Remove workaround for Unicode bug fixed in 16.0

Jules-Bertholet · Jules-Bertholet · commit 425c2506b46a · 2025-01-14T23:41:40.000-05:00
diff --git a/scripts/unicode.py b/scripts/unicode.py
@@ -506,12 +506,6 @@ def load_zero_widths() -> list[bool]:
         lambda cp: operator.setitem(zw_map, cp, True),
     )
 
-    # Unicode spec bug: these should be `Grapheme_Cluster_Break=Extend`,
-    # as they canonically decompose to two characters with this property,
-    # but they aren't.
-    for c in [0x0CC0, 0x0CC7, 0x0CC8, 0x0CCA, 0x0CCB, 0x1B3B, 0x1B3D, 0x1B43]:
-        zw_map[c] = True
-
     # Treat `Hangul_Syllable_Type`s of `Vowel_Jamo` and `Trailing_Jamo`
     # as zero-width. This matches the behavior of glibc `wcwidth`.
     #
diff --git a/src/lib.rs b/src/lib.rs
@@ -99,15 +99,6 @@
 //!         with the [`Default_Ignorable_Code_Point`] property.
 //!       - [Characters](https://util.unicode.org/UnicodeJsps/list-unicodeset.jsp?a=%5Cp%7BGrapheme_Extend%7D)
 //!         with the [`Grapheme_Extend`] property.
-//!       - The following 8 characters, all of which have NFD decompositions consisting of two [`Grapheme_Extend`] characters:
-//!         - [`'\u{0CC0}'` KANNADA VOWEL SIGN II](https://util.unicode.org/UnicodeJsps/character.jsp?a=0CC0),
-//!         - [`'\u{0CC7}'` KANNADA VOWEL SIGN EE](https://util.unicode.org/UnicodeJsps/character.jsp?a=0CC7),
-//!         - [`'\u{0CC8}'` KANNADA VOWEL SIGN AI](https://util.unicode.org/UnicodeJsps/character.jsp?a=0CC8),
-//!         - [`'\u{0CCA}'` KANNADA VOWEL SIGN O](https://util.unicode.org/UnicodeJsps/character.jsp?a=0CCA),
-//!         - [`'\u{0CCB}'` KANNADA VOWEL SIGN OO](https://util.unicode.org/UnicodeJsps/character.jsp?a=0CCB),
-//!         - [`'\u{1B3B}'` BALINESE VOWEL SIGN RA REPA TEDUNG](https://util.unicode.org/UnicodeJsps/character.jsp?a=1B3B),
-//!         - [`'\u{1B3D}'` BALINESE VOWEL SIGN LA LENGA TEDUNG](https://util.unicode.org/UnicodeJsps/character.jsp?a=1B3D), and
-//!         - [`'\u{1B43}'` BALINESE VOWEL SIGN PEPET TEDUNG](https://util.unicode.org/UnicodeJsps/character.jsp?a=1B43).
 //!       - [Characters](https://util.unicode.org/UnicodeJsps/list-unicodeset.jsp?a=%5Cp%7BHangul_Syllable_Type%3DV%7D%5Cp%7BHangul_Syllable_Type%3DT%7D)
 //!         with a [`Hangul_Syllable_Type`] of `Vowel_Jamo` (`V`) or `Trailing_Jamo` (`T`).
 //!       - The following [`Prepended_Concatenation_Mark`]s:

Original file line number	Diff line number	Diff line change
`@@ -506,12 +506,6 @@ def load_zero_widths() -> list[bool]:`
`506`	`506`	`lambda cp: operator.setitem(zw_map, cp, True),`
`507`	`507`	`)`
`508`	`508`
`509`		- # Unicode spec bug: these should be `Grapheme_Cluster_Break=Extend`,
`510`		`- # as they canonically decompose to two characters with this property,`
`511`		`- # but they aren't.`
`512`		`- for c in [0x0CC0, 0x0CC7, 0x0CC8, 0x0CCA, 0x0CCB, 0x1B3B, 0x1B3D, 0x1B43]:`
`513`		`- zw_map[c] = True`
`514`		`-`
`515`	`509`	# Treat `Hangul_Syllable_Type`s of `Vowel_Jamo` and `Trailing_Jamo`
`516`	`510`	# as zero-width. This matches the behavior of glibc `wcwidth`.
`517`	`511`	`#`