durable-workflow 2.3.5__tar.gz → 2.3.7__tar.gz

{durable_workflow-2.3.5/src/durable_workflow.egg-info → durable_workflow-2.3.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: durable-workflow
-Version: 2.3.5
+Version: 2.3.7
 Summary: Python client and worker SDK for Durable Workflow Cloud and self-hosted Server
 Author: Durable Workflow Contributors
 License-Expression: MIT

{durable_workflow-2.3.5 → durable_workflow-2.3.7}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "durable-workflow"
-version = "2.3.5"
+version = "2.3.7"
 description = "Python client and worker SDK for Durable Workflow Cloud and self-hosted Server"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -71,8 +71,8 @@ durable-workflow-replay-conformance = "durable_workflow.replay_conformance:main"
 durable-workflow-workflow-updates-conformance = "durable_workflow.workflow_updates_conformance:main"
 
 [tool.durable-workflow]
-product-train = "2.3.5"
-registry-version = "2.3.5"
+product-train = "2.3.7"
+registry-version = "2.3.7"
 supported-server-versions = "2.4.8"
 worker-protocol-version = "1.19"
 control-plane-version = "2"

{durable_workflow-2.3.5 → durable_workflow-2.3.7}/src/durable_workflow/external_storage.py

@@ -250,7 +250,7 @@ class ExternalPayloadCache:
     contract as a fresh driver fetch.
     """
 
-    def __init__(self, *, max_entries: int = 128, max_bytes: int = 16 * 1024 * 1024) -> None:
+    def __init__(self, *, max_entries: int = 1024, max_bytes: int = 16 * 1024 * 1024) -> None:
         if max_entries < 1:
             raise ValueError("external payload cache max_entries must be at least 1")
         if max_bytes < 1:

{durable_workflow-2.3.5 → durable_workflow-2.3.7}/src/durable_workflow/worker.py

@@ -98,6 +98,7 @@ from .workflow import (
 log = logging.getLogger("durable_workflow.worker")
 
 QUERY_TASKS_CAPABILITY = "query_tasks"
+WORKFLOW_HISTORY_PAGE_SIZE = 500
 UPDATE_VALIDATION_TASKS_CAPABILITY = "update_validation_tasks"
 WORKFLOW_UPDATES_CAPABILITY = "workflow_updates"
 MESSAGE_STREAMS_CAPABILITY = "message_streams"
@@ -1637,22 +1638,19 @@ class Worker:
         wf_type: str = task.get("workflow_type", "")
         history = task.get("history_events", [])
 
-        # Paginate history if needed
+        # The worker requests bounded history pages when polling. Do not replay
+        # an incomplete history if fetching a later page fails.
         next_page_token = task.get("next_history_page_token")
         while next_page_token:
-            try:
-                page_data = await self.client.workflow_task_history(
-                    task_id=task_id,
-                    next_history_page_token=next_page_token,
-                    lease_owner=self.worker_id,
-                    workflow_task_attempt=attempt,
-                )
-                if page_data and page_data.get("history_events"):
-                    history.extend(page_data["history_events"])
-                next_page_token = page_data.get("next_history_page_token") if page_data else None
-            except Exception as e:
-                log.warning("failed to fetch history page for task %s: %s", task_id, e)
-                break
+            page_data = await self.client.workflow_task_history(
+                task_id=task_id,
+                next_history_page_token=next_page_token,
+                lease_owner=self.worker_id,
+                workflow_task_attempt=attempt,
+            )
+            if page_data and page_data.get("history_events"):
+                history.extend(page_data["history_events"])
+            next_page_token = page_data.get("next_history_page_token") if page_data else None
 
         start_input: list[Any] = []
         codec = task.get("payload_codec")
@@ -2535,6 +2533,7 @@ class Worker:
                     timeout=self._poll_http_timeout,
                     build_id=self.build_id,
                     task_kinds=self._workflow_poll_task_kinds(),
+                    history_page_size=WORKFLOW_HISTORY_PAGE_SIZE,
                 )
             except asyncio.CancelledError:
                 self._release_workflow_capacity()
@@ -3224,6 +3223,7 @@ class Worker:
                         timeout=self._poll_http_timeout,
                         build_id=self.build_id,
                         task_kinds=self._workflow_poll_task_kinds(),
+                        history_page_size=WORKFLOW_HISTORY_PAGE_SIZE,
                     )
                 except BaseException:
                     self._release_workflow_capacity()

