[zhongkechen]

Issue #, if available

Follow-up to #865, merged as 1653f08. Addresses its three remaining review comments:

• Shared status mapping
• Timestamp documentation
• Literal test expectations, including the optional suggestion

Description of changes

Execution and Invocation views duplicated the operation status rules, so each correction required editing both. Both now call set_operation_span_status in a shared module. Error details still take precedence, FAILED/TIMED_OUT/STOPPED without details remain ERROR, SUCCEEDED remains OK, and other errorless statuses remain UNSET.

The README states the timestamp rules directly, including child-end extension, the 1 µs adjustments, and clock fallback. It removes the misleading statement that checkpointless contexts generally start at completion.

test_operation_status_and_exception now looks up literal expected statuses instead of recomputing the production rule. Expected values are unchanged, and iteration over OperationStatus ensures a new enum member requires an explicit expectation. test_failed_operation_without_type_or_message_uses_fallback also asserts ERROR with an empty status description, preserving the existing no-message behavior.

Design

• Design note (rule 3): not needed; this is an internal refactor and documentation clarification with no public API change.
• Simplification (rule 2): replace both copies of the three-branch mapping with one function. Exception recording uses the existing recorder's no-error return. Timestamp selection remains in each view.
• Root cause: duplicated mapping logic, a test expectation that repeated that logic, and documentation that described a fallback as general checkpointless-context behavior.
• Net source change: +6 lines across three modules, with one shared mapping replacing two copies. No public API members added.

Validation

• Before the extraction, the literal-table status suite passed: 164 tests.
• hatch run dev-otel:test -qq --tb=short -o faulthandler_timeout=60: 874 passed, including public-handler replay tests.
• hatch run dev-otel:typecheck: no issues in 46 files.
• OTel package hatch fmt --check: passed.
• Core and OTel hatch build: passed.
• hatch run test-wheel-otel:test -qq --tb=short: 874 OTel tests and 12 lifecycle compatibility tests passed; installed source bytes and wheel hashes were verified.
• Independent mutations of error precedence, failed-status handling, success handling, preservation of UNSET, and missing-message fallback caused 34, 12, 4, 16, and 2 test failures respectively. All mutations were restored; the final 164 status tests passed and the restored module matches the tested wheel.
• Commit-message lint and git diff --check: passed.
• Python 3.13.5 and Hatch 1.17.0. Full suites ran outside the sandbox after the restricted source run stalled in its first local-runner test and was interrupted.

Assumptions

• Status descriptions and exception events are checked against a real OTel TracerProvider and InMemorySpanExporter in both views. Existing public-handler tests exercise replay and caught failures.
• Deployed Lambda/exporter conformance was not rerun locally for this refactor. Runtime timing and lifecycle contracts are unchanged.

Decisions not covered by the issue or the design note

• Put the shared function in operation_span_status.py to keep status policy separate from exception event construction and each view's timestamp policy.
• Call the existing exception recorder for every operation end; its existing None check suppresses absent diagnostics without another condition in each view.
• Strengthen the existing empty-error test to check the moved message fallback; reuse lifecycle coverage rather than add duplicate direct-function tests.

Checklist

• Each moved status branch has meaningful lifecycle coverage; branch mutation results are recorded above (rule 4).
• This refactor preserves behavior; a failing bug reproduction is not applicable (rule 5).
• Every edited test expectation is listed above with its reason (rule 5).

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[github-actions]

Codex AI review

No actionable findings in the five changed files, including against CONTRIBUTING.md rules 1–9. Residual test risk: tests were inspected but not executed under the review constraints.

Reviewed commit 9b0166abf2ee5e5e12bd1539677f015651281f56. Workflow run