Make it clearer that `timeout` configuration affects mocked as well as non-mocked requests

[jennifer-shehane]

In cy.request() and also main configuration doc, make it clearer that responseTimeout applies to mocked as well as non-mocked requests.

[jennifer-shehane]

I'm closing this. This issue is a bit strange in itself. cy.request() doesn't allow mocking in any sense, so it doesn't make sense to even refer to a 'mocked' request. Maybe this should be clearer somehow? But...I'm not totally sure how to address the original users confusion.

[mitar]

What is unclear about this issue? The issue is that responseTimeout controls timeouts of XHR requests, which is not really documented anywhere. So I think there is a bit of confusion that XHR requests can be both mocked or not. And even when XHR are not being mocked, responseTimeout still is in effect. In this way cy.request by-passes any mocking, but documentation only says that responseTimeout influences cy.request. So maybe this is why this confusion about mocking and non-mocking?

It is simply surprising that if I do not want to mock anything, some arbitrary additional not-documented timeout is inserted into my XHR requests. So maybe a reference to cy.request() was made by mistake. My issue was with XHR.

[jennifer-shehane]

So maybe a reference to cy.request() was made by mistake.

Yeah, I think there was some confusion when initially opening the documentation issue. I was under the belief that you were using cy.request() as part of your code - which cannot be mocked, but if you were trying to change the responseTimeout that is passed through to cy.wait() when waiting for a defined cy.route() then the documentation can clarify that applies to mocked and non-mocked in that case.

Is this accurate?

[mitar]

No. What I have is a node module which internally uses XHR. It does long-polling to stream data, by opening and keeping connection open long. My problem is that this connection has been during testing closed prematurely, why otherwise it has not been. The culprit was that Cypress modifies XHR and adds timeout to it. I would say this should not be done at all, but that is a discussion for some other issue, like cypress-io/cypress#2592. This issue is about documenting that timeout configuration also influences those XHR requests which are not part of test suite at all.

[mitar]

I also made a new comment here: cypress-io/cypress#2592 (comment)