Context
A lot of the protocol tests are implemented in javascript, which is very error-prone - especially when it comes to its lack of type safety.
Proposal
I think that moving from javascript to typescript might be worth the effort. The priority, however, should be not to break any of the existing logic/tests. One way we could gradually tackle the javascript<>typescript refactoring might be related to https://github.com/livepeer/protocol/issues/568, whereby we could start writing the shared behaviors of the tests in typescript right away and then 'plug them in' the existing test suites & replace the javascript parts
Context A lot of the protocol tests are implemented in javascript, which is very error-prone - especially when it comes to its lack of type safety.
Proposal I think that moving from javascript to typescript might be worth the effort. The priority, however, should be not to break any of the existing logic/tests. One way we could gradually tackle the javascript<>typescript refactoring might be related to https://github.com/livepeer/protocol/issues/568, whereby we could start writing the shared behaviors of the tests in typescript right away and then 'plug them in' the existing test suites & replace the javascript parts