% mdiep_ has joined #parrotsketch
% mdiep has left mdiep!~matt@bursley-219-72.reshall.umich.edu
% mdiep_ is now known as mdiep
% smash has joined #parrotsketch
% jonathan has joined #parrotsketch
% chromatic has joined #parrotsketch
% allison has joined #parrotsketch
<obra> hi gang
<allison> hi
<mdiep> hello
<jonathan> hola
<particle> howdy
% allison_ has joined #parrotsketch
% allison has left allison!~allison@sub17-30.member.dsl-only.net
% allison_ is now known as allison
* obra pokes chip pmichaud, chromatic leo_ coke
<leo_> hi
<leo_> hi-lighting helps ;)
<obra> :)
<chromatic> Yep.
<tewk> hi
<obra> hey tewk
* obra gives people a last moment to pre-cache their reports
* tewk actually shows up on time :)
<obra> so, tewk, what's new and exciting? What have you been up to? What's on your list?
<tewk> C99 can strip out #defines of postgres spi.h and generate the starts of a corresponding spi_constants.pasm
<tewk> Trying to build parrot on windows in an attempt to use MSVS to learn more about my pir methods on dynamic pmcs bug.
<tewk> Using GDB I've learned that the method isn't in the namespaces that are searched by the parrot function find_method_direct1
<tewk> The namespaces are the correct, they just don't contain the pir methods that they should.
<tewk> Chip is also working on fixing the dynamic pmcs bug.
<tewk> EOR
<particle> tewk: lemme know if you need windows help
<obra> Anything you're blocking on?
<tewk> Chip is aware, so no.  particle I think i need to install the platform sdk which I'm doing now
<obra> leo_, then allison, mdiep, jonathan, particle, chromatic
<leo_> * contributed to various discussions on p2 or IRC:
<leo_> - packfile PDD
<leo_> - thread-sharable core structures
<leo_> - *BooleanArray
<leo_> - OO design and parents/mro
<leo_> - I/O redirection to a file
<leo_> - I/O redirection to a STRING (w/ chromatic's implementation)
<leo_> - argument passing code and behavior
<leo_> * implemented (hopefully) :maybe_flat argument passing flag a minute ago
<leo_> * I'm currently trying to let folks pick up things and doin' it on their own :)
<leo_> ^D
<obra> anything you need someone to take off your hands?
<leo_> not yet
<leo_> I'll have to prepare a few tasks
<obra> ok
<obra> allison: what's new?
<allison> Chip and I met on Wednesday and did some brainstorming on our todo lists, as well as on the interface for overloading vtable methods in Parrot Objects.
<allison> I've nearly finished a rough Threads PDD. It's more a collection of prior wisdom. (Thanks to Nicholas Clark for unearthing an older draft and discussion by Dan.) I'll check that in later today, and start a mailing list discussion.
<allison> EOR
<obra> Any luck roping lizm in?
<allison> lizm?
<allison> oh, the plan was to collect the text
<leo_> Liz M.
<allison> and then pull her into the discussion
<obra> Cool
<allison> it's nice to have a reference collection of prior art to use as a base
<obra> (Elizabeth Mattijsen for those playing the home game)
<obra> Ok. Anything you're blocking on in general?
<allison> an apartment in the SF area, so I can get some peace to work?
<allison> no, nothing useful :)
<obra> Oh. _now_ you're asking for miracles.
<allison> heh
<chromatic> She didn't say "inexpensive".
<obra> chromatic: point.
<obra> mdiep: what's new?
<mdiep>  - midterms this week; break this weekend (time to hack parrot)
<mdiep>  - working on a Forth implementation; it's easier and more relaxing
<mdiep>  - I'll get the basics done and then add to the repo
<mdiep>  - currently trying to get new_from_string working from PIR land
<mdiep> ^D
<obra> Anything that can happen to increase your fun?
<mdiep> um.. less school work?
* obra phones in a bomb threat.
<mdiep> so, no. not really. :-)
<obra> jonathan: How's Europe treating you?
<obra> er. The Continent. (You're UK-based, right?)
* obra resorts jonathan to the end.
<obra> particle. you around?
<particle> ~ i've been concentrating on calling conventions. some cases are not represented in the arg processing state machine
<chromatic> Preach it!
