% cognominal has left cognominal!~stef@cac94-1-82-67-232-89.fbx.proxad.net
% cognominal has joined #parrotsketch
% cognominal has left cognominal!~stef@cac94-1-82-67-232-89.fbx.proxad.net
% cognominal has joined #parrotsketch
% cognominal has left cognominal!~stef@cac94-1-82-67-232-89.fbx.proxad.net
% cognominal has joined #parrotsketch
% leo has joined #parrotsketch
% mdiep has joined #parrotsketch
<obra> 'afternoon
<obra> Parrot Sketch Roll Call. Who's around?
<leo> hi all
<mdiep> helol
% pmichaud has joined #parrotsketch
<mdiep> err, hello
<pmichaud> hello
<obra> right then.
<obra> let's get rolling.
<obra> What's new, Patrick?
<obra> or, say, leo
% particle has joined #parrotsketch
<leo> - Released Parrot 0.4.0 "Luthor"
<leo> - moved return value passing out of *Continuation.invoke into the set_returns opcode and created a PIC (polymorphic inline cache) version of it
<leo> - speeds up the fib benchmark by 25%
<leo> - it's also the base for not creating a plain return continuation at all, just for returning from a sub, but that's optimization aka later
<leo> - not much else last week, code cleanup, bug fixes and such
<leo> ^Z
<obra> Blocking on anything?
<leo> no, still a lot todo
<obra> More important question: Are you having fun?
<leo> a spec for named argument passing would be fun though
<leo> I have always fun hacking Parrot ;-)
<obra> Cool.
<obra> mdiep: 
<obra> what's up?
<pmichaud> sorry to interrupt, but I can only stay for a second -- I'm at home with sick kiddies
<obra> ok. then go for it, patrick.
<obra> if you have time to report, great.
<pmichaud> little to report from my end (for reason just cited), glad to see the 0.4 parrot release
<obra> I hope the kids feel better. Flu?
<pmichaud> kids should be back in school tomorrow.  Yes, probably flue
<pmichaud> flu
<pmichaud> I'm still doing parsing stuff and will work on the 800 p5regex tests
<obra> heh.
<pmichaud> I'll also see about recruiting people to help with the regex tests :-)
<obra> nice
<pmichaud> I hope to get more people hacking on the pge guts
% chromatic has joined #parrotsketch
<pmichaud> anyway, I have to go prepare lunch
<pmichaud> not blocking on anything except time at the moment.  It would be nice to have a named argument spec.  :-)
<obra> Having fun?
<pmichaud> oh yes, definitely
<obra> :)
<pmichaud> that's my report, and I'm afk for a while
<obra> cool. enjoy the lunch and the feeding juice to the little ones
<obra> popping back. matt: what's new?
<mdiep> it was a slower week for tcl as well
<mdiep> some minor code cleanup
% Coke has joined #parrotsketch
<mdiep> coke improved the compiler status this week -- we can now autogenerate builtin subs from the inline'd versions
<mdiep> Klaas-Jan Stol was kind enough to do a little hacking and provided a [flush] implementation for us
<mdiep> coke also added dummy version of some builtins that we're missing so that we can start working on being able to run tcl's test suite natively
<obra> very cool.
<obra> Let me guess. You guys want a named args spec
<leo> sounds like  a good guess ;-)
<mdiep> after a little complaining on #parrot, leo implemented errors for PIR sub parameter under/overflow, which is helpful. but we're still waiting to be able to use .param with it.
<mdiep> we will eventually need a named args spec, yes :-)
<mdiep> I don't know that it's high on our priority list though
<obra> Ok :)
<mdiep> and also, the namespace spec hit the list this week
<mdiep> there's a fair amount of discussion. I'm still trying to catch up with all of it.
<mdiep> ^D
<obra> having fun?
<mdiep> the namespace spec isn't the most fun for me. but I got to hack on partcl itself some this week, and that's fun.
<leo> mdiep: sorry ;-)
<mdiep> I'm hoping that my part in the namespace design is almost over
<obra> (In case anyone can't tell, I've recently been reminded about _why_ people contribute to opensource development. And what project management can do for them. And if we're not actually optimizing for fun, there are problems)
<mdiep> heh. obra++
<obra> mdiep: blocking on anything else?
<mdiep> nope. just the ability to use .param inside of exception handlers.
<obra> I mostly see my role as: making sure that the developers are getting what they need, enjoying themselves and actually communicating amongst themselves and with the community. (In no particular order. All are important)
<obra> Are there other things I should be doing that I'm not?
<leo> mdiep: you are writing a compiler - just emitting 'get_results' can't be a problem
<obra> (Consider this my report ;)
<mdiep> leo: not necessarily a problem, just more work.
<leo> it's one macro (last time I looked)
<leo> mdiep: and of course patches welcome ;-)
<obra> ok. next report, I suppose.
<obra> chip: you here?
<obra> or autrijus 
<obra> or chromatic 
<chromatic> I'm here.
<chromatic> I checked in the basic Test::More, sans tests.
<chromatic> I can add the like() and unlike() shortly, but for the sake of maintainability I plan to macroize most of the MMD variants, as they're all very similar except for coercions.
<chromatic> I have two sticking points.
<chromatic> First, load_bytecode only runs once.  That is, if you try it on the same file more than once, from two different places, you get an error about not being able to run initialization code.
