Camelia, the Perl 6 bug

IRC log for #parrot, 2008-08-28

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 bacek_ joined #parrot
00:15 dalek r30605 | Whiteknight++ | trunk:
00:15 dalek : [Core] add function-level documentation to src/inter_run.c, which I was not able to understand earlier.
00:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30605
00:30 donaldh joined #parrot
00:50 dalek r30606 | julianalbo++ | trunk:
00:50 dalek : fix lack of cut in a pod comment
00:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30606
01:00 dalek r30607 | Whiteknight++ | pdd09gc:
01:00 dalek : [pdd09gc] add new src/gc/gc_ms.c to MANIFEST
01:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30607
01:07 kid51 seen magnachef
01:07 purl magnachef was last seen on #parrot 50 days, 22 hours, 29 minutes and 56 seconds ago, saying: s/is/his/  [Jul  8 02:37:20 2008]
01:19 rurban_ joined #parrot
01:21 particle1 joined #parrot
01:34 dalek r30608 | cotto++ | trunk:
01:34 dalek : [pipp] phparray tests for add and i_add
01:34 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30608
01:39 kid51 Whiteknight ping
01:50 kid51 moritz ping
01:52 Whiteknight hello kid51
01:53 kid51 Ah, I just msg-ed you!
01:53 Whiteknight yeah, I just got it
01:53 Whiteknight East Coast hackathon? count me in!
01:54 dalek r30609 | Whiteknight++ | gsoc_pdd09:
01:54 dalek : [gsoc_pdd09] update to trunk from r30457
01:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30609
01:55 kid51 This was a *very* tentative discussion.  We have neither location nor date determined.  In fact, I haven't even talked about it with him or anyone in two months!
01:55 kid51 But now that cooler weather is arriving, my brain is more capable of longer-range planning.
01:56 kid51 So if you can msg me an email address, I'll add you to any discussion that develops.
01:57 Whiteknight okay, i sent it in a privmsg
01:59 kid51 Ever since attending the Perl hackathon in Chicago in Nov 2006, I've loved hackathons and wanted to organize one on the East Coast.  But in the NYC area finding affordable space is a killer.
01:59 Whiteknight yeah, I imagine
01:59 kid51 So that's why I've tried to keep in contact with people from Phila.pm, ABE.pm, Coke (Albany).
02:00 Whiteknight I just signed up for the Phila.pm and ABE.pm mailing lists
02:00 kid51 If you're ever in NYC on the 3rd Tuesday of a month between October and May, you can come to Perl Seminar NY (http://tech.groups.yahoo.com/group/perlsemny).
02:01 kid51 Our meeting location is close to Penn Station and Port Authority!
02:05 Whiteknight oh nice. I'm never in NYC, but I might make the trip one day
02:11 cotto_home how do I test error output in a pure pir test?
02:14 cotto_home I suppose catching an exception would be part of it
02:18 cotto_home answering my own questions is fun
02:18 kid51 There's this fellow on channel called cotto_work who probably has all the answers to your questions ;-)
02:19 cotto_home I'll have to msg him.
02:20 cotto_home of course, he always seems to be here when I'm not, so there'll probably be some delay.
02:24 Whiteknight goodnight
02:32 cognominal joined #parrot
03:33 dmknopp joined #parrot
03:54 petdance joined #parrot
04:19 tetragon joined #parrot
04:36 Ademan joined #parrot
04:37 tetragon joined #parrot
04:44 konobi joined #parrot
04:44 konobi yoyo
04:49 konobi are there any PGE/TGE examples of grammars that deal with interleaved language/markup ?
04:51 PerlJam konobi: I don't think so.
04:52 PerlJam Though, pipp might have something useful there since it's an implementation of php
04:57 konobi so nothing like a grammar for TT or the like?
05:19 dalek r30610 | cotto++ | trunk:
05:19 dalek : [pmc] resolve 46823 and delete spurious references to the same
05:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30610
05:25 Psyche^ joined #parrot
05:44 tetragon joined #parrot
05:48 dalek r30611 | cotto++ | trunk:
05:48 dalek : [pmc] add some exceptional tests for negative indicies
05:48 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30611
06:00 donaldh joined #parrot
06:13 uniejo joined #parrot
06:20 dalek r30612 | chromatic++ | trunk:
06:20 dalek : [dynpmc] Changed string_from_literal() to CONST_STRING() in GDBMHash PMC.
06:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30612
06:22 allison joined #parrot
06:25 dalek r30613 | chromatic++ | trunk:
06:25 dalek : [lib] Made dynpmc group generator initialize dynpmcs in order of their
06:25 dalek : inheritance hierarchy within the group.  This is step two in reducing vtable
06:25 dalek : code size, memory use, linking, and startup time.
