Gitlab is not immune to DMCA notices either since it's a US company. The RIAA probably won't try to take it down unless the youtube-dl development is actually relocated there or they become aware of it somehow. You'd have to host it in a country where DMCA can safely be ignored (anonymously of course), e.g. the Netherlands or Russia, if you want it to stay up reliably.
But to me as a long time Gitlab fan this shows the power of using Gitlab. Because if they were using Gitlab when the DMCA hits they could have easily migrated to their own Gitlab instance.
While Github would have to export things into another format. I'm sure people have worked on exporting data from Github to Gitlab but it seems much easier to just use Gitlab.
Edit: And of course besides data, you can't migrate contributors as easy. Which is also where Gitlab is powerful because it's open source and people are actually discussing federation in it. I doubt Github will ever get any kind of federation outside of microsoft services.
You're always a target of DMCA takedown requests, but having the freedom to host yourself anywhere with federation to other instances would give you the freedom to ignore it.
China has extremely strict anti-scraping laws [1] so pointing out that youtube-dl is a tool that can be used to illegally scrape copyrighted data is likely to have the same effect as a DMCA takedown request with the same justification.
It might be the case, but the repo you linked doesn't support that claim very well, and the cases cited are largely irrelevant to the case at hand.
"Forbidden area #1: providing scraping-related services to criminal organizations". Three cases listed. The first is one programmer's personal account of being arrested, which is very scant on why; the only info I can glean: "I developed some sort of ML API which is then used by a criminal enterprise against some influential company for god knows what purpose". Hard to draw any conclusion from that. The second case is ML-based CAPTCHA bypass for credential stuffing against Tencent QQ, emphasis on credential stuffing. The third case is some sort of black hat SEO campaign against Baidu, the scraping part (if any?) doesn't seem central to the conviction.
"Forbidden area #2: scraping and sale of personal info". Common sense, irrelevant.
"Forbidden area #3: commercial use of unlicensed business data" and the following untitled category list three cases, all of which are mass scraping operations either from a business competitor or that seriously affects site operations (through aggressive scraping).
AFAIK there are a lot of low hanging fruits in the Chinese piracy scene not yet targeted, and there are enough small-time commercial operations involving copyrighted media products begging to be taken down, it's highly unlikely anyone will bother to target some high-barrier-of-entry tool mostly facilitating the download of otherwise public videos.
#1 is relevant, because the developers of scraping services themselves were held liable for the use of their tools for illegal purposes. As you say, there are many small-time copyright infringers (a.k.a. criminal organizations), some of which probably use youtube-dl to circumvent anti-scraping measures (analogous to the CAPTCHA bypass case).
> it's highly unlikely anyone will bother to target some high-barrier-of-entry tool mostly facilitating the download of otherwise public videos
It's true that any particular violation is unlikely to be prosecuted because there are so many, but youtube-dl is already being targeted, so if it stays up on Gitee (or Gitlab, for that matter), then only because RIAA doesn't file a complaint.
I’d say a publicly disseminated free tool is pretty different from for-profit services with contracts. Other than the very vague case from which we can conclude nothing, all other cases are clearly for-profit, and the QQ credential stuffing case specifically established that the for-profit service provider was aware/suspicious of the criminal abuse as part of the conviction.
> only because RIAA doesn't file a complaint
Pretty sure RIAA/MPAA aren’t targeting China (yet), it’s still a bittorrenting Wild West according to my Chinese sources. They’ll need lawyers specialized in Chinese law and be prepared to navigate a foreign legal system if they want to look that way, frivolous DMCA takedowns that they file hundreds a day won’t do.
The power of a legal threat lies in the fact that there could be follow-up legal action; it’s toothless if no legal action could follow.
I don't think so. Youtube-DL can rip some popular Chinese VOD websites such as Bilibili and iQIYI, thus will first be "subjectable" to the Chinese equivalent of DMCA.
Worse, these companies are often backed/protected by powerful figures (often under the umbrella of some influential state members) and under such power, these kind of "threats" will be eradicated very efficiently, especially with Bilibili which is partially "state-monitored" (not sure if sponsored) due to the presence of the channel of the official youth branch of CCP is opened there [1].
Thus, relocating your repository to gangs aren't safe at all.
A rule of thumb: The prerequisite for your business to survive in China is that you need to have powerful affiliates, and make sure you do not piss off that powerful guy.
Hate to say about it but I'm ethically Chinese and I just happened to know some little thing about it...I know it's a great gambit so I hope I don't have to use the race card at the very least...
Then people straight out attacking me thinking I'm attacking China...
Were you born and raised in China? If not, this is the lamest reply ever. Using any "race card" is never okay, not only because it is just a perversion of appeal to authority, but also because your ethnicity does not impart any special secret knowledge about the universe that a few supporting links can't also reveal. Appealing to your ethnicity is not a proxy for being a member of some culture.
Does it includes issues and open PRs? I think this is the most concerning part about Github: when you move somewhere else you lose important parts of the history of the project (that document known bugs, how features came to be, etc), unless you specifically step up and use some import tool.
There are Git-backed distributed issue trackers like Git-bug https://github.com/MichaelMure/git-bug (and probably other tools) that should be more used.
One could perhaps convert Github issues to git-bug and store on a branch of this IPFS Git repository.
While this situation is a reason to look more closely at GitLab due to the ability to host your own instance, why not also look at a solution like Fossile: DVCS, wiki, and issue tracker all in a cross-platform, single executable file?
Anyone else here think Downloading an executable as root from a URL that follows redirects is a bad idea? Seems like they could break the (Unix) install down into a few more steps that take some caution
Curious in general how often Youtube breaks things on their end, i.e. how long roughly this specific version of youtube-dl is expected to keep working without any active development?
I use youtube-dl fairly infrequently... maybe once a month or so. Nearly every time I used it it would not download the video I wanted - so I'd do a "youtube-dl -u", then it would work.
Just spreading the repo around isn't going to help for very long.
I with there was an issue tracking thing that you could clone as easily as a git repo. My first thought was email as an ongoing distributed solution, but as far as I know you can't just download an archive of issue tickets, bug reports etc into a maildir without a fair amount of faff?
Compare the sha commit hashes of the top commit against the hashes from the old "true" repo. If they match the repo (and the history) has not been changed. Subsequent commits can be manually audited.