Add a debugging request handler

Is your feature request related to a problem? Please describe.

One of the most common issues developers experience is unmatched requests.

Describe the solution you’d like I suggest to add a debugger request handler (i.e. log) that would act as a transparent handler, meaning it would log out any requests that didn’t match preceding handlers.

Describe alternatives you’ve considered We usually suggest to do that manually, but neither of the existing handlers can be fully used as a transparent request handler, as they have a predicate built-in which limits requests captured to a certain degree.

[kettanaito]

I believe there are multiple strategies the users may be choosing when installing MSW:

• Mock-first, no backend at all, develop entirely against MSW. Here I don’t want any unhandled requests to happen.
• Incremental, there’s an existing app, there’s a backend, let’s use MSW to either extend that backend or mock a third-party service. Here I’d expect my unhandled requests to happen, so there are no distractions to my flow.
• Testing. I’ve got an app, I need to write some tests in Node.js. Here I don’t want any unhandled requests to happen.

From the use-cases, I see the default onUnhandledRequest option would be beneficial to most of them. That being said, it’d be nice to see how many people follow each of those strategies to evaluate our choice.

The fact that misconfigured request URLs cause 80% of mismatch issuses is a huge reason to set onUnhandledRequest: "warn". That’d be both non-invasive, but informative.

[kettanaito]

I think in most of the cases, when MSW is introduced to a project, it’s the intention to mock all the API requests

What about the mock-driven development? Although unexplored, I find it a promising direction and can see people developing against MSW until their backends are ready. Wouldn’t seeing a bunch of warnings be annoying?

I do share the idea to set “warn” as the default unhandled request policy, it’d ease the debugging curve for people experiencing uncaught requests. I just wish for us to think through all the implications of that change.