Join documents before claiming schema jobs
The migration version now joins documents before claiming, but the canonical schema.sql still selects directly from indexing_v3_agent_jobs. In databases rebuilt from the schema, rows for documents that are no longer indexed are updated to processing and have attempts incremented before the later patched_documents CTE filters them out, so they are consumed without being returned to the worker until stale retry handling.
Useful? React with 👍 / 👎.
Originally posted by @chatgpt-codex-connector[bot] in #134 (comment)
The migration version now joins
documentsbefore claiming, but the canonicalschema.sqlstill selects directly fromindexing_v3_agent_jobs. In databases rebuilt from the schema, rows for documents that are no longerindexedare updated toprocessingand have attempts incremented before the laterpatched_documentsCTE filters them out, so they are consumed without being returned to the worker until stale retry handling.Useful? React with 👍 / 👎.
Originally posted by @chatgpt-codex-connector[bot] in #134 (comment)