stage #4594 onto master (v0.51.550) — preserve visible transcript after manual /compress (#3133)

Manual /compress wrote the compressed payload into BOTH s.messages (visible) and s.context_messages
(model-facing) + wiped s.tool_calls -> messages vanished from the window. Fix keeps s.messages +
tool history durable, deep-copies compressed only into s.context_messages, anchors off visible msgs.
Co-authored-by: franksong2702 <franksong2702@users.noreply.github.com>

api/routes.py

@@ -18397,15 +18397,13 @@ def _handle_session_compress(handler, body):
             if _sanitize_messages_for_api(s.messages) != original_messages:
                 return bad(handler, "Session was modified during compression; please retry.", 409)
 
-            s.messages = compressed
-            s.context_messages = compressed
-            s.tool_calls = []
+            s.context_messages = copy.deepcopy(compressed)
             s.active_stream_id = None
             s.pending_user_message = None
             s.pending_attachments = []
             s.pending_started_at = None
             s.pending_user_source = None
-            visible_after = visible_messages_for_anchor(compressed, auto_compression=False)
+            visible_after = visible_messages_for_anchor(s.messages, auto_compression=False)
             s.compression_anchor_visible_idx = max(0, len(visible_after) - 1) if visible_after else None
             s.compression_anchor_message_key = _anchor_message_key(visible_after[-1]) if visible_after else None
             summary_text = None

tests/test_gateway_sync.py

@@ -1331,7 +1331,13 @@ def test_sessions_response_backfills_imported_messaging_source_metadata(cleanup_
     sid = 'gw_legacy_import_weixin_001'
     cleanup_test_sessions.append(sid)
     try:
-        _insert_gateway_session(conn, session_id=sid, source='weixin', title='Weixin Session')
+        _insert_gateway_session(
+            conn,
+            session_id=sid,
+            source='weixin',
+            title='Weixin Session',
+            session_key='agent:test:weixin:legacy-import-backfill',
+        )
         s = Session(
             session_id=sid,
             title='Legacy Imported Weixin',

tests/test_issue2028_compression_anchor_helpers.py

@@ -14,7 +14,7 @@ def test_legacy_duplicate_anchor_helpers_are_removed():
 
     assert "def _visible_messages_for_anchor" not in routes_src
     assert "def _visible_messages_for_compression_anchor" not in streaming_src
-    assert "visible_messages_for_anchor(compressed, auto_compression=False)" in routes_src
+    assert "visible_messages_for_anchor(s.messages, auto_compression=False)" in routes_src
     assert "visible_messages_for_anchor(s.messages, auto_compression=True)" in streaming_src
 
 

tests/test_sprint46.py

@@ -102,7 +102,7 @@ def _install_fake_compression_runtime(monkeypatch, agent_cls):
     )
 
 
-def _make_session(messages=None):
+def _make_session(messages=None, tool_calls=None):
     SESSION_DIR.mkdir(parents=True, exist_ok=True)
     messages = messages or [
         {"role": "user", "content": "one"},
@@ -116,6 +116,7 @@ def _make_session(messages=None):
         workspace="/tmp/hermes-webui-test",
         model="openai/gpt-5.4-mini",
         messages=messages,
+        tool_calls=tool_calls or [],
     )
     s.save(touch_updated_at=False)
     return s.session_id
@@ -130,7 +131,23 @@ def test_session_compress_requires_session_id(cleanup_test_sessions):
 
 def test_session_compress_roundtrip(monkeypatch, cleanup_test_sessions):
     created = cleanup_test_sessions
-    sid = _make_session()
+    original_messages = [
+        {"role": "user", "content": "one"},
+        {"role": "assistant", "content": "two"},
+        {"role": "user", "content": "three"},
+        {"role": "assistant", "content": "four"},
+    ]
+    compressed_messages = [original_messages[0], original_messages[-1]]
+    settled_tool_calls = [
+        {
+            "id": "call_1",
+            "name": "terminal",
+            "assistant_msg_idx": 1,
+            "done": True,
+            "result": "schema.sql",
+        }
+    ]
+    sid = _make_session(original_messages, tool_calls=settled_tool_calls)
     created.append(sid)
 
     _install_fake_compression_runtime(monkeypatch, _FakeAgent)
@@ -144,18 +161,27 @@ def test_session_compress_roundtrip(monkeypatch, cleanup_test_sessions):
     assert payload["focus_topic"] == "database schema"
     assert payload["summary"]["headline"] == "Compressed: 4 → 2 messages"
     assert payload["session"]["session_id"] == sid
-    assert payload["session"]["messages"] == [
-        {"role": "user", "content": "one"},
-        {"role": "assistant", "content": "four"},
-    ]
+    assert payload["session"]["messages"] == original_messages
+    assert payload["session"]["tool_calls"] == settled_tool_calls
+    assert payload["session"]["message_count"] == len(original_messages)
     assert payload["session"]["compression_anchor_summary"] is not None
-    assert payload["session"]["compression_anchor_visible_idx"] == 1
+    assert payload["session"]["compression_anchor_visible_idx"] == 3
     assert isinstance(payload["session"]["compression_anchor_message_key"], dict)
     assert payload["session"]["compression_anchor_message_key"].get("role") == "assistant"
+    assert payload["session"]["compression_anchor_message_key"].get("text") == "four"
     loaded = get_session(sid)
+    persisted = Session.load(sid)
+    assert loaded.messages == original_messages
+    assert loaded.context_messages == compressed_messages
+    assert loaded.tool_calls == settled_tool_calls
+    assert persisted.messages == original_messages
+    assert persisted.context_messages == compressed_messages
+    assert persisted.tool_calls == settled_tool_calls
     assert loaded.compression_anchor_summary == payload["session"]["compression_anchor_summary"]
     assert loaded.compression_anchor_visible_idx == payload["session"]["compression_anchor_visible_idx"]
     assert loaded.compression_anchor_message_key == payload["session"]["compression_anchor_message_key"]
+    assert persisted.compression_anchor_visible_idx == 3
+    assert persisted.compression_anchor_message_key == payload["session"]["compression_anchor_message_key"]
     assert _FakeAgent.last_instance is not None
     assert _FakeAgent.last_instance.context_compressor.calls[0]["focus_topic"] == "database schema"
 
@@ -185,7 +211,16 @@ def test_session_compress_start_is_async_and_reuses_running_job(monkeypatch, cle
             self.instances.append(self)
 
     created = cleanup_test_sessions
-    sid = _make_session()
+    settled_tool_calls = [
+        {
+            "id": "call_async_1",
+            "name": "read_file",
+            "assistant_msg_idx": 1,
+            "done": True,
+            "result": "config.yaml",
+        }
+    ]
+    sid = _make_session(tool_calls=settled_tool_calls)
     created.append(sid)
     _install_fake_compression_runtime(monkeypatch, BlockingAgent)
     try:
@@ -226,6 +261,15 @@ def test_session_compress_start_is_async_and_reuses_running_job(monkeypatch, cle
     assert done_payload is not None
     assert done_payload["summary"]["headline"] == "Compressed: 4 → 2 messages"
     assert done_payload["session"]["messages"] == [
+        {"role": "user", "content": "one"},
+        {"role": "assistant", "content": "two"},
+        {"role": "user", "content": "three"},
+        {"role": "assistant", "content": "four"},
+    ]
+    assert done_payload["session"]["tool_calls"] == settled_tool_calls
+    persisted = Session.load(sid)
+    assert persisted.tool_calls == settled_tool_calls
+    assert persisted.context_messages == [
         {"role": "user", "content": "one"},
         {"role": "assistant", "content": "four"},
     ]