Camelia, the Perl 6 bug

IRC log for #parrot, 2009-02-17

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:01 rg seen allison
00:01 purl allison was last seen on #parrot 3 hours, 25 minutes and 26 seconds ago, saying: NotFound: seems sensible
00:02 kjs_ joined #parrot
00:09 AndyA joined #parrot
00:31 dalek parrot: r36804 | allison++ | trunk:
00:31 dalek parrot: [install] Some source files are needed in the devel package for
00:31 dalek parrot: compiling PMCs from an installed Parrot.
00:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36804/
00:36 Whiteknight kj: inter_run.c has been moved
00:36 Whiteknight src/call/ops.c
00:41 Coke why do we build 'installable_parrot' before we try to install the shared libs?
00:42 Coke shouldn't we build and install the shared libs, then link installable_parrot, then install that?
00:42 chromatic Because building and installing are two different steps?
00:43 Coke right. I build and install the shared lib. then I build and install the executable. =-)
00:43 rg i believe the only difference between "normal" and "installable" versions is a hint to the runtime linker where to find the libraries (which doesn't need the libraries in place)
00:43 chromatic Yes, what rg said.
00:44 chromatic You may want to build but not install, so build should not depend on installed.
00:44 Coke on osx, that hint is only honored at link time. so the built installable_parrot links against the build dir.
00:44 chromatic Here's a nickel kid.
00:44 * Coke attempts not to swear.
00:45 chromatic My theory is that no one actually wrote software in the NeXT days, which explains a lot.
00:46 rg pah. not true :P
00:46 rg however, i don't have any experience with mac, so i can't speak to the current state of affairs on osx ;)
00:47 Coke (at chromatic. not swear at chromatic.)
00:48 chromatic Everyone who says "It's just FreeBSD with a shiny GUI, right?" should actually try to write software for it, armed with a working knowledge of POSIX.
00:48 Coke I'm just trying to get shit done here. snark not helpful.
00:48 Infinoid rg: Thanks for the tracwiki patch, I will apply (minus the XXX comment).
00:48 chromatic I don't know what to tell you.
00:49 Coke It may make you feel good about yourself, but it just annoys the crap out of me. Is there a particular reaction you're trying to elicit?
00:49 Coke ok. then don't say anything. =-)
00:49 chromatic There *must* be a way to add link information to the binary to look to an installed location, but I've never been able to find one.
00:49 Infinoid Also, I've got dalek plugins for 8 more parrot-languages on github (fun, gil, lua, markdown, perk, pipp, primitivearc, wmlscript), if anyone cares.
00:50 rg infinoid: sure. i just thought i'd give a reason for removing html tags.
00:50 chromatic It wouldn't entirely surprise me if it were an XCode-only option.
00:50 Infinoid rg: All rss feeds seem to have html tags, with varying levels of escaping.  It's a really common problem for these scripts
00:50 chromatic You'll probably have to read the documentation for the linker about relocations and build-time versus run-time linking.
00:51 chromatic My point is that whatever happens, it's likely weird and Mac OS X specific.
00:52 Tene_ joined #parrot
00:57 Coke ... of course, If I'm doing this in a macport, I could actually install the library first and just link against it before installing parrot.
00:59 chromatic Yeah, that would work.
00:59 GeJ anyone from the rakudo crew?
01:01 Coke shame that we can't build parrot static anymore.
01:03 Coke chromatic: http://www.codeshorts.ca/2007/nov/01/leopard​-linking-making-relocatable-libraries-movin ... looks like -rpath should work.
01:03 shorten Coke's url is at http://xrl.us/begh26
01:04 chromatic Is it a Mach-O binary now?  Or is that only for Mach-O?
01:05 chromatic That looks a lot like what we use on Linux at least.
01:10 Coke the horrors.
01:11 chromatic Yeah, but I thought we already had that in our Makefile.
01:11 chromatic Maybe we never enabled it for Darwin.
01:11 chromatic Maybe that was the source of our static linking woes from several months back.
01:12 Coke probably doesn't help that the default linker on darwin is 'c++'
01:13 chromatic That shouldn't make too much difference, unless it's doing symbol mangling.
01:13 bacek_ joined #parrot
01:14 Coke linux.pm sets rpath. darwin doesn't.
01:14 Coke checking...
01:16 * Coke wonders how hard it is to patch a distro you're mac-porting.
01:17 rg btw. openbsd doesn't build a shared libparrot either. if somebody wanted to commit http://nopaste.snit.ch/15626 it might improve things.
01:18 Coke chromatic: just adding the same rpath as linux has doesn't seem to do a thing.
01:19 Coke rg: is there a ticket for that?
01:19 chromatic Coke, did you run otool on the resulting binary?
01:19 Coke yes.
01:19 Coke rpath          => '-Wl,-rpath=',
01:19 Coke moment.
01:19 Coke installable_parrot: /Users/coke/research/parro​t/blib/lib/libparrot.dylib (compatibility version 0.0.0, current version 0.0.0)
01:20 Coke that -Wl line doesn't appear anywhere when I say "make installable_parrot"
01:21 rg coke: i don't think so. i'm not that much into openbsd. only rurban was playing with a way old version, so i tried it on current openbsd and noticed the missing shared lib.
01:21 chromatic There's an INSTALLABLEPARROT rule in my Makefile with rpath in it.  Maybe you haven't patched enough?
01:21 rg (current being the latest release)
01:22 Coke ah. parrot_is_shared.
01:22 Coke (except it's always shared, because configure.pl is a b(*&#$)
01:23 rg coke: should i make it a ticket?
01:23 Coke so it doesn't get lost, yes please.
01:24 Coke ok, adding parrot_is_shared isn't enough...
01:25 * Coke gives up for now.
01:25 Coke shame we didn't have someone who was an excellent C developer who had access to the platform...
01:26 rg let's hope for getting more widespread attention with 1.0 ;)
01:27 Coke I actually hope we don't have to drop darwin as a core platform for 1.0
01:33 * Coke adds the URL and the note about rpath to the ticket.
01:36 Whiteknight what platform?
01:37 Coke Whiteknight: OS X.
01:37 Coke I was referring to allison, who is underpaid and overworked.
01:48 Whiteknight I need to get access to more systems. I definitely want to get my hands on a Sparc or an IA64 so I can do more GC work on those
01:50 chromatic Jarkko says that AIX would be good too.
01:50 chromatic Anything Solaris with Sun Studio will help.
01:51 Coke we have someone building on AIX. there are smoke reports.
02:13 Theory joined #parrot
02:28 bacek joined #parrot
03:05 dalek parrot: r36805 | Util++ | trunk:
03:05 dalek parrot: Updated svn:ignore property for TT#284 (rename pdump to pbc_dump)
03:05 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36805/
03:12 Theory joined #parrot
03:45 cotto http://dablackmamba.blogspot.co​m/2005/07/think-wait-fast.html
03:46 cotto shorten that
03:46 purl That URL is at http://xrl.us/begiht [dablackmamba.blogspot.com]
03:46 cotto it figures that the one time shorten would be useful, it's mysteriously absent
03:46 Andy joined #parrot
03:48 janus joined #parrot
04:25 TiMBuS joined #parrot
04:31 dalek parrot: r36806 | Infinoid++ | trunk/src/packfile/pf_items.c:
04:31 dalek parrot: [cage] Fix up several codetest failures in src/packfile/pf_items.c.
04:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36806/
04:31 nopaste "Infinoid" at 75.54.150.29 pasted "TODO passes on linux/x86-64" (6 lines) at http://nopaste.snit.ch/15641
05:12 dalek parrot: r36807 | chromatic++ | trunk/t/steps/auto_macports-01.t:
05:12 dalek parrot: [t] Improved skips of Macports configuration step tests; no reason to run them
05:12 dalek parrot: anywhere other than Darwin.
05:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36807/
05:37 s1n what is the spec for the isgt opcode?
05:38 s1n is it documented in docs/
05:39 chromatic perldoc src/ops/cmp.ops
05:39 s1n chromatic: thanks
05:43 chromatic you're welcome
06:03 chid joined #parrot
06:59 uniejo joined #parrot
07:53 uniejo joined #parrot
07:57 uniejo joined #parrot
08:11 iblechbot joined #parrot
08:12 dalek parrot: r36808 | allison++ | trunk/lib/Parrot/Pmc2c/PMC.pm:
08:12 dalek parrot: [install] In an installed Parrot, the .pmc file and the .dump file
08:12 dalek parrot: generated from it may have the same timestamp. Be smarter in deciding
08:12 dalek parrot: whether to regenerate a .dump file by checking if its timestamp is
08:12 dalek parrot: greater than *or equal to* the timestamp of the .pmc file.
