There are three possible scenarios when entering a CXXTryStmt:
- No exceptions thrown.
- An exception is thrown but caught by one of the handlers.
- An exception is thrown but not caught by one of the handlers.
The region we push after a CXXTryStmt isn't useful in scenarios (1) or (2). In
these cases, if control flows to the try block it must flow to the region
following the CXXTryStmt.
This may not be the case in situation (3). However, we don't handle this
scenario anywhere else (e.g after function calls). Having a dedicated region
here could only be helpful if its counter expression differed from the try
block, and even then it's questionable.