ci: de-vendor helm/chart-releaser-action

helm/chart-releaser-action is now allowlisted upstream (ASF Actions
allowlist, apache/infrastructure-actions/actions.yml) at v1.7.0
(cae68fefc6b5f367a0275617c9f83181ba54714f), so the git-submodule fork
under .github/actions/chart-releaser-action — added as a workaround
when it wasn't allowlisted — is no longer needed. Verified with a
real CI run (install_only, no release side effects) before landing
this.

Also drops the now-dead debug step that cat'd the vendored action's
action.yml.

helm/chart-testing-action stays vendored for now: it's allowlisted at
v2.8.0 (6ec842c01de15ebb84c8627d2744a0c2f2755c9f), but that release
depends internally on astral-sh/setup-uv@v7.0.0, which isn't itself on
the allowlist (only v8.1.0+ are). De-vendoring it needs an INFRA
request to add that setup-uv SHA first — confirmed by testing it
directly, not just reading the config.

.github/workflows/superset-helm-lint-test.yml

@@ -38,6 +38,11 @@ jobs:
         with:
           install-superset: "false"
 
+      # Still vendored (not de-vendored like chart-releaser-action below): the
+      # allowlisted helm/chart-testing-action@v2.8.0 depends internally on
+      # astral-sh/setup-uv@v7.0.0, which isn't itself on the ASF Actions
+      # allowlist (only v8.1.0+ are, at apache/infrastructure-actions'
+      # actions.yml). Needs an INFRA request before this can de-vendor too.
       - name: Set up chart-testing
         uses: ./.github/actions/chart-testing-action
 

.github/workflows/superset-helm-release.yml

@@ -86,13 +86,8 @@ jobs:
           # Return to the original branch
           git checkout local_gha_temp
 
-      - name: Fetch/list all tags
-        run: |
-          git submodule update
-          cat .github/actions/chart-releaser-action/action.yml
-
       - name: Run chart-releaser
-        uses: ./.github/actions/chart-releaser-action
+        uses: helm/chart-releaser-action@cae68fefc6b5f367a0275617c9f83181ba54714f # v1.7.0
         with:
           version: v1.6.0
           charts_dir: helm

.gitmodules

@@ -30,9 +30,6 @@
 [submodule ".github/actions/chart-testing-action"]
 	path = .github/actions/chart-testing-action
 	url = https://github.com/helm/chart-testing-action
-[submodule ".github/actions/chart-releaser-action"]
-	path = .github/actions/chart-releaser-action
-	url = https://github.com/helm/chart-releaser-action
 [submodule ".github/actions/github-action-push-to-another-repository"]
 	path = .github/actions/github-action-push-to-another-repository
 	url = https://github.com/cpina/github-action-push-to-another-repository

superset/common/query_context_factory.py

@@ -318,7 +318,7 @@ class QueryContextFactory:  # pylint: disable=too-few-public-methods
             # another temporal filter. A new filter based on the value of
             # the granularity will be added later in the code.
             # In practice, this is replacing the previous default temporal filter.
-            if filter_to_remove and is_adhoc_column(filter_to_remove):  # type: ignore
+            if is_adhoc_column(filter_to_remove):  # type: ignore
                 filter_to_remove = filter_to_remove.get("sqlExpression")
 
             if filter_to_remove:

superset/common/utils/query_cache_manager.py

@@ -206,9 +206,11 @@ class QueryCacheManager:
                 )
                 query_cache.status = QueryStatus.SUCCESS
                 query_cache.is_loaded = True
-                query_cache.is_cached = True
+                query_cache.is_cached = cache_value is not None
                 query_cache.sql_rowcount = cache_value.get("sql_rowcount", None)
-                query_cache.cache_dttm = cache_value["dttm"]
+                query_cache.cache_dttm = (
+                    cache_value["dttm"] if cache_value is not None else None
+                )
                 query_cache.queried_dttm = cache_value.get(
                     "queried_dttm", cache_value.get("dttm")
                 )

superset/daos/dashboard.py

@@ -397,8 +397,7 @@ class DashboardDAO(BaseDAO[Dashboard]):
             md["color_namespace"] = data.get("color_namespace")
 
