[
Lists Home |
Date Index |
Thread Index
]
> > Excuse my ignorance here, but is there anything like a
> "SAX Test
> >Suite" that can be used to demonstrate conformance in an objective
> >manner?
>
For what it's worth, I know of one behaviors which is not required for
conformance, but which all known SAX parsers exhibit and which I (Saxon)
depend on, specifically that the parser delivers a value for the qName
argument to startElement() even when the namespace-prefixed property is
false. (I considered putting in code to cope with the absence of this
information, but (a) it would be a lot of effort to test the code, and
(b) it would require saving information from earlier calls "just in
case" it's ever needed. Too high a price to cope with a theoretical
possibility.)
The areas where I have difficulty with non-interoperable behavior across
parsers are primarily to do with exception handling, in areas where the
spec is unspecific. For example, if an exception is thrown in a callback
such as startElement(), some parsers report it to the ErrorHandler and
others don't. It requires some careful juggling to make sure that
whichever parser is being used, such errors result in exactly one error
message.
Michael Kay
|
