diff options
| author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:02:58 +0000 |
|---|---|---|
| committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:02:58 +0000 |
| commit | 698f8c2f01ea549d77d7dc3338a12e04c11057b9 (patch) | |
| tree | 173a775858bd501c378080a10dca74132f05bc50 /vendor/unicode-bidi/src | |
| parent | Initial commit. (diff) | |
| download | rustc-698f8c2f01ea549d77d7dc3338a12e04c11057b9.tar.xz rustc-698f8c2f01ea549d77d7dc3338a12e04c11057b9.zip | |
Adding upstream version 1.64.0+dfsg1.upstream/1.64.0+dfsg1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'vendor/unicode-bidi/src')
| -rw-r--r-- | vendor/unicode-bidi/src/char_data/mod.rs | 157 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/char_data/tables.rs | 510 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/data_source.rs | 16 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/deprecated.rs | 92 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/explicit.rs | 195 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/format_chars.rs | 42 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/implicit.rs | 239 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/level.rs | 385 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/lib.rs | 1066 | ||||
| -rw-r--r-- | vendor/unicode-bidi/src/prepare.rs | 368 |
10 files changed, 3070 insertions, 0 deletions
diff --git a/vendor/unicode-bidi/src/char_data/mod.rs b/vendor/unicode-bidi/src/char_data/mod.rs new file mode 100644 index 000000000..a8b452788 --- /dev/null +++ b/vendor/unicode-bidi/src/char_data/mod.rs @@ -0,0 +1,157 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! Accessor for `Bidi_Class` property from Unicode Character Database (UCD) + +mod tables; + +pub use self::tables::{BidiClass, UNICODE_VERSION}; +#[cfg(feature = "hardcoded-data")] +use core::char; +#[cfg(feature = "hardcoded-data")] +use core::cmp::Ordering::{Equal, Greater, Less}; + +#[cfg(feature = "hardcoded-data")] +use self::tables::bidi_class_table; +use crate::BidiClass::*; +#[cfg(feature = "hardcoded-data")] +use crate::BidiDataSource; + +/// Hardcoded Bidi data that ships with the unicode-bidi crate. +/// +/// This can be enabled with the default `hardcoded-data` Cargo feature. +#[cfg(feature = "hardcoded-data")] +pub struct HardcodedBidiData; + +#[cfg(feature = "hardcoded-data")] +impl BidiDataSource for HardcodedBidiData { + fn bidi_class(&self, c: char) -> BidiClass { + bsearch_range_value_table(c, bidi_class_table) + } +} + +/// Find the `BidiClass` of a single char. +#[cfg(feature = "hardcoded-data")] +pub fn bidi_class(c: char) -> BidiClass { + bsearch_range_value_table(c, bidi_class_table) +} + +pub fn is_rtl(bidi_class: BidiClass) -> bool { + match bidi_class { + RLE | RLO | RLI => true, + _ => false, + } +} + +#[cfg(feature = "hardcoded-data")] +fn bsearch_range_value_table(c: char, r: &'static [(char, char, BidiClass)]) -> BidiClass { + match r.binary_search_by(|&(lo, hi, _)| { + if lo <= c && c <= hi { + Equal + } else if hi < c { + Less + } else { + Greater + } + }) { + Ok(idx) => { + let (_, _, cat) = r[idx]; + cat + } + // UCD/extracted/DerivedBidiClass.txt: "All code points not explicitly listed + // for Bidi_Class have the value Left_To_Right (L)." + Err(_) => L, + } +} + +#[cfg(all(test, feature = "hardcoded-data"))] +mod tests { + use super::*; + + #[test] + fn test_ascii() { + assert_eq!(bidi_class('\u{0000}'), BN); + assert_eq!(bidi_class('\u{0040}'), ON); + assert_eq!(bidi_class('\u{0041}'), L); + assert_eq!(bidi_class('\u{0062}'), L); + assert_eq!(bidi_class('\u{007F}'), BN); + } + + #[test] + fn test_bmp() { + // Hebrew + assert_eq!(bidi_class('\u{0590}'), R); + assert_eq!(bidi_class('\u{05D0}'), R); + assert_eq!(bidi_class('\u{05D1}'), R); + assert_eq!(bidi_class('\u{05FF}'), R); + + // Arabic + assert_eq!(bidi_class('\u{0600}'), AN); + assert_eq!(bidi_class('\u{0627}'), AL); + assert_eq!(bidi_class('\u{07BF}'), AL); + + // Default R + Arabic Extras + assert_eq!(bidi_class('\u{07C0}'), R); + assert_eq!(bidi_class('\u{085F}'), R); + assert_eq!(bidi_class('\u{0860}'), AL); + assert_eq!(bidi_class('\u{0870}'), AL); + assert_eq!(bidi_class('\u{089F}'), NSM); + assert_eq!(bidi_class('\u{08A0}'), AL); + assert_eq!(bidi_class('\u{089F}'), NSM); + assert_eq!(bidi_class('\u{08FF}'), NSM); + + // Default ET + assert_eq!(bidi_class('\u{20A0}'), ET); + assert_eq!(bidi_class('\u{20CF}'), ET); + + // Arabic Presentation Forms + assert_eq!(bidi_class('\u{FB1D}'), R); + assert_eq!(bidi_class('\u{FB4F}'), R); + assert_eq!(bidi_class('\u{FB50}'), AL); + assert_eq!(bidi_class('\u{FDCF}'), ON); + assert_eq!(bidi_class('\u{FDF0}'), AL); + assert_eq!(bidi_class('\u{FDFF}'), ON); + assert_eq!(bidi_class('\u{FE70}'), AL); + assert_eq!(bidi_class('\u{FEFE}'), AL); + assert_eq!(bidi_class('\u{FEFF}'), BN); + + // noncharacters + assert_eq!(bidi_class('\u{FDD0}'), L); + assert_eq!(bidi_class('\u{FDD1}'), L); + assert_eq!(bidi_class('\u{FDEE}'), L); + assert_eq!(bidi_class('\u{FDEF}'), L); + assert_eq!(bidi_class('\u{FFFE}'), L); + assert_eq!(bidi_class('\u{FFFF}'), L); + } + + #[test] + fn test_smp() { + // Default AL + R + assert_eq!