Release native output after JavaScript cancellation
The caller abandons a promise while native work finishes later; its output buffer has no remaining JavaScript consumer and is leaked.
- Focused work estimate
- 2h 30m + prerequisites
- Priority in the scenario
- Medium
- Engineering practice
- Cancellation · FFI cleanup · Concurrency
Estimated field mix
- Systems programming70%
- Real-time systems30%
Field percentages are editorial estimates of the ticket's engineering focus. They total 100%; they are not measured time, proficiency scores, or ownership evidence.
Review it, then add it to your workspace.
The board opens an editable draft; nothing is saved until you confirm it. Sign-in and workspace permissions apply, and Demo boards remain ephemeral.
Project context
A fictional desktop backup tool calls a small Rust compression library from Node.js. The binding leaks buffers on exceptions and treats ABI mismatches as corrupted input. Build against generated byte arrays and a local native library; no production archive format or user files are supplied.
Setup prerequisites
- FFI ownership
- Binary compatibility
- Error handling
Preceding work
Complete these dependencies, or supply their agreed outputs before taking this ticket.
Acceptance criteria
- Operation state records whether a result still has a consumer
- Late output is released on the native allocation thread or declared safe path
- Cancellation and completion races release once
Implementation constraints
- Cancellation is cooperative; do not unload code while a native call is running.
Verification to include
- Cancel before start and after successful delivery.
- Race cancellation with native completion across repeated deterministic schedules.
Deliverables
- Cancellation cleanup protocol and race tests
Rollout and recovery
Keep native concurrency at one until release counts reconcile.
Value of the work
For the engineer: Practice language-boundary contracts, native resource safety and version negotiation.
For the team: Review a binding that fails safely and can be rolled back before native code enters additional application paths.
Evidence boundaries
Outcome Evidence: Tests, patches, and runbooks are requested deliverables. They become Outcome Evidence only through a qualified Mission and immutable Evidence IDs.
Ownership Evidence: Independent adaptation must be observed under a declared verification policy and cite immutable Evidence IDs. Completing a planning ticket establishes no Ownership Evidence.