Ooooy, can't view what makes this site a must for perverts. :V

Can't look at the new booty now. Looks like I'll have to improvise.

Any updates on what's going on? oO

4cfcdf7316da1bc29bbe41291e02e096
https://e621.net/post/index/1/nintendo

f4f224f9d7b2c5453be595f1d0b93af8
https://e621.net/post/index/1/digimon

9943d95661a1f272ef53618a508040a1
https://e621.net/post/index/1/unknown_artist

2027cc35ba1516900d07525a07749fae
https://e621.net/post/index/1/mammal

ada9fd9f033c130284e92e6da67ef780
https://e621.net/post/index/1/feline

eliotcougar said:
"Posts RSS stream", https://e621.net/post/atom , redirects me to google.com
Other RSS channels work fine...

i was going to ask if that happened in other places as well (suspecting a link hijack virus of some sort on your end) but no, i clicked that link and it led to google for me too. odd

edit: looks like the "Comments on my posts" link is dead (he's dead jim!). both when using the esix shortcut in the top right corner of the screen and from the "My Account" page.

Unable to link to a Twitter user page via the Description section of post #859437

The DText involved is

"@LupeSilverwind":https://twitter.com/LupeSilverwind

The result is

"@LupeSilverwind";:https://twitter.com/LupeSilverwind

Looks like it just adds a semicolon despite using the formatting in the DText helpfile

"A link":http://example.com/

This affects the http:// version of the same link as well.

Most 500 errors should be fixed, but there will likely be a couple pages left that won't work just yet.

RSS pages are also still offline, and will likely continue to stay that way for the time being.

Hmm....the "It's dead, Jim" errors stopped showing up, but now uploading doesn't work, it keeps giving 500 error.

A TLDR to the recent server issues:

• There's an incompatibility bug in the particular version of our software that conflicts with our caching program and causes Unicode errors with tags, so we had to disable it a couple weeks ago as a dirty fix (the é in pokémon is Unicode, so you could imagine why this might end up manifesting as a site-wide issue)
• Disabling the cache makes the site really slow
• The caching program got enabled by accident when restarting the server (to apply fixes for unrelated performance issues), which then started up the recent site issues again (the previously mentioned Unicode errors)

We're working on fixing the cache to restore everything to normal, but I don't have a timeline to give quite yet.

eliotcougar said:
"Posts RSS stream", https://e621.net/post/atom , redirects me to google.com
Other RSS channels work fine...

We had to disable feeds for the moment due to a rogue news reader making tons of requests per second. Why it redirects to google I have no idea.

Strongbird said:
Unable to link to a Twitter user page via the Description section of post #859437

The DText involved is

"@LupeSilverwind":https://twitter.com/LupeSilverwind

The result is

"@LupeSilverwind";:https://twitter.com/LupeSilverwind

Looks like it just adds a semicolon despite using the formatting in the DText helpfile

"A link":http://example.com/

This affects the http:// version of the same link as well.

The @ is causing the issue here. It's because the code that parses@somename (e.g., @parasprite) is terrible. I'll put this on my todo list.

Edit: The fact that this breaks inline code only further proves how terrible it is. :V

parasprite said:

We had to disable feeds for the moment due to a rogue news reader making tons of requests per second. Why it redirects to google I have no idea.

Because Varka is an evil genius. Dude spammed the RSS page with tens of thousands of requests per second, now redirected to Google, which means Google likely ip banned him/his network for a couple days until he takes the reader offline. Getting banned by Google is a huge fucking annoyance for most people.

well, the site certainly is loading far faster now than earlier.

Yeah, it's loading a lot faster than before for sure, whoever was spamming the RSS feeds, he sounds like he's a little too lonely.

1. Wait a minute. My subscriptions actually work?!?! Well, sort of. WHAT. I'm not sure if I ever tried using them because I'm only a member. I feel like some admin saw my 84 tag subscriptions, felt pity, and activated that feature for me, but a system bug seems more likely because the subscriptions haven't updated beyond December 21, 2015.

