The Perl 6 design team met by phone on 17 June 2009. Larry, Patrick,
Allison, Jerry, and Will attended.
Larry:
- derived types from
Buf for UTF-8, UTF-16, UTF-32
- uncertainty of statement modifiers affecting
$_
- talked about using implicit blocks
- decided that's a bad idea
- to keep scoping rules simple, statement modifiers that are topicalizers must temporize
$_
- decided that typenames that are generic types have canonical name that fits into notion of adverbs (
is Array :of(Int) instead of Array[Int]))
- defined a
.rotate method on arrays and lists
- auto-incrementing strings not limited to alphanumerics
- related to anything in
<rangechars>
- can include non-alphanumerics like dice
- viv can duplicate gimme5 translation of operator precedence parser
- next major thing to do is to duplicate translation of rules to long form that it uses to emulate in p5
- after discussion with Daniel Ruoso, decided to add types into symbol table as if they were routine names (for example,
&Int)
- coercion functions really look like functions in the symbol table
- hacked in new types
utf8, utf16 into STD / corepad
- added several other missing types
- in parsing of declarators, some could be fooled; for example
multi subfoo
- most now require whitespace
- changed variable redefinition message to say "Useless redeclaration" to indicate same variable (also an idea from Daniel Ruoso)
Patrick:
- got a grant from Vienna.pm -- like Jonathan's -- a days worth of work on Rakduo == cash. Mainly a way to get some reimbursement for outliers that aren't covered by other grants. Thanks to Vienna.pm!
- busy with personal trip, not a LOT done
- sched for PVMW in Pittsburg, prepping
- work on Rakudo build from installed Parrot
- fixed Parrot bug concating Unicode/Latin-1
- troubleshoot with chromatic on GC issues that caused leaks in Rakudo (GC was effectively disabled)
- fixed
() to return Nil, not empty list
- string incr, decr, works on Unicode ranges, a la spec
- refactor constant/container treatment; change default to immutable to simplify. (avoid changing value of
True, etc)
- seeing more GC errors with this refactor. Where running Parrot with GC disabled, Rakudo program works; without it, code completes. Related to recent work trying to clean up memory leaks/GC in Parrot.
- need a lot of bisecting to figure out which Parrot revision did this.
- getting ready for YAPC
- tomorrow is Rakudo release #18, Pittsburgh. Was going to try to get someone else, but no one to take?
Allison:
- finished editing book
- still need a few final formatting tweaks before sending to printer
- Will's POD testing scripts were incredibly helpful
- unfortunately I have a concussion, so will miss YAPC::NA, but getting better.
- can work on computer, a little distracty
Jerry:
- working on PVMW with Patrick
- that's about it
- have a bit more time to hang out on irc
- thinking about modules and library loading
- people are asking "How do I find location of a (P5) module without
evaling it?"
- thinking about how to do this in P6
- everyone seems to be progressing on GSoC
Will:
- did minor Parrot cleanups
- Andrew Whitworth did Parrot release yesterday (1.3.0)
- talked with Patrick about gutting hand-rolled call chain that Partcl is using
- came up with one case that can't be solved in Parrot today, but Rakudo has the same problem (
lift)
- when Rakudo has that solved, partcl can steal that
- ran into issues because PMCs and Parrot classes aren't interchangeable
Jerry:
- Larry, are you doing the normal talk?
Larry:
- talk is about error messages. (p5 -> p6)
Patrick:
- he really wants to know how long your talk is
Larry:
- if I go over, you can throw an exception for me to handle
