[xsd-users] namespace clash: suggestion for improvement

[Jaume Dominguez Faus]

Hi, list.

I am just upgrading to XSD 4.0 and I stumbled into this.

In file serialization.txx lines 98 to 116 we find this code

        if (!x.null_content () && x.dom_content ().present ())
        {
          // Clone the contents of the element.
          //
          using namespace xercesc;

          DOMDocument& doc (*e.getOwnerDocument ());
          const DOMElement& se (x.dom_content ().get ());
          DOMNamedNodeMap& sa (*se.getAttributes ());

          for (XMLSize_t i (0), n (sa.getLength ()); i != n; ++i)
            e.setAttributeNode (
              static_cast<DOMAttr*> (doc.importNode (sa.item (i), true)));

          for (DOMNode* sn (se.getFirstChild ());
               sn != 0;
               sn = sn->getNextSibling ())
            e.appendChild (doc.importNode (sn, true));
                }

Which is fine in non-windows operating systems (I successfully upgraded in Linux and Mac). But it happens that in Windows this might cause a name clash with MS xml library (as follows) if another library includes it or similar.

D:\SDKs\xsd-4.0.0\include\xsd/cxx/tree/serialization.txx(104) : error C2872: 'DOMDocument' : ambiguous symbol
could be 'c:\program files (x86)\windows kits\8.1\include\um\msxml.h(183) : DOMDocument'
or 'D:\SDKs\xerces-c-3.1.1\win64-vc-120\include\xercesc/dom/DOMDocument.hpp(64) : xercesc_3_1::DOMDocument'

Since XSD is basically a header-only library and it is suggested that using namespace should be avoided in includeable files, maybe it is a good idea to remove the using namespace xercesc statement in line 102 and use it in all the types like this instead

       if (!x.null_content () && x.dom_content ().present ())
       {
          // Clone the contents of the element.
          //


          xercesc::DOMDocument& doc (*e.getOwnerDocument ());
          const xercesc::DOMElement& se (x.dom_content ().get ());
          xercesc::DOMNamedNodeMap& sa (*se.getAttributes ());

          for (XMLSize_t i (0), n (sa.getLength ()); i != n; ++i)
            e.setAttributeNode (
              static_cast<xercesc::DOMAttr*> (doc.importNode (sa.item (i), true)));

          for (xercesc::DOMNode* sn (se.getFirstChild ());
               sn != 0;
               sn = sn->getNextSibling ())
            e.appendChild (doc.importNode (sn, true));
        }


Doing like that I can build without problems (except for the fact I could need to "hack" if newer versions don't consider this).


Just a suggestion.

Regards.