tnc is an expat parser object extension, that validates the XML stream against the document DTD while parsing.
package require tdom package require tnc set parser [expat] tnc $parser enable
tnc adds the C handler set "tnc" to a tcl expat parser obj. This handler set is a simple DTD validator. If the validator detects a validation error, it sets the interp result, signals error and stops parsing. There isn't any validation error recovering. As a consequence, only valid documents are completely parsed.
This handler set has only three methods:
- tnc parserObj enable
Adds the tnc C handler set to a Tcl expat parser object.
- tnc parserObj remove
Removes the tnc validatore from the parser parserObj and frees all information, stored by it.
- tnc parserObj getValidateCmd ?validateCmdName?
Returns a new created validation command, if one is avaliable from the parser command, otherwise it signals error. The name of the validation command is the validateCmdName, if this optional argument was given, or a random choosen name. A validation command is avaliable in a parser command, if the parser with tnc enabled was previously used, to parse an XML document with a valid doctype declaration, a valid external subset, if one was given by the doctype declaration, and a valid internal subset. The further document doesn't need to be valid, to make the validation command avaliable. The validation command can only get received one time from the parser command. The created validation command has this syntax:
validationCmd method ?args?
The valid methods are:
- validateDocument domDocument ?varName?
- Checks, if the given domDocument is valid against the DTD information represented by the validation command. Returns 1, if the document ist valid, 0 otherwise. If the varName argument is given, then the variable it names is set to the detected reason for the validation error or to the empty string in case of a valid document.
- validateTree elementNode ?varName?
- Checks, if the given subtree with domNode as root element is a posible valid subtree of a document conforming to the DTD information represented by teh validation command. IDREF could not checked, while validating only a subtree, but it is checked, that every known ID attribute in the subtree is unique. Returns 1, if the subtree is OK, 0 otherwise. If the varName argument is given, then the variable it names is set to the detected reason for the validation error or to the empty string in case of a valid subtree.
- validateAttributes elementNode ?varName?
- Checks, if there is an element declaration for the name of the elementNode in the DTD represented by the validation command and, if yes, if the attributes of the elementNode are conform to the ATTLIST declarations for that element in the DTD. Returns 1, if the attributes and there value types are OK, 0 otherwise. If the varName argument is given, then the variable it names is set to the detected reason for the validation error or to the empty string in case the element has all its required attributes, only declared attributes and the values of the attributes matches there type.
- Deletes the validation command and frees the memory used by it. Returns the empty string.
The validation error reports could be much more informative and user-friendly.
The validator doesn't detect ambiguous content models (see XML recomendation Section 3.2.1 and Appendix E). Most Java validators also doesn't, but handle such content models right anyhow. Tnc does not; if your DTD has such ambiguous content models, tnc can not used to validate documents against such (not completely XML spec compliant) DTDs.
It isn't possible to validate XML documents with standalone="yes" in the XML Declaration
Violations of the validity constraints Proper Group/PE Nesting and Proper Conditional Section/PE Nesting are not detected. They could only happen inside a invalid DTD, not in the content of a document.