There's a lot of duplicated code within each end-to-end tests, and semi-duplicated functionality across the two current existing ones (one which runs signing manually, and one which runs interactive signing).
Some of the changes between manual and interactive tests involve simplification of the protocol-running infrastructure (see, e.g. the process_message function and its parameters in manual, as compared to the implementation in the interactive signing test). The primary change was simplifying the inbox that holds in-process messages
Overall, each protocol execution is virtually identical up to input and output types. Consider writing a more generic function to run a single protocol, and refactoring the tests with calls to this function.
There's a lot of duplicated code within each end-to-end tests, and semi-duplicated functionality across the two current existing ones (one which runs signing manually, and one which runs interactive signing).
Some of the changes between manual and interactive tests involve simplification of the protocol-running infrastructure (see, e.g. the
process_messagefunction and its parameters in manual, as compared to the implementation in the interactive signing test). The primary change was simplifying the inbox that holds in-process messagesOverall, each protocol execution is virtually identical up to input and output types. Consider writing a more generic function to run a single protocol, and refactoring the tests with calls to this function.
Related issue: #428