The Perl 6 design team met by phone on 16 April 2008. Larry, Allison,
Jerry, Will, Nicholas, Jesse, and chromatic attended.
Jerry:
- the release went pretty smoothly
- had some help
- the
make release target is broken on Windows, and that's the only platform I had out here
- we'll fix that before the next release from Windows
- talking to Andy Armstrong about getting Test::Harness to 3.0 and subclass TAP::Parser so that we can report Rakudo's fudge tests better
- every fudged test is a failure right now, even if all subtests pass
- I talked Jonathan into implementing simple MMD in Rakudo
- chromatic wrote about it
- but it's broken in the release (and only the release)
- put some work into internationalization
- need to figure out the make rules
- then I can put more work into localization too
- trying to secure the parrot.org domain too
- TPF is likely to get five slots for GSoC
- some will be Parrot and Perl 6 related
- the official announcement is on Monday
- trying to encourage others to take on responsibility
- seems to be working
- some of the committers I've mentored are becoming mentors themselves
Patrick:
- mostly reviewing different things
- working with Jonathan on his various objects and MMD implementations
- will check in my Rakudo milestones document tonight or tomorrow morning
- cleaning up bug reports, closing tickets, catching up on patches
- need to spend a little more time on paying work this week
Will:
- trying to get the most recent release bundled as a macport
- there's apparently a build issue since the previous macport
- should have an easy way to install Rakudo as
perl6 after that gets straightened out too
- still trying to cut dead things out of Parrot
c:
- poked at optimizations per Patrick's request
- sped up OO and Rakudo by about 40%
- not as much as I wanted, but you notice it
- looked at a few more optimizations, but they're bigger and take more work
- GC is the biggest, so let's hope we get that as one of the GSoC projects
- think I can get the profiling core to emit Callgrind-compatible output for PIR
- I know mostly how to do it now
Larry:
- spending a lot of time tweaking a new laptop and getting it all set up and customized nicely
- feebly trying to keep up with the onslaught in p6l
- I haven't been keeping up, but I'm keeping my eye out for things going off track badly
- plotting how to get rid of TRE as my longest token matcher
- want something that will scale better
- the handwriting was on the wall the first time I went in it with the debugger
- for the regex matching where a token is expected, TRE was allocating 17 MB on the stack
- I have some ideas for something with better semantics and less memory usage
- TRE optimizes for running one regex a lot over a lot of data, rather than running a lot of regexes
Jesse:
- spent a lot of time in discussions about funding Perl 5 hackers
- I see François Perrad has released a Win32 binary of Parrot
Jerry:
- he's been doing that for the past few releases
- just a release in binary form, not a fork or branch
c:
- what's the status of Parrot in Debian?
Allison:
- one final build bug in IA-64
- the guy with the box should be getting to it this week
- we're planning to put up the 0.6.0 release
- if it doesn't go soon, I'll just put the F<.deb>s on the site
- we might put IA-64 builds on our platform wishlist
c:
- we need someone who knows how to fix them too
Allison:
- it was a PGE bug unrelated to the arch, I think, in 0.4.0
- we just need the bug confirmed fixed on that arch for Debian
Jerry:
- PAUSE kinda stinks
- I hate that we don't know if we'll have an authorized release of Parrot until it gets uploaded
- PARROTRE lacks permissions on eight modules
- all of which have been refactored on something else
- they're all related to the configure or test system
Allison:
- we don't have to distribute through PAUSE
Jerry:
- we should not index those modules anyway
c:
- they're not that useful outside of Parrot anyway
Will:
- the release link on the Parrotcode site, http://www.parrotcode.org/release/devel, links to the CPAN download
- there's a lag between the update and the availability
Jerry:
- we could upload to Parrotcode first
- upload to the CPAN from that site
c:
- I'm not sure why we should index the
config::* modules
- the only ones I care about are in
Parrot::Embed
