Per Node docs,

If no 'response' handler is added, then the response will be entirely discarded. However, if a 'response' event handler is added, then the data from the response object must be consumed, either by calling response.read() whenever there is a 'readable' event, or by adding a 'data' handler, or by calling the .resume() method. Until the data is consumed, the 'end' event will not fire. Also, until the data is read it will consume memory that can eventually lead to a 'process out of memory' error.

However, when the Recorder was enabled, Nock would sometimes resume the response stream after monky-patching the push methods. This doesn't align with the goal of being consistent with native Node. And has presented itself as an issue with Got because sometimes that client doesn't register the 'end' event listener fast enough.

Moreover, Nock was not consistent with the approach. It would only resume the response if .request was not provided a callback. Which meant the two following examples created different behavior.

http.request('http://example.com', res => {/*...*/})

// vs.

const req = http.request('http://example.com')
req.on('response', res => {/*...*/})

fixes: #2086