The Perl 6 design team met by phone on 03 October 2007. Larry, Allison,
Jerry, Will, Jesse, Nicholas, and chromatic attended.
Patrick:
- worked on getting PGE to work with Parrot's new object model and vice-versa
- created a branch for the pct refactor -- still need to check in some updates
- I have a draft of the AST design document; will likely check it into the repo tomorrow for comments and updates
Larry:
- didn't do anything on the spec
- mostly re-hacking the lexer optimizer
- the approach from last week was a dead end
- now the program spits out the AST as chunks of YAML
- I was trying to second-guess the information I need
- turns out I needed most of it
- I'm much happier now
Jesse:
Larry:
- making it work
- the generated lexer is an egrep-style regular expression
- I need something to take that and do lexing
- it's not clear what I'll use or if I'll roll my own
- it needs something which resembles longest-token matches
- if someone had written an egrep-clone in Perl 6 then I'd just use it
Jesse:
- have you put that challenge to #perl6 yet?
c:
- he will have when I publish these notes
Larry:
- it's probably a week too soon for that
- they're having more fun playing with the CL backend
- it's almost to the point of the Perl 5 backend now
- I'm trying to keep a low profile these days and get some work done
Allison:
- things in Parrot are excellent
- hacking on objects all week
- I'm about to check in a patch that fixes the last MMD system failures in the branch
- that's the last major broken subsystem aside from PGE
- the rest of the failing tests mostly don't instantiate new PMCs correctly
- I may have a hack that could enforce correct instantiation
- clean up the core usage, not just the surface usage
- that should clear up a lot of the tests as well
Jerry:
- why do you call that a hack?
- it sounds pretty good
Allison:
- it checks type ids to decide what it's instantiating
- it's a little more code
- eventually I'd like to get rid of that as well
- we didn't actually roll out the branch at the end of the month
- we're not too many days off though
- if Patrick doesn't get to PGE when I clean up a few tests, I'll just do the change for him
- it's a small change
Jesse:
- do you have a sense of when he'll be able to get back to Perl 6?
Allison:
- this week
- he put that aside temporarily to finish a funded milestone
- he said yesterday that he had a nearly-complete draft to check in
- this is infrastructure he needs to finish Perl 6
Jesse:
- it's all critical path stuff, I know
c:
- helping Allison by cleaning up the edges of PDD15
- adding a couple of small features
- fixing some bugs
- found and fixed some unrelated bugs that were bothering me
- looking at some optimizations
- put one in, waiting on PDD15 branch to land for another
- reviewing some benchmark code that someone said was a lot slower in Parrot
- will have more details on that soon
Nicholas:
- spent a lot of time looking at CPAN code
- lots of it is pretty mediocre
- reviewing how people use C
- I put a hack into maintperl to help
Class::DBI
- did see that Flavio has released KP6 to the CPAN (on October 4)
Jerry:
- working with Allison on the PDD 15 branch
- fixing up tests, examples, other code to use the new model
- the most important thing for me is when Patrick finishes blocking
- then I can turn my efforts to Perl 6
Will:
- just trying to give guidance on #parrot when anyone needs it
- otherwise, work keeps sucking up my free time
Jesse:
- not much going on here
- will be out of town next week