08:12 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36808/
08:14 dalek parrot: r36809 | allison++ | trunk/lib/Parrot/Pmc2c/Dumper.pm:
08:14 dalek parrot: [install] Don't hard-code paths to files, as it breaks installed
08:14 dalek parrot: versions.
08:14 purl versions are perls biggest Crap Ever or set automatically
08:14 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36809/
08:15 masak joined #parrot
08:20 Ademan joined #parrot
08:24 integral joined #parrot
08:34 martin joined #parrot
08:38 UltraDM joined #parrot
08:42 mberends joined #parrot
08:43 mberends joined #parrot
08:45 mberends joined #parrot
08:49 dalek rakudo: 00307bb | (Moritz Lenz)++ | t/spectest.data:
08:49 dalek rakudo: add zip.t to t/spectest.data
08:49 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/0​0307bbc91bb597412b5d5421b0d703a82604f7b
08:49 shorten dalek's url is at http://xrl.us/begi3q
09:14 mikehh joined #parrot
09:24 masak has anyone ever got "src/string.c:767: failed assertion 'encoding'" when building the Rakudo included in Parrot 0.9.0?
09:28 masak I received a proto bug report with an attached make.log with this error in it. would it be useful to anyone if I attached that file in a Parrot bug report?
09:28 masak I want to again emphasize that this concerns Parrot 0.9.0, so the problem might not remain.
09:36 dalek parrot: r36810 | rurban++ | trunk/config/init/hints/openbsd.pm:
09:36 dalek parrot: TT #346, openbsd hints fixed for shared. patch by rg
09:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36810/
10:01 dalek parrot: r36811 | allison++ | trunk:
10:03 moritz Infinoid: it seems that dalek choked on allison's last commit message
10:04 dalek parrot: [install] Some generated header files that need to be installed, and
10:04 dalek parrot: altering the installation directory for header files to match Parrot's
10:04 dalek parrot: prefixed include path.
10:04 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36811/
10:05 moritz Infinoid: false alarm, it worked, just took 3 minutes to produce it ;-9
10:17 dalek joined #parrot
10:50 Tene joined #parrot
11:33 masak trac?
11:33 purl trac is a web-based software project management and bug/issue tracking system emphasizing ease of use and low ceremony. It provides an interface to the Subversion revision control systems, integrated Wiki and convenient report facilities.  http://projects.edgewall.com/trac/ or Python, SQLite and ClearSilver or killing killtrac or a bug-tracking tool or at https://trac.parrot.org/parrot/ or slow or REALLY slow
12:03 AndyA joined #parrot
12:17 mberends joined #parrot
12:24 jimmy joined #parrot
12:31 rg1 joined #parrot
12:47 Infinoid moritz: ah, thanks.  Probably a trac delay, or maybe some irc server lag (dalek sees plenty of both)
12:48 Infinoid normally when dalek chokes, it won't even emit the first line...
13:06 iblechbot joined #parrot
13:40 Whiteknight joined #parrot
13:49 dalek joined #parrot
13:57 Whiteknight it's eerily quite in here today, for a release day
13:57 Whiteknight normally there are people in here all morning working out last-minute solutions to various problems
14:03 Infinoid and introducing new bugs :)
14:07 gryphon joined #parrot
14:07 Whiteknight I've been keeping all my work in branches recently, although it seems like the branches never go well
14:08 Whiteknight the morph branch went well but merging was terrible. The rename_pccinvoke branch is like marching through mud
14:11 Infinoid you're not exactly taking on easy tasks, my friend
14:20 Whiteknight you're right, I should focus on easier things for a while
14:21 Whiteknight to hell with all this C programming.
14:31 Infinoid oh, it's easy to find hard tasks in any language :)
14:32 Infinoid you're doing great stuff though
14:32 Infinoid whereas I'm just sitting around and screwing with dalek, and cleaning the occasional cage
14:37 PerlJam Infinoid++ (not breaking the build on a release day ;)
14:38 moritz I think with that reasoning we can ++ (nearly) everyone in here ;-)
14:38 PerlJam sure, why not?  :)
14:39 Whiteknight moritz: that sounds like  dare!  I bet I could break the build in less then 30 seconds
14:40 moritz Whiteknight: everybody with commit privs can
14:40 Whiteknight you take away all my fun
14:40 PerlJam breaking things is easy.
14:40 Infinoid but fun :)
14:40 moritz purl: breaking things?
14:40 purl i heard breaking things was so HOT
14:41 Whiteknight well, I could "fix" things in a way that causes weird problems later
14:41 Whiteknight or, as I like to call it "doing work on the GC", or "the calling conventions refactor" :)
14:41 dalek parrot: r36812 | Infinoid++ | trunk/tools/build/pbcversion_h.pl:
14:41 dalek parrot: [cage] Add header guards to generated pbcversion.h to pass c_header_guards.t.
14:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36812/
14:42 PerlJam Whiteknight: so ... you feel challenged to break things in ways that are inscrutable so that someone in the future (probably chromatic) will have to spend herculean effort?
14:42 PerlJam s/spend/expend/
14:42 PerlJam :-)
14:42 Whiteknight no, I'm trying my damndest not to break things
14:43 PerlJam Whiteknight: oh, then stay away from thos topics!  ;)
14:43 PerlJam Anyway, Whiteknight++ for tackling "hard" problems.
14:44 Whiteknight haha, dont' give me karma just yet. Tackling the problems is one thing, actually resolving them satisfactorily is another
14:44 PerlJam tackling them is worth karma whether you resolve anything or not IMHO
14:47 Infinoid karma points should be one per tuit, not per commit
14:47 purl points should be one per tuit, not per commit has neutral karma
14:47 Coke purl, scooby snack with hemlock.
14:47 purl Coke: huh?
14:47 Infinoid ok, lets see how badly I broke dalek
14:48 dalek joined #parrot
14:51 Infinoid I've just completely done all of the github plugins to consolidate code, and added a bunch of new ones.  if dalek doesn't report any github commits, that means I screwed up somewhere
14:51 dalek partcl: r326 | wcoleda++ | trunk/src/ (4 files):
14:51 dalek partcl: Track rename of string_length
14:52 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=326
14:52 Infinoid Coke: partcl's plugin is next on the list, got a couple more googlecode languages I'd like to track
14:56 Coke that commit is from a week ago.
14:57 Coke (just fyi)
14:57 Coke very approximately
14:57 dalek parrot: r36813 | rurban++ | trunk/t/codingstd/copyright.t:
14:57 dalek parrot: [cage] typo: transistion
14:57 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36813/
14:58 Coke r36813; if we're checking anyway, I'd make that a todo.
14:58 Coke that test isn't going to break the harness.
15:02 kj joined #parrot
15:07 dalek allison@perl.org | Debian/Ubuntu chroot Environment Setup:
15:07 dalek link: http://www.perlfoundation.org/parrot/index.​cgi?debian_ubuntu_chroot_environment_setup
15:07 dalek Duzy Chan | Parrot:
15:07 shorten dalek's url is at http://xrl.us/beesjm
15:07 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?parrot
15:07 dalek bernhard.schmalhofer@gmx.de | Pipp:
15:07 dalek link: http://www.perlfoundation.​org/parrot/index.cgi?pipp
15:08 Whiteknight Infinoid: I may be getting a new googlecode project set up too, so I'll let you know about that
15:09 Coke what part of DON'T PUT NEW CONTENT HERE is hard to understand?
15:12 davidfetter joined #parrot
15:14 Whiteknight where shouldn't we put new content?
15:15 rg i think he's refering to the old parrot wiki
15:16 Coke http://www.perlfoundation.o​rg/parrot/index.cgi?parrot
15:16 Coke which was just edited.
15:17 Coke I'll see about ruthlesslessly deleted content later in the week.
15:17 Coke *deleting
15:24 masak how's the release going?
15:28 dalek parrot: r36814 | whiteknight++ | trunk/docs/book/ch04_pir_subroutines.pod:
15:28 dalek parrot: [Book] Add a small bit of extra info about the :main subroutine. Needs major expansion on that point.
