Report site bugs here

I had to push the reverse search distance up to 0.67 in order to find >>1050070 from its source (or its direct URL on dA).

Interestingly, when I view Derpi’s file and the file from dA in Chrome, they appear identical, but when I copy/paste each into an image editor, they have a different color balance. Something to do with color profiles, I assume? How does that interplay with the calculated quadrant values, and has it changed since some time in the past (ie. do old images need recalculating)?

Go to favorite an image and the page either crashes or takes awhile to realize I clicked/tapped the ⭐.
Also uploading does the same thing.

Thanks, but that was actually an effect of just the regular code parser (~~~). :P

Okay thanks 😊 you’re welcome 😌

Off-site images crash my browser when posting.

I’m on an old phone with Bluetooth connected.

Verizon K8 VS500

Google Chrome, incognito.

First crash was in PM and I will not link.

People reporting crashes should indicate what OS and browser version, and provide a link to a page that causes a crash. Otherwise I have no idea.

Emojis don’t appear to get spoilered
⭐🌈⭐

@UmbraPone1
That’s an easy fix and the fix would also solve other smaller issues found recently. I’ll do it once I have some spare time.

@UmbraPone1 &@Princess Celestia
I never noticed that but you’re right. Thanks Celestia 👍🙂

@UmbraPone1
Agree

When I press post on longer comments and posts and probably images with a lot of data on Android with Firefox, the process freezes up. I have to press refresh and then press post again, which luckily always works, which is good. But even so, this is still a hassle of a glitch.

(Of course this post went through successfully without having to refresh it)

@Cosmas-the-Explorer
I can occasionally reproduce, but I have to imagine this is a bug in Fennec, not in the site.

@byte[]
I’m curious. What’s Fennec?

@Cosmas-the-Explorer
Firefox’s Android version.

@JP
Okay everything makes sense now. Thanks.

See if it still happens the same way in your phone’s native browser; Firefox’s Android version has gotten increasingly buggy and full of glitches since they deleted the about:config functionality.

@TexasUberAlles
Oh gee, I wonder if I should considering moving onto Brave browser for android. Firefox had better fix this issue some day.

Reverse search is broken again, it seems.

For example, if I download >>2766810 and search by uploading it, there are no results. Same if I do the search from a URL, or through the API endpoint. I’ve tried it with a variety of different posts.

Is the dupe detector broken for anyone else? I uploaded pictures like >>2771738 (merged) that turned out to be a duplicate of >>1370948, but it never detected it.

@Background Pony #62CA
@Background Pony #8E63
There are some known issues with this part of the site due to ongoing backend changes; the relevant people are aware, but a fix may yet take a while.

@The Smiling Pony
Glad to know I’m not just doing something wrong. It was pretty odd, since I didn’t see anything in the changelog that looked at all related.

@Background Pony #62CA
To my (very limited) understanding, it’s related to changes closer to how the the OS handles file routes than anything specific to philomena code.

I see that the drop-down menu used to autocomplete tags now shows up instantly, but tag aliases are not working with the menu (at least for me).
For example, typing “fullbody” does not display “full body”.

@NICOGAMER3000
There is a trade-off made between space and accuracy for the clientside tag completion.

I can add alias tracking if you find it to be a critical feature, but the extra space needed for it will decrease the number of tags that can be shipped in the dictionary while meeting the size budget. The current minimum images count is 19, and it would likely have to increase to around 27 to fit in the same size.

@byte[]
Oh, okay. I asked just because I noticed this thing.
I can still tag without aliases so you don’t need to update it.