Camelia, the Perl 6 bug

IRC log for #parrot, 2008-05-18

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:01 AndyA joined #parrot
00:02 Tene Eevee: I was showering.  Did your patch get committed?
00:02 Eevee svn bot hasn't said anything so I assume not.  but give me a sec, fixing for empty list
00:02 Tene kk
00:05 Eevee jonathan: hm.  with get_hll_global 'Failure', "<>.reduce({ ... })" stringifies to 'Failure' whereas rakudo's current undef stringifies to '' (but .WHAT is still Failure)
00:05 Tene I'll check back in 5.
00:05 pmichaud correct, because Failure is the protoobject
00:06 pmichaud try $P0 = get_hll_global 'Failure';   $P1 = $P0.'new'()
00:06 pmichaud that creates a new Failure object from the proto
00:07 Eevee ah.  get_hll_global sounded like it grabbed a singleton or something.  okay
00:07 pmichaud well, it does grab a singleton... unfortunately it's a singleton protoobject.  :-)
00:07 pmichaud I don't know if we have a singleton Failure non-protoobject lying about.  Perhaps we should.
00:08 Eevee it made sense
00:08 pmichaud come to think of it, much of Rakudo is still using Undef so it probably wouldn't hurt to use Undef for now if you prefer that.
00:08 pmichaud or even just   new 'Failure'
00:08 pmichaud (I'm assuming 'Failure' is registered as a class, which it may not be.)
00:09 pmichaud so, feel free to just go with new 'Undef' -- and we can fix it later
00:09 Eevee good old "try everything until one works" approach
00:10 Eevee just trying to reduce the amount of fixing-it-later that will be necessary
00:10 pmichaud well, there's a fair amount of that no matter what we do at this point.  :-)
00:11 Eevee yeah, Failure's not a class.  will go with new 'Undef', and change the undef()s in here so at least they work
00:14 Eevee Tene: http://pastebin.com/d5df8ad8d  <>.reduce() and <>.pop() work now
00:14 particle japhb, pmichaud: look at the Exporter PMC, it was designed for hll exporting, has a flexible, well-defined api, and should be subclassable in c and pir
00:14 particle picnic dinner &
00:15 pmichaud particle: yes, but afaict Exporter doesn't actually solve the issues I brought up earlier.
00:15 pmichaud i.e., it handles the mechanics of exporting symbols, but not the conventions.
00:16 particle subclass it to handle the conventions
00:16 pmichaud no, that's not the issue.
00:16 pmichaud we want more than just export, we also want to be able to load modules.
00:16 pmichaud I don't think Exporter should do that.
00:16 pmichaud as loading a module is a very hll-specific item.
00:16 particle ah, ok. then i suggest a Loader PMC
00:17 pmichaud nah, we're just going to make a method on the HLLCompiler object
00:17 pmichaud it doesn't need to be in C
00:18 pmichaud hllcompiler.'load_module'('name', args)
00:18 Eevee Tene, pmichaud: wouldn't fixing pointy blocks only take this?  seems to work without breaking ones with param lists: http://pastebin.com/d3be0b70f
00:19 pmichaud Eevee: that's part of it.  Rakudo has an issue that it sometimes confuses  ->  with prefix:-
00:19 pmichaud i.e., it's also a parsing issue.
00:19 Eevee well, at least the problem with -> { foo($_) }
00:20 particle really gone &
00:20 pmichaud later, enjoy picnic!
00:20 Eevee $ ../../parrot perl6.pbc -e 'say <1 2 3>.grep(<-> { $_ % 2 })'
00:20 Eevee 13
00:20 * pmichaud has a soccer game in 40 mins.
00:22 * jonathan has bed in less than that
00:33 pmichaud and laptop is out of power.
00:33 pmichaud later, all.
00:34 Eevee seeya
01:03 Tene Eevee: looks like that's at least a big part of it.  I'll commit that and then look at reduce and pop
01:04 Eevee thanks
01:04 * jonathan sleeps
01:04 jonathan night all
01:04 Tene Man, typing is weird when my fingers are all sticky.  My girl had me helping her with some craft project and now my hands are covered with some weird adhesive.
01:07 davidfetter is it that spray adhesive?
01:08 Tene nah, some weird sticky wax.
01:08 Tene ... why am I compiling on the slow box?
01:09 * davidfetter finds this a strange juxtaposition, but probably because he thinks of ski- and other lubricating waxes
01:09 Eevee Tene: dramatic effect?
01:16 Tene Eevee: another simple patch would be adding an optional test param to List::uniq
01:18 Eevee ah, will have a look
01:20 dalek r27606 | tene++ | trunk:
01:20 dalek : [rakudo]
01:20 dalek : * Pointy blocks no longer set a signature when there are no params.
01:20 dalek : * List::reduce
01:20 dalek : * Eevee++
01:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27606
01:20 Eevee woo, I am useful
01:21 Tene That should be a pretty trivial patch, I was just too lazy to look up the optional param syntax stuff
01:34 rdice joined #parrot
01:48 Tene Eevee: tests for this stuff would be good too.
01:49 Eevee quite.  I have pugs commit access
01:51 Eevee ah, there's already a reduce.t, it just doesn't parse
01:55 Andy joined #parrot
01:57 braceta joined #parrot
01:58 Tene Eevee: reduce the parse fail to a minimal example
01:58 Tene and then show me
01:58 Tene (or fix it yourself if you'd like :)
02:01 Eevee it's at the colon in 'reduce { ... }: @array'
02:01 Eevee which I assume is more of a "not done" than "broken"
02:05 Theory joined #parrot
02:12 Tene I'm not sure exactly what that is.
02:13 Tene one of the weird invocant syntaxes?
02:17 Eevee yeah, indirect object syntax.  method $obj: @args;
02:17 Eevee why it's written to use the *block* as the object, I don't know
02:31 bacek joined #parrot
02:38 Tene Yeah, that looks suspicious to me, unless blocks have a reduce method too.
02:38 Tene Or, is MMD relevant?  Is MMD insensitive to the order of arguments?