         md["expanded_slices"] = data.get("expanded_slices", {})
-        if "refresh_frequency" in data:
-            md["refresh_frequency"] = data["refresh_frequency"]
+        md["refresh_frequency"] = data.get("refresh_frequency", 0)
         md["color_scheme"] = data.get("color_scheme", "")
         md["label_colors"] = data.get("label_colors", {})
         md["shared_label_colors"] = data.get("shared_label_colors", [])

superset/db_engine_specs/databricks.py

@@ -534,7 +534,6 @@ class DatabricksDynamicBaseEngineSpec(BasicParametersMixin, DatabricksBaseEngine
         ],
     ) -> list[SupersetError]:
         errors: list[SupersetError] = []
-        connect_args: dict[str, Any] = {}
         if extra := json.loads(properties.get("extra")):  # type: ignore
             engine_params = extra.get("engine_params", {})
             connect_args = engine_params.get("connect_args", {})

superset/db_engine_specs/presto.py

@@ -1235,7 +1235,6 @@ class PrestoEngineSpec(PrestoBaseEngineSpec):
         all_columns: list[ResultSetColumnType] = []
         expanded_columns = []
         current_array_level = None
-        unnested_rows: dict[int, int] = defaultdict(int)
         while to_process:
             column, level = to_process.popleft()
             if column["column_name"] not in [
@@ -1249,7 +1248,7 @@ class PrestoEngineSpec(PrestoBaseEngineSpec):
             # added by the first. every time we change a level in the nested arrays
             # we reinitialize this.
             if level != current_array_level:
-                unnested_rows = defaultdict(int)
+                unnested_rows: dict[int, int] = defaultdict(int)
                 current_array_level = level
 
             name = column["column_name"]

tests/unit_tests/common/test_query_context_factory.py

@@ -532,42 +532,6 @@ class TestQueryContextFactory:
 
         assert query_object.columns == ["ds", "other_col"]
 
-    def test_apply_granularity_no_filter_to_remove(self):
-        """No x-axis and no temporal filters leaves the filters untouched."""
-        query_object = Mock(spec=QueryObject)
-        query_object.granularity = "P1D"
-        query_object.columns = ["other_col"]
-        query_object.post_processing = []
-        query_object.filter = [{"col": "other_col", "op": "==", "val": "value"}]
-
-        datasource = Mock()
-        datasource.columns = [{"column_name": "ds", "is_dttm": True}]
-
-        self.factory._apply_granularity(query_object, {}, datasource)
-
-        assert query_object.filter == [{"col": "other_col", "op": "==", "val": "value"}]
-
-    def test_apply_granularity_with_adhoc_temporal_filter(self):
-        """An adhoc temporal filter is matched on its SQL expression."""
-        adhoc_column = {"label": "ds_expr", "sqlExpression": "DATE(ds)"}
-        query_object = Mock(spec=QueryObject)
-        query_object.granularity = "P1D"
-        query_object.columns = ["other_col"]
-        query_object.post_processing = []
-        query_object.filter = [
-            {"col": adhoc_column, "op": "TEMPORAL_RANGE", "val": "a : b"},
-            {"col": "DATE(ds)", "op": "TEMPORAL_RANGE", "val": "a : b"},
-        ]
-
-        datasource = Mock()
-        datasource.columns = [{"column_name": "ds", "is_dttm": True}]
-
-        self.factory._apply_granularity(query_object, {}, datasource)
-
-        assert query_object.filter == [
-            {"col": adhoc_column, "op": "TEMPORAL_RANGE", "val": "a : b"}
-        ]
-
     def test_apply_filters_with_time_range(self):
         """Test _apply_filters with time_range"""
         query_object = Mock(spec=QueryObject)

tests/unit_tests/dao/dashboard_test.py

@@ -24,7 +24,6 @@ from superset.connectors.sqla.models import Database, SqlaTable
 from superset.daos.dashboard import DashboardDAO
 from superset.models.dashboard import Dashboard
 from superset.models.slice import Slice
-from superset.utils import json
 from tests.unit_tests.conftest import with_feature_flags
 