2. When I put something in the My Prices field (i.e., a joke) in User Settings, the text I entered in the About Me field gets reformatted to half its regular width when viewing my profile page. That may be normal behavior for artist accounts, but I am not an artist. To be clear, the text I entered in My Prices was properly suppressed from appearing on my profile page, but it would seem the profile page gets reformatted nonetheless. Clearing the My Prices field corrects the weird formatting.

abadbird said:

2. When I put something in the My Prices field (i.e., a joke) in User Settings, the text I entered in the About Me field gets reformatted to half its regular width when viewing my profile page. That may be normal behavior for artist accounts, but I am not an artist. To be clear, the text I entered in My Prices was properly suppressed from appearing on my profile page, but it would seem the profile page gets reformatted nonetheless. Clearing the My Prices field corrects the weird formatting.

Oh. that explains a lot. I've had mine set to "One apple" more/less since my account's creation and I was wondering why my profile looked different.

Well, I don't know if anyone else is experiencing this, but I am having problems unfavoriting things. It continues to show the green "favorite" button, and when clicked it says "error: you've already favorited this post". No current way to workaround, and I am using the iPhone normally, but my desktop is also suffering this problem.

Constant timeouts again, can't get anything done today either. Mostly 500 and 502, with some rare CloudFlare 524s.

It took several tries to post this.

Uploading still doesn't work. Uploading the usual way gives 500 error message, and uploading via link gives file errors, no matter the link, filename or file type.

Schlob said:
Uploading still doesn't work. Uploading the usual way gives 500 error message, and uploading via link gives file errors, no matter the link, filename or file type.

Seems like a bunch of files I was trying to upload had iMac Compression or something embedded in the image...after converting that out in GIMP2 I was able to get past the 500 Errors.

Bug: Tag scripting randomly deletes previously added tags from a post. Appears to only affect tags you've added yourself, ones added by other users are left alone.

Expected behavior: To have the site add the tags in your script, not deleting anything unless you've specifically ordered it to.

Actual behavior: The scripted tags get added, a varying amount of previously added tags (unrelated to the newly added ones) will get deleted. Not just from the most recent edit, sometimes it seems to grab two previous edits at once and nukes all their tags.

Steps to duplicate: Use a tag script. Duplication isn't extremely reliable, it seemingly strikes when it feels like it, but way too often to ignore.

https://e621.net/post_tag_history/index?post_id=863109

^ Example, though it's happened to me on nearly all of this comic's pages.

Jugofthat said:
Bug: Tag scripting randomly deletes previously added tags from a post. Appears to only affect tags you've added yourself, ones added by other users are left alone.

I can confirm that.
Thought that I might've just imagining it since I wasn't able to replicate it, but evidently not.

Did you use both tax scripting and manual tagging when you added those? That's what I was doing when it happened to me. Manual edit seemed to wipe the tags I had scripted.

Genjar said:
Did you use both tax scripting and manual tagging when you added those? That's what I was doing when it happened to me. Manual edit seemed to wipe the tags I had scripted.

Hmmm, then perhaps there's more going on here than meets the eye... No, I'm sure I've only used tag scripting.

Or well, I did actually input a few tags manually on one post before deciding to just begin by tag scripting the whole comic (as most of the pages had things like species and characters in common), and of course I went back to manual tagging after finding out the script function was trolling me to hell and back. But the unwanted wiping of tags definitely only happened when I used scripts.

Oh, and before I noticed how certain tags would disappear by me entering others, I also ran into the old problem where you're trying to quickly put a certain tag script on every post of the comic (a character tag in this case), and then you do a check and find out the request was only handled for like half of the posts. So that was already a sign things perhaps weren't going the way they're supposed to.

