How to handle WM_CHAR with /utf-8 and UTF-8 manifest

I'm having trouble figuring out how to update my WM_CHAR handler to work with the new /utf-8 execution mode +manifest. I have it working correctly if I use strings from source with unicode characters. The windows title and console will display them as they should, but now I want my user input to handle it correctly as well.

I can't find documentation about it anywhere, so at some point I tried to just figure it out manually but no luck.

So I try to print the data like this:

case WM_CHAR:    println("WM_CHAR: 0x{:X}[{}]", (uint32)wParam, GetUTF8CharCount((uint32)wParam));

The count function is this:

static uint GetUTF8CharCount(uint32 c) {    if (c <= 0x7F) {        /* U+0000 ... U+007F */        return 1;    } else if (c < 0x07FF) {        /* U+0080 ... U+07FF */        return 2;    } else if (c < 0xFFFF) {        /* U+0800 ... U+FFFF */        return 3;    } else {        /* U+10000 ... U+10FFFF */        return 4;    }}

If I type 'ウ' in a textbox the following info prints:

[Info]: WM_CHAR: 0x4581[3][Info]: WM_CHAR: 0x4581[3][Info]: WM_CHAR: 0xA6[2]

This website tells me that only the last byte is correct.It makes no sense to me, how do I interpret this msg?

Edit:

Additionally I've printed the following info:

CPINFOEX cpInfo;GetCPInfoEx(CP_ACP, 0, &cpInfo);println("ACP: {}, IsWindowUnicode: {}", GetACP(), IsWindowUnicode(hWnd));println("CPINFOEX: {{{}, {}, {}, {}, {}}}", cpInfo.CodePage, std::string(cpInfo.CodePageName), *((uint16*)cpInfo.DefaultChar), cpInfo.MaxCharSize, (uint32)cpInfo.UnicodeDefaultChar);

Result:

[Info]: ACP: 65001, IsWindowUnicode: 0[Info]: CPINFOEX: {65001, 65001 (UTF-8), 63, 4, 65533}

To emphasize that it does work (if I add it directly in source):

Edit 2:

It seems to be related to the active language. If I set my keyboard layout to US-international (my default), then press Win+; to type the 🎉 emoji I get F0 9F 8E 89, but switching to Japanese causes the 0x4581 to appear