Camelia, the Perl 6 bug

IRC log for #parrot, 2008-10-08

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:08 rblackwe joined #parrot
00:09 AndyA joined #parrot
00:27 dalek r31767 | infinoid++ | trunk:
00:27 dalek : [codingstd] Fix trailing whitespace in src/charset/unicode.c
00:27 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31767
00:28 Infinoid anyone else seeing this segfault?
00:28 Infinoid t/pmc/complex....47/54
00:28 Infinoid #   Failed test 'add using subclass of Complex (RT \#59630)'
00:28 Infinoid #   at t/pmc/complex.t line 1871.
00:28 Infinoid # Exited with error code: [SIGNAL 11]
00:28 Infinoid (I'm wondering if its linux/x86-64 specific.)
00:35 nopaste "Infinoid" at 75.28.79.48 pasted "backtrace of t/pmc/complex.t segfault on linux/x86-64" (47 lines) at http://nopaste.snit.ch/14249
00:48 * Infinoid posts that backtrace to the ticket.
00:50 cotto Infinoid, it's broken for me too
00:50 pmichaud Infinoid: yes, I know about the error.
00:51 pmichaud it's a known error -- we're waiting for someone to todo it or fix it.
00:54 bennett joined #parrot
00:55 allison joined #parrot
00:55 Infinoid ok, great
01:04 dalek r31768 | infinoid++ | trunk:
01:04 dalek : [include] Add include/parrot/platform_limits.h (autogenerated) to svn:ignore.
01:04 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31768
01:12 silug joined #parrot
01:15 dalek r31769 | dmknopp++ | trunk:
01:15 dalek : [cardinal]: a few Range tests
01:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31769
01:17 dalek r31770 | dmknopp++ | trunk:
01:17 dalek : [cardinal]: tests for the Math class
01:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31770
01:18 dalek r31771 | dmknopp++ | trunk:
01:18 dalek : [cardinal]: tests for the Kernel class
01:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31771
01:18 dalek r31772 | dmknopp++ | trunk:
01:18 dalek : [cardinal]: tests for the Integer class
01:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31772
01:23 dalek r31773 | dmknopp++ | trunk:
01:23 dalek : [cardinal]: adding a few tests for the Time and GC classes
01:23 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31773
01:30 dalek r31774 | dmknopp++ | trunk:
01:30 dalek : [cardinal] Adding initial support for Range, Time, Math, GC, and Kernel classes. Much more to come :-)
01:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31774
01:32 Whiteknight dmknopp++
01:32 Whiteknight lots of commits, I like that!
01:33 dmknopp :-)
02:00 dmknopp left #parrot
02:15 Debolaz joined #parrot
02:30 dalek r31775 | coke++ | extract_tcl:
02:30 dalek : Update NEWS for partcl. (last entry)
02:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31775
02:52 ab5tract joined #parrot
03:02 Coke anyone fixing cardinal smoke errors?
03:06 dalek r31776 | coke++ | trunk:
03:06 dalek : [tcl] remove partcl from the repository, see:
03:06 dalek : http://partcl.blogspot.com​/2008/10/leaving-nest.html
03:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31776
03:09 dalek r31777 | coke++ | trunk:
03:09 dalek : [codingstd] Fix file metadata for cardinal
03:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31777
03:12 ab5tract Coke: ii wrote some tests
03:14 ab5tract if my patch made it in
03:14 dalek r31778 | coke++ | trunk:
03:14 dalek : [codingstd] remove trailing spaces.
03:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31778
03:15 dalek r31779 | coke++ | trunk:
03:15 dalek : [smolder] TODO failing test
03:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31779
03:19 Ontolog joined #parrot
03:19 ab5tract doesnt look like it did
03:28 Andy joined #parrot
03:30 Theory joined #parrot
03:35 dalek r31780 | coke++ | extract_tcl:
03:35 dalek : This branch has been merged to trunk and is no longer needed.
03:35 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31780
03:40 ab5tract * U    languages/dotnet/t/harness
03:40 ab5tract sweet :)
03:40 mberends joined #parrot
03:41 * ab5tract didn't know there was a dotnet in the codebase
03:44 tetragon joined #parrot
03:54 Coke lots of things in dusty nooks and crannies.
03:57 ab5tract i'm interested to see what happens once PCT stabilizes/extends
03:58 Theory joined #parrot
03:59 Coke Tene: poke
04:02 Coke NotFound: you sad that RT#59704 was a result of one of your commits?
04:03 Coke s/sad/said/
04:47 Bzek joined #parrot
04:47 Maddingu1 joined #parrot
05:02 Tene Coke: lemme look it up
05:03 purl joined #parrot
05:04 diakopter joined #parrot
05:08 Theory joined #parrot
05:09 Tene Coke: I'm pretty sure I never said anything of the sort.
05:09 Tene Coke: *OH*!  You addressed that to NotFound.
05:10 Tene Yeah... reading comprehension is a valuable skill.
05:16 ab5tract Tene: did you do have that cardinal hackathon?
05:19 Coke tene: was that your cardinal mess i cleaned up? =-)
05:19 Coke also, do you want a partcl bit?
05:39 Tene Coke: file metadata?  No, it was dmknopp.  I'll tell him that he needs to do that the next time I see him online.
06:00 MariachiElf joined #parrot
06:18 uniejo joined #parrot
06:22 chromatic joined #parrot
06:42 cosimo joined #parrot
06:53 dalek r31781 | rurban++ | pdd30_install:
06:53 dalek : [ports] update from trunk: cygwin parrot-0.7.1-1 release
06:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31781
07:03 dalek r31782 | rurban++ | pdd30_install:
07:03 dalek : [cage] support DYNPMC_INCLUDE+DYNPMC_LINK, make msys independent of old cygwin, update to trunk
07:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31782
07:21 allison joined #parrot
07:33 iblechbot joined #parrot
07:38 dalek r31783 | fperrad++ | trunk:
07:38 dalek : [WMLScript] pdd27mmd
07:38 dalek : - second step : rewrite PMC
07:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31783
07:45 Zaba joined #parrot
08:16 dalek r31784 | rurban++ | trunk:
08:16 dalek : [cage] add myself as cygwin responsible
08:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31784
08:29 apeiron joined #parrot
09:03 tomyan joined #parrot
09:04 kj joined #parrot
09:10 Ontolog joined #parrot
09:11 apeiron joined #parrot
09:54 dalek r31785 | rurban++ | trunk:
09:54 dalek : [cage] add op docs
09:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31785
09:56 dalek r31786 | rurban++ | trunk:
09:56 dalek : [cage] RT#48260 add op docs. Fix typo.
09:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31786
10:24 Zaba_ joined #parrot
10:25 dalek r31787 | rurban++ | pdd30_install:
10:25 dalek : [docs] update book to trunk
10:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31787
10:32 bacek joined #parrot
10:58 bacek joined #parrot
11:14 dalek r31788 | fperrad++ | trunk:
11:14 dalek : [WMLScript]
11:14 dalek : - fix Float representation (like before r31402)
11:14 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31788
12:20 Zaba joined #parrot
12:26 flh joined #parrot
12:26 flh hi
12:26 purl que tal, flh.
