Skip to content

Update to Unicode 18 #96

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Manishearth merged 1 commit into from
Sep 16, 2026
+245 −72
Merged

Update to Unicode 18 #96

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions scripts/unicode.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@
from itertools import batched
from typing import Callable, Iterable, IO

UNICODE_VERSION = "17.0.0"
UNICODE_VERSION = "18.0.0"
"""The version of the Unicode data files to download."""

NUM_CODEPOINTS = 0x110000
Expand Down Expand Up @@ -138,7 +138,7 @@ def fetch_open(filename: str, local_prefix: str = "", emoji: bool = False):
def load_unicode_version() -> tuple[int, int, int]:
"""Returns the current Unicode version by fetching and processing `ReadMe.txt`."""
with fetch_open("ReadMe.txt") as readme:
pattern = r"for Version (\d+)\.(\d+)\.(\d+) of the Unicode"
pattern = r"for version (\d+)\.(\d+)\.(\d+) of the\nUnicode"
return tuple(map(int, re.search(pattern, readme.read()).groups())) # type: ignore


Expand Down
117 changes: 67 additions & 50 deletions src/gen/tables.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@

/// The version of [Unicode](http://www.unicode.org/)
/// that this version of unicode-width is based on.
pub const UNICODE_VERSION: (u8, u8, u8) = (17, 0, 0);
pub const UNICODE_VERSION: (u8, u8, u8) = (18, 0, 0);

