Report site bugs here

@dualreason  
It seems like Picarto broke their API, probably for good this time (all requests to it report 404 Not Found).

@byte[]  
… or not. They slightly changed the URL, but also now seem to be banning API clients (for example, when I try to curl it I get a 403 from Cloudflare).

@byte[]  
Are you using public/private tokens, or is the API fully open access? I would think if you’re self-identifying, they should service requests. Might I suggest poking their help desk to see if they have a recommended solution?

@dualreason  
The path is now fixed in the upstream repository to use the new URL for their API, and it works in my browser, but the problem with the 403s on API clients hasn’t gone away. Perhaps they activated a new Cloudflare feature without realizing that it indiscriminately bans clients that don’t look or act like browsers.

Hello, I have an issue of an upload that I did yesterday in which the file got replaced by an inferior format.
 
Do I have to post it here or in other thread?

@SomeponyfromCHL95  
You should be able to report the merged-to image with an explanation.

@The Smiling Pony  
Well, I did reported yesterday the image due to format difference (JPG -> PNG), however it did the opposite and deleted automatically the PNG one and JPG replace it.
 
But I’ll report it again and explain it the issue more specifically, thanks.

How many letters are allowed in a PM? Because I was sending an extremely long message to somebody else but it appeared an error every time, no matter how many times I refreshed the page or came back later. I had to split it in three so it could be sent.

The syntax for “no parse” doesn’t even need the brackets despite everything saying they are needed. For example, “This is not parsed and there is no brackets attached”:/example
 
They don’t cause issues with links so I have used this method a couple of times. Recently, I noticed that the syntax list actually intends to say the brackets are important, so now I imagine “them not requiring the brackets”:/example is a bug that needed to be fixed.

@Algebroot Neogears  
Not a bug. == is regular textile syntax just like *, _ or ``@ and just like them usually doesn’t need brackets. Brackets just explicitly tell Textile where formatting should be applied if Textile can’t figure it out by itself. For example, multiple lines require brackets.  
1  
2  
3 without brackets around it doesn’t work.
 
1  
2  
3 with brackets around it does.
 
Also if the same syntax is used multiple times and/or nested, Textile can get confused and mix up which opening/closing syntax belongs together.
 
I assume the reference lists [====] so it always works.

@Yoshimon1  
But the thing is, the “no parse” button always adds the brackets when they are not necessary. And if that is the reason the brackets exist then it is kind of inconsistent with links, which only says “On-site link”:/some-link without the opening or closing brackets, so it could end up resulting in:  
“This whole “sentence is meant to be a link.  
Which looks kind of awkward. It’s intended to be:  
This whole “sentence” is meant to be a link.

@Yoshimon1  
The parser rule for them is actually the same.
 
Also, something you might not have known is that literals have a
 
white-space: pre-wrap
 
rule applied to them, so you can actually paste code in them and not have weird issues.

@byte[]  
Ah, neat.

>>79981 (suggestive) is supposed to have 2 comments, but it has only one.

If a page on a forum doesn’t exist, the Next/Last buttons appear even though the next page obviously does not exist (most notably because the “current” page doesn’t either).

@Algebroot Neogears  
Obviously a bug that you can navigate to pages that don’t exist, though I feel I should point out that you would generally have to do this with parameter tampering (the application shouldn’t generate links to nonexistent pages).
 
But explanation of why the links still appear:  
The first/previous links are rendered when a page is not the first page. That page is not the first page.  
The next/last links are rendered when a page is not the last page. That page is also not the last page.

@byte[]  
Basically, the only reason I noticed this was because it appeared when I checked the most recent post in a different forum topic some time recently. It was the first post on a page, and the forum treated it like it (and as a result, the page itself) didn’t exist. Even though there was one post. This issue can’t be seen now, possibly because other posts were made since then (and as a result the aforementioned post can now be seen normally and the page properly exists).

@Algebroot Neogears  
Yeah, it’s the good old pagination bug, and it came up because I forgot a unique constraint. As of now exactly 3 topics are affected, and the fix is relatively straightforward, but will require some coordination to deploy.

I’m having trouble uploading an animated SVG, I’ve seen animated SVGs posted here before, so I’m not sure what may be the problem. I get the preview just fine, but when I click “upload”, I get this:
 
Image can’t be blank  
Image size can’t be blank  
Image width can’t be blank  
Image height can’t be blank  
Image mime type (text/xml) is invalid  
Image duration can’t be blank  
Image orig sha512 hash can’t be blank
 
This is the SVG: https://sta.sh/0134gz84s8bn

@Favitwink  
What an annoying bug in libmagic. Try this one, it should work.

Awesome :D thank you so much for such a quick reply!

Sorry to be back at it again, but the APNG version has a bug too (>>2588031 (merged) in the full view, though not in the thumbnail), I reported it, a mod told me to upload again, but same bug, so the mod told me to post about this here.
 
I was curious about this, and it seems that save for one exception, mine is the biggest APNG in the site, and most of them don’t even have an animated thumbnail, and one that does have it’s preview animated also seems to have a problem with the full view, so, maybe in Derpi there’s a size limit for APNGs before they start to bug out?
 
If that’s the case and it’s something hard to fix, that’s ok, for me at least, I guess I can upload a smaller APNG version, a GIF, WEBM, or something.

I forgot to give a source for the original APNG… Here:  
https://sta.sh/018yenlw5g5x

@Favitwink  
Is the bug that it doesn’t loop smoothly in the full view?

@byte[]  
Looks like optipng ends up deleting data from the file (!)
 
`** Processing: ScootPreenAni.png
Warning: fdAT: no space in chunk cache
Importing APNG
1600x1000 pixels, 4x8 bits/pixel, RGB+alpha
Can’t reliably reduce APNG file; disabling reductions.
(Did you want to -snip and optimize the first frame?)
Recoverable errors found in input. Fixing…
Input IDAT size = 121994 bytes
Input file size = 38787665 bytes

Trying:
zc = 9 zm = 8 zs = 0 f = 0 IDAT size = 118623

Selecting parameters:
zc = 9 zm = 8 zs = 0 f = 0 IDAT size = 118623

Output IDAT size = 118623 bytes (3371 bytes decrease)
Output file size = 21829329 bytes (16958336 bytes = 43.72% decrease)

** Status report
1 file(s) have been processed.
1 error(s) have been encountered.
1 erroneous file(s) have been fixed.   Testing the duration shows that it truncated it:  $ mediastat ScootPreenAni.png
21829329 201 1600 1000 336733 100000
$ mediastat ScootPreenAni_orig.png
38787665 359 1600 1000 600119 100000
`
 
first one duration: 336733/100000 = 3.36733 seconds  
second one duration: 600119/100000 = 6.00119 seconds