Even if for now, they have a thing set up for redirects to whatever the new URL is, that's still a lot of images to comb through before that redirect dies.

I'd imagine changing or redirecting the URLs via a script would be pretty easy.

peacethroughpower said:
I'd imagine changing or redirecting the URLs via a script would be pretty easy.

I reckon Mairo would not be okay with that. It would be fine with an archive.org link, but not a nitter instance link.

I think an admin could mass edit them. Maybe that would be intensive for the site if it edits hundreds of thousands at once, but it's not something that needs to be manually combed. Until then just keep them the same and wait to see what happens.

This is why a single point of failure sucks. Redundancy is crucial.

As long as the IDs remain the same, it would probably be trivial enough to run a fixer script on the database itself. I believe similar things have been done in the past, for stuff like changing http://furaffinity.net to https://furaffinity.net.

I imagine the bigger problem is going to be when Elon Musk finally runs Twitter so hard into the ground, the site just stops existing and all of the source links are irreparably broken.

faucet said:
I imagine the bigger problem is going to be when Elon Musk finally runs Twitter so hard into the ground, the site just stops existing and all of the source links are irreparably broken.

It's not a huge problem if we already have the best quality version that Twitter can provide, and any description that accompanied it. You lose some context surrounding the tweet that might be useful to somebody once in a blue moon.

You find a new link or you go back and link it again.
It's hard work but some neet has gotta do it.

wolfmanfur said:
I reckon Mairo would not be okay with that. It would be fine with an archive.org link, but not a nitter instance link.

You can just use a nitter extension that replaces the URL, anyways. I try to use archive.today or archive.org when I can on stuff like ImgDur and Fumblr but there's just so damn many.

Yeah, archive.org is trivial to redirect. Mostly automatic flow, as well. If there isn't an archive and someone clicks on it, it'll ask if you want to save it.

lance_armstrong said:
I think an admin could mass edit them. Maybe that would be intensive for the site if it edits hundreds of thousands at once, but it's not something that needs to be manually combed. Until then just keep them the same and wait to see what happens.

OK

closetpossum said:
You find a new link or you go back and link it again.
It's hard work but some neet has gotta do it.

There might not be any data to link _to_, left on the Internet. All eggs in one basket. :(

I imagine Twitter and Archive.org are dealing with a lot of people trying to keep links from going dead.

Remember when Tumblr banned porn?

That sucked.

alphamule said:

There might not be any data to link _to_, left on the Internet. All eggs in one basket. :(

I imagine Twitter and Archive.org are dealing with a lot of people trying to keep links from going dead.

then we paheal/r34 rule that shit. Link? what link? e621 is your link now!
those sites been doing that for decades. We just have more integrity.

closetpossum said:
then we paheal/r34 rule that shit. Link? what link? e621 is your link now!
those sites been doing that for decades. We just have more integrity.

Part of why MD5 became the standard... Annoying that Google won't let you search images with one. So we have to use SauceNAO!

And even Saucenao doesn't hit everywhere

For some reason It doesn't find IB unless we or gelbooru or such also has the link

Saucenao only crawls a few select sites. Any inkbunny links it has, it only has because it got them from one of those sites.

Oh fuck, it's happening already. My twitter is already showing it as X instead of the old logo like how it did with doge for like april fools...

Delusional optimism: it stays as just a logo change.

Yes but so far "x.com" redirects to twitter.com, not the reverse

All the links work perfectly fine.
Optimistically there won't be a need to mass update the sources

benjiboyo said:
Oh fuck, it's happening already. My twitter is already showing it as X instead of the old logo like how it did with doge for like april fools...

Delusional optimism: it stays as just a logo change.

LOL, given what 'X' means, that's hilarious. He's ex-ing the old logo. I guess a lot of users X'd their accounts. But X is all the things, in algebra.

alphamule said:
LOL, given what 'X' means, that's hilarious. He's ex-ing the old logo. I guess a lot of users X'd their accounts. But X is all the things, in algebra.

No wonder he banned Xanderhal. Musk is mad Xan is cooler and uses X.

kora_viridian said:
I think there's at least 56,000 posts with a source at Twitter, and probably at least a few times that...

I checked on a local copy of the post data and got 1011248 posts that contain the text "twitter" in the source string, with an additional 82k deleted posts. It would likely be better to just update it offline.

post #4470932 https://nitter.uni-sonia.com/pic/orig/media/FYnC00WacAE8v89.jpg

alphamule said:
Hmm, nitter.net probably shouldn't be a primary source (but seems to make sense for direct link, especially if archiving, because Twitter URLs are a pain).

https://pbs.twimg.com/media/FYnC00WacAE8v89?format=png&name=orig doesn't work, but https://pbs.twimg.com/media/FYnC00WacAE8v89?format=jpg&name=orig does. If someone uses more advanced tools to replace those links, they might want to make sure that archived_source remains true, or simply not do it for them if archived. i.e. use source:*nitter* source:*media* -archived_source

Oops, turns out that those can be turned into usable links at the original site's domains. For non-archived links, probably a good idea to add those with original source archived?