On: In MySQL, use “utf8mb4” instead of “utf8” (2016)

[loeg]

UTF-8 was not restricted to U+10FFFF until November 2003. Prior to that, implementers had to assume it could encode up to 6 bytes of data.

The "utf8" mode in MySQL can only encode 3 bytes of data (up to 16 bit codepoints), which is less than 4 (less than 21 bit codepoints).

"utf8mb4" correctly encodes all UTF-8 codepoints by current standards.

[daneel_w → loeg]

> "utf8mb4" correctly encodes all UTF-8 codepoints by current standards.

Yes, all good and known, but not really what the misunderstanding is about.

"utf8" follows the definitions of UTF-8 as they stood at the time, adhering to Unicode's former code space limit.

[butlerm → daneel_w]

MySQL's adoption of a utf8mb3 limitation for "utf8" was over seven years obsolete when it was adopted. It would have been obsolete in 1996.