Re: [xml] Parse dtd

[Daniel Veillard <veillard redhat com>]

On Sun, May 08, 2005 at 08:37:37PM +0200, Rene de Zwart wrote:
> I'm experimenting to make it possible to
> 1) Dtdparse (http://dtdparse.sf.net with libxml2)
> 2) do in vim as in emacs with psgml-mode. 
> I need to parse a dtd for that and since libxml2 does that
> already I use that.
> 
> I let libxml2 parse my dtd with the code below.
> I'm stuck now with getting the dtd tree from libxml2.
> I'm not getting anywhere with valid.c, hash.c, tree.c.
> My questions are 
> 	1) how do I loop/walk over the dtd structure to get the
>        information.
>        I envision something like:
>        Element        OPTIONS  ATTR     
>        book           Req,...  id,lang  
> 	       chapter    Req,...  id,lang  
> 		       title  Req,...  id,lang  
> 		       para   +,...    role,lang          

  walk the children list won't help in general. See valid.c code to
lookup informations in the element and attribute hash tables associated
to the internal and external subset of a document.

>    2) Why does 
>               xmlParseChunk(ctxt, Buff, sizeof(Buff),0)
>       report  "memory:4: parser error : Document is empty"
> 	  It seems illogical because I haven't said that I'm
> 	  finished with xmlParseChunk (That would require the
> 	  4th parameter to be !0.
> 
> 	
> static char Buff[] = {
> "<?xml version='1.0' encoding='UTF-8'?>\n"
> "<!DOCTYPE book PUBLIC '-//OASIS//DTD DocBook XML V4.3//EN' \n"
> "'http://www.oasis-open.org/docbook/xml/4.3/docbookx.dtd'\n"
> ">"
> };

  ugly, note that buff includes the 0 terminating the string so
Buff[sizeof(Buff)] == 0 and the parser find an end of file and rightly
complains.

  I strongly suggest you use a debugger while working on DTD lookup
code. A debugger would also have pointed out 2) fairly easilly I think.

Daniel

-- 
Daniel Veillard      | Red Hat Desktop team http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/