Camelia, the Perl 6 bug

IRC log for #parrot, 2008-07-01

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:03 tetragon joined #parrot
00:04 dalek r28877 | jkeenan++ | autoicu:
00:04 dalek : [configure] Add one new test file to test case where 2nd return point is
00:04 dalek : exercised.  Add explicit number of tests to plan in each test file.
00:04 dalek : _handle_autodetect() and _handle_search_for_icu_config() become internal
00:04 dalek : methods so that we can, in testing, provide a dummy value for the default
00:04 dalek : icu-config program.
00:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28877
00:05 dalek r28878 | jkeenan++ | autoicu:
00:05 dalek : Update MANIFEST.
00:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28878
00:09 AndyA joined #parrot
00:25 bacek_ http://bacek.homeip.net/cgit/parrot/tree/lang​uages/perl6/src/classes/List.pir?h=bacek&​id=98825289fb25d8a19f31e0679b9bd9ac23ab7854
00:25 bacek_ pmichaud: this my current lazy list implementation.
00:25 japhb joined #parrot
00:53 pmichaud DietCoke: (deprecated)  I'm very glad we have it.
01:14 Andy joined #parrot
01:26 cotto_work can PMC arrays be sparse?
01:26 cotto_work rather array PMCs
01:27 * bacek_ got 500 from RT...
01:28 Whiteknight cotto_work, at the moment I dont think we have sparse arrays
01:28 Whiteknight I would like to make them, if I had the time
01:28 cotto_work of course if we did have them, a standard iterator would work fine
01:30 Whiteknight we could probably rip a lot of code off from the hash pmc, but use indices instead of string hash keys
01:34 cotto_work would there be any reason to avoid accessing any of the sparse elements when iterating over the array
01:34 cotto_work also, would the sparse elements be counted in VTABLE_elements()?
01:35 cotto_work or is that implementation-dependent?
01:36 Whiteknight i think it depends how you implement it
01:37 cotto_work I'll wait until my code breaks, then.
01:37 acalhoon joined #parrot
01:38 acalhoon hey everyone
01:38 purl hmmm... everyone is stupid
01:38 acalhoon I'm back
01:39 acalhoon i've got a new and exciting issue
01:40 Whiteknight pray do tell
01:40 acalhoon ha
01:40 Whiteknight hello acalhoon, what's up?
01:41 acalhoon i'm getting some build issues related to a series of changes to remove stacks_common
01:41 magnachef_ joined #parrot
01:41 acalhoon if I back out change sets 28863:28866 I can build again
01:41 Whiteknight oh don't say that, that was my project!
01:41 acalhoon I know
01:41 acalhoon :)
01:41 Whiteknight you're on openbsd right?
01:41 acalhoon yessir
01:42 Whiteknight what's the issue? I didn't change much code, so it surprises me that there's a problem
01:42 acalhoon gimme a sec
01:42 acalhoon what's the paste site again
01:42 acalhoon ?
01:43 cotto_work nopaste?
01:43 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/
01:43 purl rumour has it nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating or App::Nopaste or tools/dev/nopaste.pl
01:53 acalhoon hmm
01:54 acalhoon now that I built without the changes I can't get it to happen again.
01:54 bacek_ pmichaud: index2.diff from RT#56214 is reworked patch.
01:54 stupidbot RT 56214: [PATCH] Implementation of Str.index. - open
01:56 acalhoon making from realclean
01:58 Theory joined #parrot
02:01 Tene seen jhorwitz?
02:01 purl jhorwitz was last seen on #parrot 5 days and 12 hours ago, saying: particle: short but sweet.  i loaded up on beers while i could.  there were about 8 or 9 people there, including waltman, jcap, geoff avery, and Tene.   [Jun 25 06:54:01 2008]
02:01 acalhoon Whiteknight you're in the clear.
02:01 Whiteknight yay!!
02:01 acalhoon I'm not sure what I did, but it wasn't your change...apparently
02:01 acalhoon I do have a problem with 'make world'
02:02 acalhoon wait wait wait...
02:03 acalhoon during make world I get:
02:03 acalhoon src/packdump.c
02:03 purl somebody said src/packdump.c was broken. Isn't it? src/packdump.c: In function 'PackFile_Constant_dump':
02:03 acalhoon src/packdump.c:40: error: nonnull argument references non-pointer operand (arg 1, operand 2)
02:03 acalhoon *** Error code 1
02:03 acalhoon oh
02:03 acalhoon ha
02:03 acalhoon well there you have it
02:03 Whiteknight I've never even heard of make world
02:03 acalhoon muhaha
02:03 * Whiteknight has a lot to learn still
02:04 acalhoon 'all' and 'parrot_utils'.
02:04 acalhoon ha, I'm WAY behind
02:04 acalhoon I've just begun
02:04 acalhoon :)
02:04 Whiteknight well, I wish you luck with it then
02:05 acalhoon any suggestions?
02:05 * purl farts
02:05 dalek r28879 | jkeenan++ | autoicu:
02:05 dalek : [configure] _handle_icushared() and _handle_icuheaders() become methods so
02:05 dalek : that we can reference more hard-coded values which have been moved inside
02:05 dalek : _init().
02:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28879
02:05 Whiteknight purl is the absolute worst bot ever
02:06 acalhoon as far as being part of the "solution"
02:06 kid51 IIRC 'make world' is something the release manager (and whoever is helping) runs just before release.
02:06 acalhoon purl can be pretty annoying
02:06 purl acalhoon: what?
02:07 acalhoon and apparently it's been broken for a while
02:07 kid51 acalhoon:  'make world' consists of several variations upon 'make'.  Do you know in which one you got that error?
02:08 bacek_ stupidbot: insult purl
02:08 stupidbot bacek_: Error calling said() for insult: Can't locate object method "new" via package "Acme::Scurvy::Whoreson::BilgeRat" (perhaps you forgot to load "Acme::Scurvy::Whoreson::BilgeRat"?) at /usr/local/share/perl/5.10.0/Bot/B​asicBot/Pluggable/Module/Insult.pm line 42.
02:08 bacek_ yak...
02:08 kid51 purl escapes ... for the moment!
02:08 purl kid51: excuse me?
02:09 kid51 u heard me, u lamebrain
02:09 acalhoon make ./pdump
02:09 acalhoon that's where it dies for me
02:09 kid51 Now *that's* a 'make' target *I've* never heard of.
02:09 acalhoon haha
02:10 acalhoon it's the first target for make world
02:10 acalhoon I didn't realize i was in such uncharted waters
02:12 kid51 pdump is insider 'make parrot_utils', which is inside 'make world'
02:12 kid51 acalhoon:  Are you having any problems with, say, a simple 'make'?
02:13 kid51 If not, then I would recommend that you file a bug ticket about this (parrotbug@parrotcode.org) and lose no sleep over it.
02:14 acalhoon no
02:14 acalhoon simple make is fine
02:15 Whiteknight i think you're making all this up. What does make world do anyway?
02:15 Whiteknight and why would anybody want to do it?
02:16 kid51 Whiteknight:  As I said, it's the "everything and the kitchen sink" target used by release managers.
02:16 kid51 Ordinary users would not use it.
