[Look at my other PR first pls!] fix(release): use official durabletask release

[sicoyle]

Description

Durabletask python has only ever had prereleases. I justtttt cut a v0.17.0 so we can use an official release here and in dapr agents instead of a pre-release. This PR builds on top of my other PR so pls review this one last as it is only 2 lines of changes to use the newest durabletask version.

Look at this other PR first pls #938

https://pypi.org/project/durabletask-dapr/#history

Issue reference

We strive to have all PR being opened based on an issue, where the problem or feature have been discussed prior to implementation.

Please reference the issue this PR will close: #[issue number]

Checklist

Please make sure you've completed the relevant tasks for this PR, out of the following list:

• Code compiles correctly
• Created/updated tests
• Extended the documentation

[Copilot]

Pull request overview

Updates dependencies and client/runtime behavior to align with the official durabletask-dapr release and newer API surface area.

Changes:

• Bump durabletask-dapr dependency to >= 0.17.0 in workflow extension and dev requirements.
• Add worker readiness polling/logging to WorkflowRuntime and expand tests accordingly.
• Extend Conversation (Alpha2) support for model, token usage details, and request fields like response_format / prompt_cache_retention.

Reviewed changes

Copilot reviewed 18 out of 18 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
ext/dapr-ext-workflow/setup.cfg	Bumps workflow extension dependency to official durabletask-dapr release.
dev-requirements.txt	Keeps dev dependencies aligned with new official durabletask-dapr version.
ext/dapr-ext-workflow/dapr/ext/workflow/workflow_runtime.py	Adds readiness waiting and improves exception logging around workflow/activity execution.
ext/dapr-ext-workflow/dapr/ext/workflow/logger/logger.py	Adds Logger.exception() wrapper used by WorkflowRuntime.
ext/dapr-ext-workflow/tests/test_workflow_runtime.py	Adds unit tests for worker readiness behavior and start/shutdown paths.
dapr/clients/grpc/client.py	Adds converse_alpha2 request fields (response_format, prompt_cache_retention).
dapr/clients/grpc/conversation.py	Adds Alpha2 model/usage dataclasses and populates them from gRPC responses.
tests/clients/test_conversation.py	Adds coverage for Alpha2 model/usage, response_format passthrough, and output parsing helper.
tests/clients/fake_dapr_server.py	Extends fake server to populate model and usage when supported by proto.
ext/dapr-ext-workflow/dapr/ext/workflow/workflow_context.py	Documentation wording change from “cross-app” to “multi-app”.
ext/dapr-ext-workflow/dapr/ext/workflow/dapr_workflow_context.py	Documentation/log wording change from “cross-app” to “multi-app”.
examples/workflow/*.py	Updates examples to rely less on sleeps and use workflow completion waiting.
examples/workflow/README.md	Renames “Cross-app” docs/commands to “Multi-app” terminology.
README.md	Improves contributor setup instructions by adding a venv step and renumbering.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

output.usage / u.completion_tokens_details / u.prompt_tokens_details are protobuf message fields; their truthiness is not a reliable indicator of presence (and may evaluate truthy even when unset). This can incorrectly populate usage (and details) with zero-valued objects when the server didn't send usage, changing API semantics from None to 'all zeros'. Prefer presence checks like HasField('usage') when available, or u.ListFields() / u.ByteSize() > 0 to detect an actually-populated submessage before constructing these dataclasses.

[Copilot]

Re-raising with raise ready_error discards the original traceback context. Use a bare raise here to preserve the original stack trace (you’re already logging with exception()).

Suggested change

	raise ready_error
	raise

[Copilot]

The timeout accounting is based on incrementing by poll_interval, which can drift from wall-clock time (sleep delays, scheduling, etc.). Using a monotonic clock (e.g., store deadline = time.monotonic() + timeout and loop until time.monotonic() >= deadline) makes the timeout behavior more accurate and avoids potential overshoot.

Suggested change

	elapsed = 0.0
	poll_interval = 0.1 # 100ms
	while elapsed < timeout:
	if self.__worker.is_worker_ready():
	return True
	time.sleep(poll_interval)
	elapsed += poll_interval
	poll_interval = 0.1 # 100ms
	deadline = time.monotonic() + timeout
	while time.monotonic() < deadline:
	if self.__worker.is_worker_ready():
	return True
	time.sleep(poll_interval)

[Copilot]

This patcher is started in setUp() but never stopped, which can leak state across tests and make ordering-dependent failures likely. Store the patcher (e.g., self._patcher = mock.patch(...)) and stop it in tearDown() or via self.addCleanup(self._patcher.stop).

Suggested change

	mock.patch('durabletask.worker._Registry', return_value=FakeTaskHubGrpcWorker()).start()
	self._registry_patcher = mock.patch(
	'durabletask.worker._Registry',
	return_value=FakeTaskHubGrpcWorker(),
	)
	self._registry_patcher.start()
	self.addCleanup(self._registry_patcher.stop)

[Copilot]

mock_worker is created with a restrictive spec that does not include is_worker_ready, so del mock_worker.is_worker_ready will raise AttributeError (the attribute doesn't exist). To model 'no is_worker_ready', just avoid creating/deleting it (the spec already achieves that), or use a plain object/MagicMock without setting the attribute at all and rely on hasattr(...) behavior in the implementation.

Suggested change

	del mock_worker.is_worker_ready
	self.runtime._WorkflowRuntime__worker = mock_worker
	self.assertFalse(self.runtime.wait_for_worker_ready(timeout=0.1))
	self.runtime._WorkflowRuntime__worker = mock_worker
	self.assertFalse(self.runtime.wait_for_worker_ready(timeout=0.1))

[Copilot]

Same issue as above: deleting is_worker_ready on a MagicMock(spec=[...]) where the attribute never existed will raise AttributeError. Remove the del ... and let the spec enforce the absence, or use a different mock strategy to represent the missing attribute.

Suggested change

del mock_worker.is_worker_ready

[Copilot]

The broad except Exception: pass will silently hide real incompatibilities/typos in the fake server and can make tests pass while behavior is broken. Prefer catching the specific expected exceptions (e.g., AttributeError, TypeError) and/or at least fail fast (re-raise) so test failures point to the root cause.

Suggested change

	except Exception:
	except (AttributeError, TypeError):
	# Ignore usage population issues due to minor API/type differences in the fake server

[codecov]

Codecov Report

❌ Patch coverage is 89.75610% with 21 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (release-1.17@f3339f6). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
...ext-workflow/dapr/ext/workflow/workflow_runtime.py	64.28%	20 Missing ⚠️
dapr/clients/grpc/client.py	83.33%	1 Missing ⚠️

Additional details and impacted files

@@               Coverage Diff               @@
##             release-1.17     #939   +/-   ##
===============================================
  Coverage                ?   86.89%           
===============================================
  Files                   ?      103           
  Lines                   ?     7305           
  Branches                ?        0           
===============================================
  Hits                    ?     6348           
  Misses                  ?      957           
  Partials                ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sicoyle]

can close this one since it is covered by #944