Camelia, the Perl 6 bug

IRC log for #parrot, 2008-09-01

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 zostay joined #parrot
00:22 Tene I thought I saw some list traffic recently about the n_* ops?
00:23 Limbic_Region joined #parrot
00:43 Tene It's nice having someone else work on cardinal with me.
00:43 Tene dmknopp++
00:43 dalek r30675 | tene++ | trunk:
00:43 dalek : [cardinal]
00:43 dalek : * Fix puts to print a newline before every item
00:43 dalek : * Fix the test harness
00:43 dalek : * Array::each_with_index
00:43 dalek : * Array::include?
00:43 dalek : * not
00:43 dalek : dmknopp++
00:43 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30675
00:44 dmknopp :-)
00:50 bacek joined #parrot
01:01 megamic joined #parrot
01:27 davidfetter joined #parrot
01:53 Zaba joined #parrot
02:15 dalek r30676 | tene++ | trunk:
02:15 dalek : [cardinal]
02:15 dalek : * First draft of 'require' and 'load'
02:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30676
02:15 bacek joined #parrot
03:02 bacek_ joined #parrot
03:09 dalek r30677 | tene++ | trunk:
03:09 dalek : MANIFEST fixes for new cardinal files.
03:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30677
03:22 dalek r30678 | tene++ | trunk:
03:22 dalek : [cardinal]
03:22 dalek : * Fix a problem with globals.
03:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30678
03:22 dalek r30679 | tene++ | trunk:
03:22 dalek : [cardinal]
03:22 dalek : * test library
03:22 dalek : * test library tests
03:22 dalek : * convert one test file to use the new test library
03:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30679
03:24 spinclad joined #parrot
03:25 dalek r30680 | tene++ | trunk:
03:25 dalek : MANIFEST and metadata fixes for new cardinal files
03:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30680
03:36 Zaba joined #parrot
04:34 bacek joined #parrot
04:41 Ademan joined #parrot
05:20 Psyche^ joined #parrot
05:23 Ademan joined #parrot
06:12 uniejo joined #parrot
07:02 Orchygirl joined #parrot
07:11 iblechbot joined #parrot
07:21 cosimo joined #parrot
07:30 Orchygirl left #parrot
07:39 Tene I'm running into some weird problems when trying to use a precompiled test.rb in cardinal.  Making a note here to discuss it later.
07:39 Tene Got a lot done in cardinal. dmknopp++ for motivating me.
07:41 moritz does ruby have a reusable test suite?
07:43 barney joined #parrot
08:39 viklund joined #parrot
11:27 Zaba_ joined #parrot
11:55 Zaba joined #parrot
12:00 Zaba_ joined #parrot
12:25 rba joined #parrot
12:40 Debolaz joined #parrot
13:38 slavorg joined #parrot
13:46 tetragon_ joined #parrot
14:08 paco joined #parrot
14:15 Whiteknight joined #parrot
14:22 dmknopp joined #parrot
14:42 masak joined #parrot
14:45 iblechbot joined #parrot
15:42 tetragon joined #parrot
15:44 japhb moritz: The twigils lesson says 'seld-declared positional parameter'.  Typo in first word there.
15:45 japhb moritz: and keep up the good work, these are very well done.
15:48 Zaba joined #parrot
15:48 moritz japhb: thank, will correct real soon
15:49 japhb moritz: hmmm,  and the next example doesn't have the twigils in the interpolating string
15:50 japhb moritz: and it looks like there are a few places later where you say 'sigil' but mean 'twigil'.
15:50 japhb Did I catch you mid-edit?  :-)
15:51 japhb And 'specical' in the second paragraph of MOTIVATION
15:52 japhb OK, all done
15:52 * japhb goes back to hacking ...
15:56 moritz japhb++
15:56 moritz fixed now
15:56 Whiteknight joined #parrot
16:16 iblechbot joined #parrot
16:24 Zaba joined #parrot
16:46 stupidbot joined #parrot
16:50 dalek r30681 | tene++ | trunk:
16:50 dalek : [cardinal]
16:50 dalek : * Support default arguments to functions
16:50 dalek : * Add a test and convert another test file to use the test library
16:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30681
16:50 dalek r30682 | tene++ | trunk:
16:50 dalek : [cardinal]
16:50 dalek : * String + String
16:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30682
16:51 dalek r30683 | tene++ | trunk:
16:51 dalek : [cardinal]
16:51 dalek : * More functions in test.rb
16:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30683
16:51 dalek r30684 | tene++ | trunk:
16:51 dalek : [cardinal]
16:51 dalek : * typo fix
16:51 dalek : * die
16:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30684
16:51 * jonathan is back from vacation
16:52 Tene HIJONATHAN!