02:39 Tene are
02:39 Tene multi sub t ( Foo $x, Bar $y ) { ... }
02:39 Tene multi sub t ( Bar $x, Foo $y ) { ... }
02:39 Tene different?
02:39 purl different is probably good
02:39 Tene purl: forget different
02:39 purl Tene: I forgot different
02:40 Eevee hm, I don't know.  I would think so
02:41 Tene Yeah, that would surprise me too.
02:42 Eevee http://perlcabal.org/syn/S0​6.html#Invocant_parameters  third set of code examples..
02:42 Eevee rationale might be reduce {...}: @array -> {...}.reduce(@array) -> reduce({...}, @array)
02:43 Eevee or maybe colon really does reduce to comma if indirect syntax doesn't work
02:43 Eevee this test also has "{ ... }.reduce: @array"
02:46 Tene Yeah, that looks like blocks have a reduce method too.
02:46 Tene If so, those should be in their own test.
02:47 Tene refactoring!
02:47 purl refactoring is http://st-www.cs.uiuc.edu/users/brant/Refactory/ or http://st-www.cs.uiuc.edu/​~droberts/tapos/TAPOS.htm or http://www.awl.com/cseng/titles/0-20​1-89542-0/techniques/refactoring.htm or Martin Fowler and http://www.refactoring.com/ or what you put on your time card when you're really on IRC. or http://search.cpan.org/~swalters/Object-Per​lDesignPatterns-0.03/PerlDesignPatterns.pm or gives you a fuzzy feeling
02:47 Eevee if so, it's not in the spec
02:47 Eevee also having three different reduce()s is possibly a bit overboard
02:48 Eevee actually, List::uniq isn't in S29 either
02:50 Tene "The colon is just a special form of the comma, and has the same precedence"
02:50 Eevee "{ ... }.reduce: @array" still doesn't make sense then
02:50 Eevee $obj.method, @args?
02:51 Tene That's a different kind of colon. :)
02:51 Tene foo: ...
02:51 Tene is the same as
02:51 Tene foo( ... )
02:52 Eevee aha, no method calls without parens?
02:52 Tene right
02:52 Tene parens or :
02:53 Eevee okay, that makes a bit more sense then
02:53 Tene Oh, you didn't create the sub form of reduce in that patch.
02:53 Tene Where's 'reduce' specced?
02:54 Tene Ah, S29
02:54 Eevee http://perlcabal.org/syn/S29.html#reduce
02:54 Eevee yes
02:55 Eevee so yes needs a sub form
02:56 Tene the sub form takes Code first.
02:56 Tene and the invocant form first tries method dispatch and if that doesn't work, looks for a normal sub
02:56 Tene so it doesn't need to be a Code method.
02:56 Eevee right: <Eevee> rationale might be reduce {...}: @array -> {...}.reduce(@array) -> reduce({...}, @array)
02:56 Tene Right.
02:57 Tene Do you have any particular desire to write the sub version, or should i just do it?
02:57 Eevee just wasn't sure if the syntax was meant to chain like that
02:57 Tene According to S06, it is.
02:57 Eevee it's trivial, so probably easier for you if you write it
03:01 Tene Hm.  If it's slurpy, it won't be a List.
03:01 Eevee oh, wait.  will it be a ResizeablePMCArray?
03:01 Tene yeah
03:03 Eevee sub grep and some others don't work for the same reason
03:03 Eevee mentioned it earlier, and: <pmichaud> hopefully this will be fixed when we do .HLL mapping
03:03 dalek r27607 | tene++ | trunk:
03:03 dalek : [rakudo]
03:03 dalek : * sub form of reduce
03:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27607
03:04 Tene For now we could manually construct a List from the resizablepmcarray
03:04 Tene call constructor, that is
03:04 Tene anyway, dinner now.
03:06 dalek r27608 | rgrjr++ | trunk:
03:06 dalek : [DOCS]
03:06 dalek : * docs/pdds/draft/pdd19_pir.pod:
03:06 dalek :    + Expand the documentation of :immediate by hijacking the revcomp.pir
03:06 dalek :      example, which illustrates a hidden feature.  Expand :postcomp
03:06 dalek :      slightly to match.
03:06 dalek :    + Also refine the descriptions of :main, :init, and :load.
03:06 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27608
03:07 Eevee okay, think I've played with rakudo enough for the moment..  doctor who taking priority
03:13 Tene sweet
03:30 japhb nopaste?
03:30 purl rumour has it nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating
03:55 * japhb can't wait for the next season of Dr. Who to come out on DVD ...
03:57 b2gills left #parrot
04:02 bacek hi there
04:03 bacek hello. I've got another stupid question: why 'cmp pmc1, pmc2' always returns -1?
04:03 bacek joined #parrot
04:04 bacek hello. I've got another stupid question: why 'cmp pmc1, pmc2' always returns -1?
04:04 bacek oops. sorry for spam...
04:20 pmichaud it depends on what pmc1 and pmc2 are
04:22 pmichaud easy way to construct a List is to call list()
04:30 * particle wishes he had more time this weekend for rakudo, but building a deck takes lots of time...
04:30 particle movie night &
04:32 Psyche^ joined #parrot
04:33 pmichaud I'm spending lots of time on it this weekend, but wish I had more.
04:33 pmichaud time is definitely my number one blocker at the moment.  Either that or coding speed.
04:38 Tene particle: "a deck"?
04:49 dalek r27609 | pmichaud++ | p6object:
04:49 dalek : [p6object]:
04:49 dalek : * Add P6object.pbc into root makefile.
04:49 dalek : * Set properties on files.
04:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27609
04:50 dalek r27610 | pmichaud++ | p6object:
04:50 dalek : [p6object]:
04:50 dalek : * Have p6object.t load P6object.pbc instead of P6object.pir.
04:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27610
04:52 dalek r27611 | pmichaud++ | p6object:
04:52 dalek : [p6object]:
04:52 dalek : * Add some documentation to P6object.pir .
04:52 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27611
05:01 PerlJam Good evening.
05:01 PerlJam How did the bug hunt go?
05:01 pmichaud good for me -- I found a few and got them closed.
