban_ joined #parrot
00:38 dalek parrot: r37148 | jkeenan++ | trunk/t/steps/auto_neg_0-01.t:
00:38 dalek parrot: _evaluate_cc_run() takes 4 arguments.  Provide the 2 missing arguments to each test of this method.
00:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37148/
00:39 wayland76 left #parrot
00:49 Coleoid joined #parrot
00:52 Limbic_Region pmichaud - response sent
00:53 dalek parrot: r37149 | jkeenan++ | trunk/t/steps/auto_neg_0-01.t:
00:53 dalek parrot: Add tests for verbose output.
00:53 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37149/
00:57 dalek parrot: r37150 | jkeenan++ | trunk/config/auto/pod2man.pm:
00:57 dalek parrot: _handle_no_pod2man() is not called anywhere inside runstep() and so may be deleted.
00:57 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37150/
00:59 ron left #parrot
01:02 GeJ Good morning everyone
01:08 bacek_ joined #parrot
01:09 Whiteknight joined #parrot
01:14 kid51 good morning Geraud
01:14 dalek parrot: r37151 | jkeenan++ | trunk/t/steps/auto_alignptrs-01.t:
01:14 dalek parrot: Mock an HPUX special case.
01:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37151/
01:14 cotto I should start a cryogenic suspension company.  I bet it's looking like an attractive option to a lot of people atm.
01:15 kid51 ?
01:15 cotto ecomony
01:15 kid51 purl: cryogenic?
01:15 purl i don't know, kid51
01:15 cotto freeze/thaw for people
01:16 cotto except that thaw works
01:16 kid51 But wouldn't that run into the Storable version problem?   ;-)
01:16 cotto If thaw doesn't work, the problem takes care of itself.
01:18 crythias joined #parrot
01:25 GeJ kid51: heya James.
01:26 Infinoid cotto: sounds like something you'd want 100% test coverage for
01:27 Infinoid happy weekend, all
01:44 Andy joined #parrot
02:03 cognominal joined #parrot
02:05 dalek parrot: r37152 | jkeenan++ | trunk/t/steps/auto_alignptrs-01.t:
02:05 dalek parrot: Last revisions duped what was being tested in auto_alignptrs-02.t.  Hence, added nothing to coverage.  Reverting.
02:05 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37152/
02:10 Coleoid joined #parrot
02:43 silug joined #parrot
02:53 diakopter Coleoid: paste your build error here...
02:53 diakopter er, nopaste
02:54 diakopter do a make realclean then try again
02:55 diakopter Coleoid: oh, it was already a fresh checkout?
02:56 Coleoid "Coleoid" at 76.251.82.128 pasted "ops2pm failure" (9 lines, 663B) at http://sial.org/pbot/35383
02:57 Coleoid Not fresh since switching to bash.
02:58 Coleoid Same error.
02:58 purl i guess same error is on MojoMojo-0.999024
02:59 Coleoid There's no #parrot choice in the sial.org channel dropdown--what's the pastebot for this channel, again?
03:00 diakopter nopaste?
03:00 clunker3 http://pasta.test-smoke.org/ or http://paste.husk.org/ or http://nopaste.snit.ch:8001/ or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/
03:00 purl nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl or at http://www.extpaste.com/ or http://paste.scsys.co.uk (for #catalyst, #dbix-class, #moose  and others)
03:00 diakopter only some of those work for #parrot
03:00 ron joined #parrot
03:01 diakopter what version of Perl did cygwin install
03:02 Coleoid ...
03:02 Coleoid 5/10/0
03:02 diakopter oh, ok, hrm.
03:04 diakopter anyone around to help diagnose a parrot build error?
03:04 Coleoid Nothing obviously (to me :) askew in the detailed version.
03:04 nopaste "Coleoid" at 76.251.82.128 pasted "perl -V" (63 lines) at http://nopaste.snit.ch/15802
03:15 ron pmichaud: ping
03:16 Coleoid seems to be griping about line 60 of bit.ops
03:22 Andy joined #parrot
03:27 Coleoid ...in parrot/lib/Parrot/OpsFile.pm:441
03:33 diakopter Coleoid: I dunno.  #parrot seems quiet tonight
03:35 Coleoid Thanks for the help so far, anyhow.
03:40 Coleoid Hm.  For some reason, some of the *.ops files are saved with PC line endings.
03:41 Coleoid As I reconvert those to UNIX line endings, the errors go away.
03:42 janus joined #parrot
03:42 diakopter interesting
03:44 Coleoid Gonna toast the whole rakudo/parrot folder and make REAL clean.  See if perhaps I had something leftover from round one, despite the earlier make realclean?
03:55 diakopter worth a shot
03:56 Coleoid Got a lot further, then borked on a chmod command, of all things.
03:57 Coleoid diakopter, just double checking--you build without error?
03:58 diakopter I haven't tried in cygwin lately... I'll try now
03:58 Coleoid Thanks!
