% integral has joined #parrotsketch
% PerlJam has joined #parrotsketch
% pmichaud has joined #parrotsketch
% kjs has joined #parrotsketch
% kjs has left kjs!~IceChat7@221.221.243.101
% smash has joined #parrotsketch
% allison has joined #parrotsketch
% Coke has joined #parrotsketch
% barney has joined #parrotsketch
% chromatic has joined #parrotsketch
% chromatic_ has joined #parrotsketch
<chromatic_> Shall we begin?
% chromatic has left chromatic!~chromatic@sub17-30.member.dsl-only.net
% chromatic_ is now known as chromatic
<leo> greetings to all
<chromatic> allison?
<allison> - Mainly worked on finishing off the new object model.
% jhorwitz has joined #parrotsketch
<allison> - Revamped the implementation of PMCProxy: filling in missing behavior, using the same core struct as Class, got inheritance from low-level PMCs by high-level classes working.
<allison> - Fixed the generated default vtable functions in Object, so they work.
<allison> - Added oo.h, oo_private.h, and oo.c, the home of the public and private interface functions for the OO subsystem (these will replace objects.h, objects.c, and have already replaced src/pmc/classobject.h).
<allison> - Added a variant of 'typeof' that returns the class object.
<allison> - More fully implemented type IDs in the new metamodel for backwards compatibility with the MMD system.
<allison> - Updated tests for the new metamodel.
<allison> EOR
<particle> hi there leo, glad to have you here
<chromatic> barney?
<barney> nothing to report
<chromatic> I threw in a small optimization for regexes, but there may be one more fix there.
<chromatic> I also helped Allison get some of the PDD15OO tests passing.
<chromatic> jhorwitz?
<jhorwitz> fixed parallel make problems, but with a small race condition. 
<jhorwitz> i rewrote c2str.pl to work better in a parallel environment, but working around makefile issues right now
<jhorwitz> need to finish up my patch to past-pm for new quoted types (i assume nobody else is working on this ATM).  this will keep mod_perl6 working.
<jhorwitz> EOR
<chromatic> leo?
<leo> Hi all - greetinx
<leo> I just svn up'd parrot a few minutes ago and did `make -j2 all` on my AMD X2 - Woot it works and it's *speedy*
<leo> really good job - I was badly looking for that in the old days - thx all - this definitely will help devel folks
<leo> EOR
<chromatic> -j 3 or higher should be even faster.
<chromatic> mdiep?
<particle> -j for maximum speed
<leo> thx
<chromatic> particle?
<particle> helping out where i can during this past week of light parrot involvement
<particle> ~ working with perldreamer++ on the unified testing branch, which is closer to done
<particle> ~ some help on pdd1500 branch for allison++, i have some pending patches i hope to submit today/tomorrow
<particle> ~ keeping in touch with tbb folks
<particle> ~ tim o'reilly has blogged about concurrency and parrot (http://radar.oreilly.com/archives/2007/09/parrot_and_mult.html) -- andy++ is interested in interviewing us about that topic for perlbuzz
<chromatic> pmichaud?
<pmichaud> kids' school activities and weekend guests ended up consuming most of my weekend :-(
<pmichaud> However, I worked on the pdd15oo activities from the PGE side of things, and also have the outline of the AST draft done.  I expect to check it in later tonight.
<pmichaud> (the outline is actually more complete than the word "outline" implies -- it's really more of a draft)
<allison> pmichaud++
<pmichaud> EOR.  I'm queuing one question.
<chromatic> smash?
<smash> no report, still trying to find time to finish pdd19.. EOR
<chromatic> tewk?
<chromatic> Okay, let's go to question time.
<chromatic> pmichaud?
<pmichaud> I'm just re-raising the question I posed to the m/l this morning -- can we get attributes to return null for uninitialized values instead of Undef ?
