Commit Graph
4 Commits
Author SHA1 Message Date
Claude ec247a96e1 test(data): PostgreSQL MD5, cleartext, trust and TLS tests can run, and must where TLS is set up
BRIEF §6.3 requires every authentication method and TLS mode against the
PostgreSQL service containers, but the MD5, cleartext and trust tests always
skipped in CI (no roles) and the TLS tests skipped at run time because the
official images start with TLS off. .github/scripts/postgres-ci-setup.sh prepares
a server container: a role per method with its pg_hba line ahead of the image's
SCRAM line (the MD5 password stored as MD5), and TLS with a self-signed
certificate for localhost and 127.0.0.1, enabled with ALTER SYSTEM and a reload
whose rejected lines fail the script. With TITANO_TEST_PG_TLS=1 a server that
refuses TLS fails the TLS tests instead of skipping them, and a new test checks
that verify-full accepts the server whose certificate chains to
TITANO_TEST_PG_ROOT_CERT.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
2026-10-12 10:06:10 +00:00
Claude fb42423346 fix(data): close concurrency and cleanup gaps found in review
- SQLite restore holds every pool slot instead of the write lock first,
  removing a lock-order deadlock with sessions that begin to write.
- PostgreSQL sessions always return their connection on dispose, a
  rollback while disposing never masks the original exception, and a
  busy session keeps its transaction until it is discarded.
- The pool disposes a connection whose health check the caller cancels.
- Listener commands keep their lock until the server answers and never
  tear a write with the caller's token; closing and pending no longer race.
- Tag replacement locks the sequence and the tag row; unused-tag cleanup
  locks candidates and re-checks them with a fresh snapshot; render
  counts and stats refreshes lock the sequence row on PostgreSQL.
- Sequence upserts keep rendered_count, which renders maintain.
- AuthenticationOk after an unfinished SCRAM exchange is rejected.
- Oversized read buffers are released; failed backups leave no .tmp;
  a zero idle lifetime is rejected.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
2026-10-02 10:27:19 +00:00
Claude 8d9b04c713 refactor(data): pin server output settings and guard busy connections
Pin bytea_output=hex and standard_conforming_strings=on in the startup
packet instead of tracking ParameterStatus values nobody read, so the
legacy bytea escape decoder goes away. A PostgreSQL connection now
refuses a new command while a row stream is still open, and the pool
discards such a connection. Drop unused SQLite declarations.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
2026-10-02 10:13:09 +00:00
Claude 14567ecac3 test(data): cover sessions, SQLite binding and PostgreSQL wire client
Shared session tests run on SQLite and PostgreSQL (types, nulls,
placeholders, transactions, constraint errors, COPY/bulk escaping,
cancellation). PostgreSQL tests cover SCRAM/MD5/cleartext/trust auth,
TLS, result decoding, bytea, LISTEN/NOTIFY, error mapping, pooling and
broken-connection recovery, and the statement cache. In-memory SQLite
now pools shared-cache connections and retries SQLITE_LOCKED.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
2026-10-02 09:50:31 +00:00