Some unicode characters stripped from autocomplete responses

[maddysrc]

Description

Some unicode characters are getting stripped from autocomplete responses, causing emoji to show up incorrectly.

Steps to Reproduce

Respond to an autocomplete interaction with

{
  "type": 8,
  "data": {
    "choices": [
      {
        "name": "🏳️‍🌈 gay_pride_flag",
        "value": "🏳️‍🌈"
      }
    ]
  }
}

Expected Behavior

The response is returned and shown to the client as-is

Current Behavior

Characters \uFE0F and \u200D get stripped from the response

Screenshots/Videos

Client and System Information

Bot info:

• Library: discord.js@13.6.0
• Node.js v17.6.0

Clients tested:

• PTB build 119911/90e94273602cbfd62e06b34dd02c270dfe73a87b
• Android beta 119.7 build 119107

[night]

This behavior appears to be working as intended at present, since we strip certain unicode control characters in many strings due to abuse patterns we've seen in the past. Many fields have this behavior as a result, including usernames, channel names, etc.

[maddysrc]

we strip certain unicode control characters in many strings due to abuse patterns we've seen in the past

That's understandable, but could an exception be made for \uFE0F and \u200D specifically, they're fundamental to making a lot of glyphs (like emoji) show up properly.