@Ciaran
@Background Pony #62CA
They continue doing that because we do nothing about it. The moment those tag histories are purged, they disappear too, because then they can’t leave a mark on the site.

There also should be a limit on how many tag changes you can submit at once, especially if you’re an anon.

@JP
That has been under development, though I don’t know what still needs to be completed.

It’s likely caused by a bug in the way Firefox for iOS handles caching, but it is EXTRAORDINARILY ANNOYING when a comment gets eaten because of some validation error. Can the contents of the message not pre-puopulate in the box after a posting error? For that matter, why do errors and forum posts trigger a full-page reload but successful image comments do not?
If spam is a concern, limit the recovery on validation failures to logged-in accounts only.

The tags page currently includes “pinot noir” and “shiraz” in a search for tags that are aliased but are implied by other tags, however, no tags actually seem to imply them?

@Background Pony #62CA
…So I took a few guesses on which tag(s) matched the search. implied_by:silver berry, implied_by:pinot noir.
No idea why the inconsistency between implies and implied_by. :T

https://derpibooru.org/tags/unicorn%25E2%2580%258B this tag has some symbol at the end and can’t be be removed from the only image

@mdashlw
The “symbol” at the end is our not-so-good old friend Zero Width Space.

Those things (and other invisible space characters) have caused weird tag breakages in the past.

Tag has been taken out back and put down.

Not sure if it’s an error or me being dumb but lately it always reloads when I’m about to send a message

I smash send, the page randomly reloads and my message is deleted, which is a bit INFURIATING _{Hulk smash!} when you’re writing a long message ^^

Okno being silly but it’s slightly annoying and I’m not sure if I can fix it or something or it’s an error on my part ^^”

@FliegerFaustToP47
IIRC, it has something to do with CSRF or session tokens expiring. VERY ANNOYING, especially since the failed message isn’t prepopulated in the box for you to try again.

@FliegerFaustToP47
Sometimes you can “go back” and many browsers will repopulate what you’d typed before hitting the submit button.

@PUBLIQclopAccountant
Yeah, I get that too. Especially if I have too many other tabs open and leave a post I’m editing for too long in the background. I’ve actually developed the habit of just copying all/reloading/pasting before submitting anything to the site that took more than a few minutes to put together.

@FliegerFaustToP47, @PUBLIQclopAccountant & @Ciaran

I experience this with Chrome reloading and losing my tabs too often >.<

But seriously, I tend to copy long tag lists (from the Traditional Editor) or any long forum posts or descriptions for images. Just in case I leave the browser (reloads a lot when I use my camera app, though I think that’s just a Google glitch - been happening with my texts a lot too 😞).

Saves you a lot of frustrations sometimes.

Not sure if the new behavior is the bug or if the old behavior was a beneficial bug, but there’s been a change in the notifications display in the past month or two.

When on the notifications list, after I click a link to view the content and press that back button, the read notification is removed from the list. This was only for desktop Firefox (didn’t try other desktop browsers); iOS Firefox has always had the “new” behavior.

Benefit: easier to track which notifications were read.

“Read” notifications are not removed from the list after pressing the back button until a full refresh occurs.

Benefit: if you want to go back two notifications ago.

This may be a change in how Firefox handles caching instead of a Derpi change.

Hello, I’m looking to do some data analysis on Derpibooru analytics. I see the “nightly builds” section and have successfully downloaded and begun using that. However, the latest data is from May, so I can’t see the last 5 months of activity. Is it possible to get the latest comprehensive pg_dump?

Also, please advise if there is a better forum to direct my inquiries to, particularly involving API content (ex. I’m curious what “image_intensity” is, but I don’t see where I can found out more about that)

@dualreason
Unfortunately the data dumps have not been updated since multiple sources were added in late May. The new feature broke the dumping system and it hasn’t been fixed since.

@dualreason
@JP
Just remembered this was still broken and fixed it

@dualreason
Intensities are a set of 4 floating point values (northwest, northeast, southwest, southeast) corresponding to the average luma value (in the sRGB color space) of the corresponding 4 quadrants of the image.
processor

@byte[]
Oh, wow. Thanks!

Rejoice!

@PUBLIQclopAccountant
Update on this: my:faves and my:upvotes actually are supposed to be supported but currently broken. I’ve made a PR to fix: https://github.com/philomena-dev/philomena/pull/193

@byte[]
Thank you so much for fixing this so quickly, epic #1!

The API page Tag Responses section is missing body-type from the list of possible category values.

I don’t remember when it broke, but the way I’d think linking on an image works is broken. Just found out you need to do the image embed then linkify that image embed instead of the old method of putting link inside the brackets of image. Not sure how many linkages this broke in descriptions, but I had to correct one I knew of manually.

Related to this and should be easy, please enable a button to auto-copy or open up the code for a description. Helps when you want to keep an artist’s “signature” intact on other uploads if one hasn’t done it before on a specific artist or at all.

EDIT: For clarification, this is what I was tweaking
(Link leads to explicit content)
Old version
“Corrected”
(Link leads to explicit content)
I went back and checked a random post, and I’m not sure how the artist had their embedded descriptions still work.

@Blissful690
It’s actually a proper JPEG image, server just says its content-type is text/html, for some reason, and browser belives that..
Btw Opera 12 (probably last adequate browser ever made) displays it properly anyway

Is there or has there been a site bug that adds tags to the image that the uploader did not?

Recently I uploaded an image, for some reason two tags were on it which I didn’t put on. They were not implied by or imply or were allied with any other tag I put on