16:52 moritz hi jonathan ;)
16:55 jonathan :-)
16:56 moritz the bug tracker welcomes you with >150 new/open tickets in the perl6 queue ;)
17:00 dalek r30685 | tene++ | trunk:
17:00 dalek : [cardinal]
17:00 dalek : * Array + Array
17:00 dalek : * Array::collect
17:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30685
17:00 dalek r30686 | tene++ | trunk:
17:00 dalek : [cardinal]
17:00 dalek : * Array::sort
17:00 dalek : * Array::uniq
17:00 dalek : * Array == Array
17:00 dalek : * Generic !=
17:00 dalek : * infix::&
17:00 dalek : * Update a few test files
17:00 dalek : * Only one test failure left
17:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30686
17:01 dalek r30687 | tene++ | trunk:
17:01 dalek : MANIFEST and metadata changes for new cardinal file.
17:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30687
17:06 jonathan 150?! :-O
17:07 moritz no, just 146
17:07 moritz I exaggerated a bit
17:08 Tene Mmm... only one test failure left on cardinal.  I need to harass dmknopp into writing more tests for me.
17:09 moritz can you reuse some of the original ruby tests?
17:09 Tene Official ruby has no tests.
17:09 cotto_home you're serious?
17:09 Tene There's ruby-spec.org, but I need some weird meta stuff before I can run those, I think.
17:09 moritz it has something that is run when you say "make test"
17:09 moritz but I guess they partly rely on assertions
17:10 Tene I also don't like the ruby-spec.org tests because they use a silly pretend DSL.
17:11 Tene Silly, I know.
17:11 * moritz can't judge that
17:11 Tene Eh, lemme see what I'd actually need.
17:14 moritz in PIR, how can I let a match of a (Perl 6) regex start from a specific position?
17:17 gmansi joined #parrot
17:23 * moritz just wrote 30 lines of PIR code without a single syntax error
17:23 moritz record for me
17:24 Tene yay!
17:25 moritz it doesn't work, of course :/
17:26 moritz and of course it doesn't work, it calls the wrong multi :/
17:27 moritz so the code I wrote is never executed
17:41 stupidbot joined #parrot
17:46 dalek r30688 | moritz++ | trunk:
17:46 dalek : [rakudo] add recursion tests to spectest_regression.data
17:46 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30688
17:48 tetragon joined #parrot
17:58 particle1 joined #parrot
18:09 gmansi joined #parrot
18:10 jan joined #parrot
18:11 stupidbot joined #parrot
18:13 jhorwitz joined #parrot
18:31 sjansen joined #parrot
18:40 Tene Okay, all tests pass now.  Need more.
18:40 dalek r30689 | tene++ | trunk:
18:40 dalek : [cardinal]
18:40 dalek : * %w[like perl's qw]
18:40 dalek : * all tests pass
18:40 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30689
18:40 tetragon joined #parrot
18:46 eternaleye moritz: I figured out why I was getting that build failure: it occurs on make -j3 but not make -j1
18:48 moritz eternaleye: that's bad
18:48 eternaleye nopasting build output now
18:48 moritz eternaleye: I use parallel builds all the time (on a dual core machine), but not 'make installable'
19:05 eternaleye Hm. It seems that 'make reallyinstall' only installs the languages/* directories, and not their contents, whether 'make world' or 'make installable' is used
19:11 eternaleye make -j1 installable output: http://rafb.net/p/OEFoNw78.html . make -j3 installable output: http://rafb.net/p/ryFiej50.html
19:21 dmknopp h3llz yeah stephen! you killed those tests
19:22 dalek r30690 | allison++ | pdd27mmd:
19:22 dalek : [pdd27mmd] Convert all the 'infix' opcodes to use the new multi_dispatch call.
19:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30690
19:37 cognominal what is the PIR way to call a  VTABLE_set_attr_str in a pmc?
19:38 jonathan setattribute the_pmc, "name", value
19:38 cognominal strange, I could not ack at that...
19:39 cognominal and the Perl 6 way ?
19:39 moritz cognominal: what does it do (high level)?
19:40 jonathan $!name = value;
19:40 cognominal suppose I want to get (not set here) the id attribute of  task. How?
19:41 cognominal > say Task.new().id
19:41 cognominal Null PMC access in get_string()
19:41 cognominal ho, that's because it is private
19:41 cognominal but what Taks.new().id returns anyway?
19:42 jonathan If it's private, there shouldn't be a method...
19:43 cognominal that's my point
19:43 jonathan Curious. How's it declared?
19:43 jonathan has $!foo ?
19:43 cognominal I don't know.
19:47 cognominal btw,  I usbmitted a parrot bug that did not show up.
19:47 cognominal Type at the rakudo prompt :  class A { has $a; }; my $a = A.new()
19:47 cognominal repeat two more times and you crash parrot.
19:48 cognominal And the stack trace is strange too, I don't think rakudo (?) does the right thing.
19:49 jonathan Doing it two times and crashing Parrot isn't right. :-)
19:50 cognominal with the first time, that makes three times
19:50 cognominal you must hit three times to kill the parrot
19:51 jonathan Third time lucky!
19:51 cognominal interestingly enoug doing the same thing  with  only class A { has $a; } gets a cleaner kill
19:52 cognominal compilers/imcc/pbc.c:993: failed assertion 'PObj_is_string_TEST(lex_name)'
19:52 cognominal Backtrace - Obtained 32 stack frames (max trace depth is 32).
19:52 cognominal Parrot_confess...
19:52 cognominal or messier depending on the point of view.