#[repr(align(32))]
pub(crate) struct Align32<T>(pub(crate) T);
Expand Down Expand Up @@ -140,38 +140,38 @@ pub(crate) static WIDTH_MIDDLE: Align64<[[u8; 64]; WIDTH_MIDDLE_LEN]> = Align64(
],
[
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x7B, 0x7C, 0x7D, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x7B, 0x7C, 0x7D, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x7E, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02,
],
[
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x7E, 0x39, 0x39, 0x7F, 0x39, 0x39, 0x80, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x81, 0x02, 0x02,
0x02, 0x7F, 0x39, 0x39, 0x80, 0x39, 0x39, 0x81, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x82, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02,
],
[
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x82, 0x02, 0x02, 0x02, 0x83, 0x84, 0x85, 0x02, 0x86, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x87, 0x88, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x83, 0x02, 0x02, 0x02, 0x84, 0x85, 0x86, 0x87, 0x88, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x89, 0x8A, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02,
],
[
0x89, 0x8A, 0x77, 0x02, 0x02, 0x8B, 0x02, 0x02, 0x02, 0x8C, 0x02, 0x8D, 0x02, 0x8E, 0x02,
0x02, 0x02, 0x8F, 0x90, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x91, 0x92, 0x02, 0x93, 0x94, 0x02, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9A, 0x9B,
0x9C, 0x02, 0x9D, 0x02, 0x02, 0x9E, 0x9F, 0xA0, 0xA1, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x8B, 0x8C, 0x77, 0x02, 0x02, 0x8D, 0x02, 0x02, 0x02, 0x8E, 0x02, 0x8F, 0x02, 0x90, 0x02,
0x02, 0x02, 0x91, 0x92, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x93, 0x94, 0x02, 0x95, 0x96, 0x02, 0x97, 0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D,
0x9E, 0x02, 0x9F, 0x02, 0x02, 0xA0, 0xA1, 0xA2, 0xA3, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02,
],
[
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0xA2,
0x39, 0x39, 0x39, 0xA4,
],
[
0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D, 0x1D,
Expand All @@ -182,17 +182,17 @@ pub(crate) static WIDTH_MIDDLE: Align64<[[u8; 64]; WIDTH_MIDDLE_LEN]> = Align64(
],
#[cfg(feature = "cjk")]
[
0x00, 0xA3, 0x02, 0x02, 0x02, 0x02, 0xA4, 0xA5, 0x02, 0x04, 0x02, 0x05, 0x06, 0x07, 0x08,
0x00, 0xA5, 0x02, 0x02, 0x02, 0x02, 0xA6, 0xA7, 0x02, 0x04, 0x02, 0x05, 0x06, 0x07, 0x08,
0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x02, 0x02, 0x1E, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x1F, 0x20, 0x21, 0x22, 0x23, 0x02, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x02, 0x2A,
0x02, 0x02, 0x02, 0x02,
],
#[cfg(feature = "cjk")]
[
0xA6, 0xA7, 0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0x2E, 0xAD, 0x39, 0xAE, 0xAF, 0xB0, 0xB1, 0xB2,
0xB3, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0xB4, 0x02, 0x02, 0x35, 0x36, 0x37, 0x02, 0x38,
0x39, 0x3A, 0x3B, 0x3C, 0x3D, 0x3E, 0xB5, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0x2E, 0xAF, 0x39, 0xB0, 0xB1, 0xB2, 0xB3, 0xB4,
0xB5, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0xB6, 0x02, 0x02, 0x35, 0x36, 0x37, 0x02, 0x38,
0x39, 0x3A, 0x3B, 0x3C, 0x3D, 0x3E, 0xB7, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39,
],
Expand All @@ -202,23 +202,23 @@ pub(crate) static WIDTH_MIDDLE: Align64<[[u8; 64]; WIDTH_MIDDLE_LEN]> = Align64(
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39,
0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x39, 0x4B, 0x02, 0x02, 0x02, 0x02, 0x02,
0xB6, 0x4D, 0x4E, 0xB7,
0xB8, 0x4D, 0x4E, 0xB9,
],
#[cfg(feature = "cjk")]
[
0x89, 0x8A, 0x77, 0x02, 0x02, 0x8B, 0x02, 0x02, 0x02, 0x8C, 0x02, 0x8D, 0x02, 0x8E, 0x02,
0x02, 0x02, 0x8F, 0x90, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x91, 0x92, 0xB8, 0xB9, 0x94, 0x02, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9A, 0x9B,
0x9C, 0x02, 0x9D, 0x02, 0x02, 0x9E, 0x9F, 0xA0, 0xA1, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x8B, 0x8C, 0x77, 0x02, 0x02, 0x8D, 0x02, 0x02, 0x02, 0x8E, 0x02, 0x8F, 0x02, 0x90, 0x02,
0x02, 0x02, 0x91, 0x92, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x93, 0x94, 0xBA, 0xBB, 0x96, 0x02, 0x97, 0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D,
0x9E, 0x02, 0x9F, 0x02, 0x02, 0xA0, 0xA1, 0xA2, 0xA3, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02,
],
]);