15:28 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36814/
15:42 dalek parrot: r36815 | fperrad++ | trunk/config/gen/makefiles/root.in:
15:42 dalek parrot: [install] add option --srcdir for install-dev
15:42 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36815/
15:45 kj masak: it's going in a few hours ;-()
15:45 kj :-)
15:46 masak kj: cool.
15:46 Infinoid Whiteknight: I've been scanning http://trac.parrot.org/parrot/wiki/Languages regularly for changes, so post it there and it'll eventually get added to dalek
15:46 kj just about to start running tests
15:49 * moritz does a 'make fulltest'
15:49 Andy joined #parrot
16:00 moritz testf still fails t/op/debuginfo.t as mentioned in my trac ticket. Apart from that it's all green here
16:02 Tene_ joined #parrot
16:06 kj moritz: I think we should SKIP or TODO that test
16:06 kj it's not a big deal there's just something wrong with line numbers
16:06 kj it fails on most cores, except, I think, CGP
16:07 kj so either most cores are wrong, or CGP and the test art
16:07 kj *are
16:07 kj afk # obligatory office coffee break
16:07 masak all green here too, apart from that debuginfo test file.
16:08 rg from what i gathered, most cores don't put their pc in the context, so the line numbers aren't tracked.
16:11 * Coke tries "make TEST_JOBS=3 fulltest"
16:12 bkuhn joined #parrot
16:16 Whiteknight rg is right, not all cores update the current pc in the context
16:16 Whiteknight that shouldn't strictly be necessary, but a real fix won't materialize until after the release
16:31 particle1 um, slow core does, so it shouldn't be wrong.
16:32 NotFound The problem is that several people think that keep it updated will slow down things.
16:32 particle maybe it's just that cgp is default on some platforms, so that's what you see. but slow, trace, gc-debug should all work
16:32 particle yes, it will slow things down
16:33 particle we're not optimizing for speed, we're optimizing for hll developers
16:34 NotFound I agree with that point view, but it seems that there is no broad consensus
16:34 dalek rakudo: f9a9ea6 | pmichaud++ | Configure.pl:
16:34 dalek rakudo: Improve error message when parrot_config not found.
16:34 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/f​9a9ea6da49fc01b4e4cea7237d8b00e712701a6
16:34 shorten dalek's url is at http://xrl.us/begjza
16:35 rg what's bothering me a bit is that for the cores where it doesn't work, the context pc is still initialized and as a result you're getting filename information, but wrong line numbers, which can be confusing.
16:35 rg however i haven't found where that initialization is happening
16:36 particle how's your gdb-fu?
16:36 particle you can set a conditional breakpoint on pc != null
16:36 particle or == 0 or whatever
16:37 rg that assumes i know where the context is created ;)
16:44 Coke t/op/debuginfo.t fails here in fulltest.
16:44 NotFound Someone knows something about examples/pge ?
16:44 Coke (testf)
16:46 NotFound The README mentions all.pir but there is not such file.
16:46 particle try svn log
16:50 * Coke updates TT#281 to "blocker" status for the release.
16:50 pmichaud examples/pge is way out of date.
16:50 Coke let's kill it.
16:51 pmichaud +1
16:51 purl 1
16:54 NotFound Now?
16:54 purl The chalice contains human blood cursed in Death's name. The hooded figure kneeling beside you passes it to you and you have to drink.  LOSE 3 HIT POINTS.
16:55 dalek rakudo: 13b84ef | pmichaud++ | build/Makefile.in:
16:55 dalek rakudo: Temporarily disable the pmc tests until we resolve the ../../parrot issue.
16:56 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​3b84ef799ce5ff12874963b3bc3857b7a51a558
16:56 shorten dalek's url is at http://xrl.us/begj3r
16:58 kj anybody knows what lib/parrot/distribution.pm checks for? I get a WARNING saying inter_run.c is not found
16:58 dalek parrot: r36816 | NotFound++ | trunk/examples/pge:
16:58 dalek parrot: [examples] remove outdated examples/pge
16:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36816/
16:58 kj Whiteknight indicated it was moved I think
16:58 NotFound Killed!
16:59 Whiteknight inter_run.c was moved to src/call/ops.c
16:59 rurban joined #parrot
16:59 Whiteknight Nothing should still be relying on it's old location, you might need to realclean
17:00 kj Whiteknight: yes, thanks for the info. What is lib/parrot/distribution.pm checking that file for?
17:00 kj it's afresh co, so it's as clean as possible
17:00 Whiteknight no idea, I haven't seen that failure.
17:00 kj it s the step: make world docs html
17:00 rurban release already out? I think I broke Sparc 64bit compilation
17:01 kj rurban: no, will do it soon after #ps
17:01 rurban Ok, I'll commit a quickfix then
17:01 kj rurban: thanks
17:01 NotFound src/call/pcc.c src/interpreter.c docs/debug.pod lib/Parrot/Docs/Section/C.pm --> Those files have mentions of inter_run
17:06 kj Whiteknight: by "moved to src/call/ops.c", that means move in the sense of a unix 'mv'?
17:06 kj as in, rename?
17:06 Whiteknight yes, a rename
17:06 purl a rename is still a marvel, as far as rra is concerned.  A Perl script that's its own man page is brilliant. or http://darkuncle.net/rename
17:06 kj ok
17:06 Whiteknight I'm updating now to fix any lingering references
17:07 kj I'd say in section/C.pm, the value should be 'call/inter_run', not 'inter_run'
17:07 kj eh
17:07 kj 'src/call/ops.c'
17:08 dalek parrot: r36817 | whiteknight++ | trunk:
17:08 dalek parrot: [Core] Remove old references to inter_run.c, which has since been renamed to src/call/ops.c.
17:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36817/
17:09 Whiteknight if somebody else could test that commit for me, make sure nothing else is screwed up by it
17:09 kj you removed the reference to it in docs/section/c.pm
17:09 kj why not rename it?
17:09 kj (sorry, lib/parrot/docs/section/c.pm, that is)
17:10 particle NotFound: please update the manifest
17:10 Whiteknight because I don't know enough about that script to know what input it expects
17:10 NotFound particle: ups, sorry
17:10 kj good point, but other equivalent code also contains slashes/ full paths
17:11 kj anyway, i'll try
17:12 kj locally I had changed it to call/ops.c, and that seems to work fine
17:12 Whiteknight okay, that may be a better solution then
17:13 dalek parrot: r36818 | NotFound++ | trunk/MANIFEST:
17:13 dalek parrot: [examples] MANIFEST update forgotten in r36816
17:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36818/
17:17 kj particle: shouldn't perl sources in parrot be copyright Parrot Foundation as well?
17:18 kj for instance, lib/parrot/docs/section/C.pm
17:21 ron joined #parrot
17:23 ron looking at rt 42962 (http://rt.perl.org/rt3/Publi​c/Bug/Display.html?id=42962).  I am on 64 bit ubuntu and get same results as Whiteknight.  Checking smolder reports for cygwin, freebsd and i686 etc all seem to be passing string.t.  Can provide smolder urls if asked.  Any reason not to close?
17:25 Infinoid I've fixed some stuff recently relating to encoding and charset initialization, which may have spontaneously fixed that.
17:26 Infinoid I'd say close it.  If it resurfaces, we can create a TT for it
17:27 Infinoid pmichaud is the original reporter, if it works on his machine then it's definitely closeable
17:32 dalek parrot: r36819 | rurban++ | trunk:
17:32 dalek parrot: TT #308: immediate fix for 64bit big-endian
17:32 dalek parrot: - add two missing functions for 64bit big-endian
17:32 dalek parrot: - change endianize transformers: first endianize, then convert
17:32 dalek parrot: - describe the three floatformats in detail
17:32 dalek parrot: - start with prototypes for 16->8, 12-16, 16->12
17:32 dalek parrot:   and the simplier casting versions for 8->16, 8->12
17:32 dalek parrot: - disable target conversions for non-matching targets (smaller obj size)
17:32 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36819/
17:33 ron left #parrot
17:36 dalek parrot: r36820 | rurban++ | trunk:
17:36 dalek parrot: TT #281: fix make testf for t/op/debuginfo.t 1,7-8
17:36 dalek parrot: - skip tests on -f runcore
17:36 dalek parrot: - nullify current_pc on fast_core on init
17:36 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36820/
17:36 rurban Ok, I'm finished. Sad that nobody looked at my tt312-win32-linking.patch. libparrots.lib is really needed
17:37 kj rurban: would it help if I look at that? (am on windows)
17:37 rurban sure, but it's too late for that now.
