onemoreanonymous said:
This artist tag seems to be considered a general tag in their own artists page: https://e621.net/artists/77241

I fixed it
Simply edited the tag to general and back to artist but I seriously don't know how the hell did that happened

Noticed that under Flags, sometimes the approver will be listed as "None" even though someone clearly approved it. Is it a bug, or is e621 haunted?

https://e621.net/post_flags?search%5Bcreator_id%5D=94124

wolfen32 said:
Noticed that under Flags, sometimes the approver will be listed as "None" even though someone clearly approved it. Is it a bug, or is e621 haunted?

https://e621.net/post_flags?search%5Bcreator_id%5D=94124

The approver will be listed as none if it's uploaded by a user with unrestricted uploads, or if it was flagged prior to being approved.

kora_viridian said:
There might be a missing artist on the Avoid Posting list.

I was (belatedly) checking that list earlier, and down in the Conditional DNP section, I noticed a line with no artist listed, between shawoo and shishibyte:

I don't know if that line should have an artist listed, or if there was an artist that was removed from the "avoid posting" list, and their name got removed, but not the rest of the line.

archive.org shows that as of October 2022, that partial line wasn't there; shawoo was immediately followed by shishibyte.

After looking at the diff it looks more like a mistake was made when the conditions were changed rather than it being a missing artist.

• shawoo - only the artist is allowed to post

became

• shawoo - only the artist and goodusername^[1] are allowed to post

- only the artist is allowed to post

Did "Preview Final Tags" on the upload page completely break for anyone else? It just told me "Error loading tag preview [object Object]".

wat8548 said:
Did "Preview Final Tags" on the upload page completely break for anyone else? It just told me "Error loading tag preview [object Object]".

For me "Preview Final Tags" on both the upload page and on ordinary posts returns a 500 error & displays that message if and only if the tag list is empty (and for the upload page, no tag buttons have been clicked).

The site news section, when closed, is now slightly too large - the top row of pixels of the "Site Maintenance" text is bleeding through at the bottom. I'm assuming this is caused by the DText upgrade. (Firefox on Linux if that matters.)

Speaking of the news ticker, why does it keep reappearing for me? I've had to click on 'Dismiss' like four times in two days, even though there's no actual new news.

sexygriffon said:
Speaking of the news ticker, why does it keep reappearing for me? I've had to click on 'Dismiss' like four times in two days, even though there's no actual new news.

Same reason any notice would appear again, your browser is removing the data that hides it. Specifically, the local storage key hide_news_notice.

donovan_dmc said:
Same reason any notice would appear again, your browser is removing the data that hides it. Specifically, the local storage key hide_news_notice.

Okay. Why?

When trying to use nsfw mode in the Wolf's Stash app, cloudfare blocks the api and the captcha test loops. Pretty sure cloudfare is blocking the nsfw content because if you disable nsfw it works fine.

When trying to interact with the E621 API, Cloudflare is flagging my requests and returning 403 with a recaptcha challenge (which my code obviously cannot complete), despite me having the correct User-Agent header ("KemonoBot/1.0 (by imafurryxdlol)").

i guess the latest clownflare shenanigans belong here instead?

i'm an api user, and like all of us i'm getting their captcha challenge page (not your "Please confirm you are not a robot" page) instead of json data from the api endpoints.

i'm not personally affected, but some of us are either unable to pass or having trouble passing their captcha challenge page in heavy web browsers.

current threads, newest to oldest: topic #38079 - topic #38076 - topic #36318 - topic #30192

cinnamoncrunch said:
i guess the latest clownflare shenanigans belong here instead?

i'm an api user, and like all of us i'm getting their captcha challenge page (not your "Please confirm you are not a robot" page) instead of json data from the api endpoints.

i'm not personally affected, but some of us are either unable to pass or having trouble passing their captcha challenge page in heavy web browsers.

current threads, newest to oldest: topic #38079 - topic #38076 - topic #36318 - topic #30192

Yeah, knew it wasn't a problem with my code when I found out a friend was having trouble accessing the API, and some Discord bots were also having trouble too.