Jugofthat said:
Bug: Tag scripting randomly deletes previously added tags from a post. Appears to only affect tags you've added yourself, ones added by other users are left alone.

I haven't tested it, but I suspect this is only if you don't refresh the post index before tag scripting. Something like this:

Tab 1: Post index
Tab 2: post #1

• Open the post index (Tab 1)

post #1 tags: cat solo

• Open post #1 in a new tab (Tab 2)
• Add fish dragon to the tags

post #1 tags: cat dragon fish solo

• Close Tab 2
• On post index, tag script -cat horse
• since post #1 tags on the index last loaded cat solo on the index, the script removes cat and adds horse

post #1 tags: horse solo

But I did refresh the index. I went back to the index from one post where I was manually slotting some tags in, and I believe I did so through the "Posts" button at the top. Maaaaybe I used my browser's back button, but at least I certainly didn't have multiple tabs open.

And it happened to nearly every page of that comic, while prior to that I had only manually edited one.

Self-edit: See forum #191658 for a workaround

Self-edit: Semi-intentional/Fix WIP (see post below)

Probably not a bug, and somewhat minor, but:

\h4. This is a main header

\h5. This is a sub header

This is regular text

This is regular text

This is regular text

This is regular text

----------------------------------------------------

Also, idk if this is intentional either, but unescaped headers seem to do..something inside \

tags:

[code]
h4. This is a main header

h5. This is a sub header

This is regular text

titanmelon said:
Probably not a bug, and somewhat minor, but:

Bug:Header dtext has no proceeding linebreak

Expected behavior: ???

Actual behavior:

-

Steps to duplicate:

\h4. This is a main header

\h5. This is a sub header

This is regular text

This is regular text

This is regular text

This is regular text

----------------------------------------------------

Also, idk if this is intentional either, but unescaped headers seem to do..something inside \

tags:

[code]
h4. This is a main header

h5. This is a sub header

This is regular text

Regarding newlines and headers

This is semi-intentional. Previously headers would swallow the line before/after if you didn't use 2 line breaks. This was a way for me to fix that bug (it caused a lot of "hidden text" to show up in random wiki pages to anyone unaware of the bug).

Code block and block-level tags

TLDR answer: Writing text parsing code that works as intended in all cases is a huge pain in the ass. I've got it mostly working but there's still some buggy parts of the code that need to be rewritten.

Here is a list of known issues:

• Block-level tags (headers, lists, tables, etc.) inside \

tags tend to alter spacing and newlines
* Code, quote, and section tags have an unintended extra newline at the end
* @ username code is buggy and unreliable (especially inside custom link titles)
* Link code is much better than it used to be but is still easy to break
* Tables are still pretty messy, and unpredictable with block-level code. I may have to disable those for tables if I can't figure out a solution.
* Backslash escaping breaks certain DText (especially links)
* Shortcuts like -> (arrows) don't work inside lists right now (fix is ready)
* (JavaScript-related) Unicode outside the 8-bit range (e.g., emoji) doesn't get encoded properly when quoting someone (by hitting the respond button)

Workaround for DText inside \[code] tags: Use backslash escapes for headers, lists, tables, etc.

Now it seems that I've been hit by the 500 error on upload bug, trying to upload a local file. I'm able to browse and edit tags on existing posts, however.

Spoilers: If there's an empty line in a spoiler, the entire spoiler is always visible.
Example: forum #190060

Munkelzahn said:
Spoilers: If there's an empty line in a spoiler, the entire spoiler is always visible.
Example: forum #190060

It seems to happen with all formatting.If you add a blank space you have to add the formatting to each block of text.

Munkelzahn said:
Spoilers: If there's an empty line in a spoiler, the entire spoiler is always visible.
Example: forum #190060

Solution:
Instead of leaving a blank line
-
Use hyphens to denote linebreaks.

Munkelzahn said:
Spoilers: If there's an empty line in a spoiler, the entire spoiler is always visible.
Example: forum #190060

DragonFox69 said:
It seems to happen with all formatting.If you add a blank space you have to add the formatting to each block of text.

