aboutsummaryrefslogtreecommitdiff
path: root/test/test_download.py
diff options
context:
space:
mode:
authorFilippo Valsorda <filippo.valsorda@gmail.com>2013-06-17 19:28:18 +0200
committerFilippo Valsorda <filippo.valsorda@gmail.com>2013-06-17 19:28:18 +0200
commit8bc7c3d858c9e2dcf6cf65a8306f3ed2bc9109f5 (patch)
treea057ff983256938a9a9497b3f5d120e31d91d3ca /test/test_download.py
parent36ed7177f01d278935ab5eac3d44fcea421c5325 (diff)
parentaf44c9486255f16ab180a9e45aaab06a6b38bdde (diff)
Merge branch 'search_regex' - PR #872 - closes #847
Diffstat (limited to 'test/test_download.py')
-rw-r--r--test/test_download.py23
1 files changed, 21 insertions, 2 deletions
diff --git a/test/test_download.py b/test/test_download.py
index 3eca333f2..577bcdbf2 100644
--- a/test/test_download.py
+++ b/test/test_download.py
@@ -7,8 +7,8 @@ import os
import json
import unittest
import sys
-import hashlib
import socket
+import binascii
# Allow direct execution
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
@@ -38,11 +38,16 @@ def _try_rm(filename):
if ose.errno != errno.ENOENT:
raise
+md5 = lambda s: hashlib.md5(s.encode('utf-8')).hexdigest()
+
class FileDownloader(youtube_dl.FileDownloader):
def __init__(self, *args, **kwargs):
self.to_stderr = self.to_screen
self.processed_info_dicts = []
return youtube_dl.FileDownloader.__init__(self, *args, **kwargs)
+ def report_warning(self, message):
+ # Don't accept warnings during tests
+ raise ExtractorError(message)
def process_info(self, info_dict):
self.processed_info_dicts.append(info_dict)
return youtube_dl.FileDownloader.process_info(self, info_dict)
@@ -121,7 +126,21 @@ def generator(test_case):
with io.open(tc['file'] + '.info.json', encoding='utf-8') as infof:
info_dict = json.load(infof)
for (info_field, value) in tc.get('info_dict', {}).items():
- self.assertEqual(value, info_dict.get(info_field))
+ if isinstance(value, compat_str) and value.startswith('md5:'):
+ self.assertEqual(value, 'md5:' + md5(info_dict.get(info_field)))
+ else:
+ self.assertEqual(value, info_dict.get(info_field))
+
+ # If checkable fields are missing from the test case, print the info_dict
+ test_info_dict = dict((key, value if not isinstance(value, compat_str) or len(value) < 250 else 'md5:' + md5(value))
+ for key, value in info_dict.items()
+ if value and key in ('title', 'description', 'uploader', 'upload_date', 'uploader_id', 'location'))
+ if not all(key in tc.get('info_dict', {}).keys() for key in test_info_dict.keys()):
+ sys.stderr.write(u'\n"info_dict": ' + json.dumps(test_info_dict, ensure_ascii=False, indent=2) + u'\n')
+
+ # Check for the presence of mandatory fields
+ for key in ('id', 'url', 'title', 'ext'):
+ self.assertTrue(key in info_dict.keys() and info_dict[key])
finally:
for tc in test_cases:
_try_rm(tc['file'])