<allison> just saw that in the irc log
<allison> I also prefer PMCNULL to Undef, because it allows the return check to make a distinction between elements that don't exist and elements that have an undef value
<allison> it's a one-line change to ResizablePMCArray
<pmichaud> also, for reference and more detail, my post from July:  http://www.nntp.perl.org/group/perl.perl6.internals/2007/07/msg39161.html
<pmichaud> note that there are a lot of pieces of code that assume that ResizablePMCArray returns undef for nonexistent elements, however
<allison> but, will likely break other code
<pmichaud> in fact, returning "None" used to be the default for non-existent values of Hash, but that was changed some time ago
<allison> yes
<pmichaud> can we just have get_attribute return null but leave ResizablePMCArray alone?
<allison> at the time of the original discussion, I decided to make the spec match the current behavior
<allison> we can, by explictly checking 'exists'
<allison> before returning the value from the attribute lookup
<allison> but I think array is broken
<pmichaud> anyway, my preference is that getattribute return null for uninitialized attributes.  I don't have a strong opinion about ResizablePMCArray, other than to note that a lot of code expects uninitialized elements to return Undef
<allison> it's actually creating an undef PMC and inserting it into the array every time you do a lookup on an element that doesn't exist
<particle> can we make either change without deprecation?
<allison> hash returns PMCNULL
<particle> i say getattribute, yes, rpmca no
<pmichaud> having getattribute return null is the current trunk behavior -- pdd15oo changes it
<chromatic> I'm testing the ResizablePMCArray change right now.
<pmichaud> so, getting getattribute to return null in pdd15oo doesn't require a deprecation cycle
<particle> pmichaud needs getattribute to get pge building again
<pmichaud> I just need to know what the answer will be for getattribute :-)
<allison> if we can make the array change in the branch, and fix all the failing tests that result from it, that's my preference
<pmichaud> I can work with either version of getattribute -- I just don't want to have to go through and switch back later on if I can avoid it.  And returning null is more consistent with the behavior of all other symbol lookups
<allison> but I'm happy to stick a temporary hack into class attributes, and then change resizablepmcarray on a standard deprecation cycle
<allison> either way, undefined attributes will return PMCNULL
<pmichaud> outstanding.  allison++
<pmichaud> when that's done, it should be very quick to get the rest of pge working in pdd15oo, as I think all of the pieces are in place
<pmichaud> it may be difficult to get the other hll's working with PGE, because of the way hll namespaces are organized, but we'll cross that bridge after pge works again :-)
<chromatic> I'm only seeing a couple of test failures for the rpmca change.  It looks easy to fix.
<chromatic> Any other questions?
<chromatic> I have one: can I get some Windows testing for the c2str.pl patch I suggested which fixes the race condition?
<pmichaud> (I expect most of the past-pm, pct, and nqp stuff to fail as a result of the rpmca change -- this includes languages such as abc and perl6 that depend on past-pm)
<particle> chromatic: it works. there's no -j equivalent on windows, so no problem with the code
<particle> (no -j option for nmake)
<chromatic> Okay, then expect a commit there very quickly.
<chromatic> Is anyone blocking on anything that we can easily fix?
<pmichaud> I'm only blocking on time, and as of this morning that appears to have been fixed for this week :-)
<pmichaud> also, forgotten from my report:  I've created a temporary branch for the pct/nqp refactor changes I'm making this week
<pmichaud> in branches/pct-devel
<pmichaud> I'm guessing we're done?
<chromatic> Looks like it.
<chromatic> Talk to you all next week.
<pmichaud> see you next week!
<particle> ta
% chromatic has left #parrotsketch
% jhorwitz has left #parrotsketch
% allison has left #parrotsketch
% barney has left barney!~bernhard@p549A2461.dip0.t-ipconnect.de
% Coke has left #parrotsketch
% particl1 has joined #parrotsketch
% particle has left particle!~particle@c-24-22-165-145.hsd1.mn.comcast.net
% tewk has left tewk!~tewk@ekstrom.org
% particl1 has left particl1!~particle@c-24-22-165-145.hsd1.mn.comcast.net
% particle has joined #parrotsketch
% particle has left particle!~particle@c-24-22-165-145.hsd1.mn.comcast.net
% particle has joined #parrotsketch