PostgreSQL Compatibility Roadmap
SydraDB currently treats PostgreSQL compatibility as a narrow bridge, not as the identity of the product.
Short-term goal:
- let simple PostgreSQL clients connect
- translate a small SQL subset into sydraQL
- expose enough status/diagnostic information to make that bridge debuggable
See also:
./compat-matrix-v0.1./testing
What exists today
- a PG v3 listener and startup/auth flow
- simple-query execution
- a small SQL translator subset
- SQLSTATE mapping helpers
- compatibility counters and logs
- a small in-memory catalog snapshot/debug surface
What this cycle is optimizing for
- startup/auth-ok flow that is honest and testable
- simple query execution with the current translator subset
- standard
CommandCompletebehavior for the implemented path - docs and tests that describe the current bridge accurately
Explicitly unsupported this cycle
- TLS
- SCRAM-SHA-256
- extended protocol
- prepared statements and portals
- COPY
- full
pg_catalog/information_schemacoverage - migration tooling and CDC-based cutover flows
- broad ORM unlock work
Near-term backlog
1. Make the simple-query path honest
- keep the current protocol surface clearly documented
- tighten smoke coverage for startup and simple query handling
- make unsupported paths fail with explicit SQLSTATE/documented messages
2. Keep translator/docs/tests in sync
- describe the supported SQL subset in one place
- keep translator fixtures aligned with current behavior
- document the escape hatch for direct sydraQL where relevant
3. Treat the catalog shim as scaffold, not parity
- maintain the current debug catalog snapshot for visibility
- avoid over-claiming parity until real coverage lands
Deferred until after the current alpha reset
- deeper catalog emulation
- prepared statement/session state compatibility
- ORM trace harnesses and broader regression suites
- schema/data migration tooling
- low-downtime cutover support