17:38 rurban btw: did my glut patch TT#240 help you?
17:38 rg rurban: i think the close for TT #281 is a bit premature
17:39 kj rurban: I think you mean TT#340. Sorry hadn't seen that until now. Will try immediately
17:40 rg also, could it be possible you've got some leftovers of my patch2 in your patch and therefor committed those?
17:40 rurban No, that was on puspose.
17:40 rurban purpose...
17:40 purl purpose is (assuming every channel -must- have a purpose) This channel is for Perl programmers to discuss Perl, and occasionally help new people,  Our poupose is -not- to serve as a substitute for documentation.
17:40 dalek parrot: r36821 | kjs++ | trunk/docs/pdds/pdd19_pir.pod:
17:40 dalek parrot: [pdd19] clarify some text; update text for .HLL_map and add a deprecation note for that.
17:40 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36821/
17:42 rg the problem is that my regexp containing -g also matches --gc-debug and breaks the regular test :(
17:42 rurban I see.
17:43 rurban well, it's not a stable fix. but it fixes make testf, which was considered blocking
17:44 rurban can you note the -g problem in the ticket?
17:45 chromatic joined #parrot
17:46 wknight8111 joined #parrot
17:48 kj rurban: building now with your patch for tt#340; did it work for you? ( i mean, you didn't have the issue, but still works this way?)
17:48 barney joined #parrot
17:49 rg it's failing *now* you will need to back that out.
17:50 rg or at least remove all -g
17:50 rurban ok, will do.
17:50 rg that is |-g from the regexps
17:54 dalek parrot: r36822 | rurban++ | trunk/t/op/debuginfo.t:
17:54 dalek parrot: TT #281, remove -g from r36820
17:54 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36822/
17:54 rurban yes, it's  t/harness --gc-debug --running-make-test -f --runcore-tests
17:54 Coke now the same test fails in 'testb'
17:55 rg yes, and it will also fail in testj
17:55 rurban disable the tests 1,7-8 on which cores now?
17:56 rg uhm testb is the standard core
17:56 rg do you have r36822?
17:56 rurban for me only -g fails now
17:57 nopaste "coke" at 65.91.151.195 pasted "I kid you not." (14 lines) at http://nopaste.snit.ch/15644
17:57 Coke rurban: ah, no, I was one behind.
17:57 rurban ok, jit fails also
17:57 rurban so it's: -f -j and -g to disable
17:58 rurban -j only for 7-8.
17:58 rg that's the ones i know of so far. i'm not sure i've tried all.
17:58 rurban -S fails also 7-8
17:58 rurban fast test: perl  t/harness --gc-debug --running-make-test -S --runcore-tests t/op/debuginfo.t
18:00 ron joined #parrot
18:05 rg i'm not sure this is as important as the ticket implies. we should probably (re?)move the milestone and lower the priority
18:06 Whiteknight Coke: It's going to fail for most runcores I think.
18:07 geof joined #parrot
18:08 dalek parrot: r36823 | kjs++ | trunk/compilers/imcc:
18:08 dalek parrot: [imcc] fix copyright; now Parrot Foundation. minor cleanups (spaces etc.), no functional changes.
18:08 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36823/
18:08 Coke testg
18:08 Coke rg: 'make fulltest' should not be failing on this.
18:08 Coke once those failures are not happening (by hook or by crook) then we can lower the priority and bump the milestone.
18:09 rurban I have it now to /[^-]-[fgjS]/
18:10 rg i think that should be fine for now.
18:11 rurban ok, for 1 its: /[^-]?-[fg]/, for 7-8 /[^-]?-[fgjS]/
18:12 rg did you check jit for 1 ?
18:12 kj rurban: (opengl patch tt#340) didn't work; I have even more undefined references now
18:12 Coke kj: we're still failing a ton of tests in "examples_tests"
18:12 Coke I had submitted a patch before 0.9.0 that was never applied.
18:12 Theory joined #parrot
18:13 rurban jit works for me on 1
18:13 kj Coke: I was making fulltest, but apparently those tests aren't run.
18:13 Coke https://trac.parrot.org/parrot/ticket/200
18:13 kj (unlike the name implies)
18:13 rurban faster test:_ TEST_PROG_ARGS="-j" prove t/op/debuginfo.t
18:13 Coke no, they're not. we're just shipping examples that don't work. =-)
18:13 kj mmmr
18:13 rg rurban: then i guess it works
18:15 Coke rant: the default test harness I'm running doesn't seem to give me a way to figure out how many of the total # of tests passed or failed.
18:15 Coke this seems like a pretty basic metric.
18:15 rurban kj: which glut do you have installed? we have no auto-detection yet. it's on my todo list
18:15 Coke (I can add up the individual files' results, but meh.)
18:16 kj rurban: it said freeglut 4 or something
18:16 rurban maybe you need DISPLAY set, because you have freeglut on X
18:16 kj mm
18:16 kj but why the undefined references?
18:16 rurban okay, you need to set DISPLAY then
18:16 kj can try of course. Any idea how to do that?
18:16 kj export something right?
18:16 rurban because you were missing the libs
18:17 rurban export DISPLAY=:0.0
18:17 kj I don't understand that, really... but that's ok. there's more I don't get :-)
18:17 rurban �h the other way round.
18:17 kj :=?
18:17 purl := is pascal or U+2254
18:17 rurban I haven't touched that in a while
18:17 dalek parrot: r36824 | rurban++ | trunk/t/op/debuginfo.t:
18:17 dalek parrot: [t] TT #281. more needed exceptions for 1,7-8
18:17 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36824/
18:18 pmichaud #parrotsketch in 12
18:18 rurban "export DISPLAY=localhost" e.g. should work also
18:18 kj I just love the speed of gcc compared to msvc
18:19 rurban I love the speed of linux compared to cygwin :)
18:19 kj the last command didn't work.. trying the other one
18:19 kj nope. still same errors
18:19 kj or should I apply the patch as well?
18:19 rurban just set something, the examples will not work then, but compilation will succeed
18:19 rurban the patch is not needed then, no
18:20 rurban thsi is described in README_cygwin.pod btw
18:20 kj mm ok. will read that
18:20 kj I stopped reading after "parrot builds out of the box on cygwin" :-)
18:21 rurban ...but...
18:22 kj ooh. first set DISPLAY, THEN run configure
18:22 rurban yep.
18:23 rurban or fix the libs on config_lib.pasm and lib/Parrot/Config/Generated.pm
18:23 kj I'd rather have it working through the normal way.
18:24 kj for ccache, would I do: perl Configure.pl --cc="ccache gcc" ?
18:25 NotFound This is a problem that need to be addressed. Configure sets libs to be linked, when they are linked the functions are located even if the libraries are not found at runtime.
18:25 rurban yes, there's even a ticket for this
18:25 rurban this annoyed me a lot.
18:26 rurban But I had not time yet to implement it
18:26 NotFound We need a design decision about values to be returned to mean 'failed' and the way to mean 'current process'
18:26 rurban kj: yes
18:27 kj rurban: (yes) well somehow it doesn't always work then. It just started from scratch. weird
18:27 kj rurban++ # cygwin build cleanly!!
18:27 rurban particle tested it last
18:27 NotFound And don't link at compile time things intended to be linked at runtime
18:27 chromatic #ps in 3
18:28 rurban the whole nci header parsing is wrong.
18:29 rurban we need to check if the required libs are working, and need matching header sets then. As with static linkage
18:29 rurban I mean for opengl.pm only
18:30 chromatic #ps in 1
18:30 dalek tracwiki: v59 | barney++ | Languages
18:31 dalek tracwiki: Add matrixy
18:31 dalek tracwiki: https://trac.parrot.org/parrot/wiki/​Languages?version=59&action=diff
18:31 shorten dalek's url is at http://xrl.us/begkdn
18:31 dalek parrot: r36825 | kjs++ | trunk/README_cygwin.pod:
18:31 dalek parrot: [README] add fool-proof message to set DISPLAY before running Configure.pl. That makes noobs such as myself very happy.