04:09 dalek parrot: r37153 | jkeenan++ | failed to fetch changeset:
04:09 dalek parrot: Consolidate former 024-version.t into 023-version.t.  In new 024-version.t,
04:09 dalek parrot: test installed-Parrot case.
04:09 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37153/
04:10 diakopter Coleoid: this may be complicated since (I think) I'm using non-cygwin svn to checkout... I'll check in a minute and fix my path if so, and use cygwin svn
04:13 dalek parrot: r37154 | jkeenan++ | trunk/t/configure (6 files):
04:13 dalek parrot: Eliminate an unused variable.
04:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37154/
04:24 dalek parrot: r37155 | jkeenan++ | trunk/t/configure/023-version.t:
04:24 dalek parrot: Add tests for Parrot::BuildUtil::get_bc_version().
04:24 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37155/
04:29 diakopter Coleoid: I had a chmod error also
04:30 Coleoid Hm!
04:30 diakopter CHMOD 0775 "*.dll"
04:30 Coleoid Yes.
04:30 diakopter can't put * inside "" in bash.  you can in cmd.exe
04:30 Coleoid Removing the quotes, it succeeds.
04:32 Coleoid Thanks again.  Not quite a solution, but I feel like you helped me past where I was causing my own problems.
04:33 Coleoid The only chmod in the makefile isn't that one.
04:33 diakopter yeah I'm trying to trace it
04:35 Coleoid How do you build Rakudo's Parrot with -rHEAD instead of -r37000?  Perhaps it's already been traced?
04:35 diakopter found it
04:35 Coleoid :D
04:35 diakopter kid51: are you around
04:35 diakopter I see you committing stuff
04:35 diakopter but you're not on channel
04:36 diakopter anyway, it's in config/gen/makefiles/dynoplibs.in
04:36 diakopter also config/gen/makefiles/dynpmc.in
04:40 Coleoid Thanks again again.  Fixed, rebuilding, going to bed.  I'll see how it worked in the morning.
04:46 japhb joined #parrot
05:10 ascent joined #parrot
05:19 sjn joined #parrot
05:19 mj41_ joined #parrot
05:19 dalek joined #parrot
05:27 silug joined #parrot
05:39 Theory joined #parrot
05:41 Theory_ joined #parrot
05:45 Theory joined #parrot
06:09 Theory joined #parrot
06:30 rurban joined #parrot
06:36 TiMBuS joined #parrot
07:21 bacek joined #parrot
07:38 Tene joined #parrot
07:57 rurban I love the new smolder addon: http://smolder.plusthree.com/app/pu​blic_projects/report_details/18667 tuxdna failed a test because he changed the src
07:57 shorten rurban's url is at http://xrl.us/bein45
09:26 rurban joined #parrot
09:29 barney joined #parrot
10:12 * barney is investigation http://gist.github.com/75298
10:14 cotto interesting
10:16 barney That started between 37060 and 37069
10:18 cotto Is it still happening?  One of those commits was later reverted.
10:18 rurban_ joined #parrot
10:18 cotto nm.  confirmed
10:19 barney 37069 allowed building of Pipp again, like Rakudo
10:19 cotto looks like the Key code is highly suspect
10:19 barney 37060 was last working version
10:19 cotto (or something in that PMC that messes with it)
10:19 barney yep
10:20 cotto I'm testing an unrelated change.  Once it's done I'll be glad to look at the PhpArray break.
10:21 cotto should be a just couple minutes
10:22 cotto Yeah.  I'm pretty sure I know what the problem is there.
10:22 barney cotto++
10:22 cotto PippArrays use a special Key type flag to iterate correctly when used with Iterators.
10:23 cotto The key_integer code is probably choking on that.
10:27 barney looks like an infinite loop between key_integer() and VTABLE_get_integer()
10:28 NotFound Again?
10:28 purl That trick never works! or see "yet again"
10:29 NotFound yet again?
10:29 purl yet again is full too
10:30 cotto If you're surprised, you don't know the Key code very well.  Stay that way.  You'll be happier.
10:36 cotto Who has rakudo commit bits?
10:39 barney I don't
10:40 cotto barney, super easy fix:
10:40 nopaste "cotto" at 96.26.202.243 pasted "phparray fix" (13 lines) at http://nopaste.snit.ch/15805
10:42 cotto It passes the phparray tests, although it might be buggy.  Let me check some stuff.
10:43 dalek parrot: r37156 | rurban++ | trunk/config/gen/makefiles/root.in:
10:43 dalek parrot: [cage] clean parrot_config
10:43 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37156/
10:46 Tene cotto: I have rakudo commit bit.
10:47 dalek parrot: r37157 | rurban++ | trunk/lib/Parrot/Harness/Smoke.pm:
10:47 dalek parrot: [t] TT #323 fix sparc smolder report +32/64
10:47 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37157/
10:48 cotto Tene, good to know.  I may bug you a little later.
