The Perl 6 design team met by phone on 28 November 2007. Larry, Allison,
Patrick, Will, Jesse, Richard, and chromatic attended.
Larry:
- recovering from Thanksgiving; took the week off
- working on getting the State of the Onion ready for Perl.com
- gave a talk at work on the history of Perl in terms of development, revision control, et cetera
- stuff you guys all know
- I had to organize it in a way to present it to work
Jesse:
- could the slides stand alone?
- other people not tied into the community might find them interesting
Larry:
- mostly just reiterating what's out there
- continued to fiddle with the longest token parser
- just want to keep the cache loaded as long as possible
- I feel like I'm getting my brain back now from the last month's sleeplessness
- nothing exciting coming up soon in terms of Perl
Allison:
- I have always-on ADSL now in my flat
- this week I've been doing a first implementation of the concurrency PDD
- necessary for the events deadline for Saturday
- adding in a few details to the concurrency PDD
- there are some pieces I didn't realize we needed until I started coding
Jesse:
- how much is threading and how much is the more interesting models?
Allison:
- it includes what some older systems would call threading
- it also includes Intel's TBB
- we have multiple concurrency models: exceptions, events, async IO
- there's one centeral concurrency model
- it's more modern than people would call the classic threading model
- that's why the name is "concurrency", to escape the threads-only box
- I'll finish off the events PDD next
- then think about the multiple dispatch PDD and implementation for January
- have to work on some talks for January
- probably need to work on the Python implementation for that
Patrick:
- that's part of my grant specifically
- I planned to get it started and back to its current point with NQP
- you can pick it up from there
- I'll probably present that to the local Python group again
- it'll be a bigger example of NQP that's not Perl 6
- that's easier
- there are also a couple of local people interested in hacking on it who didn't like the looks of all-PIR right away
- given that I wrote the original Pynie in eight hours, I think the new version shouldn't take too long to put together
Allison:
- I'll be speaking about Parrot at Microsoft in March
- it's the weekly DLR meeting
- John Lam's putting out a broader invite in the company as a whole
Patrick:
- published a roadmap for the Perl 6 compiler
- getting feedback on that
- basically finished the AST implementation for now
- basically finished NQP for now
- posted an example of driving SDL from NQP
- reimplemented abc in NQP
- the only PIR code it has is for subroutines that are trickier in NQP than in PIR (for now)
- they're all basically short five-line subroutines
- added a bunch of comments to the grammar yesterday
Jesse:
- the things that are easier in PIR than NQP, does that indicate a lack in NQP?
Patrick:
- they're primitives
- to do an add
- NQP can do one, but it's just actually easier to write the code that does it because it's straight to go to the right opcode there
- those kinds of things
- basic, built-in primitives are easier to write in PIR
- once you get above that layer, it's much easier to write in NQP
- there are 256 lines in all of the transformations for abc
- half of that is comments
- working on some Parrot cleanups
- next up is reimplementing Perl 6 in terms of NQP
- that shouldn't take too long
- maybe two or three days, but I have some other work to do
- I don't want to lose my good momentum here
- trying to keep my journal up to date daily
c:
- not a lot of progress, just some cleanups
- if anyone working on a big task has some smaller bits to break off, I'll do them
- I'm happy to write the compiler tutorial for you, Patrick
Jerry:
- glad to be away for a month then come back with nothing on my plate
- but I'm looking forward to writing the Perl 6 compiler with Patrick
- possibly working on a tutorial, the abc one
- might be able to do that one
Patrick:
- I already wrote the comments for the transformation
- we could use some explanation in the grammar
- people look at that
- it's not always clear even if you've read S05
- a common question is "What's the difference between
rule and token?"
- I put a little cheat sheet in the comments
Will:
- doing a lot of cleanup this week, especially in languages/
- cleaned up a lot of cruft
- trying to get the RT queue in a more manageable state
- nowhere near complete, but Patrick and I have made a dent in the past few days
- I had some stuff in hand, but chromatic's already working on it
- I can focus on converting some other languages, like APL
- also want to get Patrick's journal added to Planet Parrot so they can see his work
Richard:
- a bit of additional work after the Mozilla Foundation grant
Jesse:
- CJ Collier mentioned interest in hacking Perl 6 on .NET
- not sure where that's going to end up
- it's nice to have someone hacking on it to see if it'll work though rather than assuming it won't
