X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/4a8883858028d4068b52710acd9c349108b6fc7c..55063ad6fc06f74392338802dd70cd315abf3a1a:/sdk/python/tests/test_arv_put.py

diff --git a/sdk/python/tests/test_arv_put.py b/sdk/python/tests/test_arv_put.py
index d8c1725e29..c2eaf12bbb 100644
--- a/sdk/python/tests/test_arv_put.py
+++ b/sdk/python/tests/test_arv_put.py
@@ -19,6 +19,7 @@ import yaml
 import threading
 import hashlib
 import random
+import uuid
 
 import arvados
 import arvados.commands.put as arv_put
@@ -268,12 +269,39 @@ class ArvPutUploadJobTest(run_test_server.TestCaseWithServers,
             with open(os.path.join(self.small_files_dir, str(i)), 'w') as f:
                 f.write(data + str(i))
         self.arvfile_write = getattr(arvados.arvfile.ArvadosFileWriter, 'write')
+        # Temp dir to hold a symlink to other temp dir
+        self.tempdir_with_symlink = tempfile.mkdtemp()
+        os.symlink(self.tempdir, os.path.join(self.tempdir_with_symlink, 'linkeddir'))
+        os.symlink(os.path.join(self.tempdir, '1'),
+                   os.path.join(self.tempdir_with_symlink, 'linkedfile'))
 
     def tearDown(self):
         super(ArvPutUploadJobTest, self).tearDown()
         shutil.rmtree(self.tempdir)
         os.unlink(self.large_file_name)
         shutil.rmtree(self.small_files_dir)
+        shutil.rmtree(self.tempdir_with_symlink)
+
+    def test_symlinks_are_followed_by_default(self):
+        cwriter = arv_put.ArvPutUploadJob([self.tempdir_with_symlink])
+        cwriter.start(save_collection=False)
+        self.assertIn('linkeddir', cwriter.manifest_text())
+        self.assertIn('linkedfile', cwriter.manifest_text())
+        cwriter.destroy_cache()
+
+    def test_symlinks_are_not_followed_when_requested(self):
+        cwriter = arv_put.ArvPutUploadJob([self.tempdir_with_symlink],
+                                          follow_links=False)
+        cwriter.start(save_collection=False)
+        self.assertNotIn('linkeddir', cwriter.manifest_text())
+        self.assertNotIn('linkedfile', cwriter.manifest_text())
+        cwriter.destroy_cache()
+
+    def test_passing_nonexistant_path_raise_exception(self):
+        uuid_str = str(uuid.uuid4())
+        cwriter = arv_put.ArvPutUploadJob(["/this/path/does/not/exist/{}".format(uuid_str)])
+        with self.assertRaises(arv_put.PathDoesNotExistError):
+            cwriter.start(save_collection=False)
 
     def test_writer_works_without_cache(self):
         cwriter = arv_put.ArvPutUploadJob(['/dev/null'], resume=False)
@@ -286,12 +314,14 @@ class ArvPutUploadJobTest(run_test_server.TestCaseWithServers,
             f.flush()
             cwriter = arv_put.ArvPutUploadJob([f.name])
             cwriter.start(save_collection=False)
-            self.assertEqual(3, cwriter.bytes_written - cwriter.bytes_skipped)
+            self.assertEqual(0, cwriter.bytes_skipped)
+            self.assertEqual(3, cwriter.bytes_written)
             # Don't destroy the cache, and start another upload
             cwriter_new = arv_put.ArvPutUploadJob([f.name])
             cwriter_new.start(save_collection=False)
             cwriter_new.destroy_cache()
-            self.assertEqual(0, cwriter_new.bytes_written - cwriter_new.bytes_skipped)
+            self.assertEqual(3, cwriter_new.bytes_skipped)
+            self.assertEqual(3, cwriter_new.bytes_written)
 
     def make_progress_tester(self):
         progression = []
@@ -761,7 +791,8 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
         datadir = self.make_tmpdir()
         with open(os.path.join(datadir, "foo"), "w") as f:
             f.write("The quick brown fox jumped over the lazy dog")
-        p = subprocess.Popen([sys.executable, arv_put.__file__, datadir],
+        p = subprocess.Popen([sys.executable, arv_put.__file__,
+                              os.path.join(datadir, 'foo')],
                              stdout=subprocess.PIPE,
                              stderr=subprocess.PIPE,
                              env=self.ENVIRON)
@@ -808,7 +839,40 @@ class ArvPutIntegrationTest(run_test_server.TestCaseWithServers,
         self.assertEqual(col['uuid'], updated_col['uuid'])
         # Get the manifest and check that the new file is being included
         c = arv_put.api_client.collections().get(uuid=updated_col['uuid']).execute()
-        self.assertRegex(c['manifest_text'], r'^\. .*:44:file2\n')
+        self.assertRegex(c['manifest_text'], r'^\..* .*:44:file2\n')
+
+    def test_upload_directory_reference_without_trailing_slash(self):
+        tmpdir1 = self.make_tmpdir()
+        tmpdir2 = self.make_tmpdir()
+        with open(os.path.join(tmpdir1, 'foo'), 'w') as f:
+            f.write('This is foo')
+        with open(os.path.join(tmpdir2, 'bar'), 'w') as f:
+            f.write('This is not foo')
+        # Upload one directory and one file
+        col = self.run_and_find_collection("", ['--no-progress',
+                                                tmpdir1,
+                                                os.path.join(tmpdir2, 'bar')])
+        self.assertNotEqual(None, col['uuid'])
+        c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
+        # Check that 'foo' was written inside a subcollection
+        # OTOH, 'bar' should have been directly uploaded on the root collection
+        self.assertRegex(c['manifest_text'], r'^\. .*:15:bar\n\./.+ .*:11:foo\n')
+
+    def test_upload_directory_reference_with_trailing_slash(self):
+        tmpdir1 = self.make_tmpdir()
+        tmpdir2 = self.make_tmpdir()
+        with open(os.path.join(tmpdir1, 'foo'), 'w') as f:
+            f.write('This is foo')
+        with open(os.path.join(tmpdir2, 'bar'), 'w') as f:
+            f.write('This is not foo')
+        # Upload one directory (with trailing slash) and one file
+        col = self.run_and_find_collection("", ['--no-progress',
+                                                tmpdir1 + os.sep,
+                                                os.path.join(tmpdir2, 'bar')])
+        self.assertNotEqual(None, col['uuid'])
+        c = arv_put.api_client.collections().get(uuid=col['uuid']).execute()
+        # Check that 'foo' and 'bar' were written at the same level
+        self.assertRegex(c['manifest_text'], r'^\. .*:15:bar .*:11:foo\n')
 
     def test_put_collection_with_high_redundancy(self):
         # Write empty data: we're not testing CollectionWriter, just