 
@@ -118,53 +117,3 @@ def test_set_dash_metadata_preserves_soft_deleted_members(
     )
     # And the position slot kept its UUID rather than being nulled.
     assert positions["CHART-trashed"]["meta"]["uuid"] == str(trashed_chart.uuid)
-
-
-def test_set_dash_metadata_preserves_refresh_frequency(session: Session) -> None:
-    """set_dash_metadata must not reset refresh_frequency when absent from data.
-
-    Regression test for #42116: ``data.get("refresh_frequency", 0)`` would
-    unconditionally overwrite the existing value with 0 whenever the caller
-    did not include ``refresh_frequency`` in the data dict.
-    """
-    Dashboard.metadata.create_all(session.get_bind())
-
-    dashboard = Dashboard(
-        dashboard_title="refresh_test_dash",
-        json_metadata=json.dumps({"refresh_frequency": 30}),
-    )
-    db.session.add(dashboard)
-    db.session.flush()
-
-    # Simulate a save that does NOT include refresh_frequency
-    # (e.g. changing only the title via the PropertiesModal).
-    DashboardDAO.set_dash_metadata(dashboard, {"color_scheme": "superset"})
-
-    md = json.loads(dashboard.json_metadata)
-    assert md["refresh_frequency"] == 30, (
-        "refresh_frequency should be preserved when not present in data"
-    )
-
-
-def test_set_dash_metadata_updates_refresh_frequency_when_present(
-    session: Session,
-) -> None:
-    """set_dash_metadata must update refresh_frequency when it IS in data."""
-    Dashboard.metadata.create_all(session.get_bind())
-
-    dashboard = Dashboard(
-        dashboard_title="refresh_test_dash_2",
-        json_metadata=json.dumps({"refresh_frequency": 30}),
-    )
-    db.session.add(dashboard)
-    db.session.flush()
-
-    # Simulate a save that explicitly sets refresh_frequency to 0.
-    DashboardDAO.set_dash_metadata(
-        dashboard, {"refresh_frequency": 0, "color_scheme": "superset"}
-    )
-
-    md = json.loads(dashboard.json_metadata)
-    assert md["refresh_frequency"] == 0, (
-        "refresh_frequency should be updated when present in data"
-    )

tests/unit_tests/result_set_test.py

@@ -571,53 +571,3 @@ def test_stringify_values_non_serializable_dict_falls_back_to_str() -> None:
     # Must not raise — falls back to str()
     result = stringify_values(data)
     assert result[0] == str({"key": _Unserializable()})
-
-
-def test_empty_result_set_preserves_column_metadata() -> None:
-    """
-    Test that column metadata is preserved when query returns zero rows.
-
-    When a query returns no data but has a valid cursor description, the
-    column names and types from cursor_description should be preserved
-    in the result set. This allows downstream consumers (like the UI)
-    to display column headers even for empty result sets.
-    """
-    data: DbapiResult = []
-    description = [
-        ("id", "int", None, None, None, None, True),
-        ("name", "varchar", None, None, None, None, True),
-        ("created_at", "timestamp", None, None, None, None, True),
-    ]
-
-    result_set = SupersetResultSet(
-        data,
-        description,  # type: ignore
-        BaseEngineSpec,
-    )
-
-    # Verify column count
-    assert len(result_set.columns) == 3
-
-    # Verify column names are preserved
-    column_names = [col["column_name"] for col in result_set.columns]
-    assert column_names == ["id", "name", "created_at"]
-
-    assert result_set.columns[0]["type"] == BaseEngineSpec.get_datatype(
-        description[0][1]
-    )
-    assert result_set.columns[1]["type"] == BaseEngineSpec.get_datatype(
-        description[1][1]
-    )
-    assert result_set.columns[2]["type"] == BaseEngineSpec.get_datatype(
-        description[2][1]
-    )
-
-    # Verify the PyArrow table has the correct schema
-    assert result_set.table.num_rows == 0
-    assert len(result_set.table.column_names) == 3
-    assert list(result_set.table.column_names) == ["id", "name", "created_at"]
-
-    # Verify DataFrame conversion works
-    df = result_set.to_pandas_df()
-    assert len(df) == 0
-    assert list(map(str, df.columns)) == ["id", "name", "created_at"]