10:50 Tene Then I may be awake a little later. :)
10:51 rurban great: nmake (win32) clean cmdline too long
10:56 barney cotto: Looks fine, Pipp tests are passing
10:57 cotto Yeah.  I
10:58 dalek parrot: r37158 | rurban++ | trunk/config/gen/makefiles/root.in:
10:58 dalek parrot: [cage] win32 fix for r37156: cmdline too long
10:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37158/
10:58 cotto I'm not entirely sure if they should, but it's certainly reliable enough to let stand (until a Key refactor breaks something, at least).
10:59 rurban I'm off now for the afternoon. back in the evening
10:59 cotto bye barney
10:59 cotto and now to break some hlls...
10:59 barney thanks cotto
11:00 cotto np
11:00 cotto barney++ for pointing it out
11:02 cotto Tene, looks like a false alarm.  I won't be breaking any HLLs or submitting any patches.
11:02 dalek parrot: r37159 | cotto++ | trunk (3 files):
11:02 dalek parrot: [PMC] partially convert ParrotLibrary to ATTRs (no dynops breakage yet)
11:02 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37159/
11:02 Tene cotto: okay
11:04 dalek pipp: a7e420c | (Bernhard Schmalhofer)++ | src/pmc/phparray.pmc:
11:04 dalek pipp: Fix shifting from PHPArray iterator.
11:04 dalek pipp: Courtesy of Christoph Otto
11:04 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/a7e420c5059060cf139dfdd2522384c1dc653f27
11:04 shorten dalek's url is at http://xrl.us/beioa7
11:06 dalek parrot: r37160 | rurban++ | trunk/t/pmc/packfiledirectory.t:
11:06 dalek parrot: [t] fix new win32 packfiledirectory failure
11:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37160/
11:21 dalek parrot: r37161 | NotFound++ | trunk/examples/tools/pbc_checker.cpp:
11:21 dalek parrot: [examples] refactoring, cleaning and more diagnostics in pbc_checker
11:21 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37161/
11:42 * barney is looking into using lighthouse for tracking Pipp bugs and features
11:52 dalek pipp: 9cc74dd | (Bernhard Schmalhofer)++ | build/PARROT_REVISION:
11:52 dalek pipp: Pipp now works with Parrot r37155
11:52 dalek pipp: cotto++
11:52 dalek pipp: review: http://github.com/bschmalhofer/pipp/commit​/9cc74dde2b16d43f391acb9f30dd0a4d797608dc
11:52 shorten dalek's url is at http://xrl.us/beiocg
11:56 * barney barney likes lighthouse+github,  but trac+github should work as well
11:56 moritz is lighthouse distributed as well?
11:59 barney no. But I just need any free, hosted issuetracker that can communicate with github
12:01 barney An march 19th I'll meet some PHP guys at DevDusk München. I'll ask there, what is commonly used in the PHP world
12:02 baest joined #parrot
12:15 dalek parrot: r37162 | cotto++ | trunk (6 files):
12:15 dalek parrot: [PMC] finish switching ParrotLibrary to ATTRs, also with no HLL breakage
12:15 dalek parrot: I'm not 100% sure about how to deal with the need for pmc_parrotlibrary.h by
12:15 dalek parrot: dynops code, but the naive approach doesn't appear to break anything.
12:15 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37162/
12:20 barney cotto: got http://gist.github.com/75315 after make realclean
12:26 cotto checking...
12:26 Whiteknight joined #parrot
12:27 dalek parrot: r37163 | cotto++ | trunk/MANIFEST.generated:
12:27 dalek parrot: [MANIFEST] fix a creatively sorted filename
12:27 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37163/
12:29 cotto thanks.  looks like time for a reversion
12:37 cotto all better now
12:39 dalek parrot: r37164 | cotto++ | trunk (6 files):
12:39 dalek parrot: [PMC] revert previous ParrotLibrary commits, which did in fact break something
12:39 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37164/
12:41 cotto sleep.  want.
13:00 dalek tpfwiki: bernhard.schmalhofer@gmx.de | Pipp
13:00 dalek tpfwiki: http://www.perlfoundation.​org/parrot/index.cgi?pipp
13:12 dalek parrot: r37165 | jkeenan++ | trunk/t/configure/023-version.t:
13:12 dalek parrot: Switch order of two calls in order to hit cache differently.
13:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37165/
13:23 kid51 joined #parrot
13:24 alvar joined #parrot
13:55 dalek parrot: r37166 | jkeenan++ | trunk/lib/Parrot/BuildUtil.pm:
13:55 dalek parrot: Since pattern match will only succeed if both major and minor versions are captured, we only need to check for one of them before dying.
13:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37166/
14:15 rurban joined #parrot
14:15 dalek parrot: r37167 | rurban++ | trunk/lib/Parrot/Harness/Smoke.pm:
14:15 dalek parrot: [t] TT #323 smolder 32/64 wordsize oops
14:15 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37167/
14:31 dalek parrot: r37168 | jkeenan++ | trunk/t/op/01-parse_ops.t:
14:31 dalek parrot: Adjust inline comment so that syntax highlighting works better.
14:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37168/
14:56 integral joined #parrot
14:56 braceta joined #parrot
15:05 mikehh I was just looking at the smolder tests and 18680 passes 40 more tests that I do on 18681
15:06 mikehh he has 4 more skips
15:07 mikehh he has similar platform at r37167 vs r37168
15:08 mikehh sorry r37185 vs r37167
15:18 kid51 joined #parrot
15:19 Andy joined #parrot
15:19 kid51 mikehh:  There can be different numbers of tests executed/skipped depending on how you configured Parrot before running the smoke test.
15:21 kid51 Prior to report 18680, I configured:  perl Configure.pl --test --configure_trace, which guaranteed that t/postconfigure/05-trace.t was not skipped.
15:22 mikehh ok got you - haven't tried the trace configure yet, was just wondering if I was missing some libraries or perl modules
15:25 mikehh can you run trace as well as optimize?
15:26 kid51 Yes.  They have nothing to do with one another.  configure_trace merely (well, it's not so mere) logs the state of the Parrot::Configure object at the completion of each config step.  Then you can write programs using methods from Parrot::Configure::Trace to extract that state data.
15:26 dalek parrot: r37169 | NotFound++ | trunk/examples/tools/pbc_checker.cpp:
15:26 dalek parrot: [examples] more pbc_checker refactoring
15:26 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37169/
15:27 kid51 --optimize affects how 'make' runs.  --configure_trace is over and done with by the time 'make' is called.
15:30 mikehh I was wondering about - I couldn't find anything about trace in the documentation
15:31 mikehh not that I have gone through everything yet
15:31 kid51 Well, it's probably not in the PDDs because it's just a diagnostic tool for configuration.
15:31 gryphon joined #parrot
15:32 rurban_ joined #parrot
15:32 mikehh so I really don't need it unless I am working on the configuration files
15:33 kid51 No.  perldoc lib/Parrot/Configure/Trace.pm
15:33 kid51 ... which is why that postconfiguration test is usually skipped.
15:36 mikehh ok found it in the html under Perl Modules
15:41 kid51 link?  which file?
15:44 mikehh index -> Perl Modeles -> (under Cofiguration)  lib/Parrot/Configure/Trace.pm
15:45 mikehh s/Modeles/Modules/
15:46 mikehh I haven't read through that documentation yet - I will now
15:51 Tene joined #parrot
15:53 mikehh I might be able to use that info to track down what is going wrong with my AMD64 build
15:55 kid51 Think of P::C::Trace as being a step-by-step approach to the info that ultimately ends up in %Config.
15:55 rurban mikehh: do you have your smolder url?
15:55 kid51 Some attributes are first grabbed from Perl 5 defaults, then overwritten or modified one or more times during configuration.  If you're concerned about such attributes, that's when P::C::Trace might be useful.
15:57 diakopter kid51: did you see the backlog about building on cygwin
15:58 rurban perl -d Configure.pl is also helpful. b postpone auto::xxx::runstep c
15:58 diakopter it might be a Perl thing, so I thought I'd ask you before filing a bug report
15:58 diakopter (a change to a .pl file I mean, to fix it)
15:59 rurban on cygwin I currentyl have an exception.pmc failure and the known pcre-config bug
15:59 rurban oops: t/pmc/exception TODO passed:   30
16:00 mikehh rurban: which one?
16:00 rurban http://smolder.plusthree.com/app/publ​ic_projects/smoke_reports/8?tag=amd64 shows no fails
16:00 shorten rurban's url is at http://xrl.us/beioq6
16:00 diakopter rurban: are you addressing my cygwin statement?  (see the irclog backlog for details.... there's a * inside a double quoted shell expansion...
16:00 mikehh rurban: the AMD64 one?
16:01 rurban yes. we talked about it days ago but I forgot which failure you had. I improved now the smolder attributes
16:03 rurban diakopter: about CHMOD 0775 "*.dll" in bach?
16:03 diakopter yeah..
16:03 rurban I believe I already fixed that a few days ago...
16:05 diakopter oh
16:05 Coleoid rurban, it may be fixed in HEAD, but it was tripping me last night trying to build Rakudo, which is grabbing rev 37000.
16:05 rurban I remember now: chmod is /perl -MExtUtils::Command -e ExtUtils::Command::chmod 0775 "*.dll"
16:06 diakopter it's not fixed in HEAD afaict
16:06 rurban which platform? hpux maybe?
16:06 Coleoid Cygwin.
16:06 diakopter #IF(cygwin or hpux):$(CHMOD) 0775 "*$(LOAD_EXT)"
16:06 rurban ok, so you have not ExtUtils::Command. It is awful anyway and I already added a ticket to fix this.
16:07 diakopter I don't know why the double quotes are necessary in any shell...