{durable_workflow-2.3.5 → durable_workflow-2.3.7/src/durable_workflow.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: durable-workflow
-Version: 2.3.5
+Version: 2.3.7
 Summary: Python client and worker SDK for Durable Workflow Cloud and self-hosted Server
 Author: Durable Workflow Contributors
 License-Expression: MIT

{durable_workflow-2.3.5 → durable_workflow-2.3.7}/tests/test_external_storage.py

@@ -1,6 +1,6 @@
 import hashlib
 from pathlib import Path
-from unittest.mock import AsyncMock
+from unittest.mock import AsyncMock, patch
 
 import pytest
 
@@ -740,6 +740,26 @@ def test_fetch_external_payload_cache_reuses_verified_bytes(tmp_path: Path) -> N
     assert len(cache) == 1
 
 
+def test_default_cache_reuses_large_replay_reference_set(tmp_path: Path) -> None:
+    storage = LocalFilesystemExternalStorage(tmp_path)
+    cache = ExternalPayloadCache()
+    references = [
+        store_external_payload(storage, f"signal-{index}".encode(), codec="avro")
+        for index in range(1000)
+    ]
+
+    with patch.object(storage, "get", wraps=storage.get) as get:
+        for _ in range(2):
+            for index, reference in enumerate(references):
+                assert fetch_external_payload(storage, reference, cache=cache) == (
+                    f"signal-{index}".encode()
+                )
+
+    assert get.call_count == len(references)
+    assert len(cache) == len(references)
+    assert cache.current_bytes <= cache.max_bytes == 16 * 1024 * 1024
+
+
 def test_delete_external_payload_removes_blob_and_cache_entry(tmp_path: Path) -> None:
     storage = LocalFilesystemExternalStorage(tmp_path)
     cache = ExternalPayloadCache(max_entries=2, max_bytes=1024)

{durable_workflow-2.3.5 → durable_workflow-2.3.7}/tests/test_replay_regression_corpus.py

@@ -4,11 +4,12 @@ import json
 from collections.abc import Mapping, Sequence
 from pathlib import Path
 from typing import Any
+from unittest.mock import AsyncMock
 
 import pytest
 
-from durable_workflow import Replayer, serializer, workflow
-from durable_workflow.client import WorkflowStreamAppendItem
+from durable_workflow import Replayer, Worker, serializer, workflow
+from durable_workflow.client import Client, WorkflowStreamAppendItem
 from durable_workflow.errors import NonDeterministicReplayError, WorkflowPayloadDecodeError
 from durable_workflow.workflow import WorkflowContext, commands_to_server_commands, query_state
 from tests.test_golden_history_replay import (
@@ -156,6 +157,17 @@ class RecordedSideEffectWorkflow:
         return (yield ctx.side_effect(unexpected_execution))
 
 
+@workflow.defn(name="tests.replay.paged-recorded-side-effects")
+class PagedRecordedSideEffectsWorkflow:
+    def run(self, ctx: WorkflowContext):  # type: ignore[no-untyped-def]
+        def unexpected_execution() -> int:
+            raise AssertionError("recorded side effect callable ran during replay")
+
+        first = yield ctx.side_effect(unexpected_execution)
+        second = yield ctx.side_effect(unexpected_execution)
+        return first + second
+
+
 @workflow.defn(name="tests.replay.message-stream-consumer")
 class MessageStreamConsumerWorkflow:
     def run(self, ctx: WorkflowContext):  # type: ignore[no-untyped-def]
@@ -212,6 +224,7 @@ WORKFLOWS = [
     NestedParallelPathWorkflow,
     ParallelMetadataProducerWorkflow,
     ParallelResultBindingWorkflow,
+    PagedRecordedSideEffectsWorkflow,
     PostConditionReceiversWorkflow,
     RecordedSideEffectWorkflow,
     SelectionAwaitMarkerWorkflow,
@@ -394,6 +407,41 @@ def test_checked_in_replay_regression_corpus_uses_official_replayer(
     _execute_fixture(fixture)
 
