% coke has joined #parrotsketch
% pmichaud has joined #parrotsketch
<obra> who's around?
<obra> My network here today turns out to hurt really badly.
<PerlJam> I'm here, but I don't count :-)
<PerlJam> 0, 1, ... infinity!    See?
<obra> particle: can you ringlead today?  my report: http://fsck.com/~jesse/parrot.xls  can people pleasse send me thir lists of what cells they believe to be done
<obra> I'll compile
<pmichaud> obra: what do you want filled in -- i.e., who's responsible, completion rate, target date, ...?
<pmichaud> (have to run errand, back in 87)
<particle> pm: just initial with what's complete
<pmichaud> got it
% jonathan has joined #parrotsketch
% allison has joined #parrotsketch
<allison> hello?
<jonathan> 'lo
<leo_> hi @all
<chip> hi
<leo_> he - hi chip 
<allison> Looks like no jesse, so let's get started.
<allison> I'll paste in chromatic's report.
% particle has left particle!~particle@144.81.84.198
<allison> I fixed a couple of compiler warnings; that should be everything I've seen 
<allison> except for the GDBM hash warnings.
<allison> I think I've fixed Parrot::Embed to build at least on Unix platforms.  There 
<allison> might be a -L flag missing.  I don't know about Windows; reports would be 
<allison> nice.
<allison> I also added documentation to Parrot::Embed.  If it builds and passes tests, I 
<allison> can release it to the CPAN.
<allison> We have too many failing tests; we need to mark them as TODO or fix them in 
<allison> preparation for a release.
<allison> I'll try to talk to a guy at OOPSLA about Smalltalk on Parrot.
<allison> We'll go chip, leo, jonathan, me. (Let me know if I've missed someone.)
* pmichaud missed.
<chip> Good Day.  (OOPSLA, funny to know they're still around)
<chip> I have returned to live after an extended period of suspended animation.
<chip> That whole ZOmbie thing?  Not recommended.
<pmichaud> .oO( Too long in the matrix? )
<chip> I have a busy week ahead of me preparing for Chi-town.
<chip> . # EOR
% mdiep has joined #parrotsketch
<allison> okay, leo?
<leo_> * fixed some corner cases like accessing arrays by string index
<leo_> * not much else except contributions to various discussions on p2p or IRC
<leo_> ^D
<leo_> there are also some notes of obra in the scrollback I can paste
% particle has joined #parrotsketch
<allison> leo: okay, paste obra's notes
<leo_> < obra> My network here today turns out to hurt really badly.
<leo_> < obra> particle: can you ringlead today?  my report: http://fsck.com/~jesse/parrot.xls  can people pleasse send me thir lists of what cells they believe to be done
<leo_> .
* particle finally gets reconnected to irc.perl.org
<jonathan> OK, I guess it's me now.
<jonathan> What's news...
<jonathan> * Back from Spain in the early hours of Monday morning. While not having any time for Parrot stuff beyond reading the lists, I did read some interesting papers and find some other interesting ones to read, that may be helpful in Parrot/Perl 6 stuff.
<jonathan> * Yesterday finally made time to write and ci updates to the Bytecode PDD - it needs some further tweaks but I've hopefully dealt with the main issues brought up so far.
<jonathan> * As discused on list, started a document to outline what different languages need from Parrot's OOP support; I'm initially populating it with as much of the Perl 6 requirements as I understand, plus the .Net ones. Then I'll ci it and encourage other language implementers to add to it, in hope that when Allison gets to that PDD this helps her.
<jonathan> What's next...
<jonathan> * I proposed ages ago a couple of PMCs, I think for referencing an array element and an attribute in a class. .Net needs them (my implementation has some code for this too) and leo said other languages too. Nobody commented on my PDD patch sent to the list; I'll take that to mean nobody majorly objects and Just Write the Darn Code.
<jonathan> * Otherwise, looking for coding tasks. If anyone is waiting on anything that's spec'd and not implemented, I'm happy to have a crack. Otherwise, I'll be hacking on .Net translator and seeing if there are any RT tickets I can handle.
<jonathan> .end
<pmichaud> I think I'm next, so...
<pmichaud> * Worked on namespaces in PGE, sent messages to p2p
<pmichaud> * Several parts of PGE won't be able to be pdd21 compliant until the HLL classname issues are resolved
<pmichaud> * The HLL classname issues will also pose medium-term challenges for the perl6 implementation, since there are classname conflicts between Perl 6 and Parrot
<pmichaud> * Reviewed latest versions of pdds in docs/pdds/
<pmichaud> * All of those looked pretty good to me... the first features of those pdds that I'm likely to want to have available (if they aren't already) are the opcodes in pdd23 (exceptions)
<pmichaud> * I'll be sending a detailed list to p2p
<pmichaud> * Next on the hitlist will likely be I/O
<pmichaud> * To me, I/O is currently far more important than threading, although I recognize they may be related
<pmichaud> * On the other hand, perl6 can work with the existing IO features
<pmichaud> * Since PGE ns support is stalled, I'm now working on PAST
<pmichaud> * Due to a conflict on Sat 11/11, I won't be able to make Chicago hackathon
<pmichaud> * However, I've already cleared my schedule such that I'll be on IRC/hacking perl6/Parrot all day for at least 11/10, 11/12, and likely 11/13
<pmichaud> * So, I'll be virtually present at the hackathon
<pmichaud> EOR
<allison> okay, me, particle, mdiep (let me know if anyone else has joined)
<allison> I had productive time on the concurrency PDD.