12:28 flh I've started playing with parrot and pct (read the squaak tutorial...), and I have a few questions: where's the best place to ask them?
12:32 kj hi flh. that would be here
12:32 flh ok :)
12:32 kj or the mailing list, for a bigger audience
12:32 kj i'm afk every now and then, just so you know.
12:33 kj like now :-) but just ask your questions, and I'll see them if I get back.
12:33 flh let's try here first... some time ago, I read a post on a mailing-list saying that there can be only one optable in a grammar
12:34 flh is this still true?
12:35 kj I'm not sure if that's true, but I *think* it's true
12:35 kj (I can't imagine how that would work, with 2 or more such optables)
12:36 flh actually, i'm trying to implement (a subset of) the ocaml language, which has two different kinds of expressions ("usual" expressions, and type expressions)
12:36 kj btw, the squaak tutorial is updated, and maintained in parrot/languages/squaak/doc
12:36 kj so you should check out that one instead of the one on parrotblog.org
12:36 flh thx, i saw it yesterday, and thanks to pod2latex i have wonderful printed version
12:37 flh so I'll try defining two optables and see what happens :)
12:37 kj If you have any comments or experiences that you'd like to share while using the PCT (as a newcomer), please send them to list
12:38 kj then we can do something about the documentation to prevent any problems in the future.
12:38 flh and is there a way to define operators in a more complicated way than just with their arity (much like what would be allowed by yacc)?
12:39 kj not sure what you mean... could you elaborate?
12:39 flh ocaml has a "let x = <expr> in <expr>" construction
12:39 kj ok, you would define that as a rule, i guess
12:39 flh the "let ... in" is an operator, with a precedence, etc.
12:39 kj yeah, but it has rich syntax, in the sense it's all keywords
12:40 kj I'd do it as a rule I think
12:40 kj rule let_statement { 'let' <ident> '=' <expr> 'in' <expr> }
12:40 flh so i have to switch back to top-down parsing, and define a rule for each precedence after the let...in
12:41 kj well, I don't know ocaml... so not sure what are the options that go in between 'let' and 'in'
12:41 kj I guess for advanced questions like that, you might want to ask pmichaud, when he wakes up :-)
12:42 flh your rule let_statement is (almost) the good one
12:43 flh there could have been a clever trick (like defining, by hand, a PIR sub which would behave like a "proto" of the grammar)
12:44 kj I have to admit that my knowledge here is limited on these advanced tricks :-)
12:45 flh i knew i should have chosen another language than caml for my first compiler :)
12:45 kj that would be helpful yes,to get started.
12:45 kj but what you can do, is implement a subset
12:45 kj just the easy bits
12:45 kj and then later extend it
12:46 flh that's actually what i am doing, but "let...in" for example is really the core of the language
12:47 flh a similar case happens with "if..then..else.." which is also an operator, and i would be happy if the optable could understand it
12:48 flh btw, does the optable syntax come from perl6 regexes, or was it made specifically for parrot?
12:48 kj I think a lot of it is based on perl6
12:49 flh (my idea: would it be easy to suggest some improvements? :) )
12:50 kj ideas are always welcome ;-) Please send them to the list
12:52 flh ok :) thx for your answers, i'll try to take some time to send a message to the list (it's parrot-dev@lists.parrot.org, isn't it?)
12:52 kj I think it has changed quite recently; sorry, don't remember which one it is now...
12:53 kj the website is being moved to a different domain; parrot.org
12:53 Lorn joined #parrot
12:55 Hinrik joined #parrot
13:19 Coke flh: yes, that's the new list
13:19 Coke pperl6-internals?
13:19 Coke perl6-internals?
13:19 purl rumour has it perl6-internals is quite slow
13:19 Coke no, perl6-internals is now parrot-dev@lists.parrot.org
13:19 purl okay, Coke.
13:19 Coke parrot-porters?
13:19 purl parrot-porters is now superceded by mailto:parrot-dev@lists.parrot.org
13:19 Coke p2?
13:19 purl i think p2 is mailto:parrot-porters@perl.org or known as mailto:perl6-internals@perl.org
13:20 Coke no, p2 is <reply>See parrot-porters
13:20 purl okay, Coke.
13:21 iblechbot joined #parrot
13:33 particle commitbit me, coke!
13:40 gryphon joined #parrot
13:41 masak joined #parrot
13:45 Coke for partcl?
13:46 Coke (done)
13:53 * Coke does like that the wiki is also editable via svn and a text editor.
13:55 Coke ls
13:57 * Coke pokes particle to join the mailing lists if he's got commit bits.
13:59 particle done
14:00 Coke particle: I'm assuming any commits here are done under the same terms as your CLA with TPF.
14:00 particle ok by me. you should state that somewhere in your copious docs :)
14:01 Coke ... Guess what your first job is.
14:10 particle coke: you should have a credits or authors file
14:11 particle do you want a manifest, too?
14:11 particle hopefully not
14:11 Coke (manifest) hell no.
14:12 particle :)
14:12 Coke Credits is fine. Feel free to start off with whatever there is for tcl in parrot's CREDITS.
14:13 pmichaud hll mapping _really_ makes class handling complicated :-|
14:14 Coke for anyone looking to move their language out of the repo, turns out I just needed a very small Configure.pl, btw.
14:15 Coke (you can use reconfigure.pl to generate your makefile even if you're not one of the core languages, it seems.)
14:16 particle sweet
14:19 PerlJam Is this the start of a mass exodus?  :)
14:20 dalek r31789 | moritz++ | trunk:
14:20 PerlJam make it easy to leave and they'll all go.
14:20 dalek : [cage] clean POD in cardinal's globals.pir
14:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31789
14:22 Coke tcl is a trendsetter.
14:23 Coke I hope some of the big languages stay in core for a bit longer.
14:23 Coke makes it easier on core developers, I think, to test things.
14:24 PerlJam Coke: excellent ... now all you have to do is release your language into the wild so that it's  "real" and then when the others follow suit, parrot will be "real"   :-)
14:25 moritz parrot will never be "real" without working lexicals :/
14:25 moritz but I guess I'm just spreading FUD
14:25 pmichaud tcl doesn't use Parrot's lexicals :-)
14:29 PerlJam moritz: it's an odd statement that something on the order of a CPU or assembly language should have lexicals.
14:30 moritz PerlJam: parrot is much more high-level then a mere CPU emulator
14:30 PerlJam But I guess, if you're used to CISC architectures, you probably want more than the average VM anyway  :)
14:31 pmichaud given that parrot has hashes, arrays, methods, classes, coroutines, etc,  lexicals isn't that odd.
14:34 PerlJam yeah, the really odd thing is that it's a VM  from that perspective :)
14:35 PerlJam pm: so ... how does the hll mapping make class handling complicated?
14:36 pmichaud in lots of ways, but here's a ticket... :-)
14:36 pmichaud oh, I'll do a nopaste first.