02:16 Whiteknight if i'm going to be a release manager, I'm going to have to learn that one day I guess
02:16 kid51 Yup
02:16 acalhoon it's really "make parrot_utils"
02:16 spinclad release manager would run 'make release' before releasing
02:16 * Whiteknight starts taking notes
02:17 Whiteknight acalhoon, the only tests that fail for me now are the ones caused by a half-patch I'm working on
02:17 acalhoon no no no, you're fine Whiteknight
02:18 spinclad as an ordinary developer i've generally run 'make world'
02:18 acalhoon I don't think this is your problem
02:18 Whiteknight okay, that makes me worry less
02:19 acalhoon sorry sorry
02:19 kid51 as an even more ordinary developer i've generally just run 'make' :-)
02:19 acalhoon I'm in the "honey moon" phase of the parrot development things, a recent hobby. So I had a problem with an obscure feature and "freaked" out :)
02:20 Whiteknight i dont even run "make test" every day like I should
02:20 * acalhoon gasps! ;)
02:20 kid51 Whiteknight:  No one does.  That's evident from some of the breakage ;-)
02:21 Whiteknight I usually try to test before I submit a patch or a trunk commit
02:21 kid51 But I've written some scripts which automate much of my local 'make' and 'make test'
02:21 Whiteknight I don't even re-read my work before committing to branch though :)
02:21 kid51 That's cool.
02:22 acalhoon the only reason I thought to run "make world" was that "make help" displayed it under the Executable heading.
02:22 Whiteknight "make help"? I've learned two things today!
02:22 acalhoon haha
02:22 acalhoon make help is your friend
02:25 acalhoon so if I wanted to get more serious about participating, is parrotcode.org/docs the best place to start?
02:26 Whiteknight that's where I started, but there are a lot of omissions
02:26 acalhoon ha, like any good software project :)
02:28 acalhoon fair enough, that's where I'll start
02:45 tetragon joined #parrot
02:47 Infinoid acalhoon: (backlogging) so you're building successfully again?
02:50 acalhoon make  works fine
02:51 acalhoon thnx
02:52 Infinoid Whiteknight's change required a "make realclean", I'm betting that's what your earlier issue was about
02:52 Infinoid looking at the packfile.c thing now
02:53 acalhoon I bet you're right
02:54 acalhoon rm line 40 of packdump and it works
02:54 acalhoon whether or not that's the right fix....I have no idea :)
02:55 Whiteknight I should have put a note about realclean in the svn log
02:56 Infinoid I think there's something in the docs that says you should do a realclean after *every* update... and if not, maybe there should be
02:56 Whiteknight I think you would only need a realclean if the makefile was changed, or if you change something weird with a dependency that isn't in the makefile correctly
02:58 Ademan joined #parrot
02:59 Infinoid or if you change the vtable numbering
02:59 Infinoid or the opcode numbering
03:00 Limbic_Region joined #parrot
03:00 Infinoid after having plenty of problems of that nature, I realclean religiously
03:00 acalhoon hah
03:00 * acalhoon commits that to memory :)
03:02 * kid51 must sleep
03:02 purl $kid51->sleep(8 * 3600);
03:03 acalhoon I second that. sleep it is
03:03 acalhoon g'night
03:03 acalhoon left #parrot
03:03 Tene I'm having problems with mod_parrot.
03:04 Infinoid src/thread.c:1292: warning: suggest braces around empty body in 'do' statement
03:04 Infinoid that's caused by my system headers, and it seems I can't do anything about it
03:07 Infinoid Andy: I have a question about headerizer stuff, if you're around.
03:07 Andy ask me
03:07 Andy but I'm in the middle of Work work, too
03:08 Infinoid ok.  compilers/imcc/pbc.c has a static function within an #ifdef HAS_JIT, and a prototype above which is also inside an #ifdef HAS_JIT
03:09 Infinoid a second prototype was apparently generated by a headerizer run a couple days ago, which is not inside the #ifdef HAS_JIT
03:09 Infinoid thus, on non-JIT architectures like amd64, a warning results:
03:09 Infinoid compilers/imcc/pbc.c:188: warning: 'old_blocks' declared 'static' but never defined
03:09 Infinoid is there some special markup I can use to make headerizer smart about this?
03:38 apeiron joined #parrot
03:53 tetragon joined #parrot
03:57 * Tene sends email to jhorwitz and gives up on mod_parrot for the night.
04:04 apeiron joined #parrot
04:05 magnachef joined #parrot
04:14 tetragon joined #parrot
04:25 magnachef_ joined #parrot
04:39 Ademan joined #parrot
04:40 tetragon joined #parrot
04:43 dalek r28880 | chromatic++ | trunk:
04:43 dalek : [pdb] Cleaned up the debugger, making sure that it's always initialized when
04:43 dalek : needed and moving IMCC-specific debugging functions into IMCC (NotFound, RT
04:43 dalek : #56346).
04:43 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28880
04:58 dalek r28881 | chromatic++ | trunk:
04:58 dalek : [t] Added 'make smolder' target, as well as Smolder support code to Parrot's
04:58 dalek : test harness.  Note that you need to install a recent version of TAP::Harness
04:58 dalek : to use this code.  We need to add that dependency to Bundle::Parrot.
04:58 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28881
05:05 contingencyplan joined #parrot
05:14 bacek_ joined #parrot
05:14 dalek r28882 | chromatic++ | trunk:
05:14 dalek : [src] Fixed segfault when using the get_class opcode with a NULL STRING
05:14 dalek : (NotFound, RT ##55640).
05:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28882
05:33 Ademan joined #parrot
05:47 contingencyplan joined #parrot
05:51 dalek r28883 | allison++ | pdd25cx:
05:51 dalek : [pdd25cx] A more general approach to ensuring a particular exception handler is
05:51 dalek : never invoked more than once.
05:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28883
05:52 dalek r28884 | chromatic++ | trunk:
05:52 dalek : [PMC] Fixed a segfault when accessing the class of a NameSpace without an
05:52 dalek : associated class; now all NameSpaces have PMCNULL in their class slots when
05:52 dalek : created (reported by Vasily Chekalkin and fixed by NotFound in RT #55620).
05:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28884
05:52 stupidbot Error calling said() for rt: Malformed RT response received from http://rt.perl.org/rt3/
05:57 Psyche^ joined #parrot
06:03 Ademan joined #parrot
06:09 dalek r28885 | chromatic++ | trunk:
06:09 dalek : [config] Rearranged Darwin hints, so that -bundle doesn't come first (Paco, RT
06:09 dalek : #54986).
06:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28885
06:13 zostay joined #parrot
06:21 bacek_ joined #parrot
06:37 Ademan joined #parrot
06:44 dalek r28886 | allison++ | pdd25cx:
06:44 dalek : [pdd25cx] Rename 'pbj' and 'pbr' opcodes to 'local_branch' and 'local_return'.
06:44 dalek : Remove unneeded 'branchcc' opcodes.
06:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28886
06:57 Ademan joined #parrot
07:16 masak joined #parrot
07:45 dalek r28887 | tene++ | trunk:
07:45 dalek : [lolcode]
07:45 dalek : * First draft of adding proper block handling
07:45 dalek : * First draft of support for globals
07:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28887
07:49 dalek r28888 | tene++ | trunk:
07:49 dalek : MANIFEST and svn keywords.
07:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28888
07:50 Tene I really need to get a hold of jhorwitz for help getting mod_parrot working to see if those updates can help get lolcode working better for oscon.
07:59 Eevee joined #parrot
08:16 dalek r28889 | japhb++ | trunk:
08:16 dalek : [PROFILE] Fix op profiling calibration
08:16 dalek : * Fix divide by zero in op profiling calibration which made noop appear
08:16 dalek :   to take infinite time, thereby making all other ops instantaneous
08:16 dalek : * Display calibrated time for noop when printing op profile results
08:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28889
09:04 cognominal joined #parrot
09:07 cognominal joined #parrot
09:42 barney joined #parrot
10:28 Ademan joined #parrot
10:37 kid51 joined #parrot
11:23 bacek_ joined #parrot
11:50 kid51 Anyone else having a problem with 'make test' subsequent to r28881?
