> > If the generator was used as the primary way to testing the regex
> > engine, do you really think that any of these limitations would
> > exist?
>
> Sigh. [Because] seems to have flown right by you.
Ok, I think this thing has pretty much played itself out, but I hate ending
on a misunderstanding. So I hope that this post clears things up. Two points:
1) [Because] has not 'flown right past me'. I realize that "we're an
open source project" and that putting it in the core doesn't make
it "magically implemented".
2) I realize that there are three options: operator, core-module and
CPAN module.
Just to be clear, what I'm arguing is that, by putting it in the core,
and by adopting a rigorous methodology that requires testing regex and
generator - feature for feature - you *do* pretty much get the development
of the generator for free.
Why? Because - collectively - in the long run you save more time with the
rigorous regression testing than you spend in developing the generator.
Yes, I'm willing to put together a prototype, and extend Regexp::Genex for perl5.
And I'm also willing to bounce ideas and syntax off of the perl6 list, and put a
prototype together for perl6. And, yes, if the module/operator was accepted for
core I'd help maintain it.
But since its a timesaver in any case, I *shouldn't* need to be intimately
tied to it. An idea is an idea - and there are lots of good 'vaporware'
ones listed in perl6's apocalypses right now. At some point, Larry is going to
have to prioritize them. I sincerely doubt that rev 1 of the language is going
to include everything listed there.. So why not list it if not just as a
reminder of its existence?
Anyway, I'm going to give this a bit more of a think. I still believe its better
off as an operator, but I see that this idea needs some work to be truly convincing.
Ed
(
ps - is YAPE::Regex the best regular expression parser out there? I sort
of shudder at the idea of parsing perl5's regular expression syntax from
scratch...
)