14:37 nopaste "pmichaud" at 76.183.97.54 pasted "hlls can't have duplicate class names" (37 lines) at http://nopaste.snit.ch/14252
14:38 pmichaud in addition to that, if I have a string (e.g. "Foo::ABC" in Perl 6) and I want to get its corresponding class, there's not an easy way to do that.
14:38 bennett joined #parrot
14:39 Coke pmichaud: why are you doing get_root_namespace and not get_hll_namespace?
14:40 pmichaud Coke:  I'm taking the position of a library (i.e,. PCT) that needs to be able to create classes in other hll namespaces
14:40 pmichaud but even if I did get_hll_namespace I'd end up with the same issue.
14:40 * pmichaud writes a demo.
14:41 Coke hurm. I regularly get things back and forth between 'tcl' and '_tcl' hlls.
14:41 pmichaud with identical class names?
14:41 Coke that, probably not.
14:42 nopaste "pmichaud" at 76.183.97.54 pasted "hlls can't have duplicate class names, using get_hll_namespace" (42 lines) at http://nopaste.snit.ch/14253
14:44 Coke Probably helps me that I don't do user facing classes, either. =-)
14:45 NotFound BTW I'm wondering, supposing a module written in some HLL that wants to publish a language agnostic module, where must it put the subs and calss of his interface? ['parrot';'module_name] ?
14:45 NotFound a language agmostic interface, I mean
14:46 pmichaud it would probably still live in the HLL namespace (so it can access builtin functions), but then export to other HLL namespaces.  That module would also need to be HLL aware if it's anything but method-based.
14:47 pmichaud (by HLL aware I mean "able to know how to access things in other HLL namespaces", probably through those HLL's compiler objects.)
14:47 NotFound Yes, by language agnostic I mean that the caller does not need know nothing about the language the module is written and does not need to use his types to using it.
14:48 Coke holy crap.
14:48 purl holy crap is, like, only in the Vatican
14:48 Coke particle++
14:49 * Coke chuckles.
14:49 Coke particle: you named the page, PageName.
14:49 pmichaud and a darn good pagename PageName is for naming a page.
14:49 Coke I don't think you can rename it via the web, but you can probably do an svn rename.
14:50 Coke particle: https://partcl.googlecode.com/svn/wiki
14:51 particle er, that's default, and i had no idea i had to change that
14:51 pmichaud I had originally stuck pynie stuff (issue tracking) into launchpad, should we see about switching over to googlecode instead?
14:52 Coke pmichaud: I don't think it matters.
14:52 particle fixed
14:53 pmichaud RT #59720 is the no-duplicate-classes bug
14:53 Coke particle++
14:59 Andy joined #parrot
15:01 particle credits file added
15:06 Coke particle++
15:07 masak rakudo: my $a = "hello"; my $x=0; say $a.trans("l" => {++$x})
15:07 polyglotbot OUTPUT[he12o␤]
15:08 pmichaud that is totally scary.
15:08 masak rakudo++
15:08 Coke tcl: puts "hi"
15:08 polyglotbot OUTPUT[Parrot VM: Can't stat languages/tcl/tcl.pbc, code 2.␤main: Packfile loading failed␤]
15:08 * Coke cries a bit. =-)
15:08 moritz pmichaud: why scary? ;)
15:08 moritz cjfields++
15:09 pmichaud it's scary that it even works.
15:09 masak aye
15:09 Coke scary++
15:09 moritz come on, it's *tested*
15:09 pmichaud I'm happy that it works, but it's still like "wow, I can't believe rakudo just did that"
15:09 masak moritz++
15:09 masak pmichaud: I know the feeling :)
15:09 moritz but I don't know if the combination of regex + closure works
15:09 moritz rakudo: my $a = "hello"; my $x=0; say $a.trans(/l/ => {++$x})
15:09 polyglotbot OUTPUT[he12o␤]
15:10 moritz rakudo: my $a = "hello"; my $x=0; say $a.trans(/l/ => { $_ x 3 })
15:10 polyglotbot OUTPUT[hellllllo␤]
15:10 masak ooh
15:10 pmichaud rakudo:  my $a = "hello";  my $x = 0;  say $a.trans(/./ => {++$x})
15:10 polyglotbot OUTPUT[12345␤]
15:10 pmichaud rakudo:  my $a = "hello world";  my $x = 0;  say $a.trans(/./ => {++$x})
15:10 polyglotbot OUTPUT[1234567891011␤]
15:11 masak rakudo: my $a = "hello"; my $x=0; say $a.trans(/(l)/ => { $0 })
15:11 polyglotbot OUTPUT[Null PMC access in get_pmc_keyed_int()␤current instr.: '_block22' pc 407 (EVAL_12:157)␤called from Sub 'parrot;Any;trans' pc 11904 (src/gen_builtins.pir:7368)␤called from Sub '_block11' pc 79 (EVAL_12:28)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 801
15:11 polyglotbot ..(src/PCT/HLLCompiler.pir:480)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles'...
15:11 moritz rakudo: my $a = "hello"; my $x=0; say $a.trans(/(l)/ => { $_[0] x 2 });
15:11 polyglotbot OUTPUT[hellllo␤]
15:11 masak haha -- I broke it first! :)
15:11 pmichaud masak++   # expert rakudo breaker
15:11 moritz masak: trans doesn't set $/, but passes it as an argument to the block
15:11 moritz masak: I don't know if that's a feature or a bug
15:12 * masak considers every Null PMC access a bug
15:12 moritz rakudo: say $0
15:12 polyglotbot OUTPUT[get_pmc_keyed() not implemented in class 'Undef'␤current instr.: '_block11' pc 21 (EVAL_12:15)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 801 (src/PCT/HLLCompiler.pir:480)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1075 (src/PCT/HLLCompiler.pir:609)␤called from Sub
15:12 polyglotbot ..'parrot;PCT::HLLCompiler;command_line' pc 1254 (src/PCT...
15:12 moritz masak: there's a ticket for that already
15:12 masak aye. just realized.
15:12 masak it segfaults over here, too :/
15:13 moritz currently allison++ is the expert rakudo breaker ;)
15:13 masak really? I have to work harder, then.
15:14 moritz masak: well, the MMD merge broken spectest_regression
15:14 pmichaud okay, I should rephrase
15:14 pmichaud allison breaks rakudo, masak finds where it's broken.
15:14 masak :)
15:15 moritz jonathan fixes masak's tickets, and I write tests for them, and pmichaud makes snarky comments - aren't we a happy family? ;-)
15:16 pmichaud I fix jonathan's fixes
15:16 moritz right
15:16 moritz ... and everything else ;)
15:16 pmichaud I need lunch, bbiah
15:19 NotFound A null PMC access may not be a bug, just the result of "I don't care at what exact point this function throws, just need to know if it throws or not"
15:19 moritz NotFound: if it's exposed to the Perl 6 programmer, it's a bug
15:20 NotFound moritz: that was the second part I was written: if it never gets catched, almost sure is a bug.
