Publish durable Session artifacts from hosted native outputs - #693
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hosted Turns currently leave their outputs only in the live workspace, so clients cannot retrieve a durable result once the Environment expires. This adds the pinned Session Artifacts list, retrieve, content and delete operations and publishes immutable
/workspace/outputscopies with successful Turn completion.Capture uses the same authorized Runtime/workspace and bounded streaming into private execution-database storage. Publication requires confirmed native completion; failed/cancelled captures remain invisible and are reclaimed. Inputs received while outputs are being captured wait for the next native Turn. Artifact deletion preserves workspace/source Files, while Session deletion reclaims its copies.
Validation: fixed OpenAI Python 3.13.0 and raw HTTP against independent Core, Docker, stock Codex 0.153.4 and real Kimi; binary/empty/nested outputs, immutable versions, pagination and tenant isolation; 500 MiB batches with 200 MiB members and exact streamed hashes; concurrent input ordering, native credential-link rejection, restart/cancellation/expiry/deletion. SQLC/OpenAPI regeneration and focused Go race/PostgreSQL/Rust checks passed. Final server
make checkpassed (459.01s); all 61 source hashes match the candidate. Actual PostgreSQL/Core restart retained exact metadata and SDK/HTTP bytes (41.14s). The final real 500 MiB run also verified simultaneous directory reads and queued next-Turn input (145.66s). Fresh independent whole-diff review of all 61 files found no actionable issues; focused Runtime/gateway/API tests also passed during review. Failed fixture attempts remain preserved; the PostgreSQL restart fixture rediscovers its ephemeral Docker port.This is the qualified Codex/Docker hosted workflow, not full protocol compatibility. Exact upstream default/error/cancellation edge behavior and unchanged-output republishing parity remain recorded gaps. User-managed enrollment, Claude/E2B, V2, and an external storage framework are outside this batch. Board: SESSION-ARTIFACTS-001 under ATOM-005; parent tasks remain open.