(bidi_class('\u{10800}'), R); + assert_eq!(bidi_class('\u{10FFF}'), R); + assert_eq!(bidi_class('\u{1E800}'), R); + assert_eq!(bidi_class('\u{1EDFF}'), R); + assert_eq!(bidi_class('\u{1EE00}'), AL); + assert_eq!(bidi_class('\u{1EEFF}'), AL); + assert_eq!(bidi_class('\u{1EF00}'), R); + assert_eq!(bidi_class('\u{1EFFF}'), R); + } + + #[test] + fn test_unassigned_planes() { + assert_eq!(bidi_class('\u{30000}'), L); + assert_eq!(bidi_class('\u{40000}'), L); + assert_eq!(bidi_class('\u{50000}'), L); + assert_eq!(bidi_class('\u{60000}'), L); + assert_eq!(bidi_class('\u{70000}'), L); + assert_eq!(bidi_class('\u{80000}'), L); + assert_eq!(bidi_class('\u{90000}'), L); + assert_eq!(bidi_class('\u{a0000}'), L); + } +} diff --git a/vendor/unicode-bidi/src/char_data/tables.rs b/vendor/unicode-bidi/src/char_data/tables.rs new file mode 100644 index 000000000..502ae9e90 --- /dev/null +++ b/vendor/unicode-bidi/src/char_data/tables.rs @@ -0,0 +1,510 @@ +// NOTE: +// The following code was generated by "tools/generate.py". do not edit directly + +#![allow(missing_docs, non_upper_case_globals, non_snake_case)] +#![cfg_attr(rustfmt, rustfmt_skip)] + +/// The [Unicode version](http://www.unicode.org/versions/) of data +pub const UNICODE_VERSION: (u64, u64, u64) = (14, 0, 0); + +#[allow(non_camel_case_types)] +#[derive(Clone, Copy, Debug, PartialEq, Eq)] +/// Represents values of the Unicode character property +/// [`Bidi_Class`](http://www.unicode.org/reports/tr44/#Bidi_Class), also +/// known as the *bidirectional character type*. +/// +/// * <http://www.unicode.org/reports/tr9/#Bidirectional_Character_Types> +/// * <http://www.unicode.org/reports/tr44/#Bidi_Class_Values> +pub enum BidiClass { + AL, + AN, + B, + BN, + CS, + EN, + ES, + ET, + FSI, + L, + LRE, + LRI, + LRO, + NSM, + ON, + PDF, + PDI, + R, + RLE, + RLI, + RLO, + S, + WS, +} + +#[cfg(feature = "hardcoded-data")] +use self::BidiClass::*; + +#[cfg(feature = "hardcoded-data")] +pub const bidi_class_table: &'static [(char, char, BidiClass)] = &[ + ('\u{0}', '\u{8}', BN), ('\u{9}', '\u{9}', S), ('\u{a}', '\u{a}', B), ('\u{b}', '\u{b}', S), + ('\u{c}', '\u{c}', WS), ('\u{d}', '\u{d}', B), ('\u{e}', '\u{1b}', BN), ('\u{1c}', '\u{1e}', B), + ('\u{1f}', '\u{1f}', S), ('\u{20}', '\u{20}', WS), ('\u{21}', '\u{22}', ON), ('\u{23}', + '\u{25}', ET), ('\u{26}', '\u{2a}', ON), ('\u{2b}', '\u{2b}', ES), ('\u{2c}', '\u{2c}', CS), + ('\u{2d}', '\u{2d}', ES), ('\u{2e}', '\u{2f}', CS), ('\u{30}', '\u{39}', EN), ('\u{3a}', + '\u{3a}', CS), ('\