Windows build instructions fetch gcc 5.x, resulting in segfault during build

[SingingTree]

I believe the Windows build instructions will currently lead to a segfault as outlined in these issues: #28260 #28445 #28962. Previously when running pacman -S mingw-w64-i686-toolchain or pacman -S mingw-w64-x86_64-toolchain MSYS2's pacman would fetch a 4.x version of gcc. However, the packages have been bumped, now giving gcc 5.2. This results in the issues outlined above.

This can be worked around by using MSYS2 but obtaining an alternative version of gcc. I've had success using MSYS2 but putting gcc 4.9 from mingw-w64 on the path.

I imagine it may be possible to remedy this by passing different args when using gcc 5.2, but have not experimented with this. I'd be interested in if anyone has had any success with this method.

Should the Windows build instructions be updated to reflect a work around for the above issues? I'd be happy to do so, but would like to have some discussion about the best direction to go with this.

This is a possible dupe of #28843, but I believe it is a separate issue.

[NobbZ]

I'd prefer if the build process could distinguish between older and newer versions of gcc, but I could live with installing an older version of gcc until the build process has catched up and builds with current version of gcc.

Still I do think, that this should be mentioned in the README, so that noone wastes his time by hunting bugs on his system where actually is none (as I did the last 4 hours).

edit
Sidequestion: Can I force to use clang for the build when gcc and clang are both available?

[alexcrichton]

Yeah for now I'm going to close this as a dupe of #28260, but I'd be totally fine updating the build instructions temporarily, so feel free to send a PR!