Limit max GitRef and DataImport name fields from Gitlab provider

arkindex/dataimport/providers.py

@@ -69,7 +69,8 @@ class GitProvider(ABC):
             ref.revision = revision
             ref.save()
         except GitRef.DoesNotExist:
-            repo.refs.create(name=name, type=type, revision=revision)
+            # Limit the name length to the max size of GitRef name field
+            repo.refs.create(name=name[:250], type=type, revision=revision)
 
     def get_or_create_revision(self, repo, sha, save=True):
         from arkindex.dataimport.models import Revision
@@ -111,7 +112,8 @@ class GitProvider(ABC):
         refs = ", ".join(rev.refs.values_list('name', flat=True))
         if not refs:
             refs = rev.hash
-        name = f"Import {refs} from {rev.repo.url}"
+        # Limit the name length to the max size of DataImport name field
+        name = f"Import {refs} from {rev.repo.url}"[:100]
 
         if rev.repo.type == RepositoryType.Worker:
             # Build revision workers in a new process

arkindex/dataimport/tests/test_gitlab_provider.py

@@ -583,7 +583,8 @@ class TestGitLabProvider(FixtureTestCase):
         # Add 2 commits on 2 branches
         self.gl_mock().projects.get.return_value.branches.list = lambda: [
             Ref("master", {"id": "commit1", "message": "A commit on master", "author_email": "someone@teklia.com"}),
-            Ref("killer-feature", {"id": "commit2", "message": "My fancy feature", "author_email": "another@teklia.com"})
+            # Create a commit with a very long branch name
+            Ref("A" * 1000, {"id": "commit2", "message": "My fancy feature", "author_email": "another@teklia.com"})
         ]
 
         # Add 1 new commit on a tag, and reuse a commit on another tag
@@ -623,7 +624,8 @@ class TestGitLabProvider(FixtureTestCase):
             ('v0.1', GitRefType.Tag),
         ])
         self.assertListEqual(list(commit2.refs.values_list('name', 'type').order_by('name')), [
-            ('killer-feature', GitRefType.Branch),
+            # Name has been restricted to 250 chars due to GitRef model constraint
+            ('A' * 250, GitRefType.Branch),
         ])
 
         # We should now have 3 dataimport for these refs
@@ -631,5 +633,6 @@ class TestGitLabProvider(FixtureTestCase):
         self.assertListEqual(list(DataImport.objects.values_list('name', 'mode', 'revision__hash').order_by('revision__hash')), [
             ("Import v0.1 from http://gitlab/repo", DataImportMode.Repository, 'commit0'),
             ("Import master from http://gitlab/repo", DataImportMode.Repository, 'commit1'),
-            ("Import killer-feature from http://gitlab/repo", DataImportMode.Repository, 'commit2'),
+            # Last import name has been limited to 100 chars due to DataImport model constraint
+            (f"Import {'A' * 93}", DataImportMode.Repository, 'commit2'),
         ])