The Perl 6 design team met by phone on 24 October 2007. Larry, Allison, Patrick, Jerry, Will, Jesse, and chromatic attended.
Larry:
- mostly under the weather this week for one reason or another
- most of my progress has been thinking, with not much to show for it
- difficult to type with tendinitis and one arm in a sling
- starting to think about spec tweaks
- nothing terribly major
- a bunch of pending feedback from Wolfgang Luan to handle
- just hanging out and being available for questions on IRC
- kp6 is now making more progress on hooking up to a Parrot backend
Jesse:
- who's doing the work on that?
Larry:
- they're just getting up to speed now
Jerry:
- Daniel Loca
- I've been in touch with him
- he's targeting PIR right now
- we're talking about doing transformations at the AST level
- that would involve a materialization of the AST
- we've talked to Ingy about doing that in YAML
- Patrick's AST PDD draft is a good starting point
- I consolidated a bunch of Parrot's runcore command-line options
- just need to decide how to deprecate the old options
- improved the parallel make performance too
- I put the process that took the longest first in the list, not last
- seem to be making one step forward and getting two TODOs
- I'm gearing up to help with progress toward Perl 6, but I have vacation in November
Jesse:
- what are you blocking on?
Jerry:
- the remainder of the AST implementation
- that's to work on Perl 6
- I'll also be working on the exceptions implementation with Kevin Tew
- he submitted a patch that Allison and I are looking at
Patrick:
- things are pretty good
- mostly cleaning things up and making progress on NQP and the AST representation
- cleanup with respect to PDD 15 and S05 over the past few months
- inside of NQP, I removed the notion of using
return to return the AST; it now uses make
- also wrote a preliminary implementation of protobjects
- it's a nice abstraction layer
- externally everything looks the same at that layer, even as we muck with the internals
- we can split things into namespaces and the like without affecting customers of that interface
- also talking about YAML representations of the AST
- that doesn't depend on PDD 26 from my perspective
- we just need a YAML representation where each node has a type, a hash, and an array
Allison:
- continued cleaning up a few last failing tests after the PDD 15 branch merge
- decided not to clean up some failures in prototyped systems that don't have specifications, so marked those as TODO
- continued removing old object code including ParrotObject and ParrotClass
- preparing for the 0.5.0 release
- started looking at the exceptions PDD
- aside from the stuff Kevin has done and adding the introspection interface, there are two things left: writing comprehensive tests and reviewing at the exception hierarchy
- started a discussion of the PIR PDD draft
- should be out of draft within the week
c:
- checked in my GC runcore after Jerry consolidated the argument options
- found several problems
- fixed most of those, but there are a couple of tough ones
- somehow we reclaimed most of Tcl, for example, but Will found another hairy problem there
Will:
- trying to find ways to remove bookkeeping that we need to manage
- mostly reclaimed Tcl
- plan to run language tests through this runcore after we get Tcl fixed up
- think we'll be able to find and fix more memory issues there
- probably some PMC-related stuff that's undertested in the core
Jesse:
- talking to the folks working on kp6 and the Lispish backends
- the performance and compliance are impressing me
- I'll drop Damian a line this afternoon
Jerry:
- someone asked on #perl6 about .NET developers interested in Perl 6 on .NET?
Larry:
- CJ Collier noted that there was some interest in getting Perl 6 running there
- he interrogated me as to some of the difficulties I saw there
- he forwarded them Eric Lippert
- mostly Eric agreed to my concerns
- I'm willing to help if someone else drives it, but I won't code it myself
Jerry:
- that VM has some challenges to overcome to implement Perl 6
Larry:
- if there's interest from within Microsoft in doing that port, I don't think they should be discouraged
Jerry:
- I'm happy to meet with them in person if they want
Jesse:
- I'm sure Jonathan would have interesting input as well
Allison:
- which languages are working after the merge?
c:
- Pheme, but I also tried Tcl, Lua, and Perl 6
Jerry:
- APL and Punie are working
Patrick:
- NQP is working
- I figure I'll end up redoing Perl 6 in NQP in a few weeks
- we'll have a working version somehow before 0.5.0
Will:
- we can just run
make test and get that working
Jerry:
- we should work on getting all languages back to their previous state before 0.5.0
Will:
- I'll start a migration guide on the wiki
Patrick:
- the compiler toolkit change is going to protobjects which solves a lot of problems
- I haven't looked at PAST-pm
- if someone looks at that or wants my help, that's fine
- I don't think it would take long to fix it
Allison:
- we might just migrate them to PCT anyway
Patrick:
- it's time to start migrating things
- I'm trying to get PCT running so people can migrate to it
- if we're within a week of release date and the migration doesn't look solid, I'll backport things so that the other languages can work
- so far the object model seems to be working really well for all of the stuff I need
