On Wed, Oct 03, 2001 at 09:48:32AM +0200, Josi Luis Sancho wrote:
> I am using XML::Grove and Data::Grove::Visitor to parse SGML/XML
> documents; the visitor package generates calls back by element names;
> unfortunately I face dot-named elements and I MUST parse them by means of
> identically dot-named subroutines.
You don't -have- to. There are various alternatives.
You could encode all of your subroutine names in a consistent manner. For
example, "foo_bar" would become "foo_5fbar", and "foo.bar" would become
"foo_2ebar" (the encoding, btw, is basically URL encoding, but with _ as the
prefix, instead of %). However, that's ugly, so I much prefer this upcoming
solution.
Use a hash. You're basically using the symbol table as a hash, make the
relationship more formal. The keys of the hash are your element names, the
values code references.
%elements = (
"foo.bar" => sub { ... },
"foo_bar" => sub { ... },
);
Then, instead of having to say "&{'foo.bar'}(...)" you say
"$elements{'foo.bar'}->(...)".
With the solution you eventually found, even though you can create a
subroutine with a dot in it, you still can't call that subroutine as you
would one without such a dot. the code "foo.bar(...)" still causes warnings
and errors; you'd have to say "&{'foo.bar'}(...)". Using a hash, to me, is
much preferable because it has the added benefit of not polluting the symbol
table.
Michael
--
Administrator www.shoebox.net
Programmer, System Administrator www.gallanttech.com
--
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
