Preserve payloads in v1 status query shims

[andystaples]

Summary

• fetch payloads in the deprecated v1 get_status(), get_status_all(), and get_status_by() shims
• apply show_input locally so disabling input does not discard output, custom status, or failure details
• preserve native durabletask payload-fetch defaults and add running, completed, failed, and collection regression coverage

Testing

• python -m pytest tests\azure-functions-durable -m "not dts and not azurite and not functions_e2e" --quiet
• python -m flake8 azure-functions-durable
• python -m flake8 tests\azure-functions-durable
• python -m pyright -p azure-functions-durable\pyrightconfig.json

Fixes #231

[Copilot]

Pull request overview

This PR fixes the deprecated v1-compatible status query shims in azure-functions-durable so they fetch orchestration payloads needed to reproduce v1 responses, while applying show_input locally to suppress only the input field (without discarding output, custom status, or failure details).

Changes:

• Updated get_status(), get_status_all(), and get_status_by() shims to ensure payloads are fetched (without changing native durabletask defaults).
• Extended DurableOrchestrationStatus to support suppressing input via an internal include_input flag while preserving other payload fields.
• Added regression coverage for running/completed/failed status projections and updated the package changelog.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
azure-functions-durable/azure/durable_functions/client.py	Forces payload fetch in v1 status shim APIs and projects show_input in the wrapper.
azure-functions-durable/azure/durable_functions/internal/compat/durable_orchestration_status.py	Adds include_input support to hide input in v1 JSON/status surface without affecting other payload fields.
tests/azure-functions-durable/test_client_compat.py	Adds/updates tests to validate payload-fetch and v1 projection behavior across runtime statuses and collection shims.
azure-functions-durable/CHANGELOG.md	Documents the user-visible fix under FIXED.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (3)

tests/azure-functions-durable/test_client_compat.py:321

• test_get_status_by_maps_statuses inspects mock.await_args but doesn’t assert the shim awaited get_all_orchestration_states() exactly once. Adding a call-count assertion makes the delegation contract explicit and prevents false positives if the shim calls the API multiple times.

        query = mock.await_args.args[0]
        assert query.runtime_status == [OrchestrationStatus.RUNNING]
        assert query.fetch_inputs_and_outputs is True

tests/azure-functions-durable/test_client_compat.py:202

• test_get_status_all_delegates no longer asserts that get_all_orchestration_states() was awaited exactly once. Without an explicit call-count assertion, the test could pass even if the shim makes multiple calls (or retries) and only the last call has the expected query.

This issue also appears on line 319 of the same file.

        query = mock.await_args.args[0]
        assert query.fetch_inputs_and_outputs is True

tests/azure-functions-durable/test_client_compat.py:589

• The regression test for failed instances validates to_json()["output"] == "boom", but it doesn’t validate the compatibility object’s status.output property. If v1 callers read status.output directly (common), they could still miss the failure message even though to_json() includes it. Consider asserting status.output == "boom" here and updating DurableOrchestrationStatus.output to fall back to failure_details.message when serialized_output is None.

        status_json = status.to_json()
        assert status_json["runtimeStatus"] == "Failed"
        assert status_json["output"] == "boom"
        assert status_json["customStatus"] == "cs"
        assert "input" not in status_json

[berndverst]

One compatibility gap remains: failed statuses expose the failure text only through to_json(), not through the documented .output property.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (2)

tests/azure-functions-durable/test_client_compat.py:638

• Similar to test_get_status_all_returns_wrapped_list, this test validates output on a default RUNNING fake state. durabletask typically doesn't populate serialized_output for RUNNING instances (durabletask/client.py:268-270), so using a COMPLETED fake state here better reflects the terminal-state scenario this shim is meant to preserve (output/custom status/failure details).

async def test_get_status_by_returns_wrapped_list():
    client = _make_client()
    try:
        with patch.object(client, "get_all_orchestration_states",
                          new=AsyncMock(return_value=[_fake_state()])):
            with pytest.warns(DeprecationWarning):
                statuses = await client.get_status_by(
                    runtime_status=[df.OrchestrationRuntimeStatus.Running])
        assert statuses[0].instance_id == "abc"
        assert statuses[0].input_ == {"in": 1}
        assert statuses[0].output == {"out": 2}
        assert statuses[0].custom_status == "cs"

tests/azure-functions-durable/test_client_compat.py:622

• _fake_state() defaults to a RUNNING instance but the assertions in this test expect an output payload. In durabletask, non-terminal instances typically have serialized_output=None (see durabletask/client.py:268-270), so asserting output for RUNNING here makes the test less representative and potentially brittle. Consider exercising a COMPLETED state when validating output preservation.

This issue also appears on line 627 of the same file.

async def test_get_status_all_returns_wrapped_list():
    client = _make_client()
    try:
        with patch.object(client, "get_all_orchestration_states",
                          new=AsyncMock(return_value=[_fake_state()])):
            with pytest.warns(DeprecationWarning):
                statuses = await client.get_status_all()
        assert len(statuses) == 1
        assert statuses[0].runtime_status == df.OrchestrationRuntimeStatus.Running
        assert statuses[0].input_ == {"in": 1}
        assert statuses[0].output == {"out": 2}
        assert statuses[0].custom_status == "cs"

[berndverst]

Re-reviewed the follow-up at this exact head. The failed-status output property now preserves the failure message, to_json() reuses the same projection, and both prior threads are resolved. Focused projection probes, unit tests, lint, strict typing, and compatibility checks pass.