15:23 jhorwitz joined #parrot
15:35 johbar joined #parrot
15:46 Theory joined #parrot
15:48 particle masak: can you name a few killer features of rakudo that have made november-wiki development fun? i have a rakudo workshop tonight, and it would be nice to have fun examples to show
15:49 moritz I'm not masak, but there's one I can name - grammars/regexes
15:49 zostay joined #parrot
15:49 moritz I did the initial HTML::TEmplate grammar, and that was *real* fun
15:50 masak particle: aye, grammars
15:51 masak particle: also, a general feeling of joy when realizing that rakudo actually implements Perl 6 -- the language we've been reading about for all these years -- to a significant degree
15:52 masak particle: also, I was really happy when I realized that you can send a closure as a parameter: http://use.perl.org/~masak/journal/37529
15:53 grim_fandango joined #parrot
15:53 masak encapsulating a method and sending it to a method in another class in order to minimize coupling... doing that in rakudo felt like science fiction
15:54 Zaba joined #parrot
15:55 NotFound masak: A method name or a method sub?
15:56 masak NotFound: not sure I understand the question. I sent a closure containing a method call sending along an implicit parameter. see blog post.
15:57 NotFound masak: ah, sorry, I understood 'science fiction' as 'can do that yet'
15:57 NotFound can't
15:58 masak NotFound: sorry, my view of what counts as science fiction has been skewed since about 2002. nowadays I'm deeply convinced that we already live in the future
15:59 NotFound Certainly, seeing '2002' as a great example of sci-fi in 2008 change points of view :D
15:59 masak aye
16:00 NotFound masak: I also think that I live in the future when I see at work people having problems with things that have simple solutions from the '90
16:00 masak I'm psychologically stuck at kilobytes for RAM and megabytes for hard drives. all higher prefixes tend to strike me with awe.
16:01 masak NotFound: the future is already here. it's just not evenly distributed.
16:01 masak I can't remember who said that
16:01 * moritz should really buy a 1TB hard disk at some point
16:01 * masak 's eyes open wide
16:02 masak TB? I think I saw that unit in a game from the 90's once, as an in-joke
16:02 masak they're here now?
16:04 NotFound masak: maybe that is the reason for python to have the import from future thing
16:05 masak hardly needed nowadays, I'd say
16:06 NotFound masak: very nice metaphor, the bathroom as callback :D
16:06 masak there's another Perl 6 feature suggestion, btw: allowing `use future:Some::Module`, which pulls a bugfree future version of the module through a wormhole, and loads it
16:07 TimToady quantum foam is not your friend here
16:07 NotFound masak: Will that feature require a quantum computer?
16:07 masak NotFound: thank you. :) that's the first feedback I get on that post. I wasn't sure whether it went above everyone's heads or was just completely inane...
16:08 NotFound masak: maybe few people read "The number of the Beast" on this times of neo conservadurism
16:08 masak NotFound: maybe, but it could perhaps be simulated on a Turing machine, with the caveat that it's provably faster to just wait :)
16:08 Tene I got no feedback at all on my parrotblog post.  I didn't really expect to, though.
16:09 TimToady the problem is that you can't specify "bug-free" through a wormhole.  all you can do is check the result after you get it back, and see if you've been shrödingerized
16:09 NotFound (If they have no idea of the content, just for the title)
16:09 masak NotFound: I can't claim to see the political side of that book even though I'm aware that Heinlein is generally explicitly political
16:09 * Infinoid wants a corresponding patch to perldoc, so he can read the future documentation
16:10 NotFound masak: I was alluding to sexual morality in the novel, not political
16:10 masak TimToady: :)
16:10 Infinoid if I find a documentation bug in future software, am I allowed to put a patch in a time-release capsule?
16:10 masak NotFound: sure, there's that. I don't even think about that anymore when reading his books. I just find it refreshing.
16:10 masak NotFound: then again, I'm from Europe. :P
16:11 NotFound Infinoid: we already do that, a lot of times I read things in the docs that are no yet implmented.
16:11 Infinoid parrot is better than most, in that regard.  usually I'm stuck reading docs that were written for 10 API revisions ago
16:11 masak Infinoid: I'd suggest fixing the bug at once and deal with the causality paradoxes as they come.
16:11 NotFound masak: I'm from Spain, have heard this place is in Europe ;)
16:12 masak NotFound: allegedly. but I'm from Sweden, the land of liberty. let's see you trump that. :)
16:13 NotFound masak: but you and me have readed the book, then excluded to the group of people alluded.
16:13 masak mm
16:15 Tene TimToady: obviously the solution is to just destroy the universe when you get the wrong answer.  Then you can ensure you end up in the right everett branch.
16:16 masak Tene: yes, that's it. but maybe not in 6.0.0 :P
16:16 NotFound I wonder if that book have some translation able to pass censure in the Franco dictartoship times, BTW X-)
16:17 pyrimidine joined #parrot
16:18 pyrimidine moritz: the .trans issue with $/ being passed is a bug (I saw it too).
16:18 pyrimidine .trans does pass it in to a Closure if it's present, though, so it should work.
16:19 NotFound "The global type registry will go away eventually" I suppose that means that RT#59720 is eventually not a bug
16:23 NotFound Dropping the call to fail_if_type_exists (the function that have that comment) te example works.
16:27 NotFound I don't see anything related to global type registry or old object metamodel in DEPRECATED.pod
16:28 NotFound pmichaud: ping
16:38 sjansen joined #parrot
16:40 rurban joined #parrot
16:41 pmichaud NotFound: pong
16:41 Tene pmichaud: check out src/pmc/class.pmc:194
16:42 Tene Oh, wait, nevermind.
16:42 pmichaud use future:Perl6;
16:42 pmichaud oh, wait.
16:43 Tene Wait, yeah, that might be it.  Lemme look up Parrot_oo_register_type...
16:44 NotFound pmichaud: I was taking a look at RT#59720
16:46 NotFound Tene: yes, that function is the place I was looking
16:46 Tene pmichaud: Yeah, if you comment out class.pmc:194 that test case passes.  You'll probably need to update some other usages in the tree.
16:47 pmichaud Tene: it's probably worse than that, as it changes the whole meaning of some of the opcodes.
16:47 NotFound Maybe is time to completely drop oo.c
16:48 Tene pmichaud: the opcodes could become .hll-aware?
16:48 pmichaud I'm not expecting a solution to RT #59720 anytime soon, I just wanted to make sure the ticket got posted
16:48 pmichaud Tene:  making the opcodes .hll-aware I think breaks a lot of existing code.
16:48 Tene Hmm.
16:49 pmichaud for example, if I'm in the 'perl6' HLL and my code has   $P0 = new 'Integer', do I get Parrot's Integer PMC or some other PMC?
16:51 Tene Should there be a difference between new 'Integer' and new ['Integer']?
16:51 pmichaud possibly.
16:51 Tene Also, if you defined a perl6 Integer , isn't that likely to be what you want it to do?
16:52 pmichaud I would think so, yes.  But then how do I get a Parrot Integer ?
16:52 pmichaud ("you don't need a Parrot Integer" is probably an okay answer for now.)