I'm not even having problems accessing the API endpoints through my browser, but through my bot's code is a no-go until those problems are fixed lol

The wolfs stash is completely down for me, hoping it gets to be fixed soon :/

cinnamoncrunch said:
i'm an api user, and like all of us i'm getting their captcha challenge page (not your "Please confirm you are not a robot" page) instead of json data from the api endpoints.

the api is working normally again. thank you.

cinnamoncrunch said:
i'm an api user, and like all of us i'm getting their captcha challenge page (not your "Please confirm you are not a robot" page) instead of json data from the api endpoints.

they broke the api again.

cinnamoncrunch said:
i'm not personally affected, but some of us are either unable to pass or having trouble passing their captcha challenge page in heavy web browsers.

now i am. they're giving me the captcha challenge page in my web browser nearly every time i open a tab to e621.

When using "the wolf's stash" app (downloaded from Google play store, using android version 12 latest update with latest updated app) I am prompted to verify with captcha, so I tap "OK" and the cloudflare verification beings loading, it says "verify you are human" with a box to tap on, I tap on the box and it loads and says "failed" for a millisecond and then beginning reloading the captcha. In the end "the wolf's stash" is unusable as the cloudflare verification fails every time causing a neverending verification loop.

sexygriffon said:
Okay. Why?

Ask your browser 🤷‍♂️

translation mode dose not work on mobile

I can't vote when connecting with tor (had been like this for the past week or so). It pops up "Posts updated" after I clicked a vote button but the vote doesn't change. In dev tools the POST request to "votes.json" gives 403 Forbidden with the body {"success":false,"reason":"Access Denied"}. In the same way, adding to favorite usually pops up "Error: undefined" with the same HTTP status and body with the request to "favorites.json".

Implications sometimes don't show up in the suggestions dropdown when searching.

When searching for futanari, the dropdown shows:

futanari -> intersex    184k
futanari/female -> intersex/female    33k
futanari/penetrating -> intersex/penetrating    28k
etc...

This works.

But when searching for floating_island, the dropdown shows:

floating island -> 166
floating_islands -> 3

However, floating_islands is actually aliased to floating_island. When searching for floating_islands, the implication properly shows up

floating islands -> floating island    166
floating islands    3

But floating_islands still shows up, which it shouldn't. So, that's a bit weird and confusing.

mononatriumglutamat said:
Implications sometimes don't show up in the suggestions dropdown when searching.

When searching for futanari, the dropdown shows:

futanari -> intersex    184k
futanari/female -> intersex/female    33k
futanari/penetrating -> intersex/penetrating    28k
etc...

This works.

But when searching for floating_island, the dropdown shows:

floating island -> 166
floating_islands -> 3

However, floating_islands is actually aliased to floating_island. When searching for floating_islands, the implication properly shows up

floating islands -> floating island    166
floating islands    3

But floating_islands still shows up, which it shouldn't. So, that's a bit weird and confusing.

I suspect the tag not having zero posts is throwing it off. A janitor creating a tag correction would probably fix the issue.

Looking at the sql used, tags with zero posts are excluded. The post count being incorrect is the issue.
a b

donovan_dmc said:
I suspect the tag not having zero posts is throwing it off. A janitor creating a tag correction would probably fix the issue.

Looking at the sql used, tags with zero posts are excluded. The post count being incorrect is the issue.
a b

Thank you for looking into it!

mononatriumglutamat said:
Thank you for looking into it!

Of course. Now that I can actually do tag corrections myself, I've done one for it. The issue with that tag specifically should be fixed, though with big systems discrepancies like this can happen pretty often.

[b]Line 1

Line 2[/b]

currently renders as one line, as if you had typed

[b]Line 1Line 2[/b]

Italics, strikethrough, etc. work similarly.

forum #363435 has the voting buttons under it despite there being no BUR visible. From what I can tell, it seems BUR #4590 was posted there by mistake, then was removed, but the voting buttons remained regardless.

crocogator said:
forum #363435 has the voting buttons under it despite there being no BUR visible. From what I can tell, it seems BUR #4590 was posted there by mistake, then was removed, but the voting buttons remained regardless.

This makes sense to me, the poster might have removed the DText that embeds the BUR preview but it's still linked to a pending BUR. If the BUR is rejected, which should be done if the BUR wasn't supposed to be there, the voting buttons will disappear.

crocogator said:
forum #363435 has the voting buttons under it despite there being no BUR visible. From what I can tell, it seems BUR #4590 was posted there by mistake, then was removed, but the voting buttons remained regardless.

