Context

This isn't really a feature request, as what I need is possible with nock as-is. But I spent several hours searching, reading old issues and searching through the source code to find the solution, so I thought this might help others.

I am testing code that accesses a service that sets the statusMessage of the response, as well as the statusCode. I am using nock to mock the server

What

A new ctx.forward utility may be added to forward the currently captured request to its original destination (bypass).

Why

• See #655

Currently bypassing is done via implicit/explicit return in the resolver. This is not intuitive and deterministic action, which also makes it hard to distinguish when the user forgot to return a mocked response, and when they meant to b

I'd like a configuration option to be able to disable the default mock map on this line here and get an error instead:
https://github.com/apollographql/graphql-tools/blob/5c5418cec88f1d5520eccc1d2e6dfaa511547e4d/src/mock.ts#L282

We have tests that use mock data and I'd rather have an explicit error if a value could not be resolved. I find the implicit default value unhelpful in this case and m

The test suite is currently split between two organization models: some tests are organized by the component, or the feature, they test (like this), and a lot of tests are organized by the number of the GitHub issue they verify (like [this](https://github.com/mockk/mockk/blob/master/mockk/jvm/s

I would welcome a feature when, at the same url which is set to respond with OK status, I could have an error response generated with a random seed.

I am developing an Observable pattern to perform a series of retries to call an endpoint, with increasing delay. The problem is that during testing, I cannot really mock up the condition by randomly switching between two mocky.io urls (error and ok

A "normal" event on an instance ignores null subscriptions. If I do so on a substituted interface, a null subscription leads to an exception. I have a not so nice legacy code with different listeners to an event an one class that does null-attach internally under some conditions, would be nice if this behaviour would be able to be tested as well. Example code for the problem:

        publ

What you did:

Create a route:
/orders/*
then set the delay to 2000ms.

What happened:

route not delayed when hit with another param beside *
/orders/1 not delayed

/orders/900 not delayed
/orders/* delayed
![image](https://user-images.githubusercontent.com/19

Today, my project's build failed in CI because of a problem downloading the cuckoo_generator executable in the run script. I'm not sure what the actual issue was (there was an issue extracting the URL from the GH API response), but it started working again by itself sometime later, so I'm assuming that it was a GH blip.

I find it a little strange that the SPM installation for Cuckoo depends