u{3b}', '\u{40}', ON), ('\u{41}', '\u{5a}', L), ('\u{5b}', '\u{60}', ON), + ('\u{61}', '\u{7a}', L), ('\u{7b}', '\u{7e}', ON), ('\u{7f}', '\u{84}', BN), ('\u{85}', + '\u{85}', B), ('\u{86}', '\u{9f}', BN), ('\u{a0}', '\u{a0}', CS), ('\u{a1}', '\u{a1}', ON), + ('\u{a2}', '\u{a5}', ET), ('\u{a6}', '\u{a9}', ON), ('\u{aa}', '\u{aa}', L), ('\u{ab}', + '\u{ac}', ON), ('\u{ad}', '\u{ad}', BN), ('\u{ae}', '\u{af}', ON), ('\u{b0}', '\u{b1}', ET), + ('\u{b2}', '\u{b3}', EN), ('\u{b4}', '\u{b4}', ON), ('\u{b5}', '\u{b5}', L), ('\u{b6}', + '\u{b8}', ON), ('\u{b9}', '\u{b9}', EN), ('\u{ba}', '\u{ba}', L), ('\u{bb}', '\u{bf}', ON), + ('\u{c0}', '\u{d6}', L), ('\u{d7}', '\u{d7}', ON), ('\u{d8}', '\u{f6}', L), ('\u{f7}', '\u{f7}', + ON), ('\u{f8}', '\u{2b8}', L), ('\u{2b9}', '\u{2ba}', ON), ('\u{2bb}', '\u{2c1}', L), + ('\u{2c2}', '\u{2cf}', ON), ('\u{2d0}', '\u{2d1}', L), ('\u{2d2}', '\u{2df}', ON), ('\u{2e0}', + '\u{2e4}', L), ('\u{2e5}', '\u{2ed}', ON), ('\u{2ee}', '\u{2ee}', L), ('\u{2ef}', '\u{2ff}', + ON), ('\u{300}', '\u{36f}', NSM), ('\u{370}', '\u{373}', L), ('\u{374}', '\u{375}', ON), + ('\u{376}', '\u{377}', L), ('\u{37a}', '\u{37d}', L), ('\u{37e}', '\u{37e}', ON), ('\u{37f}', + '\u{37f}', L), ('\u{384}', '\u{385}', ON), ('\u{386}', '\u{386}', L), ('\u{387}', '\u{387}', + ON), ('\u{388}', '\u{38a}', L), ('\u{38c}', '\u{38c}', L), ('\u{38e}', '\u{3a1}', L), + ('\u{3a3}', '\u{3f5}', L), ('\u{3f6}', '\u{3f6}', ON), ('\u{3f7}', '\u{482}', L), ('\u{483}', + '\u{489}', NSM), ('\u{48a}', '\u{52f}', L), ('\u{531}', '\u{556}', L), ('\u{559}', '\u{589}', + L), ('\u{58a}', '\u{58a}', ON), ('\u{58d}', '\u{58e}', ON), ('\u{58f}', '\u{58f}', ET), + ('\u{590}', '\u{590}', R), ('\u{591}', '\u{5bd}', NSM), ('\u{5be}', '\u{5be}', R), ('\u{5bf}', + '\u{5bf}', NSM), ('\u{5c0}', '\u{5c0}', R), ('\u{5c1}', '\u{5c2}', NSM), ('\u{5c3}', '\u{5c3}', + R), ('\u{5c4}', '\u{5c5}', NSM), ('\u{5c6}', '\u{5c6}', R), ('\u{5c7}', '\u{5c7}', NSM), + ('\u{5c8}', '\u{5ff}', R), ('\u{600}', '\u{605}', AN), ('\u{606}', '\u{607}', ON), ('\u{608}', + '\u{608}', AL), ('\u{609}', '\u{60a}', ET), ('\u{60b}', '\u{60b}', AL), ('\u{60c}', '\u{60c}', + CS), ('\u{60d}', '\u{60d}', AL), ('\u{60e}', '\u{60f}', ON), ('\u{610}', '\u{61a}', NSM), + ('\u{61b}', '\u{64a}', AL), ('\u{64b}', '\u{65f}', NSM), ('\u{660}', '\u{669}', AN), ('\u{66a}', + '\u{66a}', ET), ('\u{66b}', '\u{66c}', AN), ('\u{66d}', '\u{66f}', AL), ('\u{670}', '\u{670}', + NSM), ('\u{671}', '\u{6d5}', AL), ('\u{6d6}', '\u{6dc}', NSM), ('\u{6dd}', '\u{6dd}', AN), + ('\u{6de}', '\u{6de}', ON), ('\u{6df