The BUR is still referenced to the post and is still pending, the poster just removed the [bur:xyz] marker to display it.

stupidmelon said:
I can't vote when connecting with tor (had been like this for the past week or so). It pops up "Posts updated" after I clicked a vote button but the vote doesn't change. In dev tools the POST request to "votes.json" gives 403 Forbidden with the body {"success":false,"reason":"Access Denied"}. In the same way, adding to favorite usually pops up "Error: undefined" with the same HTTP status and body with the request to "favorites.json".

This is still happening.

stupidmelon said:
This is still happening.

E621 blocks api requests from tor. This isn't a bug, it's intentional.

donovan_dmc said:
E621 blocks api requests from tor. This isn't a bug, it's intentional.

Is there a reason for this decision? And will it be permanent? The only thing I can think of would be DDoS/abuse related, which I would hope could be resolved or at least alleviated by changes other than outright banning traffic from a major service such as that.

manwithnoplan021 said:
Is there a reason for this decision? And will it be permanent? The only thing I can think of would be DDoS/abuse related, which I would hope could be resolved or at least alleviated by changes other than outright banning traffic from a major service such as that.

My words are a bit misleading, since there's no official statement on it. Tor itself is not necessarily entirely blocked, most exit nodes end up blocked for one reason or another. You sometimes might get lucky and get an exit node that isn't blocked. Tor is a major outlet for abuse, and I seriously doubt anything's changing there.

_{This isn't an official statement. I'm speculating.}

The database dumps of the posts have a problem.

https://e621.net/db_export/

There are 70 lines in the latest posts download (https://e621.net/db_export/posts-2023-05-01.csv.gz) where the last character of the field is an unescaped backslash.

This escapes the " and make php's fgetcsv cry, since now the quotation mark is not properly terminating the field.

Case in point:
Line #9566084:@trinityfate62\",,2023-03-23 20:55:36.24986,f,f,f,10,10,0,f,f,f

That \" is killing the import.

whiskeyfur said:
The database dumps of the posts have a problem.

https://e621.net/db_export/

There are 70 lines in the latest posts download (https://e621.net/db_export/posts-2023-05-01.csv.gz) where the last character of the field is an unescaped backslash.

This escapes the " and make php's fgetcsv cry, since now the quotation mark is not properly terminating the field.

Case in point:
Line #9566084:@trinityfate62\",,2023-03-23 20:55:36.24986,f,f,f,10,10,0,f,f,f

That \" is killing the import.

Quotes in e621's CSV files are escaped by doubling, not by backslashes. No character has a special meaning inside double quotes other than the double quote character itself. If you were parsing the CSV under the assumption this line was invalid then you definitely have a lot of other mistakes in your import.

(my code handles this case fine, just saying)

I've always just used the npm module csv-parse for parsing the exports. E621's exports follow standard csv rules, so any competently built library should be able to handle them.

You can see the mess of code I use to grab exports, parse them, stream them, then dump them into an sql database for easy manipulation here

whiskeyfur said:
There are 70 lines in the latest posts download (https://e621.net/db_export/posts-2023-05-01.csv.gz) where the last character of the field is an unescaped backslash.

Keep in mind that some of the fields for each post, most notably the description and sources, can have embedded newlines. In other words, when you hit a newline, you haven't necessarily hit the end of the record for the current post.

The description will have embedded newlines if whoever typed in the description used them; the source will have embedded newlines if there is more than one URL. If someone happened to end a line in the description with a \ , then there will be a line ending with a \ and a newline in the dump.

I have a copy of the posts file from last summer and the first few posts in it that have a line that ends with a backslash are due to the description field: post #103975, post #407329, post #438294, post #545478. Sometimes the description writer was attempting ASCII art; sometimes they ended a line with a :\ smiley. There are probably many more - those are just the first few in the dump.

The newlines in the source field can be seen in the very first post in the database dump, post #14. Since those are almost always URLs, many of them end with a forward slash / , so they might not be confusing PHP's parser.

I write code in Perl to handle the database dumps; I use Perl's Text::CSV_XS module to parse them, which seems to do the right thing. A simple example follows; adjust the subscripts in the "print" line to pick out different fields from the posts file.