06:25 dalek : The next step may be murkier; the no_init flag for PMCs may need to go away.
06:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30613
06:29 tetragon joined #parrot
07:05 tetragon joined #parrot
07:05 magnachef_ joined #parrot
07:14 dalek r30614 | cotto++ | trunk:
07:14 dalek : [pipp] phparray get_repr test
07:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30614
07:16 iblechbot joined #parrot
07:24 dalek r30615 | allison++ | trunk:
07:24 dalek : Slate the n_* opcode variants for deprecation.
07:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30615
07:40 Ademan joined #parrot
09:03 dalek r30616 | allison++ | trunk:
09:03 dalek : [pdd] Adding a draft installation PDD, thanks to Reini Urban.
09:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30616
09:18 rurban_ joined #parrot
09:23 dalek allison@perl.org | milestones:
09:23 dalek link: http://www.perlfoundation.org​/parrot/index.cgi?milestones
09:35 rba joined #parrot
10:16 Debolaz joined #parrot
10:21 megamic joined #parrot
10:29 dalek allison@perl.org | milestones:
10:29 dalek link: http://www.perlfoundation.org​/parrot/index.cgi?milestones
10:39 masak joined #parrot
11:01 iblechbot joined #parrot
11:12 Zaba joined #parrot
11:31 kid51 joined #parrot
11:59 dalek r30617 | julianalbo++ | trunk:
11:59 dalek : allow struct and pointer to pointer in pmc ATTR
11:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30617
12:09 Ademan joined #parrot
12:10 dmknopp left #parrot
12:24 dalek r30618 | julianalbo++ | trunk:
12:24 dalek : clean ATTR workarounds in perl6multisub
12:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30618
12:47 tswall joined #parrot
13:01 ejs_ joined #parrot
13:03 dalek rurban@cpan.org | dtrace:
13:03 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
13:09 dalek rurban@cpan.org | parrot_installation:
13:09 dalek link: http://www.perlfoundation.org/par​rot/index.cgi?parrot_installation
13:11 dalek rurban@cpan.org | parrot_installation:
13:11 dalek link: http://www.perlfoundation.org/par​rot/index.cgi?parrot_installation
13:15 jan joined #parrot
13:30 gryphon joined #parrot
13:50 jdv79 joined #parrot
13:59 dalek St�phane Payrard | dtrace:
13:59 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
14:02 jhorwitz joined #parrot
14:06 dalek St�phane Payrard | dtrace:
14:06 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
14:07 AndyA joined #parrot
14:10 ejs__ joined #parrot
14:18 masak can NCI be used in Parrot today to make Rakudo talk to mySQL?
14:19 particle1 iirc donaldh was looking into this after tewk's merge
14:19 masak so there's some hope, in other words
14:20 particle yes, much hope
14:20 masak nice
14:20 particle if it doesn't work today, it could work within a week or four
14:20 particle it's totally solvable
14:20 masak sounds great
14:20 * masak reads up on pdd16 in the meantime
14:22 * particle notices 'pdd16' doesn't become a link in the irc log :(
14:22 moritz particle: in general your changes are better if you hilight me ;)
14:22 particle i knew you were nearby anyway :)
14:25 particle m/(?i)\bpdd\d\d(_\w+)*\b/ ought to match it
14:25 moritz qr{\bpdd(\d\d)\b}
14:25 moritz oki
14:25 particle actually, you probably want to capture the pddNN part
14:26 moritz actually pdd(\d\d)(_\w+)?\b is enough
14:26 moritz \w matches _ also
14:26 particle ah, yes
14:27 particle fine with me if they all link to http://www.parrotcode.org/docs/pdd/
14:27 particle if you can't get the svn linking right due to the draft/ dir
14:27 moritz I think in the browser it's easier to read HTML anyway
14:28 particle yep
14:28 moritz sadly I need a list of their names, because I can't link to them only by number
14:28 particle right, unless you scrape html or something nasty
14:29 moritz I'll just put a hash into my module, basta!
14:29 particle ack -ag pdd docs/pdds/ will work
14:31 Zaba ack <3
14:36 particle ack++
14:39 moritz how do I replay a macro in vim?
14:40 masak @r
14:40 masak where r is the register
14:40 moritz thanks, masak++
14:40 masak or @@ if you played it last
14:40 * masak wishes there were a way involving only one keypress
14:41 moritz you just saved me 300+ key strokes
14:41 moritz :bind yourkey @@ ?
14:41 masak moritz: might work, yes :)
14:41 particle yep, like <F5> or something
14:41 * masak has never bound keys in vim before
14:42 * particle binds ` to :bn for presentations
14:42 particle yes, i do presentations in vim
14:42 moritz woooo
14:46 masak particle: Damian Conway too. he uses something called vimpoint.
