Twitch fails when --download-archive present.

[Hajitorus]

I have a --download-archive clause present in ~/.config/youtube-dl.conf, and use --download-archive /dev/null to circumvent this when needed. However, twitch URLs appear to not produce an ID that the archiving system can use, producing this error:

% youtube-dl --download-archive /dev/null http://www.twitch.tv/therealtakeshi/b/516100842
[justin.tv] 516100842: Extracting information
[justin.tv] 516100842: Downloading video info JSON
Traceback (most recent call last):
  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/runpy.py", line 162, in _run_module_as_main
    "__main__", fname, loader, pkg_name)
  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/runpy.py", line 72, in _run_code
    exec code in run_globals
  File "/usr/local/bin/youtube-dl/__main__.py", line 18, in <module>
  File "/usr/local/bin/youtube-dl/youtube_dl/__init__.py", line 836, in main
  File "/usr/local/bin/youtube-dl/youtube_dl/__init__.py", line 826, in _real_main
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 1033, in download
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 522, in extract_info
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 656, in process_ie_result
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 559, in process_ie_result
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 814, in process_video_result
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 839, in process_info
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 479, in _match_entry
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 1099, in in_download_archive
  File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 1092, in _make_archive_id
TypeError: coercing to Unicode: need string or buffer, int found

Moving my youtube-dl.conf aside and retrying (without any --download-archive option) works fine. I pulled down a local copy and was able to fix this bug (for me) by coercing info_dict[id] to string:

diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py
index 5794fdb..c8e1482 100644
--- a/youtube_dl/YoutubeDL.py
+++ b/youtube_dl/YoutubeDL.py
@@ -1089,7 +1089,7 @@ class YoutubeDL(object):
                 extractor = info_dict.get('ie_key')  # key in a playlist
         if extractor is None:
             return None  # Incomplete video information
-        return extractor.lower() + ' ' + info_dict['id']
+        return extractor.lower() + ' ' + str(info_dict['id'])

     def in_download_archive(self, info_dict):
         fn = self.params.get('download_archive')

I'm not sure if this is the right fix or if extractors should be forced to ensure id is a string?

[Hajitorus]

… on second thought, justintv might need to return something more unique to work with download archive, probably at least <username>/b/<id> or whatever is appropriate based on the input URI?

[phihag]

My understanding is that the numeric Id is already unique (it's certainly large enough). The justin.tv extractor is at fault here, not the download_archive code, since video IDs are supposed to. Fixed, thank you very much for the report!