16:52 Tene You'd need the ['perl6';'Integer'] namespace, I'd think.
16:52 pmichaud right, so it'd have to be
16:53 pmichaud $P0 = get_root_namespace ['parrot';'Integer'] ;   $P1 = get_class $P0;  $P2 = new $P1
16:53 pmichaud which is okay with me
16:54 dalek r31790 | pmichaud++ | hllmagic:
16:54 dalek : [p6object]:  Fix 'Foo::Bar' lookups in get_parrotclass.
16:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31790
16:55 NotFound pmichaud: or you can store the classes  frequenlty uses somewhere in the HLL namespace.
16:55 pmichaud NotFound: sure, that's an option also.
16:55 pmichaud anyway, it's a design decision for allison.
16:55 pmichaud until then we do workarounds.
16:55 Tene pmichaud: It might be nice to support new ['parrot';'Integer']
16:55 pmichaud but then how do I distinguish between non-HLL and HLL versions?
16:56 pmichaud i.e., what if I have a parrot::Integer class in the perl6 HLL?
16:56 Tene ['perl6';'parrot';'Integer'] is what I'd lean towards.
16:56 NotFound Dropping the fail_type_exists call breaks PGE building.
16:57 pmichaud right, but how does Parrot know if  ['parrot';'Integer'] refers to the 'parrot' HLL namespace or a namespace within the current HLL namespace?
16:57 pmichaud or, if you're saying that C<new> should always require the HLL namespace, that means lots of code to update.
16:58 Tene Hmm.
16:59 pmichaud as I said, I don't expect it to be solved anytime soon.  RT #43419 has been around for 15 months, and I identified it as a problem over two years ago.
16:59 particle we could have a pragma, but i don't like that
16:59 pmichaud (when Chip and Leo were trying to get me to do things with .HLL)
17:00 Tene Perhaps we should harass allison about making a decision here.
17:00 Tene I'll do all the heavy lifting in updating old code if needed.
17:00 pmichaud yes, if we at least had a design for how this would be resolved, that would help.
17:00 TimToady joined #parrot
17:00 pmichaud because we could at least know how to make the workarounds.
17:04 Tene Yeah, I lean toward requiring getnamespace/getclass for getting from a different .hll than current.
17:06 Tene Well, no, I guess it depends on whether we optimize for writing by hand or for generation/execution.
17:07 pmichaud I think it's also a bit deeper question that it might appear at first.
17:07 pmichaud Last time I looked at the OO code, it seemed to me that a lot of "isa" and "does" checks were basically doing string compares.
17:07 Tene if we always require the hll be specified, that's more consistent
17:07 pmichaud i.e., checking to see if the class name matched.
17:07 pmichaud that's also terribly broken in a multi-HLL environment.
17:08 pmichaud (even if using the version of C<isa> that accepts a class as an argument, internally Parrot was converting it to a string comparison.)
17:11 particle joined #parrot
17:11 coke joined #parrot
17:12 * Tene drives to work.
17:15 dalek r31791 | particle++ | trunk:
17:15 dalek : [t] fix config pmc test to deal with case-insensitive platform
17:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31791
17:16 ruoso joined #parrot
17:20 pyrimidine perl6: my $i = 0;say 'hello'.trans(/l/ => {ord('A')})
17:20 polyglotbot OUTPUT[he6565o␤]
17:20 pyrimidine my $i = 0;say 'hello'.trans(/l/ => {ord($0)})
17:20 pyrimidine perl6: my $i = 0;say 'hello'.trans(/l/ => {ord($0)})
17:20 polyglotbot OUTPUT[Null PMC access in get_pmc_keyed_int()␤current instr.: '_block20' pc 302 (EVAL_12:114)␤called from Sub 'parrot;Any;trans' pc 11904 (src/gen_builtins.pir:7368)␤called from Sub '_block11' pc 58 (EVAL_12:21)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 801
17:20 polyglotbot ..(src/PCT/HLLCompiler.pir:480)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles'...
17:20 pmichaud $0 doesn't quite work there.
17:20 pyrimidine nope
17:21 pmichaud I'm not even sure if it should, or if it does, then how it does.
17:21 moritz pmichaud: by setting the contextual $/ for the block before calling it?
17:21 pyrimidine The error is similar to one reported before that jonathan fixed.
17:21 * pyrimidine looking in RT
17:21 pmichaud moritz: so, trans should set the contextual $/ ?
17:21 particle rakudo: say 'hello'.trans(/./ => {$_.ord});
17:21 polyglotbot OUTPUT[10448525256␤]
17:22 moritz pmichaud: I just say that if it should work, it'd work that way
17:22 pmichaud right.
17:22 pmichaud I'm thinking not, since $0 has to refer to the $/ of the block that contains the .trans call, not the .trans method itself.
17:22 pyrimidine No, S05 states that the resulting Match object is passed into the closure
17:23 pmichaud pyrimidine: yes, but is it passed as $_ or $/ (or both?)
17:23 pyrimidine It doesn't mention setting $/
17:23 pyrimidine (passed as $_)
17:23 particle "passed as current topic" would be more clear
17:23 pmichaud if passed as $_, then that's not $0
17:23 pyrimidine Actually, the example is $0, so I'm wrong
17:24 pmichaud which example?
17:24 purl somebody said which example was that?
17:25 pyrimidine actually, the example I'm looking at is in t/spec/S05-transliteration/trans.t
17:25 pyrimidine so it may not be correct
17:25 pmichaud oh, that's a test.  The test could be wrong.
17:25 pyrimidine yes
17:26 pyrimidine so it could (should?) be $_
17:26 particle that's my best guess
17:26 pyrimidine right now (via PIR) it is passed to the closure directly, so that would be $_
17:27 pyrimidine perl6: rakudo: say 'hello'.trans(/./ => {ord($_)});
17:27 polyglotbot OUTPUT[Statement not terminated properly at line 1, near ": say 'hel"␤␤current instr.: 'parrot;PGE::Util;die' pc 119 (runtime/parrot/library/PGE/Util.pir:82)␤called from Sub 'parrot;Perl6::Grammar;eat_terminator' pc 26373 (src/gen_grammar.pir:3187)␤called from Sub
17:27 polyglotbot ..'parrot;Perl6::Grammar;statementlist' pc 25029 (src/gen_grammar.pir:2654)␤called f...
17:27 pmichaud so, $_[0] would work, but not $0
17:27 pyrimidine d'oh
17:27 pyrimidine yes
17:27 pyrimidine (to $_[0]
17:28 pyrimidine rakudo: say 'hello'.trans(/./ => {ord($_)});
17:28 polyglotbot OUTPUT[10448525256␤]
17:28 particle i guess rakudo doesn't understand labels yet :)
17:28 pyrimidine rakudo: say 'hello'.trans(/l/ => {ord($_)});
17:28 polyglotbot OUTPUT[he10848o␤]
17:28 particle rakudo: label: 1;
17:28 polyglotbot OUTPUT[Statement not terminated properly at line 1, near ": 1;"␤␤current instr.: 'parrot;PGE::Util;die' pc 119 (runtime/parrot/library/PGE/Util.pir:82)␤called from Sub 'parrot;Perl6::Grammar;eat_terminator' pc 26373 (src/gen_grammar.pir:3187)␤called from Sub 'parrot;Perl6::Grammar;statementlist' pc
17:28 polyglotbot ..25029 (src/gen_grammar.pir:2654)␤called from Su...