11:50 moritz Can't locate object method "new" via package "TAP::Harness" at t/harness line 91.
11:50 moritz make: *** [test] Error 255
12:04 barney joined #parrot
12:05 kid51 moritz:  Ah, so it's not just me!  I've filed in RT 55954.
12:05 stupidbot Error calling said() for rt: Malformed RT response received from http://rt.perl.org/rt3/
12:09 kid51 moritz:  I think this needs Test::Harness 3.0+ to work.  IIRC, that's part of TAP::Parser.
12:09 moritz kid51: then it's a missing dependency
12:10 * moritz tries again with perl5.10.0 which has Test::Harness 3.something installed
12:10 kid51 I recently had to reinstall the OS on my Mac.  In the process, all my /usr/local/ libraries were wiped out and I haven't restored all of them.  So I suspect this occurs because I'm reduced to the Test::Harness 2.64 that came with the vendor Perl.
12:10 masak pmichaud: how goes the work on refactoring $_, $/ et al?
12:11 moritz kid51: yes, with TH 3.0 it works fine
12:11 dalek r28890 | bernhard++ | trunk:
12:11 dalek : [Plumhead]
12:11 dalek : svn merge -r 28822:28889 https://svn.perl.org/parrot/bran​ches/libs4php/languages/plumhead
12:11 kid51 moritz:  I don't think 5.10 came with Test::Harness 3.  Test::Harness 3 was a bit too late developing to make the cut IIRC.
12:11 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28890
12:12 moritz kid51: yes, but i've installed it for my perl5.10
12:12 kid51 Well, then, the project will have to decide whether we want to make T:::H 3 a dependency now.
12:12 kid51 moritz:  Yes, I have it the same way on my Linux box.
12:13 barney (T:::H 3 )++
12:14 moritz indeed
12:15 kid51 I do like T::H 3.   I even had a patch added in 3.11 to facilitate Parrot testing!  But we're going to have massive breakage unless we include it as a dependency and declare T::H 2 dead.
12:15 * kid51 must go to $job.
12:36 dalek r28891 | bernhard++ | trunk:
12:36 dalek : [Plumhead]
12:36 dalek : Add t/php/functions.t with a single TODO test.
12:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28891
12:54 Andy joined #parrot
12:58 dalek r28892 | bernhard++ | trunk:
12:58 dalek : [Plumhead PCT]
12:59 dalek : Add support for function definitions without args.
12:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28892
13:07 Psyche^ joined #parrot
13:15 Whiteknight joined #parrot
13:15 dalek r28893 | Whiteknight++ | gsoc_pdd09:
13:15 dalek : [gsoc_pdd09] A few pointer arithmetic fixes, and improved pointer protection for headerizer
13:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28893
13:18 dalek r28894 | Whiteknight++ | gsoc_pdd09:
13:18 dalek : [gsoc_pdd09] update to trunk r28892
13:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28894
13:25 dalek r28895 | pmichaud++ | trunk:
13:25 dalek : [rakudo]:
13:25 dalek : * spectest-progress update:  75 files, 1172 passing tests
13:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28895
13:38 Khisanth joined #parrot
13:49 dalek r28896 | Whiteknight++ | gsoc_pdd09:
13:49 dalek : [gsoc_pdd09] A few fixes:
13:49 dalek : * Better bounds checking on the allocation loop
13:49 dalek : * Improved pointer arithmetic
13:49 dalek : * Improved GC_IT_CARD_ALL_* macros
13:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28896
14:00 dalek r28897 | coke++ | trunk:
14:00 dalek : The quiet option we're using in the config file requires a newer version
14:00 dalek : of P::C to use. Require a newer version or skip gracefully.
14:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28897
14:01 gryphon joined #parrot
14:07 DietCoke (you should not have to do a realclean after every update. don't put that in the docs.)
14:08 jan joined #parrot
14:09 DietCoke regarding the T:H 3 breakage, I believe we already had a plan on that, but I was expecting more bits of the plan to be working before the patch that broke make test got applied.
14:13 particle and make world is not for release managers only
14:13 DietCoke ... they're the only ones who ever use it, though. =-)
14:13 particle it's for anybody who wants the extra tools included in the make world target
14:14 DietCoke yes, but since we don't build or test those tools by default...
14:14 DietCoke "anything not built by default will be broken by the time you try it." =-)
14:18 * DietCoke wonders who added him to the RT commit list.
14:18 Andy joined #parrot
14:22 Whiteknight who added who to the commit list?
14:26 DietCoke added whom. me.
14:26 DietCoke and I don't know. that's why I asked.
14:27 DietCoke note I said the -rt- commit list. like, for the software that runs our ticketing system.
14:34 iblechbot joined #parrot
14:44 dalek r28898 | Whiteknight++ | gsoc_pdd09:
14:44 dalek : [gsoc_pdd09] Fixed problem with cards overlapping arena headers. Added a number of new assertions and diagnostics messages, some of which might prove to be unnecessary
14:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28898
14:45 Whiteknight maybe your reputation has proceeded you. maybe you're so freaking awesome that everybody wants to add you to their projects
14:47 NotFound Everybody wants to have a Coke at hand.
14:48 DietCoke ^_o
14:48 DietCoke as long as that's not "in hand", I suppos I'm ok.
14:48 * DietCoke self-referentially drinks a Coke Zero
14:49 Infinoid at least action figures are readily available
14:49 Whiteknight Yeah, I have the whole set of DietCoke action figures
14:49 Whiteknight with karate chop action
14:51 Whiteknight cuts through old RT  tickets like a tornado
14:52 DietCoke <sam the eagle>You are all weirdos!</sam>
14:57 Theory joined #parrot
15:07 Khisanth joined #parrot
15:20 barney joined #parrot
15:31 sjansen joined #parrot
15:35 dalek r28899 | bernhard++ | trunk:
15:35 dalek : [Plumhead]
15:35 dalek : Add a TODO test for OO PHP.
15:35 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28899
15:38 Auzon joined #parrot
15:39 cotto_home DietCoke, if I can remove some #includes from a couple PMCs and still keep make test happy, is it fine to commit that change?
15:42 DietCoke very likely. which includes? =-)
15:42 DietCoke (these are individual includes from a .pmc versus those automagically added in by the pmc2c compiler?)
15:44 cotto_home they're some includes that would be included anyway by pmc2c
15:44 DietCoke oh, yah, kill 'em.
15:45 DietCoke and in general, cleanup that doesn't break 'make test' is safe.
15:45 DietCoke ... and you're not at work today? =-)
15:45 cotto_home good to know
15:45 particle ...unless it's in an alternate runcore
15:45 particle then you need fulltest
15:46 DietCoke Does anyone use alternate runcores aside from release managers? =-)\
15:46 DietCoke (and those trying to track down a GC issue.) Hey, why doesn't make fulltest test the gcdebug runcore? =-)
15:47 dalek r28900 | cotto++ | trunk:
15:47 dalek : [pmc] Remove some #includes that are taken care of by pmc2c.  All tests still
15:47 particle DietCoke: you know you have commit rights to bundle-parrot
15:47 dalek : pass.
15:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28900
15:47 DietCoke yes. I also don't know how to add something cleanly.
15:47 NotFound DietCoke: maybe because of speed.
