All the Perl that's Practical to Extract and Report

Perl 6 Design Meeting Notes for 22 November 2006

chromatic on 2006-11-27T03:34:51

The Perl 6 Design team met via phone on 22 November 2006. Damian, Allison, Patrick, Jesse, Nicholas, and chromatic attended. These are the notes:

Damian:

  • a bit of a challenging week
  • family stuff here
  • also had apocalyptic Biblical weather: snow, hail, 40 C temperatures, gale-force winds, a plague of snakes
  • haven't worked as hard as I meant to work on S26
  • working on it at night, when I can't see the snakes
  • 60% of the way through the final read-through
  • I swear that I will post it today
  • spent a lot of time implementing a POD6 to XHTML parser
  • very close to having that done
  • might be out in a couple of days
  • it's about 600 lines of code
  • mostly package declarations
  • one of the big changes was that POD6 is descriptive, not presentational
  • that made XHTML much easier
  • it was a Basic decision
  • the I, U, and B mark Unusual, Important, and Basis of the surrounding text
  • that actually helps!
  • Synopsis itself is just shy of 10,000 words
  • I removed a lot of stuff from the parser that wasn't in the Synopsis
  • the Synopsis itself will be out before I sleep today
  • the parser, realistically, will be probably later this week

c:

  • fixed a Parrot bug for Patrick yesterday: Patrick's test suite issue with Test::Harness
  • looked at MiniPerl6 for Parrot
  • did a bit of poking there, but Jerry has had more luck

Allison:

  • having fun with a Parrot week
  • ran into a TGE bug with vtable methods
  • still isn't fixed
  • working on the IO PDD today

Jesse:

  • not much to report here either
  • want to hear from Larry what's left on the horizon
  • what are big things left needing definition?
  • there are a few undefined holes

c:

  • concurrency

Jesse:

  • that's at the top of the list

Nicholas:

  • the part that scares me at the moment is the thread about dynamic scope, continuations, and STM
  • how does STM fit with continuations?
  • is there a language with both of them implemented or are we trailblazing here?

Allison:

  • the IO language does something similar, if I recall correctly
  • do you see something problematic?

Nicholas:

  • you seem to have trouble reconciling those and dynamic scope

Allison:

  • the dynamic binding proposal?

Nicholas:

  • if you open a transaction, take a continuation, then commit the transaction, what happens when you resume the continuation?

Allison:

  • you essentially have two separate transactions
  • you have the initial state of things
  • in both the original execution and the resumed continuation, you have the hypothetical state of the STM transaction within the continuation
  • all you've done by reinvoking the continuation is to restart the transaction

c:

  • seems like you can ask the same question about doing IO in continuations