Wednesday March 26, 2008
Perl 6 Design Notes for 19 March 2008
The Perl 6 design team met by phone on 19 March 2008. Larry, Allison, Will, Mitchell, Jesse, and chromatic attended.
- indisposed last week
- getting half a gallon of blood put into me in the hospital last week
- had various tests, all negative, so probably something not too terribly obvious
- going in for a lower endoscopy in a week or two
- will probably know more then
- probably something benign that's just leaking blood slowly
- I do feel better than I did a week ago
- it's amazing how much more quickly I can walk, when my blood doesn't think that I'm up at 14,000 feet altitude
- hopefully thinking more clearly too
- or at least less unclearly
- mostly getting caught up and doing some spec clarifications about the semantics of list assignment to arrays
- what you can or can't assume about the batching of asynchronous read operations
- did throw in one new feature this morning
- I wanted badly when I wrote a Perl 5 program last week
- there's no easy way to say "match as much of this string as you can, then pretend as if you succeeded, ignoring the rest of the string"
- it's very useful for matching abbrev. and such
- each subsequent character is optional, but only if the previous one matched
- nested optional characters
- we've had some horrible hacks to work around that
- I put in syntactic sugar for that
- we'll see how it works in practice
- tried to figure out which rules can be rewritten and which can't, reliably
- thinking about the future of the bootstrap I've been working on with standard 5
- trying to figure out what the blockers are on bootstrapping
- next travel is to YAPC::Asia
- will stop in Seattle
- struggling to get to a usable Perl 6 implementation
- tried to use the Rakudo parser and dump the parse tree, to go to a separate backend
- making pretty good progress
- the parser is a bit slow, though
- 40 seconds to parse 700 lines of code
- not quite tractable
- back to looking at a hand-transliteration of STD.pm into Ruby
- it's starting to gel
- may be a new attempt this week to use a slapped on backend to get to a usable Perl 6
- whenever standard 5 matures, it can be swapped in for the Ruby parser
- is there anything anyone else can do to help you on this?
- once it can do Perl 6, there's a world of possibilities
- until it actually runs, there's not much for anyone else to do
- working on the Unicode/character set PDD
- gradually changing its name to just "strings"
- Simon had a very good first draft
- does he know you're working on it?
- I'll mail him
- gave a talk at Microsoft to their DLR/CLR groups about Parrot
- then met with their open source lab
- they'll do smoke testing for us on several different versions of Windows
- we released Parrot 0.6.0
- most of the big changes are in the PMC PDD, which landed
- we've also had a big push to close new and old bugs in our system
- should give us a lot of energy in the near future
- right now I'm looking at a way to profile Parrot apps at the PIR level
- good optimizations available there only after we can measure them
- did you look at KCachegrind?
- I did, also the Devel::DProf format
- the big work so far seems to be instrumenting Parrot appropriately to detect spots where we need to measure things
- haven't yet decided which output format to use
- Valgrind is necessary if you use the Cachegrind format from KCachegrind
- TPF is now an official part of GSoC 2008
- haven't seen any Parrot or Perl 6 specific hits for the Summer of Code
- I'll send out something to the Parrot and compiler lists trying to recruit students
- is there a way to send mail to Perl Mongers groups?
- I think Eric sent mail to the monger managers
- are there any Perl 6 or Parrot talks that aren't on the schedule yet for YAPC?
- maybe "getting started with a little language"
- or "implementing a Perl 6 feature"
- I'll send mail to the lists