Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upGitHub is where the world builds software
Millions of developers and companies build, ship, and maintain their software on GitHub — the largest and most advanced development platform in the world.
[cbs] Colbert Ep. 107 yields 404 failures on some m3u8 segments #12416
Comments
|
By the way, here's how it ends (when run without
That is, it results in a 16-minute video with a lot of content omitted. By contrast, the |
|
This is happening to me as well. |
|
So, if we assume for the moment that this is a mistake on the content-provider's side, rather than a deliberate choice to frustrate youtube-dl and other HLS clients, what should youtube-dl do?
|
|
The latest episode (the one after this one) downloads fine by the way, so this doesn't seem to be something they did to discourage downloads. |
|
The default Although I don't know if it's worth tracking the #12416 (comment) thoughts on how to handle this condition somehow... |
Attempting to download the Thursday March 9 episode of The Colbert Report yields a variety of errors on most but not all m3u8 segments. Perhaps this is a one-off issue with this particular recent episode; the prior episode 106 downloads just fine. Also, testing with an non-HLS format (rtmp-1928) works fine.
ffmpeg reports the failures like this, over and over for each segment:
(At first I thought the presence of "crypto" in the string was suggesting a cryptography issue, but on the basis of the
404 Not FoundI suspect that's a misinterpretation.)Note also that the partial segment success means you can open the downloaded file and have itshow a clip from a few minutes into the show. I'm not sure if perhaps there would have been a better diagnostic had I let it run to completion.
Here's the format list:
Both the
hls-2642andhls-3653fail in similar ways. Problem occurs without specifying--hls-use-mpegtsas well as with it.Finally, here's the copious output of the failure. I don't know if it's long enough that it should go into a gist instead. Sorry, if so.