How pgedge/lolor works
Project type and purpose
• This is a PostgreSQL extension (“lolor”) written in C using PGXS.
• Purpose: override Postgres’s native large-object API (lo_*) so that large objects are stored in user-managed tables (lolor.pg_largeobject and lolor.pg_largeobject_metadata) and can be logically replicated via pgedge/spock.
File structure
• Top-level build artifacts and config:
• Makefile (PGXS build)
• .github/workflows/*.yml (CI pipelines)
• docker/ (Dockerfiles, compose, entrypoint, scripts to spin up Postgres clusters for testing)
• src/ (C source: lolor.c, API stubs, large-object management)
• tests/ multi-language test suites:
• tests/go (Go client tests)
• tests/javascript (Node.js tests)
• tests/junit (Java JUnit tests for single- and multi-node replication)
• tests/python (Python client tests)
• SQL/control files: lolor–1.0.sql, lolor.control (Postgres extension metadata)
README for additional insights
• Describes extension installation (pgEdge CLI or PGXS), configuration (lolor.node, search_path), function overrides, and replication-set integration via pgedge spock.
• Shows example calls to lo_creat, lo_import, lo_unlink, and queries against lolor schema tables.
Elements
• from Build → Postgres (install)
• from Test Clients → Postgres (API calls)
• from Postgres primary → Replication Engine → Postgres replica
• from CI → Docker Compose (orchestrator)
• Blue for build/CI artifacts
• Green for runtime DB components
• Orange for client/test harnesses
• Purple for replication subsystem