#!/usr/bin/perl

# print md5 and tags fields from e621 all-posts csv file

use Text::CSV_XS;

binmode STDOUT, ":utf8";

my $csv = Text::CSV_XS->new ({ binary => 1, auto_diag => 1 });

open my $fh, "<", "posts.txt" or die "posts.txt $!";

# skip first line
$junk = <$fh>;

while (my $row = $csv->getline ($fh))
  {
  print "@$row[3] @$row[8]\n";
  }
close $fh or die "posts.txt: $!";

kora_viridian said:
Keep in mind that some of the fields for each post, most notably the description and sources, can have embedded newlines. In other words, when you hit a newline, you haven't necessarily hit the end of the record for the current post.

I had already taken that into account and php's fgetcsv DOES understand that newlines can be a part of a field. So that's already taken into account. Perl is a bit stupider in that regard in that it doesn't make that assumption. It's understanding of CSV files is more limited.

kora_viridian said:
The description will have embedded newlines if whoever typed in the description used them; the source will have embedded newlines if there is more than one URL. If someone happened to end a line in the description with a \ , then there will be a line ending with a \ and a newline in the dump.

Except as the sample I posted, there is 70 instances where the escape character follows immediately in front of the terminating quote for the field. so, no there will not always be a newline.

kora_viridian said:
I have a copy of the posts file from last summer and the first few posts in it that have a line that ends with a backslash are due to the description field: post #103975, post #407329, post #438294, post #545478. Sometimes the description writer was attempting ASCII art; sometimes they ended a line with a :\ smiley. There are probably many more - those are just the first few in the dump.

The newlines in the source field can be seen in the very first post in the database dump, post #14. Since those are almost always URLs, many of them end with a forward slash / , so they might not be confusing PHP's parser.

PHP's parser is not even parsing urls in this. The CSV import doesn't care what the contents of the fields are, as long as it's text and the CSV record is well formed. there's 70 instances where it is not well formed.

kora_viridian said:
I write code in Perl ...

I do not have the option of using perl right now.

donovan_dmc said:
I've always just used the npm module

NPM is for nodejs. Not php. PHP is a lot more strict about well formed csv formats than either perl or nodejs is.

wat8548 said:
Quotes in e621's CSV files are escaped by doubling, not by backslashes. No character has a special meaning inside double quotes other than the double quote character itself. If you were parsing the CSV under the assumption this line was invalid then you definitely have a lot of other mistakes in your import.

(my code handles this case fine, just saying)

It is not the quotes that are not being escaped, but the actual escape character itself.

Ergo, when a line has \" as a terminating quote mark, PHP does not see it as such.

You also wrote it in python, not php. I am restricted to PHP right now.

PHP is a lot more strict about what passes for a CSV record, but it does allow new lines mid-field.

whiskeyfur said:
It is not the quotes that are not being escaped, but the actual escape character itself.

Ergo, when a line has \" as a terminating quote mark, PHP does not see it as such.

You also wrote it in python, not php. I am restricted to PHP right now.

PHP is a lot more strict about what passes for a CSV record, but it does allow new lines mid-field.

Going back and looking at https://datatracker.ietf.org/doc/html/rfc4180 now to see if perhaps this is a PHP issue, not the file. It could very well be.

whiskeyfur said:
Perl is a bit stupider in that regard in that it doesn't make that assumption. It's understanding of CSV files is more limited.

If you're rolling your own parser in Perl with split(), then yes, newlines in the middle of the record will confuse you. If you're using a Perl CSV-parsing library, it has a knob where you can tell it whether to expect embedded newlines or not. That's what the binary => 1 piece does in the code I posted.

Except as the sample I posted, there is 70 instances where the escape character follows immediately in front of the terminating quote for the field. so, no there will not always be a newline.

\ isn't engraved in stone as an escape character. In a lot of C and Unix-derived string-handling things, it is treated that way, but there's nothing that makes it inherently an escape character. The string-processing stuff in most languages I know of can be told to treat it as an escape character, or not.

I downloaded the 2023-05-01 posts file just now and looked through it.

The oldest post by post number that has a \ immediately followed by a " is post #78421. It comes from the DeviantArt link in the sources. That link was put there a little over 8 months ago.

The next oldest post with the same condition is post #163350. It comes from the description, which was added about 9 months ago.

Next oldest is post #214787. Description again, added about 10 months ago.

Next oldest is post #447860. Description again, added about 9 years ago.

Next oldest is post #478326. Description again, added at initial upload, 9 years ago.

Next oldest is post #594943. Description again, added at initial upload, 8 years ago.