14:48 particle yeah, i need to steal some of his ideas
14:49 particle wish i could get vim fullscreen on windows, too
14:55 ilbot2 joined #parrot
14:55 Topic for #parrotis now http://www.parrotcode.org | Last Release: 0.7.0 "Severe Macaw"
14:55 moritz logger should be back now
15:15 dalek r30619 | moritz++ | trunk:
15:15 dalek : [rakudo] added test for statement modifying for to spectest_regression
15:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30619
15:16 bpphillips joined #parrot
15:44 particle1 joined #parrot
15:54 uniejo joined #parrot
16:10 dalek r30620 | allison++ | trunk:
16:10 dalek : [pdd] Adding an early draft PDD for HLL interoperability, from Bob Rogers.
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30620
16:16 davidfetter joined #parrot
16:16 iblechbot joined #parrot
16:38 konobi left #parrot
16:46 Ademan joined #parrot
17:01 dalek rurban@cpan.org | parrot_installation:
17:01 dalek link: http://www.perlfoundation.org/par​rot/index.cgi?parrot_installation
17:05 cognominal joined #parrot
17:18 rurban_ joined #parrot
18:00 Zaba_ joined #parrot
18:28 dalek r30621 | tene++ | trunk:
18:28 dalek : [cardinal]
18:28 dalek : Add a file that I forgot to include earlier
18:28 dalek : sleephacking--
18:28 dalek : dmknopp++
18:28 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30621
18:33 dalek rurban@cpan.org | Platforms:
18:33 dalek link: http://www.perlfoundation.or​g/parrot/index.cgi?platforms
19:13 gryphon joined #parrot
19:44 dalek r30622 | allison++ | trunk:
19:44 dalek : [pdd] Adding a series of architectural comments to the HLL Interop PDD.
19:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30622
19:45 itz joined #parrot
19:51 dalek moritz@faui2k3.org | Platforms:
19:51 dalek link: http://www.perlfoundation.or​g/parrot/index.cgi?platforms
19:59 NotFound Somenone fluent with nci?
19:59 * moritz hopes tewk is ;)
20:00 NotFound I've found that the signature for strings does not allow for NULL strings, and I wonder if is intentional.
20:00 NotFound Just trying a simple test with XDisplayName
20:02 NotFound With this change, it works: "t_$temp_num = string_to_cstring(interp, GET_NCI_S($reg_num));";  ---> "t_$temp_num = GET_NCI_S($reg_num) ? string_to_cstring(interp, GET_NCI_S($reg_num)) : (char *) NULL;"
20:03 NotFound tools/build/nativecall.pl:505
20:08 grim_fandango joined #parrot
20:12 peepsalot joined #parrot
20:30 Tene I was going to write mail to the list about cardinal HLL interop design questions, but I need to do more research first.
20:30 Tene The big issue is that ruby defines keyed lookup and assignment as calling methods named '[]' and '[]=
20:30 Tene ' on objects.
20:31 Tene Should I make set_pmc_keyed look for those methods specifically?  Should I watch for those names on method definition in the compiler and override set_pmc_keyed in the class?
20:31 Tene I haven't sat down to do the thinking about design yet, and haven't tried the later to see if it's possible.
20:33 particle1 .sub '[]=' :method :vtable('set_pmc_keyed')
20:34 particle is that what you're getting at?
20:34 particle i don't know ruby at all
20:35 Tene That might be it.
20:35 Tene Can I do that with p6object for user-defined classes, though?
20:35 particle if not, it's a missing feature
20:36 particle that is, if you're looking for some special syntax p6object provides
20:36 particle user-defined classes should work just like core classes in that regard
20:37 Tene Okay.
20:37 Tene I'll investigate that tonight.
20:38 particle what's the latest on resumable exceptions?
20:38 Tene They work fine still.
20:38 particle i'd like to put warnings into rakudo
20:38 particle great
20:38 Tene Or do you mean automatically resuming on non-fatal?
20:39 uniejo joined #parrot
20:39 Tene Exception handlers right now can only handle a single exception before the handler is marked as already-used and then ignored in the find_exception_handler logic.
20:39 Tene So, that's a bug.
20:40 Tene http://pleasedieinafire.net​/~tene/parrotexceptionnotes are the notes I've made so far for myself about the exception system.
20:42 particle thanks
20:43 Whiteknight joined #parrot
20:44 Tene The fix for that bug is to get every exception handler to check the type of their exception and rethrow exceptions they don't intend to catch.
20:44 Tene The current behavior is a workaround for invinite loops caused by exception handlers throwing exceptions.
20:45 particle does the pdd have an example of a properly formed handler?
20:59 NotFound I've sended the question about NCI null strings and a patch to parrotbug.