15:47 particle 4525/18864
15:47 purl 0.239874893977947
15:47 DietCoke NotFound: it was a joke.
15:47 DietCoke particle: do I update the version? Can I specify a minimum required version? How do I cut a release?
15:48 particle DietCoke: i've already updated svn head
15:48 particle just need to make cpan release
15:48 particle which is, .tar.gz and upload to pause
15:48 DietCoke I will check the recent commit log.
15:48 particle i have a mostly crafted draft email response
15:53 dalek r28901 | bernhard++ | trunk:
15:53 dalek : [Plumhead]
15:53 dalek : Add a TODO test for namespaces.
15:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28901
15:53 NotFound In #56508, if I change "abc" for 'abc' in the lexid or in the string, the problem disappears, changing both they reappears as abc'
15:57 particle i suspect a lexer bug
15:58 particle or maybe it's a string vs cstring bug
15:58 jonathan I suspect it's probably because the two are sharing the string constant, and the one constant gets modified (for lexid)
15:58 NotFound I think I got it.
15:59 jonathan See pbc.c, search for lexid
15:59 NotFound unit->lexid->name[strlen(unit->lexid->name) - 1] = 0;
15:59 NotFound unit->lexid->color = add_const_str(interp, unit->lexid);
15:59 NotFound These two lines were in reverse order.
16:00 jonathan Hmm. I wonder if that was in c's patch to fix the double free issue in my initial one.
16:01 NotFound Going to rebuild and make a full test with this change.
16:09 dalek r28902 | moritz++ | trunk:
16:09 dalek : [rakudo] one more file for spectest_regression
16:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28902
16:12 Whiteknight what's the difference between miniparrot and parrot?
16:13 Auzon joined #parrot
16:13 NotFound (bad joke omitted)
16:14 particle miniparrot has no embedded config info
16:14 Whiteknight ...so what does that mean?
16:15 particle miniparrot is something we don't need anymore
16:15 particle it should go away
16:15 Whiteknight oh, because miniparrot is the only thing working in my branch. It (mostly) runs, but regular parrot segfaults terribly
16:15 particle heh
16:16 particle ok, well glad we have it, then :)
16:16 particle take a look at Makefile to see the difference between parrot and miniparrot
16:16 NotFound It was supposed to be an auxiliar tool to build the rest of parrot using it, it isn't?
16:18 moritz NotFound: probably very much like miniperl
16:20 particle yes
16:20 dalek r28903 | moritz++ | trunk:
16:20 dalek : [rakudo] tests for ternary now in spectest_regression
16:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28903
16:20 particle but we've decided we're going to depend on perl for parrot 1.0
16:20 particle so no need for miniparrot
16:21 moritz no own dogfood
16:21 moritz at least not in this
16:21 NotFound And there is a problem to build it using only standard c file io, wich accoriding some docs was the intention.
16:21 particle the problems can be overcome, but we have a schedule
16:22 particle we want parrot 1.0 out the door
16:22 particle it'll take too long to get a build system built around miniparrot
16:23 moritz aye
16:24 particle i'd love it if somebody proved that wrong
16:24 NotFound Patch sended to RT#56508
16:24 stupidbot Error calling said() for rt: Malformed RT response received from http://rt.perl.org/rt3/
16:24 particle that bot is too stupid
16:25 dalek r28904 | Whiteknight++ | gsoc_pdd09:
16:25 dalek : [gsoc_pdd09] parrot, libparrot, and miniparrot all build, runtime/parrot/include/config.fpmc builds. parrot segfaults, miniparrot (mostly) runs.
16:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28904
16:33 tetragon joined #parrot
16:41 pmichaud #parrotsketch in 109
16:42 NotFound Seconds?
16:42 purl Seconds is second
16:43 Whiteknight pmichaud, I'm planning on it, yes
16:43 Whiteknight oh wait, you weren't asking the question NotFound was
16:46 NotFound Forget it.
16:46 purl NotFound: I forgot it
16:50 cotto_work now we'll never know what it was
16:52 NotFound Ask Stephen King about that.
16:57 dalek r28905 | Whiteknight++ | gsoc_pdd09:
16:57 dalek : [gsoc_pdd09] build process nearly completes, segfaults when building Perl6Grammar.pir
16:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28905
17:01 japhb joined #parrot
17:03 dalek r28906 | moritz++ | trunk:
17:03 dalek : [rakudo] two more files for spectest_regression
17:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28906
17:07 buff3r joined #parrot
17:22 barney joined #parrot
17:31 dalek r28907 | bernhard++ | trunk:
17:31 dalek : [Plumhead]
17:31 dalek : Add test script for references.
17:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28907
17:57 dalek r28908 | infinoid++ | pdd13pbc:
17:57 dalek : [pdd13] Merge changes from trunk (-r28100:28907) into pdd13pbc branch.
17:57 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28908
18:00 moritz can I resolve a ticket in RT without writing a new mail?
18:00 pmichaud if you have admin privileges, yes
18:01 pmichaud (put another way -- writing a reply isn't required -- just leave the reply box blank.)
18:01 moritz ok, closed RT#55164
18:01 stupidbot Error calling said() for rt: Malformed RT response received from http://rt.perl.org/rt3/
18:02 pmichaud stupidbot got more stupid.
18:02 szbalint how fitting.
18:02 particle opbots, names
18:02 moritz bacek: there's no need for stupidbot to report rt errors, we know it's broken sometimes ;)
18:02 was kicked by particle: stupid bot got more stupid
18:06 tco joined #parrot
18:11 dalek r28909 | fperrad++ | libs4php:
18:11 dalek : [php] PCRE (very first step)
18:11 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28909
18:14 barney fperrad++
18:16 allison joined #parrot
18:16 peepsalot joined #parrot
18:18 chromatic joined #parrot
18:20 cotto_work barney, it looks like the delay in transferring pipp.org is because of the 5-day delay for a paypal payment to clear
18:20 cotto_work it should be complete later this week
18:24 * DietCoke tries to quickly write a perl script to say "here are all of coke's commits to parrot in the last week." and fails in time for parrotsketch.
18:24 dalek r28910 | allison++ | pdd25cx:
18:24 dalek : [pdd25cx] Add range checking to 'local_return' opcode, to require it be within
18:24 dalek : the current code segment.
18:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28910
18:25 moritz DietCoke: should I expand http://moritz.faui2k3.org/svn-search/ to include the parrot repo?
18:25 moritz DietCoke: I won't make it 'till #ps, but it would help you in future ;)
18:25 barney cotto: k
18:25 pmichaud I just use  "svn log -r '{2008-06-24}:head' | less"  and then search for 'pmichaud'
18:26 barney I look in http://cia.vc/
18:26 NotFound In what perl? 6?
18:26 Whiteknight joined #parrot
18:28 moritz http://cia.vc/stats/project/parrot perhaps?
18:28 pmichaud #parrotsketch in 2
18:29 DietCoke pmichaud: that's basically what I ended up doing. i was just trying to save myself the "search" bit. =-)
18:29 DietCoke First week I've committed this much in a while. =-)
18:29 jhorwitz joined #parrot
18:30 Tene I just run 'git log' and search for 'tene' and keep going until I pass the date of the previous #ps
18:31 wknight8111 joined #parrot
18:31 jhorwitz Tene: re mod_parrot, did you export PARROT_RUNTIME in envvars?
18:32 tetragon joined #parrot
18:32 Tene I added a SetEnv line to httpd.conf
18:33 Tene I also tried exporting it in the shell I launched httpd from.