<chromatic> Patrick has run into this too.
<chromatic> Second, I don't have a good place yet to stick the tests for modules.  Perl 5 uses the ext/ directory to store modules, but it's a terrible name and it doesn't have to run tests with two different programs (Parrot and Perl).
<chromatic> I think the right solution is to create a t/modules directory and make sure that all tests run through a little harness file there.
<chromatic> I haven't looked at the test target in the Makefile though, so it may be even easier than that.
<obra> anything else?
<obra> chromatic: having fun?
<chromatic> I wish Parrot ran on my DS.  That would be Very Fun.
<obra> Nintendo DS?
<chromatic> That's right.
<obra> What does the devkit look like?
<chromatic> I haven't looked yet; I doubt there's one available to non-publishers such as you and I.
% allison has joined #parrotsketch
<obra> ok. aside from your sticking points, do you have blockers?
<chromatic> The library loading is going to block me very soon.
<obra> chromatic: http://www.metrowerks.com/MW/Develop/Games/Nintendo_DS.htm
<chromatic> The other won't.  It sounds like Jerry Gay has a solution.
<obra> cool.
<leo> chromatic: the lib loading stuff is very high in priority of fixes for next release
<chromatic> I'm happy to write a test for it if that would help
<leo> double loading prevention and such - would be great to have tests
<obra> ok. anything else before I turn to allison?
<obra> Allison: what's new and exciting?
<allison> I pushed Punie to the point where it's emitting PIR and compiling it
<allison> the whole PAST->POST->PIR phase went much faster than I expected.
<allison> That encourages me that the design is heading in the right direction.
<allison> Next is to build up more and more complex syntax in Punie.
<allison> starting with "print 'ok 1'"
<obra> nice.
<obra> got a failing test yet?
<allison> Well, all the language tests were failing for the past month. :)
<obra> :P
<allison> Still have several TODOs that Autrijus checked in.
<obra> Whatcha blocking on?
<allison> nothing at the moment
<allison> have a Q for later
<allison> EOR
<obra> Having fun?
<allison> yes! :)
<obra> woot!
<obra> chip, autrijus: Last Chance!
<mdiep> obra: I've got one more blocking item for Tcl
<obra> mdiep: what is it?
<mdiep> coke reminded me that we have a weird segfault (#37814) when trying to run the tcltest harness natively
<mdiep> http://rt.perl.org/rt3/Ticket/Display.html?id=37814
<leo> mdiep: that's the lib loading stuff
<mdiep> leo: ah, excellent. we'll be looking forward to the fix. :-)
<mdiep> leo: perhaps you can add a note to the ticket stating such?
<obra> ok. question time.
<obra> who has questions?
<leo> (and for sure a reason that I don't even look at .param for exception handlers, it there is a working solution ;-)
<leo> mdiep: will do
<mdiep> question: what will the process look like for creating a PDD from the namespace spec? I suspect that it's a question for chip
<leo> looks like, yes
<allison> my Q was the obj.method() syntax
<leo> well, as answered in the mail, it's sometimes useful, it's ambiguous, and we might need a better syntax for the variable case
<leo> e.g. like the proposed obj.$method()
<leo> OTOH obj."method"() is always ok
* allison searching p6i mailbox for reply
<allison> mmmm... ok
<allison> $ works
<allison> but obj.$$P1() is a little confusing
<leo> then we sould forbid the bare word?
<allison> I like the bare word
<leo> there is already a $ - we special case that one probably
<allison> it's just the multiple meanings that cause problems
<leo> hmm
<leo> that's with all the bare words in imcc sytax
<allison> If I were to choose one meaning for the bareword it would be "call the method with this name on this obj"
<leo> sure, but in your case you additionally have .local pmc method
<leo> now what
<leo> then the bare word is a declared variable
<leo> and it is correct sill, but doesn't DWIM for you
<allison> I suspect it's a bigger problem than just method calls
<leo> it is for sure
<allison> not a "today" solution
<leo> I fear not - no
<obra> Ok. anything else for today?
<obra> Right. It was pointed out to me that I hadn't said "hello" to our listening audience in a while.
<obra> So, hello, lurkers. 
<obra> We've got a few ground rules here
<obra> Conveniently, everybody's been honoring em ;)
<obra> So. the deal is:
<obra> * ParrotSketch is a weekly meeting. 
<obra> * It's open to the public - for listening only
<obra> * It's publically archived
<obra> * If you aren't an invited participant (IE, the moderator has asked you to join and report), you should keep quiet, though feel free to message the moderator privately.
<obra> * Right now, we don't enforce any of that technically and as long as the rest of the audience stays as polite as you all do, we never will.
<obra> So, thanks! And I hope it's interesting and useful.
<obra> See y'all next week, gang.
<leo> see ya and thanks obra 
<obra> (And yes, I'm usually the moderator, though occasionally, I ask someone else to fill in for me)
<mdiep> thanks, obra
% allison has left allison!~allison@sub17-30.member.dsl-only.net
% chromatic has left #parrotsketch
% Coke has left #parrotsketch
% pmichaud has left pmichaud!~chatzilla@c-24-1-26-255.hsd1.tx.comcast.net
% particle_ has joined #parrotsketch
% particle has left particle!~particle@144.81.84.172
% particle_ is now known as particle