05:02 pmichaud mostly I've been doing devel work today, though.
05:03 PerlJam looks like bernhard and fperrad have been busy updating plumhead and lua too (reading the log now)
05:05 PerlJam And I have to say that I'm always amazed and very happy when I see chromatic do something like "16.23% improvement in the Rakudo-building benchmark"
05:05 PerlJam chromatic++
05:05 PerlJam chromatic++
05:05 pmichaud I'm surprised at how much GC is a factor.
05:06 PerlJam really?  Why's that?
05:07 pmichaud I knew GC was a factor -- I just didn't think it was as big as it's turning out to be.
05:07 pmichaud I mean, when a 3-line change results in a 20% speed improvement... well, something's weird there.  :-)
05:07 PerlJam GC is pervasive; seems to me like it would be the biggest factor.
05:08 pmichaud yeah, I suppose so.
05:08 pmichaud anyway, I agree.  chromatic++
05:08 pmichaud actually, I'm quite pleased and impressed with all of the contributors we've had recently
05:09 PerlJam Jonathan++ is amazing too :)
05:09 pmichaud It's been very good to see things like .* implemented, and placeholder vars
05:10 pmichaud and Eevee's recent contribution of .reduce was a very pleasant surprise
05:10 PerlJam heh, I was just looking at that
05:13 PerlJam When's the next release?
05:13 pmichaud Tuesday.
05:13 purl i think tuesday is pick-on-purl day or ask us to give you the syntax of perldoc -f day or http://us.imdb.com/Title?0064471 or little-tolerance-for-stupidness day or devoice dynweb at will day or love masque day or random body mods day or I'll-use-colloquialisms-if-I-wanna day or FAQ day. But then so is every other day. or overbearing sexual reference day. (NO, not SOFT references.) or homoerotica day or see tuesday2
05:16 PerlJam pm: What else have you been up to lately?  And what's this Solstice framework I see you're talking about at yapc?  :)
05:16 pmichaud I've been mainly rakudo and parrot hacking all week.
05:17 pmichaud and begrudging sleep because it takes away from my rakudo and parrot hacking.
05:17 pmichaud other than that, just planning trips and vacations.  It's going to be a very busy summer.  :-)
05:19 pmichaud and Solstice isn't me (but I think you knew that).
05:20 pmichaud oooh, one week to get my YAPC::EU submission in :-)
05:20 pmichaud .....30 minutes?!?!?
05:20 pmichaud how am I supposed to do a talk in just 30 minutes?
05:21 PerlJam talk real fast
05:21 pmichaud maybe I'll do talks as part 1, part 2, etc :-)
05:22 PerlJam You could do a mix of Bob Dylan and the Takahashi method of presentation. You wouldn't even have to speak, just flip through slides.
05:22 PerlJam Dr. Seuss might be funnier though :)
05:25 dalek r27612 | pmichaud++ | p6object:
05:25 dalek : [p6object]:
05:25 dalek : * Allow new_class method to create classes with multiple parent classes.
05:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27612
05:33 pmichaud found another bug.  :-)
05:35 PerlJam I thought bug day was about squashing them, not finding them :)
05:36 pmichaud it's both.  :-)
05:36 pmichaud but I'm working on something else at the moment, so I'll just file my ticket and move on.
05:42 PerlJam hmm.  interesting.  "git svn rebase" says it's up to date, but I've only got to r27608
05:42 obra 3
05:42 pmichaud 2
05:43 pmichaud 1
05:43 PerlJam 0
05:43 pmichaud *
05:43 pmichaud (whatever.)
05:43 PerlJam -Inf
05:46 Eevee okay, I have to know
05:46 Eevee purl, tuesday2?
05:46 purl tuesday2 is Beat-the-fuck-out-of-{coworker,client}-day or kick Yaakov day or not kick dynweb day or free pr0n day
05:46 Eevee and a thrilling conclusion.  good to know not everything that happens on tuesday is terrible
05:47 pmichaud purl, tuesday is also #parrotsketch day.
05:47 purl that is too long, pmichaud
05:47 pmichaud aha.  tuesday is also too long, apparently.
05:47 Eevee I was about to say you should add that to tuesday
05:48 Eevee but tuesday is too long..  so you'd have to add it to tuesday2..  and then it would be wrong..
05:50 pmichaud I guess purl has outsmarted us once again.
05:54 pmichaud argggh.  I really want to stay up and code more, but I'm falling asleep.  :-(
05:55 pmichaud guess I'd better sleep -- sleepy code is often not good code.
05:55 PerlJam pm: nap, get up early and code some more.
05:55 pmichaud I pretty much need to dedicate tomorrow to finishing grading for my classes.
05:56 * spinclad just relieved purl of 'correct.' and 'sounds good.' responses.  key was 'forget correct. ' <- the final space.
05:56 pmichaud really?  excellent!
05:56 * pmichaud tests.
05:56 pmichaud correct.
05:56 pmichaud yay!!!!!
05:56 * purl beats pmichaud to a pulp
05:57 pmichaud 05:50 <pmichaud> I guess purl has outsmarted us once again.
05:58 pmichaud spinclad, thanks a lot for finding that.  It's been bugging me and jonathan (as you undoubtedly saw)
05:59 pmichaud okay, time for sleep.  be back tomorrow sometime
05:59 Eevee night
06:11 Zaba_ joined #parrot
06:16 spinclad pmichaud, my pleasure.  been bugging me a lot, too.
06:17 spinclad (and now that i'm caught up to thistime):
06:17 spinclad .oO { PAST::Var.new( :name('xyz'), :scope('register'), :register_class('int'), :isdecl(1) )  # generates  C< .local int xyz >  } Oo.
