fix: algo-prefix sprint follow-up — test and impl fixes across 12 files
Implementation fixes - code_check.py: _resolve_commit now expands short hex prefixes to full commit IDs via find_commits_by_prefix; short IDs no longer raise ValueError - commit_graph.py: wrap read_commit in try/except ValueError so bad --tip args (invalid IDs, ANSI strings) return a graceful error instead of crashing - conflicts.py: expose _use_color alias for use_color so tests can import it - coordination.py: create_reservation now generates UUID4 for reservation_id instead of calling compute_reservation_id (which returns sha256: IDs that fail _validate_reservation_id's UUID check)
Test fixes - test_cmd_code_add.py: update commit/snapshot path globs to algo-prefixed layout (sha256/<hex>) after the sprint added a directory level - test_cmd_code_check.py: fix commit_id[:8] slice on sha256:<hex> string (yielded only 1 hex char); now strips prefix before slicing - test_cmd_conflicts.py: add _use_color = use_color alias; relax duration_ms < 1.0 threshold to < 5000 (sub-ms was flaky on real hardware) - test_cmd_coord_lifecycle.py: replace fmt='json'/'text' with json_out=True/ False in all 14 Namespace constructions across TestReleaseCmdSingle, TestReleaseCmdBatch, TestHeartbeatCmd, TestCoordGcCmd, TestListCoordLifecycle - test_cmd_coord_list.py: replace --format json with --json flag - test_cmd_dag.py: fix _reserve_ns helper to use json_out=True instead of fmt='json' (already correct in _dag_ns) - test_cmd_describe.py: _REPO_ID and tag_id must be sha256: content-addressed IDs; replace plain strings with content_hash(); drop stale tags/{_REPO_ID} mkdir (write_tag creates its own parent dirs) - test_cmd_describe_hardening.py: same three fixes as test_cmd_describe.py
0 comments
muse hub commit comment sha256:28cad493cb957316dfaaeb80f26f9ffc3d0afa8ff9fd3fd744d09e17a6a874c4 --body "your comment"
No comments yet. Be the first to start the discussion.