In bug 1679929, we're capping the version number reported for macOS. I suggest we also cap the Windows version to 10.0. There isn't a change in sight to the version number reported by Windows, which means that if it ever changes, the potential for disruption to user agent sniffing is going to be large.

If we cap it to 10.0 now, we can then evaluate whether we want to change the Web-visible value if the OS-reported value ever changes instead of experiencing what happened with macOS: The OS API changing the value and us having to react to the breakage.

https://searchfox.org/mozilla-central/rev/0379f315c75a2875d716b4f5e1a18bf27188f1e6/netwerk/protocol/http/nsHttpHandler.cpp#951

No tests, because proper testing would need CI with Windows reporting a a version above 10.0.

https://treeherder.mozilla.org/#/jobs?repo=try&revision=7c783a2c27a42278cf044c2b1e001df56f970a3b

Even if Windows changes the version number, the reported version number will not change until we update supportedOSes in the manifest, I guess. Nevertheless I totally agree with you. We don't have to rely on operating system's compatibility shim.

I ran the test build on Windows 10 and Windows 7, and it didn't interfere with the pre-existing functionality. (I.e. reported 10.0 and 6.1, respectively, both via HTTP and navigator).

https://hg.mozilla.org/mozilla-central/rev/3b60746765d3