17:28 pyrimidine rakudo: say 'hello'.trans(/./ => {ord($_[0])});
17:28 polyglotbot OUTPUT[Use of uninitialized value␤Cannot get character of empty string␤current instr.: 'parrot;Any;ord' pc 12153 (src/gen_builtins.pir:7503)␤called from Sub '_block18' pc 295 (EVAL_11:111)␤called from Sub 'parrot;Any;trans' pc 11904 (src/gen_builtins.pir:7368)␤called from Sub '_block11' pc 40
17:28 polyglotbot ..(EVAL_11:16)␤called from Sub 'parrot;PCT::HLLCompiler...
17:28 pmichaud well, there is no $_[0] there, since no captures.
17:28 moritz pyrimidine: you need a capture in your regex
17:29 particle ()
17:29 pyrimidine Ah
17:29 pyrimidine rakudo: say 'hello'.trans(/(l)/ => {ord($_[0])});
17:29 polyglotbot OUTPUT[he10848o␤]
17:29 pyrimidine bingo
17:30 * pyrimidine mildly surprised purl didn't respond to that
17:30 dalek r31792 | pmichaud++ | hllmagic:
17:30 dalek : [pct]:  Update PCT::HLLCompiler to look in correct place for <TOP>.
17:30 dalek : nqp now compiles and passes its tests.
17:30 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31792
17:31 particle why that's 10848 i don't know
17:31 pmichaud I think it's not incrementing properly
17:31 pmichaud I think 'l' is 108, and the the '0' (from the '108') is being translated into '48'
17:31 particle rakudo: 'l'.trans( /(l)/ => {$_[0].ord} ).say
17:31 polyglotbot OUTPUT[108␤]
17:32 pmichaud rakudo: say 'hoops'.trans(/(o)/ => {ord($_[0])});
17:32 polyglotbot OUTPUT[h11149ps␤]
17:32 particle rakudo: 'll'.trans( /(l)/ => {$_[0].ord} ).say
17:32 polyglotbot OUTPUT[10848␤]
17:32 pmichaud 'o' is 111, '1' is 49.  :-)
17:32 particle ick
17:32 pmichaud so the internal pointer isn't being updated properly when the substitution is more than two chars.
17:33 pmichaud rakudo: say 'xyozz'.trans(/o/ => '123')
17:33 polyglotbot OUTPUT[xy1zz␤]
17:33 pmichaud rakudo: say 'xyozz'.trans(/o/ => {'123'})
17:33 polyglotbot OUTPUT[xy123zz␤]
17:33 pmichaud hmmm
17:33 pmichaud rakudo: say 'xyoozz'.trans(/o/ => {'123'})
17:33 polyglotbot OUTPUT[xy123123zz␤]
17:33 pmichaud okay, I don't know then.
17:34 particle wtf
17:34 pmichaud rakudo: say 'xyoozz'.trans(/o/ => {123})
17:34 polyglotbot OUTPUT[xy123123zz␤]
17:34 pmichaud rakudo: say 'xyoozz'.trans(/(o)/ => {123})
17:34 polyglotbot OUTPUT[xy123123zz␤]
17:35 particle rakudo: say 'xaoozz'.trans(/(o)/ => {123})
17:35 polyglotbot OUTPUT[xa123123zz␤]
17:36 particle rakudo: say 'hello'.trans(/(o)/ => {123})
17:36 polyglotbot OUTPUT[hell123␤]
17:36 particle rakudo: say 'heool'.trans(/(o)/ => {123})
17:36 polyglotbot OUTPUT[he123123l␤]
17:36 pmichaud tene: as of r31792, nqp and perl6 pass their respective test suites in hllmagic.  (Haven't tried spectest_regression yet.)
17:36 particle rakudo: say 'hello'.trans(/(l)/ => {123})
17:36 polyglotbot OUTPUT[he123123o␤]
17:36 * Coke is reminded why he left earlier today. =-)
17:37 mberends joined #parrot
17:37 particle something to do with the result from ord, then
17:37 moritz particle: mind to open ticket and add test?
17:38 particle i think i remember how to do that ;)
17:39 pyrimidine say ord('l')
17:40 pyrimidine perl6: say ord('l')
17:40 polyglotbot OUTPUT[108␤]
17:40 pyrimidine perl6: say ord('o')
17:40 polyglotbot OUTPUT[111␤]
17:50 Theory joined #parrot
17:51 pmichaud argggh, with spectest_regression failing, it's hard for me to figure out if my changes cause more or less failures.  :-(
17:53 Coke are they failing because of you, parrot, or teh spec tests?
17:53 pmichaud the mmd merge.
17:53 pmichaud primarily the Complex PMC bug
18:05 particle pmichaud: yes, i've had that problem all week
18:05 particle ...so i gave up.
18:05 pmichaud particle: which problem, the Complex PMC ?
18:05 particle the failing spectests
18:05 particle i was fighting an uphill battle with is export() anyway
18:06 pmichaud I'll ask Allison in today's design meeting about an eta for fixing that one, if it looks as though it won't be fixed soon I think we'll have to skip them in spectest_regression for a while.
18:06 particle i have that workshop tonight, i hate having failing tests for it
18:06 pmichaud particle: as soon as tene and I can get the hllmagic branch merged into trunk, I think we can fix Rakudo's namespaces.
18:06 pmichaud particle: well, we should have an answer right after the meeting.  :-)
18:07 pmichaud I also don't want to go into pghpw with failing rakudo tests.
18:07 pmichaud so, I'd rather regress than have failing tests.
18:08 particle yes, me too
18:11 particle can i help with hllmagic?
18:18 pmichaud sure, we just need to see what's working and what's broken now after the changes.
18:18 pmichaud I think we're almost ready to merge back to trunk.
18:20 Ademan joined #parrot
18:24 Tene pmichaud++
18:26 Theory joined #parrot
18:27 pmichaud for the time being I think I'm going to declare that string specifiers for class names aren't always guaranteed to work outside of the 'parrot' HLL namespace.
18:27 pmichaud that the correct approach is to use a protoobject or possibly a namespace.
18:30 pmichaud afk for a bit
18:34 Tene Is anyone else running into problems with a compiler fialure in compilers/json?
18:35 particle in trunk?
18:35 pmichaud in hllmagic?
18:35 particle no
18:35 pmichaud Tene: you might need to svn up and rebuild tge
18:35 pmichaud or make clean in compilers/json, it may be using an old version
18:35 Tene In trunk.
18:35 particle i'm checking out hllmagic now
18:35 * Tene realcleans
18:37 Tene Yeah, it doesn't fial anymore after a realclean. :)
18:37 Tene fail
18:37 Tene I need to review dmknopp's cardinal stuff.