18:33 jhorwitz SetEnv won't work
18:33 pmichaud I find the search bit is worthwhile because I also find where I was mentioned in others' commits :-)
18:33 slightlyoff joined #parrot
18:33 Tene Okay, I can't find any reference to an 'envvars' config entry.
18:33 jhorwitz exporting it in the shell should work though...
18:33 jhorwitz can you send me the strace from the latest?
18:34 slightlyoff left #parrot
18:34 dalek r28911 | chromatic++ | trunk:
18:34 dalek : [IMCC] Prevented :lexid from interfering with identical string constants (RT
18:34 dalek : ##56508, test case by Patrick Michaud, patch by NotFound).
18:34 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28911
18:34 jhorwitz Tene: you can also just paste the interpinfo.pasm stat calls.  i want to see the paths
18:35 Tene ... and now it works.
18:35 Tene I guess I must have had a typo in the path the last time I exported it.
18:35 * jhorwitz rejoices
18:39 pmichaud jonathan: just fyi -- I'll be unavailable July 16th, and will have only sporadic access July 19th-22nd.  (Jul 23-27 is oscon, of course)
18:41 jonathan pmichaud: Thanks good to know...I'm not doing OSCON this year, which I guess you already knew.
18:41 Tene jhorwitz: you sticking around after #ps for a bit?
18:41 jhorwitz yes
18:42 Tene fantastic.
18:42 pmichaud reminds me... I need to book YAPC::EU tickets
18:42 moritz jonathan: as always, when you tell me your topics for rakudo day in advance, I might be able to contribute a few tests beforehands
18:42 pmichaud I'm thinking of arriving on Aug 11 and departing Aug 17 -- does that sound about right?
18:42 pmichaud that gives me a day before and a day after YAPC::EU for hacking
18:43 * Tene is distracted by $awesomejob_yeswearehiring for a bit.
18:45 pmichaud I always think of "RT#" as a sigil, much like "r" is a sigil for SVN revision numbers.  :-)
18:46 pmichaud pipp++
18:46 moritz pmichaud: shameless cross-poster ;)
18:46 pmichaud I'm saving my back-channel comments for #parrot
18:47 jonathan moritz: Thanks, good to know. MMD is upcoming this month, as are signature objects as I kinda need those to do MMD.
18:47 moritz jonathan: not exactly my area of expertise, but I'll see what I can do
18:47 jonathan moritz: Also, enums.
18:48 chromatic You don't find the digraph T# horribly ugly?
18:48 jonathan I just about worked out how to do those the other day in the showever, but now I've forgotten again. :-|
18:48 pmichaud it's not ugly in my font.
18:49 pmichaud having enums might be very useful; we could see about replacing Bool::* and Order::* with them
18:51 cotto_work barney, do you know why fperrad is working in a branch?
18:52 Ron joined #parrot
18:52 barney because some tests are working only with the PHC variant
18:52 jonathan pmichaud: Dates - look OK. I don't know what if anything is happening with the hackathon that was proposed in the south of Sweden.
18:53 barney This way there are no failures in the default PCT variant
18:53 jonathan I will certainly be around somewhat on the day before, but I may have a lunchtime meeting that day.
18:54 pmichaud given that I'm traveling jul 19-27, jul 30-aug 7, I think I'll be glad to keep it a bit shorter trip
18:56 jonathan OK
18:57 jonathan I'm taking a week's break at the end of August too, so I wouldn't mind getting back after the conference to get a week or so's working in.
18:58 jonathan If there's not going to be much Perl 6 hacking going on at the Lund hackathon, I'll probably skip it...
18:58 jonathan Anyway, day after YAPC is fine with me too.
18:58 jonathan I'll arrange to be around then.
19:11 davidfetter joined #parrot
19:14 wknight8111 allison still here?
19:15 allison yup
19:15 wknight8111 I have a debugging question that chromatic said I should ask you, in hopes there would be a quick answer
19:15 allison ok
19:16 wknight8111 in parrot initialization, in src/pmc_freeze:run_thaw, it's calling a DOD run
19:16 wknight8111 but in that initial run, a lot of important values are null, like interp->dynamic_env
19:16 wknight8111 or interp->exception_list
19:17 wknight8111 I'm wondering what would have caused this, is this a big problem, etc?
19:18 allison it means the interpreter wasn't fully initialized
19:18 allison is this a test file with a dummy interpreter?
19:18 wknight8111 no, it's the parrot executable. I traced it during the build process
19:19 wknight8111 I was getting segfaults because I was trying to call pobject_lives on those null values
19:19 chromatic Are we perhaps having a mark phase run earlier in the process than we expected?
19:19 barney So T::H 3 is now a real requirement ?
19:19 allison the simple answer is to have the interpreter's gc check if those exist before it marks them
19:20 particle barney: for smoking, yes
19:20 allison interp->exception_list is going away anyway
19:20 NotFound I think we have a potential problem with the duality NULL/PMC_NULL
19:20 chromatic I posted a patch that should relax the T::H 3 requirement for simple 'make test'.
19:20 DietCoke barney: apparently we can avoid it for make test, I think the harness was just not smart enough about the require.
19:20 DietCoke I don't have T:H2 and therefore cannot easily test it. I'm hoping jim will be able to report back shortly after work.
19:20 wknight8111 allison, I did add conditionals to avoid those if they are null, I was assuming they weren't supposed to be null in any case
19:21 wknight8111 but if it's ok, I'll ignore it (improves my test results anyway)
19:21 barney When I have some tuits, I'll look into languages tests aggregation with T::H 3
19:22 moritz barney: would that imply the ability to test in parallel?
19:22 moritz I'd very much welcome that!
19:22 allison at certain stages of initialization, it's fine for those to be null
19:23 barney Maybe, first of all I want to get rid of that suboptimal 'unified languages testing'
19:23 wknight8111 allison, thanks! You've saved me lots of time
19:24 particle moritz: yes, and note there's a branch for this work already, however i'm sure there's bitrot
19:24 particle http://svn.perl.org/parrot​/branches/unified_testing/
19:24 DietCoke As i said, i would not really mind making t:h 3 a requirement for 'make test'.
19:25 DietCoke (if we have to, which it sounds like we can avoid for now.)
19:25 particle i think we should try it, and roll back upon failure
19:26 particle 1) parallel testing, 2) test report aggregation, 3) smoke testing with smolder
19:26 particle dc?
19:26 purl i heard dc was at language.perl.com/ppt/src/dc or Damian Conway or dc(1) or the district of columbia :p or Daimler Chrysler or Data Center
19:27 particle dc is also DietCoke
19:27 purl okay, particle.
19:29 * Tene prods jhorwitz
19:29 * jhorwitz jumps
19:29 * particle heats up the branding iron
19:30 Tene jhorwitz: Can you give me an example of what you'd like to run in mod_lolcode?
19:31 Tene jhorwitz: Can you show me the apache config I need to use to run lolcode with mod_parrot?
19:32 jhorwitz Tene: don't need much.  basically i need treat undeclared variables as undefined instead of bombing out with an unknown lexical error.
19:33 jhorwitz nopaste?
19:33 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/
19:33 purl nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating or App::Nopaste or tools/dev/nopaste.pl
19:34 nopaste "jhorwitz" at 96.245.16.45 pasted "Tene: mod_lolcode config" (4 lines) at http://nopaste.snit.ch/13440
19:39 Tene jhorwitz: show me an example of a lolcode file you'd like to run?
