% particle has left particle!~particle@144.81.84.203
% particle has joined #parrotsketch
% allison has joined #parrotsketch
% pmichaud has joined #parrotsketch
<obra> hi gang!
<obra> Leo, how are you feeling?
<leo> hi @all
<leo> better
<leo> much better
<obra> Excellent
<chip> yo dude, very happy to hear that.  takes a big hammer to put somebody down for weeks
<obra> So glad to hear it
<particle> hi
<obra> (Leo, you're up first)
<leo> nothing to report ;)
<obra> Fast and easy!
<obra> Chip?
<leo> except for a summary:
<leo> <@obra> anything you want to see done in your abcess?
<leo> < particle_> then he had another hole drilled in his head
<leo> EOR
<leo> (all mentioned on irc already ;)
<obra> *nod* chip. how's things?
% chromatic has joined #parrotsketch
<obra> Failing chip, Patrick how's it going?
<chip> I had a network error
<chip> back now
* pmichaud falls back to chip
<obra> ok :)
<pmichaud> (I'm still composing my report )
<chip> Well, I have bad and good news.  The bad news is that up through this morning, the status was unchanged.  My TODO remains as it was: pending minor cleanups in namespaces, major PDD list from Allison etc.
<chip> The good news is that today is shaping up to be a good Parrot Day for taking a bite out of crime ^W that list.
<chip> Also, I've settled on an architecture criterion that will speed things - making the system be what I want to use as a programmer.
<chip> Mind you, this is the notional 'me', in the place of and doing the tasks that others are doing.
<chip> (i.e. I know I'm not the sole use case.  <grin>)
<pmichaud> chip:  with a lot of work, you could be.  :-)
<chip> pmichaud: a worthy goal to aspire to :-)
<chip> }
<obra> What's your next week look like? Who should we be hassling for you?
<chip> Well, if somebody could take the ball on the YAPC talk scheduling and run with it, that'd be a TODO off hte list
<obra> YAPC talk scheduling?
<chip> Also, some more reminders about verifying RT bugs might shake loose a volunteer or two
<obra> we're sending out weekly rt reminders now
<particle> (weekly reminders)++
<chip> obra: My talks fell through the cracks for YAPC scheduling, so four intrepid Parroteers are giving up ten minutes each so I can speak on Wednesday AM
<chip> weekly reminders work for me
<pmichaud> chip:  (yapc talks)  how about you just propose a revised schedule?  I'm pretty flexible.  Or, lacking that, I say just take your 40 minutes at the front of the block and we'll keep our ordering after that
<chip> if patrick/allison/particle/coke coordinate with the YAPC guys and give them the new start/end times that'd be helpful
<chip> the "start" suggestion is great
<obra> I'll poke josh
<particle> agreed.
<pmichaud> I'll forward a schedule on that basis, then
<chip> thank you
<obra> or ok. go for it, patrick
<pmichaud> cc's to all concerned
<chip> that's it for the bug(ging) list
<obra> ok.
<obra> patrick: want more composition time?
<pmichaud> no, I'm ready now :-)
<pmichaud> so, this week...
<pmichaud> perl6: Added lexical and package scalars
<pmichaud> perl6: Added support for numeric literals (0x, 0b, 0d, etc.) and string conversions of those
<pmichaud> perl6: Added a --dump option to make it easier to see the translation phases
<pmichaud> pgc: Added support for 'rule' and 'token' productions
<pmichaud> pge: Added initial support for :pos, :cont, and :ratchet
<pmichaud> pge: Added some more regex code gen optimizations
<pmichaud> pheme: Worked with chromatic to find out why pheme fails on x86, reported as #39072
<pmichaud> those are accomplishments -- current buggy things
<pmichaud> with some of the recent changes perl6 seems to be tickling a gc bug somewhere; it shows up in various tests at odd times, but running with --no-gc or with perl6's --dump option seems to circumvent it (which is why I suspect gc)
<allison> I have a question for patrick when we get to questions.
<pmichaud> there are a few bugs relating to named parameter passing (w/ tests in t/op/calling.t), just wanted to note them so leo can get to them when he gets some tuits
<pmichaud> in particular, I'm curious if parrot has any problems with passing empty hashes as :flat :named, because that seems to cause an occasional segfault
<pmichaud> (but I haven't got a good test case for it yet)
<chip> pmichaud: shouldn't, of course, but if so it's a regular bug
<pmichaud> #39072 (load_bytecode with :multi) seems to be particularly troubling on the horizon
<pmichaud> in doing some of the perl6 work this last week I tried setting it up as a .HLL, and PGE just doesn't want to play that way (yet).  So, some of the PGE changes this week have been to try to resolve that and understand the use-case scenarios better so I can write something up for chip/p6i
<pmichaud> <cut>, but I can backtrack to allison's and others' questions anytime :-)
<obra> ok. what's next for you?