Non-block level formatting is all inline currently. That is, it doesn't play well with line breaks. I may or may not change that at some point but that's going to have to be part of a bigger rewrite.

Furrin_Gok said:
Solution:
Instead of leaving a blank line
-
Use hyphens to denote linebreaks.

The black_theme -> dark_theme alias seems to need refreshing, since both are now listed in the master tag list:
https://e621.net/tag?name=*_theme&type=&order=count

Not the first time something like this has happened. Someone recently edited the wiki entry for the defunct black_theme. Could that be what's breaking these: wiki edits?

Genjar said:
Someone recently edited the wiki entry for the defunct black_theme. Could that be what's breaking these: wiki edits?

I'll need to test this one, but based on what I know about the alias/tag code I doubt it. The two parts of the site aren't actually directly linked.

Any estimate on RSS re-enabling?.. It's been 12 days... I'm already down to 50000 un-fapped-to images in my RSS-reader... Does the attacker still spam requests?..

Just a little visual bug:

When a post has both a parent post and a child post, clicking the child post section to expand it covers up the parent post link. Expected behavior is that it expands downward without moving position.

o.O i can't seem to get any of commissionedbutts pics to come up. neither tag search or clicking the artist tag from the tag index work. it just gets stuck loading longer than usual then gives a "he's dead, jim" error message.

1c014eb585facf075a4ac754821b8ba4

Bug: Blacklisted images flash on the screen briefly when a page with blacklisted images on it loads.

Expected behavior: The images shouldn't load at all.

Actual behavior: Thumbnails load, then they disappear.

Steps to duplicate: Blacklist some search terms, then load any page with thumbnails such as /posts

The current implementation of the blacklisting mechanism uses javascript to remove images that have already loaded.

This is bad for 3 reasons:

1. The user still has to see the thumbnails they have blacklisted no matter how heinous the content.
2. This is a waste of resources for both the user and the server because the images load and then are removed from the browser viewport.
3. The content, because it has already loaded, remains in the browser cache.

Dason said:
2. This is a waste of resources for both the user and the server because the images load and then are removed from the browser viewport.

I brought this up a while ago (too lazy to find my post)
but they said they're not changing it

Dason said:
Bug: Blacklisted images flash on the screen briefly when a page with blacklisted images on it loads.

Expected behavior: The images shouldn't load at all.

Actual behavior: Thumbnails load, then they disappear.

Steps to duplicate: Blacklist some search terms, then load any page with thumbnails such as /posts

The current implementation of the blacklisting mechanism uses javascript to remove images that have already loaded.

This is bad for 3 reasons:

1. The user still has to see the thumbnails they have blacklisted no matter how heinous the content.
2. This is a waste of resources for both the user and the server because the images load and then are removed from the browser viewport.
3. The content, because it has already loaded, remains in the browser cache.

That's the way the blacklist works: It's on YOUR end, not theirs, and runs on cookies to work. Your computer can't hide the images until they're loaded up, so if your internet isn't fast enough, then just move to a different tab while it's loading.

Munkelzahn said:
I brought this up a while ago (too lazy to find my post)
but they said they're not changing it

We'd like to change it at some point, but there are a lot of other things that have taken priority.

Furrin_Gok said:
That's the way the blacklist works: It's on YOUR end, not theirs, and runs on cookies to work. Your computer can't hide the images until they're loaded up

That isn't actually true. Your computer can hide the images before they are loaded. The simplest way to achieve this is not to have the images there to begin with. That is, for e621 to send a page that includes all the post information used in blacklisting, but no actual post thumbnail images. Then the blacklisting code can insert images into the DOM if the post is not on the blacklist, and hidden 'blacklisted' thumbnail if the post is on the blacklist.

Yeah, the blacklist being clientside has nothing to do with whether or not the image must be displayed before being hidden.