18:38 Coke and tell him about 'make codetest'. =-)
18:39 Wknight8111 joined #parrot
18:39 Tene cardinal doesn't compile in hllmagic... haven't checked in trunk.
18:40 moritz it did last time I tried
18:40 moritz (after dmknopp's commits)
18:40 Tene Okay, thanks.
18:41 Coke bah. need to fix a pge->past->pir issue in partcl. hate those. :|
18:42 Tene Hmm.  Maybe I should fork tcl back into parrot's tree. ;)
18:51 pmichaud oh, a bunch of languages/abc tests are failing too (in trunk)
18:52 Tene cardinal's make fails with something weird... I should try to track it down right now, though.
18:52 pmichaud ah, abc is still using the n_* opcodes.
18:56 Wknight8111 joined #parrot
19:00 dalek r31793 | pmichaud++ | trunk:
19:00 dalek : [abc]:  Update to avoid n_* infix opcodes.
19:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31793
19:02 pmichaud lots of tests failing in APL, too.
19:02 pmichaud oh, mostly having to do with precision issues.
19:02 pmichaud that can be fixed...
19:09 dalek r31794 | pmichaud++ | trunk:
19:09 dalek : [APL]:  Clean up tests to allow for more precision in output.
19:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31794
19:12 dalek r31795 | pmichaud++ | trunk:
19:12 dalek : [pynie]:  Update to avoid obsolete n_* opcodes.
19:12 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31795
19:12 Wknight8111_ joined #parrot
19:13 particle-split moritz: http://moritz.faui2k3.org/en/ilbot has a broken link for #perl6 irc log
19:13 dalek r31796 | Whiteknight++ | trunk:
19:13 dalek : [BCG] add some function-level documentation for /compilers/bcg/bcg.c
19:13 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31796
19:13 particle-split should be s/irc/irclog/
19:13 moritz particle-split: I'll take look, thanks
19:14 silug joined #parrot
19:16 moritz particle++ # fixed
19:16 dalek r31797 | pmichaud++ | trunk:
19:16 dalek : [pynie]:  Remove some obsolete n_* opcodes.
19:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31797
19:19 Wknight8111_ joined #parrot
19:20 * Coke rolls EPIC FAIL trying to understand parsing again.
19:21 Coke (the pge/past/pir parsing going on in tcl)
19:22 pmichaud Coke:  in trunk, or hllmagic?
19:22 Coke https://partcl.googlecode.com/svn/trunk/
19:24 Coke the special {*} syntax doesn't work when for "{*}{puts hi}". But trying to track down what's going on is just a maze of twisty passages for me.
19:26 * particle-split heads towards portland &
19:29 pmichaud I'm a bit distracted at the moment, I'd have to look at tcl stuff later
19:31 ruoso rakudo: multi foo (Int $a, Num $b) { say 1 }; multi foo (Num $a, Int $b) { say 2 }; foo(1,1);
19:31 polyglotbot OUTPUT[No applicable candidates found to dispatch to.␤current instr.: '_block11' pc 31 (EVAL_13:15)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 801 (src/PCT/HLLCompiler.pir:480)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1075 (src/PCT/HLLCompiler.pir:609)␤called from Sub
19:31 polyglotbot ..'parrot;PCT::HLLCompiler;command_line' pc 1254 (src/PCT/H...
19:32 dalek r31798 | Whiteknight++ | trunk:
19:32 dalek : [BCG] Added some brief function-level and file-level documentation to /compilers/bcg/src/bcg_emitter_pasm.c
19:32 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31798
19:32 moritz ruoso: rakudo doesn't know that 'Int' also conforms to the 'Num' type
19:32 pmichaud at the moment rakudo doesn't consider an Int to be a Num.
19:32 moritz there's a ticket for that...
19:32 ruoso hm...
19:32 ruoso so the other test I made was a false positive
19:32 rurban cygwin has now http://subversion.tigris.or​g/svn_1.5_releasenotes.html
19:33 moritz ruoso: aye
19:33 rurban (merge tracking, interactive conflict resolution)
19:34 ruoso rakudo: class A {}; class B is A {}; sub foo (A $a) { say 1 }; sub foo (B $a) {say 2 }; foo(A.new()); foo(B.new())
19:34 polyglotbot OUTPUT[1␤1␤]
19:34 ruoso rakudo: class A {}; class B is A {}; multi foo (A $a) { say 1 }; multi foo (B $a) {say 2 }; foo(A.new()); foo(B.new())
19:34 polyglotbot OUTPUT[1␤Ambiguous dispatch.␤current instr.: '_block11' pc 64 (EVAL_14:22)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 801 (src/PCT/HLLCompiler.pir:480)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1075 (src/PCT/HLLCompiler.pir:609)␤called from Sub
19:34 polyglotbot ..'parrot;PCT::HLLCompiler;command_line' pc 1254 (src/PCT/HLLCompiler.pir:698)␤calle...
19:35 ruoso a false positive indeed.
19:35 ruoso is there any idea on how to solve that? I know this issue is on the table for some time already...
19:36 * ruoso remembers that discussion in YAPC::EU::2007
19:36 pmichaud jonathan++ is the one handling multi dispatch in rakudo.
19:37 * Wknight8111 misses jonathan's weekly blog updates
19:38 pmichaud I need to get back to blogging again.  I just haven't found the motivation or energy.
19:38 * Wknight8111 misses pmichaud's blogs too
19:38 rurban_ joined #parrot
19:38 Wknight8111 I'm just an avid blog reader
19:45 Lorn joined #parrot
19:49 johbar joined #parrot
19:50 sjansen joined #parrot
19:50 sjansen http://dave.fumberger.com/2008/​10/08/cisco-networks-new-album/
19:50 sjansen oops
19:51 * Coke hurls http://partcl.blogspot.com/ for Wknight8111
19:52 Tene pmichaud: my post is still visible without scrolling on planet.parrotcode.org.  Therefore you need to blog more.
19:52 Wknight8111 Trust me coke, I'm an avid reader of yours
19:53 Wknight8111 And you haven't been posting often enough on cokefloats either!
19:53 * Coke will endeavor to blog more.
19:54 Coke especially now that I'm avoiding RPI's homegrown chat server.
19:54 Wknight8111 I probably write two or three blog posts per day, on average.
19:54 Wknight8111 of course, it's my own damn fault for participating in so many blogs!
19:54 Tene RPI?
19:54 purl i heard RPI was Rensselaer Polytechnic Institute in Troy, New York, USA. or RensselInstiPolyTechnitute or where Coke and japhy went to school. or and Josh
19:55 PerlJam purl: RPI is also repetitive perling injury
19:55 purl okay, PerlJam.
19:57 Andy joined #parrot
19:59 Wknight8111 heading home. later
20:36 jhorwitz hm, r31577 and r31758 cause any embedded app loading perl6.pbc to segfault
20:36 jhorwitz reverting those fixes my mod_perl6 segfaults
20:36 jhorwitz will open a ticket
20:38 moritz jhorwitz: isn't 31577 just a documentation fix?