18:31 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36825/
18:34 rurban kj: DISPLAY is only needed if you have freeglut installed. the win32 native (and faster and simplier) glut 3 needs DISPLAY to be unset
18:35 kj rurban: oh. ehm. I'll add a note for tha, ok?
18:36 chromatic jhorwitz, ping
18:37 kj nopaste?
18:37 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/
18:37 purl nopaste is, like, 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)
18:37 nopaste "kjs" at 193.1.104.7 pasted "rurban: how does this look?" (4 lines) at http://nopaste.snit.ch/15645
18:40 dalek parrot: r36826 | barney++ | trunk/config/gen/makefiles/languages.in:
18:40 dalek parrot: [languages] add targets co-matrixy and up-matrixy
18:40 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36826/
18:44 dalek parrot: r36827 | barney++ | trunk:
18:44 dalek parrot: [codingstd] let svn ignore languages/matrixy
18:44 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36827/
18:53 dalek tracwiki: v113 | chromatic++ | ParrotRoadmap
18:53 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Par​rotRoadmap?version=113&action=diff
18:53 shorten dalek's url is at http://xrl.us/begkgm
19:02 dalek tracwiki: v114 | chromatic++ | ParrotRoadmap
19:02 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Par​rotRoadmap?version=114&action=diff
19:02 shorten dalek's url is at http://xrl.us/begkis
19:02 dalek parrot: r36828 | rurban++ | trunk/docs/pdds/draft/pdd14_numbers.pod:
19:02 dalek parrot: [docs] pdd14: gmp. bignum <=> bigint
19:02 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36828/
19:11 dalek tracwiki: v60 | kjs++ | Languages
19:11 dalek tracwiki: https://trac.parrot.org/parrot/wiki/​Languages?version=60&amp;action=diff
19:11 shorten dalek's url is at http://xrl.us/begkjk
19:11 dalek parrot: r36829 | rurban++ | trunk/PLATFORMS:
19:11 dalek parrot: [cage] update PLATFORMS from my latest smokes
19:11 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36829/
19:17 * Coke ah craps, as he probably missed PS.
19:17 Coke time?
19:17 purl time is 19:17:13 2009 and (did you mean "clock"?) or flowing like a river
19:17 Coke clock?
19:17 purl Coke: LAX: Tue 11:17am PST / CHI: Tue 1:17pm CST / NYC: Tue 2:17pm EST / LON: Tue 7:17pm GMT / BER: Tue 8:17pm CET / IND: Wed 12:47am IST / TOK: Wed 4:17am JST / SYD: Wed 6:17am EST /
19:17 Coke cruses.
19:17 kj Coke: still chatting there..
19:18 NotFound Coke: blame irc lag ;)
19:20 dalek tracwiki: v115 | chromatic++ | ParrotRoadmap
19:20 dalek tracwiki: https://trac.parrot.org/parrot/wiki/Par​rotRoadmap?version=115&amp;action=diff
19:20 shorten dalek's url is at http://xrl.us/begkkw
19:22 kj Whiteknight: ping
19:22 Whiteknight kj: pong!
19:25 Whiteknight kj?
19:25 purl i guess kj is Klaas-Jan Stol or mailto:parrotcode@gmail.com
19:28 kj Whiteknight: Hi! I just thought of an extra chapter for the book, maybe it's useful, about history
19:28 kj of parrot project
19:29 Whiteknight There is some history in the chapter 1
19:29 Whiteknight Probably needs to be more though
19:29 Whiteknight I also thought that we need a chapter about extending and embedding too
19:29 kj Whiteknight: ok, I was just browsing some docs listing release managers, and figured that might as well go
19:29 kj but losing it altogheter might be a shame, so we could put it in the book somewhere
19:30 kj perhaps. optionally. if you like :-)
19:30 kj Whiteknight: yes, so we can get rid of docs/embed etc
19:31 Whiteknight I don't know if we're supposed to be merging the existing docs into the book. Allison seems to want both, I think
19:31 NotFound Whiteknight: wait some days before starting to write that ;)
19:32 allison history is more of a project docs item then a book item
19:32 allison Whiteknight: and I've been merging book chapters into the project docs, rather than merging project docs into the book
19:35 Whiteknight yeah, I can't keep track of where all information should be going
19:35 Whiteknight so I just write in the book, because I know that the best
19:37 dalek parrot: r36830 | kjs++ | trunk/docs/pdds/pdd19_pir.pod:
19:37 dalek parrot: [pdd19] minor cleanups. remove {{ and }}, and remove vagueness.
19:38 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36830/
19:38 kj rurban: ping
19:42 rurban hi kj
19:42 kj rurban: hi, I nopasted a correction for readme_cygwin
19:42 kj based on your comment. could you check that?
19:42 rurban do you want an updated cygwin release in ports? there's none yet
19:43 kj http://nopaste.snit.ch/15645
19:43 rurban Configure.pl will detect freeglut if freeglut is installed .. is wrong. Configure detects nothing. wait, I check
19:44 kj rurban: but parrot builds fine on cygwin, why would we need that?
19:44 rurban see config/auto/opengl.pm: This is detected if freeglut is installed and DISPLAY is set.
19:44 rurban It requires an X server.
19:45 kj ok, I'll revert my earlier commit then.
19:45 rurban getting opengl right on cygwin tricky currently.
19:45 rurban no, your comment is right.
19:45 purl okay, rurban.
19:46 kj ok
19:46 Whiteknight your comment?
19:46 Whiteknight purl, your comment?
19:46 purl purl's comment is probably right.
19:46 rurban http://nopaste.snit.ch/15645  is right, sorry
19:46 kj rurban: ok but I still have to commit that :-)
19:47 rurban but we already have a similar sentence: Configure.pl will detect freeglut if freeglut is installed and DISPLAY
19:47 rurban is set, otherwise it checks for w32api and opengl with GLUT. Note that DISPLAY
19:47 rurban must be set I<before> running Configure.pl.
19:48 rurban I'll commit your added sentence.
19:48 kj rurban: yes I just added that I<before> part :-)
19:48 kj that's the fool-proof message for people like me
19:49 Coke allison: do you know familiar with command line linker flags for OSX that could address our install issues there?
19:49 NotFound If you try to detect the library by compiling and running something, any program that uses X fails if no DISPLAY
19:49 Coke *know anyone
19:49 Coke fulltest fails on testj on osx/x86
19:49 Coke (why it's even running, I have no idea.)
19:49 NotFound Except for basic things like getting Xlib version and such
19:52 Infinoid allison: https://www.parrot.org/users/infinoid
19:52 NotFound And by the way, is not unusual to compile from a terminal or an ssh without X tunneled and still want to use Xlib, depending on a working DISPLAY is wrong
19:53 Infinoid I'll gather up a list of tasks and tickets and start banging away at them
19:54 rurban Notfound: I know that's its stupid and fragile. That's why I wanted to check that in a better way.
19:54 rurban xemacs started with this logic on cygwin.
19:54 NotFound rurban: I'd like better to have opengl out of core Configure and have his own build process
19:55 rurban sure
19:55 dalek parrot: r36831 | rurban++ | trunk/README_cygwin.pod:
19:55 dalek parrot: [docs] clarify freeglut DISPLAY requirements. kj++
19:55 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36831/
19:55 rurban should I fix 64-bit issues or bignum first?
19:56 NotFound Or maybe better to kill direct opengl usage and let SDL take care of it.
19:57 rurban opengl is a useful core library
19:58 NotFound rurban: is useful, but can be used throug sdl
19:58 rurban and we have the open issue with wchar paths for src/library.c
19:59 rurban fix that in src/library.c, or workaround it with $ENV{TEMP}
19:59 dalek parrot: r36832 | allison++ | trunk/src:
19:59 dalek parrot: [cage] Updating all copyrights in src/ to Parrot Foundation.
19:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36832/
20:00 NotFound rurban: I think that issue is more complex and not just library.c related. Filepaths can come from environment and from command line arguments, for example.
20:01 rurban I thought we got already the encoding support in the path handling
20:01 rurban shouldn't we add more tests for that?
20:02 rurban ha: Parrot_add_library_path only takes const char *path, not STRING* path
20:03 rurban Parrot_locate_runtime_file also takes only char*
20:03 NotFound rurban: we don't know what encoding a path has. It can even not have a valid one.
