Re: C++ version of regexprep.cc

[Andy Adler]

On Thu, 4 May 2006, Alois Schloegl wrote:
> I was able to read it using a modified version of XMLTREE. I've modified
> XMLTREE in such a way that the colons ':' are replaced by underscores
> '_'. Then, the names became valid fieldnames such that
> <abc:de value> converts to xml.abc_de = 'value'. An alternative solution
> could be, that the colons are used to define subfields such that the
> results is xml.abc.de = 'value'. Perhaps, this would be a nicer solution.

Alois,

Note that this is a very non standard way to parse XML.

The colon is a name space designator. Consider thse documents

    <doc xmlns:ABC="http://my.xml.name.space";>
      <ABC:de value="1"/>
    </doc>

and

    <doc xmlns:XYZ="http://my.xml.name.space";>
      <XYZ:de value="1"/>
    </doc>

The XML infoset spec requires that these documents be
considered to be EXACTLY equivalent.  This means that
any XML processing application which treats these
documents differently is techically NOT VALID.

Do we want octave XML parsers to correctly deal with
these kinds of issues.

--
Andy Adler <address@hidden> 1(613)562-5800x6218