Re: [xml] XML validation

[Daniel Veillard <veillard redhat com>]

On Fri, Dec 19, 2008 at 06:47:34PM +0100, Francesco Gennai wrote:
> Scenario:
> 
> I need to validate an XML file received via e-mail.
> It has no <!DOCTYPE element.
> The DTD file is stored locally (on the receiving system).
> Reading some thread on this list and the documentation I have understood that 
> the best validation can be obtained if the DTD file is referenced by the 
> XML file.
> 
> Since the XML file is a short file, my idea is to add the line
> 
>  <!DOCTYPE eulotnw SYSTEM "/dtd/xxxx.dtd">
> 
> to the file, before to pass it to 
> 
>   xmlCtxtReadFile(pctxt, xmlfilename, NULL, XML_PARSE_DTDVALID);
> 
> Is this a good idea to get the best validation of the XML file ?

  That's one possibility, the most correct from an XML perspective
except you need to make some check on where to insert the DTD.

> Any other suggestion?

  Do a normal parsing, load the DTd separately with xmlParseDTD
and then do an a posteriori validation of the parsed XML with
xmlValidateDtd(). Search for dtdvalid options and strings in the
xmllint.c code, it does just this.

Daniel

-- 
Daniel Veillard      | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
daniel veillard com  | Rpmfind RPM search engine http://rpmfind.net/
http://veillard.com/ | virtualization library  http://libvirt.org/