Unicode characters in chapter name

[Kyuzo]

I tried to add some Unicode characters, e.g. 🅻 , in the name of a chapter, but I this failed with a "Warning: Something went wrong with your upload".
I didn't have a problem when I tried this with Manga titles.

Edit 1: Some technical information: The AJAX POST request failed from the server side with status 500.

Edit 2: Only characters in plane 0 (0000–​FFFF) are currently supported.

Edit 3: Manga titles have a slightly different bug. The update is correctly reflected in the manga page in

<span class="mx-1">

. However, once 🅻 was added, the tags

<title>

,

<meta property="og:title" ...>

and

<meta name="keywords" ...>

don't update anymore. I'm not sure if this is caused by cache. This doesn't appear to be caused by cache.

 

[BzzBzz]

This is obviously not a unicode issue. Check english chapters 107-117.

 

[Kyuzo]

I don't see chapter 107 (english) in that page.

It certainly is a Unicode issue. I merely tried to change the name of an already uploaded chapter. The update works if 🅻 is not present.

 

[BzzBzz]

Settings -> Unavailable chapters

 

[Kyuzo]

Ok, I can see the chapter names now. But those were uploaded 2 years ago. This means that the issue appeared after an update of the website's source code.

 

[BzzBzz]

Well I'm sure I saw circled numbers not so long ago in some new chapters.

Looks like unicode plane issue. Symbols from plane 0 (like ♡ or ⑲) works fine but plane 1 (like your 🅻) is problem.

 

[Kyuzo]

Confirmed. This is a plane issue.
I checked to see if this is also a Unicode version issue, however it's not. I tried characters from up to version 13.0 from plane 0.

 

[BzzBzz]

There's a possibility that is not unicode issue but DB issue. If strings stored there not in UTF8/16/32 but in old fixed 16bit per character format then it obviously can't store anything past Plane 0 (U+0000...U+FFFF).

 

[Kyuzo]

The strange part is that manga titles don't have this issue. So assuming the database/tables were configured with the same character set from the beginning, this leaves us with the sole possibility that it's a backend code issue.
I rarely encounter character set issues, but in the times I did, the causes included:
- The character set of the file in which the code for creating/updating entries is saved.
- The string handling function used by the code (or that was set for the database handling library).