20:59 NotFound RT#58438
20:59 Whiteknight so what does that require, that exception handlers have some kind of signature?
21:00 particle no, it requires exception handler writers to use 'rethrow'
21:01 particle in pseudocode: eh_mytype: .get_params pmc e; unless e.type = 'mytype' rethrow; #handle 'mytype' exception here
21:03 Whiteknight That seems like it could be ripe for some kind of syntactic sugar or macros
21:03 Ademan joined #parrot
21:04 particle yes, it does
21:04 Whiteknight Like HANDLER_CATCH_ONLY(handler_name, type_expected)
21:05 moritz and you pass in a junction of handlers to catch ;)
21:36 Tene particle: how much of that did you read?
21:37 particle if you mean your notes, none yet
21:37 Tene Ah.
21:37 Tene I need to formalize them a bit into a mail for allison.
21:37 * particle is debugging javascript
21:37 Zaba joined #parrot
21:37 Tene Maybe I can talk myself into doing that tonight, though.
21:38 Tene There are issues with exceptions thrown from C.
21:38 Tene According to comments in that file, you should be able to resume just by returning from throw_from_c, but in practice, there are issues.
22:06 dalek r30623 | Whiteknight++ | pdd09gc:
22:06 dalek : [gsoc_pdd09] update to trunk
22:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30623
22:10 idemal joined #parrot
22:23 kid51 joined #parrot
22:24 bacek joined #parrot
22:25 bacek morning everyone
22:26 PerlJam afternoon here :)
22:26 moritz not quite morning here ;)
22:26 Whiteknight Good evening
22:29 cotto_work good night
22:29 purl sleep well too
22:29 cotto_work (not really)
22:36 bacek joined #parrot
22:45 bacek joined #parrot
23:09 tetragon joined #parrot
23:14 evan joined #parrot
23:14 evan how does parrot allow for profiling? (or will do)
23:14 evan opcode replacement like perl5?
23:15 moritz perhaps a different runloop
23:15 PerlJam indeed, I think it is a different runloop
23:16 evan ok, so it does it now?
23:16 evan I wanted to know what to look for
23:17 PerlJam parrot has profiling, PIR does not.
23:18 PerlJam i.e., you can know how much time is spent in particular opcodes and how often they are executed, but you can't easily know how many times a .sub is executed or how much time is spent inside.
23:19 bacek joined #parrot
23:19 evan PerlJam: ah! so it's opcode profiling
23:19 evan but no facility for tracking sub entry/exit
23:20 PerlJam evan: right.  That's an outstanding problem as far as I know.  Something that would be *really* nice to have if only someone would implement it.
23:21 evan sadly, i'm not your man for that job.
23:21 evan i'm asking because i'm surveying how other VMs do it
23:21 evan to add to my VM.
23:22 PerlJam well, -p is for profilling parrot if you ever happen to want to do that
23:22 evan that outputs opcode usage stats?
23:23 PerlJam aye
23:23 evan ok, thanks for the info!
23:25 particle evan: i suggest you use callgrind-compatible output
23:25 evan yeah, thats what I'm thinking too
23:25 evan it's so dead simple
23:25 evan i'm not sure how I couldn't
23:25 particle yep
23:26 particle what's your vm for?
23:26 PerlJam yeah, why aren't you using parrot?  :)
23:26 particle hee
23:29 evan particle: Rubinius
23:29 evan a new ruby environment
23:31 particle ah yes, i know of it
23:33 PerlJam Will this one make ruby faster?
23:35 particle evan: rubini.us mentions "spec-driven", but i know ruby is like perl 5 in that the implementation is the spec
23:36 particle so what's spec-driven, the vm?
23:36 PerlJam doesn't ruby have multiple implementations?
23:36 evan particle: both
23:36 evan we're writing a spec for MRI, the main implementation
23:37 particle yes, ruby has multiple implementations
23:37 particle ah, i didn't know that
23:37 particle *finally* :)
23:37 evan and we then make Rubinius match that spec
23:37 evan otherwise it's just not possible to write another implementation
23:38 PerlJam for some reason I parsed that domain as ruby-in-us in my head.
23:38 dalek r30624 | Whiteknight++ | gsoc_pdd09:
23:38 dalek : [gsoc_pdd09] add some file-level documentation to explain what the hell I think I'm doing here. Also, add a quick stub for a quick duplication function I'm going to add later.
23:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30624
23:39 evan particle: http://www.rubyspec.org/
23:40 particle evan: have you considered using llvm components (like their jit) or are you building all your subsystems from scratch?
23:40 evan we're currently finishing up a new VM arch that uses llvm
23:40 evan so yes!
23:40 particle i'll have to keep an eye on you guys
23:55 bacek joined #parrot

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

Parrot | source cross referenced