}', '\u{6e4}', NSM), ('\u{6e5}', '\u{6e6}', AL), ('\u{6e7}', + '\u{6e8}', NSM), ('\u{6e9}', '\u{6e9}', ON), ('\u{6ea}', '\u{6ed}', NSM), ('\u{6ee}', '\u{6ef}', + AL), ('\u{6f0}', '\u{6f9}', EN), ('\u{6fa}', '\u{710}', AL), ('\u{711}', '\u{711}', NSM), + ('\u{712}', '\u{72f}', AL), ('\u{730}', '\u{74a}', NSM), ('\u{74b}', '\u{7a5}', AL), ('\u{7a6}', + '\u{7b0}', NSM), ('\u{7b1}', '\u{7bf}', AL), ('\u{7c0}', '\u{7ea}', R), ('\u{7eb}', '\u{7f3}', + NSM), ('\u{7f4}', '\u{7f5}', R), ('\u{7f6}', '\u{7f9}', ON), ('\u{7fa}', '\u{7fc}', R), + ('\u{7fd}', '\u{7fd}', NSM), ('\u{7fe}', '\u{815}', R), ('\u{816}', '\u{819}', NSM), ('\u{81a}', + '\u{81a}', R), ('\u{81b}', '\u{823}', NSM), ('\u{824}', '\u{824}', R), ('\u{825}', '\u{827}', + NSM), ('\u{828}', '\u{828}', R), ('\u{829}', '\u{82d}', NSM), ('\u{82e}', '\u{858}', R), + ('\u{859}', '\u{85b}', NSM), ('\u{85c}', '\u{85f}', R), ('\u{860}', '\u{86a}', AL), ('\u{86b}', + '\u{86f}', R), ('\u{870}', '\u{88e}', AL), ('\u{88f}', '\u{88f}', R), ('\u{890}', '\u{891}', + AN), ('\u{892}', '\u{897}', R), ('\u{898}', '\u{89f}', NSM), ('\u{8a0}', '\u{8c9}', AL), + ('\u{8ca}', '\u{8e1}', NSM), ('\u{8e2}', '\u{8e2}', AN), ('\u{8e3}', '\u{902}', NSM), + ('\u{903}', '\u{939}', L), ('\u{93a}', '\u{93a}', NSM), ('\u{93b}', '\u{93b}', L), ('\u{93c}', + '\u{93c}', NSM), ('\u{93d}', '\u{940}', L), ('\u{941}', '\u{948}', NSM), ('\u{949}', '\u{94c}', + L), ('\u{94d}', '\u{94d}', NSM), ('\u{94e}', '\u{950}', L), ('\u{951}', '\u{957}', NSM), + ('\u{958}', '\u{961}', L), ('\u{962}', '\u{963}', NSM), ('\u{964}', '\u{980}', L), ('\u{981}', + '\u{981}', NSM), ('\u{982}', '\u{983}', L), ('\u{985}', '\u{98c}', L), ('\u{98f}', '\u{990}', + L), ('\u{993}', '\u{9a8}', L), ('\u{9aa}', '\u{9b0}', L), ('\u{9b2}', '\u{9b2}', L), ('\u{9b6}', + '\u{9b9}', L), ('\u{9bc}', '\u{9bc}', NSM), ('\u{9bd}', '\u{9c0}', L), ('\u{9c1}', '\u{9c4}', + NSM), ('\u{9c7}', '\u{9c8}', L), ('\u{9cb}', '\u{9cc}', L), ('\u{9cd}', '\u{9cd}', NSM), + ('\u{9ce}', '\u{9ce}', L), ('\u{9d7}', '\u{9d7}', L), ('\u{9dc}', '\u{9dd}', L), ('\u{9df}', + '\u{9e1}', L), ('\u{9e2}', '\u{9e3}', NSM), ('\u{9e6}', '\u{9f1}', L), ('\u{9f2}', '\u{9f3}', + ET), ('\u{9f4}', '\u{9fa}', L), ('\u{9fb}', '\u{9fb}', ET), ('\u{9fc}', '\u{9fd}', L), + ('\u{9fe}', '\u{9fe}', NSM), ('\u{a01}', '\u{a02}', NSM), ('\u{a03}', '\u{a03}', L), ('\u{a05}', + '\u{a0a}', L), ('\u{a0f}', '\u{a10}', L), ('\u{a13}', '\u{a28}', L), ('\u{a2a}', '\u{a30}', L), + ('\u{a32}', '\u{a33}', L), ('\u{a35}', '\u{a36}', L), ('\u{a38}', '\u{a39}', L), ('\u{a3c}', + '\u{a3c}', NSM), ('