Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

use Perl Log In

Log In

[ Create a new account ]

chromatic (983)

chromatic
  (email not shown publicly)
http://wgz.org/chromatic/

Blog Information [technorati.com] Profile for chr0matic [technorati.com]

Journal of chromatic (983)

Wednesday May 21, 2008
03:01 PM

Perl 6 Design Minutes for 14 May 2008

[ #36487 ]

The Perl 6 design team met by phone on 14 May 2008. Allison, Richard, Patrick, Jerry, Will, Nicholas, and chromatic attended.

Allison:

  • working on the exceptions implementation for the concurrency branch
  • also the Parrot foundation

Richard:

  • trying to drum up funding for the Perl 6 paupers
  • something's happened along these lines
  • an anonymous member of the Perl community brokered a deal with a philanthropic business acquaintance
  • I put together a pitch describing what we could do productively with a donation
  • the money arrived yesterday
  • half of it will go toward supporting Perl 6 development in a variety of ways
  • mostly direct developer granting, like Patrick's grant
  • the other half is for TPF use to improve organizational capabilities internally
  • there's a lot more potential for funding
  • orders of magnitude more
  • that'll take us over the finish line
  • though TPF is not a paragon of activity and organization
  • we'll use it to improve to make us more credible to the rest of the world
  • and potentially more developers
  • want to go from a handful of grants like this to something larger
  • I'll spend the next week or so making plans
  • we'd like to start getting some of that out over the next few weeks

Patrick:

  • outstanding work
  • you'll go down in history for this

Jerry:

  • will they money go out through the grant committee?

Richard:

  • I don't think so
  • this will be as directed by the board

Jerry:

  • is there a PR plan for spreading this news?

Richard:

  • I hadn't thought of that
  • I'd like ideas for helping with that

Jerry:

  • it's easy to post to web sites
  • it might be good to give brian d foy a heads up for TPJ, too

Richard:

  • I'm trying to do similar things for Perl 5
  • this particular donor had a specific interest in Perl 6
  • the half of this we're using to improve TPF will help Perl 5 as well

Nicholas:

  • I hope the positive feedback will encourage other people to do the same for Perl 5
  • presumably they're already using it and want to contribute back

Patrick:

  • busy week here
  • focusing specifically on Rakudo
  • helping other people hack themselves
  • added fat-arrow parsing, so now that works for passing named parameters
  • added signature default values
  • no more crazy MMD stuff to get that working
  • worked on the test suite
  • started at the beginning, Synopsis 2 cleanup
  • making notes for others on things to do and not to do
  • tried to close out old tickets, did about 18 Parrot-only tickets
  • some updates to PGE to bring in S05 changes
  • Tene added placeholder parameters to Rakudo
  • surprised and pleased me to see that
  • he found a problem with hashes versus arrays in PCT
  • I did quite a bit of work to improve PCT code generation
  • substantial improvements there
  • generates much nicer code

c:

  • there are some speed improvements there
  • they're hard to measure, but they're there

Patrick:

  • we're creating many fewer elements there
  • especially in places like loops

c:

  • we don't have many long-running benchmarks that would demonstrate that
  • if we did, you'd really notice it

Patrick:

  • it's been in the back of my mind since I started writing PCT
  • it makes for nicer looking PIR code, if you like PIR code
  • now I'm writing some journal posts to describe this to the wider community
  • I don't know if anyone saw it, but there was a Twitter post that singled out TPF for its organization for Google's Summer of Code
  • Jerry should get some credit for that for Parrot

Richard:

  • I've talked to Eric Wilhelm a lot over the past few weeks
  • he's doing a great job
  • this has been his full-time job over the past few weeks
  • that's why we're not always this organized
  • very seldom do we have someone who can make such a thing a full-time job
  • simple matters of organization often aren't

Jerry:

  • worked with Allison on research for the Parrot Foundation
  • she interviewed a lawyer to help set up the incorporation
  • work has been lighter this week
  • spent more time on Parrot and Perl 6
  • haven't had any significant commits
  • began to rewrite mod_parrot's configure system for flexibility and portability
  • hope to get it compiling on Windows
  • shouldn't be difficult
  • Patrick and I discussed making HLL type mappings available in IMCC

Patrick:

  • PIR doesn't seem to have a way of finding HLL type maps
  • PCT needs to move something from a register into a PMC
  • equivalent types, as far as the HLL is concerned
  • there doesn't appear to be a way to do that
  • without the HLL person repeating the mappings in PIR
  • violates DRY

Jerry:

  • I think I see a way of implementing that
  • the major problem is that, internally, the PMC types are still type IDs
  • not string names
  • I might have a short-term fix
  • the long term is to get rid of type IDs internally
  • it's easy enough to expose that through interp_info.pasm or something

c:

  • let me know if you need help

Patrick:

  • also remember that class names are not always strings
  • they are now, but the won't always be

Will:

  • I'll try to unstick the type ID removal branch later this week
  • mostly I've been doing queue management, which is not exciting

Patrick:

  • I'm excited about seeing the number go down

Will:

  • we've closed an enormous number of tickets in the last week
  • yet more tickets keep coming in
  • what's wrong with these users?

c:

  • spent some time closing bugs and applying patches
  • need to go through open patches and make sure they're all closed
  • probably won't make much of the Bug Day
  • backported yada yada yada operators to Perl 5, waiting for someone to apply them
  • managed to speed up the Rakudo-building benchmark by about 40% yesterday
  • hope to get onto the concurrency implementation soon

Nicholas:

  • talked to Daniel Ruoso and stuck his response on the wiki
  • otherwise busy with a potential move

Patrick:

  • Jerry, are you still planning to reimplement abc to use the language shell tool?

Jerry:

  • yes
  • I plan to update the tool to include an operator precedence parser
  • not sure if I should couple those changes

Patrick:

  • I'm thinking of updating the tutorial
  • kjs's tutorial is great
  • I just want another one
  • I don't want to step on your toes

Jerry:

  • I plan to make my changes
  • then apply the revised tool to abc
  • I'll turn my attention to that next

Patrick:

  • we should knock that out pretty quickly

Jerry:

  • what's your plan for this release?

Nicholas:

  • are you in a position to type things from this release to last release?
  • or should I ask Patrick?
  • you've sped up a lot of things

Patrick:

  • Rakudo gets bigger over time
  • there's quite a bit more code there than there was last month

c:

  • there are a couple of ways to do that
  • we could plot the ratio of time per feature over the release numbers
  • that's not too difficult
  • otherwise, we could take the current PGE/PCT/NQP/Rakudo and run it on this month's release versus last month's release
  • or run last month's toolkit on this month's Parrot

Patrick:

  • I can do that
  • already thinking of ways to do that

Nicholas:

  • Damian was talking about negative time last year
  • you guys are doing it!

c:

  • the big improvement is likely in the new garbage collector

Jerry:

  • Andrew and Patrick should have a race

Patrick:

  • longest token matching will help languages
  • that's on my agenda for later in the summer though

c:

  • remember that these are only Parrot-level improvements
  • we haven't even started on PIR-level improvements
  • because we can't profile it easily yet
  • I suspect there are plenty of trivial 10 - 20% optimizations we'll find in hotspots there
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More | Login | Reply
Loading... please wait.
  • > Jerry:
    >
    > * it's easy to post to web sites
    > * it might be good to give brian d foy a heads up for TPJ, too

    please, for "$foo" [perl-magazin.de], too.

    BTW: I think you meant TPR and not TPJ ;-)