Currently we stop default server in worker run loop only before the next test start. So the previous test does not account whether the Tarantool executing the test exits successfully or not. Thus we do not fail on memory leaks for example.
Let's stop the server and account its exit status in test.
coverage: 62.721% (+0.1%) from 62.623%
when pulling 7aa25f02edcda25e2b26d7bce0afba3054f5f70b on nshy:fail-on-tarantool-server-failure
into 240cdeadf736a96a41c3d98a5a10dad2015f5135 on tarantool:master.
Currently we stop default server in worker run loop only before the next test start. So the previous test does not account whether the Tarantool executing the test exits successfully or not. Thus we do not fail on memory leaks for example.
Let's stop the server and account its exit status in test.
Closes #416