Document CLI I/O timeout behavior - #62
Conversation
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
✨ Simplify code
Comment |
|
Closing as stale: this documents the I/O timeout behaviour of the commander shell, which was deleted in #139 — the behaviours it describes (controller-level fetch timeouts, the old streaming paths) no longer exist as written. The engine has its own timeout story (for example the bounded identity enrichment in |
Draft PR for visibility only. This contains the CLI I/O timeout spec and implementation plan; it is not intended to be merged.
@luanvdw tagging you for awareness.
Changes
.agents/projects/cli-io-timeouts.spec.md, defining the intended CLI behavior for stalled Prisma-controlled I/O, includingOPERATION_TIMEOUT, cancellation separation, generous defaults, and out-of-scope boundaries..agents/projects/cli-io-timeouts.plan.md, sequencing the work through a shared timeout contract, Prisma API/SDK request boundaries, long-lived inactivity boundaries, and product-doc alignment.Why
The spec chooses step-scoped and inactivity-scoped timeouts instead of a root command-wide timeout. That keeps long-running workflows such as
app logs,app domain wait,auth login, deploys, and local app runs usable while still preventing Prisma-controlled API, SDK, polling, and stream boundaries from hanging forever.