Hi Alberto,
Thanks for the prompt and detailed response.
Whenever a system is implemented in terms of itself, you always reach a
point where there are low-level primitives that cannot be expressed
"normally". I am not surprised to see this applies to the "schema for
schemas", but thanks for clarifying the issue for me.
What I want to do is write some code that will:
/1/ Parse a ".xml" or ".xsd" file and do as much checking
as Xerces will allow for that type of file, and then...
/2/ Build a DOM tree for the parsed file.
I had a look at the API documentation for loadGrammar(). If I understand
your email and the documentation correctly then my algorithm should be
something like the following pseudo-code:
parser = new XercesDOMParser();
parser->setDoNamespaces(true);
parser->setErrorHandler(...);
if (inputFilename ends in ".xsd") {
parser->setDoSchema(false);
parser->setValidationScheme(XercesDOMParser::Val_Never);
parser->loadGrammar(inputFilename, Grammar::SchemaGrammarType);
} else {
parser->setDoSchema(true);
parser->setValidationSchemaFullChecking(true);
parser->setValidationScheme(XercesDOMParser::Val_Always);
}
parser->parse(inputFilename);
In effect, I should use full schema checks for ".xml" files, but for
".xsd" files I should turn off schema checking and compensate for this
by calling loadGrammar(). Is this correct?
Regards,
Ciaran.
--
Ciaran McHale, Ph.D. Email: [EMAIL PROTECTED]
Principal Consultant Tel: +44-7866-416-134 (mobile)
IONA Technologies, UK WWW: www.iona.com
