Re: Namespace conformance & SAX2

[Rob Lugt <roblugt@elcel.com>]

Jonathan Borden wrote:

> > Yes, perhaps my wording was unfortunate.  How about if I reword my
> > explanation like so:-
> >
> > "Syntax errors are well-formedness errors in XML 1.0 and well-formedness
> > errors are reported as fatal errors.  Therefore, it follows that syntax
> > errors on the QName or NCName productions should be reported as
> > fatal errors
> > to be consistent."
> >
>
> no. XML 1.0 does not define a 'syntax error' as a well-formedness error.
see
> http://www.w3.org/TR/REC-xml#sec-well-formed. Well-formedness constraints
> applicable to XML are defined in the XML 1.0 rec, not in any other
document.
> These are labelled "WFC" in the spec. Similarly "Namespace Constraints"
are
> labelled "NSC" in the namespace spec. Do a grep and you'll see.
>

Oh boy this is as easy as stepping through treacle.

My words above "Syntax errors are well-formedness errors in XML 1.0" were
chosen carefully.  Notice I didn't write "well-formedness constraint".
Proof for the above statement can be obtained from XML 1.0, section 4.3.2
which states: "The document entity is well-formed if it matches the
production labeled document.".  So, if there are syntax errors, the document
won't match the document production and therefore will not be well-formed.
I would expect an XML processor to report this as a fatal error.

Back to the subject of the Namespaces recommendation.  This spec changes the
document production somewhat, replacing the Name production with QName or
NCName.  If a document fails to match one of these productions, then it does
not match the [modified] document production and (I'd like to say it isn't
well-formed but I'll resist) this should be reported as a fatal error to be
consistent.

I'm not too concerned about the name given to the error.  But the severity
of the error is important and should be agreed upon.

Regards
Rob Lugt
ElCel Technology