<allison> I think jonathan is in spain this week
<particle> ~ i committed docs/pcc_state.pod, an incomplete state table diagram, which should shed light on the dark corners
<particle> ~ and i've *just* committed t/op/cc_state.t, which will eventually test all the states in the table
<particle> ~ plus the usual small patches and fixes
<particle> .end
<obra> allison: yes. and he said hi 18 minutes ago ;)
<obra> ok. chromatic what's new?
<chromatic> * fixed up Parrot::Embed and moved to classes and objects (still needs documentation)
<chromatic> * talking to various groups about embedding strategies
<chromatic> * implemented some of the namespace spec (have a question about find_sub())
<chromatic> * did some work on the SDL examples (but the object model is frustrating me)
<chromatic> * working on a port of the Perl 6 object instantiation model (question about parameter passing posted to list)
<chromatic> That's it.
<obra> neat.
<obra> blocking on anything?
<chromatic> Random bugs I keep posting.
<obra> The random ones are always the hardest.
<obra> jonathan: you're up
<chromatic> SDL Minesweeper works, sort of, but the math is incorrect for drawing the board.
<jonathan> * Working on-site in Spain, which is quite tiring, so hard to find any Parrot time. :-(
<jonathan> * Answering the odd thing on the mailing list to do with bytecode PDD, still thinking over leo's posts (thanks for input).
<jonathan> * Will try and steal a moment to check in some changes to the PDD later this week.
<jonathan> }
<jonathan> EOR
<obra> cool.
<obra> anyone whom I missed?
<obra> failing that, chromatic, you had a question?
* particle reserves (a) comment
<chromatic> Right.
<chromatic> I added the get_var() method on namespace PMCs, but there's no easy way that I could tell if a PMC really is a var.
<chromatic> It makes more sense for HLL implementations of NameSpace, where you might have a sigil prefix or a base class to check.
<chromatic> I didn't really want to hard-code a list of variable PMCs, and there's no role we can check in the code yet.
<chromatic> Is it okay for now, or is that a TODO bug?
<particle> is there a vtable method that only vars do?
<leo_> indeed, chip always wante dto have the current namespace pmc totally untyped, therefore get_var() and such would only be implemented in subclasses of the NameSpace PMC
<leo_> I thinkm the default just should do it, i.e. be fully typed
<chromatic> Right now it just returns the PMC with the given name, if it exists, regardless of the PMC type.
<leo_> (now itĂt's half typednamespace / not-namespace )
<particle> time for a TypedNameSpace subclass?
<leo_> wouldn't work yet
<leo_> too much of NameSpace is hard-wired and non subclassable
<particle> blech
<obra> other questions?
<leo_> I'm worried about lang interop if everyone does it's own namespace pmc
<chromatic> I'm not thinking of it necessarily from a HLL implementation perspective.
* particle thinks lang interop should be pushed to the hll implementors
<leo_> yes - but how do you import then, if namespaces are really different - that might be a problem
<allison> if you push it to the hll implementors, they'll implement it in 100 incompatible ways
<chromatic> There are a lot of metaprogramming opportunities that Parrot could make easier in PIR; I think that might help interoperability and HLL implementation.
* mdiep thinks the existing namespace pdd will be sufficient. but we might as well wait until we have two languages to try it with
<tewk> Could a list of metaprogramming opprotunities show up as a message on p2? I'm just curious of what your thinking.
<obra> ==tewk
<particle> yeah, coke's been waiting on a second language to try hll-interop
<leo_> even HLL - lib isn't interacting yet in a nice way
<chromatic> I'll send in my object instantiation library.
<leo_> you can take 'lib' as the 2nd language, which is 'parrot' - the default
<obra> anyone got anything else?
<particle> yes.
<obra> Go for it
<particle> it seems there hasn't been enough input from designers on the mailing list of late. i've tried to handle/defer some requests, but we have both core committers and *new volunteers* that aren't getting the attention i believe they require.
<obra> allison: do you know what's up with chip?
<allison> real life
<obra> ok. same with you? is there a way to get a better parrot timeslice out of either or both of you?
<allison> last week he said he was coming into some more free time
<particle> patrick's got real life problems, too.
<leo_> is there some forseeable time frame for that?
<allison> I'm not good at monitoring the mailing list
<leo_> allison you should ;)
<allison> but if you ping me off list, I'll always jump into the thread
<particle> allison, is there a better way to bug you? copy you directly?
<allison> don't cc, since it'll just disappear into the mailing list folder, but drop me a separate email
<allison> or AIM me
<allison> or even call me
<leo_> allison: btw I've mailed re ci bits for cage cleaners
* particle requests allison's contact info
<allison> (I'm happy to hand out my cell phone number for parrot developers)
<obra> allison: what could make it easier for you to monitor the mailinglist?