16:07 diakopter that's from the makefile .in
16:07 rurban does it work without the quotes? what is your @chmod@?
16:07 Coleoid diakopter, btw, that was the last parrot failure for me.
16:08 dalek parrot: r37170 | NotFound++ | trunk/examples/tools/pbc_checker.cpp:
16:08 dalek parrot: [examples] check annotations segment in  pbc_checker
16:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37170/
16:08 diakopter yes it works w/o the quotes on cygwin (it's bash).  the only shell I know of where it might work would be cmd.exe
16:09 * diakopter runs off to build another runtime
16:09 rurban there's no shell involved since it's using ExtUtils::Command
16:10 diakopter but it's a Makefile. the command line arguments still go through the shell.
16:11 diakopter I mean, I guess I could have a large misunderstanding or wrong assumption here.
16:12 rurban tested and works fine. it's cotto code
16:12 rurban I mean my fix which should appear here soon...
16:12 dalek parrot: r37171 | rurban++ | trunk/config/gen/makefiles (2 files):
16:12 dalek parrot: [cage] fix "*.dll" quoting issue with non-ExtUtils::Command chmod and cp
16:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37171/
16:12 rurban here it is, please try
16:12 diakopter oh, lol; heh.
16:13 diakopter yeah, that's the identical fix Coleoid tried last night; it works.
16:14 mikehh rurban: it was 18602 and 18605
16:15 rurban mikehh: you mean your amd64 failures?
16:15 Coleoid Grand!  I know the separation of Rakudo and Parrot is still recent.  Do we have a Parrot maintenance branch that we can get this fix into, so that Rakudo can point there?
16:15 NotFound Note that "*.dll" is not the same as *.dll when expanded by the shell. "*.dll" is a file with an asterisk in his name
16:15 rurban mikehh: I remember now. it was set_integer_keyed_str
16:17 rurban mikehh: can you paste your myconfig please
16:17 mikehh rurban: yes - and I seemed to be the only one failing that test - it aborted for some reason
16:18 rurban and maybe submit another smoke.
16:21 mikehh I will switch to AMD64 - I am on my i386 platform at the moment - I am not running a VM
16:21 rurban ok, I'll reboot to a knoppix also
16:22 mikehh cu in a minute
16:28 dalek parrot: r37172 | jkeenan++ | trunk/lib/Parrot/Harness/Smoke.pm:
16:28 dalek parrot: Add documentation:  expanded explanation of exported subroutines.
16:28 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37172/
16:29 mikehh joined #parrot
16:30 kid51 diakopter:  No, until now I didn't see that cygwin discussion.
17:00 dalek parrot: r37173 | NotFound++ | trunk/src/exceptions.c:
17:00 dalek parrot: [core] show annotated source file and line in die_from_exception
17:00 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37173/
17:14 mikehh where's nopaste - I just pasted my myconfig there:  http://nopaste.snit.ch/15807
17:15 mikehh still failing the same tests on AMD64: http://smolder.plusthree.com/app/pu​blic_projects/report_details/18688
17:15 shorten mikehh's url is at http://xrl.us/beiowd
17:18 mj41_ joined #parrot
17:23 braceta left #parrot
17:25 rg joined #parrot
17:26 rurban joined #parrot
17:28 mikehh rurban: still failing the two tests: http://smolder.plusthree.com/app/pu​blic_projects/report_details/18688
17:28 shorten mikehh's url is at http://xrl.us/beiowd
17:28 rurban mikehh: thanks
17:29 rurban did you try without --optimize?
17:29 mikehh i pasted myconfig at: http://nopaste.snit.ch/15807
17:30 rg rurban++ # closely checking smoke reports
17:30 mikehh yes a couple of days ago - still failed
17:30 rurban yes, rg: your ppc0 gave me trouble
17:31 rg no, the sparc without bytesize was mine
17:31 mikehh I am going to try with a completely clean co
17:31 rg i was going to give you a patch, but you beat me to it ;)
17:31 rurban btw rg: cyn you please update some native_pbc's. simply do the mk_native_pbc --noconf
17:32 rurban rg: it was an obvious stupid error
17:32 rurban I just interesting microknoppix errors with gcc but gave up
17:32 rg i know. i'm angry at myself i missed it when i checked the suggested patch
17:33 rg i'm currently compiling a sparc32 parrot. after that i'll rebuild the pbc files.
17:33 rg where do you want me to put them?
17:34 rurban thanks. the hints patch should work also I hope.
17:34 rurban hmm, maybe Configure args should also be added to smolder.
17:37 rg i don't think we really want to debug everything from a smoke report. otherwise we'd probably need the whole configuration structure.
17:39 rurban yes. I'm adding myconfig also to the tar
17:41 rurban We need to verify certain configurations with these reports. They give us false assumptions
17:43 rurban TAP::Harness::Archive extra_files      => [ 'myconfig' ] it is
17:44 rg notfound: ping?
