David Vergin wrote:
That way, part of the development of a new template would be
to document its data needs in a leading comment in the template itself
or in the companion file. Then the rest would take care of itself.
That last part is kind of key...
I'm not sure exactly what problem you're trying to solve, but various
tries have been made at a system where the templates control what data
gets pulled.
The simplest way is to use DBI directly from your templates. The
arguments for and against are pretty well worn, and I expect you know
what they are and chose not to do this.
A more complicated way is to have some sort of data access layer in perl
which your templates talk to. This would mean something like
Class::DBI, Rose::DB::Object, DBIx::Class, Tangram, etc. You load the
data class you need from your template and call methods on it to get
data. If you auto-generate these classes from your database, you
theoretically can have nothing but a database, some templates, and a
generic driver script. It's ultimately the same effect as your template
metadata idea.
Is this where you wanted to go with it?
- Perrin
_______________________________________________
templates mailing list
[email protected]
http://lists.template-toolkit.org/mailman/listinfo/templates
