Re: [Sax-devel] Should I use sax?

[Roberto C. <li...@ro...>]

Hi David. Thank you for your clarification.
It is clear, SAX is the way to go for me.
I confirm that sax is reading the doctype and doing the validation
accordingly. It was missleading that the XMLSerializer was outputting
the "Strict" doctype unregarding of the input one.

I have one more question:
Parsing a document which contains the fragment
<a href="mylink.html" >mylink</a>, generates events also for the
a-attribute 'shape="rect"', which is not given in input!!
Is there any way to turn this behavior off? It seems like the parser is
inserting some implicit attributes here and there...
/Roberto

David Megginson wrote:
> On 22/05/06, Roberto Cosenza <li...@ro...> wrote:
> 
>> I'm writing an application that has to parse and MODIFY some XHTML.
>> The result of the operation has to be the modified XHTML.
>> I started using SAX (xerces) and, though seemed to work, I found out
>> that SAX does not keep my doctype (all parsed documents take the XHTML
>> Strict doctype). Besides, I was using the XHTMLSerializer class which is
>> deprecated.
>> Maybe I should use DOM for this job? The document is usually small. I'm
>> missing some point?
> 
> There are two conditions for using an in-memory object tree like the DOM:
> 
> 1. Your documents are relatively small.
> 2. Operations are relatively infrequent.
> 
> For example, if you're dealing with 100 MB documents, the DOM is
> definitely out; likewise, if you're working on a server that will
> handle hundreds or thousands of requests per second or a
> batch-processing system that will open tens of thousands of XML
> documents quickly, then DOM is probably out.  If neither of these
> applies, then go ahead and use the DOM or another in-memory object
> tree.
> 
> If you do need to use SAX, then the first thing to understand is that
> SAX does not actually include a writing component -- whatever Xerces
> uses is outside of SAX proper.  SAX2 does support *reading* DOCTYPE
> information through its optional LexicalHandler interface:
> 
> http://www.saxproject.org/apidoc/org/xml/sax/ext/LexicalHandler.html
> 
> Finally, unless you do something special inside the DOCTYPE
> declaration (like declaring entities or extending XHTML in some way),
> you can just generate a boilerplate DOCTYPE when you write out your
> document.
> 
> 
> All the best,
> 
> 
> David
> 

-- 
Roberto Cosenza
ICQ 12231605, MSN & Jabber robcos AT robcos.com
Tel: +46-(0)70-4660928
Work Tel: +46-(0)8-55576860, Fax: +46-(0)8-55576861