17:44 NotFound pong
17:44 rg oh good :)
17:45 rg i take it you didn't see my note here that you could optimize check_overlap if you ran it after sorting the directory
17:45 tetragon joined #parrot
17:46 rg since you're still working on the pbc_checker, you should consider that
17:47 rurban NotFound: since I have currently no 64bit machine it's hard for me to test which change changed the padding. Did you found out?
17:48 NotFound I thinked about that, but I think is better to do check before doing anything else to avoid unexpected failures. And premature optimization...
17:48 NotFound rurban: I don't have 64 bit machine available.
17:48 rurban bad bad
17:49 rurban well, next weekend I'm back home.
17:49 NotFound Will be good to have a collection of pbc from lots of releases in several platforms.
17:50 rg notfound: i don't think it's a premature optimization nor a bad assumption.
17:50 rurban we have those in svn
17:50 rurban we just disabled the tests, so we don't know which of those worked
17:51 NotFound rg: now it works and it's not slow, so any optimization is premature IMO
17:51 NotFound rurban: The t/pbc_native, you mean?
17:51 rurban yes
17:53 NotFound rurban: How can I know from what release where the files generated?
17:54 rurban it's in the header
17:54 rg well i think it would make check_overlap a lot easier to read. but i was just making a suggestion.
17:54 rurban simpliest is perl tools/dev/pbc_heacder.pl pbc
17:55 rurban but ./pbc_dump -h is good also
17:57 NotFound $ ./pbc_dump -h t/native_pbc/integer_3.pbc
17:57 NotFound PackFile_unpack: This is not a valid Parrot bytecode file
17:57 rurban yes, forgot that. I have to fix that. there's even a ticket
17:57 rurban TT #384
17:59 rg notfound: the _3 pbc files are probably ancient
17:59 NotFound rg: the assumption is that sort cannot fail, I don't take that for granted, even if I don't see ant reason to fail.
17:59 rurban not so. kid51 updated them recently
17:59 rg he did? i missed that. sorry.
17:59 rurban and we have only problems >= 4
18:00 rurban for 0.9.0 or so
18:00 rg so it is ancient
18:00 rurban a fairly recent update...
18:01 geof joined #parrot
18:01 rg i wonder if sparc32 produced _3 pbc files
18:01 rurban and endianize works reliably. no problem with that
18:01 NotFound ./pbc_dump -h t/native_pbc/integer_6.pbc
18:01 NotFound PackFile_unpack: This Parrot cannot read bytecode files with version 3.36.
18:01 rurban sparc32 should produce 32
18:01 NotFound pbc_dump is not a great help
18:01 rurban NotFound: sure use perl tools/dev/pbc_heacder.pl
18:02 rurban oops: perl tools/dev/pbc_header.pl
18:02 rg gotta go. bbl.
18:03 rurban pbc_dump is better if you use TRACE_PACKFILE 2 and pbc_dump --debug
18:04 NotFound What can be the reason to have a 3.36 pbc version in that file?
18:04 rurban well, someone (me probably) uploaded them
18:05 mikehh rurban: It is now passing at r37173
18:05 rurban mikehh: strange
18:06 mikehh it still failed at r37172
18:06 NotFound So the change to 3.37 was after 0.9.1 ?
18:06 rurban mikehh: same Configure.pl args?
18:06 rurban NotFound: 3.36 and 3.3.7 were after 0.9.1
18:06 rurban 0.9.1 had 3.34 or 3.35 I believe
18:07 mikehh the onlr difference is I used perl Configure.pl --optimize --test --configure_trace
18:07 rurban did you use --optimize before?
18:07 mikehh yes for the smoke
18:08 NotFound rurban: will not be easier to fix the problem without worrying abouy previous versions?
18:08 rurban well, your logic says, that the bc writer changed the padding for the debug segment around 3.37
18:09 rurban we should understand that and maybe revert that change
18:09 mikehh I did a clean co at r37173 and I used perl Configure.pl --test --configure_trace and the tests prooved
18:09 rurban but better check if the writer really conforms to our specs
18:09 NotFound rurban: no, the logic in the program just says that it does don't know if in later versions will be different.
18:10 rurban ah, <= 0x0325
18:11 rurban ok, so the whole 64bit writing was wrong all the time, which I suspected also with my analysis
18:11 rurban if so then we must fix it.
18:12 rurban before 1.0 if possible
18:12 NotFound Mmmm.... wait a moment... I must recheck that logic.
18:13 mikehh let me try without the --configure_trace and see if it makes a difference
18:14 Theory joined #parrot
18:14 NotFound rurban: anyway, there is some problem with the segments header independent of 64
18:16 rurban yes, I know. But I wanted to postpone that after 1.0. It only affects cross-platform
18:16 NotFound rurban: It looks like after the segment size field there is an alignment to 16 byte, that is not in the spec. And the debug segment does not have it, or has some other field.
18:17 rurban cross-version will be much trickier
18:19 rurban I'll make a new test smolder now for cygwin with added config_lib.pasm. and I have to leave a few hours.
