develooper Front page | perl.perl6.language | Postings from February 2004

This week's summary

From:
The Perl 6 Summarizer
Date:
February 17, 2004 10:40
Subject:
This week's summary
Message ID:
m2hdxpczx7.fsf@obelisk.bofh.org.uk
The Perl 6 Summary for the week ending 20040215
    In his talk on Parrot at OSCON last year, Dan said that Parrot would
    have objects 'by the end of the week'. It's been a very long week, and
    it's not finished yet.

    Right, venting over, time for the summary. I'll start with
    perl6-internals, as always. Actually, we're starting slightly before the
    usual start of the week because about half an hour after I sent last
    week's summary off I realised I'd completely missed a couple of threads.

  Some minor decisions and timetables
    Dan made a few design decisions and outlined a release timetable for the
    next release of Parrot (which was initially set to be Valentine's Day,
    but the appeal of February 29th for a Leap Parrot release proved to be
    greater).

    http://tinyurl.com/2y8fz

  Why didn't I mention this last week?
    Jens Rieks earns a "Wow, cool!" from me with his parrot version of
    Data::Dumper. Way to go Jens. People were impressed.

    http://tinyurl.com/2uebj

  Supporting simple PMCs as keys
    Stéphane Payrard posted a patch to support using simple PMCs instead of
    full on key PMCs for simple lookups. He and Leo discussed it, but the
    patch hasn't been accepted yet.

    http://tinyurl.com/2fru2

  Keyed access to UnManagedStruct
    Everyone's favourite differently capitalized O'Reilly employee,
    chromatic (Maybe the new running joke should be my struggles to avoid
    using his name at the start of a sentence so we don't have a falling out
    cases), continued his discussion with Leo about keyed access to
    Unmanaged (and presumably managed) C structs.

    http://tinyurl.com/yuk3m

    http://tinyurl.com/2c6z6 -- An example of chromatic's problem

  Backward branch, fixed
    Remember the problem with a backward branch that I said was Warnocked
    last week. It turned out that it was fixed before I posted the summary.
    They don't call Leo Tötsch the Patchmonster for nothing.

    http://tinyurl.com/24tre

  Parrot "system"
    Dan apologized for his recent silence on list, but thought he'd be more
    active from now on. He also announced that he was in the process of
    adding in perl 5 style "system" support

    http://tinyurl.com/2jg8v

  Parrot sort
    Jens Rieks garnered more kudos with his implementation of a simple _sort
    function in IMC. Leo checked it in, and you'll find it in
    library/sort.imc in an up to date CVS distribution. Then, seemingly in
    his spare time, he fixed up some SDL related stuff for chromatic.

    http://tinyurl.com/3fm4w

  RFD: Symbol naming and IMCC2
    Melvin Smith posted a Request For Discussion about how IMCC2 will handle
    symbols and sigils. The issue seems to be that, for humans writing code,
    quoting symbols is a pain. Various options were suggested, but nothing
    appears to have been decided yet. (I quite like Leo's suggestion
    involving declaring lexicals with a mangled name for use in parrot, but
    marking it as being associated with an $unmangled_original_name.)

    http://tinyurl.com/2ovqp

  JIT and branches under the Sun
    Stephane Peiry found with the JIT on Suns; it seems that a loop that
    should terminate carries on forever instead. He and Leo worked on
    tracking it down, but the bug was still there by the end of the week.

    http://tinyurl.com/2acc8

  Rules for method resolution?
    Okay, so I may have kvetched about the continuing lack of Parrot objects
    earlier, but Dan's definitely working on it. This time he's thinking
    about method resolution. In particular he wondered when

       findmethod P4, Pobject, "methodname"

    should do the name resolution. In other words, should P4 contain the
    method that was current at the time of the "findmethod" call, or a proxy
    PMC that waits until the method is invoked to find the appropriate
    method. The discussion seemed to come down in favour of doing both.

    http://tinyurl.com/232lg

  Object spec
    Dan noted that he'd checked in some major revisions to PDD15 (objects!)
    and suggested that people take a look. Simon Glover, Leo Tötsch and LF
    had questions.

    http://tinyurl.com/2hm3x

Meanwhile in perl6-language
  The thread that ate the mailing list
    Luke Palmer doesn't like the repetition inherent in

       @sorted = sort { $^a.foo('bar').compute <=> $^b.foo('bar').compute }
                   @unsorted;

    He wondered if there might be some way of rejigging "sort" to eliminate
    the duplication and proposed a couple of spectacularly ugly
    possibilities. And then the world exploded. It seems that nobody (on
    perl6-language) actually likes the current signature of "sort", but
    nobody can quite agree on what it *should* look like. You know things
    are getting hand when Dan proposes just using SQL's "ORDER BY" syntax
    (even though he was joking). Simon "Rubyometer" Cozens proposed:

       @sorted = @unsorted.sort(op => &infix:<=>, 
                                key => { $^a.foo('bar').compute });

    Which, I have to confess, I rather like. Rod Adams demonstrated a truly
    painful looking sort function which this summary is too succinct to
    contain, see the link for details.

    Personally, I'm amazed that a single library function could generate so
    much discussion. And there's more to come next week.

    http://tinyurl.com/3drwy

    http://tinyurl.com/2sk5l

    http://tinyurl.com/2exo4 -- Feel Rod's pain.

  Haven't I had this feeling of déja vù before?
    Dmitry Dorofeev pointed everyone at Scharli et al's paper on Smalltalk
    Traits and proposed that Perl 6 use something similar. Which is what
    Larry proposed a few weeks back, except he called Traits Roles and
    propose that they also have (or mandate) state (the original Smalltalk
    Traits are stateless). Aaron Sherman wondered if the Emperor was
    actually wearing clothes by saying that Roles didn't seem to be
    substantially different from Java interfaces. I particularly liked
    chromatic's answer that "interfaces completely suck while [Roles]
    don't." My other favourite quote of the thread was from (who else)
    Larry: "The Traits paper didn't go into traits with state, though it did
    mention it as a future research topic. We're just doing that future
    research for them."

    http://tinyurl.com/32ttd

Announcements, Acknowledgements, Apologies
    Well, I've thought about it some more, and I'd like you all to welcome
    chromatic as the Perl 6 Summary's official running joke. I don't
    guarantee to mention him every week, but I do guarantee that I'll always
    jump through hoops to avoid starting a sentence with his name when I do
    mention him.

    If you find these summaries useful or enjoyable, please consider
    contributing to the Perl Foundation to help support the development of
    Perl. You might also like to send me feedback at
    mailto:p6summarizer@bofh.org.uk, or drop by my website.

    http://donate.perl-foundation.org/ -- The Perl Foundation

    http://dev.perl.org/perl6/ -- Perl 6 Development site

    http://www.bofh.org.uk/ -- My website, "Just a Summary"




nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About