Hi! Consider the following use case: preventing accidental WebRTC-style information leaks. These leaks used to happen because WebRTC JS API exposes IP enumeration even if no packets get sent over the corresponding interfaces (i.e., even though the default route is the VPN endpoint, WebRTC API would "betray" information about other interfaces visible to the browser.)
In Linux, an elegant way around such leakage is to run your application in a separate network namespace a la https://www.wireguard.com/netns/ . For example, you can launch your browser/BitTorrent client/etc in a separate netns that only sees wgN so that even if there were WebRTC-style leaks, the application would not immediately see interfaces outside its network namespace. What would one do to achieve a similar result for WireGuard clients on Windows? I'd be happy to write a little bit of code / accept solutions that are not production-grade (this is all meant for a developer workstation). Madars