The Perl 6 design team met by phone on 05 December 2007. Larry, Allison,
Patrick, Jerry, Nicholas, Jesse, Richard, and chromatic attended.
Allison:
- still working on the events implementation
- a few days past the deadline
- I have the event loop working, but I'm just cleaning up some backend stuff
- improved the scheduler runloop just this morning
- having a blast
Richard:
- what is included? Signal handling?
Allison:
- yes
- also user-generated events
- can be keypresses, any event you want, messages from async IO
c:
Allison:
- it was the December 1 milestone
- I'm also implementing the concurrency PDD to help with the events milestone
- that puts us ahead of the concurrency deadline
- looks like I won't be doing the release this month
- we're going to pull in some new hackers to do some release work
- I'll work on the next PDD
Jesse:
- who are the new people you're dragging in?
Allison:
- smash is one of them, kjs is on the list
- maybe Jim Keenan?
- Will's putting out a call for more people
Larry:
- things are pretty okay
- the English-speaking part of my brain has turned back on
- I polished up the State of the Onion and sent that off for Perl.com
- I finally wrote the interview for Federico Biancuzzi that's six months overdue
- answering a whole bunch of mailing list e-mail I've been putting off for weeks
- no actual hacking this week
Patrick:
- Perl 6 on Parrot compiler is now mostly NQP
- really really cool
- just a strong set of hacking over the past couple of weeks
- I'm very happy with the way it works and the way it looks
- I think the slowness is the parsing
- likely something in the rules engine
- Jerry has been hacking on the new version of the compiter quite a bit
- he says it's much nicer than it was before, and I agree
- I no longer feel like I'm writing code that no one else will be able to read
- it's not always obvious, but it's easier to explain than most of the PIR code
- more updates to the compiler tools suite
- now you can put PIR inline with the Perl 6 functions in NQP
- that came out much better than I thought it would have
- we can write everything in Perl 6 syntax, except for a few things
- with a few PIR instructions, the rest seems to work nicely
- dropping in and out of PIR is compatible with Perl 6 syntax
- it's really fun
- we get to write Perl 6 code, and then we just put in some PIR to make it all better until we can write that part in Perl 6
- spent the past couple of days rewriting the coding rules for the parser
- interpolation, splitting on space, not splitting on space
- that can get a little hairy
- I'm pleased with how the code came out
- it's understandable, even though it's all PIR
- my next step is to help Jerry and others get the compiler passing all of its tests
- at least where it was before
- primarily working on documentation
- that's the big thing people come in and look for
- I'll work on docs or Pynie tomorrow on a four-hour plane ride
- Pynie's not as primitive as abc or as complex at Perl 6
Jesse:
- how much of the Perl 6 implementation is filling in small stubs?
- is there a large amount of that to do?
Patrick:
- we need a few more of the basics in there
- they need more intimate knowledge of what's going on behind the scenes
- we have the proper Perl 6 object model in place now
- we couldn't do that easily before
- there will be a lot of those things
- adding generic functions and generic methods
Jesse:
- I'd really appreciate a tutorial showing adding a medium complexity feature
c:
- I'd be happy to write that
Patrick:
- I'll identify one and then we'll put something together
- useful and not trivial
Jerry:
- things are wonderful
- I'm having more fun now than I had in all of last year
- the setup we have now with PCT and the Perl 6 compiler is a joy to work with
- everything falls into place
- we're making good progress
- I'm looking forward to being able to add more contributors
- getting various tests to pass from the sanity tests
- the PIR to NQP conversion is still going
- I'm working on the
for statement now
- Patrick and I talked earlier today
- I have a good handle now on the structure of PAST
- PDD 26 is descriptive
- can't think of a single blocker
Patrick:
- I'm letting Jerry do most of the implementation
- just answering questions about it
- that gives us someone else who understands it
- gives me experience leading others through it
- lets me focus on deeper issues
- it's working out very well
Nicholas:
- I think Perl turns 20 on the 18th of December
c:
- hope to get the task scheduler algorithm in place in the next couple of days
- it's a simple array-based heap
- basically a solved problem in OS-based task schedulers
Richard:
- does it seem worthwhile to find someone who will fund many hackathons?
- I talked to Patrick about this a while back
- I'd like to ship him around every couple of months to crystallize developer effort in useful directions
- if you get one person fanaticized per hackathon, that's good progress
Jesse:
- is this Perl 6 in specific or TPF?
Richard:
- good question
- I've noticed that grassroots hackathons have never had a specific Perl 6 focus
- personally I'd prefer to focus on Perl 6
- maybe saying that at least half of it will be Perl 6
- we'll get Allison on a plane or Patrick
- I'd like to at least do that much
- we'll still need local organizers with time and effort to make it happen
- it'll depend on where they want their volunteer efforts to go
Jesse:
- might be good to offer the organizers of a workshop or conference money to organize a hackathon
- lots of developers, done with talks, existing space
Patrick:
- if someone can get me somewhere, I'm happy to go and present or organize
- part of the Mozilla Foundation grant is to evangelize the project
- I've tentatively committed to offer a Perl 6 track at the PPW next October
- my biggest problem is conflicts with existing dates
- I'll be in Brussels in February for FOSDEM
- it's not too late to change my tickets to swing through somewhere else for a hackathon or a mongers group
- if anyone hears about that, let me know
Richard:
- we'll find new money to fund your travel and lodging
Nicholas:
- given jet lag and the hate of transatlantic flights, is this a North American thing?
Richard:
- we've funded things like the Amsterdam hackathon
- I'll happily plan one wherever it makes sense
- as for the key people, it might mean their plane ticket is more expensive and they'll arrive more jetlagged
Nicholas:
- it looks like Rafael will have one more release candidate for 5.10
- kp6 showed a big segfault in a massive hunk of generated code
- not sure how to simplify that
- is there a Perl 5.10 press release?
Jesse:
- Simon Wistow has one just about ready to go
Nicholas:
- we had a lot of translations for the one for 5.8
- it was nice padding for newspapers and things
Jesse:
- ask Simon to send that around for translation
- commit it to the trunk
- somehow get translators without turning p5p into a press release debating service
Nicholas:
- some interest in bug bounties from the London Perl Workshop for the Winter of Code
- is that something Perl 6 needs?
Jesse:
- let's see how well the bounties work before complicating it first
- no call on Boxing Day
- we'll see about the second of January
Richard:
- about the inlining of PIR into NQP...
- how does that work?
- doesn't the PIR have to be compatible with what's generated on either side of it?
- needs to be aware of its context?
Patrick:
- a little bit
- it's not too bad
- PIR has intstructions to fetch and store lexical variables
- you say that you need
$x into PIR, fetch it into a register, then manipulate it
- you do need to know the context
- but it's easy to do
- you can create lexical variables and the remainder of the subroutine can access them
- once you see an example or two, you can do it for yourself
- you can write an entire subroutine in PIR but let NQP/Perl 6 handle all of the calling conventions
- we'll have some examples of that
- I'll answer this in more detail in a journal post
