Parrot aka Perl 6 info
rdsteph at earthlink.net
Sun Dec 30 10:26:49 EST 2001
This is a follow on to thread from here that Dan Sugalski (perl
implementer) posted over on comp.lang.ruby, that didn't show up here. I
copy it here fyi only, just to show the firm intent of the Perl
implementers to create a runtime that will run both ruby and Python as
well as Perl. FWIW:::
The folks over at comp.lang.ruby ae quite interested in this.
I suppose if anyone has any follow up questions, they could be directed
to Dan at his email address listed at the end of his post. I wonder how
valuable such a "new runtime environment" for Python might be? I wonder
if the Python core team is well aware of the Perl 6 efforts?
At 07:57 PM 12/29/2001 -0800, ptkwt at aracnet.com wrote:
>In article <188.8.131.52.2.20011229222259.0206fe80 at pop.sidhe.org>,
>Dan Sugalski <dan at sidhe.org> wrote:
> >At 09:57 AM 12/30/2001 +0900, Ron Stephens wrote:
> >>However, for what its worth, the PERL 6 implements, whom I met and
> >>with at the MIT Lightweight Languages conference, are definitely
> truing, as a
> >>goal, to create a PERL 6 runtime environment that can easily and
> >>someday later on, run Perl, Ruby, and Python. I suppose that could
> >>runtime environments for these languages, just like Jython is an
> >>runtime environment for Python. Maybe that could work, and give some
> >>advantages of "bringing the tow (or , in this case, three) streams
> >>development back togethers into one mighty river.
> >>I am a doubter about this, btu who knows?
> >Oh, don't doubt. We've pulled it off already, to an extent, with the
> >engine, and that's a mess as these things go.
> >Technically, an engine capable of running all three languages (and
> >pretty trivial as these things go. Inter-language calling conventions
> >the toughest and that's not really tough at all. The politics
> >by far the biggest hurdle, and they exceed the technical difficulties
> >many orders of magnitude.
> >Having said that, Parrot will be *capable* of running perl, python,
> >ruby. (Whether there's a full-blown language parser for python and
> >up in the air--I'm all for the ruby parser in ruby, as that'd make my
> >really easy.
>So what do we have to do? Parse Ruby and spit out Parrot bytecode,
For now, yep. Once we get things situated right, you won't even need to
spit out bytecode. Spitting out a full AST would be sufficient, which
gets fed into the compiler module.
>I think there is a project or two out there to create a Ruby parser in
>Ruby. JRuby could always be modified, I suppose, to become PRuby -
>might be another option.
Sure, and if you have a ruby parser in ruby, we can run it through the
engine as it stands now and spit out ruby bytecode. We run that through
ruby bytecode->parrot bytecode converter (which is already partially
written) and get parrot bytecode which we then freeze to disk and poof!
have a working parser.
> >We'll have full bytecode translators, but those are trivial
> >compared to parsing the languages) Even if it's not fully used, it'll
> >Parrot a better engine.
> >And anyway, at this point I can wave the "Hey, look, we have a JIT!"
> >to entice people into writing a ruby parser for Parrot... :)
>I think that there are a lot of Rubyists out there who want to see a
>parser for Parrot and some who would even be willing to work on it.
>wouldn't it be cool if we could get there before the Pythonistas?
Yep, darned cool. Heck, we don't have a full perl parser yet--you could
even beat us! (Which, personally, I think would be amusing)
Dan Sugalski even samurai
dan at sidhe.org have teddy bears and even
teddy bears get drunk
More information about the Python-list