#[cfg(feature = "cjk")]
pub(crate) const WIDTH_LEAVES_LEN: usize = 186;
pub(crate) const WIDTH_LEAVES_LEN: usize = 188;
#[cfg(not(feature = "cjk"))]
pub(crate) const WIDTH_LEAVES_LEN: usize = 163;
/// Autogenerated. 186 sub-table(s). Consult [`lookup_width`] for layout info.
pub(crate) const WIDTH_LEAVES_LEN: usize = 165;
/// Autogenerated. 188 sub-table(s). Consult [`lookup_width`] for layout info.
pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32([
[
0x55, 0x55, 0x75, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
Expand Down Expand Up @@ -247,7 +247,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0x55, 0x55, 0x55, 0x55, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x10, 0x41, 0x10, 0x55, 0x55, 0x55, 0x55, 0x55, 0x57, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x10, 0x41, 0x10, 0x50, 0x55, 0x55, 0x55, 0x55, 0x57, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55,
],
[
Expand Down Expand Up @@ -302,7 +302,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0x51, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x04, 0x01, 0x54, 0x55, 0x51, 0x55, 0x01, 0x55, 0x55, 0x05, 0x55, 0x55, 0x55, 0x55, 0x55,
0x04, 0x01, 0x54, 0x55, 0x51, 0x15, 0x00, 0x55, 0x55, 0x05, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55,
],
[
Expand Down Expand Up @@ -407,7 +407,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x50, 0x00, 0x00, 0x00, 0x55, 0x55, 0x55,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x54, 0x55,
0x55, 0x55,
],
[
Expand Down Expand Up @@ -657,8 +657,8 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x15, 0x54, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x05, 0x00,
0x55, 0x55, 0x55, 0x15, 0x00, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x00, 0x00,
0x00, 0x00,
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
Expand Down Expand Up @@ -787,7 +787,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0x55, 0x55, 0x55, 0x55, 0x50, 0x11, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x54, 0x55,
0x55, 0x55,
],
[
Expand Down Expand Up @@ -837,11 +837,11 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA,
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x5A, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x6A, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x95,
],
[
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x6A, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x56, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55,
],
Expand All @@ -850,14 +850,19 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x6A, 0x55,
0x55, 0x55,
],
[
0xAA, 0xAA, 0xAA, 0xAA, 0x5A, 0x55, 0x55, 0x55, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA,
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x6A, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55,
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0xAA, 0xA9,
0xAA, 0x69,
],
[
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x6A, 0x55, 0x55, 0x55, 0x65, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x6A, 0x59, 0x55, 0x55, 0x55, 0xAA, 0x55, 0x55, 0xAA, 0xAA,
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0x56, 0x55, 0x65, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x6A, 0x59, 0x55, 0x55, 0x55, 0xAA, 0x56, 0x55, 0xAA, 0xAA,
