Scripted Testing

[jameshopkins]

Whilst it's relatively straightforward capturing metrics from a stateless page, it's conceptually trickier capturing the same metrics from a stateful application (i.e an SPA). This is because a certain state representation is required in order for the appropriate UI to be rendered, so that the test can be carried out.

This can be achieved by either using:

• an application API (if it's exposed to the browser) that we can call directly through some sort of interface in WebAppTest. In which case, we'd need some sort of eval mechanism.
• or by either carrying out a realistic user journey in a programmatic way, or an interstitial UI that makes these API calls itself.

Trying to execute such a sequenced flow in a programmatic way, is incredibly tricky.

Any conventional 'web driver' that offers a waitOn API are non-deterministic by their very nature. Cypress looks like something we should look at when it's released, given it's use of promises