[YongGoose]

The Redis and PostgreSQL persister integration tests left their data behind, so re-running them against the same databases failed 22 tests with duplicate-key errors. Each test now removes what it wrote.

Closes #1034

Changes

• test_postgresql.py: the sync and async fixtures delete the rows in their test table on teardown. I used DELETE rather than DROP TABLE because DROP needs an exclusive lock and waits indefinitely behind any connection still idle in a read transaction (for example the unpickled persister in test_serialization_with_pickle, until it is garbage collected). DELETE doesn't conflict with readers.
• test_b_redis.py: an autouse fixture records the keys in db 0 and db 1 before each test and deletes only the keys that appeared during it. This covers the fixture-based tests and test_redis_persister_class_backwards_compatible, and leaves any existing data in db 0 (Redis' default db) alone, so there's no FLUSHDB.

The persisters themselves are unchanged.

How I tested this

With postgres:15, mongo:7.0 and redis:7 (the CI images):

• Ran BURR_CI_INTEGRATION_TESTS=true pytest tests/integrations/persisters three times in a row against the same databases: 68 passed each time. Before this change: 68 passed on the first run, then 22 failed on every later run.
• Afterwards no test keys are left in Redis and both Postgres test tables are empty. A key I set in db 0 and db 1 before the runs was still there.
• Got the same result running the suite twice inside a Linux container (Python 3.12).
• Checked the DELETE vs DROP TABLE reasoning: with another connection idle in transaction after a SELECT on testtable, DELETE FROM testtable returned immediately, while DROP TABLE testtable was still waiting after 5 minutes.
• pre-commit run passes on the changed files.

Notes

• If your local databases already hold leftovers from earlier runs, reset them once (for example, recreate the containers). The Redis cleanup deliberately never deletes keys that existed before a test started.
• fix: end each psycopg2 persister call's transaction #950 and fix(redis): support decoded client responses #1012 also add tests to these two files. They only append tests and don't touch the fixtures, so this merges cleanly with both (checked with git merge-tree). The new Postgres tests in fix: end each psycopg2 persister call's transaction #950 use the same fixture, so they get cleaned up too.

Checklist

• PR has an informative and human-readable title (this will be pulled into the release notes)
• Changes are limited to a single goal (no scope creep)
• Code passed the pre-commit check & code is left cleaner/nicer than when first encountered.
• Any change in functionality is tested
• New functions are documented (with a description, list of inputs, and expected output)
• Placeholder code is flagged / future TODOs are captured in comments (n/a)
• Project documentation has been updated if adding/changing functionality. (n/a, test-only change)

🤖 Generated with Claude Code