Rakudo Perl: January 2008 Archives

The Perl 6 design team met by phone on 30 January 2008. Larry, Allison, Patrick, Will, Jerry, Jesse, Nicholas, and chromatic attended.

Larry:

• double-booked at work, so I'm kinda busy there
• still doing Perl 6 clarifications as requested
• starting firestorms that I've subsequently ignored on p6l
• some of the $work I'm doing is in the same mindspace as the longest token matcher
• might bootstrap me back into that

c:

• still working on the PMC PDD branch with Allison
• have a couple of bugs to fix there
• a couple of cleanups
• one minor optimization that will reduce (generated) code size somewhat
• another similar optimization to look at, should be more valuable

Patrick:

• unexpected family and house issues came up
• I'm back now
• the plan is to continue on last week's work
• mostly conversions to Rakudo to make it fit what Parrot expects (namspaces and the like)

Jerry:

• made some more spec tests pass with Rakudo
• added type constraint parsing
• thinking about a lot of things
• list context, return exceptions, Perl 6 docs
• very close to committing perl6doc executable for Rakudo

Jesse:

• according to Damian's spec?

Jerry:

• according to Larry's grammar to POD6
• it's small
• it's a start for people who want to hack on the POD6 grammar
• had a discussion with Alias with respect to deterministic parsing
• there's a small patch to prevent Rakudo from running BEGIN blocks
• not sure it's worth applying

Larry:

• see my latest diatribe on p6c

Patrick:

• grammar changes aren't implemented with BEGIN blocks

Larry:

• I don't care if you add the switch

Jerry:

• is it worth adding something to prevent BEGIN blocks from running

Larry:

• I don't see its value, but it shouldn't hurt anything

Patrick:

• just derive a new grammar that disables BEGIN blocks and use that
• shouldn't have to do anything in the compiler to make that happen

Allison:

• spent some time on Mac OS X compilation problems
• seem to have a solution for that
• still working on the PMC implementation
• giving a talk at Linux.conf.au on the Pynie implementation
• it's interesting poking in the guts of NQP and PCT and comparing that to the old PGE/TGE approach

Will:

• my RT viewer is pretty much done
• just need to get an RT instance to play with extracting data in a timely fashion
• working on the Mac port for Parrot, hope to get 0.5.2 in MacPorts in a couple of weeks
• should be simple to keep it updated from then on

Nicholas:

• Rafael wants a lot more test cases for switch in 5.10.1

Patrick:

• we're looking for hints about how to do list context parsing in STD.pm

Larry:

• particle and I talked about that the other day

Patrick:

• that's the next big thing we're missing
• I don't want to hack something together if there's something better in STD.pm

Jesse:

• any blocking?

c:

• I really would love more platform coverage
• a few patches lately that need attention on something more than 32-bit Linux with GCC
• any version of OpenSolaris, just about any BSD would be good, Mac OS X, Windows

Patrick:

• we can set up VMs
• ask people to create their own VM images

c:

• that would work
• ask for freely-redistributable VM images
• and then get non-redistributable OSes to submit to the smoke farm

Patrick:

• I can post that in my journal
• or post a request on Perlbuzz

Check it out! Rakudo Perl has classes with methods and attributes!

uniqua:~/parrot $ cat woot.pl
class Foo {
    has $a;
    method set {
        $a = "w00t";
    }
    method get {
        say $a;
    }
}

my $x = Foo.new();
$x.set();
$x.get(); # prints w00t

uniqua:~/parrot $ ./perl6 woot.pl
w00t

They're not complete by any means, but it's a big step forward. Thanks, Jonathan! Details in his journal.

chromatic has been putting Valgrind to work lately, checking for memory leaks in Rakudo and Parrot. This is the kind of detective work I love to do myself.