Repository navigation
Test PostgreSQL cancellation, transactions and pool recovery across event loops - #114
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds 36 real PostgreSQL cases using asyncpg and SQLAlchemy AsyncEngine across asyncio, rsloop and uvloop. Tests in-flight cancellation/timeouts with server-observed advisory lock barriers, rollback after application exceptions, concurrent transaction visibility, pool exhaustion and backend PID reuse, and replacement connections after terminating the active test backend (real server-side TCP close).
Adds isolated PostgreSQL 16 services on Linux/Python 3.10 and 3.14, locked asyncpg dependency and environment/version recording. Existing cross-platform SQLite/HTTP/WebSocket jobs remain intact. PostgreSQL tests opt in through RSLOOP_POSTGRES_DSN and use unique tables/locks, bounded cleanup and the existing task-leak fixture.
Validation actually executed:
40a9ad8: Linux x86_64 / Python 3.10.22 36 passed; Linux x86_64 / Python 3.14.8 36 passed. Both used PostgreSQL 16.15, asyncpg 0.32.0, SQLAlchemy 2.0.52 and uvloop 0.23.0, comparing all three loops. No PostgreSQL case skipped in these jobs.No rsloop production defect was reproduced; no production implementation change was needed. PostgreSQL ran in CI, not locally; PostgreSQL macOS/Windows and production-load compatibility are not claimed. The final commit updates only documentation with these actual execution environments and links. See docs/async-ecosystem-tests.md for synchronization and limits.
Final-head
40183adCI: Async ecosystem completed successfully, including both PostgreSQL jobs (36 passed each), Linux/Python 3.10 and 3.14, macOS/Python 3.14 and Windows/Python 3.14. PostgreSQL jobs used the same server/client versions listed above. MkDocs passed. The earlier workflow was superseded by this documentation commit; those cancelled jobs are not test failures.Final-head Tests also completed successfully: the core OS/Python matrix (including free-threaded interpreters), Rust checks, Clippy, Miri, profiling, tooling, AnyIO and Redis/hiredis are green. All three final-head workflows completed successfully; no CI failures remain.