Bug:Invalid day numbers in date: metatag search are not properly clipped.
Expected behaviour:Searching with a date that is past the end of the month -- eg 2016-02-31 -- should be treated the same as searching with a date that is exactly the end of the month.
Actual behaviour: Some weird date newer than all posts is returned .. meaning that with > operator, no posts are returned; with < operator, all posts (having the relevant tags) are returned.
Example: Compare feral date:>2016-02-30 and feral date:>2016-02-28

On my first attempt to upload post #873493 (warning: absurd res), the post failed to upload, and the error message printed was 'Error: Sample couldn't be created: IDAT: CRC error'.

On my second attempt, using the same image with the same tags, sources, etc., the post uploaded fine. Not sure what happened there, since posting other material of even larger res and file size from the same artist and gallery worked fine.

I've been trying to upload http://www.hentai-foundry.com/pictures/user/PaintUm/404252/Hanging-Demons

When I upload from the URL, e621 displays:
Error: File is an invalid content type:

When I upload the file from my hard disk, e621 displays:
500 Internal Server Error
If you are the administrator of this website, then please read this web application's log file to find out what went wrong.

Munkelzahn said:
I've been trying to upload http://www.hentai-foundry.com/pictures/user/PaintUm/404252/Hanging-Demons

When I upload from the URL, e621 displays:
Error: File is an invalid content type:

When I upload the file from my hard disk, e621 displays:
500 Internal Server Error
If you are the administrator of this website, then please read this web application's log file to find out what went wrong.

I've encountered times when an image wasn't actually the proper file type, though it's usually a corruption on my end, not the artist's. If all else fails, copy the image into a new image file.

When adding a direct image link from DeviantArt, an incorrect userpage link is automatically added, even when the correct one is already included.

See post #875836. The correct userpage link is re-sublimity-kun.deviantart.com, but re_sublimity_kun.deviantart.com is added automatically, which is incorrect.

Think I already mentioned this, but can't find it found it, but it's different behaviour- with bullet/list dtext

self edit: (currently) Known behaviour, fix may(?) be possible in future. See forum #191435
Bug:

Munkelzahn said:
I've been trying to upload http://www.hentai-foundry.com/pictures/user/PaintUm/404252/Hanging-Demons

When I upload from the URL, e621 displays:
Error: File is an invalid content type:

When I upload the file from my hard disk, e621 displays:
500 Internal Server Error
If you are the administrator of this website, then please read this web application's log file to find out what went wrong.

How can I find out what's wrong with the image?
It must be corrupted somehow if e621 says "Error: File is an invalid content type:" when I upload it from a URL.
But why doesn't e621 give me same error when I upload it from my hard disk?

Granberia said:
-set:186 order:score -animated search result looks like this - http://snag.gy/afOm9.jpg
-set part seems to be not working correctly.

This is still my most-wanted fix.
Being able to exclude sets from searches would be insanely useful for tagging projects. It'd make it easy to spot tag wars for problematic tags, and it'd make it easy to keep track of the ones that I've already cleaned. Lately I've ended up checking the same posts repeatedly, which is a complete waste of time.

Bug:
Unnecessary filtering of colon(":") symbol, becomes unremovable.

Expected behavior:
Tags should list as:
:o abc def tabme tag etc

Actual behavior:
Tags list as:
&colon;o abc def tabme tag etc
(Non-valid html entity symbol)

In the wild: post #876938

Steps to duplicate:
I didn't add it but this is how I believe it done got added.
Add tag: ":o" to any post.
Observe it cannot be removed.

titanmelon said:

Think I already mentioned this, but can't find it

Bug:
Anchor tags do weird things when:

• linked from a separate page
• linked from the same page

Expected behavior:
???