19:53 jonathan Ugh.
19:53 jonathan Thing is, that should work...
19:53 cognominal Sorry for throwing that on your lap :)
19:55 cognominal on Mac OS X and Cygwin, so that should be on many platform
19:57 cognominal what is strange is that getting an error in the rakudo prompt seems to give you a new runloop so things stacks up.
19:57 tetragon joined #parrot
19:57 cognominal This one seems a rakudo problem.
19:57 cognominal And that in return triggers a bug in parrot.
19:59 cognominal doing a dod_register within a dod_register is probably very healthy
20:00 cognominal dod_register_pmc I mean.
20:00 dalek r30691 | cotto++ | trunk:
20:00 dalek : [pmc] PIRified and added documentation to t/pmc/resizablepmcarray.t
20:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30691
20:00 jonathan Seems Rakudo problem, but maybe also Parrot one.
20:01 cognominal bugs like to live in groups :)
20:17 cotto_home is there any reason it'd be a bad idea to move the fp equivalence macros into their own file somewhere under runtime/parrot?
20:18 cotto_home currently they're generated from a static string in lib/Parrot/Test.pmc
20:18 cotto_home s/pmc/pm/
20:18 tetragon joined #parrot
20:30 dalek r30692 | cotto++ | trunk:
20:30 dalek : [pmc] comment and whitespace fix
20:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30692
20:50 Zaba joined #parrot
20:53 Theory joined #parrot
20:57 tetragon joined #parrot
21:08 Tene So, I copied the 'use' stuff from rakudo for cardinal's 'require'
21:09 Tene So theoretically it should support precompiled pir.
21:09 Tene However, stuff breaks badly when I try to use it.
21:09 Tene But then it works fine if I add a :load to the main sub in the generated pir.
21:10 Tene I don't see a :load in rakudo's generated pir, though, so that can't be the problem.
21:16 megamic joined #parrot
21:27 tetragon joined #parrot
21:39 tetragon joined #parrot
22:02 tetragon joined #parrot
22:24 dalek r30693 | tene++ | trunk:
22:24 dalek : [cardinal]
22:24 dalek : * few more tests
22:24 dalek : * Boolean class
22:24 dalek : dmknopp++
22:24 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30693
22:30 TiMBuS joined #parrot
22:35 Zaba_ joined #parrot
22:54 Whiteknight joined #parrot
22:55 AndyA joined #parrot
22:57 gmansi joined #parrot
23:17 cognominal It seems that parrot has a major competitor : http://blogoscoped.com/google-chrome/
23:17 cognominal it is based on webkit for the rendering and v8 a virtual machine.
23:18 Whiteknight joined #parrot
23:22 cognominal http://blogoscoped.com/archive/2008-09-01-n47.html
23:29 Whiteknight You'd think they could just release an extension to firefox
23:31 cognominal webkit is acid while firefox is not.
23:32 cognominal but what worries me is that they use v8 that has the same agenda as parrot, running many languages.
23:32 diakopter reading the Chrome comic book.. IE8 already has per-tab processes.
23:35 Ademan joined #parrot
23:40 diakopter cognominal: how is Chrome a competitor to parrot?
23:40 cognominal because they use their own vm
23:40 cognominal open source
23:40 purl well, open source is another term for "Free Software" that doesn't scare the suits, or see http://www.opensource.org or just another way to say "Buy my own beer!" or a club for dispossessed teens or not another term for "Free Software"
23:41 cognominal and they have an application on top of it. Parrot has not any.
23:41 TiMBuS maybe if parrot had its own comic
23:42 TiMBuS :D
23:43 diakopter cognominal: are you saying Chrome is a parrot competitor, or v8 is a parrot competitor?
23:43 cognominal both
23:44 diakopter Chrome seems like just another browser.
23:44 Whiteknight cognominal, where would one go to learn more about v8?
23:44 diakopter I don't see anything new, as it's described in the comic
23:44 Whiteknight it's wikipedia entry is empty
23:44 cognominal what is a vm without appliations that uses it.
23:45 cognominal Whiteknight,  google will release everything tomorrow
23:45 Whiteknight ok
23:45 cognominal it seems that the comic release happened one day too early.
23:45 cognominal If you follow links there are many 404.
23:48 cognominal ironically it seems that their webkit base chrome does not run yet on macosx.
23:48 cognominal s/base/based/
23:50 dalek r30694 | cotto++ | trunk:
23:50 dalek : [pmc] PIRified t/pmc/pmcproxy.t
23:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30694
23:54 diakopter cognominal: what makes you think v8 wants to run more than javascript/ecmascript?
23:55 cognominal may be I misread.
23:55 Whiteknight it doesnt matter if it does. If we work to be the best VM, it doesn't matter what competition we have
23:56 cognominal and if their vm api is similar enough we could plug our vm in chromes
23:56 Whiteknight I would love to get libparrot into a browser
23:57 Whiteknight imagine the cool scripting possibilities in webpages with that!
23:57 cognominal that's my wet dream
23:57 cognominal need a security model though
23:59 cognominal sure enough this is linked to android

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

Parrot | source cross referenced