Next oldest is post #630016. Description again, added at initial upload, 8 years ago.

I do not have the option of using perl right now.

The idea is not that you should switch languages. The idea is to ask "what is more likely"?

Three different people have said that they are able to successfully parse the posts file. One of them rolled their own parser in Python. Two of them used a library available to the language they use - Node.js in one case, Perl in the other.

Some of the lines in the posts file with the alleged defect have been there for nine years or more.

Two of those people told you, in this thread, that they use their code to parse the posts file for further processing.

The third person didn't mention anything in this thread, but the thread they linked to with their code shows examples of how they build on their parsing code to answer questions about the ranking of various posts, the counts of various kinds of tags, and how far certain tag implication and alias chains go. All of that code was posted about two and a half months ago, so it presumably worked with the posts file as it was at that time, which would have included all of the posts listed above.

So, what is more likely?

1. Nobody else who has attempted to parse the posts file, at any time over the past nine years, has noticed the alleged defect.

2. Three different people are all lying to you about the capabilities of the code they have written and/or the language they wrote it in.

3. Your code is wrong.

whiskeyfur said:
NPM is for nodejs. Not php.

I'm aware of that.

whiskeyfur said:
You also wrote it in python, not php.

They are also aware of that.

We aren't giving you code to fix the problem. We're telling you we don't have the problem, so it's something you alone need to solve.

That issue is a bug in the library you are using. E621's csv files have been able to be parsed by numerous libraries across various languages, so coming back to us and complaining about it isn't going to fix it.

kora_viridian said:
If you're rolling your own parser in Perl with split(), then yes, newlines in the middle of the record will confuse you. If you're using a Perl CSV-parsing library, it has a knob where you can tell it whether to expect embedded newlines or not. That's what the binary => 1 piece does in the code I posted.

As I've already explained three times now, newlines mid field are not the problem.

donovan_dmc said:
We aren't giving you code to fix the problem. We're telling you we don't have the problem, so it's something you alone need to solve.

I understand that.

donovan_dmc said:
That issue is a bug in the library you are using.

Yes, I know that now.

More to the point I thought that php's fgetcsv function was RFC compliant. Thought, as in past tense.

I'm writing up a bug report to submit to the php devs now, because it's not processing the line as it should, because it is indeed escaping the last quote when it shouldn't.

kora_viridian said:

So, what is more likely?
1. Nobody else who has attempted to parse the posts file, at any time over the past nine years, has noticed the alleged defect.
2. Three different people are all lying to you about the capabilities of the code they have written and/or the language they wrote it in.
3. Your code is wrong.

This is a language bug, the worst kind to find, so #4.

a,b,c,d,e,f
1,2,"3\","tt",55, "tt\"

should NOT yield

Array
(
[0] => 1
[1] => 2
[2] => 3\",tt"
[3] => 55
[4] => tt\"
)

damned bug existed all the way back to php 5.6... I can't be this lucky.

whiskeyfur said:
damned bug existed all the way back to php 5.6... I can't be this lucky.

Par for the course with php. (yes this is very outdated, but it should still get the point across)

donovan_dmc said:
Par for the course with php. (yes this is very outdated, but it should still get the point across)

PHP is good for a lot of things, but like any development effort, match the language to the task after requirements have been hammered out and taking into account the skills of your programmers and what resources you have.

whiskeyfur said:
As I've already explained three times now, newlines mid field are not the problem.

The reason why I said that is that you were extolling the virtues of PHP's superior CSV parsing over Perl, and, later, over Node.js. I was trying to point out that naively-written Perl would indeed have trouble parsing e621's CSV, but using Perl's CSV library would not have the same trouble.

As it turns out, PHP's default CSV parsing may be worse than either Perl or Node.js.

Shifting gears...

Here is some free advice unrelated to coding. You may not want it, but here it is anyway.

I feel like you have been trained by something to lead off with arguing and complaining when you post on the Internet. I don't know exactly what it is, but my guess would include the alternatives of Twitter (about 85% chance - the algorithm promotes controversy), brogrammer culture (about 10% chance), or something else (about 5% chance).

In my experience, the Internet, in general, does not respond well to that approach.

If you're within one of the spaces where that's a normal mode of discourse, like Twitter, that approach probably works, more or less.

If you use that approach, consciously or not, on the wider Internet, you will probably find that your interactions with other posters are less than satisfying.