06:57 IllvilJa joined #parrot
08:12 Zaba_ hmm
08:12 Zaba_ seems like some svn metadata in trunk is messed
08:12 Zaba_ test fails
08:12 Zaba_ http://rafb.net/p/JgyRQf55.html
08:26 Senaka joined #parrot
08:26 Senaka left #parrot
08:54 barney joined #parrot
09:08 dalek allison@perl.org | YAPC::NA::2008:
09:08 dalek link: http://www.perlfoundation.org/​parrot/index.cgi?yapc_na_2008
10:40 rdice joined #parrot
11:01 lichtkind joined #parrot
11:01 lichtkind pmichaud: haha i offended you 30 sec ago muahaha
11:04 TimToady it's hard to offend most people at 4:00
11:04 TimToady well, I guess maybe it's 6:00 there...
11:11 Zaba joined #parrot
11:12 cognominal lunch time here
11:13 TimToady dinner time in 日本
11:13 jonathan Lunch time here too, apart from I only woke up an hour ago...
11:14 jonathan Apparently, the culture here suggests early mornings and un-late nights, though I've only lived here for a few days, so I guess I've got some de-Spanishing to go.
11:15 jonathan TimToady: When you get a spare moment, feedback on the "Private methods in classes and roles" thread would be helpful...I'm planning to look at the role composition stuff soonish.
11:16 jonathan (p6l thread)
11:18 lichtkind joined #parrot
11:19 lichtkind TimToady: yes but by  offended i mean calling him an camel, (i suggest him for white camel award)
11:22 cognominal Timtoady is there a way to add a variable to a typeclass?  I wanted to be too smart and I would get something like :    token attvalue  {  $delim= <['"]>  <-[<&"$delim]>* $delim   }
11:23 cognominal modulo the syntax of the expansion of the variable in the typeclass
11:24 cognominal heu, that would be :   token attvalue  {  $delim= <['"]>  <-[&"$delim]>* $delim   }
11:30 Ivatar joined #parrot
12:12 lichtkind there is a bug with the win32 build of parrot
12:14 lichtkind readline5.dll not found
12:24 iblechbot joined #parrot
12:31 allison joined #parrot
12:33 Limbic_Region joined #parrot
12:53 Debolaz joined #parrot
13:04 tetragon joined #parrot
13:32 paco joined #parrot
13:36 teknomunk joined #parrot
14:09 TimToady cognominal: not currently, but you can always write [<!before $delim> .]*
14:10 TimToady and maybe <-$delim> could be made to mean the same thing
14:10 TimToady but usually you just do .*? $delim
14:14 jonathan TimToady: Giving a Perl 6 talk soonish that will have some theory-ish folks there, and have just been playing with doing dependent types in Perl 6. That is, types that depend on their values.
14:15 jonathan Perl 6 is wonderful for this because of "subset" types or especially anonymous ones.
14:15 jonathan I've just written this:
14:15 jonathan sub add_all(List where { our $len = $^xs.elems; 1 } $a, List where { our $len == $^xs.elems } $b) { }
14:15 jonathan Which type-checks to make sure that you passed two lists of the same length.
14:16 jonathan However, it's not really right; using "our" means this won't work out with recursive calls or in multi-threaded environments.
14:16 jonathan Using "my" won't work as that scopes the variable inside the block.
14:16 jonathan I think "temp" will have the same issue.
14:16 jonathan Is there something I should be using here?
14:17 cognominal TimToady, thx.  anyway it easier to do a |  (or)  to get separately the simple and double quoted stuff.
14:18 TimToady jonathan: why not just List $b where { $a.elems == $b.elems }
14:19 jonathan Oh!
14:19 jonathan Yes. :-)
14:19 jonathan Thanks. Brain only half-functioning today.
14:19 TimToady mine's kinda in sideways at the moment...
14:20 TimToady long 2nd day at the hackthon plus residual jetlag
14:20 TimToady I'm sure the sake has nothing to do with it.  :)
14:21 jonathan :-)
14:21 pmichaud ....someone please talk me out of adding protoregexes to PGE  :-|
14:21 pmichaud (just figured out a relatively easy way to do it.)
14:21 jonathan pmichaud: Is it hard?
14:21 jonathan Ah, no.
14:21 jonathan :-)
14:22 pmichaud I don't think it'll be that hard, but it requires refactoring a bunch of PGE stuff (and perhaps languages that rely on it)
14:22 cognominal what is the standard way of calling a same named method in a parent class. Does this work for grammars?
14:22 jonathan Well, it gives you a COOL thing to blog about.
14:22 pmichaud cognominal: at present I don't know that there is a way to call same-named method in a parent class
14:22 jonathan cognominal: I *think* nextwith
14:22 jonathan Maybe...
14:23 pmichaud cognominal: if you know the identity of the parent class, it can be done as a sub call
14:23 pmichaud $P0 = get_global ['ParentClass'], 'methodname'
14:23 pmichaud $P0(self, args)
14:24 pmichaud jonathan: yes, it would give me a COOL thing to blog about.
14:24 jonathan Do it! Do it!
14:24 pmichaud well, I doubt I'd be able to do it in time for the release (or that that would be a good idea), so it'll be branch work for a bit
14:25 jonathan Yes, I suggest maybe do it in a seperate branch to the protoobject stuff?
14:25 pmichaud on the plus side, it could really speed up parsing
14:25 jonathan You're making it real hard for me to talk you out of this.
14:25 pmichaud it wouldn't be full ltm yet, but it would allow us to get rid of     rule statement { <for_statement> | <if_statement> | <while_statement> | ... }
14:26 ambs joined #parrot
14:26 jonathan Nice
14:26 pmichaud as well as
14:26 purl as well as is, like, entirely self-organizing
14:26 jonathan purl forget as well as
14:26 purl jonathan: I forgot as well as
14:26 TimToady that's more or less what I just hacked into STD using Moose last night
14:27 * ambs waves
14:27 pmichaud I'm only half-joking about putting protoregexes into PGE -- it's just that I have a bunch of other "high-priority" things to implement as well
14:27 jonathan Too many high-pirority things. :-)
14:28 pmichaud yes, it'd be a separate branch from p6object -- I'd want to get p6object finished first
14:28 pmichaud but I'm also having a major re-think about how PGE handles grammars
14:28 pmichaud although I think that fixing the grammar stuff would be somewhat less intrusive than protoregexes
14:29 pmichaud actually, if I get the grammar stuff in place, then protoregexes aren't so much of a pain on external systems
14:29 pmichaud hrm.
