I found that using Text::CSV_XS as a parent class didn't really work.
For one, it's XS, so debugging is trickier, and for another I need to store
extra information.
It turned out that extending by encapsulation was much simpler and cleaner.
As for exposing fields, I'm happy to take that feature.
With your permission, I'll add a MARKSTOS@cpan.org account to my repository.
Adam K
On 12/12/2007, MARKSTOS via RT <bug-Parse-CSV@rt.cpan.org> wrote:
Show quoted text>
>
> Tue Dec 11 17:01:36 2007: Request 31389 was acted upon.
> Transaction: Ticket created by MARKSTOS
> Queue: Parse-CSV
> Subject: WISH: expose 'fields' method
> Broken in: 1.00
> Severity: Wishlist
> Owner: Nobody
> Requestors: MARKSTOS@cpan.org
> Status: new
> Ticket <URL:
http://rt.cpan.org/Ticket/Display.html?id=31389 >
>
>
> When using new( fields => 'auto' ), it can be useful to know which
> fields were discovered before a row is fetched.
>
> (For example: to use the fields to create a table based on the CSV field
> names before inserting into it).
>
> Text::CSV_XS can already do this with the fields() method. That method
> just needs to be exposed.
>
> Alternately, the whole Text::CSV_XS object could be exposed, or used as
> a parent class.
>
