1.9 KiB
Prompt
After rebasing the TIPC backend branch onto current main, inspect the result before the human force-pushes it. Repair and verify any integration problems, then provide a proper /commit-plan rather than an ad hoc staging and commit suggestion.
Response
The range-diff showed that the two omitted pformat commits were already present on main, while the conflict resolutions correctly retained current platform-aware transport behavior. It also exposed a lost uv.lock update: pyproject.toml retained the unreleased WG-capable multiaddr source pin, but the lock still selected the PyPI release.
git diff HEAD~1..HEAD -- uv.lock
Regenerated the lock against current main, preserving its newly landed dependencies while selecting py-multiaddr revision f86519da.
The full test suite then found that eagerly importing Multiaddr from the new TIPC module violated main’s cold-import contract.
git diff HEAD~1..HEAD -- tractor/ipc/_tipc.py
Moved Multiaddr under TYPE_CHECKING and supplied the runtime Any alias used by the sibling TCP and UDS backends, retaining lazy multiaddr loading and runtime annotation resolution.
git diff HEAD~1..HEAD -- tests/test_lazy_imports.py
Extended lazy annotation checks to the TIPC stream and updated mk_maddr()’s expected runtime return type for the branch’s interim Multiaddr|str TIPC grammar.
Validation completed:
uv lock --checkpassed.- Focused lazy-import, TIPC, and multiaddr checks: 60 passed.
- Full default transport suite: 548 passed, 9 skipped, 7 xfailed, 3 xpassed.
- Full
--tpt-proto=tipcsuite: 548 passed, 9 skipped, 8 xfailed, 2 xpassed.
The human corrected the initial ad hoc commit suggestion and required the repository’s complete /commit-plan workflow.