Front page | perl.perl6.language |
Postings from December 2005
This week's summary
From: The Perl 6 Summarizer
December 8, 2005 12:33
This week's summary
Message ID: email@example.com
The Perl 6 Summary for the week ending 2005-12-04
I heard a rumour on the London.pm mailing list week. Apparently the Perl
6 Summaries are no longer being published. As I'm sure you can imagine,
it came as something of a surprise to me.
This week has been all about Parrot, Leo's got the new lexical scheme,
calling conventions and exception handlers working and made Parrot
stricter about arguments. The end of the week saw the release of
'Luthor', version 0.4.0 of Parrot. Read on for more details...
This week in perl6-compiler
Um... one post in perl6-compiler this week. And that was crossposted to
perl6-language. And because it got posted at the end of the week, none
of the actual discussion occurred this week.
Moving swiftly on...
This week in perl6-internals
Much more going on here as everyone rushed towards the release of Parrot
0.4.0 "Luthor" at the end of the week.
Exception handlers and calling conventions
As I predicted last week, Leo's brain dump about exception handling got
discussed this week. It was well liked, and after a small bit of sugar
was sprinkled on to make ParTCL's life a little easier (and possibly
unsprinkled later) all manner of things were well.
Subs may not contain dynamic call info, ever
Chip posted a clarification of his comments on what data could and
couldn't be hung off a Sub object at runtime. Let's be reentrant people.
Will Coleda announced that ParTCL is now working with the new lexically
lovely Parrot calling conventions. There was much rejoicing.
Test::More and Tests in PIR
Leo showed the love for chromatic's shiny pure parrot implementation of
Test::More. So the patch was applied.
Leo announced the scratchpad's impending doom and outlined the planned
change for comment. Nobody commented, and the changes went in.
Parrot directory reorganization
Quick quiz: where would you expect to find tests in the parrot
distribution? How about generated source files?
Jerry Gay proposed a reorganization to make things a little more lovely.
The consensus seemed to be that a reorg along Jerry's lines wouldn't be
a bad idea, but Chip pointed out that, whatever gets done it should be
done 'cautiously so as to minimize unpleasantness'. So Jerry is
proceeding cautiously, starting with a host of new TODO tickets in RT.
Solving "=" confusion: ':=' for aliasing
There are those of us who are wondering why this one took so long...
Chip proposed that people start to spell aliasing as ":=" and assignment
as "=". I think it's a really good idea, but then I don't have a large
amount of PIR code to maintain so what do I know. Some other folks
weren't so sure, but Chip is not to be denied. Discussion then span off
into what language to write the automagical translator in.
I believe this may involve writing it in PIR then converting it to PIL,
which would be converted to Perl 5 using pugs and then Larry's Perl 5 to
Perl 5 project could be used to convert it to XML, which could then be
modified using XSLT and converted back into PIR using some scary voodoo
Or they could just write it in Perl 5. Prosaic, but possible right now.
Chip announced that he'd put up another revision of PDD03 on Parrot
calling conventions. Most of the changes are simple clarifications and
flag renaming, but he's also proposing a new "READONLY" flag for
"get_params" to make it easy to support the default Perl 6 argument
mode. Response was muted, but favourable.
PDD03 and Overflow/Underflow
It's been mandated for ages that Parrot should throw an exception when
functions get called with the wrong number of arguments. It's always
been one of those things that will be implemented 'some day'. Well this
week had a someday in it as Leo made parrot do what it's supposed to do.
And broke PGE for a while...
Jonathan Sillito is a class act. He didn't just ask a bunch of questions
about the new PDD20 on lexical variables, he promised to take the
answers he received and use them to patch PDD20 to make things clearer.
Spurred on by this promise, Chip was unstinting in his answers and
clarifications of them. Which is nice.
PIR methods and local variables occupy the same 'namespace'
Allison Randal used Snarks, Boojums and Thingies to demonstrate a
possible problem with the way Parrots local variable and method
namespaces overlap. Leo pointed out that this can sometimes be useful.
So, for the time being, Parrot continues as is in this area. If you go
getting the bowsprit and the rudder mixed up, it's your own silly fault.
After "many months and lots of work," Matt Diephouse returned from the
mountain top bearing a couple of stone tablets. One was labelled
'Namespace Spec' and the other '[Draft]'. Muttering dark incantations
about giving a discussion eyes, he laid out their contents and asked for
comments. And low, there were comments in abundance. I would attempt to
summarize both the spec and the ensuing discussion but, alas, the margin
here is narrow and my time is short. Suffice it to say that if
namespaces are your bag, you should read this thread.
HLL and autoboxing
Okay, who else has an image of two of those transformer robot things
from the Renault Megane advert punching each others' lights out?
I'm afraid I didn't quite understand François Perrad's question about
maps and autoboxing. Luckily Roger Browne and Leo did.
Punie's demo likes its memory
Jerry Gay worried that Punie is using way too much memory, running out
of the stuff on his 512Mb machine. He wondered if it was a bug or to be
expected. Luke thought it might be an issue with the algorithm used in
the attribute grammar implementation which is somewhat less than
abstemious with memory. Nick Glencross noted that, after he'd done a
nuke of the distribution and rechecked it out, the problem seemed to go
away. So, it's either fixed, or there was something odd going on
Parrot 0.4.0 "Luthor"
Leo announced the usual freezes and parrot proceeded smoothly down the
road to release. After the release he discussed goals for the next
release and asked for comments.
Meanwhile, in perl6-language
A quiet week here too.
Capabilities in Perl 6
Rob Kinyon has read slides on CAPerl and had some thoughts about how to
build the idea into Perl 6. Larry and Yuval Kogman commented on it.
Acknowledgements, apologies and everything else
It's not exactly Monday today is it? Well, I could lie and claim that I
was waiting for the Perl Foundation blog to be announced before I posted
the summary, but that would be Wrong. Instead I shall claim (with some
justification) that there's an awful lot of things that need sorting out
when you're moving house, and they take far longer than you expect.
My ankle and shin are progressing nicely, thanks for asking, and there
has been a good deal less random swearing in the Cawley household this
<http://blog.perlfoundation.org/> -- The Perl Foundation blog
<http://geeksunite.org/> -- Chip still needs help.
The usual coda
If you find these summaries useful or enjoyable, please consider
contributing to the Perl Foundation to help support the development of
<http://donate.perlfoundation.org/> -- The Perl Foundation
<http://dev.perl.org/perl6/> -- Perl 6 Development site
Check out my website, it's lovely.
Vaguely pretty photos by me can be found at:
Piers Cawley <firstname.lastname@example.org>