14:29 jonathan It may be better to do intrusive stuff earlier, and further back from a release.
14:29 pmichaud unfortunately I *should* be grading papers at the moment.
14:29 jonathan Ah, ugh.
14:30 pmichaud oh well, after today (tomorrow?) I'll be done with that job and can focus full-time on rakudo/parrot.
14:30 jonathan I *should* be writing slides, instead I'm playing with dependant types and being happy that Rakudo kinda does 'em.
14:31 pmichaud TimToady: I know people probably tell you this all the time, but Perl 6 (the design/spec) is just incredibly awesome.
14:31 jonathan Admittedly, I started doing that after I was planning the types section of my talk and thought, "hey..."
14:31 pmichaud jonathan: yes, I had a similar problem.  I started cleaning up a few other things in PGE and then thought "hey...."
14:32 jonathan I'm vaguely tempted to put return type checks in.
14:32 pmichaud eek!  can I at least implement 'return' first?  ;-)
14:32 TimToady well, given we've been thinking about it for close onto 8 years, I'd hope it has a few nice features in it by now.  :)
14:32 pmichaud or you can, if you wish.  :-P
14:32 jonathan I kinda wanted that in first, so I didn't have to re-do whatever I do now. ;-)
14:32 TimToady me, I *should* go to sleep now.
14:33 TimToady and in fact, I don't think I have much of a choice, if I don't want to drool into my keyboard.
14:33 pmichaud sometime this week I also need to get talk proposals into YAPC::EU.  Although how I'm going to make a talk fit in 30 mins... I don't know.  :-)
14:33 * TimToady aims head at pillows &
14:33 pmichaud sleep well, TimToady
14:34 jonathan enjoy sleep :-)
14:34 * TimToady hopes not to notice whether he is enjoying sleep...
14:34 pmichaud jonathan: did you notice that spinclad++ got rid of the "sounds good." reply from purl?
14:34 jonathan pmichaud: Yeah, I saw that in backscroll!
14:34 DietCoke sounds good.
14:34 jonathan Speakers so loud that they blow womens clothing off!
14:35 iblechbot joined #parrot
14:35 jonathan Er, I mean, spinclad++
14:36 jonathan pmichaud: 30 minutes isn't so short...I implemented and checked in "given" in a lightning talk (5 minutes) once... ;-)
14:37 jonathan I'm going to submit two: one on Perl 6 OO, one on Perl 6 types.
14:37 pmichaud for YAPC::EU?
14:37 jonathan Yes.
14:37 pmichaud Sounds good.
14:37 pmichaud :-)
14:37 jonathan Oh, on the Pittsburgh Perl Workshop
14:37 pmichaud I'll do one for Rakudo Perl and one for PCT.
14:38 jonathan I can't make it. :-( I'd already agreed to go to another workshop. On the same date.
14:38 pmichaud But I'm thinking of having part1/part2 for one of the talks.
14:38 pmichaud no problem on PPW.
14:38 jonathan Also, there is talk of a post-YAPC::EU hackathon being organized in Sweden
14:38 jonathan Just over the border.
14:38 pmichaud on mailing list there was discussion of hackathon day before and day after YAPC::EU.  Is this in addition to that?
14:38 ambs jonathan, don't you come to PPW?
14:39 ambs :D
14:39 jonathan ambs: When is that one?
14:39 jonathan I'm doing FPW and NPW this month!
14:39 ambs 6 and 7 June
14:39 ambs we have Daniel talking on SMOP and smash talking a little on Parrot
14:39 ambs would love that have some more on parrot :)
14:39 ambs or Perl 6
14:39 * pmichaud wants to go to FPW, NPW, PtPW
14:40 jonathan It's not that I wouldn't *like* to come, but I'd like to go to pretty much every Perl event going.
14:40 ambs jonathan, hehehe
14:40 jonathan It'd make it the third weekend running when I was at a workshop.
14:41 cognominal being decentralised, fpw have few talkers so half of it will be about parrot, rakudo, pct and Perl 6
14:41 cognominal thx to jonathan++
14:41 jonathan cognominal: Yeah, I've started on my two hour talk!
14:42 jonathan pmichaud: I think this is maybe in addition, it's not in Copenhagen...let me try and find out some more.
14:42 cognominal jonathan, I am still chasing a room for you in the Pastelier
14:42 jonathan cognominal: thanks
14:43 cognominal this would be nice for you to be in the same hotel than most of people
14:43 jonathan (having somewhere to sleep)++
14:44 cognominal I even know one who choosed the camping  (or is camping a faux-ami)
14:45 jonathan Camping is great fun, in good weather.
14:45 jonathan And great un-fun, in un-good weather.
14:46 pmichaud for YAPC::EU I'll already be away from home from Aug 11-17, so I don't know how many more days I can tack onto that trip.
14:46 cognominal and I don't know if there is nearby power plugs.
14:47 pmichaud although there's also talk of a post-YAPC::EU excursion to Legoland... :-)
14:47 jonathan Legoland++
14:48 jonathan cognominal: Oh yeah! Need power! No power, no code. :-O
14:48 cognominal if someone commented and applied #52276, I would be motivated to do the hash composer
14:48 allison Legoland++
14:49 ambs LegoLand++
14:49 jonathan That said, I should probably get plenty of hacking done at FPW during the day, when I'm not giving talks. Since I won't understand anyone else's. ;-)
14:49 jonathan allison: You coming to YAPC::EU?
14:49 ambs not sure if I will be able to go, but hope so
14:49 ambs jonathan, hope so :)
14:49 DietCoke going to yapc::eu?
14:49 purl well, going to yapc::eu is (: going to YAPC::EU 2007) or joel, $joel->gf
14:49 pmichaud is it possible for a class to override the 'name' vtable entry?
14:49 DietCoke going to yapc::na?
14:49 purl going to yapc::na is, like, particle, Coke, cotto, rjbs, confound, hobbs or Alias or jjore
14:49 ambs going to YAPC::EU 2007?
