Re: [xml] question about how to set namespace in libxml

[Michael Ludwig <milu71 gmx de>]

Dan Quach schrieb am 01.03.2011 um 12:05 (-0800):
> Currently I am reading in the following xml document (through a ruby
> wrapper)
> LibXML::XML::Document.string( saml_plain)
>
> <ds:SignedInfo>
> <ds:CanonicalizationMethod Algorithm='
> http://www.w3.org/2001/10/xml-exc-c14n#'/>
> [â]
> </ds:SignedInfo>

That document is not namespace-well-formed.

> The problem when it is read in that all of the ds namespaces are
> stripped and errors occur

This is sort of a fallback parsing as namespaces are an add-on to XML,
which is not in the XML recommendation.

> Is there a way I can set the 'ds' namespace as '
> http://www.w3.org/2000/09/xmldsig#' so it would be read in as:
> <ds:SignedInfo xmlns:ds="http://www.w3.org/2000/09/xmldig#";>
> [â]
> </ds:SignedInfo>

You can simply edit your document.

If you want to edit your document programmatically, you could parse it
using SAX with namespace-awareness turned off and have a filter insert
an xmlns:ds attribute, which in non-NS mode really is just an attribute
with a name that happens to contain a colon. You would then build a
document *with* namespaces from the result of that preprocessing step.

Well, I'm not quite sure this is possible with LibXML2. It is with Java
Xerces ( void setDoNamespaces(const bool) ) [1], but I can't find such
an option for Perl's LibXML2 bindings. [2] Don't know the C API. Anyone?

Namespace ignorance does not seem to be possible with the XmlReader
interface. I'm saying this because I can't find the option in the Perl
interface [3].

You could also use a non-XML process, such as Perl and regular
expressions to fix thousands of such documents.

-- 
Michael Ludwig

[1] http://xerces.apache.org/xerces-c/program-sax-3.html
[2] http://search.cpan.org/~pajas/XML-LibXML-1.70/lib/XML/LibXML/Parser.pod
[3] http://search.cpan.org/~pajas/XML-LibXML-1.70/lib/XML/LibXML/Reader.pod