Failed to remove directory lock /home/runner/.proto/bin/.lock

[jfinch-pp]

We're using this action in a number of our workflows, and we're frequently hitting lock-related panics when restoring from the tooling cache. Our flows are usually multi-stage and use parallel running for things like test suites, so we've broken the initial action run out into it's own dependent job, this has reduced the frequency of the issue, but when we've got a block of parallel jobs we hit the issue.

Retrying the job will often fix the issue.

 Error: 
  × Main thread panicked.
  ├─▶ at /home/runner/.cargo/registry/src/index.crates.io-6f17d22bba15001f/
  │   starbase_utils-0.8.13/src/fs_lock.rs:53:13
  ╰─▶ Failed to remove directory lock /home/runner/.proto/bin/.lock: Failed to
      remove path ~/.proto/bin/.lock.
  help: set the `RUST_BACKTRACE=1` environment variable to display a
        backtrace.

For reference, here's our workflow graph:

I'll have a look at this myself when I get a minute, but if there's an immediate fix please let me know.

[milesj]

@jfinch-pp This is weird. Can you set PROTO_LOG=trace in the pipeline and see if the logs uncover anything else.