Camelia, the Perl 6 bug

IRC log for #parrot, 2008-10-25

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 particle you may be able to modify it not to use rakudo components
00:06 allison The exception "Parameter type check failed" comes from Rakudo
00:07 ruoso I guess that the MMD in Rakudo is failing because it is not considering that the multi must check variants outside itself before dispatching
00:07 ruoso this is similar to the problem of lexically-declared sub variants
00:08 davidfetter 'sup, ruoso
00:09 ruoso davidfetter, boa noite ;)
00:10 AndyA joined #parrot
00:10 davidfetter 1710 aqui ;)
00:10 ruoso :P
00:13 ruoso allison, speaking of MMD... how do you plan to solve the issue of lexically-declared multi variants?
00:14 allison what problem? ticket #?
00:14 ruoso it's more of a conceptual problem
00:14 allison how so?
00:14 allison mmd subs are just subs
00:15 ruoso right... but then you have a new variant lexically declared in an inner scope...
00:16 allison aye, and inner scopes take priority over globals
00:16 ruoso but then the inner version needs to find the outer declarations...
00:17 allison the semantics of the interactions between lexicals and globals differs from one HLL to the next
00:17 ruoso right... so let's consider only lexicals
00:17 allison Parrot doesn't have lexical MMD at the core
00:18 allison only global MMD
00:18 ruoso hmm... ok...
00:18 allison so, if Rakudo does lexical MMD, then we need to know what semantics they're trying to support
00:18 allison there's a good chance they can do it without any core Parrot changes, since Rakudo has its own MultiSub PMC
00:19 ruoso in one line: { my multi foo (Int $a) {...}; { my multi foo (Str $a) {...}; <you can see foo(Str) here> }; <but not here> }
00:19 allison but, if they need some hooks added to core Parrot to make it simpler, we can certainly add them
00:20 allison so, Rakudo's MultiSub needs to check the lexical scope for MMD variants before checking the global namespace
00:21 ruoso ok... that's also how I plan to do...
00:21 ruoso it's good to know I'm choosing a sane path ;)
00:21 allison it's up to the HLL to decide if lexical MMD's should always have priority over global MMDs (so, a lexical MMD means the global is ignored), or should just be added as another alternate compared in the total collection of MMD variants
00:22 ruoso hmm... actually lexical does not hide global in Perl 6... but I think it doesn't make much difference...
00:23 allison either way works, it's just a slight change in the search order
00:54 TimToady joined #parrot
01:22 daxelrod joined #parrot
01:34 mberends joined #parrot
01:36 Coke down to 3330 lines of tcl to cause the segfault. :|
01:37 cotto I guess that's progress.
01:48 Coke bah.
01:53 Andy joined #parrot
02:01 Coke where do constant strings get created?
02:02 Coke hurm. const STRING * is probably not the same as a CONST_STRING, izzit.
02:04 Infinoid nope, CONST_STRING is a function that calls string_make_direct
02:08 Coke maybe I can use the trick from http://www.oreillynet.com/onlamp/blog/200​7/10/how_to_debug_a_gc_problem_in_p.html against string_make_direct instead of pmc_new
02:09 Coke but not now; if anyone's bored (ha), it's reproducable on feather with latest parrot and partcl, with "../../parrot tcl.pbc t_tcl/namespace-old.test" (hve to run make spectest first to get the .test file)
02:09 Coke ->
02:38 Andy joined #parrot
02:38 Psyche^ joined #parrot
03:01 Psyche^ joined #parrot
03:08 tetragon joined #parrot
03:25 rblackwe joined #parrot
04:42 chromatic joined #parrot
04:45 Debolaz joined #parrot
04:58 davidfetter joined #parrot
05:01 chromatic joined #parrot
06:28 Ademan joined #parrot
07:17 particle joined #parrot
07:36 johbar joined #parrot
08:21 barney joined #parrot
09:23 mj41 joined #parrot
09:26 iblechbot joined #parrot
10:17 chromatic joined #parrot
11:13 dalek r32152 | bernhard++ | trunk:
11:13 dalek : [Pipp] Add a TODO test for simple parameter passing.
11:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32152
11:28 dalek r32153 | bernhard++ | trunk:
11:28 dalek : RT #60108: [PATCH] cleanup chapter 3 of book
11:28 dalek : Courtesy of Michael Stephens.
11:28 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32153
12:09 dalek r32154 | bernhard++ | trunk:
12:09 dalek : [Pipp] Require whitespace after 'echo'.
12:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32154
12:39 dalek r32155 | bernhard++ | trunk:
12:39 dalek : [Pipp] Rephrase check for 'array'.
12:39 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32155
12:41 iblechbot joined #parrot
12:47 dalek r32156 | bernhard++ | trunk:
12:47 dalek : [docs] Note that development of Pirate seems to be stalled.
12:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32156
12:52 Bzek joined #parrot
13:31 Andy joined #parrot
14:27 rhr joined #parrot
14:33 moritz barney: I get an error when running 'make test' in pipp
14:34 moritz cd .. && /usr/bin/perl -I../lib -I pipp/lib pipp/t/harness --with-pct
14:34 moritz Not an ARRAY reference at /usr/local/share/perl/5.8.8/TAP/Harness.pm line 593.
14:34 moritz make: [test-pct] Error 255 (ignored)
14:34 omega joined #parrot
14:35 moritz barney: it seems that sometimes the perl from $PATH is used, and sometimes it uses the one that ran Configure.pl
14:38 Andy joined #parrot
14:50 barney moritz: looking into it
14:59 barney moritz: what's your version on TAP::Harness ?
15:03 barney TAP::Harness might look at the shebang line of the test scripts, when  'exec => undef'
15:07 Limbic_Region joined #parrot
15:09 moritz 3.14 or something
15:09 moritz sorry, no time... bbl
15:23 dalek r32157 | bernhard++ | trunk:
15:23 dalek : [Pipp] Explicitly specify the program for running the perl test scripts
15:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32157
15:24 Theory joined #parrot
15:26 jsut joined #parrot
15:50 dalek r32158 | tewk++ | gsoc_nci_001:
15:50 dalek : [gsoc_nci_001] Delete, its already been merged
15:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32158
16:25 jhorwitz joined #parrot
16:28 apeiron joined #parrot
17:10 particle joined #parrot
17:26 Psyche^ joined #parrot
17:41 dalek r32159 | kjs++ | trunk:
17:41 dalek : [pirc/new] integrate the macro layer in pirc/new. Mostly done, .macro_locals need work.
17:41 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=32159
17:56 kj joined #parrot
18:18 chromatic joined #parrot
18:38 grim_fandango joined #parrot
19:10 cognominal joined #parrot
19:43 particle joined #parrot
20:21 Andy joined #parrot
21:03 particle joined #parrot
21:16 rdice joined #parrot
21:23 Andy joined #parrot
22:07 Theory joined #parrot
22:11 TiMBuS joined #parrot
23:52 davidfetter joined #parrot

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

Parrot | source cross referenced