20:38 jhorwitz er, sorry
20:38 jhorwitz fatfingered that
20:38 jhorwitz 31757 and 31758
20:38 purl 63515
20:38 peepsalot joined #parrot
20:46 pyrimidine joined #parrot
20:47 pyrimidine perl6: say 'ABC123DEF456GHI'.trans(/<alpha>/ => { say $_; my $x = $_.ord; chr(--$x)} , /\d/ => { say $_; my $x = $_.ord; chr(++$x) });
20:47 polyglotbot OUTPUT[A␤B␤C␤1␤2␤3␤D␤E␤F␤4␤​5␤6␤G␤H␤I␤@AB234CDE567FGH␤]
20:48 pyrimidine perl6" say 'ABC123DEF456GHI'.trans(/(<alpha>)/ => { say $_; my $x = $_[0].ord; chr(--$x)} , /(\d)/ => { say $_; my $x = $_[0].ord; chr(++$x) });
20:49 pyrimidine perl6: say 'ABC123DEF456GHI'.trans(/(<alpha>)/ => { say $_; my $x = $_[0].ord; chr(--$x)} , /(\d)/ => { say $_; my $x = $_[0].ord; chr(++$x) });
20:49 polyglotbot OUTPUT[A␤B␤C␤1␤2␤3␤D␤E␤F␤4␤​5␤6␤G␤H␤I␤@AB234CDE567FGH␤]
20:49 pyrimidine meh
20:49 pyrimidine perl6: say "hello".trans(/(l)/ => { $_[0].ord })
20:50 pmichaud that's not correct?
20:50 polyglotbot OUTPUT[he10848o␤]
20:50 pmichaud or you're trying to figure out why it fails?
20:50 pmichaud (in the "hello" case)
20:50 pyrimidine I'm trying to work out the fail
20:50 Tene Hmm.  Perhaps I should update polyglotbot to include a 'tene++' in its output every time.
20:50 Tene ... that's too much factoring, though.  Nevermin.d
20:50 Tene purl: karma
20:50 purl rumour has it karma is just as useful as the Perlmonks XP :-) or at http://ircxory.jrock.us/
20:50 Tene purl: karma tene
20:50 purl tene has karma of 259
20:51 pmichaud I'm going to go ahead and mark the failing spectest_regression tests with skip.
20:51 Tene 37*7, nice
20:51 pmichaud for particle's presentation tonight, and for pghpw later in the week.
20:52 pyrimidine It's definitely odd.
20:52 Tene Where's he presenting, and what's pghpw?
20:52 pmichaud pghpw?
20:52 pmichaud pghpw is the Pittsburgh Perl Workshop
20:53 pmichaud I think his presentation tonight is somewhere in Portland, OR
20:54 pmichaud pdx.pm, it looks like.   http://pdx.pm.org/kwiki/
20:54 Tene We need to get cardinal working again before we merge back into trunk.
20:54 Tene Oh, these are mmd failures
20:54 pmichaud if you decide to go, be sure to become a member first.
20:55 pmichaud I can't do much about the mmd failures, sorry.
20:55 Tene No, I meant that the failures you're marking spikked are mmd-related
20:55 pmichaud right.
20:55 pmichaud is cardinal not working in hllmagic?
20:55 Tene nope, doesn't even compile.
20:56 pmichaud works in trunk?
20:56 Tene so I'm told
20:56 pmichaud normal test is "make test"?
20:56 Tene yes
20:56 Tene but it fails on 'make'
20:56 Tene make clean first
20:57 pmichaud Method 'newtok' not found for invocant of class 'CardinalHash'
20:57 Tene yes
20:57 Tene look where that's happening
20:57 pmichaud is cardinal doing HLL type mapping?
20:57 Tene I think it sets up one for array, but I commented that out and it didn't fix anything
20:58 pmichaud okay, I'll see about fixing.
20:58 pmichaud I'm not sure why it's getting a 'CardinalHash' there, though.
20:58 Tene Me either, but I haven't had time to look.
20:58 Tene I'll have time after work, though.
21:01 donaldh joined #parrot
21:08 pmichaud oh, I see the problem.
21:08 Tene What is it?
21:08 purl it's it!
21:09 pmichaud cardinalmeta.'register'('Hash', 'parent'=>'CardinalObject', 'protoobject'=>mappingproto)
21:09 Tene Are you sure?  I deleted that and it didn't help.
21:09 pmichaud that means that all of CardinalObject's methods (including 'new') are being poked into the Hash class
21:09 pmichaud and since OPTable is a subclass of Hash, it's getting the wrong 'new' method.
21:10 peepsalot joined #parrot
21:10 pmichaud should probably not do the 'parent'=> option.
21:10 pmichaud you're correct, though, it doesn't seem to help.
21:11 Tene That *should* be removed, though.
21:14 NotFound I'm looking at the complex thing and seems like in the inherited case vtable methods GET_ATTR use the Complex attributes values but in the multis it uses setattribute on the child object.
21:19 Whiteknight joined #parrot
21:20 pmichaud okay, I think I see more of the problem.  Essentially we don't want PGE always using the protoobject for 'new'
21:20 pmichaud (for its own internal classes.)
21:20 pmichaud that will get us through things for now, at any rate.
21:20 pmichaud I'll fix the rakudo spectests, first though.
22:00 dmknopp joined #parrot
22:03 TiMBuS joined #parrot
22:09 dalek r31799 | pmichaud++ | hllmagic:
22:09 dalek : [pge]:  Some more cleanups for P6object and HLL usage in PGE.
22:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31799
22:09 dalek r31800 | pmichaud++ | hllmagic:
22:09 dalek : [cardinal]:  Some p6object/class name updates.
22:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31800
22:15 Lorn joined #parrot
22:16 raiph joined #parrot
22:23 ruoso joined #parrot
22:27 pyrimidine joined #parrot
23:15 TiMBuS joined #parrot
23:23 silug joined #parrot
23:27 dalek r31801 | julianalbo++ | trunk:
23:27 dalek : Complex pmc: get rid of RE and IM macros (unfinished)
23:27 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31801
23:29 Andy joined #parrot
23:30 peepsalot joined #parrot
23:34 dalek r31802 | chromatic++ | trunk:
23:34 dalek : [src] Avoided segfault in interpinfo_s() when nothing has set the name of the
23:34 dalek : executable embedding libparrot (see RT #59732, reported by Jeff Horwitz).
23:34 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31802
23:37 dalek r31803 | julianalbo++ | trunk:
23:37 dalek : Complex pmc: fix a big introduced in r31801
23:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31803
23:43 cotto is the CallSignature PMC complete?
23:44 cotto nm.  It's definitely not.
23:46 Andy Quinn is watching an episode of Pinky Dinky Doo, and the episode revolves around "cage cleaning day."
23:54 Limbic_Region joined #parrot
23:56 dalek r31804 | cotto++ | trunk:
23:56 dalek : [t] add stub tests for the CallSignature and CPointer PMCs
23:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31804

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

Parrot | source cross referenced