* Joshua Hoblitt ([EMAIL PROTECTED]) [25 Jun 2003
16:05]:
> > I've just checked some experimental code into CVS. Take
> > a read of t/dispatch.t and
> > lib/DateTime/Format/Builder/Parser/Dispatch.pm (damn
> > these paths are getting long).
> >
> > Something like that?
> Wow - Yes.
The idea came to me while in the shower. I said it would
have to percolate for a while =)
> The possibility for recursion is just Evil{TM}.
It's actually a key feature of Builder. It used to be far
more evil and recursive but it's been toned down a bit.
Now, however, you could have a method that ends up calling a
Dispatch, that tries assorted things, ends up calling
another Dispatch, etc.
It gets quite messy in one's head, really. My next step is
to reintroduce verbose mode since debugging is getting to be
somewhat complex.
> I'm really really impressed with the DT::F::B::Parser
> framework, this module required so little code.
Though there is an evil global floating around now.
There's a lot of latent power in the system. After the
internals redesign to allow the separate Parsers, I can
mostly leave it at "if someone wants something, I'll
implement it".
> Now the question is should I change my ISO8601 parser to
> use multi-length or group parsing? :)
If something is of a static length, use lengths.
Note that a Dispatch parser can have a length. Which works
as one might expect it to. Though I'm starting to have
difficulty keeping potential combinations of things in my
head =)
> Ps. Tell Simon to say this is cool too.
He's away =)
http://blog.simon-cozens.org/bryar.cgi/id_6461
cheers,
--
Iain. <http://eh.org/~koschei/>