14:49 purl going to YAPC::EU 2007 is probably TimToady, TheDamian,  MJD, BooK, Maddingue, ambs, cog, toor, abeltje, Abigail & JettaMakka, Tux, markov, liz, woolfy, champs, Nicholas, acme, Ranguard, Barbie, pfig, davord, Trelane, Greg, DrHyde, Ovid, richardc, BinGOs, sjn, Juerd, dollarbar, dha, castaway, avar, mauke, kane & grum, ash or joel & lauren, beatnik, rGeoffrey, or not audreyt or mst or stevan and nothingmuch or abraxxa or gbjk
14:49 cognominal jonathan, but you will have to deal with me peering over your shoulder
14:50 DietCoke purl, you fickle collection of bits.
14:50 purl DietCoke: what?
14:50 pmichaud purl, going to yapc::na is also pmichaud
14:50 purl okay, pmichaud.
14:50 ambs purl, going to YAPC::EU 2007 is also smash
14:50 purl okay, ambs.
14:50 pmichaud ambs... time travel?
14:50 pmichaud or history?
14:50 jonathan pmichaud: I've also been asked, if I'll go to a Hackathon in Russia sometime in autumn.
14:50 purl it has been said that history is rewritting itself in very bad episodes of Survivor...
14:51 pmichaud Russia would be interesting.
14:51 cognominal purl,  going to YAPC::EU 2007 is also cognominal
14:51 purl that is too long, cognominal
14:51 jonathan pmichaud: Come! :-)
14:51 allison jonathan: yeah
14:51 cognominal I am built that way, purl
14:51 purl ...but cognominal is stef or ESR's padawan : "qd je comprend pas, je sors mon gun" or doesn't know shit about star Wars or crapulous...
14:52 ambs cognominal++
14:52 jonathan pmichaud: It's just a long train journey for me, mind...or there's always Aeroflot. ;-)
14:52 jonathan pmichaud: Maybe there would be someone willing to sponsor you to come or something, if you were interested.
14:52 allison purl, going to yapc::na is also allison
14:52 purl okay, allison.
14:52 pmichaud I'm interested but it really depends on the timing.
14:52 jonathan It's a long flight for you too.
14:53 jonathan The timing is something that I/we can greatly influence.
14:53 pmichaud I already have a ton of travel.
14:53 allison purl, going to YAPC::EU 2008?
14:53 purl well, going to YAPC::EU 2008 is ambs, smash
14:53 allison purl, going to YAPC::EU 2008 is also allison
14:53 purl okay, allison.
14:54 pmichaud going to OSCON 2008?
14:54 cognominal purl,  going to YAPC::EU 2008 is also cognominal
14:54 purl okay, cognominal.
14:54 pmichaud purl, going to SOCON 2008?
14:54 purl pmichaud: i haven't a clue
14:54 pmichaud purl, going to OSCON 2008?
14:54 purl pmichaud: i don't know
14:54 ambs stupid me.. 2008
14:54 cognominal I did not noticed that was 2007
14:54 jonathan pmichaud: Sure, well, I'll keep you updated on that one...I just got an email titled "Another crazy idea!" yesterday from a bunch of folks enjoying lunch at YAPC::Russia. :-)
14:54 ambs purl, going to YAPC::EU 2008 is also cog
14:54 purl okay, ambs.
14:54 ambs purl, going to YAPC::EU 2008 is also root
14:54 purl okay, ambs.
14:54 pmichaud purl, going to YAPC::EU 2008 is also pmichaud
14:54 purl okay, pmichaud.
14:55 ambs no, going to yapc::eu is (: going to YAPC::EU 2008)
14:55 purl okay, ambs.
14:55 ambs there!
14:55 ambs thanks, alisson, for noticing year problem :)
14:56 jonathan nopaste?
14:56 purl i guess nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://paste.husk.org/ or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or don't bother me while I'm eating
14:56 pmichaud so far my travel for 2008 looks like:   Fossil Rim, YAPC::NA, Northern Calif, OSCON, YAPC::EU, and Pittsburgh Perl Workshop
14:56 pmichaud oh, and Hawai'i between OSCON and YAPC::EU
14:56 pmichaud (some are vactions, some are conferences)
14:57 pmichaud *vacations
14:57 nopaste "jonathan" at 85.216.151.226 pasted "plan for advanced, trying to get your head around Perl 6 style talk - any comments?" (49 lines) at http://nopaste.snit.ch/13005
14:57 jonathan That's an outline of a talk I'm planning to give at the Nordic Perl Workshop, to an audience with some more theory-ish, academic-ish folks in it.
14:57 pmichaud In p6object I actually have HOW/WHAT transposed
14:58 jonathan It's marked as advanced rather than beginners talk.
14:58 pmichaud On an instance,  .WHAT actually redispatches to .HOW.WHAT
14:58 jonathan Oh!
14:58 pmichaud that turns out to be far simpler to manage.
14:58 jonathan I think you're doing it right, good catch
14:59 pmichaud and it means that protoobjects continue to be lighter-weight and don't require any additional attributes.
14:59 jonathan *nod*
14:59 jonathan I think S12 even says somewhere that all meta-programming is done through HOW.
14:59 pmichaud I really wish I could do per-object role composition
14:59 AndyA joined #parrot
14:59 jonathan Like in the "but" operator?
15:00 pmichaud partially -- mainly so that I don't have to create a separate class for protoobjects
15:00 jonathan I think under the hood it will still really be a separate class.
15:00 pmichaud right now the separate-protoobject-class is causing a problem because the <typeof> opcode is returning an empty string for the name
15:01 pmichaud $S0 = typeof object   # works
15:01 pmichaud $S0 = typeof protoobject   # returns ''
15:01 pmichaud and I can't seem to override the 'name' vtable method.
15:01 jonathan There is a name v-table method?
15:01 pmichaud yes.
15:02 pmichaud That is what <typeof> calls, and it's also mentioned in docs/vtables.pod
15:02 jonathan Oh, yes.
15:02 jonathan You're right.
15:02 pmichaud I'm not sure why I can't override it though.
