>>That's a very old version of Xalan-C
We are working for the U.S. Government. It was the newest version when we
started air force certification. We haven't finished certification yet and would have to
restart if we changed the parser out at this time.
The people writing the system that is having this problem have given me only a little more
information about the problem when I asked them the questions you asked.
>What error?
When they say numeric_error they are refereeing to a ADA numeric_error. ADA actually returns "numeric_error" and doesn't give more specific information. This could be an system exception or some other kind of error code coming back from C++.
The system they are working on is written in Ada and uses a C++ module that my company wrote. They know they are crashing on the first recursive call to transform. ( they used printf statements to track that down ). I've asked them to put
try-catch statements into transform and attempt to actually get the error code back.
>Well, that's bound to fail because you are dereferencing a null pointer
What the last test the message described was supposed to demonstrate is that the original call , the first recursive call to transform, is crashing during the cast of the variable. By changing the variable name, they were able to get the second call to transform to proceed past the initial call to it. They expected it to crash later.
Any further help would be appreciated. you don't happen to know if IBM , or anybody else, sells support contracts for xerces/xalan/xml?
Regards,
Christopher