<pmichaud> perl6 subs if I can manage it
<pmichaud> that's about it (should be enough for the week)
<obra> Sounds great.
<obra> Allison? How's tricks?
<allison> I'm still on refactoring. Patrick and I have been bouncing ideas back and forth on PAST/POST, gradually moving toward convergence.
<allison> hopefully I'll wrap that up this week (at least, as much as we're going to wrap at the moment)
<allison> EOR
<obra> Cool. Convergence is definitely a plus :)
<obra> particle: what's new and exciting?
<particle> been studying the compiler tools
<particle> and (re)implementing perl6 operators
<pmichaud> particle++
<particle> i've seen quite a few questions from non-coms on #parrot lately
<particle> been trying to answer them
<particle> also, there's more crosstalk between #parrot and #perl6 now, which is great
<obra> Woot
<particle> and from the fun department, i'd like to add that it's great to be working on Perl6. i'm really excited.
<particle> i have some tasks, like adding to the test suite for perl6, as well as apl
<obra> How close is the p6c to being able to run pugs smokes?
<pmichaud> need subs :-)
<obra> right. thanks :)
<particle> EOR
<obra> chromatic: how's things?
<chromatic> Pheme is blocking somewhat on the multi/PBC bug (#39072).
<chromatic> In lieu of that, I released P5NCI to the CPAN.
<obra> Woot.
<chromatic> TPF received a Summer of Code proposal to work on that, so I hope we can accept the proposal.
<chromatic> C-c C-x
<obra> ok. question time. Allison, you had one?
<allison> yup, question for patrick:
<allison> What prompted the change from 'rule' to 'regex'?
<pmichaud> S05
<pmichaud> what we used to call 'rule' is now 'regex'
<allison> Okay, so my question goes to p6l. Thanks.
<pmichaud> 'rule' is now 'regex :ratchet :words'
<pmichaud> well, I can answer the reasoning behind the change, too :-)
<allison> (That's truly awful.)
<pmichaud> the idea is that we have regex, token, and rule
<pmichaud> regex is the traditional, always backtracking form of an expression (unless explicitly cut)
<pmichaud> token is a regex that defaults to non-backtracking (unless explicitly requested)
<allison> Okay, what are rules if they aren't the parts of a grammar?
<pmichaud> rule is a token (defaults to non-backtracking) that also assumes the :words flag is in effect (i.e., ws significant between components of the regex)
<pmichaud> all three can be parts of a grammar -- the only difference is the handling of backtracking and whitespace in the regex
<allison> The categories of distinction make sense, but the terminology is poorly chosen.
<allison> But, not really relevant here.
<pmichaud> I think that the idea is that "rule" harkens back to the BNF sense of rule
<allison> I'll take it to p6l
<chip> regex nurse :ratchet {...}  # sorry, it had to be written
<allison> :)
<pmichaud> okay.  It was also discussed on 6p a bit
<pmichaud> anyway, I had suggested   rule/token/phrase, but that didn't work out :-)
<allison> I don't see the thread, but then I've just switched email clients, which makes searching interesting.
<pmichaud> searching
<obra> (Are there other questions today?)
<chromatic> Where's Leo?  Is he okay?
<leo> I'm here and ok again
<chromatic> Good.
<pmichaud> (allison: I misspoke -- it apparently was on p6l and not 6p. sorry)
<pmichaud> (still getting the reference for you)
<allison> (patrick: ah, that makes more sense, I've just started reading p6l again with the new client)
<pmichaud> allison: http://www.nntp.perl.org/group/perl.perl6.language/25075 has damian's post (but there are replies on either side worth reading)
<obra> anything else today?
<obra> If not, I'll catch y'all next week
<chromatic> Thanks, Jesse.
<obra> thanks, all
<leo> thx jesse
<pmichaud> thanks, jesse!
<pmichaud> allison: btw, one can *still* use "rule" in a grammar, and that's the expected case
% chromatic has left #parrotsketch
<chip> thanks
<pmichaud> I simply switched everything from "rule" to "regex" to make sure things continued to work, figuring that the individual language implementors could then decide what level of backtracking/wordsplitting was needed
<pmichaud> i.e., in the perl6 grammar, I switched everything from rule to regex, and now I'm going through and replacing 'regex' with 'token' or 'rule' as appropraite
% pmichaud has left #parrotsketch
% allison has left #parrotsketch
% leo has left #parrotsketch
% particle is now known as practice
% practice is now known as particle
% particle_ has joined #parrotsketch
% particle has left particle!~particle@144.81.84.203