15:03 jonathan I've just looked at the source, and I know why. :-)
15:03 pmichaud bug?  design?
15:04 pmichaud where are you looking, so I can tag it for future reference in my head?
15:04 cognominal joined #parrot
15:04 allison bug. or, more accurately, not yet completed implementation
15:04 jonathan pmichaud: I'll implement it for you now, it should be a 5 minute job.
15:05 pmichaud oh, that would be a huge help, yes.
15:05 jonathan pmichaud: Did the talk outline look OK overall?
15:05 pmichaud oh, still looking.
15:05 pmichaud (got distracted.)
15:05 jonathan :-)
15:06 pmichaud how long is the talk?
15:07 jonathan An hour.
15:07 purl an hour is half full! or long enough for lunch
15:07 pmichaud Looks like a great talk.
15:07 jonathan Cool, thanks.
15:07 pmichaud An hour should be about right for what you have.
15:08 jonathan I'm trying to give people more of the "big picture" view of the langauge.
15:08 pmichaud yes, the focus on mmd/types should be good for theory-ish/academic-ish folks.
15:09 pmichaud is this purely Perl 6 or are you also going to mention Rakudo and what it has/doesn't have thus far?
15:10 jonathan Purely Perl 6.
15:10 jonathan I have a second talk the next day where I will talk about Rakudo.
15:11 jonathan I'm covering PCT a bit in the version of that at the Nordic Perl Workshop.
15:11 pmichaud excellent.
15:11 jonathan Probably less so at the French Perl Workshop because cognominal is already talking about PCT there.
15:11 cognominal that'a good,
15:12 jonathan My talk at Stockholm uni will be a bit on Parrot architecture, but more on PCT than anything.
15:13 jonathan Sell it as a kinda "playground" for making compilers for dynamic languages.
15:13 pmichaud the various times I've presented PCT I always getting the academics asking if it could be used for teaching a compilers course :-)
15:13 jonathan Should be quite appropriate, for students who've just done a dynamic languages course.
15:13 jonathan "And today we'll build our own language"
15:14 pmichaud or, as I do in the presentations...  "And today, we'll build our own language.  From scratch.  During the talk."
15:14 jonathan Yeah, something like that.
15:14 purl i guess yeah, something like that is what I'd do as well
15:15 * pmichaud is impressed by how many purl factoids jonathan++ manages to hit.  :-)
15:15 * jonathan probably just talks too much.
15:17 jonathan pmichaud: Got overriding name working, just smoking it then I'll commit
15:18 nopaste "jonathan" at 85.216.151.226 pasted "this now works locally" (15 lines) at http://nopaste.snit.ch/13006
15:18 pmichaud wonderful.  That saves me quite a few headaches.
15:22 * pmichaud comes up with another cool idea for protoregexes.
15:23 pmichaud Paula just told me "sounds good" and internally I mentally added "Speakers so loud..."   *sigh*
15:24 * jonathan wonders if he dares to ask what it is
15:24 jonathan lol!
15:25 pmichaud oh, I just figured out how I can automatically register the protoregexes without a lot of :load :init code.
15:25 jonathan OMG my patch involving C passed the codingstd tests!
15:26 jonathan Even trailing whitespace!
15:27 jonathan pmichaud: Committed, r27621, enjoy.
15:28 pmichaud jonathan: many thanks
15:29 pmichaud time for me to do some shopping, some lunch, and then grading
15:29 jonathan OK
15:29 jonathan Slide writing, slovak studying and dinner cooking for me.
15:29 pmichaud I'll probably be afk most of the rest of today.
15:29 jonathan OK, see you tomorrow then.
15:29 pmichaud as soon as my grading is done, though, I'll be back here again.  Too much cool code to write.
15:29 pmichaud see you tomorrow.
15:33 Theory joined #parrot
15:34 Senaka joined #parrot
15:34 Senaka seen chromatic
15:34 purl chromatic was last seen on #parrot 4 days and 11 hours ago, saying: Some sort of mediated access anyway.  [May 13 21:49:02 2008]
15:47 Tene purl: airports?
15:47 purl hmmm... airports is the most likely method.
15:47 Senaka lol
15:47 Tene purl: pungenday?
15:47 purl tene: i don't know
15:50 Senaka left #parrot
16:05 Zaba_ joined #parrot
16:09 dalek r27613 | pmichaud++ | p6object:
16:09 dalek : [p6object]:
16:09 dalek : * Make :name option work for registering classes.
16:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27613
16:09 dalek r27614 | bernhard++ | trunk:
16:09 dalek : [commit]
16:09 dalek : driver_nqp.pbc no longer needs to be built
16:09 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27614
16:09 dalek r27615 | fperrad++ | trunk:
16:10 dalek : [Lua]
16:10 dalek : - remove useless .local (when vararg)
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27615
16:10 dalek r27616 | bernhard++ | trunk:
16:10 dalek : Set svn:eol-style to native for languages/c99/src/cpp/t/comment_01.t
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27616
16:10 dalek r27617 | bernhard++ | trunk:
16:10 dalek : [regex]
16:10 dalek : make test target work under Unix:
16:10 dalek : t\harness  => t/harness
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27617
16:10 dalek r27618 | allison++ | pdd25cx:
16:10 dalek : [pdd25cx] Changing the names of 'rethrow_exception' and 'rethrow_c_exception'
16:10 dalek : to 'Parrot_ex_rethrow_from_op' and 'Parrot_ex_rethrow_from_c'.
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27618
16:10 dalek r27619 | allison++ | pdd25cx:
16:10 dalek : [pdd25cx] Fixing constant strings that should be CONST_STRINGs.
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27619
16:10 dalek r27620 | fperrad++ | trunk:
16:10 dalek : [Lua]
16:10 dalek : - bytecode translation : split in & out register (binop)
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27620
16:10 dalek r27621 | jonathan++ | trunk:
16:10 dalek : [core] Allow object's name vtable method to be overridden.
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27621
16:10 dalek r27622 | fperrad++ | trunk:
16:10 dalek : [Lua]
16:10 dalek : - bytecode translation : split in & out register (call)
16:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27622
16:10 moritz that's feather's svn working again ;)
16:48 japhb allison: around?
