I\'m trying to get my code coverage in java, using Eclipse and EclEmma.
My tests are using JUnit 4 and I\'ve got some tests looking like this :
@
I faced the same issue, and offered a pull request dealing with the most prominent cause of this annoyance. Instead of adding way too many probes, I only deal with the opcodes of method invocations. In other words, the problem of (previously) wrong coverage results caused by an exception thrown by a method invocation is solved, but not if the exception is thrown by, for example, a division by zero, a wrong cast, or an array index problem.
https://github.com/jacoco/jacoco/pull/261
Feel free to comment on the pull request or participate in the discussion I started on the JaCoCo mailing list.