An example that the old extractor does not work properly:
Fix 2 bugs in youtube_dl/extractor/sohu.py
1. Highest quality is the first one in ('ori', 'super', 'high', 'nor')
Their chinese names are 原画, 超清, 高清, 标清, respectively
2. Sohu's server can't identify duplicate slashes in the url.
Seems Sohu has fixed the bug related to duplicate slashes, but a more concise URL should lead to less bugs.
[Sohu] Update the MD5 checksum of the test case
[Sohu] Fix a test failure due to md5 errors
Only the first 10240 bytes are calculated in tests