Actual behavior:

• Navigating to an anchor tag from a different page works, even if the anchor tag is hidden in a section
• Trying to navigate to the same anchor tag from within the same page does nothing if the section tag is hidden

• When navigating from a different page, the anchor tag skips just a bit lower than if it were from the same page
• (see steps to duplicate below if this is confusing)

Steps to duplicate:

1. Click on these, and note the scroll distance
e621:ambiguous tags#gen_am
avoid posting#conditional_dnp

2. Then, while on the same wiki page, try visiting the section again via anchor links(2d. General: A-M, and Conditional Do Not Post respectively)

This is known behavior and it's due to browsers "ignoring" what is under display:none (the collapsed section is just hiding, it's still part of the page). In order to fix this we need to rewrite sections to use some other method of hiding so the browser can always see it.

In the meantime my current recommendation is to always place the anchor outside of sections unless you are using ,expanded. There's really no other way around it.

Known bug:
[API] JSON/XML requests for post indexes randomly generate exceptions because of changes in preloading child posts.

Hoping to have this resolved and deployed by Monday.

Steps to reproduce:
Any XML/JSON request to /post/index.(xml/json) that includes a tag.

Edit:
Now fixed and deployed.

Bug: Unable to upload images, whether from PC or websites

Expected behavior: Picture is uploaded.

Actual behaviour:
- Uploading file directly from harddrive gives a 500 error
- Uploading from Imgur returns "Source couldn't be opened: Permission denied"
- Trying to upload from Dropbox results in message saying the action is not permitted despite Dropbox being on the whitelist

Alm-Pe said:
Bug: Unable to upload images, whether from PC or websites

Expected behavior: Picture is uploaded.

Actual behaviour:
- Uploading file directly from harddrive gives a 500 error
- Uploading from Imgur returns "Source couldn't be opened: Permission denied"
- Trying to upload from Dropbox results in message saying the action is not permitted despite Dropbox being on the whitelist

I have the same problem BUT THEY WON'T BELIEVE ME :-(

Alm-Pe said:
Bug: Unable to upload images, whether from PC or websites

Expected behavior: Picture is uploaded.

Actual behaviour:
- Uploading file directly from harddrive gives a 500 error
- Uploading from Imgur returns "Source couldn't be opened: Permission denied"
- Trying to upload from Dropbox results in message saying the action is not permitted despite Dropbox being on the whitelist

I find trying it a few times when that happens usually gets it through.

However, what's the particular URL? Does it have any non-alphanumeric symbols?

Furrin_Gok said:
I find trying it a few times when that happens usually gets it through.

However, what's the particular URL? Does it have any non-alphanumeric symbols?

It's an i.imgur one, and aside from the usuals dots and a slash, it's letters and numbers only. I also probably tried directly uploading the picture at least a dozen times and always get the same result.

Alm-Pe said:
It's an i.imgur one, and aside from the usuals dots and a slash, it's letters and numbers only. I also probably tried directly uploading the picture at least a dozen times and always get the same result.

Have you tried copying the image into a new file and "Save as" it? Sometimes it's a matter of the file that ended up on the site got corrupted, so a standard download keeps a bit of corrupted data, possibly. I mean, I don't know enough to know for sure, but maybe it'll work?

Furrin_Gok said:
Have you tried copying the image into a new file and "Save as" it? Sometimes it's a matter of the file that ended up on the site got corrupted, so a standard download keeps a bit of corrupted data, possibly. I mean, I don't know enough to know for sure, but maybe it'll work?

Tried that now, but it still refuses to work.

Alm-Pe said:
Tried that now, but it still refuses to work.

Shoot me a copy of the links through a dmail or something and I'll take a look at it. See what I can figure out.

When attempting to submit an alias request- Pacmancorp to CocoFox -an error appears stating that 'Pacmancorp is already aliased to something', even though there is no wiki entry for the tag and the e621 artist page makes no mention of an alias.

Even when there's no wiki page it still lists all aliases and implications: pacmancorp

Whether it should list aliases on artist pages is another matter of course.

