Fabian Meumertzheim

Yes, I think that without a dedicated platform constraint value for the CPU we can't solve this properly on the rules_go side.

Adding a dep on `bazel_features` isn't easy with rules_python's WORKSPACE setup since it doesn't use two levels of macros. Can we add a call to `bazel_features_deps` to the relnotes?

PRs are welcome. With Bzlmod, Gazelle should no longer need to call the `go` tool, so the issue shouldn't apply to that.

I checked `git blame` and this table's content hasn't changed in two years. Either it's been broken forever or something about GitHub's `.rst` rendering changed.

@illicitonion Didn't you file an issue for this a while ago? Can't find it right now, but it also had relevant discussion.

Nvm, found it: https://github.com/bazelbuild/bazel/issues/23139

Reviewing this change is probably quite easy compared to validating its design choice (@tjgq suggested an alternative above, https://github.com/bazelbuild/bazel/issues/23139 has another one). I would personally benefit from a short write-up...

This is what I (personally) have been looking for. It's very rich in information, thanks! Regarding `$(dirname ...)`: Since paths obtained via location expansion are always slash-separated, I wonder whether...

@jmillikin Not sure whether you intentionally deleted your fork and don't want to continue to work on this, but I'm now firmly in favor of approach C (which aligns with...

@matts1 @armandomontanez Friendly ping @jmillikin Yes, I actually think that handling slash-separated paths only also on Windows is the only behavior that actually makes sense: all Make variables and location...