% particl1 has left particl1!~particle@c-24-22-165-145.hsd1.mn.comcast.net
% particle has joined #parrotsketch
% mdiep_ has left mdiep_!~mdiep@c-71-205-39-103.hsd1.mi.comcast.net
% kjs has joined #parrotsketch
% pmichaud has joined #parrotsketch
% jhorwitz has joined #parrotsketch
% kjs has left kjs!~IceChat7@221.218.214.101
% Andy has joined #parrotsketch
<Andy> Are we not sketching?
% jisom has joined #parrotsketch
<pmichaud> I'm here to sketch
<particle> i'll take the helm today, since coke and chromatic seem absent. who's here to report?
<jhorwitz> present
<particle> it's a light week due to yapc::eu
<particle> okay, then. andy, wanna start things up?
<Andy> OK
<Andy> EOR
<Andy> Next?
<particle> jhorwitz: what's new?
<jhorwitz> a plethora of updates to mod_parrot to remove deprecated ops
<jhorwitz> and more importantly migrate to pdd15 objects
<jhorwitz> all tests pass once again
<jhorwitz> looking forward to nqp enhancements
<jhorwitz> EOR
<particle> okay, i'll go
<particle> ~ some cleanup commits this week, more reading than anything else
<particle> ~ reading on tree ssa form, thinking about optimizing passes for pct
<particle> ~ #perl6 now has an evalbot for pugs, kp6, nqp, and perl6 (moritz++ and rhr++)
<particle> ~ considering a LOGO implementation once we have hll->nci bindings
<particle> .end
<particle> pmichaud?
<pmichaud> I didn't pre-type my report today, so on-the-fly
<particle> we've got time :)
<pmichaud> * Issued the 0.4.15 release, that went fairly smoothly
<pmichaud> * More updates to NQP, many from perlDreamer++
<pmichaud> * I've been working on handling namespaces in PAST/POST/NQP
<pmichaud> * Also have been explaining why the 'return' statement has to be handled as an exception :-)
<pmichaud> EOR
<particle> anybody not having fun?
<particle> anybody blocking?
<pmichaud> my only block is energy.  :-(
<particle> we'll try to move you somewhere hotter :)
<pmichaud> heh, no thanks.  That's probably what has been zapping my energy :-)
<particle> okay. comment/question time. anyone?
<jisom> I upgraded my system so my dvd drive doesn't work and I broke debuggging, and jit
<jisom> s/doesn't/does/
<pmichaud> which jit?
<jisom> amd64, just changed my system software, not parrot
<particle> odd.
<pmichaud> well, 64-bit jit hasn't worked for me yet
<pmichaud> (but I'm intel)
<particle> i'd *love* to get 64bit parrot working on win32
<particle> still having miniparrot segfaults
<particle> then i can help with 64bit jit
<jisom> tried more than one compiler?
<particle> no, i haven't. i dunno about other 64bit compilers for win32
<particle> just msvc pro (can't use the free one for 64bit)
<particle> pmichaud: would implementing warnings be any easier than implementing returns? i know they're both exceptions, but the handlers are in a different place
<pmichaud> particle: what sort of warnings?
<particle> eg. deprecated perl 5 syntax
<pmichaud> you mean at a compile-time level?  isn't that just  <panic: ...?>
<particle> basically a resumable exception
<particle> i don't want 'die' behavior.
<pmichaud> what behavior do you want?  You want to be able to throw/catch an exception?
<particle> like perl 5's use warnings
<particle> so, yes.
<pmichaud> I'm lost -- I don't see the relationship with returns
<particle> if we can enable/disable warnings lexically, it's a step closer to returns in my mind (but perhaps i'm incorrect)
<particle> { no warnings 'foo'; generate_foo_warning; } say "ok";
<particle> { no warnings 'foo'; } generate_foo_warning; say "ok";
<particle> tests like that
<pmichaud> I still don't understand... but perhaps this answers your question:   the issue with returns/exceptions has nothing to do with the display of warning messages
<pmichaud> essentially, PAST has to have a way for a lexical sub to be able to catch an exception of type 'return'
<pmichaud> the exception payload has to have a return value, which the lexical sub can then return to its caller (via a Parrot .return statement)
<pmichaud> if a lexical sub catches an exception that isn't of type 'return', it has to re-throw it or pass it to its outer handler
<particle> PAST also has to have a way for a lexical scope to catch an exception of type 'warning'. the payload is a message, and execution resumes at the throw point.
<particle> i see in this some of the pieces needed for returns
<pmichaud> PAST already has the ability to do try/catch blocks
<particle> okay, so warnings are already implementable at the past level. cool.
<pmichaud> we just need to add an Exception class that can represent the various types of exceptions we'll have in PAST
<pmichaud> (off-the-cuff list:   return, warn, die, next, last, redo)
<particle> goto?
<particle> i think it's about time we extend the exception model to allow subs as handlers, rather than only labels
<particle> would that help here?
<pmichaud> not really -- labels works out easier for me
<particle> i guess i'm missing the bit that's hard. it's not the payload. it's not the throw.
<pmichaud> it's the api
<pmichaud> it's the defining of the exception class
<particle> ah, okay. now i see. i'd be happy to talk that out with you on #parrot or via email
<pmichaud> right -- I just want to get the namespaces stuff resolved first
<pmichaud> it's not particularly hard, just time consuming
<particle> np
<pmichaud> (and people's initial thought for implementing "return" is to try to translate it directly to a PIR .return statement, which doesn't work for most HLLs.  Although Coke repeatedly explains how Tcl does it the way it's supposed to be done :-)
<particle> oh, by way of update, parrot 64bit on win32 no longer segfaults with miniparrot. now it segfaults when building pge.
<particle> yes, tcl uses exceptions for return, so there's probably something useful there :)
<particle> oh, good to see ptc working on coverity. i need to get an account, so i can do more than watch.
<particle> anything else this week?
<particle> if not, see you next week.
<pmichaud> see you all next week
<pmichaud> or on #parrot
% jhorwitz has left #parrotsketch
% jisom has left jisom!~jisom@74-134-230-123.dhcp.insightbb.com
% Coke has joined #parrotsketch
% Andy has left #parrotsketch
% Coke has left #parrotsketch
% PerlJam has left PerlJam!duff@feather.perl6.nl
% leo has left leo!lt@feather.perl6.nl
% pmichaud has left pmichaud!pmichaud@feather.perl6.nl
% tewk has left tewk!~tewk@72.8.94.63
% tewk has joined #parrotsketch
% contingencyplan has joined #parrotsketch