The Perl 6 design team met by phone on 24 September 2008. Larry, Patrick, Jerry, Nicholas, and chromatic attended.
Larry:
- spent most of the week hacking
- neither golf nor computers
- haven't had much brain to think about anything other than work
- made a few comments on IRC
- thinking that the direction I need to head now is turning the standard grammar into a Cfront for Perl 5
- get rid of
gimme5 by taking the standard parse tree
- use that instead of the faked up parse
- spit out the same thing
- then I can translate some subset of Perl 6 back to Perl 5
- then I can think about rewriting other things
- Cursor.pm in Perl 6
- getting ready to go to a conference in Urbana in a week and a half
- ACM-ish thing called Reflections Projection
- giving two talks there
- most of my discretionary time is getting my video output suitable for running Firefox and OpenOffice.org Impress
Patrick:
- not a lot of actual code
- lots of design work and shepherding other code
- getting container design right in Rakudo
- Jonathan and I talked about it
- won't require major Parrot rework
- plan to start implementing that in the next week or so
- will solve some problems with array, hash, and object references
- Moritz has been working like crazy on spec tests
- finding things that Rakudo will pass
- we pass well over 3700 tests
- that's 400 more than last week
- over 200 yesterday alone
- few of these tests are really changes to Rakudo itself
- have about five new contributors to Rakudo in the past two weeks
- spent Monday rethinking how lexicals should work in Parrot
- that's blocking a lot of things in Rakudo
- re-read S04 and various threads on the Parrot lists
- wrote a draft of how I think things ought to work
- awaiting feedback on that
- the more I think about it, the more I think it's the way things ought to go
- ought to simplify things and get rid of a lot of ugly code in Parrot
- unless I'm overlooking things
- might try implementing it myself in a branch unless someone (hint, chromatic) wants to do it instead
- not a lot of changes to PIR itself
- seems to work pretty well in all of the mental cases I've thrown at it thus far
Jerry:
- setting travel plans for fall conferences and meetings
- patched a few HLLs today
- someone added SMOP to the #perl6 eval bot
Nicholas:
- Jesse's going to the Beijing Perl Workshop to talk about Perl 6
- they found a sponsor for him
c:
- fixed Parrot's float output precision
- makes test output different in some cases, so it's a big patch there
- want to get testing on Cygwin, Darwin, and MSVC
- it's a well-understood and well-supported part of
sprintf
- also looking at Patrick's lexicals proposal
- had some clarifications and questions
- will get that out tonight
- one question is about an enclosing lexical scope for multiple globally-visible subs
- initialization time is tricky
Patrick:
- I explicitly disclaimed that for now
- I'll look at S04 and take examples from there
c:
- also had a question about recursion
- that can get tricky
Patrick:
- I had that thought too
- I think it works
c:
- I'll definitely send my thoughts
- want to make sure they get in the archives
Nicholas:
- is a Cfront to STD.pm another bootstrap to get Perl 6 running on Perl 5?
- or does it only bootstrap the parser, not the runtime?
Larry:
- you have to have a runtime, or you can't bootstrap
- it just uses Perl 5 for the runtime
- that set of semantics which are easy to do in Perl 5
- or at least possible
- it won't easily do all of Perl 6 semantics
- including lazy lists and such
- not without internal hacking to the regex engine
Nicholas:
- the first thing I thought about was continuations, which Perl 5 won't be doing
c:
Larry:
- I'd settle for lazy lists
Patrick:
- I'm looking forward to a
Cursor rewritten in Perl 6
- Perl 6 implementations make it easier to bootstrap on whatever implementation we have