0xAA, 0xAA,
],
[
Expand All @@ -876,7 +881,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
0x55, 0x55,
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x15, 0x54, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x01, 0x50, 0x01, 0x00, 0x00,
0x00, 0x00,
],
Expand All @@ -887,7 +892,12 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x05, 0x54, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x05, 0x54, 0x55, 0x55, 0x40, 0x55, 0x15, 0x14, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55,
],
[
0x50, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55,
],
[
Expand Down Expand Up @@ -956,7 +966,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
0x55, 0x55,
],
[
0x55, 0x55, 0x55, 0x75, 0xFD, 0xFF, 0x7F, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x75, 0xFD, 0xFF, 0x7F, 0x55, 0x55, 0x55, 0x55, 0x65, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0xF5, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF,
],
Expand Down Expand Up @@ -1002,12 +1012,12 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
[
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0x5F, 0x55, 0x57, 0x7F, 0xFD, 0x57, 0xFF, 0x55, 0x55, 0xD5, 0x57, 0x55, 0xFF,
0xFF, 0xFF, 0x5F, 0x55, 0x57, 0x7F, 0xFD, 0x5F, 0xFF, 0x55, 0x55, 0xD5, 0x57, 0x55, 0xFF,
0xFF, 0x57,
],
[
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0xFF, 0xFF, 0xFF, 0x55, 0x57, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x65, 0x55, 0xFF, 0xFF, 0xFF, 0x55, 0x57, 0x55,
0x55, 0x55,
],
[
Expand All @@ -1026,9 +1036,9 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
0xFF, 0x57,
],
[
0xFF, 0xFF, 0x7F, 0xF5, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0x7F, 0x57, 0xFD, 0xFF, 0xFF, 0xFF, 0xD7, 0xFF, 0xFF, 0x7F, 0xD5, 0xFF, 0xFF,
0x57, 0x55,
0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0x7F, 0x57, 0xFF, 0xFF, 0xFF, 0xFF, 0xDF, 0xFF, 0xFF, 0xFF, 0xD5, 0xFF, 0xFF,
0x7F, 0x55,
],
[
0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA, 0xAA,
Expand Down Expand Up @@ -1169,7 +1179,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
],
#[cfg(feature = "cjk")]
[
0xAA, 0xAA, 0xAA, 0xBA, 0xFE, 0xFF, 0xBF, 0xAA, 0xAA, 0xAA, 0xAA, 0x56, 0x55, 0x55, 0x55,
0xAA, 0xAA, 0xAA, 0xBA, 0xFE, 0xFF, 0xBF, 0xAA, 0xAA, 0xAA, 0xAA, 0x66, 0x55, 0x55, 0x55,
0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0x55, 0xF5, 0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF,
],
Expand All @@ -1178,7 +1188,7 @@ pub(crate) static WIDTH_LEAVES: Align32<[[u8; 32]; WIDTH_LEAVES_LEN]> = Align32(
/// Sorted list of codepoint ranges (inclusive)
/// that are zero-width but not `Joining_Type=Transparent`
/// FIXME: can we get better compression?
pub(crate) static NON_TRANSPARENT_ZERO_WIDTHS: [([u8; 3], [u8; 3]); 71] = [
pub(crate) static NON_TRANSPARENT_ZERO_WIDTHS: [([u8; 3], [u8; 3]); 74] = [
([0x05, 0x06, 0x00], [0x05, 0x06, 0x00]),
([0x90, 0x08, 0x00], [0x91, 0x08, 0x00]),
([0xE2, 0x08, 0x00], [0xE2, 0x08, 0x00]),
Expand Down Expand Up @@ -1246,6 +1256,9 @@ pub(crate) static NON_TRANSPARENT_ZERO_WIDTHS: [([u8; 3], [u8; 3]); 71] = [
([0xF0, 0x6F, 0x01], [0xF1, 0x6F, 0x01]),
