Description

As is known, it is better to avoid Fuzzer executions when we generate parametrized tests.
Sometimes they are really required to increase coverage - the we put a separate region with them under a parameterized test
Sometimes we have a pair of absolutely similar executions, but one of them is symbolic and the second is fuzzer one. Thus we should better prefer symbolic.

Expected behavior

After the discussion with Sergey Pospelov and Dmitry Ivanov we decided to do the following:

• first of all we minimize symbolic executions
• if some instructions are not still covered, we use fuzzer executions

Potential alternatives

The alternative is to miss all fuzzer executions in parameterized tests. It seems to be worth.