<allison> I'll msg particle
<allison> reducing my email inflow down from 1000 messages a day?
<leo_> hrmmpfff
<tewk> Tewk would love to exchange IM addresses with parrot developers.
<leo_> filter 999 ;)
<allison> leo: :)
<obra> allison: I know you despise IRC, but I wonder if having a designated weekly community parrot irc hour wouldn't suck
<obra> It might suck.
<allison> isn't that what this is?
<particle> imo #parrot is fine, it's p2 that's lacking input
<leo_> indeed - current chiefs aren't visible at all
<allison> aye
<allison> I'll keep trying to manage reading every message
<obra> allison: this isn't a public meeting. I was thinking "some way to let _everybody_ talk." But I see the point about it being p2
<allison> but in the mean-time, any help you can give me to make sure I don't miss the really important ones are greatly appreciated
<particle> allison, just filter for your name, i'll mention you when i think you should reply :)
<allison> (chromatic often did that for me, drew my attention to mesages that needed my attention)
<leo_> you can't read 1000 mails a day - that's 86 seconds for one - if you don't sleep ;)
<allison> yup, which is why I have a backlog of several thousand messages, and will never catch up
<leo_> rm -rf backlog
<allison> email is the bane of my existence
<particle> you should make spamassassin smarter or just more aggressive :)
<obra> I've found that spending 6 hours on one day offline killing my backlog has done wonders for me.
<obra> "Threads older than 6 weeks, where I read the first 3 and they're not relevant can die"
<leo_> especially in the case of ', and will never catch up
<allison> particle: I'll try an 'allison' filter
<allison> obra: hmmm... I suspect you're missing the sheer magnitude of the problem
<allison> obra: but thanks for the suggestions :)
<obra> allison: I had O(3k) items in my inbox ;)
<obra> I have less than a hundred now. Though that's up from the 50 I managed to get it down to. 
<obra> But understood
<obra> anyway.
<obra> Anyone have anything else?
<particle> allison, thank you for the extra effort.
<leo_> well, some parrot release would be nice
<particle> i spoke to chip about that
<particle> it's on his list for last week
<leo_> 'k
<chromatic> We'll have to get the failing tests down for that though.
<particle> okay, we can do that without him. i'll make efforts in that direction, and would appreciate help.
<leo_> we can pugs-practice-em for release (i.e silence)
<chromatic> The sprintf() tests yes.
<chromatic> The others... it'd be nice to fix them up, if possible.
<particle> oh, chromatic, your load_bytecode tests are fragile (rely on being run from a particular dir)
<particle> should be an easy fix to use File::Temp
<leo_> sure, but we still can just todo some
<chromatic> Works for me.
<leo_> todo => release # /me always wanted to have some silence filter
<particle> test::harness::releasefilter or something...
<particle> good idea
<leo_> chromatic: re MakeObject - see also my experiments re instantiate (opcode and mails on p6i)
<chromatic> Will do.
<leo_> http://groups.google.de/group/perl.perl6.internals/browse_frm/thread/e68dc0a0a96585b7/462023a5939b0844?lnk=gst&q=toetsch+instantiate&rnum=2#462023a5939b0844
<leo_> it's basically combining full argument passing with object.new (aka __init)
<chromatic> Does it walk down the inheritance chain?
<leo_> it should of course, but as chip didn't like it, it's stalled
<leo_> there are sdl examples even ;)
<leo_> IMHO we should unit a lot of the current new opcodes into a reasonable scheme
<chromatic> That would be nice.
<chromatic> The objects PDD is still in clip/
<chromatic> FWIW.
<allison> yeah, the objects pdd needs some attention
<allison> it needs composition (roles) for one thing
<allison> I told jonathan I'd work on it after getting out at least a first draft of events and threads
<obra> ok. is it time to wrap this up for the week?
<allison> thanks all, I'm off to lunch
<leo_> thanks all - cya next week or at #parrot
% chromatic has left #parrotsketch
% jonathan has left #parrotsketch
% allison has left allison!~chatzilla@nat-147-1.oreilly.com
% allison_ has joined #parrotsketch
% allison_ is now known as allison
% allison has left allison!~chatzilla@adsl-69-232-231-227.dsl.pltn13.pacbell.net
% allison_ has joined #parrotsketch
% allison_ is now known as allison
% allison has left allison!~chatzilla@209.58.173.176