I just tried uploading a pic but got a duplicate warning with this deleted pic. It, and other pics by that artist, emeralds_(artist), were taken down at their own request. Yet they're not on the DNP list, there's no warning tag implied to their artist tag, and other pics of theirs have been uploaded since then.

I'm guessing they rescinded their DNP? If so, can the pics deleted at their takedown request be automatically recovered or would they require re-uploading?

kora_viridian said:
The reason why I said that is that you were extolling the virtues of PHP's superior CSV parsing over Perl, and, later, over Node.js...

Don't start. That was not what I was saying at all.

kora_viridian said:
As it turns out, PHP's default CSV parsing may be worse than either Perl or Node.js.

As we already figured out. You're late to the party, please try to keep up.

kora_viridian said:
I feel like you have been trained by something to lead off with arguing and complaining when you post on the Internet.

I've been trained to trust that the libraries and language work, and with what testing I was able to do, the only piece that remained at the time was the file itself I was trying to parse. Yes, 99.999999% of the time it is user error and that is normally my default mode. I had tested the script with sample files (which I now realize worked because they did not have that \" character sequence), and I had tested other pieces of the entire system to figure out if it was a disk error, memory or otherwise.

That left the only piece I hadn't written myself, the csv file in question. It did not occur to me at the time that the language itself was at fault.

So, again... please try to keep up. We already covered all this.

As for where I got my training? I started on mainframes and was on the internet years before mosaic came out as a browser for the common man.

Stop looking for an argument.

onemoreanonymous said:
I just tried uploading a pic but got a duplicate warning with this deleted pic. It, and other pics by that artist, emeralds_(artist), were taken down at their own request. Yet they're not on the DNP list, there's no warning tag implied to their artist tag, and other pics of theirs have been uploaded since then.

I'm guessing they rescinded their DNP? If so, can the pics deleted at their takedown request be automatically recovered or would they require re-uploading?

Deleted content shouldn't be reuploaded, if you want the content back you'll have to ask the artist to request the content be undeleted. Chances are that it's not a mistake and they're supposed to stay deleted.

onemoreanonymous said:
I'm guessing they rescinded their DNP? If so, can the pics deleted at their takedown request be automatically recovered or would they require re-uploading?

faucet said:
Deleted content shouldn't be reuploaded, if you want the content back you'll have to ask the artist to request the content be undeleted. Chances are that it's not a mistake and they're supposed to stay deleted.

Looking through the artist's Twitter, it appears as though they had a bad run-in with one of their images, where a janitor falsely removed the image because they mistook Gardevoir as being too human. They later said that the appeal went all the way up to the admins, and that one of the admins said that "skin color doesn't matter, a green human is still a human", and that the artist was removing all of their content because it's "just stress I could do without". I don't understand why they haven't applied to be DNP.

Videos with sound (using mobile to browse).

While I can hear the audio, the screen goes completely green the moment the video starts. The video is apparently playing according to the progress bar, but no actual video can be seen. Help.

anontea said:
Videos with sound (using mobile to browse).

While I can hear the audio, the screen goes completely green the moment the video starts. The video is apparently playing according to the progress bar, but no actual video can be seen. Help.

As someone who also primarily uses E6 on mobile. I've only encountered this problem once. I'd recommend you try to log out, clear out the cache/data for whatever browser your using, and log back in. It should fix the issue. That or add E6 to your home screen for quicker access if you haven't already. Hope that would help.

I don't know whether it's my own network or my mobile phone. Now I can't watch the video, and I need to refresh it repeatedly before I can watch it.

I looked over the posts of the last year, but found nothing.
Firefox 112 on Windows 11: The shortcut "f" (as announced in the tooltip) does not work to favorite a post.

neki said:
I looked over the posts of the last year, but found nothing.
Firefox 112 on Windows 11: The shortcut "f" (as announced in the tooltip) does not work to favorite a post.

