Even if not done, if would be great to provide a "path to follow" for the community in the sense of a task break down and/or pending dependencies form clang/llvm that could make this not possible at the moment.
If there is a Windows port being worked on, why spend effort on anything except for *-windows-msvc?
I would have thought the goal of a language platform is to be as seamless as possible and requiring the least amount of "moving parts" as possible. Considering a port to *-windows-mingw, *-windows-cygnus, etc. just means that the port becomes dependent on more third-party frameworks.
Not sure how valuable that would be. It complicates things to have different upstream for the ports. Furthermore, most of the MSVC support is merged into mainline (theres some work needed to fix the last few things). I think we can do the same with cygwin, and just work from the same tree.
The MSVC port is pretty much done. There is ongoing work to get the tests passing, but a complete build works. Furthermore, the MSVC build has been verified to be able to build and run programs written in swift on Windows. The cygwin port seems to be pretty similarly far along. I'm marking this as done for now.