Re: Subclassing HTML::Parser to support $p->include()

On 24 Feb 2006, at 17:46, Terrence Brannon wrote:
> If you want to stay pure HTML, then do this:
>
> <span myparser_action="include" myparser_args="header.html" />

The syntax was just chosen for conciseness but the actual syntax  
looks like

<ido:include src="blah.html" />

I /assumed/ (perhaps mistakenly) that HTML::Parser would be OK with  
arbitrary tag names provided everything was syntactically correct. Is  
that a stupid assumption?

> No need to use non-HTML and expect an HTML parser to parse it. If  
> you want
> non-HTML, then maybe XML parsing is more appropriate.

That's certainly a possibility - although a quick glance at  
XML::Parser suggests that I'd have at least the same - and possibly  
worse - problems implementing an include() method on it.

> Also, what you want appears to be do-able using the formerCPAN module
> HTML_Tree by Paul J Lucas:
> http://homepage.mac.com/pauljlucas/software/html_tree/

OK, that may be an option - although I have a working template  
compiler built around HTML::Parser already so I'd rather keep going  
in that direction unless it's a dead end :)

> And you might like the CPAN module PeTaL or my own HTML::Seamstress
>
> and XML::LibXML and XML::LibXSLT is XML floats your boat :)

OK :)

> why not use HTML::Tree on CPAN (HTML::Tree is on CPAN, HTML_Tree is  
> at the
> URL I gave you previously).

Build parse trees and then merge them? I could do - altenately I  
could introduce a preprocessor stage using HTML::Parser that expanded  
the include tags and generated HTML source with the included chunks  
included and then parse that a second time. Ideally though I'd like  
to handle the include directive at the same stage where I handle my  
other tags - for various architectural reasons I won't bore you with.

Thanks :)

-- 
Andy Armstrong, hexten.net