Are you using re621 by chance? If so, disable it and try again
re621 changes around keybinds and such

If you don't know what re621 is, you can safely ignore this

donovan_dmc said:
Are you using re621 by chance? If so, disable it and try again
re621 changes around keybinds and such

If you don't know what re621 is, you can safely ignore this

F is still the default hotkey for favorites in re621 >_<
https://github.com/re621/re621.Legacy/blob/master/src/js/modules/post/PostViewer.ts#L77

bitwolfy said:
F is still the default hotkey for favorites in re621 >_<
https://github.com/re621/re621.Legacy/blob/master/src/js/modules/post/PostViewer.ts#L77

Hey, I don't know how much you break in there (╯°□°)╯︵ ┻━┻
In other news, it works both with and without re621 for me ¯\_(ツ)_/¯

donovan_dmc said:
Hey, I don't know how much you break in there (╯°□°)╯︵ ┻━┻
In other news, it works both with and without re621 for me ¯\_(ツ)_/¯

I don't have re621, but it could be one of my addons. I'll investigate.

neki said:
I don't have re621, but it could be one of my addons. I'll investigate.

You should also make sure that the keyboard shortcuts are enabled in the settings.

bitwolfy said:
You should also make sure that the keyboard shortcuts are enabled in the settings.

Ah, that was it, shortcuts were disabled. I didn't know that is a thing you can do.

Anyways, thanks for the tip, it works now.

For what ever reason I don't know If it's a bug or it's because I almost dropped my phone when I was scrolling but a yellow theme as been added to my side of the site and I can't tap/click on post. When I try to it only brings up the ability to change the tags and thats it

Adding a tilde to a set search does not work.

Example:
~set:cartoony ~set:cartoonyjp

Expected result:
Posts from both sets should show up.

Actual result:
Nobody here but us chickens!

apotatothatflew said:
For what ever reason I don't know If it's a bug or it's because I almost dropped my phone when I was scrolling but a yellow theme as been added to my side of the site and I can't tap/click on post. When I try to it only brings up the ability to change the tags and thats it

That's not a bug, you're in edit mode. I don't know how the site works on mobile, but you should be able to find a button to go back to view mode.

wolfmanfur said:
Adding a tilde to a set search does not work.

Example:
~set:cartoony ~set:cartoonyjp

Expected result:
Posts from both sets should show up.

Actual result:
Nobody here but us chickens!

That isn't a bug, it's how the search works. The same thing happens with id and similar. The tilde can't be used with meta tags, and meta tags cannot be doubled up. The rightmost is used, and all others are discarded.

The tag kitsunebi_(artist) seems to simultaneously be aliased to demireality and to be its own tag. See this search link.

jmvryt said:
Looking through the artist's Twitter, it appears as though they had a bad run-in with one of their images, where a janitor falsely removed the image because they mistook Gardevoir as being too human. They later said that the appeal went all the way up to the admins, and that one of the admins said that "skin color doesn't matter, a green human is still a human", and that the artist was removing all of their content because it's "just stress I could do without". I don't understand why they haven't applied to be DNP.

I think the artist was bothered because they were the one who uploaded their own works here and this drove them off, and it doesn't mean they don't want their stuff here.

Well, the deleted images are preventing them from being reuploaded since they're duplicates, despite not being a DNP. Can an admin weigh in on this?

The tag kitsunebi_(artist) seems to simultaneously be aliased to demireality and to be its own tag. See this search link.

jmvryt said:
Looking through the artist's Twitter, it appears as though they had a bad run-in with one of their images, where a janitor falsely removed the image because they mistook Gardevoir as being too human. They later said that the appeal went all the way up to the admins, and that one of the admins said that "skin color doesn't matter, a green human is still a human", and that the artist was removing all of their content because it's "just stress I could do without". I don't understand why they haven't applied to be DNP.

I think the artist was bothered because they were the one who uploaded their own works here and this drove them off, and it doesn't mean they don't want their stuff here.

Well, the deleted images are preventing them from being reuploaded since they're duplicates, despite not being a DNP. Can an admin weigh in on this?

Edit: it seems Emeralds is the new nickname of Mricantdraw, whose works were taken down at the same time and who also wasn't put on the DNP list.

admins pls respond

I have found a new bug that is bothering me.
The keyboard shortcut R is no longer working.
To quote the Keyboard Shortcuts page:
r Go to random post
I use it often when editing tags or when I am looking at my own posts & favorites. I sometimes hit R until the acceptable post needs the current tag I am working on added, and then I add it.
Please, fix/bring it back!

I noticed that I cannot navigate pages using A/D anymore. Was working fine yesterday.
It still works when viewing a post, just not for navigating between pages.

I made a small change to the shortcut code yesterday, I will take a look later

Your shortcuts should work again, sorry about that.