-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
from __future__ import absolute_import
from __future__ import division
from builtins import str
from builtins import range
import apiclient
-import io
import mock
import os
import pwd
import threading
import hashlib
import random
-
-if sys.version_info >= (3, 0):
- from io import StringIO
-else:
- from cStringIO import StringIO
+import uuid
import arvados
import arvados.commands.put as arv_put
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)
def test_writer_works_with_cache(self):
with tempfile.NamedTemporaryFile() as f:
- f.write('foo')
+ f.write(b'foo')
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 = []
def test_progress_reporting(self):
with tempfile.NamedTemporaryFile() as f:
- f.write('foo')
+ f.write(b'foo')
f.flush()
for expect_count in (None, 8):
progression, reporter = self.make_progress_tester()
arv_put.human_progress(count, None)))
-class ArvadosPutTest(run_test_server.TestCaseWithServers, ArvadosBaseTestCase):
+class ArvadosPutTest(run_test_server.TestCaseWithServers,
+ ArvadosBaseTestCase,
+ tutil.VersionChecker):
MAIN_SERVER = {}
Z_UUID = 'zzzzz-zzzzz-zzzzzzzzzzzzzzz'
def call_main_with_args(self, args):
- self.main_stdout = StringIO()
- self.main_stderr = StringIO()
+ self.main_stdout = tutil.StringIO()
+ self.main_stderr = tutil.StringIO()
return arv_put.main(args, self.main_stdout, self.main_stderr)
def call_main_on_test_file(self, args=[]):
super(ArvadosPutTest, self).tearDown()
def test_version_argument(self):
- err = io.BytesIO()
- out = io.BytesIO()
- with tutil.redirected_streams(stdout=out, stderr=err):
+ with tutil.redirected_streams(
+ stdout=tutil.StringIO, stderr=tutil.StringIO) as (out, err):
with self.assertRaises(SystemExit):
self.call_main_with_args(['--version'])
- self.assertEqual(out.getvalue(), '')
- self.assertRegexpMatches(err.getvalue(), "[0-9]+\.[0-9]+\.[0-9]+")
+ self.assertVersionOutput(out, err)
def test_simple_file_put(self):
self.call_main_on_test_file()
def test_api_error_handling(self):
coll_save_mock = mock.Mock(name='arv.collection.Collection().save_new()')
coll_save_mock.side_effect = arvados.errors.ApiError(
- fake_httplib2_response(403), '{}')
+ fake_httplib2_response(403), b'{}')
with mock.patch('arvados.collection.Collection.save_new',
new=coll_save_mock):
with self.assertRaises(SystemExit) as exc_test:
result = arv_put.desired_project_uuid(arv_put.api_client, BAD_UUID,
0)
except ValueError as error:
- self.assertIn(BAD_UUID, error.message)
+ self.assertIn(BAD_UUID, str(error))
else:
self.assertFalse(result, "incorrectly found nonexistent project")
[sys.executable, arv_put.__file__, '--stream'],
stdin=subprocess.PIPE, stdout=subprocess.PIPE,
stderr=subprocess.STDOUT, env=self.ENVIRON)
- pipe.stdin.write('stdin test\n')
+ pipe.stdin.write(b'stdin test\n')
pipe.stdin.close()
deadline = time.time() + 5
while (pipe.poll() is None) and (time.time() < deadline):
elif returncode != 0:
sys.stdout.write(pipe.stdout.read())
self.fail("arv-put returned exit code {}".format(returncode))
- self.assertIn('4a9c8b735dce4b5fa3acf221a0b13628+11', pipe.stdout.read())
+ self.assertIn('4a9c8b735dce4b5fa3acf221a0b13628+11',
+ pipe.stdout.read().decode())
def test_ArvPutSignedManifest(self):
# ArvPutSignedManifest runs "arv-put foo" and then attempts to get
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],
- stdout=subprocess.PIPE, env=self.ENVIRON)
- (arvout, arverr) = p.communicate()
- self.assertEqual(arverr, None)
+ p = subprocess.Popen([sys.executable, arv_put.__file__,
+ os.path.join(datadir, 'foo')],
+ stdout=subprocess.PIPE,
+ stderr=subprocess.PIPE,
+ env=self.ENVIRON)
+ (out, err) = p.communicate()
+ self.assertRegex(err.decode(), r'INFO: Collection saved as ')
self.assertEqual(p.returncode, 0)
# The manifest text stored in the API server under the same
# manifest UUID must use signed locators.
c = arv_put.api_client.collections().get(uuid=manifest_uuid).execute()
- self.assertRegexpMatches(
+ self.assertRegex(
c['manifest_text'],
r'^\. 08a008a01d498c404b0c30852b39d3b8\+44\+A[0-9a-f]+@[0-9a-f]+ 0:44:foo\n')
[sys.executable, arv_put.__file__] + extra_args,
stdin=subprocess.PIPE, stdout=subprocess.PIPE,
stderr=subprocess.PIPE, env=self.ENVIRON)
- stdout, stderr = pipe.communicate(text)
+ stdout, stderr = pipe.communicate(text.encode())
+ self.assertRegex(stderr.decode(), r'INFO: Collection (updated:|saved as)')
search_key = ('portable_data_hash'
if '--portable-data-hash' in extra_args else 'uuid')
collection_list = arvados.api('v1').collections().list(
- filters=[[search_key, '=', stdout.strip()]]).execute().get('items', [])
+ filters=[[search_key, '=', stdout.decode().strip()]]
+ ).execute().get('items', [])
self.assertEqual(1, len(collection_list))
return collection_list[0]
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.assertRegexpMatches(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
"Test unnamed collection",
['--portable-data-hash', '--project-uuid', self.PROJECT_UUID])
username = pwd.getpwuid(os.getuid()).pw_name
- self.assertRegexpMatches(
+ self.assertRegex(
link['name'],
r'^Saved at .* by {}@'.format(re.escape(username)))