Option to NOT check file existence?

[CrendKing]

Checklist

• I'm asking a question
• I've looked through the README and FAQ for similar questions
• I've searched the bugtracker for similar questions including closed ones

Question

Is there a way to tell youtube-dl NOT to check file existence and always tries to download? My scenario is this. I setup youtube-dl so that

• the output format is just C:/%(title)s.%(ext)s, which is convenient but prone to duplicate.
• uses aria2c as external downloader. For aria2c, I enable the "auto-file-renaming" feature, so that if file exists it automatically rename it.

To test, I create an empty file test.mp4 (imagine it as a completely unrelated video). Then invoke youtube-dl to download a video titled "test" (not Youtube). Here's output:

[debug] System config: []
[debug] User config: []
[debug] Custom config: ['--external-downloader', 'aria2c', '--external-downloader-args', '--conf-path=C:/aria2/aria2.conf', '--output', 'C:/%(title)s.%(ext)s', '-v']
[debug] Command-line args: ['--config-location', 'C:\\youtube-dl\\ydl.conf', 'https://www.test.com/test/']
[debug] Encodings: locale cp65001, fs utf-8, out utf-8, pref cp65001
[debug] youtube-dl version 2020.09.14
[debug] Python version 3.8.5 (CPython) - Windows-10-10.0.19041-SP0
[debug] exe versions: ffmpeg git-2020-08-31-4a11a6f, ffprobe git-2020-08-31-4a11a6f
[debug] Proxy map: {}
[generic] test-mp4: Requesting header
WARNING: Falling back on generic information extractor.
[generic] test-mp4: Downloading webpage
[generic] test-mp4: Extracting information
[download] Downloading playlist: test.mp4
[generic] playlist test.mp4: Collected 1 video ids (downloading 1 of them)
[download] Downloading video 1 of 1
[debug] Default format spec: bestvideo+bestaudio/best
[debug] Invoking downloader on 'https://www.test.com/video/test.mp4'
[download] C:\test.mp4 has already been downloaded
[download] 100% of 0.00B
[download] Finished downloading playlist: test.mp4

It seems youtube-dl checked the file existence by itself and never sent a request to aria2c. I see the "--no-overwrites" option, but I think what I need is the opposite of it. "--continue" also doesn't help. I know adding something like %(epoch)s to the template will break the duplicate, but I don't want random number in the filename.

Reading the code at https://github.com/ytdl-org/youtube-dl/blob/master/youtube_dl/downloader/common.py#L348, I don't see any way to change the behavior. Or am I missing anything? Thanks.

[pukkandan]

Doesn't --no-continue do this?

[CrendKing]

It's one step closer. However, youtube-dl always adds -c to the arguments, which defeats the purpose of --no-continue. If I remove -c locally, I get the exact behavior I need. So I suggest changing the beginning of Aria2cFD._make_cmd() to something like

        cmd = [self.exe]
        if self.params.get('continuedl', True):
            cmd += ['-c']

[reallyuniquename]

A bug with internal handler for aria2c then? It's better to rename this issue.

[CrendKing]

I created PR #26607 . I'll close this.