20:04 rurban on clisp we specify the encoding via cmdline arg. -E
20:04 rurban -E for encoding
20:05 rurban and specializers for teh various encoding types: paths, strings, ...
20:05 chromatic Not every platform has SDL installed, where most every platform on our main platforms has OpenGL installed.
20:06 rurban $ clisp -E --help
20:06 rurban WARNING: *DEFAULT-FILE-ENCODING*: no encoding --HELP, using locale encoding
20:06 rurban WARNING: *PATHNAME-ENCODING*: no encoding --HELP, using locale encoding
20:06 rurban WARNING: *TERMINAL-ENCODING*: no encoding --HELP, using locale encoding
20:06 rurban WARNING: *FOREIGN-ENCODING*: no encoding --HELP, using locale encoding
20:06 rurban WARNING: *MISC-ENCODING*: no encoding --HELP, using locale encoding
20:08 NotFound rurban: if we take the locale encoding and for example you have an utf8 locale and take a path with some part in latin-1 we get an encoding error
20:09 NotFound And if you transcode it, you won't find the directory later
20:09 rurban but the user should be able to specify the path encoding for his arg somehow. ./parrot "my chinese path.pbc"
20:11 rurban can we say that he must use the default encoding utf8 and use all paths as utf8?
20:11 NotFound rurban: not at all
20:11 rurban do we need the default encoding as Configure.pl argument?
20:12 NotFound I'll build you a path completely unusable that way in a moment
20:12 rurban I'll hand over this ticket to you :), okay?
20:13 NotFound rurban: I don't have a solution, I just have some knowledge of the problem
20:13 dalek parrot: r36833 | allison++ | trunk:
20:13 dalek parrot: [cage] Updating copyright in whole repository to Parrot Foundation.
20:13 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36833/
20:13 rurban I would simply check that in the perl5 lib first. A workaround, but no solution.
20:14 rurban TT #326 it is
20:15 kj allison: was DONORS.pod also supposed to change?
20:15 NotFound Maybe in windows the easier solution is to use utf16 for all that things
20:15 rurban do we support that also?
20:16 allison kj: yes, the Copyright line was supposed to change, just not the donor entry
20:16 rurban onyl with icu
20:16 allison kj: (last time I reverted the whole file, which was wrong)
20:17 * japhb rezzes in ...
20:17 kj mmm. the object model picture is changed as well. that's a big chunk in the diff.
20:17 japhb I see a lot of highlighted lines in my IRC ... was there something OpenGL related that needed my input, or was it just general discussion?
20:18 rurban just the usual opengl rambling
20:18 japhb heh
20:20 rurban_ joined #parrot
20:22 dalek parrot: r36834 | allison++ | trunk/docs/submissions.pod:
20:22 dalek parrot: [cage] Caught a reference to "Perl Contributor License Agreement".
20:22 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36834/
20:22 rurban_ hmm, make fulltest also runs make testj on non-jitcapable platforms
20:24 japhb rurban: Given that NCI is broken on supposedly jitcapable platforms, I'm not sure it should be running anywhere ...
20:25 Infinoid Tene_: Looks like fperrad forked your "gil" repo at github.  Should dalek track yours or fperrad's?
20:25 Tene_ fperrad's
20:25 Infinoid oki, thanks
20:25 Tene_ I haven't touched mine since I created it as a demo for a presentation.
20:25 Tene_ about a year ago
20:26 Tene_ I don't know what he's planning to do with it.  It has no goals besides being simple enough to describe the development in about an hour, with questions.
20:27 dalek joined #parrot
20:28 Infinoid Thanks, updated.
20:28 dalek allison@perl.org | Debian/Ubuntu chroot Environment Setup:
20:28 dalek link: http://www.perlfoundation.org/parrot/index.​cgi?debian_ubuntu_chroot_environment_setup
20:28 dalek Duzy Chan | Parrot:
20:28 shorten dalek's url is at http://xrl.us/beesjm
20:28 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?parrot
20:28 dalek bernhard.schmalhofer@gmx.de | Pipp:
20:28 dalek link: http://www.perlfoundation.​org/parrot/index.cgi?pipp
20:28 Tene_ Aw, he's not on IRC for me to ask him.
20:31 dalek partcl: r326 | wcoleda++ | trunk/src/ (4 files):
20:31 dalek partcl: Track rename of string_length
20:31 dalek partcl: review: http://code.google.com/p/p​artcl/source/detail?r=326
20:31 Infinoid Coke: I'm going to fix that.  Honest.
20:35 dalek parrot: r36835 | kjs++ | trunk/ports/debian/copyright:
20:35 dalek parrot: [copyright] fix wrongly changed copyright.
20:35 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36835/
20:39 kj anything that definitly needs to go in 0.9.1?
20:39 Infinoid more cowbell!
20:39 purl I got a fever. And the only prescription is... more cowbell. http://mknx.com/v/cowbell.wmv
20:39 kj I'm going to get myself a new co for the release now
20:40 dalek parrot: r36836 | kjs++ | trunk/examples/subs/jsr_ret.pasm:
20:40 dalek parrot: [copyright] fix a copyright notice.
20:40 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36836/
20:44 NotFound Someone wanted to win the prize for the biggest commit ever? ;)
20:45 dalek parrot: r36837 | kjs++ | trunk/examples/mops/mops.scheme:
20:45 dalek parrot: [copyright] fix a copyright notice.
20:45 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36837/
20:48 Coke NotFound: I think I did that with the string mergeback from six years ago. but I could be wrong.
20:49 dalek parrot: r36838 | kjs++ | trunk/languages/lisp/LICENSE:
20:49 dalek parrot: [copyright] fix yet another copyright notice. allison++ for finding.
20:49 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36838/
20:53 kj should in release.json, web.repository be changed to svn.parrot.org/parrot/trunk, instead of svn.perl.org/parrot/trunk
20:53 kj ?
20:53 chromatic Probably.
20:53 purl Really? Probably? Are you Certain it's not certain? Are you sure it's unsure? I think you need to look harder.
20:54 allison kj: yes
20:54 Coke nclude/parrot/datatypes.h.orig
20:54 Coke whoops.
20:54 purl You just lost the game.
20:54 dalek joined #parrot
20:54 rurban solaris fails to jit, because the cpuarch name is not matching: uname -a SunOS opensolaris 5.11 snv_86 i86pc i386 i86pc Solaris => set P0["jitcpu"], "i86pc"
20:55 * Infinoid dares dalek to output old revisions any more
20:55 rurban i86pc => x86_64 or i386
20:55 Coke is it reasonable to expect jit to work on solaris at all anyway?
20:55 Coke (e.g. jit fails to work on darwin, even though it's i386)
20:56 rurban I'll try now, both 32 and 64 bit.
20:57 rurban but why not. with hcc it shoudl do. with sunpro cc i'm not so sure.
20:57 japhb Coke: JIT doesn't even work on Linux/i386 ...
20:57 rurban gcc of course
20:57 chromatic JIT doesn't work on x86 Linux?
20:57 japhb chromatic: I consider the fact that it seriously breaks NCI to be "not working"
20:58 Coke licensing: docs/project/debian_packaging_guide.pod
20:58 Coke hey, jit
20:58 rurban I also find it rather funny, that jit goes the debian name: amd64, but cpu goes the official name x86_64
20:58 Coke hey, jit was there first, no? =-)
20:59 Coke licensing: src/string.c
20:59 Coke bah.
20:59 Coke (old file)
20:59 Coke apparently 'make distclean' doesn't. =-)
21:00 rg distclean just does realclean
21:02 dalek parrot: r36841 | allison++ | trunk/docs/project/debian_packaging_guide.pod:
21:02 dalek parrot: [cage] Caught stray Perl Foundation copyright mention (thanks to Coke).
21:02 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36841/
21:02 kj I'm at step 2f, but "g" will be to run tools/dev/mk_native.pbc; earlier today there was some mentioning of not doing that. is there a final though t about that?
21:03 bacek joined #parrot
21:03 NotFound You want to skip the g point?
21:03 kj well, there were some people arguing that
21:03 chromatic There's little value in doing that step if we're going to skip the tests.
21:04 Coke are the tests always skipped, though?
21:04 Coke or are they only conditionally skipped?
21:04 NotFound I was making a bad joke X-)
21:05 allison that step was just added to the release guide this month. If it's not stable, skip it.
21:05 kj well, shall I run the "if not possible" alternative?