([0x65, 0xD1, 0x01], [0x66, 0xD1, 0x01]),
([0x6D, 0xD1, 0x01], [0x72, 0xD1, 0x01]),
([0x50, 0xD2, 0x01], [0x52, 0xD2, 0x01]),
([0x5F, 0xD2, 0x01], [0x5F, 0xD2, 0x01]),
([0x80, 0xD2, 0x01], [0x81, 0xD2, 0x01]),
([0x00, 0x00, 0x0E], [0x00, 0x00, 0x0E]),
([0x02, 0x00, 0x0E], [0x1F, 0x00, 0x0E]),
([0x80, 0x00, 0x0E], [0xFF, 0x00, 0x0E]),
Expand All @@ -1257,15 +1270,15 @@ pub(crate) static NON_TRANSPARENT_ZERO_WIDTHS: [([u8; 3], [u8; 3]); 71] = [
/// (mostly ccc > 1).
/// FIXME: can we get better compression?
#[cfg(feature = "cjk")]
pub(crate) static SOLIDUS_TRANSPARENT: [([u8; 3], [u8; 3]); 208] = [
pub(crate) static SOLIDUS_TRANSPARENT: [([u8; 3], [u8; 3]); 212] = [
([0x00, 0x03, 0x00], [0x33, 0x03, 0x00]),
([0x39, 0x03, 0x00], [0x6F, 0x03, 0x00]),
([0x83, 0x04, 0x00], [0x87, 0x04, 0x00]),
([0x91, 0x05, 0x00], [0xBD, 0x05, 0x00]),
([0xBF, 0x05, 0x00], [0xBF, 0x05, 0x00]),
([0xC1, 0x05, 0x00], [0xC2, 0x05, 0x00]),
([0xC4, 0x05, 0x00], [0xC5, 0x05, 0x00]),
([0xC7, 0x05, 0x00], [0xC7, 0x05, 0x00]),
([0xC7, 0x05, 0x00], [0xC9, 0x05, 0x00]),
([0x10, 0x06, 0x00], [0x1A, 0x06, 0x00]),
([0x4B, 0x06, 0x00], [0x5F, 0x06, 0x00]),
([0x70, 0x06, 0x00], [0x70, 0x06, 0x00]),
Expand Down Expand Up @@ -1334,8 +1347,7 @@ pub(crate) static SOLIDUS_TRANSPARENT: [([u8; 3], [u8; 3]); 208] = [
([0x75, 0x1A, 0x00], [0x7C, 0x1A, 0x00]),
([0x7F, 0x1A, 0x00], [0x7F, 0x1A, 0x00]),
([0xB0, 0x1A, 0x00], [0xBD, 0x1A, 0x00]),
([0xBF, 0x1A, 0x00], [0xDD, 0x1A, 0x00]),
([0xE0, 0x1A, 0x00], [0xEB, 0x1A, 0x00]),
([0xBF, 0x1A, 0x00], [0xF0, 0x1A, 0x00]),
([0x34, 0x1B, 0x00], [0x34, 0x1B, 0x00]),
([0x44, 0x1B, 0x00], [0x44, 0x1B, 0x00]),
([0x6B, 0x1B, 0x00], [0x73, 0x1B, 0x00]),
Expand Down Expand Up @@ -1398,7 +1410,8 @@ pub(crate) static SOLIDUS_TRANSPARENT: [([u8; 3], [u8; 3]); 208] = [
([0x69, 0x0D, 0x01], [0x6D, 0x0D, 0x01]),
([0x74, 0x0E, 0x01], [0x74, 0x0E, 0x01]),
([0xAB, 0x0E, 0x01], [0xAC, 0x0E, 0x01]),
([0xFA, 0x0E, 0x01], [0xFB, 0x0E, 0x01]),
([0xCB, 0x0E, 0x01], [0xCF, 0x0E, 0x01]),
([0xF0, 0x0E, 0x01], [0xFB, 0x0E, 0x01]),
([0xFD, 0x0E, 0x01], [0xFF, 0x0E, 0x01]),
([0x46, 0x0F, 0x01], [0x50, 0x0F, 0x01]),
([0x82, 0x0F, 0x01], [0x85, 0x0F, 0x01]),
Expand Down Expand Up @@ -1441,12 +1454,16 @@ pub(crate) static SOLIDUS_TRANSPARENT: [([u8; 3], [u8; 3]); 208] = [
([0x2F, 0x61, 0x01], [0x2F, 0x61, 0x01]),
([0x30, 0x6B, 0x01], [0x36, 0x6B, 0x01]),
([0xF0, 0x6F, 0x01], [0xF1, 0x6F, 0x01]),
([0x27, 0xD1, 0x01], [0x28, 0xD1, 0x01]),
([0x65, 0xD1, 0x01], [0x66, 0xD1, 0x01]),
([0x6D, 0xD1, 0x01], [0x72, 0xD1, 0x01]),
([0x7B, 0xD1, 0x01], [0x82, 0xD1, 0x01]),
([0x85, 0xD1, 0x01], [0x8B, 0xD1, 0x01]),
([0xAA, 0xD1, 0x01], [0xAD, 0xD1, 0x01]),
([0x42, 0xD2, 0x01], [0x44, 0xD2, 0x01]),
([0x50, 0xD2, 0x01], [0x52, 0xD2, 0x01]),
([0x5F, 0xD2, 0x01], [0x5F, 0xD2, 0x01]),
([0x80, 0xD2, 0x01], [0x81, 0xD2, 0x01]),
([0x00, 0xE0, 0x01], [0x06, 0xE0, 0x01]),
([0x08, 0xE0, 0x01], [0x18, 0xE0, 0x01]),
([0x1B, 0xE0, 0x01], [0x21, 0xE0, 0x01]),
Expand Down Expand Up @@ -1742,5 +1759,5 @@ pub(crate) static EMOJI_MODIFIER_LEAF_6: [(u8, u8); 12] = [
#[rustfmt::skip]
pub(crate) static EMOJI_MODIFIER_LEAF_7: [(u8, u8); 2] = [
(0xC3, 0xC5),
(0xF0, 0xF8),
(0xF0, 0xFA),
];
Loading
Loading