19:43 jhorwitz Tene: mod_lolcode maps all query params to globals.  a simple "VISIBLE FOO" would be an example.  you can see it crash if you don't supply FOO as a parameter
19:44 Tene ahh
19:44 jhorwitz even *testing* for truthiness crashes it
19:44 particle VISIBLE TRUTHINESS
19:44 jhorwitz so "FOO/O RLY?" still bombs out
19:45 * jhorwitz flips particle's smartass switch
19:45 Tene That should be a simple fix for now.
19:45 * jhorwitz watches out for particle's branding iron
19:53 Whiteknight joined #parrot
19:59 dalek r28912 | tene++ | trunk:
19:59 dalek : [lolcode]
19:59 nopaste "tene" at 67.137.148.11 pasted "lolcode demo for jhorwitz" (11 lines) at http://nopaste.snit.ch/13441
19:59 dalek : Initialize tokens that can't be found to 0.
19:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28912
20:00 Tene jhorwitz: that work for you?
20:02 jhorwitz hm, can we initialize to an Undef instead?
20:02 jhorwitz that way it exists and can be tested against
20:03 jhorwitz but has no value
20:03 Tene Sure.
20:03 DietCoke chromatic: looks like your patched worked, can we get it applied?
20:04 Whiteknight DietCoke, how does one delete a branch?
20:05 chromatic DietCoke, let me fix one bug in it.
20:05 Whiteknight please_kill_worthless_branch("/gmc")
20:06 nopaste "tene" at 67.137.148.11 pasted "another lolcode demo for jhorwitz" (18 lines) at http://nopaste.snit.ch/13442
20:07 chromatic DietCoke, r28913.
20:07 jhorwitz Tene: FARAWAY, LOL.  Alias and i were considering using INVISIBLE for something like this.  :)
20:07 DietCoke Whiteknight: you svn dev "repo address to the branch"
20:07 DietCoke er, 'svn del'
20:07 Tene jhorwitz: I looked all over the forums, but couldn't find any proposed syntax for globals, so just made up my own.
20:07 dalek r28913 | chromatic++ | trunk:
20:07 dalek : [t] Re-allowed Test::Harness 2.x for simple make test, after r28881
20:07 dalek : inadvertently required TAP::Harness.
20:07 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28913
20:08 DietCoke so something like svn del https://svn.perl.org/parrot/branches/gmc
20:08 DietCoke WARNING! DO NOT TYPO THIS COMMAND LINE.
20:08 Whiteknight okay. I've looked all through it and found nothign to salvage. I am going to delete it now then
20:08 DietCoke you could cause some serious damage. someone inadvertantly deleted all the tags at one point, e.g.
20:09 DietCoke if you're paranoid, you could do an svn co of branches/ and then svn del gmc from inside that one. =-)
20:09 dalek r28914 | tene++ | trunk:
20:09 dalek : [lolcode]
20:09 dalek : * Undef, not zero.
20:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28914
20:09 DietCoke always be extra careful working directly agaginst the repository.
20:09 Whiteknight I'll do it your way, I'll just be very careful
20:09 DietCoke ... ok, or I'll sic chromatic on you. =-)
20:09 Whiteknight !!!
20:10 DietCoke DietCoke: !: event not found
20:13 Whiteknight done
20:14 dalek r28915 | Whiteknight++ | gmc:
20:14 dalek : [GMC] Killing old, obsolete branches/gmc.
20:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28915
20:14 confound you can always just revert your change if you delete the wrong thing
20:14 confound though it's probably slow and clunky
20:14 * Whiteknight looks, doesnt appear to be any bad side effects...
20:15 NotFound ...famous last words.
20:15 particle1 joined #parrot
20:15 dalek r28916 | fperrad++ | libs4php:
20:15 dalek : [php] declare PCRE constants
20:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28916
20:17 japhb For the profiling runcore, what is the intended profile?  The "empty loop" calibrate() routine is currently nonsensical, but given the POD right above it, I can see two possible ways to correct it: 1) assumes that the goal is to profile the normal speed of the standard runcore; in this case, the empty loop should time just the profiling overhead, nothing else.  2) assumes that the goal is to profile the speed of ops *with dispatch ignored*
20:17 japhb , in which case the empty loop should time both the profiling overhead and the cost of dispatching to 'noop'.  Thoughts?
20:20 japhb Oh yeah, nothing like stopping conversation ...
20:21 Whiteknight japhb, whatever you said, I agree.
20:21 NotFound I even didn't know that the profiling runcore existed ;)
20:21 dalek r28917 | Whiteknight++ | gsoc_pdd09:
20:21 dalek : [gsoc_pdd09] updating to trunk r28915
20:21 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28917
20:21 japhb Whiteknight: Thank you.  That's one vote.  :-)
20:21 japhb NotFound: :-)
20:23 DietCoke Whiteknight: when you close out the ticket, be sure to resolve it and cc the list.
20:23 Whiteknight done and done
20:23 DietCoke my apologies as I didn't check the ticket, just the list. =-)
20:24 japhb DietCoke: any thoughts on my profiling question above?
20:24 DietCoke Other than, "I'd ask chromatic"?
20:24 japhb If no one really cares, I'm going with option 1, which makes more sense to me.
20:24 japhb summon chromatic
20:25 DietCoke chromatic: you awaken in a small, dank room. There is no source of light.
20:25 japhb DietCoke: been playing a lot of RPGs lately?  I'm noticing a theme this morning ....
20:28 NotFound Don't know if this helps, but I've bee able to build pdd25cx with c++ , and pass same test as with c, except for a problem already present in trunk.
20:30 chromatic japhb, #2 sounds more useful, but I'd like to replace the whole thing.  Profiling ops per ops doesn't seem as useful as profiling PIR (or PASM or whatever).
20:31 NotFound By the way, the problem is commented in RT#56440 Has someone looked at it?
20:32 japhb chromatic: certainly, replacement is the longer-term goal.  I wanted to unbreak the current thing before starting something new.
20:32 japhb But it wasn't obvious which was the proper 'unbreak'.
20:32 chromatic NotFound, my first impression was to say "That'll probably change with ParrotIO changes"
20:32 chromatic japhb, pick the smallest change that gets you useful information for now.
20:33 japhb chromatic: OK, thanks.
20:34 DietCoke japhb: sadly, I haven't played an RPG since about 1992.
20:34 NotFound chromatic: the implementation, sure, but the usefullness to have functions not dependant of io header I think will be the same.
20:34 chromatic NotFound, although if users of the external API need access to Parrot's printing systems... that's a strong argument for adding them here.
20:35 japhb DietCoke: I'm guessing high school or college, then ...
20:35 chromatic (though they could use tests in t/src/)
20:35 DietCoke Unless you count a phone call I got about the turn of the millenium from my old GM (who still GMs to this day) asking me how I ruled on one of my original characters worshippers offerings.
20:35 DietCoke (I thought _i_ was a geek. =-)
20:36 japhb yikes
20:37 japhb I know some people who are still GMing, and have by now built up fully populated worlds, with thousands of pages of notes, maps, history, etc.
20:37 NotFound Well, if users of the extend interface are supposed to not use that functions, using them in his tests looks to me like a bad idea.
20:39 chromatic If we can figure out some use case for why external interface users need to use those functions, adding them is fine.
20:42 NotFound By the way, all extend test need the embed header to be compiled. That does not help to have a clean separation between extend and embed.
20:44 NotFound The reason to use PIO_eprintf in the test looks to be in some cases to be able to print parrot strings, and the others by inertiae.
