fake_value cookie removed, which breaks a test in requests

[kevin1024]

This commit: 791e31c removed the cookie that was always set on the digest endpoint, which breaks a test in requests (see also: this travis failure)

There was some old (never-resolved) discussion of this odd feature here: kennethreitz/httpbin#124

Paging @sigmavirus24 - should we add this behavior back?

See also: psf/requests#4263

[sigmavirus24]

Perhaps it'd be better to do that in a different way. Not ever HTTP client wants to handle Cookies as part of HTTP Digest even though that's patently wrong.

Requests has a test dependent upon that because we broke that at one point and were relying on being able to use HTTPbin to provide us the behaviour so we could verify it.

[kennethreitz]

Pull requests break everything around me.