Provide language construct for "save some state, then do some stuff, then use that state as the input for a test" kind of testing

[S1artie]

When doing testing of asynchronous processes, one often has to work like in this pattern:

• First, save some state (maybe a counter value)
• Then trigger some async process which eventually changes the state in question (but which does not guarantee to do that during execution of the fixtures that triggered this process)
• Then wait for the state change to happen (for example wait for the counter to increase further than the original value stored earlier)

You can explicitly code this in Integrity already by saving call output in variables, then explicitly passing the output back into the test fixture at the end. However, it would be super-awesome if this regular pattern would somehow be supported by Integrity, for example by allowing to create a "bracket" of some sort around instructions which at the start implicitly allows a fixture to save some state which it reuses later to perform the final check/wait.