20:45 Debolaz joined #parrot
20:46 contingencyplan joined #parrot
20:48 jhorwitz NotFound: while writing both the embedding pod and mod_parrot, i had great difficulty finding a logical separation between extend and embed.  it's even more confusing when you consider that most extensions will probably use PIR & NCI.
20:48 chromatic I'm not sure there should be a separation sometimes.
20:49 jhorwitz Allison and I talked about it at frozen perl, and we both though they should be merged.
20:50 jhorwitz obviously wasn't a pressing issue at the time
20:50 * particle1 agrees
20:50 DietCoke chromatic: do you have a particular plan of attach to get us below 600 tickets?
20:50 NotFound IMO the only important difference is that an embedder creates the main interpreter instance.
20:51 chromatic DietCoke, the same thing we do every night: try to take over the world.
20:51 DietCoke Great. Where do I find the lederhosen?
20:51 chromatic Also, find someone who can test patches on Windows again.
20:51 DietCoke chromatic: I can do that in a pinch.
20:51 chromatic If your lederhosen are pinchy, try some corn starch.
20:51 dalek r28918 | japhb++ | trunk:
20:51 dalek : [PROFILE] Make profiling runcore calibration more useful
20:51 dalek : * Include fetch of high-precision timestamps in the profiling overhead
20:51 dalek : * Clean up calibration message to be more readable and correct
20:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28918
20:52 DietCoke I'll see about going through and cleaning some of that up.
20:52 jhorwitz NorFound: right.  but as an embedder i still use the functions declared in extend.h.  very confusing.
20:52 chromatic If extend.h should do anything in particular, it should be useful for people writing their own PMCs.
20:52 DietCoke chromatic: I'm going to cogitate on how to get all deprecated vtable calls to respect the new deprecation flag.
20:53 DietCoke chromatic: like HLL authors who currently cheat?
20:53 chromatic Lure them with cupcakes.  It always works on me, unless I have to develop on Windows or with the Solaris C compiler.
20:53 NotFound jhorwitz: the reason to have a separation can be to have several functions that an extender is supposed to never use. Don't know if there is one.
20:53 ruoso joined #parrot
20:54 jhorwitz exactly.  i can't even imagine a parrot extension in C.  that's what NCI is for.  of course there's an exception to every rule...even the ones i make up on the spot.  ;-)
20:55 NotFound Well, to solve the current practical problem, I can just drop the PIO_eprintf usage in t/src/extend, I don't see any need to maintain it.
20:56 NotFound The string printed contains plain asciic chars, no problem to convert them to cstrings before printing.
20:58 DietCoke ->
20:58 * jhorwitz reads back and sees cupcakes.
21:01 cotto_work php arrays have an internal pointer that's used for iterating through the array's elements.
21:01 cotto_work when an iterator PMC wants to know how many elements there are to be iterated, it uses VTABLE_elements()
21:01 cotto_work if the internal pointer is pointing at an element other than the first, it screws the iterator up
21:01 cotto_work the iterator works if I make elements() count the elements starting from the internal pointer
21:01 chromatic You get one iterator per array in PHP?
21:02 cotto_work this doesn't seem proper, though
21:02 cotto_work I can hear you laughing now
21:02 chromatic $POST['please_dont_modify_me_in​_a_request_iterator_pointer']?
21:02 szbalint ha-ha
21:02 confound isn't it $_POST?
21:02 PerlJam yikes!
21:02 chromatic No wonder I only rooted six out of twelve PHP installations trying that out a second ago.
21:02 PerlJam I come here to get *away* from PHP
21:03 confound PerlJam: mysql_escape_string_real!
21:03 szbalint quick, antidote!
21:04 * jhorwitz throws up a little in his mouth
21:04 NotFound cotto_work: Why an iterator wants to know that?
21:04 cotto_work would it be ok to do VTABLE_elements() based on the internal pointer and have a method that does the right thing, or is there a better solution
21:04 chromatic Oh hey, I could tell you about Perl 5's hash iterators, or filehandle iterators, or... you don't want to know.
21:04 cotto_work NotFound, to know where to start iterating
21:04 NotFound An iterator mus just iterate, that's his mission in life.
21:05 chromatic cotto_work, you could create your own PHP Iterator.
21:05 dalek r28919 | jkeenan++ | autoicu:
21:05 dalek : As there is no code in config/auto/icu.pm to process the '--icudatadir' option
21:05 dalek : mentioned in lib/Parrot/Configure/Options/Conf.pm and README_win32.pod, delete
21:05 dalek : references thereto from those files.  Begin to touch up the POD in
21:05 dalek : config/auto/icu.pm.  Add a fifth test file to test previously untested branch.
21:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28919
21:06 NotFound Maybe I'm too C++ oriented.
21:06 cotto_work subclass the iterator pmc?
21:06 cotto_work I suppose that would work
21:06 dalek r28920 | jkeenan++ | autoicu:
21:06 dalek : Update MANIFEST.
21:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28920
21:07 cotto_work chromatic, thanks
21:14 dalek r28921 | moritz++ | trunk:
21:14 dalek : [rakudo] first junction tests in spectest_regression
21:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28921
21:16 dalek r28922 | jkeenan++ | autoicu:
21:16 dalek : Improve POD in module and test files.
21:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28922
21:18 Auzon joined #parrot
21:36 NotFound It's not so easy to remove PIO_eprintf usage from t/src/extend.t as I thinked. There are different tests for getting parrot strings and getting cstrings, and string_to_cstring is not available from extend or embed headers.
21:38 japhb string_to_cstring sounds like the kind of thing an embedder might want access to ... but I haven't actually looked at the code.
21:39 NotFound There is a Parrot_free_cstring function in extend.h
21:41 chromatic It sounds like the easist approach to cleaning up some of that for now is to apply your patch then.
21:43 cjfields joined #parrot
21:44 Whiteknight any reason to delay rt56410?
21:44 Whiteknight it's just a patch to add a JSON data dumper language, doesnt change or remove anything existing
21:46 chromatic As far as I know, no one has cared either way on it.
21:46 chromatic I thought kjs had commit privs, but if he doesn't have access at home....
21:47 Whiteknight I'll add it nowish and close the ticket if there are no problems
21:47 Whiteknight we can delete it later if nobody wants it
21:47 chromatic +1
21:47 purl 1
21:47 chromatic +i
21:47 chromatic 1+i
21:47 chromatic 1+1
21:47 purl 2
21:47 chromatic For once, purl is rational.
21:48 Whiteknight haha, nice
21:48 allison whiteknight: sounds good
21:48 Whiteknight yay! I love reassurances!
21:49 szbalint maybe purl just can't express the irrationality. That doesn't mean it doesn't exist ;)
21:49 szbalint *-ity
21:49 NotFound By the way, someone can check if my CLA has arrived? According post office guy, he must, but spain's post is not known by his speed or his sincerity ;)
21:50 chromatic I'll know by Thursday.  Good thing, too -- I'm getting sick of applying your patches!  We should get you commit access to fix that.
21:50 chromatic Sick as in "I vomit with approval and desire for more free time", by the way.
21:51 chromatic Probably only amusing to Andy though.  Sorry.
21:51 confound I'm mildly amused
21:51 Andy I don't get it.
21:51 dalek r28923 | Whiteknight++ | trunk:
21:51 dalek : [languages/json]  Add languages/json, a PCT-based data dumper for JSON format. Thanks kjs++, from rt56410.
21:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28923
21:52 NotFound chromatic: I was trying to find an answer in the same mood, but my jokes are hard to translate.
