Proposal: A more robust solution for truncating file names

[yan12125]

Currently in this line, info_dict['title'] is truncated to 200 characters. This is first done by @phihag in 0599ef8. This implementation can't handle wide characters correctly. An examples are #5740. I think YoutubeDL should switch to a utf-8 friendly implementation.

There's an implementation at a stackoverflow answer, although I haven't tested it. Any opinions?

[dstftw]

As an addition: it still may not be enough since max filename limitation in general can be arbitrary (see #5908).

[yan12125]

Another side note: on Windows file system encoding is not always utf-8, which make thinks more complicated. For example, on Traditional Chinese Windows file system uses big5 (cp950), where each Chinese character occupies 2 bytes rather than 3 bytes in utf-8.

[Invertex]

This is still an issue, at least in Windows, and can be further exacerbated depending on the character length of the directory path that youtube-dl is outputting files in. --restrict-filenames doesn't trim it and there's no option to truncate the filename either still. A simple -truncate ## parameter would suffice just fine, or perhaps even an auto value that checks total file path length.

[shillshocked]

Also an issue in Linux, and moderator keeps closing new issues on this without referencing or addressing the original issue.

[wizzup]

I've got the same error message, add here for reference and seaching

#23702