<allison> And several great conversations with chromatic on the OO PDd.
<allison> I'm going to dump my parrot-internals folder in an archive and start over.
<allison> That should make it more possible to keep up.
<allison> (But if you have any urgent questions still lingering, bring them up gradually over the next few weeks.)
<allison> EOR
<particle> ~ added more smartlink functionality--it is almost complete now (just need to merge data streams and emit html)
<particle> ~ coding standard tests and code file cleanups
<particle> ~ discussed cage cleaner wrangling tasks with allison
% smash has joined #parrotsketch
<particle> btw re: jesse's spreadsheet--just initial the areas you consider 'done' and email to him
<particle> .end
<mdiep>  - worked on exhaustive ResizableStringArray tests and checked them in today, along with some improvements to ResizableStringArray itself.
<particle> mdiep++
<mdiep>  - discussed .HLL + classes on the list
<mdiep> ^D
<particle> i think allison needs to run off soon, so, any questions for her?
<particle> okay, any other questions?
<allison> I have a question for jesse, maybe someone else can answer: what's "jg" in the spreadsheet?
<pmichaud> allison: when dumping parrot-internals, please keep HLL classnames in mind
<particle> jg is my initials
<mdiep> that spreadsheet is much larger than I was anticipating
<allison> pmichaud: I'm only archiving it, not trashing it
<particle> okay, why don't i discuss the spreadsheet jesse created for a bit, after allison and pmichaud are done
<allison> pmichaud: you mean classnames and HLL namespaces?
<pmichaud> allison: okay, I'll rephrase :)   HLL classnames are a semi-urgent question for me
<allison> I'll pull that thread back into the "current" box
<pmichaud> until that's resolved I'm having to do a lot of workarounds or things that don't follow the pdd21 spec
<allison> pmichaud: okay, I'll dive back into that thread this afternoon (thanks for the heads up)
<chip> that's a problem with a known answer, albeit one that isn't implemented [mea culpa]
<chip> assuming you mean the problem of two HLLs both of which have classes named "Foo"
<pmichaud> chip: or even an HLL that has a classname that matches a parrot classname
<allison> chip: particularly, parrot and an HLL that both have classes named Foo
<chip> yeah, in this context, ['parrot'] counts as an HL
<particle> pm: did you commit a test for this?
<pmichaud> particle:  no, if only because I'm not certain what the answer is
<allison> it's pretty clear to me that the various  languages should have isolated namespaces, but there's some question about how accessible the 'parrot' namespace should be from the various HLLs
<allison> I lean toward saying it's just as isolated as the other namespaces, but that leads to some pretty clunky code
<pmichaud> I'm not so sure it's all that clunky
<leo_> namespaces *are* isolated, classnames aren't
<particle> it may warrant a special op or macro, but in't not that clunky
<pmichaud> especially if there's a good exporting mechanism
<allison> well, classnames should follow the same pattern as namespaces
<pmichaud> I'd be very happy with saying that parrot classnames are isolated
<allison> whatever pattern we decide that is
<leo_> well and that specs are just missing
% chromatic has joined #parrotsketch
<leo_> i.e. a class isa namespace # .t or .f ?
<chip> leo_: .f now
<leo_> I know
<allison> pmichaud: let's call that the quick answer, and I'll go through the mailing list messages
<pmichaud> I think that if we go with the notion that one can get to the built-in Parrot classes using .ClassXYZ constants, then it's not clunky at all
<chip> I mean, .f is the decision now for the future
<pmichaud> then 'ClassXYZ' and ['ClassXYZ'] are treated like namespace items
<chip> classes need to be registered anyway, which means namespace manipulation can never be enough to make a new class, so we may as well keep the current separation of metaobject in hidden hash and namespace out in public
<mdiep> maybe this can be added to the tasks list for Chicago?
<allison> good idea, mdiep
<particle> pm: can it wait that long?
<chip> mdiep++ # if I haven't done it by then ## no outbursts from the peanut gallery
<particle> s/it/you/
* particle offers chip some BRAINZ
<pmichaud> particle:  well, I'd prefer it sooner, but I don't feel like my track record warrants pushing the point :-)
<pmichaud> so, I'd just say that if it's not done by hackathon, that becomes a big priority for the hackathon
<chip> +1
<particle> ok, great. let's move on...
<allison> I have to go, but I'd like to see particle's discussion of jesse's spreadsheet. I'll read the archive.