Strongbird said:
When attempting to submit an alias request- Pacmancorp to CocoFox -an error appears stating that 'Pacmancorp is already aliased to something', even though there is no wiki entry for the tag and the e621 artist page makes no mention of an alias.

It should be listed in other names though (it also lists pending ones, even though it probably shouldn't).

And the site shat itself when I tried to upload a pic with a 504 error. So much for them being resolved.

This makes no effing sense seeing as people uploaded five minutes before me without an issue.

Edit: Finally got it up, only took three times.

Can't seem to add an artist tag to an image with two dashes in them, for example here-kitty--kitty turns into here-kitty-kitty.

Blind_Guardian said:
Can't seem to add an artist tag to an image with two dashes in them, for example here-kitty--kitty turns into here-kitty-kitty.

I fail to see what the problem is. Two dashes or one dash doesn't make a difference, it's still the exact same name, and if you're trying to use extra dashes to use the same name as somebody else, you're trying to use the same name as somebody else and should just stop.

Furrin_Gok said:
I fail to see what the problem is. Two dashes or one dash doesn't make a difference, it's still the exact same name, and if you're trying to use extra dashes to use the same name as somebody else, you're trying to use the same name as somebody else and should just stop.

The problem is that is the artists tag, with over 100 posts. It worked before and now it doesn't. I didn't choose their alias. A couple other artists and character tags use two dashes and I suppose aliasing them together would be a solution, but if I had intended for that I wouldn't have posted here in the bug report thread...

Trying to send API requests in XML/Json format seems to return a 403 error, and did so for the past day. Any ideas why?

dds_ said:
Trying to send API requests in XML/Json format seems to return a 403 error, and did so for the past day. Any ideas why?

You need to send a valid and "unique"(for your client, please don't randomize it per request!) User Agent with your requests. Requests without a user agent were disabled because of blatant and repeated abuse.

Please pick something unique that identifies your project and respect that you should respond to http error codes and rate limit your requests. You are encouraged to pick something that contains your e621 username so that you can be contacted if your project causes problems.

Something like "ProjectName/1.0 (by username on e621)"

If you have any questions, send me a DMail.

JackalDarkfox said:
not sure if it's a bug or something I did. There is this green background on everything and when I cick on A picture instead of opening it. It adds it to favorites not sure how to get ride of it ? http://imgur.com/pZW1TXN

On the left you changed the view mode to add to favorites. Change it back to "View Posts".

kiranoot said:
On the left you changed the view mode to add to favorites. Change it back to "View Posts".

Thanks I can't believe I missed that X/

I believe I said this before, but there is an issue where after favoriting an image, you cannot unfavorite it after reloading or revisiting the page. It shows the favorite button, in green, but clicking it gives the error "you already favorited this".

Siral_Exan said:
I believe I said this before, but there is an issue where after favoriting an image, you cannot unfavorite it after reloading or revisiting the page. It shows the favorite button, in green, but clicking it gives the error "you already favorited this".

This appears to be javascript related, but a solid reproduction case hasn't been found. It's on my list of things to look at.

Blind_Guardian said:
Can't seem to add an artist tag to an image with two dashes in them, for example here-kitty--kitty turns into here-kitty-kitty.

I recently made it so that multiple underscores and dashes are automatically converted into one to cut down on dupe tags.

However, since this was never really a problem with dashes I'm going to go ahead and remove that from the code. It should be ready for the next update.

Edit: We may end up just fixing the tags instead.

Siral_Exan said:
I believe I said this before, but there is an issue where after favoriting an image, you cannot unfavorite it after reloading or revisiting the page. It shows the favorite button, in green, but clicking it gives the error "you already favorited this".

Tracked down the cause of this, hopefully the fix will go out by the end of the week. In the meantime, this can be fixed by logging out and then back in again.

Trying to edit the wiki article for missingno. results in the 404 error.
...
How appropriate.