Improve validation for import session (#12492)

- Ensure that validated data is a dict shape

src/backend/InvenTree/importer/serializers.py

@@ -98,7 +98,10 @@ class DataImportSessionSerializer(InvenTreeModelSerializer):
         if type(defaults) is not dict:
             try:
                 defaults = json.loads(str(defaults))
-            except:
+            except json.JSONDecodeError:
+                raise ValidationError(_('Invalid field defaults'))
+
+            if type(defaults) is not dict:
                 raise ValidationError(_('Invalid field defaults'))
 
         return defaults
@@ -111,7 +114,10 @@ class DataImportSessionSerializer(InvenTreeModelSerializer):
         if type(overrides) is not dict:
             try:
                 overrides = json.loads(str(overrides))
-            except:
+            except json.JSONDecodeError:
+                raise ValidationError(_('Invalid field overrides'))
+
+            if type(overrides) is not dict:
                 raise ValidationError(_('Invalid field overrides'))
 
         return overrides
@@ -124,7 +130,10 @@ class DataImportSessionSerializer(InvenTreeModelSerializer):
         if type(filters) is not dict:
             try:
                 filters = json.loads(str(filters))
-            except:
+            except json.JSONDecodeError:
+                raise ValidationError(_('Invalid field filters'))
+
+            if type(filters) is not dict:
                 raise ValidationError(_('Invalid field filters'))
 
         return filters

src/backend/InvenTree/importer/tests.py

@@ -152,7 +152,47 @@ class ImporterTest(ImporterMixin, InvenTreeTestCase):
         self.assertEqual(n + 12, Company.objects.count())
 
     def test_field_defaults(self):
-        """Test default field values."""
+        """Test default field values.
+
+        Regression test for bug #30 (dev/todo/bugs.md): validate_field_defaults
+        must reject any parsed JSON value that isn't actually a dict, not just
+        values which fail to parse as JSON at all.
+        """
+        from django.core.exceptions import ValidationError as DjangoValidationError
+
+        def data_file():
+            # full_clean() reads the file eagerly, so each session needs its own
+            # freshly-opened (binary) copy
+            content = self.helper_file('companies.csv').read()
+            return ContentFile(content.encode(), 'companies.csv')
+
+        # A dict value is accepted as-is
+        session = DataImportSession(
+            data_file=data_file(), model_type='company', field_defaults={'name': 'Test'}
+        )
+        session.full_clean()
+
+        # A JSON-encoded dict string is also accepted (de-stringified)
+        session = DataImportSession(
+            data_file=data_file(),
+            model_type='company',
+            field_defaults='{"name": "Test"}',
+        )
+        session.full_clean()
+
+        # A JSON-encoded list is valid JSON, but not a dict - must be rejected
+        session = DataImportSession(
+            data_file=data_file(), model_type='company', field_defaults='[1, 2, 3]'
+        )
+        with self.assertRaises(DjangoValidationError):
+            session.full_clean()
+
+        # Garbage which does not even parse as JSON must also be rejected
+        session = DataImportSession(
+            data_file=data_file(), model_type='company', field_defaults='not valid json'
+        )
+        with self.assertRaises(DjangoValidationError):
+            session.full_clean()
 
     def test_lookup_field_ambiguous_match(self):
         """Test the behavior of lookup_related_field for ambiguous and pinned matches."""

src/backend/InvenTree/importer/validators.py

@@ -6,7 +6,7 @@ from django.core.exceptions import ValidationError
 from django.utils.translation import gettext_lazy as _
 
 # Define maximum limits for imported file data
-IMPORTER_MAX_FILE_SIZE = 32 * 1024 * 1042
+IMPORTER_MAX_FILE_SIZE = 32 * 1024 * 1024
 IMPORTER_MAX_ROWS = 5000
 IMPORTER_MAX_COLS = 1000
 
@@ -46,8 +46,11 @@ def validate_field_defaults(value):
         return
 
     if type(value) is not dict:
-        # OK if we can parse it as JSON
+        # OK if we can parse it as JSON - but the result must be a dict
         try:
             value = json.loads(value)
-        except json.JSONDecodeError:
+        except (TypeError, json.JSONDecodeError):
+            raise ValidationError(_('Value must be a valid dictionary object'))
+
+        if type(value) is not dict:
             raise ValidationError(_('Value must be a valid dictionary object'))