<particle> cya allison
<allison> cya
<pmichaud> bye, allison
<chip> ciao
% allison has left allison!~chatzilla@216-31-236-219.static-ip.telepacific.net
<particle> last week, i asked if we had any milestones that we're working towards. jesse thought we should, and that the current milestone doc was way out of date. he put together a spreadsheet of parrot subsystems, and related items.
<particle> this is a first cut, where core committers should put their initials next to items they consider DONE
<particle> it will be given to the design team for review and comment
<pmichaud> I think I would like to see a transpose of the bottom half of the table
<particle> we'll probably all end up discussing finer points via email and possibly at #parrotsketch in the future
<pmichaud> the feature list is likely to be longer than the "completion stage" list
<pmichaud> so, put the features on the vertical axis, and the requirements/api/stub/subsystem/etc on the horizontal
<particle> ok, great, and i agree. i'll make sure jesse gets that feedback. additionally, you can send him comments in the email with your response
<pmichaud> this also makes it easier for many of the perl6 items at the top of the page
<leo_> I'd prefer a patch compatible text format
<chromatic> How about "complete as specified, but fairly broken"?
<particle> ok, i don't want to hold up the feedback too much. do you think these formatting changes could be done within a day, and the link resent out?
<particle> or should we just move forward with current format, and reformat later?
<pmichaud> I could make a text format pretty quickly, I think
<pmichaud> I'd say move forward with current format, and we'll backport in changes as needed.  I don't think there will be many
<particle> ok, great. any other questions/comments about that spreadsheet?
<particle> if nothing immediate, you can address them to jesse via email.
<particle> next: chip, you've been missed.
* pmichaud celebrates chip's return
<mdiep> particle: can you elaborate what is meant by "Stub functional (spec based) tests" and "Subsystem stubs"?
<particle> mdiep: sure
<particle> one level of basic testing is... did you test the spec?
% chromatic has left chromatic!~chromatic@ip150.fa1-0-1.occ.iinet.com
<particle> if the test is adequately tested, then it can be marked done
<particle> subsystem stubs i'm less comfortable explaining
<mdiep> but how does "stub" functional tests compare to the "complete" tests?
<particle> missing edge cases, etc
<particle> invald input
<mdiep> so the stub tests is just the basics?
<particle> yep
<mdiep> okay
<particle> please ask jesse about subsystem stubs
<particle> ok. this is the part where we get to tell chip WHY he's been missed.
<particle> anyone have rusty nails that need the mighty hammering power chip's steady hand provides?
<obra> Subsystem stubs: incomplete implementations, often in advance of not-complete specs
* obra is on _dialup_ with satellite return
<obra> this is pretty special
<particle> thx, obra!
<obra> thanks for dealing today, particle
<chip> particle: thanks [on two counts]
% chromatic has joined #parrotsketch
<particle> ok, no rusty nails?
<particle> chip: have you run 'make test' lately?
<particle> there are quite a few failing tests. some of us have been working towards release prep in your absence. your guidance as to which tests should be passing, and which todo/skip would be appreciated
<particle> oh, and of course a target release date, when you have one.
<leo_> 32/6685 subtests failed, 99.52% okay.
<leo_> x86_64 linux
<particle> ~96 tests fail on win32
<chromatic> Linux/PPC is around those numbers.
<chromatic> Some of the benchmark tests fail as there's no PPC float JIT, but most of the other failures have RT tickets or are sprintf problems.
* tewk wants to help out with object spec, I'm trying to get objects working in cardinal ( Sorry  I'm Late )
<chip> I have, & agree
<particle> oh, i have a feeling next week's meeting will be time-shifted
<particle> ...for most of us, anyway.
<coke> ~~, belatedly
<pmichaud> we currently meet at 1800 UTC -- same time next week?
<particle> pmichaud: yes, unless otherwise directed by obra
<coke> might be worth doing an audio call at some point, perhaps with less frequency.
<pmichaud> okay.  I will likely be 10 minutes late on some weeks, due to an errand I have to run each tuesday at 1145
<pmichaud> (localtime, 1745 UTC)
* mdiep tries to figure out if that means #parrotsketch is "earlier" or "later"
<pmichaud> mdiep:  you're in CDT now?
<particle> mdiep: fall back
<pmichaud> 1800 UTC == 1300 CDT == 1200 CST
* mdiep is in EST, but the same applies. thanks.
<pmichaud> 1800 UTC = 1400 EDT == 1300 EST
<particle> okay, great, looks like we're done. see you all in a week
<pmichaud> I'll post my draft text version of spreadsheet to #parrot in a few mins
<pmichaud> see you all next week
<jonathan> cya all
% chromatic has left chromatic!~chromatic@ip150.fa1-0-1.occ.iinet.com
<leo_> see ya
% coke has left #parrotsketch
% pmichaud has left #parrotsketch
% jonathan has left #parrotsketch
% mdiep_ has joined #parrotsketch
% mdiep has left mdiep!~matt@bursley-219-72.reshall.umich.edu
% smash has left smash!~smash@gil.di.uminho.pt
% smash has joined #parrotsketch