Perl 6 summary for week beginning 2002-09-30
This is yet another Perl 6 summary, documenting what has happened over
on the perl6-internals (where Parrot, the virtual machine that will run
Perl 6 is discussed) and perl6-language (where Perl 6 language design is
discussed) mailing lists. Piers is off on holiday (snowboarding and
parachuting, no doubt), so I will be your host for the next two weeks. A
particularly average week, so let's start off with the perl6-internals
list as usual.
Getting started guide
Erik Lechak offered to help write a getting started guide to Parrot - he
thought (correctly) that there may be "a pressing need for a document
helping newbie developers figure out how to get started".
Unfortunately, Erik isn't a big fan of POD and attempted to code up a
better replacement. While POD may not be perfect, an awful lot of Parrot
developers are also Perl developers and POD is simple and good enough
for Parrot for now.
<http://makeashorterlink.com/?U22B21502>
Debugger debugged
Juergen Boemmels noticed that hitting the Enter key in the Parrot
debugger (pdb) caused a segfault and that this wasn't quite ideal
behaviour. He provided a patch which ignored empty lines, and after a
nudge from Aldo Calpini provided another patch which followed the
suggested behaviour in "docs/debugger.pod", instead repeating the last
command entered.
<http://makeashorterlink.com/?O13B26502>
Patch Master
The main patcher these recent weeks has been Leopold Toetsch, and he
kept it up this week as well. He provided patches to stop Parrot
permanently allocating increasing amounts of memory, fix a parser error
in imcc, add stats to life.p6 and more. In fact, so many patches that I
have a feeling he'll get CVS write access soon ;-)
New allocator
Leopold had a play with plugging in the Doug Lea memory allocator,
providing a patch to configure in support for the current Parrot malloc,
the Doug Lea malloc, or ptmalloc in libc. He also attempted to plug in
Perl 5's malloc and found similar speed results. Later on, he created a
document to explain memory internals, continuing the recent trend for
more docs.
<http://makeashorterlink.com/?D23E216F1>
<http://makeashorterlink.com/?H64B46502>
<http://makeashorterlink.com/?D25B21502>
Patches, patches everywhere
There have been a lot of pending patches recently, and Robert Spier
prodded us about them and hopes to have an automated weekly email nudge
too.
<http://www.parrotcode.org/openpatches>
Library name collisions
Steve Fink reported a problem with library name collisions. For example,
having "intlist.c" as well as "classes/intlist.c" and "key.c" as well as
"classes/key.c". Dan proposed a naming convention where classes start
with a CL_prefix, encodings with an EN_ prefix and character set stuff
with a CS_prefix, although it looks like a simpler solution may be
taken.
<http://makeashorterlink.com/?U26B51502>
core.ops ate my (miniscule) RAM
David Chan is using an itty bitty box (Cyrix with 32 MB of RAM) and
found that compiling Parrot failed due to lack of memory. Nicholas Clark
reminded us that you can disable the (faster, but needs more memory to
compile) computed goto core at configure time using "./Configure.pl
--cgoto=0".
<http://makeashorterlink.com/?Y27B32502>
Parrot file list
There are a great number of source files in the Parrot distribution and
Mark Sparshatt gathered ideas for how to have a file list which
explained what each actually does.
<http://makeashorterlink.com/?Q18B21502>
Interfaces
Over in perl6-language, Michael Schwern was asked at a recent Java
conference whether Perl 6 will support interfaces. He tried to describe
a likely interface and raised some problems with interfaces. This
spawned a huge thread about OO, interfaces, and some of the weirder OO
languages. Do we enforce strict interfaces? Are some methods allowed to
be optional?
One of the more interesting ideas would be to move from just using
prototypes to using Design-By-Contract features: pre- and
post-conditions and invariants. Dave Whipp seemed a big fan of the
Eiffel model, where it is possible to rename methods in a derived class.
There was lots of scary Perl 6 pseudocode including cars, MP3 players,
birds, and wings.
A little later, Schwern came back and worried about enforcing interfaces
on subclasses. One interesting approach is to follow Eiffel and only
allow subclasses to weaken the preconditions or strengthen the
postconditions of the parent. Weakening involves adding ORs,
strengthening ANDs. (More Foo and Bar pseudocode. Oh, and apparently
Schwern only has 2 toes).
There was a short sidetrack on having multiple versions of modules
running at the same time, and Allison Randal informed us that it looks
like the full name of classes will include their version number, ie
"Acme::N-1.1".
In fact, a lot of the discussion is a bit in the air: as Dan Sugalski
pointed out, things like object attributes aren't until Apocalypse 12
and it may be a little early to worry about such things.
<http://makeashorterlink.com/?Q19B21502>
<http://makeashorterlink.com/?H5AB54502>
Subject-Oriented Programming
Schwern also asked about subject-oriented programming, which looked
interesting but which he couldn't quite understand. Andy Wardley
explained that all these "advanced" programming techniques are all
attempting a clear separation of concerns, and went on to describe and
give pointers to more info.
<http://makeashorterlink.com/?K1BB21502>
Matching
Someone mysteriously known only as "Ed" asked what the favoured syntax
would be to match negative multi-byte strings in Perl 6. It wasn't
entirely clear what the question was, but one thing is sure: the Perl 6
pattern matching engine will have a lot of scope for optimisation.
<http://makeashorterlink.com/?S5CB11502>
In brief
Tim Bunce asked if there were any good tools for static code analysis
around. None, apparently.
Brent Dax reminded Andy Dougherty and others about
"/Parrot_v?sn?printf(_[sc])?/" in "misc.c" - a reimplementation of the
printf functions for portability reasons.
Simon Glover added tests for the assign opcode.
It looks like Michael Lazzaro will be writing a list of issues with OO
as well as a tutorial so that everyone is clear what exactly we are
talking about.
Larry cringes every time someone says "Parens construct lists in Perl
6".
Who's Who in Perl6
Once more we get to meet people involved in the development of Perl 6.
Who are you?
Simon Cozens
What do you do for/with Perl 6?
I was the Parrot pumpking up until 0.0.4, and wrote much of the PMC
and string infrastructure. I then escaped to get a life, play more
Go and be a nicer person.
As perl.com editor and occasional author, I pop in from time to time
just to check on the correctness of things I'm writing about and
make sure Parrot compiles on at least one platform I possess so I
can test code for articles. Rumours of my return to development have
been greatly exaggerated. :)
Where are you coming from?
I'm not, I'm already here!
When do you think Perl 6 will be released?
I don't think Perl 6 - as we understand it - ever will be released.
How's that for a Delphic pronouncement?
Why are you doing this?
Because I like food.
You have 5 words. Describe yourself.
Less vehemently obnoxious than before.
Do you have anything to declare?
Watch out for some surprises in the near future.
Acknowledgements
This summary was brought to you with much distraction indeed from Super
Mario Sunshine. Thanks to Kate Pugh and #london.pm for proofreading.
As Piers says: One more, if you think this summary has value send money
to the Perl Foundation <http://donate.perl-foundation.org> and feed back
and/or T?iBooks to me, <mailto:[EMAIL PROTECTED]>. As usual, the fee
paid for publication of this summary on perl.com has been donated
directly to the Perl Foundation.
Enjoy, Leon
--
Leon Brocard.............................http://www.astray.com/
scribot.................................http://www.scribot.com/
.... Bioengineers wear designer genes
