Perl 6 Design Minutes for 05 March 2008

The Perl 6 design team met by phone on 05 March 2008. Larry, Allison, Jerry, Will, and chromatic attended.

Larry:

• very little official work appeared to get done this week
• corresponded with Simon, Allison, and Patrick on a possible implementation of grapheme-oriented encoding
• looks good there
• no real spec work thanks to health issues, day job, and my daughter and son in law visiting
• working on the gimme5 translation of the standard grammar
• working on the correct way to extract matches, given a current match state
• how to present them to the user
• ... through the constraints of the Perl 5 type system which really dislikes having capture objects...
• went down one dead end
• digging my way out of that
• also working with putter, who's trying to pull together threads around the concept of a unified AST
• that's a valuable work
• trying to encourage him in that

Allison:

• launched the security PDD
• reviewed Simon's first draft of the character sets PDD
• includes an interesting new way of representing graphemes
• we've gone through various names
• UCS-4/NFG seems to be the most comprehensive name
• finalizing a few details now on the PMC branch
• we plan to merge that in on Saturday
• tracking down one segfault on Windows
• prevents PGE from compiling to bytecode

Jerry:

• focusing on Summer of Code applications
• I volunteered to manage Parrot and Rakudo projects
• submitted as part of TPF
• working with Eric Wilhelm, who really has his act together
• Will will be my backup
• good to have backups for all of our positions
• trying to offload some of my tasks to other willing hackers
• one of those is the YAML serialization
• hope to give that over to Will or Stephen Weeks this week
• spending the rest of my time debugging PDD 17 branch on Windows with Allison

Will:

• working on Tcl on the PDD 17 branch
• nothing particularly Perl 6-y

c:

• fixed the HLL segfaults
• fixed a segfault in the copy opcode
• this seems to have fixed most of Tcl and Lua
• did more work on the PDD 17 PMC update
• made a list of potential tasks for sponsored development

Jerry:

• does Moose buy you anything on Perl 5, Larry?

Larry:

• possibly
• can it easily produce an object which has all of hash, scalar, and array interfaces simultaneously?

Jerry:

• I don't know that offhand
• I know someone who does though

Larry:

• my problems are not so much the implementation as on the conceptual end
• getting my brain around what a match object is
• when you say make something or other, how much that replaces what's there and how much that augments the scalar part of what's there
• trying to think through that
• the main issue that I have is confusing the harvesting of subnodes with the stacking up of subnodes at the same level
• the latter should make lists of matches rather than embedding them
• just a matter of getting a translation to Perl that will do that once I've thought it through
• maybe Moose will help me there
• it's a good idea
• it might run slower
• my purpose in translating to Perl 5 is not to get a perfect translation but just to get one that will run the standard grammar
• doesn't have to be pretty
• only correct enough in the areas that matter to get it to run
• if correctness were a bigger goal, I'd certainly use Moose