[ravi-pplx]

Summary

Expose a nameable FilesystemHandle with native write_file. Share admission, frame limits and supervised execution with gRPC; bound native input waits with a configurable 30-second idle timeout. Borrowed/non-Unpin streams remain supported. Share one idle default and document zero-duration and config-literal behavior. Daemon gRPC upload behavior is unchanged. Add a state-free Debug implementation for FilesystemHandle and clarify the experimental Rust compatibility policy.

Implements native uploads from the embedding RFC.

Validation

418 affected tests passed on the rebased stack, including five privileged probes. Formatting, affected all-feature/no-default-feature Clippy and doctests passed on Linux. Native macOS and Kind/Firecracker were not rerun locally; CI is pending on this head.

Severity and risk

Severity does not apply: new native API. Slow native uploads can reach the idle timeout; callers can configure it. Revert with dependent changes to roll back.

AI assistance: Codex contributed code, tests and this description.

[zibi-pplx]

The shared upload path looks right. Admission happens before the first poll, the per-message limits and bounded relay are preserved, and accepting borrowed and !Unpin streams is nicely ergonomic. The gRPC Write refactor keeps its semantics. I'm requesting changes for two API issues on the native surface (inline):

1. Split upload API / unnameable handle. filesystem_service() returns an opaque type whose write can't be called natively, so uploads need a separate ExecutionService clone.
2. No idle timeout on native uploads. A stalled stream pins a payload slot and blocks same-key retries.

Non-blocking: environment validation is inconsistent between configure_process_env and native StartRequest.env (inline, README). The same Kind/Firecracker CI note as #702 applies.

Review prepared with AI assistance (Claude Code); findings checked against the PR head.

[cla-assistant]

All committers have signed the CLA.

[zibi-pplx]

Follow-up review at f41757cc8ca8, including the full stack through #833.

The earlier nameable-handle and upload-idle-timeout findings are addressed, and the shared native/gRPC environment sanitization is now documented. I found no additional upload correctness issue.

Two non-blocking API points remain:

• Add Debug for the newly public FilesystemHandle. It currently derives only Clone (source). Our API guidelines require Debug for public types. A minimal implementation describing the handle is enough; it need not expose requests, environments, or internal state.
• Clarify the embedding API's compatibility promise. The new public ExecutionServiceConfig field breaks complete external struct literals, which the README correctly documents. Our current guidance against #[non_exhaustive] assumes workspace crates evolve together. If this embedding surface is intended for independently maintained external consumers, that assumption should be revisited explicitly. In that case, consider constructor/builder-based configuration and non-exhaustive public error enums. ExecutionServiceConfig is a struct; adding #[non_exhaustive] to it would prevent external struct literals, including ..Default::default(), so it needs a supported construction path. This is a compatibility-design question, not a request to add the attribute mechanically.

Codex-assisted follow-up review. Findings were checked against the referenced head; shutdown/drain findings are posted on their owning PRs.

[ravi-pplx]

Addressed the follow-up review in 30011f39: added minimal Debug for FilesystemHandle and documented experimental Rust compatibility with revision pinning. Stable configuration construction and error evolution remain a follow-up before promising cross-version Rust compatibility.