21:05 kj which is
21:05 allison (and maybe remove it from the the release guide, if it's not going to be stable for 1.0)
21:05 kj run perl tools/dev/pbc_header.pl
21:06 allison kj: that's the old instructions, should be safe
21:07 kj allison: okay, will do that
21:07 allison (but, most of those tests are skipped anyway)
21:07 rurban 1,2,4,5 are alread updated
21:07 rurban 3 is missing, macbook would be fine.
21:07 rurban kid51?
21:07 purl kid51 is now going to live dangerously.
21:07 Coke is 1.0 not going to have portable bytecode?
21:08 Coke then lets just skip it.
21:08 batman joined #parrot
21:08 rurban _3 crashes currently because its too old.
21:08 batman left #parrot
21:08 rurban I'm working on making it portable again
21:09 batman joined #parrot
21:09 rurban we even have full 16-byte long double support in two days
21:09 kj i got failures in make distro_tests (more)
21:09 kj nopaste?
21:09 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/
21:09 purl i guess 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)
21:09 rurban ok, I fixed jit on solaris, but I will wait after the release
21:09 nopaste "kjs" at 193.1.104.7 pasted "make distro_tests failures" (34 lines) at http://nopaste.snit.ch/15649
21:10 kj could anybody have a look at the paste, and give me a pointer?
21:10 rurban true, these are missing
21:10 kj how to fix?
21:11 rurban skip the test. we miss a couple of test files
21:11 rurban at least 2 days work
21:11 chromatic kj, you can ignore those.
21:12 kj chromatic: so the output like this is good enough, to go to the next step?
21:12 chromatic Yes.
21:12 NotFound The test title is misleading
21:13 kj oki
21:18 kj the rel.man.guide is pretty neat
21:19 kj what does this mean: perl Configure.pl --test ... ?
21:19 kj i mean, the --test ... part
21:19 kj (for people who've seen the relmanguide)
21:20 kj or are the ... just there for no reason?
21:21 chromatic If you use --maintainer or --cc= or whatever flags normally.
21:22 kj oh right, i see
21:23 rurban chromatic: what's missing on amd64 jit?
21:23 chromatic I don't know.
21:25 rurban most tests work fine on solaris ia64
21:25 rurban tailcall fails
21:25 chromatic I thought most JIT tailcalls failed.
21:26 rurban it's not native ia64, it's amd64 on solaris, sorry
21:28 batman left #parrot
21:30 rurban I havea bit of a problem when perl is 32bit but parrot 64bit. the jit arch e.g is not matching.
21:31 rurban maye we shoudl check the arch override on the cmdline, like -m64 in ccflags
21:33 dalek parrot: r36842 | fperrad++ | trunk/tools/install/smoke_languages.pl:
21:33 dalek parrot: [install] add language matrixy
21:33 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36842/
21:34 Util_away kj: re: your nopaste (warning: dumping from memory!)
21:34 Util_away For test # 2 in t/distro/test_file_coverage.t, this problem is that the test expects a one-to-one correspondence between each PMC and .t file. The crypto PMCs are not one-to-one with tests, or even with objects!
21:34 Util_away Several crypto .pmc files are generated from config/gen/crypto/digest_pmc.in, which then all combine into a single src/dynpmc/digest_group.bundle file.
21:34 Util All the functions of all the different PMCs in the bundle file are then tested by a single .t file: t/dynpmc/digest.t.
21:34 Util Either test_file_coverage.t needs to become smart enough to handle these "split/bundle" cases (yea!),
21:34 Util or the .t file needs to be split into multiple (nearly identical and redundant; boo!) .t files corresponding to the PMCs.
21:35 rurban splitting the test is not much work
21:36 kj Util: I see, thanks for the explanation
21:36 Util Oh, wait; you are doing the release! Never mind, I will make a trac ticket (as I should have last month when I saw the problem).
21:46 rurban well, there's a lot missing for amd64 jit
21:47 mikehh joined #parrot
21:47 Infinoid which isn't news... amd64 isn't a core platform anyway
21:53 Coke by that measure, neither is darwin. =-)
21:53 NotFound Is a multicore? ;)
21:55 rurban I've cleaned up jit a bit...
21:56 rg you're gonna make it work on amd64? ;)
21:56 Whiteknight joined #parrot
21:56 rurban nope. I need a better debugger there. dbx is awful
21:56 rurban I have to install a better multi-arch platform, like gentoo
21:57 rurban not debian!
21:57 Infinoid gcc-config ftw.
21:57 bacek joined #parrot
21:57 rurban I did something to jit 386 now and wait for the smokes
21:59 rurban there was duplicate code on some areas
22:15 Util Ticket submitted: TT#348 - test_file_coverage.t fails to find tests for MD2 and other crypto PMCs.
22:32 kj got a failure in t/op/string
22:34 kj allison: there's a string failure.
22:34 allison kj: in fulltest? which platform?
22:34 kj i'm on cygwin now
22:34 kj i didn' tkeep track which core this is
22:35 allison ok, I can reboot into windows if needed, what's the failure?
22:35 kj I think -f
22:35 rurban t/op/string fail is expected.
22:35 kj ok
22:35 kj t/op/sprintf2............................ok
22:35 kj t/op/string..............................
22:35 kj #   Failed test 'ge_s_sc_ic'
22:35 kj #   at t/op/string.t line 983.
22:35 kj #          got: ''
22:35 kj #     expected: 'ok
22:35 rurban we should skip it for now.
22:35 kj # '
22:35 kj ok 103 # SKIP Pending rework of creating non-ascii literals
22:35 kj ok 104 # SKIP Pending rework of creating non-ascii literals
22:36 kj rurban: couldyou SKIP it?
22:36 rurban uups. sorry, not this should pass. I mixed it up
22:36 kj ok
22:36 rurban t/native_pbc/string.t can be skipped
22:40 kj @$# more failures.
22:40 kj nopaste?
22:40 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/
22:40 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)
22:40 nopaste "kjs" at 193.1.104.7 pasted "failures :-(" (19 lines) at http://nopaste.snit.ch/15652
22:42 rurban I found a small bug when TRACE_PACKFILE is enabled. harmless for most.
22:42 bacek_ joined #parrot
22:43 rurban hmm, is your machine under heavy load?
22:43 kj well, it's running tests :-)
22:43 kj it's kinda very green
22:43 kj on the icon of the task manager
22:43 kj btw, running perl t/harness -f t/pmc/stringhandle.t passes
22:43 kj so I think it was the heavy load...
22:43 rurban it's falling into a dll load timeout
22:44 Coke_afk kj: I'm sorry you're going to lose your evening to this. Thanks.
22:44 rurban can you do prove t/op/string.t
22:44 kj ah who has time for a social life anyway
22:45 rurban sorry, it was t/pmc/stringhandle.t
22:45 NotFound The test results are complete, looks like an external error.
22:45 rurban ok, go on
22:45 rurban for me all tests pass on cygwin
22:45 kj rurban: I think I did already prove, but that worked out fine
22:45 kj but then I realized it wasnt' the correct core
22:45 chromatic kj, that looks like a Perl harness error, not Parrot.
22:45 rurban very unusual error though, only under very heavy load
22:45 kj but running with fast core works fine
22:46 kj (using perl t/harness)
22:46 dalek parrot: r36843 | rurban++ | trunk:
22:46 dalek parrot: [pbc] Fix when TRACE_PACKFILE is enabled. Use special macro for the converters,
22:46 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36843/
22:46 kj both prove and perl t/harness -f on t/pmc/stringhandle works fine
22:46 kj passes
22:47 NotFound I'll put the blame on fork emulation
22:47 kj yes, sounds reasonable, especially with the cygwin virtual environment
22:47 rurban yes
22:48 nopaste "kjs" at 193.1.104.7 pasted "final output of test" (22 lines) at http://nopaste.snit.ch/15653
22:48 rurban it need to remap the dll into the parent address space. a big memcpy
22:48 kj I'll run the tests again with prove and perl t/harness
22:49 rurban prove t/op/string.t passes?
22:49 kj so far so good.... (running)
22:49 kj my windows task manager thing in the icon tray is mostly green all the time
22:49 kj PASS
22:50 Coke OUT
22:50 kj well I will soon :-)
22:50 rurban blame perl many dll's to be loaded
22:50 rurban maybe I make a static perl once for cygwin.
