{"version":"2.1.238","anchor":"transcript-re-stamp-on-exit-tries-the-storage-layer-first","canonical_anchor":"transcript-re-stamp-on-exit-tries-the-storage-layer-first","heading":"Transcript re-stamp on exit tries the storage layer first","tier":"internal","area":"Sessions","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.238\/e\/transcript-re-stamp-on-exit-tries-the-storage-layer-first","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.238","markdown":"### Transcript re-stamp on exit tries the storage layer first\n\nOn exit, the final transcript update goes through the storage layer first, falling back to a direct write.\n\n**What**\n\nOn shutdown, the final transcript update is attempted asynchronously through the storage layer under a timeout, and falls back to the old synchronous write when that lapses or the layer is not in use.\n\n**Details**\n\n- The async path only runs when the newer storage backend is pinned on; otherwise the synchronous re-stamp is used directly.\n- Timeouts are logged rather than silently swallowed.\n- Exit lines belonging to another session are tracked so they are not appended twice.\n- A new transcript record type, `artifact-autoreact-ledger`, is registered as always persisted with last-write-wins.\n\n**Evidence**\n\n`Exit re-stamp through the storage backend did not finish within`\n\n- Area: Sessions\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 2\/5"}