⻝ is a Unicode character at U+2EDD, named CJK Radical Eat One. It is not a Chinese character in ordinary use. If you arrived here after copying a glyph out of a radical chart or a dictionary index, the character you almost certainly want is 食 at U+98DF.

The two look identical in most fonts and mean the same thing, but they are separate code points, and no normalization form will ever convert one into the other.

The CJK Radicals Supplement block holds the radical shapes that the Kangxi Radicals block leaves out: variant forms, simplified forms, and the shapes a radical takes in one position rather than another. ⻝ is such a form, indexed under Kangxi radical 184.

Its purpose is the same as that of the Kangxi Radicals block. A table of radicals is a list of abstract indexing elements rather than a passage of text, and giving those elements their own code points lets a font apply chart-specific shapes and spacing without disturbing how the same glyph is drawn in running prose. The block also serves round-trip compatibility with older East Asian character sets that had made the same separation.

Neither purpose involves writing words. ⻝ should not appear in article text, page titles, search queries, filenames, or any other place where the character 食 is meant.

⻝ carries no decomposition of any kind. No normalization form — NFC, NFD, NFKC or NFKD — will convert it to 食, and none ever will, because normalization is stability-guaranteed and decompositions are not added to existing characters.

This makes it a harder trap than the Kangxi Radicals block, whose 214 characters at least fold under NFKC. Of the 115 characters in the CJK Radicals Supplement, only two decompose. The remaining 113, including this one, are permanently distinct from the characters they depict. Text containing ⻝ will fail every search and comparison against 食, in every system, indefinitely.

• 食
• Kangxi radicals
• Unicode normalization

• The Unicode Standard, Chapter 18: East Asia — defines the Kangxi Radicals and CJK Radicals Supplement blocks and their intended use
• Unicode Standard Annex #38 — the Unihan database, including the radical-stroke indexing this block supports