16:48 purl nope.
16:57 Ron joined #parrot
16:58 dalek r27623 | smash++ | trunk:
16:58 dalek : [t]
16:58 dalek :  * make t/distro/file_metadata.t happy
16:58 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27623
16:59 japhb debian++ # Updating testing to perl 5.10
16:59 japhb debian-- # Removing SVK if you accept perl 5.10
17:04 Ron_ joined #parrot
17:20 allison japhb: hi
17:20 purl privet, allison.
17:21 barney joined #parrot
17:21 japhb allison: I sent my question about licensing of translated sample code to parrot-porters, but then I realized that may not be the right place.
17:22 japhb Since you were the visible face of Artistic 2.0, I guessed you would be the right person to ask, er, who to ask.
17:22 allison I'll go find the message
17:23 japhb thx
17:24 allison Hmmm... better not to include it in the repository
17:24 japhb harumph.  Ah well
17:25 * japhb is going to have to come up with a lot of samples from his own brain
17:25 japhb Damn having to think
17:26 allison you don't really want to distribute large sections of example GUI code in the distribution anyway, so put them up on a  page for OpenGL in Parrot
17:31 japhb I accept the decision, but can you explain why it wouldn't be OK from a license perspective to include it with Parrot?  That would inform my reasoning in the future about such issues ....
17:35 japhb I guess I should have put allison: in front of that, given my delay in asking ...
17:42 allison we need to keep the licensing of Parrot as clean as possible, so the license is the main reason not to
17:45 allison japhb: but it really is true that a few focused examples will go a lot further toward getting people started
17:46 allison so, even if licensing wasn't an issue, I would encourage taking the time to come up with your own examples
17:49 allison (more on licensing): the main problem is that when people use open source code, they have to check the terms of the license to make sure they comply. the more different licenses mixed within the project, the harder their job is, and the more "sketchy" the project appears.
17:53 japhb allison: OK, that makes sense to me.
18:08 ambs slavorg, trust me
18:08 slavorg But I already trust me
18:09 ambs slavorg, op me
18:09 Ron joined #parrot
18:10 ambs slavorg, op
18:10 * ambs shrugs
18:10 * ambs was trying to update the number of tickets.
18:11 japhb ambs, what's the number right now?
18:11 ambs 694
18:12 Topic for #parrotis now Devel: 0.6.1 | http://parrotcode.org/ | 694
18:12 japhb grrr
18:12 Topic for #parrotis now Devel: 0.6.1 | http://parrotcode.org/ | 694 new/open tix
18:12 ambs japhb, thanks
18:13 ambs slavorg, trust
18:13 slavorg huh?
18:13 ambs slavorg, do op
18:13 japhb ambs: If you have trust, I think it should auto-op you on join.  Try that?
18:13 cognominal what is this number?
18:13 japhb er, try leaving and joining?
18:14 ambs left #parrot
18:14 ambs joined #parrot
18:14 japhb cognominal: Since the "700 bug challenge", people have been wanting to keep track of that.
18:14 ambs whatever, was trying to understand slavorg
18:14 japhb It's not automated, so occasionally someone updates it by hand
18:15 cognominal you can fix one by commiting my patch  http://rt.perl.org/rt3/Tic​ket/Display.html?id=52276
18:15 japhb cognominal: I'm still in the commitbit process, not a committer yet
18:16 japhb slavorg, help
18:16 slavorg I'm slavorg, an op-bot.Commands: trust, distrust, believe, disbelieve, leave, join. See also http://jerakeen.org/programming/slavorg
18:17 ambs slavorg, leave #parrot
18:17 slavorg left #parrot
18:17 slavorg joined #parrot
18:17 cognominal slavorg,  join #perlfr
18:17 slavorg I won't join there, because I don't trust you there.
18:18 japhb slavorg has rather compartmentalized notions of trust
18:18 ambs slavorg, do op
18:19 slavorg joined #parrot
18:24 ambs japhb, please op slavorg
18:24 Infinoid opbots, names
18:25 ambs there
18:25 Infinoid next question, where is clunker
18:26 clunker3 joined #parrot
18:26 ambs there
18:26 ambs :D
18:29 Topic for #parrotis now Devel: 0.6.1 | http://parrotcode.org/ | 19/675 new/open tix
18:30 ambs easier that way
18:30 ambs than to sum all the time
18:43 Tux32 joined #parrot
18:43 Tux32 left #parrot
19:58 Psyche^ joined #parrot
19:58 japhb ambs++ # easy is good
19:58 japhb clunker3, help
20:00 japhb Wow, I wasn't expecting all that
20:03 dalek r27624 | allison++ | pdd25cx:
20:03 dalek : [pdd25cx] Update rethrow for new semantics, untodo rethrow tests.
20:03 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27624
20:04 dalek allison@perl.org | Concurrency Tasks:
20:04 dalek link: http://www.perlfoundation.org/pa​rrot/index.cgi?concurrency_tasks
20:04 shorten dalek's url is at http://xrl.us/bjqpy
20:11 Zaba joined #parrot
21:02 bacek joined #parrot
21:52 guru joined #parrot
21:53 tedkat joined #parrot
22:02 Theory joined #parrot
22:06 tetragon joined #parrot
22:26 guru left #parrot
23:15 kid51 joined #parrot
23:26 dalek r27625 | jkeenan++ | trunk:
23:26 dalek : Applying patch submitted in
23:26 dalek : http://rt.perl.org/rt3/Tic​ket/Display.html?id=54374.  Provide a little bit
23:26 dalek : more configuration information so that tests of inter::libparrot emit no
23:26 dalek : 'uninitialized value' warnings.
23:26 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=27625
23:52 japhb Any windows or Mac OS X people about?
23:53 Eevee I love that you actually have to ask around for a Windows user
23:54 japhb heh
23:57 Tene I'm installing XP in a VM right now.  Does that count?
23:58 tetragon I'm here with PPC OS X 10.5
23:59 japhb Great

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

Parrot | source cross referenced