How would I convert this encoded emoji to a string...

# announcements

[Paul Woitaschek]

How would I convert this encoded emoji to a string? U+1F3CA-200D-2642-FE0F

[diesieben07]

val codePoints = "U+1F3CA-200D-2642-FE0F"
    .substring(2)
    .split('-')
    .map { it.toInt(16) }

val string = buildString {
    for (cp in codePoints) {
        appendCodePoint(cp)
    }
}

[diesieben07]

Could probably be improved efficiency-wise, but that's the general idea.

[Paul Woitaschek]

Thanks, thats super helpful, trying to figure this out for an hour now.
Performance doesn't matter at all

[Paul Woitaschek]

This is how it ended up in my code

buildString {
  input.drop(2).split('-').forEach {
    appendCodePoint(it.toInt(16))
  }
}

[Paul Woitaschek]

But what exactly is happening here? Is it actually 4 emojis that are mangled together?

[spand]

I only see two

[diesieben07]

Emoji are made up of multiple unicode code points. For example this one has a "male modifier" at the end

[Paul Woitaschek]

And how would I write the code points if I didnt want them to be merged but wanted to display them one after another in a string?

[diesieben07]

Here you have the "person swimming" emoji plus the "man modifier" codepoint which gives you the "man swimming" emoji

[diesieben07]

You can't, the modifiers are not designed to be displayed.

[diesieben07]

They modify what comes before them (similar to how you can turn a into ä with a following modifier code point)

[diesieben07]

You will get them shown (♂️) if they don't follow a supported emoji or the font doesn't support the merged emoji

[Paul Woitaschek]

And if I wanted to unit test on android that all these work correctly, would it work to create a paint; set the typeface I use and call Paint#hasGlyph?

[diesieben07]

I can't really help you with Android, sorry.

[Paul Woitaschek]

But this thing (swimmer+male) could be concidered as a glyph?

[Paul Woitaschek]

I'm heavily confused by all of this toInt, radix, glyph, surrogate topic thing 😄

[diesieben07]

The technical term is "grapheme", as far as I know.
Theroretically there can be as many as you want into one single "emoji" (to e.g. form a family of 2 children and 2 parents from individual person emojis)

[diesieben07]

well, the toInt and radix is just to parse the hexadecimal unicode codepoint representation

[diesieben07]

surrogates are an artifact of how UTF-16 (which Java uses!) works

[diesieben07]

UTF-16 characters are 16 bits, but that's not enough to represent all of unicode (particularly emoji). So that's why there are surrogate characters, that represent one unicode code point as 2 UTF-16 characters

[diesieben07]

Then on top of that you have the fact that multiple unicode code points can be merged into one "grapheme" (what a user would call "character"). These code points can themselves be made up of 2 utf-16 surrogate characters.

[Paul Woitaschek]

danke