Mikaeri said:

Just... hold off for a moment replacing them if you can. An approver will get to yours in due time. In the meantime, @BrokenEagle98 let's not tag them "tumblr samples" just yet, albeit you can comment that link in for now.

Replacing samples is currently a fairly complicated issue since it inflates the deletion count for users, and opens up the possibility for stealing credit from other users if done outside of the post replacement function. Currently, we don't have any sort of regular cleanup job for these, so it's a really, really strange situation.

I'm fine with replacing my own uploads, since I took the time tagging and commenting all these myself (a lot of the Matilda Vin tumblr posts had creation videos and associated comments from me pointing out their youtube videos). I really don't want to give away all credit to an approver for my effort. At this point are we simply not allowed to manually re-upload and flag, as before?

D1ce said:

I'm fine with replacing my own uploads, since I took the time tagging and commenting all these myself (a lot of the Matilda Vin tumblr posts had creation videos and associated comments from me pointing out their youtube videos). I really don't want to give away all credit to an approver for my effort. At this point are we simply not allowed to manually re-upload and flag, as before?

Your name will still remain on it, just that we have a current replacement method for it so you don't have to create a new post entirely.

Benefits are:

• All favorites are still there
• We don't have to deal with moving a post to immediate deletion when we don't have to

There's likely more I'm missing, but just take a look at any post that's been replaced in place at mod actions when you get the chance.

evazion said:

To sum things up:

post #2762115
Best available: http://data.tumblr.com/9bfef40fd3ff6d849263b6a0520691ee/tumblr_orrvvh7F8a1rrqh2lo1_raw.jpg
Sample: https://68.media.tumblr.com/9bfef40fd3ff6d849263b6a0520691ee/tumblr_orrvvh7F8a1rrqh2lo1_1280.jpg

post #1455864
Best available: http://25.media.tumblr.com/tumblr_lxbzel2H5y1r9yjhso1_500.jpg or http://data.tumblr.com/tumblr_lxbzel2H5y1r9yjhso1_500.jpg
Sample: https://25.media.tumblr.com/tumblr_lxbzel2H5y1r9yjhso1_1280.jpg (metadata altered)
Doesn't work: http://data.tumblr.com/tumblr_lxbzel2H5y1r9yjhso1_1280.jpg
Doesn't work: http://data.tumblr.com/tumblr_lxbzel2H5y1r9yjhso1_raw.jpg

In other words, there are cases where 1) changing to _raw doesn't work and 2) the _500 version is more complete than the _1280 version. Let's just be careful what we do until we sort all this out.

Why does Tumblr have to be such a damn headache, Christ.

@Mikaeri, I see now. Thanks for clearing that up.

Switched the scanner back on but with commenting only, i.e. no tagging...

BrokenEagle98 said:

I've stopped the full scan for now, although weren't reuploads fixed so that the original uploader still has the credit...? At least, that's what I thought...?

If they go through the replacement process, at least, but only approvers+ can do that, and "replacing ugoiras, posts with notes, or posts below #100000 aren't allowed yet due to some current limitations" (topic #14063).

I've been limiting my automatic replacements to just pixiv posts, which I can verify myself, but perhaps I can work out something special for this case... though @reiyasona seems to have already kicked off something like that?

RaisingK said:

If they go through the replacement process, at least, but only approvers+ can do that, and "replacing ugoiras, posts with notes, or posts below #100000 aren't allowed yet due to some current limitations" (topic #14063).

I've been limiting my automatic replacements to just pixiv posts, which I can verify myself, but perhaps I can work out something special for this case... though @reiyasona seems to have already kicked off something like that?

I believe he's only doing posts that match these for now:

• height:1920 source:http*://*.media.tumblr.com/
• width:1280 source:http*://*.media.tumblr.com/

Since they're direct links to begin with, they can easily be replaced, but for ones with HTML links to Tumblr posts I think there will be a little more to do. Multi-image posts come to mind...

Mikaeri said:

...but for ones with HTML links to Tumblr posts I think there will be a little more to do. Multi-image posts come to mind...

Just scrape the comments I leave behind. I already do the heavy lifting when I process multi-image sources.

I do an MD5 match to validate a sample is a sample, and I do an IQDB/SauceNAO match to validate an MD5 mismatch is a mismatch.

Fair enough. I don't mind whoever replaces any of them as long as they're all replaced in the end.

Then we'll have to get to the problem of deleting _1280 images that already have a full-size with the same _raw md5 uploaded... There's no going around it for now, so this'll be tough.

In all honestly I wish there was a way to at least change uploader names, so we could move all sample uploads to some dummy account if needed.

This sounds like it's gonna take a long time replacing all those Tumblr images that have raw versions. Has anyone here started doing that yet?

reiyasona's working on it right now (has a script running), and we're discussing some stuff in the discord. See mod actions for more fun shenanigans.

I was going to get around to parsing BrokenEagle98's comments eventually, but maybe this will already be settled by then...

I think we still might need your help if something happens, so just stick around.

Besides, evazion still has to get in contact with albert in case we want to start looking at direct image links for bad tumblr id posts since tumblr never truly 'deletes' an image, just the post associated with it. It's the reason why we can recover full-size images just as long as someone has reblogged it or we have the direct image link, one way or another.

Tfw practically every forum-active approver knows how to script and runs their own bots.

CodeKyuubi said:

Tfw practically every forum-active approver knows how to script and runs their own bots.

Welcome to danbooru, where we bring out our active nerd when things need to be done.

This does quite reinforce our status as pretty much the penultimate reference booru board for anime art though. The way it is now, once we discover something we get working on it fast. Not to mention we easily have the best site support by a long shot, going by topic #14119.

Only thing we'll need working on now after this is still cleaning up the md5 mismatch mess on Twitter. More on that in topic #13646.

CodeKyuubi said:

Tfw practically every forum-active approver knows how to script and runs their own bots.

Well I can assure you not all of them do.

Is there any way to have a source of both _raw and also a link to the post? I like to reblog good Tumblr art and not providing a link to the post itself (unlike Twitter, Pixiv, etc sources) makes that more difficult. Or are all of these source changes just temporary?

Temporary. If the source was the link to a Tumblr post to begin with (and you can see it in comments/edit history), then feel free to revert the change.

Do not upload _raw version of existing posts.

I'm reemphasizing what Mikaeri said earlier. Doing this creates needless duplicates. Existing posts with be replaced with their _raw versions via the post replacements system (topic #14063) as soon as RaisingK/reiyasona/BrokenEagle98 are able to automate the process.

Is there a thread to request post replacements if you don't have the privilege? I went ahead and tracked down my own bad_id tumblr uploads (assuming the still live ones will be automated). Ah well, since I'm posting anyway, might as well just request here, but it'd be nice to know for the future if there's a thread.

Please replace post #2276553 with https://68.media.tumblr.com/b578fc702ce0bd407cee71824e701e68/tumblr_nutqb3lzDO1uv6edpo1_raw.png

For the one other bad_id from tumblr of my uploads, I found the _raw but it's byte-wise identical to its pixiv parent. I'm going to flag it, since it is a redundant sample, but mentioning that here in case someone wants to say that we shouldn't be doing that. (There will be plenty more posts like that, as Mikaeri said on page 1.)