This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
In the entire BPEL, a correlation set can be initiated only once. (A correlation set name is unique within it's scope). If a correlation set is attempted to initiate a second time, editor needs to throw an error saying that this leads to runtime issues and this is mostly a user error. See the attached project.
Created attachment 44966 [details] validation project
The tricky part is that if there is a "if" block or some such equivalent constructs, we can't easily determine if the correlation will result in a duplicate initiation or not. We need to take care of such a situation.
in the attached project's BPEL, <correlation set="CorrelationSet1" initiate="yes"/> is defined twice and it is wrong.
Implementation of this feature will take much time. Will not be able to fix it in current release.
I won't disagree on you pushing back this validation, but we should at least do a minimum validation that will help the user. As a simplified version of this validation, we will see if there are any duplicate initialization of the correlation. Forget about "if" and other paths which complicate this validation. If there is a duplicate validation, we point to the user saying that the during runtime it could potentially have a problem. This will be a warning and not an error. If this corrects 70% of the users we are definitely better off. I am CC-ing Sergey and Gabe so they understand this issue to allocate time for this. I am guessing that this sort of validation shouldn't take too much time.
fixed in Sierra.