Avoid exception when trying to load invalid text files

.github/workflows/tox.yml

@@ -59,7 +59,7 @@ jobs:
       WSLENV: FORCE_COLOR:PYTEST_REQPASS:TOXENV:GITHUB_STEP_SUMMARY
       # Number of expected test passes, safety measure for accidental skip of
       # tests. Update value if you add/remove tests.
-      PYTEST_REQPASS: 794
+      PYTEST_REQPASS: 795
     steps:
       - name: Activate WSL1
         if: "contains(matrix.shell, 'wsl')"

.pre-commit-config.yaml

@@ -17,6 +17,7 @@ exclude: >
     .config/requirements.*|
     .vscode/extensions.json|
     .vscode/settings.json|
+    examples/broken/encoding.yml|
     examples/broken/encoding.j2|
     examples/broken/yaml-with-tabs/invalid-due-tabs.yaml|
     examples/playbooks/collections/.*|

src/ansiblelint/file_utils.py

@@ -191,7 +191,7 @@ def __init__(
         self.dir: str = ""
         self.kind: FileType | None = None
         self.stop_processing = False  # Set to stop other rules from running
-        self._data: Any = States.NOT_LOADED
+        self.state: Any = States.NOT_LOADED
         self.line_skips: dict[int, set[str]] = defaultdict(set)
         self.exc: Exception | None = None  # Stores data loading exceptions
 
@@ -359,17 +359,17 @@ def __repr__(self) -> str:
     @property
     def data(self) -> Any:
         """Return loaded data representation for current file, if possible."""
-        if self._data == States.NOT_LOADED:
+        if self.state == States.NOT_LOADED:
             if self.path.is_dir():
-                self._data = None
-                return self._data
+                self.state = None
+                return self.state
             try:
                 if str(self.base_kind) == "text/yaml":
                     from ansiblelint.utils import (  # pylint: disable=import-outside-toplevel
                         parse_yaml_linenumbers,
                     )
 
-                    self._data = parse_yaml_linenumbers(self)
+                    self.state = parse_yaml_linenumbers(self)
                     # now that _data is not empty, we can try guessing if playbook or rulebook
                     # it has to be done before append_skipped_rules() call as it's relying
                     # on self.kind.
@@ -380,19 +380,24 @@ def data(self) -> Any:
                         # pylint: disable=import-outside-toplevel
                         from ansiblelint.skip_utils import append_skipped_rules
 
-                    self._data = append_skipped_rules(self._data, self)
+                    self.state = append_skipped_rules(self.state, self)
                 else:
                     logging.debug(
                         "data set to None for %s due to being of %s kind.",
                         self.path,
                         self.base_kind or "unknown",
                     )
-                    self._data = States.UNKNOWN_DATA
-
-            except (RuntimeError, FileNotFoundError, YAMLError) as exc:
-                self._data = States.LOAD_FAILED
+                    self.state = States.UNKNOWN_DATA
+
+            except (
+                RuntimeError,
+                FileNotFoundError,
+                YAMLError,
+                UnicodeDecodeError,
+            ) as exc:
+                self.state = States.LOAD_FAILED
                 self.exc = exc
-        return self._data
+        return self.state
 
 
 # pylint: disable=redefined-outer-name

src/ansiblelint/runner.py

@@ -124,12 +124,14 @@ def run(self) -> list[MatchError]:  # noqa: C901
                 self.lintables.remove(lintable)
                 continue
             if isinstance(lintable.data, States) and lintable.exc:
+                lintable.exc.__class__.__name__.lower()
                 matches.append(
                     MatchError(
                         lintable=lintable,
                         message=str(lintable.exc),
                         details=str(lintable.exc.__cause__),
                         rule=LoadingFailureRule(),
+                        tag=f"load-failure[{lintable.exc.__class__.__name__.lower()}]",
                     ),
                 )
                 lintable.stop_processing = True

test/test_load_failure.py

@@ -1,11 +1,23 @@
 """Tests for LoadFailureRule."""
+import pytest
+
 from ansiblelint.rules import RulesCollection
 from ansiblelint.runner import Runner
 
 
-def test_load_failure_encoding(default_rules_collection: RulesCollection) -> None:
+@pytest.mark.parametrize(
+    "path",
+    (
+        pytest.param("examples/broken/encoding.j2", id="jinja2"),
+        pytest.param("examples/broken/encoding.yml", id="yaml"),
+    ),
+)
+def test_load_failure_encoding(
+    path: str,
+    default_rules_collection: RulesCollection,
+) -> None:
     """Check that we fail when file encoding is wrong."""
-    runner = Runner("examples/broken/encoding.j2", rules=default_rules_collection)
+    runner = Runner(path, rules=default_rules_collection)
     matches = runner.run()
     assert len(matches) == 1, matches
     assert matches[0].rule.id == "load-failure"