The Perl 6 design team met by phone on 02 April 2008. Larry, Allison, Patrick, Jerry, Will, Jesse, Nicholas, and chromatic attended.
Larry:
- things are very busy
- working on a very interesting problem at work that's eating my brain
- sneaked in a few e-mails and synopsis changes
- not much work on the STD parser, just a few nips and tucks
- nothing earth-shaking, mostly just tune-ups in response to John M. Dlugosz and others
Patrick:
- things are going well
- wrote a draft of the second report for the Mozilla Foundation grant
- will post it soon, perhaps tonight
- there has been a ton going on
- this'll be the first in a series of postings to catch up on what's happened the last three months
- closed a few old tickets related to a deprecated feature
- everything that uses PGE is on the new syntax
- reviewed the PCT tutorial and sent kjs some comments and clarifications
- halfway through that
- the changes are all pretty minor
- the tutorial pointed out some places where PCT needs to evolve, and I'm working on that
- getting ready for some heavier Rakudo hacking and reviewing the test suite
Allison:
- finished the main edits to the Strings PDD
- still adding the API reference
- need to ask Simon and Larry to review it
- this is the last PDD in one of the milestones
- working on the next milestone schedule, prioritizing those
- also breaking up tasks for the concurrency milestone
Jerry:
- doing a lot of meta-work, not a lot of work
- four Parrot proposals and one Perl 6 proposal in GSoC
- maybe a few more on the way
- preparing for conference season, especially the Parrot and Perl 6 workshop at YAPC::NA
- hope to catch up with Allison while she's in town
- though we'll probably do more meta work
Will:
- lots of Parrot stuff
- cleaning up build dependencies
- removing some long-deprecated features
- hope to merge back our branch after the next release
- keeping on track with RT
- talked to kjs about shoehorning his tutorial into a compiler tools PDD
- it's a milestone, and one of the last ones
- Allison's looking after the remaining PDDs which aren't part of the milestones
c:
- working with Coke to remove deprecated features
- fixing a few segfaults and crashes, going through open tickets for ideas
- working with a few new contributors on IRC
Jerry:
- they came from Jim Keenan's visit to Toronto
- he's preparing for the workshop at YAPC::NA
- found people with various untested platforms
c:
- that's been very helpful
- adding a few new features
- talked with Mitchell about his plans for a while yesterday
Jesse:
- Patrick and Jerry, what do you think about my request to describe the future of Rakudo in chunkable pieces?
Patrick:
- that's one of my goals
- I'm doing a small part of that in my report to the Mozilla Foundation
- in a longer-term goal, I want to describe what will get us through the summer
Will:
- that's been on my TODO list for Parrot as well
- Google Docs now supports Gantt charts in their spreadsheets
- I've had positive feedback on the RT widget I posted in there
- having something shiny and available might be good for marketing
Jesse:
- even having something internal that's plain text would be good
Will:
- we have existing milestones to look at as a start
Patrick:
- who's planning to go to OSCON?
- will we have Perl 6 meetings before or after OSCON?
Allison:
- the weekend before I have a meeting
- the weekend after might work
- are other people around?
Jesse:
Patrick:
- I was thinking of flying out Saturday afternoon
- might need to find a place to stay
Will:
- I won't be at OSCON
- is anyone going to YAPC::EU this year?
Jesse:
Patrick:
- arranging for funding, would like to go
Allison:
- hadn't planned on it
- but if there's a group of hackers who'd like to hack, I can arrange it
Will:
- there's a chance I might be able to go
Nicholas:
- what's the replacement for IMCC?
Jerry:
- PIRC
- much cleaner
- still lex and yacc
Patrick:
- would the bytecode PMCs be a good Summer of Code project?
Allison:
- it's pretty big
- a C-based bytecode generation library might be a good project
Patrick:
- PCT could use that instead of going to PIR and then through IMCC
- I'll add that to the wiki
Nicholas:
- if the end date is in August, what if it fails and we don't know until then?
Patrick:
- it's not on the critical path
c:
- we want it eventually, but...
Jerry:
- we can go to Parrot 1.0 without PIRC
Allison:
Patrick:
- but if someone wants to do it outside of GSoC that's fine too
c:
- I'll convince TPF to give them a medal
Jerry:
- we have five days to convince someone to do this project though
- is the longest token matcher on the critical path to Perl 6.0?
- any estimate on that, Larry?
Larry:
- what I have almost works in prototype
- it works most of the time
- getting it robust enough to where I can use it on the grammar itself is harder to estimate
- the main problem right now is that I'm prototyping it with tagged regular expressions if the longest token actually matches and succeeds
- one of the ways we've defined the matching to work is that you can put assertions after the match which can fail and nullify the match
- so the engine has to have a way of backing off on a match
- TRE isn't up to that in its current form
- I might have to read the code and tweak it
- long-term, we should write something that does longest token matching in Perl 6 which does a parallel NFA traverse of the longest toking rules with the requisite backoff semantics
- I had one of those half-written some months ago before I decided to prototype it with TRE
Jerry:
- what would it take to get concurrency working in Rakudo based on what's in Parrot already?
Larry:
- I meant character matching in parallel, not OS-level concurrency
Patrick:
Allison:
Patrick:
- a lot of things in Perl 6 depend on MMD working
- or solving problems for you
- otherwise you have to emulate it to make it look like what's happening
- that's where Jonathan's holding up on stuff
- Rakudo will too
- MMD, strings, and concurrency are my priorities
- may not be true from the IO perspective, though
- my working plan for Rakudo is that I want to get it along the way with features
- then step back and revisit PGE and the regex engine, starting with Larry's prototype
- I'd like to do that before the summer conferences to sit down with Larry face-to-face
- may or may not come to pass
- but my working plan is to look at longest-token matching in PGE in mid-to-late summer
Allison:
- what's the blocker with Rakudo and MMD?
Patrick:
- Rakudo doesn't implement it yet
- the next blocker is that we'll use MMD this month
- then I'll be able to identify problems and potential problems
Allison:
- those notes'll feed in to the real, final implementation
