Lots of unicode emojis being broken due to ZWJ being filtered out
I've been looking into an issue where quite a few Unicode emojis break down when being send in the chat. This seems can be observed on both the website as the mobile client, as well as in the chat client I'm working on. The root cause seems to be that the "Zero Width Joiner" character (200D in UTF-16 hex representation) gets filtered out somewhere inside the IRC system. This will cause for example the following emoji (🧑🤝🧑) to break down into (🧑🤝🧑)
This issue eventually results in 1329 of the 3512 emojis specified in Unicode 13.1 (https://www.unicode.org/Public/emoji/13.1/emoji-test.txt filtered on "fully-qualified") to break down when being send in chat.
15
votes