18:22 mikehh the --configure_trace definately makes a difference
18:22 mikehh I have to go now will be back in a couple of hours
18:23 NotFound rurban: I was fooling myself, the extra padding in 64 bits is in 3.36 and not in 3.37
18:24 rurban yep. 3.36 and before. that's what I read. so we have that fixed now at least
18:24 schobes1 joined #parrot
18:25 NotFound At some point I lose the track about what I was doing %-)
18:25 NotFound rurban: then the problem that reamins is the misalignment in the debug segment.
18:34 Psyche^ joined #parrot
18:34 dalek parrot: r37174 | fperrad++ | trunk/src/pmc/string.pmc:
18:34 dalek parrot: [pmc] revert r37095.
18:34 dalek parrot: What fixes it ? No new test.
18:34 dalek parrot: But r37095 breaks ~80 tests on Lua.
18:34 dalek parrot: See http://smolder.plusthree.com/app/pu​blic_projects/report_details/18464 (r37071) and http://smolder.plusthree.com/app/pu​blic_projects/report_details/18555 (r37106).
18:34 dalek parrot: Segfault seems due to GC early freeing.
18:34 shorten dalek's url is at http://xrl.us/beio3t
18:34 shorten dalek's url is at http://xrl.us/beio3v
18:34 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37174/
18:42 dalek parrot: r37175 | NotFound++ | trunk/src/exceptions.c:
18:42 dalek parrot: [cage] drop an incorrect const
18:42 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37175/
19:00 s1n joined #parrot
19:35 Eevee joined #parrot
19:42 rurban joined #parrot
19:43 rurban what do say about this new smolder: http://smolder.plusthree.com/app/pu​blic_projects/report_details/18699
19:43 shorten rurban's url is at http://xrl.us/beipah
19:46 rg i really don't think the configure args are all that helpful
19:47 rurban I also added myconfig and config_lib.pasm to the tar
19:51 rg that's probably more useful
19:51 rg ah look, we have sparc32 :)
19:51 rg http://smolder.plusthree.com/app/pu​blic_projects/report_details/18701
19:51 shorten rg's url is at http://xrl.us/beipa2
19:52 rurban good, and even update native pbc's
19:52 rurban can you send them to the list?
19:52 rg yes. btw. why was string_2 not updated? it's failing :(
19:53 rurban hm, string_2 is not treated with the new logic
19:54 rurban string > 1 is some kind of quirks, untested
19:55 rg if we're not working on it, i think the whole string test should be skipped for now.
19:55 rurban I'm working on it, but I need foreign material, like from you :)
19:56 rg ok, hang on. i'll send you _3 _4 and _6
19:56 rurban And later we should add encoding and charset files there
19:56 davidfetter Util, you rang?
20:02 dalek parrot: r37176 | rurban++ | trunk/lib/Parrot/Harness/Smoke.pm:
20:02 dalek parrot: Document smolder cmdline sending of reports
20:02 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37176/
20:04 rg rurban: sent. happy testing ;)
20:04 rurban rg++ #thanks
20:06 dalek parrot: r37177 | rurban++ | trunk/t/harness:
20:06 dalek parrot: add myconfig and config_lib.pasm to smolder tar reports
20:06 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37177/
20:09 nopaste "rg" at 91.13.140.173 pasted "another Smoke.pm suggestion" (15 lines) at http://nopaste.snit.ch/15809
20:10 rg since you're at it, how about this patch?
20:11 rurban this is already in, I thought. We just miss gccversion when we use cc
20:11 rg this is supposed to improve that
20:11 rg if your compiler is just cc, but we know gccversion we can tell it's gcc and what version.
20:12 rg currenty it only adds the version if the compiler is actually called gcc
20:12 rurban I see what you mean. ccache cl => msvcversion
20:12 rurban ccache cc ...
20:13 rurban rg: I downloaded a opendarwin image for PPC and a PPC simulator on intel, but had no time yet to try it out.
20:15 rg well right now things are looking pretty good on sparc.
20:16 rurban do have also a static lib there? do you use -R? (rpath)
20:16 rurban sorry for my typing mistakes, I'm lying in bed with the laptop on my chest...
20:17 rg btw the test to add -xmemalign=4s doesn't work because byteorder isn't yet known.
