Radu, Cliff, All, first of all, many thanks for your replies :-)
Myself as well as my current customer, we are perfectly aware of the complexity of this issue: This is exactly why I stopped digging even deeper into after having found out out that there is no "easy" way to fix it properly (i.e. efficiently) for me as a newbie to XMLBeans internals. Instead, we decided to move forward based on my temporary "fix" of which I perfectly know that it performs poorly (especially on a large schema with lots of substitution groups as we happen to have it here).
So we did not at all expect any miracles, but were just wondering why we didn't receive any single comment or at least notice that somebody had a look into our issue...
So far, we are using my patched version (slow, but working correctly) successfully during development & integration testing. From my customer's point of view, it will still be fine if we will have a patch with good performance before the end of January (i.e. when we will start deploying this new functionality into production), which should be plenty of time to fix the issue efficiently.
I have done my best to include a detailed description with the JIRA issue of what I think must be improved when starting from my patch:
Basically, the problem is that we need to add all the legal substitutions for any of those elemnts contained in the QNameSet as returned by getJavaSetterDelimiter() that are heads of substitution groups to this set, but don't currently have any means other than an inefficient QNameSet#contains(QName candidate) check in order to determine which QNames in the set match this criterion.
I am definitely willing to help with fixing this issue, so please get back to me in case I can do anything in addition to help understand the issue and find a more efficient fix.
Many thanks again & best regards, Andreas Radu Preotiuc-Pietro wrote:
It is unusual, a lot of issues filed in the same period got at least a preliminary comment. I have looked at this a bit, but it's quite involved, writing out correct element order in the presence of substitution groups and in an efficient manner (the patch provided does point out that it's not an efficient solution, so I can't submit that directly). Will look at it again, Thanks, Radu
-- Andreas Loew Java Architect Sun Microsystems (Germany) --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
