Tiago Quelhas

@alexjski Could you take a first pass since you're the expert on tree artifacts? Feel free to reassign back to me to shepherd it in if/once you're happy with the...

> Let me first clarify the intention behind this PR: I view #15789 and #15901 as serious correctness issues in a supported feature that is enabled by default (directory outputs)....

@fmeum Yes, I think you are correct and we either have to extend the definition of `TreeFileArtifact` to also cover empty directories, or introduce a new artifact type (`TreeDirectoryArtifact`?) for...

Hi Paul, Unfortunately, I don't have a recommendation. I have a downstream project that would also benefit from latency, though - I can't make any promises, but it's likely that...

It won't work just like this, because the 'end' event on the source end of the pipe must be relayed to the destination end. I've coded up a more complete...

The remote execution spec definitely supports actions that create a symlink, but it doesn't say whether the symlink must point to an existing file or directory or not (at least...

I've confirmed with folks who work on Google RBE that output symlinks are not currently implemented (specifically, symlinks that resolve successfully are reported as the regular files or directories that...

This was fixed by 5b46a48.

> @tjgq PTAL > > Ideally, I'd like to flip this before Bazel 6.0.0 to limit who depends on the old behavior. Not sure how feasible that is though. AIUI...

While we're at it, we should also remove the limitation in https://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/remote/RemoteActionFileSystem.java;l=222;drc=79b5b5ddaa3919db48b9143e0afdd4d2cddfdb74 (which AIUI would only apply when a remotely produced declare_symlink is an input to a local action).