There are currently 13 images on the site with a completely blank tag. Here's one of them: https://e621.net/post/show/559852/-2014-clothing-equine-female-feral-friendship_is_m

This seems more like a bug really, but if it is possible to alias nothing, a "" (minus the quotes) -> invalid_tag alias could be a quick fix.

It's not searchable as far as I can tell so it's hard for the users to clean up as it stands.

In case someone edits that post, the problem is caused by a leading space in the tag field:

" 2014 clothing equine female feral [etc...]"

purple.beastie said:
There are currently 13 images on the site with a completely blank tag. Here's one of them: https://e621.net/post/show/559852/-2014-clothing-equine-female-feral-friendship_is_m

This seems more like a bug really, but if it is possible to alias nothing, a "" (minus the quotes) -> invalid_tag alias could be a quick fix.

It's not searchable as far as I can tell so it's hard for the users to clean up as it stands.

In case someone edits that post, the problem is caused by a leading space in the tag field:

" 2014 clothing equine female feral [etc...]"

It looks like they already tried to alias it to test_tag, and it didn't work. Also, I don't think it's caused by a leading space. I thought it was caused by tagging a hyphen "-" and maybe some other symbols.

I've seen it‬ ha‭ppen with control 
chara‭cters. Such as left-to-right mark (U+200E). It might be a good i‏dea to ‭preemptively‭ alias those. ‭Inclu‪ding "‎", "󠀁" and "‬‭"‮.‬‫

Yeah it's a bug. Most of the ones tagged are me playing with it and aliasing them does nothing (though I'm going to try a few more things tonight). Here's a TLDR explanation of what is essentially going on:

1. Someone adds - as a tag (or others as Genjar pointed out).
2. The system removes the - part because a tag can't start with - (otherwise it would conflict with -tag searches).
3. A tag is made, but that tag literally contains nothing. It also creates an entry for nothing.
4. Since that tag entry contains nothing, there's no way to put it into any tool, the site won't acknowledge it whatsoever, and is impossible to remove once added.
5. This thread.

That being said, I'll play with it again tonight to see if I can trick the system into removing them and/or get them to actually alias properly.

Edit: By the way, a leading space (or any number of spaces anywhere) won't cause the bug, it's only when certain characters are added. The leading space is just the result of how the system alphabetizes the list (control characters and symbols, 0-9, a-z, and then non-latin—or something like that).

Edit 2: It looks like some of these are allowing me to alias them. I don't think they trigger the bug, but I'll go ahead and alias them to invalid_character just in case.

Edit 3: Lol at U+200E (first line)

Edit 4: Yeah, as far as I can tell there isn't much we can do about this at the moment. The software we use is fairly ancient and has generally poor Unicode support. Any fixes we would make would be workarounds and likely temporary, the final solution will ultimately have to upgrade the software and get proper Unicode support (which will make fixing this significantly easier).

Maybe also alias U+202A, U+202B, U+202C, U+202D, U+202E? I could imagine someone using those to troll, and there's still an old glitch that gets some browsers stuck in infinite loop if you use certain sequence...

Genjar said:
Maybe also alias U+202A, U+202B, U+202C, U+202D, U+202E? I could imagine someone using those to troll, and there's still an old glitch that gets some browsers stuck in infinite loop if you use certain sequence...

And after seeing the U+200E, I'm kind of curious to see how those would show up on the alias list...

I'm not sure that these will accomplish anything as you can still type them in the middle of tags and get the same effect, but at least it will be easier to spot them if they were typed accidentally (and remove them).