[jdom-interest] Parsing a MODS-document with validation fails

[Thomas Scheffler]

Am 14.08.2011 07:15, schrieb Jason Hunter:
> I agree with Rolf's assessment.

I too. It is important what comes out at the end. If the input is valid, 
than the output should be valid, too. Blaming the user for not providing 
a custom filter working around a erroneous specification, would not help 
the user and not help the JDOM project, as traffic on the mailing list 
will be higher.

Hope to see 1.1.2. soon.

regards,

Thomas

> On Aug 13, 2011, at 2:14 PM, Rolf Lear wrote:
>
>> Hi Brad.
>>
>> In this case I think the issue is too 'grey' to squarely lay the blame
>> on any particular component. The use-case is such that it relies on a
>> number of different shortcomings in a number of tools, including JDOM.
>>
>> Although the 'test cases' I put together illustrates the underlying
>> problem (that there's no namespace prefix given to attributes that are
>> 'sourced' from the XSD), the initial symptom as reported by Thomas
>> is/was that JDOM was overwriting the value of his 'type' attribute
>> with the value of the 'type' attrribute in the xlink namespace.
>>
>> In other words, if there is no change made to SAXHandler, then we make
>> a 'broken' situation even worse. Saying "when XMLSchema specification
>> is revised, and all the parsers do the 'right thing' we will give the
>> right result" feels a little cheap.
>>
>> At a minimum, the SAXHandler has to be able to identify attributes
>> that are in a namespace, but without a prefix, and then to ignore
>> *that* attribute instead of potentially corrupting a different attribute.
>>
>> On the other hand, the fix/workaround for the problem is relatively
>> trivial, easy to confine to just the scope of the issue, and it is
>> safe for when we do (eventually) have parsers that provide a correct
>> (or generated) prefix.
>>
>> The concept of the user having to apply a filter or wrapper to get
>> sensible results is not pretty.
>>
>> One other thought, in order to get to this state, you have to declare
>> your Parser to be namespace-aware... and then, after asking for the
>> namspaces JDOM then does its processing based on the prefix, not the
>> namespace...
>>
>> Rolf
>>
>> On Fri, 12 Aug 2011 17:51:21 Bradley S. Huffman wrote
>>
>>> However JDOM, or any other project can't and shouldn't be adding
>>> kludges because another group won't fix there code and keeps 'passing
>>> the buck'.  Therefore, I support Michael's suggestion of handling it
>>> with a filter or wrapper, but I do not support changing SAXHandler in
>>> any way.
>>>
>>> On Wed, Aug 10, 2011 at 3:14 PM, Rolf Lear<jdom at tuis.net>  wrote:
>>>
>>> >  That issue is marked "Won't fix", with the comment: I agree it's a bug, but
>>> >  it's not that easy to fix. The problem is what prefix  to use for the new
>>> >  attribute. I'll think about it more.