20:17 rurban oops, so we have to postpone that even further
20:17 rg i wanted to look into those callbacks anyway, but i haven't had a chance yet
20:17 rurban its a adougherty thing
20:18 rg yes, kid51 said so too
20:18 dalek parrot: r37178 | coke++ | trunk (17 files):
20:18 dalek parrot: APL is now available in its new home:
20:18 dalek parrot: https://paraplegic.googlecode.com/svn/trunk
20:18 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37178/
20:18 rg i think it's using -R
20:20 rg the -Wl,-rpath (or so) didn't work, but i think adougherty had that fixed
20:20 rurban I asked, because we have strange aix failures which I suspect it's caused by using a static lib, and dynpmc's without using rpath
20:20 rurban same as on windows weeks ago
20:21 rg sorry, i've never touched an aix box
20:21 rurban well, we can only hope the one on his aix will show up somewhen
20:21 rurban at least we have his name now :)
20:22 dalek tracwiki: v67 | coke++ | Languages
20:22 dalek tracwiki: https://trac.parrot.org/parrot/wiki/​Languages?version=67&amp;action=diff
20:22 shorten dalek's url is at http://xrl.us/beipeo
20:22 rg coke: tss, no comment ;P
20:25 janus joined #parrot
20:26 rurban It's obvious even without looking at that wiki page: coke moved APL away and noted the new url
20:27 rg right now it is. when you're at some point looking only at a list of page edits, you might wonder.
20:29 rg i'm only teasing him, because some time ago he was the one complaining about bad commit messages ;)
20:29 rurban I have to hack Test::Builder now for the string.t tests...
20:39 alvar_ joined #parrot
20:41 dalek parrot: r37179 | rurban++ | trunk/t/native_pbc/string.t:
20:41 dalek parrot: [t] fix string.t logic by adding dummies for unneeded files
20:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37179/
20:45 dalek parrot: r37180 | rurban++ | trunk/t/native_pbc (6 files):
20:45 dalek parrot: [t] add native_pbcs for 4_be and 8_le by Rolf Grossmann
20:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37180/
20:48 jdv79 joined #parrot
20:49 dalek parrot: r37181 | rurban++ | trunk/t/native_pbc (3 files):
20:49 dalek parrot: [t] add string pbcs for 4_be and 8_le by Rolf Grossmann
20:49 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37181/
20:53 dalek parrot: r37182 | fperrad++ | trunk/runtime/parrot/library (10 files):
20:53 dalek parrot: [SDL] update PIR syntax (part of TT #411)
20:53 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37182/
20:54 allison joined #parrot
20:57 dalek parrot: r37183 | rurban++ | trunk/lib/Parrot/Harness/Smoke.pm:
20:57 dalek parrot: [t] catch more cc or ccache goodies. Patch by Rolf Grossmann
20:57 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37183/
21:00 rurban rg++ # for the gcc+msvc versions
21:01 rg thanks :)
21:03 rurban_ joined #parrot
21:06 rurban__ joined #parrot
21:10 rurban we really should move pbc_info.c to examples/tools
21:11 rurban it's a pure pbc sample
21:12 contingencyplan joined #parrot
21:13 dalek parrot: r37184 | NotFound++ | trunk/src/spf_render.c:
21:13 dalek parrot: [core] avoid segfaulting when printf'ing PMC that return NULL from get_string
21:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37184/
21:30 Whiteknight joined #parrot
21:30 rurban NotFound: How did you find this? Who printed a NULL string?
21:31 NotFound rurban: me while doing some embedding tests
21:31 rurban Ah ok. Thought there's seem deeper problem hidden somewhere
21:32 mikehh rurban: the test passes if I use --configure_trace but fails if I don't
21:32 NotFound ASSERT_ARGS helped
21:33 rurban mikehh: now this deserves a ticket!
21:34 rurban I have no idea why though. It should be completely independent
21:34 rurban please add the smolder urls to the ticket
21:34 mikehh ok I will set one up
21:35 Theory joined #parrot
22:05 rurban_ joined #parrot
22:17 dalek rakudo: 7559fc9 | (Cory Spencer)++ |  (3 files):
22:17 dalek rakudo: Removed PIR of grep and replaced with P6 version.
22:17 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
22:17 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​559fc9e099496e7850aa57e48e59953e31d3ab1
22:17 dalek rakudo: ed4cd14 | (Moritz Lenz)++ | src/setting/Any-list.pm:
22:17 dalek rakudo: whitespaces and modeline in Any-list.pm
22:17 shorten dalek's url is at http://xrl.us/beipph
22:17 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​d4cd1488f6c3481aa7d826bf9299e5561f49438
22:17 shorten dalek's url is at http://xrl.us/beippj
23:36 mikehh rurban: I set up TT#412
23:42 mikehh BTW the only diff in the myconfig files is the date
23:43 mikehh http://smolder.plusthree.com/app/pu​blic_projects/report_details/18710
23:43 shorten mikehh's url is at http://xrl.us/beipxu
23:44 mikehh is with --configure_trace
23:44 mikehh http://smolder.plusthree.com/app/pu​blic_projects/report_details/18712
23:44 shorten mikehh's url is at http://xrl.us/beipxy
23:44 mikehh is without
23:46 mikehh without I get the 2 tests failing, with no failures
23:46 TiMBuS joined #parrot
23:52 rurban mikehh: thanks, got the reports and configs
23:55 rurban mikehh: about your configs: I want to compare it against others
23:58 Limbic_Region joined #parrot

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

Parrot | source cross referenced