 
+@pytest.mark.asyncio
+async def test_worker_replays_recorded_side_effects_across_history_pages() -> None:
+    fixture = json.loads((FIXTURE_DIR / "paged-recorded-side-effects.json").read_text(encoding="utf-8"))
+    history = fixture["history"]
+    client = AsyncMock(spec=Client)
+    client.workflow_task_history.return_value = {
+        "history_events": history[1:],
+        "next_history_page_token": None,
+    }
+    worker = Worker(client, task_queue="paged-history", workflows=[PagedRecordedSideEffectsWorkflow])
+
+    commands = await worker._run_workflow_task_core({
+        "task_id": "task-paged-side-effects",
+        "workflow_id": "workflow-paged-side-effects",
+        "run_id": "run-paged-side-effects",
+        "workflow_type": fixture["workflow"]["type"],
+        "workflow_task_attempt": 1,
+        "history_events": history[:1],
+        "next_history_page_token": "second-page",
+        "arguments": serializer.encode([], codec="avro"),
+        "payload_codec": "avro",
+    })
+
+    client.workflow_task_history.assert_awaited_once_with(
+        task_id="task-paged-side-effects",
+        next_history_page_token="second-page",
+        lease_owner=worker.worker_id,
+        workflow_task_attempt=1,
+    )
+    client.complete_workflow_task.assert_awaited_once()
+    assert commands is not None
+    assert [command["type"] for command in commands] == ["complete_workflow"]
+    assert serializer.decode_envelope(commands[0]["result"]) == 198
+
+
 @pytest.mark.parametrize(
     "fixture",
     [

{durable_workflow-2.3.5 → durable_workflow-2.3.7}/tests/test_worker.py

@@ -836,6 +836,7 @@ class TestWorkerRegistration:
         workflow_kwargs = mock_client.poll_workflow_task.call_args.kwargs
         assert workflow_called.is_set()
         assert workflow_kwargs["build_id"] == "release-2026.04.22-a1"
+        assert workflow_kwargs["history_page_size"] == 500
 
         worker._stop.clear()
         mock_client.poll_activity_task.side_effect = activity_poll_once
@@ -1189,6 +1190,31 @@ class TestWorkerRegistration:
 
 
 class TestWorkflowTaskExecution:
+    @pytest.mark.asyncio
+    async def test_history_page_failure_does_not_replay_partial_history(
+        self, mock_client: AsyncMock
+    ) -> None:
+        mock_client.workflow_task_history.side_effect = RuntimeError("history page unavailable")
+        worker = Worker(mock_client, task_queue="q1", workflows=[TestWorkflow])
+        task = {
+            "task_id": "t-paged-history",
+            "workflow_type": "test-wf",
+            "workflow_task_attempt": 1,
+            "history_events": [{"event_id": "first-page"}],
+            "next_history_page_token": "next-page",
+        }
+
+        with pytest.raises(RuntimeError, match="history page unavailable"):
+            await worker._run_workflow_task_core(task)
+
+        mock_client.workflow_task_history.assert_awaited_once_with(
+            task_id="t-paged-history",
+            next_history_page_token="next-page",
+            lease_owner=worker.worker_id,
+            workflow_task_attempt=1,
+        )
+        mock_client.complete_workflow_task.assert_not_awaited()
+
     @pytest.mark.asyncio
     async def test_local_activity_runs_in_workflow_worker_and_records_heartbeats(
         self, mock_client: AsyncMock
@@ -4948,6 +4974,7 @@ class TestRunUntil:
         assert worker._stop.is_set()
         mock_client.register_worker.assert_awaited_once()
         assert mock_client.describe_workflow.await_count == 2
+        assert mock_client.poll_workflow_task.call_args.kwargs["history_page_size"] == 500
 
     @pytest.mark.asyncio
     async def test_run_until_times_out_and_stops_worker(self, mock_client: AsyncMock) -> None: