This web platform test is checking that an empty Content-Type header is being sent by the UA, which both Firefox and Edge are actually doing. However, the test fails because it's returning a responseText with an extra space.

That is, when I have the test's python script XMLHttpRequest/resources/inspect-headers.py print out the raw_headers.headers it actually receives, the results are what the test expects: 'Content-Type: \r\n'

However, when I ask that same script to spit out the byte sequence (in decimal) it's going to pass back as the responseText it will be checking on the client-side, it gives: 67, 111, 110, 116, 101, 110, 116, 45, 84, 121, 112, 101, 58, 32, 32, 10

Note that extra space at the end. The test is accidentally fooling itself into thinking that Firefox (and probably Edge as well) are failing.