"George T. Joseph" wrote: > Jeez, you're making me sorry I even spoke up. I set the test > code to end at 128 because that is what YOU guys do in > DeferredDocumentImpl.
I'm sorry if I gave you that impression. I'm extremely busy at the moment and just quickly checked to make sure that the binarySearch() method was working correctly. > The schema updated to 12/17 is appended below. Move your > debug code in binarySearch to just before "int value = > values[middle];" (where the exception is thrown) and run it > through DOMCount. Thanks! This reproduced the bug for me using the latest source code. The problem was not in binarySearch but in the *call* to binarySearch. Simple off-by-one caused the problem. I wrote a simple test case to run the sample file through the parser and verified that the number of identifiers that the document *thinks* it has is the same as the number of identifiers in the document: 128. Even though I may have seemed a little gruff through this whole process, I really do appreciate your help and persistence. :) Please update your copy of DeferredDocumentImpl.java and that should solve the problem. Let me know if it doesn't. -- Andy Clark * IBM, JTC - Silicon Valley * [EMAIL PROTECTED]
