% cognominal_ has left cognominal_!~stef@cac94-1-82-67-232-89.fbx.proxad.net
% cognominal_ has joined #parrotsketch
% leo has joined #parrotsketch
% pmichaud has joined #parrotsketch
<obra> hi
<pmichaud> good afternoon
<mdiep> hello
<pmichaud> I wanted to check in quickly but cannot stay long -- I have a little girl at home from school today and we need to go out and get her something to eat
<obra> home sick?
<pmichaud> yes, something's wrong with her sinuses or ear today -- nothing serious
<obra> poor girl. :/
<obra> ok. want to blat out a quick update and then run off to take care of more important things?
<pmichaud> but she probably needs some hot chicken soup :-).  Anyway, my quick update for this week is that I'm continuing to work on PGE parsing and a perl 6 parser, and I'm following along with the various lexical pad implementations from chip and leo
<pmichaud> I'm not blocking on anything
<pmichaud> (except time :-)
<obra> unfortunately, I've done all I can to stretch time for you. A couple of weeks ago, I managed to get you an extra hour late one night. But I haven't managed anything since then
<pmichaud> and I made good use of that extra hour, let me tell you.  :-)
<obra>   heh
<mdiep> pmichaud: there's no TODO file yet, right?
<pmichaud> anyway, I'm also working on the updated draft of S05 and hope to release that before wednesday
<pmichaud> at present there's not a TODO file, I'll put that on my TODO list :-)
<mdiep> :-)
<pmichaud> (I'll have a TODO file released by wednesday also)
<obra> cool
<pmichaud> it'll be easier to do the TODO file once I have the S05 draft in good shape, since we can go through the S05 draft and see what's "to do"
<pmichaud> (the S05 is already in pretty good shape, I just want to reorder a couple of sections)
<pmichaud> all of this will get announced to p6c and p6i as I complete them
<obra> Nice!
<pmichaud> oh, and since last week I added a few more optimizations to the rules engine, so that it's smarter about skipping certain things that obviously cannot match
<obra> at what point does it start to make sense to put pge on a monthly release schedule to match parrot?
<pmichaud> we can do it starting this month if desired
<pmichaud> as it was it was nice that many of pge's newest features were completed at about the same time as the wart release
<obra> From my pointy-haired perspective, that'd be very cool.
<obra> it helps people get the word out that things are moving along
<obra> and it starts to give the pugs guys something to compete with ;)
<pmichaud> let's just plan on it that way, then, pge's release cycle will match parrots
<obra> does it make sense to have it a week later or so to give you time to tweak to match their final?
<pmichaud> no, not really -- the two are fairly closely tied anyway
<obra> ok
<obra> cool.
<pmichaud> since we know when the releases are occuring (beginning of each month) that's all the heads-up time that I need
<obra> I need to vanish for 120 seconds to feed a meter. when patrick is done, leo, you're up ;)
<obra> *
<obra> -
<obra> +
<leo> ?
<leo> anyway
<leo> print <<'REPORT'
<pmichaud> and I try to keep pge closely aligned to the parrot head/trunk.  End of report for me, I'll check on the logs later.
* pmichaud away &
<leo> - substantial rework of unicode charsets and encodings with a lot of simplifications: from 4 charset conversion interfaces to 1, from 2 enconding conversions interface to 1
<leo> - more code reuse: e.g. charset.unicode.to_charset => encodings.utf8.to_encoding
<leo> - new encodings: utf16, ucs2 including new encoding conversion functionality
<leo> - opcodes: downcase, upcase, titlecase, find_encoding, trans_encoding, ...
<leo> - and there are *lexicals* and *closures* according to pdd20 specs
<leo> - folks are currently improving config; beautifying & testing examples; and HLL debug segments
<leo> REPORT
<leo> end
<obra> Wow
<obra> that's quite a week
<leo> yeah
<mdiep> indeed. leo++
<obra> leo: whatcha blocking on?
<leo> nothing
<obra> you guys rock.
<obra> mdiep: what's up?
<mdiep> Tcl is once again passing 100% of its tests
<obra> 100% of the ParTCL test? How's it doing on the TCL test corpus?
<mdiep> obra: correct. it's been a while since we've run the actual tcl tests, but it's in the 10-15% range, I believe. we got sidetracked by the conversion to a compiler and we've been waiting for namespace support.
<obra> *nod*
<mdiep> leo was very helpful this week, tackling the unicode implementation as well as a couple GC bugs that were plaguing ParTcl
<mdiep> but we did have to cheat a little bit, because PGE::Glob doesn't escape unicode currently
<obra>  is this a bug or a design difference?
