EUC(5)

💖 EUC(5) — the manual page, but make it group-chat 💖

okay so this is euc, bestie — the EUC encoding of wide characters page, and it's lowkey iconic once you get the vibe. it lives in the File Formats Manual section of the man pages, iykyk 🎀

---

✨ SYNOPSIS ✨

like, here's the shape of the thing, exactly as written:

ENCODING "EUC"

VARIABLE len1 mask1 len2 mask2 len3 mask3 len4 mask4 mask

that second line is the whole personality btw — four codesets, four lengths, four masks, and one big global mask to rule them all 👑

---

🌸 DESCRIPTION 🌸

so EUC implements a system of 4 multibyte codesets, okay? four of them. not three, not five. four. and each one has its own lil starting-byte vibe check:

• a multibyte character in the first codeset is made of len1 bytes, starting with a byte in the range of 0x00 to 0x7f. and because we still want plain ASCII to work, len1 is always 1. cute, right? 🫶
• a multibyte character in the second codeset is len2 bytes, starting with a byte in the range of 0x80-0xff, excluding 0x8e and 0x8f. sorry bestie, those two are taken 💅
• a multibyte character in the third codeset is len3 bytes, and it starts with the byte 0x8e. that's literally its whole thing ✨
• a multibyte character in the fourth codeset is len4 bytes, starting with the byte 0x8f. also its whole thing 🌈

---

🧚 the wchar_t glow-up (aka how the bytes get turned into a wchar_t) 🧚

the wchar_t encoding of EUC multibyte characters is dependent on the len and mask arguments. first, the bytes get moved into a wchar_t like this:

byte0 << ((lenN-1) * 8) | byte1 << ((lenN-2) * 8) | ... | bytelenN-1

then — and this is the part that matters — the result is ANDed with ~mask and ORed with maskN. so it's a little mask-and-shift situation. very demure, very mindful 🎀

and here's the spicy bit: codesets 2 and 3 are special 🔥 because the leading byte (that's 0x8e or 0x8f) is first removed, and the lenN argument is reduced by 1 before all that shifting happens. so like... the leading byte did its job and then dipped 🙈

---

🍓 the example, because an example ATE 🍓

the ja_JP.eucJP locale has this exact VARIABLE line:

VARIABLE        1 0x0000 2 0x8080 2 0x0080 3 0x8000 0x8080

and here's what each codeset looks like in that case, no cap:

• Codeset 1 consists of the values 0x0000 - 0x007f.
• Codeset 2 consists of the values who have the bits 0x8080 set.
• Codeset 3 consists of the values 0x0080 - 0x00ff.
• Codeset 4 consists of the values 0x8000 - 0xff7f excluding the values which have the 0x0080 bit set.

and notice — notice it, bestie — that the global mask is set to 0x8080. that implies that from those 2 bits the codeset can be determined. two bits. that's all it takes to know which codeset you're in. main character energy honestly 👑

---

🦋 SEE ALSO 🦋

go bother these pages next, they're the friend group:

• localedef(1)
• setlocale(3)

---

nothing dangerous here to warn you about, but you do have to keep your len and mask arguments straight or the whole wchar_t vibe check falls apart. literally. careful with those. 💅

---

and that's the page!! 💖✨🎀🦋🌸🫶

this lil beauty comes to you from Debian, dated September 9, 2019, and it's EUC(5) all day every day. mwah 💋 slay, bye 👋