21:53 chromatic Don't worry.  Your English has to be better than my Spanish.  estas malo grande
21:53 NotFound Next time I put them in spanish and let to you the translation
21:54 NotFound But many times native spanish also are unable to understand my jokes... sometimes even myself don't undesrtand it.
21:55 teknomunk joined #parrot
21:55 chromatic Let me put it this way.  Sometimes I confuse ojos with huevos, and as it turns out, that's not something to do in polite company.
21:56 japhb Can bytecode move in memory once it's been loaded?  In other words, is it reasonable to record PC values during execution, and then only at exit translate those PC values to source PBC file offsets?
21:56 NotFound I confuse breasts with breath, but that is more anatomic comprehensible.
21:56 chromatic That should work.
21:57 japhb chromatic: was that to me?
21:57 NotFound I hope that was :D
21:58 * japhb hasn't backlogged enough to see if someone else was asking a question .... :-)
21:58 NotFound ETOOMANYJOKES again.
21:59 * japhb finishes backlogging
22:00 japhb OK, good.  Now I have an idea how to *approach* doing PIR/PBC-level profiling.  Of course, there's the SMOP to do it ....
22:00 chromatic japhb, yes.
22:01 chromatic My best idea so far was to emit Callgrind-compatible output.
22:01 japhb Happen to have a link handy to what that format is?  If not, I'll just google for a while ...
22:02 chromatic http://kcachegrind.sourceforge.net/cgi-​bin/show.cgi/KcacheGrindCalltreeFormat
22:07 japhb thx, reading
22:13 NotFound A quick look reveals a possible problem: the format is ASCII based and we can have non-ascii function names.
22:15 pmichaud I'm guessing the test harness changed today?  "make spectest_regression" in rakudo gives a lot of
22:15 pmichaud t/spec/S03-operators/numeric-context.............dubious Test returned status 1 (wstat 256, 0x100) after all the subtests completed successfully
22:15 pmichaud where it didn't used to do that.
22:15 moritz pmichaud: I had the same problem
22:15 chromatic Does Rakudo have its own test harness?
22:15 pmichaud t/harness, yes.
22:16 japhb NotFound: I'm not through reading it, but it looks from the early bits like I might be able to specify locations in terms of .PBC file and offset, ignoring higher-level location concepts entirely
22:16 chromatic Okay.  I touched t/harness a couple of times, but that's the main harness, not the Rakudo harness.
22:16 moritz but it delegates to Parrot::Test::Harness in the end
22:17 dalek r28924 | moritz++ | trunk:
22:17 dalek : [rakudo] Implement prefix:+, prefix:-, sqrt() and exp() for Complex.
22:17 dalek : Patch curtesy of bacek++, Vasily Chekalkin <bacek@bacek.com>
22:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28924
22:17 NotFound japhb: no problem in that case.
22:17 moritz but I think it was a different test that behaved like that
22:18 japhb NotFound: of course, it means I would need to force generation of PBC files when profiling is on, but that seems like less a hardship.
22:19 NotFound japhb: if not, using pir escaping will be a simple solution.
22:19 dalek r28925 | moritz++ | trunk:
22:19 dalek : [rakudo] S29-num/complex.t is now pure. bacek++
22:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28925
22:19 japhb NotFound: roger that
22:19 pmichaud moritz++   # applying Complex patch
22:19 pmichaud we're going to have another jump in passing tests today :-)
22:20 Whiteknight (passing tests)++
22:20 pmichaud last I checked about an hour ago we were at 1287
22:20 Whiteknight wow, that's a huge jump. you just broke 1000 yesterday, right?
22:20 bacek_ moritz: thank, but you shouldn't commit this patch... pmichaud didn't like it :)
22:20 pmichaud friday, actually.
22:20 moritz http://rakudo.de/progress.png
22:20 pmichaud bacek_: moritz fixed the patch.
22:20 moritz bacek_: we discussed it on #perl6
22:21 bacek_ pmichaud: ok, thank. I didn't read commit diff
22:21 bacek_ (morning everyone)
22:21 moritz bacek_: I moved the builtins to Complex.pir
22:21 chromatic Ah, the first patch touched Parrot::Test::Harness.
22:24 bacek_ pmichaud: btw, I rewrote patch in RT#56214
22:24 bacek_ moritz: there is but in S29-str/index.t (pmichaud++ for pointing it).
22:25 bacek_ Str.index doesn't return -1 in case of failure. It should return 'bare StrPos' which evaluated to False
22:25 bacek_ s/but/bug/
22:26 moritz bacek_: an undef value of type StrPos or what?
22:26 bacek afk # $kids->school(); $self->dayjob()
22:26 pmichaud moritz: it's not entirely clear from S29
22:27 bacek 'bare StrPos'. Evalutes to False in boolean context, to 0 in scalar context (with emiting warning).
22:27 bacek definitely afk
22:28 moritz but it's enough to test for ok !$result; and for isa_ok($result, StrPos)
22:28 moritz but I wonder if that still makes it right to write is(index(...), 3)
22:29 pmichaud well, testing for !$result may be problematic if the string being looked for happens to be at position 0
22:29 dalek r28926 | pmichaud++ | trunk:
22:29 dalek : [rakudo]:
22:29 dalek : * Fix list context flattening when a lazy element (e.g., a Range)
22:29 dalek :   increases the number of items in the resulting list.
22:29 dalek : * cjfields++ for finding this and providing useful test examples.
22:29 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28926
22:30 moritz then for DWIMminess it has either to be undef, or ?StrPos always true for any defined StrPos
22:30 pmichaud PHP has this problem also with its strpos() function -- if one wants to determine that  substring is not within a given target, we have to check explicitly for === FALSE and not just !strpos(...)
22:31 pmichaud however, strpos could return   0 but True   if the substring is found at position zero :-)
22:31 pmichaud sorry, Str.index (forgot to switch back into Perl 6 mode)
22:31 moritz pmichaud: wouldn't it make sense to always return a StrPos?
22:32 pmichaud okay,  StrPos(0) but True
22:32 cotto_work PHP has many problems
22:32 jonathan Mmm...the "but" operator.
22:32 moritz "If the substring is not found, a bare StrPos containing no position is returned. This prototype StrPos evaluates to false because it's really a kind of undef. "
22:33 moritz ok, so it *is* undef
22:33 pmichaud it's not defined, yes, but I wasn't sure if the StrPos that gets returned is the StrPos protoobject or some other undef'd StrPos
22:35 * moritz wonders why the tab key in vim can't complete IRC nicks (for the commit messages) ;-)
22:41 japhb chromatic: The callgrind spec on the link you found is truncated.  This one looks like a fixed version of the same spec: http://valgrind.org/docs/manual/cl-format.html
22:41 chromatic Yeah, it is.
22:47 Ademan joined #parrot
22:47 moritz pmichaud: I hope you don't mind that I steal some of your tickets (for closing)
22:48 moritz well, I don't really steal them, I just close them ;)
22:50 pmichaud they're not my tickets.  :-)
22:50 pmichaud so, close away :-)
22:51 wknight8111 joined #parrot
22:56 moritz well, you'r the "owner" ;)
22:56 moritz at least from some of them
23:53 cotto_work If an old ticket refers to a chunk of code that no longer exists, should it be marked as resolved or rejected?
23:53 chromatic Rejected.
23:55 bacek_ joined #parrot
23:55 dalek r28927 | chromatic++ | trunk:
23:55 dalek : [t] Made TAP::Harness actually *run* tests during make smolder_test.
23:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28927

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

Parrot | source cross referenced