All the Perl that's Practical to Extract and Report

Perl 6 Design Minutes for 23 January 2008

brian_d_foy on 2008-01-25T23:03:00

The Perl 6 design team met by phone on 23 January 2008. Larry, Allison, Patrick, Jerry, Will, Jesse, Nicholas, and chromatic attended.

Larry:

  • just fudging things
  • developing the fudge preprocessor and making sure it's in the right place such that it's usable by different implementations
  • hopefully in a platform-independent fashion
  • today I just added another declaration to set the test count of the following construct
  • been thinking about that for quite a while
  • the usual spec damage
  • things have been pretty hectic at work
  • going to Japan in May for YAPC::Asia

Allison:

  • when?

Jesse:

  • the 15th through 17th

Larry:

  • if you count the hackathon

Allison:

  • launched the GC PDD
  • had some helpful comments from chromatic
  • back to implementing the PMC PDD
  • the two big things left are roles and role composition and the automatic generation of the core PMC structs
  • aside from that it's a lot of small cleanups
  • we're also having interesting linking problems on Mac OS X
  • GCC on Mac OS X has different assumptions about when to add libraries to the static lib
  • you must explicitly include a library
  • I finally found some documentation on why we have this difference between Linux GCC and Apple's GCC

c:

  • did you find a fix?

Allison:

  • I figured out how to make it pass
  • I haven't figured out how to put the .o files into pbc_to_exe
  • I think they're not getting stored in the config library
  • that's what it needs to know to get pbc_to_exe to work on the Mac

Jerry:

  • working to switch the name "Perl 6" to "Rakudo" in the source and documentation
  • made fudge work with Rakudo and am adding directives
  • there's a lot more work to do there
  • added more tests to Rakudo's spectest make target
  • it's a lot noisier
  • but it will hopefully inspire us to clean that up
  • applied a bunch of patches from various contributors
  • we've had three or four contributors over the past week, which is great to see
  • working on some Perl 6 rule syntax for PGE
  • also working on a grammar for serialized PAST
  • should let us read PAST in from the same or different implementations
  • finally trying to get more Rakudo tests passing

Patrick:

  • returned to Perl things over the weekend
  • mostly helping others: answering questions, giving guidance
  • not a lot of actual coding besides switching APL to use PCT instead of the old tools
  • didn't feel like diving into Rakudo right away
  • plan to write some reports this week, especially the Mozilla Foundation report
  • then I'll review recent commits, especially to Rakudo
  • I'll make comments on those
  • will work on some of the infrastructure things
  • improve the Parrot integration
  • switch to Parrot's HLL features, especially its own namespace
  • also need to do more review of tests and test suite things

Will:

  • trying to keep an eye on the RT queue and apply some patches
  • we're still getting patches for LOLCODE
  • sjansen has patched that, and he's patching PCT as well
  • seems like that's having the effect we hoped
  • starting to think about next month's release
  • we're shooting for 0.6.0
  • is the Mac OS X issue just for 10.5?

Allison:

  • 10.4.11, GCC 4.01, Intel

Will:

  • hmm
  • I'll talk to you more offline then

c:

  • applied several patches
  • fixed a couple of nasty bugs
  • Andy and I are working on cleaning up Valgrind output
  • fixing some warnings, hopefully they'll catch some dodgy memory uses which cause segfaults on other platforms
  • oh, and deleted a big chunk of old object code
  • plan to work on the PMC PDD plans; Allison gave me a nice list

Patrick:

  • make test in NQP regularly gets failures for me
  • looks like GC problems
  • running with -G cleans that up
  • it's 32-bit Intel on Ubuntu
  • seems consistent
  • four tests fail regularly
  • you should be able to reproduce it

c:

  • okay, I'll poke at those and see what I can find

Nicholas:

  • how's Damian doing?

Patrick:

  • I've been communicating with him
  • I get the impression he's busy, but he's around some

Roles Systems

Stevan on 2008-01-26T16:45:33

Allison:
... the two big things left are roles and role composition ...

Allison, please give a look at the latest version of Moose roles, after having written and re-written a role system about 5 times now (Class::Trait was the start, a couple ones in Pugs, the original Moose role system, and then this latest version), I am pretty confident in this approach. It may not be 100% translatable to Parrot, but at the very least it is prior art with which to learn from.

As always, feel free to email me or stop by #moose to discuss any of this, I am more then happy to pass on what I have learned so you don't have to repeat my mistakes.

- Stevan