[libxml++] RE: [xml] SaxParser: getEntity() callback

[Murray Cumming Comneon com]

> From: Daniel Veillard [mailto:veillard redhat com] 
> On Wed, Oct 22, 2003 at 11:16:03AM +0200, 
> Murray Cumming Comneon com wrote:
> > Has anyone ever successfully used the getEntity() callback? 
> It needs 
> > to return an xmlEntity instance, but I can't imagine that 
> anyone could 
> > create one with all the right voodoo contents: 
> > http://xmlsoft.org/html/libxml-tree.html#xmlEntity
> > 
> > I can't find any way to get the xmlEntity that would have 
> been created 
> > by libxml automatically if I had set the getEntity() to NULL. That 
> > might at least be a starting point.
> > 
> > I know this is difficult, but I hope it isn't impossible.
> 
>   1/ Look at SAX2.c implementation 

I think this might be easier if that default implementation was publically
accessible. I think that's why it's easier with the DOM Parser. The DOM
Parser lets us resolve the entity reference if we like, or do our own thing.
With the SAX parser it's all or nothing.

I'm not sure exactly what part of SAX2.c would need to be public API.

>   2/ Look at entities.c in 2.6.0 for the predefined static entities
>   3/ Look at the XML-1.0 spec about the various processing of entities
>      http://www.w3.org/TR/REC-xml#entproc
>      consider the complexity of the given table
>   4/ Consider the fact that libxml2 may either substitute entities or
>      keep them in the tree (including in attributes content)
> 
>  You will get an idea of what's difficult, that it's not 
> impossible (since libxml2 does it !), but that you will have 
> to duplicate code behaviour from libxml2 in your SAX handler. 
>  If you want proper entities processing the tree or xmlReader 
> based approach are definitely simpler since libxml2 take care 
> of the complexity.

Thanks. I think I'll just explain the complexity in the libxml++ reference
documentation, and leave it to future generations to make it simpler and/or
provide an example.

Murray Cumming
www.murrayc.com
murrayc usa net