22:50 NotFound Release right now, I want to apply #343 before going to bed ;)
22:50 Coke kj++
22:51 rurban or just copy away
22:51 kj I'll make tarball soon and upload that (as release candidate, only temp. available for the wiki cleaning purists here ;-)
22:53 nopaste "rurban" at 212.183.63.161 pasted "NotFound: #343 patch -p0 < patches/coretypes.patch" (20 lines) at http://nopaste.snit.ch/15654
22:53 rurban 3 hunks failed
22:53 rurban just the manifests
22:53 nopaste "kjs" at 193.1.104.7 pasted "t/op/string.t and t/pmc/stringhandle.t pass!" (64 lines) at http://nopaste.snit.ch/15655
22:54 rurban did you fix VERSION?
22:54 Coke yay! docs.parrot.org exists.
22:54 rurban did you change the version in parrot.spec?
22:54 kj I did..
22:54 NotFound rurban: I had that problems before, just clean the MANIFEST parts before applying and regenerate later
22:54 kj all according to the guide
22:55 rurban did you fix NEWS: New in February 2009 release (r35855 to r36724) => 0.9.1
22:55 Limbic_Region joined #parrot
22:55 kj rurban: yes otherwise you can't make the fancy release message
22:55 kj (found that out)
22:55 kj I guess I'll be committing then
22:56 rurban is the tarball of almost equal size than the 0.9.0?
22:56 kj dunno. haven't made the tarball yet
22:56 kj (that step comes after commit..)
22:56 rurban go on.
22:56 NotFound Coke: is an A record. I think I've heard it has to be an alias of parrot.org :?
22:57 allison NotFound: it's a separate site
22:57 NotFound Ah, nice
22:57 allison NotFound: just static generated html, at the moment
22:57 allison Coke: huzzah!
22:57 kj here goes..
22:58 rg are you going to put the book up on docs too?
22:58 NotFound Confirmed it has crossed the ocean ;)
22:58 dalek parrot: r36844 | kjs++ | trunk:
22:58 dalek parrot: [release] Update for release 0.9.1.
22:58 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36844/
22:59 rurban The NEWS line was different before: New in January 2009 release (0.9.0)
22:59 Coke I will put out a copy of the 0.9.1 docs in the next day or so.
23:00 rurban But the others are like yours
23:00 kj Coke: yes the previous was wrong
23:00 kj it doesn't work if you run crow.pir
23:00 NotFound Thanks for the ecmascript notes in NEWS, i forgot to put it.
23:01 kj rurban: should I include your recent fix with TRACE_PACKEFILE?
23:01 kj it touches c code...
23:02 rurban well, it is only relevant for debugging.
23:02 kj is it safe to include?
23:02 rurban but without the fix debugging info can not be compiled in
23:02 rurban yes, it is safe
23:02 kj cause I just waited an hour and a half for tests to complete; I prefer not to do that again :-)
23:03 rurban my tests ran fine right now
23:03 kj good. I'll include then
23:03 kj seems safe yes
23:03 rurban that's how I found out
23:03 rurban bug_day_2009_03_14 is okay?
23:03 kj 4 weeks from now, the saturday
23:03 kj before
23:03 kj must be same, as feb has 28 days
23:04 kj = 4 week
23:04 rurban ah right
23:04 rurban sounded weird, but feb is weird
23:05 rurban I'm testing your release now...
23:05 kj perhaps I should upload the tarball?
23:07 NotFound kj: one second please, I'm checking it builds with g++
23:07 rurban I'm testing msvc, cygwin and linux i386
23:09 NotFound Done, no problem other than a few warnings in nci_test
23:10 kj should I upload a tarball, or do you think it's all  ok?
23:10 NotFound Ok for me
23:10 kj i guess there's no more people around here :-)
23:11 rurban I think it's okay
23:11 kj what platforms are you guys on?
23:11 rurban my tests last longer, but it builds fine
23:11 NotFound linux i386
23:11 kj msvc, cygwin, and linux
23:11 kj any more?
23:11 rurban solaris, openbsd, freebsd
23:11 kj rurban: you have all of them?
23:11 rurban sure, vmware
23:12 kj ah right
23:12 rurban mingw ++
23:12 kj I'll just wait a bit till my msvc build is completed. then I'll continue
23:13 rurban you'll see the updated smolder reports then
23:23 rurban NotFound: your #343 patch  also builds fine
23:23 dalek parrot: r36845 | kjs++ | tags/RELEASE_0_9_1:
23:23 dalek parrot: tagged release 0.9.1
23:23 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36845/
23:23 NotFound Good
23:25 rurban I'll continue with 64bit then
23:28 tewk looking for a tool to grep my perl code base and tell me all the non-core modules that are used. ideas?
23:29 kj joined #parrot
23:29 NotFound Let's start breaking things ;)
23:29 kj I'm nto sure what it means to 'assign yourself co-maintainer permission
23:29 kj who knows PAUSE?
23:29 rurban do we still upload to PAUSE also?
23:29 kj well according to the release managers guide, we do ;-)
23:30 rurban https://pause.perl.org/pause/authenquery
23:30 rurban this will fail for you. do you have an account there?
23:30 kj yes, what option?
23:30 purl rumour has it option is a separate key. or http://www.investopedia.com/terms/o/option.asp
23:32 rurban change permissions
23:32 purl rurban: that doesn't look right
23:32 rg forget option
23:32 purl rg: I forgot option
23:33 kj what button then?
23:33 kj I tried the 4th; didn't work
23:33 kj the 2nd/3rd don't make sense
23:33 rurban https://pause.perl.org/pause/authenquery?pause​99_peek_perms_by=a&amp;pause99_peek_perms_quer​y=PARROTRE&amp;pause99_peek_perms_sub=Submit
23:33 shorten rurban's url is at http://xrl.us/begmjd
23:34 rurban that's a lot of work, needed?
23:34 kj that's what I got before chrome crashed :-)
23:34 kj allison: ping
23:35 rurban well, only those which are active are needed.
23:35 kj which are those?
23:35 rurban Parrot::% and friends e.g.
23:36 kj aaah. apparently university network here won't allow me to ssh :-S
23:36 allison kj: here
23:37 kj allison: hi i just wanted to ask whether the ftp-osl server was up, cause my connection keeps being closed by 'remove host'
23:37 kj *remote
23:37 kj but trying from moritz' box works fine, so it's just my connetion
23:37 allison kj: ah, ok
23:37 kj will do upload from there
23:39 rurban all my smokes passed successfully
23:40 dalek parrot: r36846 | rurban++ | trunk/tools/dev/install_files.pl:
23:40 dalek parrot: [install] Fix make install for cygwin, exe magic.
23:40 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36846/
23:41 rurban well, one solaris cfg fails with a missing pcre lib, but I had this since weeks.
23:47 NotFound Commits allowed?
23:47 kj created tarball and uploaded
23:48 kj so yes :-)
23:49 rurban but I don't see it on the pause queue
23:50 dalek parrot: r36847 | NotFound++ | trunk:
23:50 dalek parrot: [core] forward declare core types in a new header file, move Parrot_compile_string to embed.h, unskip a embed test, TT #343
23:50 dalek parrot: review: https://trac.parrot.org/parrot/changeset/36847/
23:50 kj rurban: I skipped that step for now, will look again in few minutes.
23:51 kj it's not really straightforward.. never used that thing
23:51 rurban I'm running perl tools/dev/parrot_coverage.pl now
23:54 rurban my DEVEL config is still "-devel"
23:55 rurban does it check for the .svn subdir?
23:55 kj where is that DEVEL config?
23:55 rurban ./parrot_config DEVEL
23:56 rurban or just look into config_lib.pasm
23:56 kj I don't have a clue what that means.
23:56 rurban that means that parrot thinks I'm not on a released version.
23:56 kj how should that be removed?
23:57 kj because I know only of DEVELOPING file, but accroding to the guide that shouldn't be included
23:58 rurban config/init/defaults.pm
23:58 rurban 227:        DEVEL   => ( -e 'DEVELOPING' ? '-devel' : '' ),
23:58 rurban yep.
23:59 kj mmm. so it was there apparently
23:59 rurban all okay.
23:59 kj i mean, DEVELOPING was there
23:59 kj it this a bad thing?
23:59 rurban for me bad, for you good.

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

Parrot | source cross referenced