Perl 6 Design Meeting Notes for 08 August 2007

chromatic on 2007-08-10T01:30:08

The Perl 6 design team met by phone on 08 August 2007. Larry, Damian, Allison, Patrick, Jesse, Nicholas, and chromatic attended. These are the minutes.

c:

  • I think I found the Tcl/Parrot problem
  • unfortunately, I need either Leo or Jonathan to help
  • otherwise it'll probably take another couple of weeks to track down
  • fixed a couple of bugs found through Coverity scans
  • looking for help from a NetBSD hacker to write a good Parrot pkgsrc package to pull and build from Subversion
  • that'll get more frequent scans

Damian:

  • a week away from my northern hemisphere speaking tour
  • fairly focused on getting ready for that
  • preparing for the YAPC::EU keynote
  • thinking heavily about Perl 6 documentation features
  • awful lot of effort to prepare for an eight-week tour
  • I'll have a lovely uninterrupted 24 hour block of time in the air in which I hope to make documentation progress

Larry:

  • did a major upgrade to the standard grammar to remove all obsolete special variables
  • decent failures about what people ought to use instead
  • made a language restriction: placeholder names cannot be single uppercase letters
  • mostly otherwise working with Patrick about either the semantics of autoincrement or the interplay of blocks and semicolons
  • we've made progress on both
  • if you have any construct in the middle of a line, a construct that could terminate there, you need a semicolon if it doesn't look like there's a terminator
  • partly for consistency
  • mostly for visual reasons
  • simplifies the grammar slightly

Nicholas:

  • the grammar inferred semicolons there?

Larry:

  • yes
  • if you have two statements on a line, you have to terminate the first one well
  • right brackets can count as termination
  • makes list comprehensions easier (blocks can contain statements)
  • that's occupied most of my time

Patrick:

  • things are going well
  • NQP is coming along nicely
  • have a new contributor working on it (Colin Kuskie)
  • I'm trying to make sure he's not bottlenecked
  • that's part of the reason for me working with Larry this week
  • the grammar doesn't always make it obvious how to accomplish certain things

c:

  • how much work does NQP need before we can translate other languages?

Patrick:

  • two or three ideal days
  • not sure how many real days that is
  • certainly hope it's before the next Parrot release
  • I know people are waiting for it
  • Larry's response time has been incredibly helpful on that too

Larry:

  • I'm trying to look omniscient

Allison:

  • implemented vtable overrides for the new object metamodel
  • should be the last big thing Patrick needed before switching over PGE
  • that has PDD 15 passing most of the old test suite
  • only have about 70 tests failing there
  • have a couple of final bits
  • the new metamodel still doesn't allow you to do a vtable override on find_method
  • the bulk are working
  • that's a three line change
  • after Patrick shifts PGE, we can remove most of the old system
  • or at least shunt it away somewhere

Larry:

  • how's the garbage collector coming along?

Patrick:

  • we have a lot of people looking into the problem, but no real fixes yet
  • it's difficult to diagnose....