<leo> it's a tcl compiler issue honestly
<obra> heh.
<mdiep> obra: leo and I have different opinions on that. it depends on if I should be able to pass a utf16 string to PGE and have it generate code correctly.
<mdiep> I'm of the opinion that PGE should handle it.
<leo> a rule is code and code is ascii (with possibly escaped unicode) - but not utf16
<obra> code is ascii? even perl6?
<leo> as said on #parrot, please folks feed you favorite C compiler with utf16
<mdiep> but I'm also unsure of what I'm supposed to do on our end to fix it like leo is suggesting. (charsets/encodings make my head spin)
<leo> inside strings there can be more stuff, if it's declared properly, e.g. utf8:unicode:"string"
<mdiep> it seems like this will be an issue for any language using PGE.
<obra> ok. this sounds like it's likely a matter for the lists.
<leo> well, the tests have no problem with it and there are tests for e.g. unicode whitespace
<mdiep> leo: could you fill in an pir example for me if I create the skeleton?
<leo> sure
<mdiep> okay, I'll do that following the meeting
<mdiep> on the namespace front, I'm still waiting to hear from chip. he assures me that I'm high on his list of people he owes emails to.
<mdiep> ^D
<obra>   ok
<obra> cool
<obra> who else is here?
<obra> autrjus is jetlagged. nick is on the tube.
<mdiep> as a note, I'd be interested in seeing an explanation of allison's TGE (what it is, how it works, why I'd want to use it, etc)
<obra> ok. if it's just us today, I guess we're done for the day. I know allison is working hard to get TGE to the point where all that will make sense and be obvious ;)
<mdiep> excellent :-)
<obra> hang on a sec, guys? I may have managed to rope in one of our hooky-playing compatriots
% autrijus has joined #parrotsketch
<autrijus> yo
<leo> hi autrijus 
<mdiep> autrijus!
<obra> Autrijus! what's new?
<autrijus> finally fought of jetlag as of today
<autrijus> s/of/off/
<autrijus> did a lot of objspace designs and ways to utilize the new lexpad on paper
<autrijus> previously couldn't focus on the monitor without dozing off
<autrijus> today discovered iblech and luqui had been chasing larry's new batch of changes effectively anyway, so I didn't have to adapt
<autrijus> so should be able to go straight to codegen this week.
<autrijus> also, instead of pairing with dconway, next month I'll be staying at Alias's instead
<autrijus> so expecting more work on the p5integ direction than p6 design work, I think.
<autrijus> otherwise, nothing much.
<leo> autrijus: in a fresh parrot please check t/op/lexicals.t (first 20 tests - rest is old syntax still)
<leo> I hope it is sane
<autrijus> I took a look 1hr ago
<autrijus> will report insanity if and when I write codegen
<leo> great - thanks
<autrijus> so :outer is all done?
<leo> yes
<autrijus> wunderbar
<leo> just the .lex from pdd20 is missing
<leo> :lex - sorry
<autrijus> goodie, then I can stop using the apparently-still-not-very-specced label-as-closure trick
<autrijus> nod, but I'm not genning Tcl :)
<leo> if there is one .lex (or :outer) LexInfo and Pad are created
<autrijus> yup, I grok the spec
<autrijus> what do you think frame-as-obj? caller() in p6 can really use that
<leo> as said on a list there are serious GC issues with this approach imho 
<leo> s/a/the/
<autrijus> I've read the reply, but I think making them weak makes sense
<autrijus> when weakptr drops off, just raise exception
<autrijus> but it's not of neccessity at this moment -- I'll see what I can dig out from the codegen work
<leo> well we don't have weak PMCs yet
<leo> but that could help, yes
<autrijus> and any sane GC system need to have WeakPtr types
<autrijus> anyway, it's probably for later :)
<leo> yup
<autrijus> eof... hopefully didn't intrude too much
<obra> nope. we were4 basiclally done
<obra> anything else you're blocking on?
<autrijus> finally nothing, what a fix
<autrijus> :)
<autrijus> leo++ chip++
<obra> nice
<obra> ok.
<obra> I think we're set for today
<leo> thanks autrijus 
<autrijus> cool, thanks obra also for dragging me in
<obra> np. later, all
<autrijus> I might need prompting next week also, as my brain isn't adapting to 2am sketch yet
<autrijus> see ya all :)
<leo> see you all - bye
% autrijus has left #parrotsketch
% leo has left #parrotsketch
% pmichaud has left pmichaud!~chatzilla@c-24-1-26-255.hsd1.tx.comcast.net