General Punctuation

Three zero-width characters U+200B through U+200D (space, non-joiner and joiner) differ in how they affect ligation and shaping of adjacent letters such as contextual forms in Arabic.

Eleven invisible characters U+200E, U+200F (left-to-right and right-to-left mark), U+202A through U+202E (embeds, pops and overrides) and U+2066 through U+2069 (isolates) control the directionality of text unless higher-level markup overrides them.

Starting with Unicode 16 (2024), the block has variation sequences defined for East Asian punctuation positional variants of the curly quotation marks ‘...’ and “...”.

They use U+FE00 VARIATION SELECTOR-1 (VS01) and U+FE01 VARIATION SELECTOR-2 (VS02):[3] The non-fullwidth forms are expected to be separated with a space on one side, the fullwidth forms are not: In vertical text, the fullwidth forms should display somewhat differently, and even as regular CJK quotation marks 「...」 and 『...』 if the vertical orientation property is set to "Hans": The General Punctuation block contains two emoji: U+203C and U+2049.

[6] The following Unicode-related documents record the purpose and process of defining specific characters in the General Punctuation block:

The red registration corners mark the glyph metrics and show how the glyph aligns within the space allotted to the character. For variable-width display (left), an adjacent space is expected; for full-width CJK display (right), a space is not necessary.
CJK behaviour of generic quotation marks in horizontal and vertical text when variation selector VS02 is appended. The 'horizontal' column at left is the 'VS2' column of the preceding table.