develooper Front page | perl.perl5.porters | Postings from February 2007

forked Perl or patched Perl / Re: Future Perl development

Thread Previous | Thread Next
From:
Scott Walters
Date:
February 6, 2007 12:07
Subject:
forked Perl or patched Perl / Re: Future Perl development
Message ID:
20070206130122.GL2993@straylight
Seems to me like the usual schism between "official" and "unofficial".
Matthijs van Duin has patch that gives you a JavaScript . like operator
that allows you to write $foo`bar`baz instead of $foo->{bar}{baz}; it
was rejected, but no one said anything about it being banished to the
nether reaches of the 'net.  For a while, autobox required a patch to
core.  Bundling patches with modules is one avenue, but the adventurous
would likely rather get all of them at once rather than rebuild all of
the time.  Perhaps we need a http://patched.perl.org that distributes
copies of the source with "community maintained" patches (in the
requestor's desired combination).  I'd love to see an increasely 
self-hosting Perl 5, along the line of David Nicol's work.  I'd love
to play with a lot of things, or even work on them, that aren't backwards
compatiable (yet) or production ready.  

Most projects that fork vanish quickly.  Keeping up with all of the
changes in the main branch is a lot of work.  Automatically and
implicitly importing all changes from the main branch and allowing
patches to break (hopefully to be fixed soon afterwards) is easiler.
Couldn't find the link, but someone set up such a patch server for
dillo (whose maintainers were notorious about ignoring patches).
People wrote patches for tabs, ssl, i18n, and other things, and
eventually, someone set up a server that tracked the status of the
patches (build cleanly? dillion passes tests with it installed?
last update?).

In short, present possible future development lines as what they
are -- potentially unmaintained, experimental patches -- give the
idea some official status without giving any official status to
any particular patch, and let the adventurous sort it out.

I might have some room on slowass.net to host this ;)

-scott

On  0, David Nicol <davidnicol@gmail.com> wrote:
> On 2/6/07, demerphq <demerphq@gmail.com> wrote:
> >On 2/6/07, chromatic <chromatic@wgz.org> wrote:
> >> On Tuesday 06 February 2007 02:11, demerphq wrote:
> >/I/ am /not/ going to fork Perl5 anytime soon, likely never,
> >certainly its not on my todo list for the next decade. I'm just
> >pointing out the obvious, unless our community sorts this schism out
> >thats whats going to happen. Goosen has already made a first step, and
> >I know that in private rumblings others have mentioned it as well.  If
> >Perl5 does get forked I would think very seriously about working on
> >it. And I don't think im alone.
> >
> >Yves
> 
> I tried forking perl some time ago; the general plan was to write a perl 
> parser
> in perl -- no not as a wrapper around an C<eval> like Acme::Inline::PERL, 
> but
> as a serious implementation of the language as specified, using pure perl to
> implement it.  Towards dumping the resulting structures to something that 
> GCC
> could handle; as well as having an easier-to-hack-on internals.
> 
> I called the project "TERN" which was an abbreviation for TERN Extensible
> Rewriting Notation and even gave a talk on it at YAPC St. Louis, which
> absolutely
> sucked due to the convention neatly lining up with the perigee of my
> regularly scheduled
> moodswing.
> 
> Anyway anyone is welcome to sign up for the TERN mailing list at
> http://savannah.nongnu.org/mail/?group=tern
> 
> there is no traffic on it.
> 
> 
> Basically implementing Perl 5 is a lot of work as there are a lot of
> little pieces to it
> that have stuck on over the years.  And there really isn't a whole lot
> of motivation to
> do it aside from wouldn't-it-be-cool which rarely lasts more than a few 
> hours.

Thread Previous | Thread Next


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