Added
* Add the `--hypothesis-strategy` flag to the `run` command. It defines how the
steady-state hypothesis is applied. One of:
* `default` is the classic mode where the hypothesis is run before and after
the method
* `before-method-only` runs the hypothesis once only before the method
* `after-method-only` runs the hypothesis once only after the method. This is
useful when you know your environment is not in the appropriate state
before the conditions are applied
* `during-method-only` runs the hypothesis repeatedly during the method but
not before nor after
* `continously` runs the hypothesis repeatedly during the method as well as
before and after as usual
* Add the `--hypothesis-frequency` flag to the `run` command. This flag is
only meaningful with `--hypothesis-strategy=during-method-only|continously`.
It takes a floating number indicating how many seconds to wait between two
executions of the hypothesis
* Add the `--fail-fast` flag to the `run` command. This flag is
only meaningful with `--hypothesis-strategy=during-method-only|continously`.
If set, this indicates the experiment should be marked as deviating
immediately. When not provided, the hypothesis runs until the end of the
method without terminating the experiment
Changed
- Bump dependency on chaostoolkit-lib to 1.13.0 to support the steady state
strategy