\u{a3e}', '\u{a40}', L), ('\u{a41}', '\u{a42}', NSM), ('\u{a47}', '\u{a48}', + NSM), ('\u{a4b}', '\u{a4d}', NSM), ('\u{a51}', '\u{a51}', NSM), ('\u{a59}', '\u{a5c}', L), + ('\u{a5e}', '\u{a5e}', L), ('\u{a66}', '\u{a6f}', L), ('\u{a70}', '\u{a71}', NSM), ('\u{a72}', + '\u{a74}', L), ('\u{a75}', '\u{a75}', NSM), ('\u{a76}', '\u{a76}', L), ('\u{a81}', '\u{a82}', + NSM), ('\u{a83}', '\u{a83}', L), ('\u{a85}', '\u{a8d}', L), ('\u{a8f}', '\u{a91}', L), + ('\u{a93}', '\u{aa8}', L), ('\u{aaa}', '\u{ab0}', L), ('\u{ab2}', '\u{ab3}', L), ('\u{ab5}', + '\u{ab9}', L), ('\u{abc}', '\u{abc}', NSM), ('\u{abd}', '\u{ac0}', L), ('\u{ac1}', '\u{ac5}', + NSM), ('\u{ac7}', '\u{ac8}', NSM), ('\u{ac9}', '\u{ac9}', L), ('\u{acb}', '\u{acc}', L), + ('\u{acd}', '\u{acd}', NSM), ('\u{ad0}', '\u{ad0}', L), ('\u{ae0}', '\u{ae1}', L), ('\u{ae2}', + '\u{ae3}', NSM), ('\u{ae6}', '\u{af0}', L), ('\u{af1}', '\u{af1}', ET), ('\u{af9}', '\u{af9}', + L), ('\u{afa}', '\u{aff}', NSM), ('\u{b01}', '\u{b01}', NSM), ('\u{b02}', '\u{b03}', L), + ('\u{b05}', '\u{b0c}', L), ('\u{b0f}', '\u{b10}', L), ('\u{b13}', '\u{b28}', L), ('\u{b2a}', + '\u{b30}', L), ('\u{b32}', '\u{b33}', L), ('\u{b35}', '\u{b39}', L), ('\u{b3c}', '\u{b3c}', + NSM), ('\u{b3d}', '\u{b3e}', L), ('\u{b3f}', '\u{b3f}', NSM), ('\u{b40}', '\u{b40}', L), + ('\u{b41}', '\u{b44}', NSM), ('\u{b47}', '\u{b48}', L), ('\u{b4b}', '\u{b4c}', L), ('\u{b4d}', + '\u{b4d}', NSM), ('\u{b55}', '\u{b56}', NSM), ('\u{b57}', '\u{b57}', L), ('\u{b5c}', '\u{b5d}', + L), ('\u{b5f}', '\u{b61}', L), ('\u{b62}', '\u{b63}', NSM), ('\u{b66}', '\u{b77}', L), + ('\u{b82}', '\u{b82}', NSM), ('\u{b83}', '\u{b83}', L), ('\u{b85}', '\u{b8a}', L), ('\u{b8e}', + '\u{b90}', L), ('\u{b92}', '\u{b95}', L), ('\u{b99}', '\u{b9a}', L), ('\u{b9c}', '\u{b9c}', L), + ('\u{b9e}', '\u{b9f}', L), ('\u{ba3}', '\u{ba4}', L), ('\u{ba8}', '\u{baa}', L), ('\u{bae}', + '\u{bb9}', L), ('\u{bbe}', '\u{bbf}', L), ('\u{bc0}', '\u{bc0}', NSM), ('\u{bc1}', '\u{bc2}', + L), ('\u{bc6}', '\u{bc8}', L), ('\u{bca}', '\u{bcc}', L), ('\u{bcd}', '\u{bcd}', NSM), + ('\u{bd0}', '\u{bd0}', L), ('\u{bd7}', '\u{bd7}', L), ('\u{be6}', '\u{bf2}', L), ('\u{bf3}', + '\u{bf8}', ON), ('\u{bf9}', '\u{bf9}', ET), ('\u{bfa}', '\u{bfa}', ON), ('\u{c00}', '\u{c00}', + NSM), ('\u{c01}', '\u{c03}', L), ('\u{c04}', '\u{c04}', NSM), ('\u{c05}', '\u{c0c}', L), + ('\u{c0e}', '\u{c10}', L), ('\u{c12}', '\u{c28}', L), ('\u{c2a}', '\u{c39}', L), ('\u{c3c}', + '\u{c3c}', NSM), ('\u{c3d}', '\u{c3d}', L), ('\u{c3e}', '\u{c40}', NSM), ('\u{c41}', '\u{c44}', + L), ('\u{c46}', '\u{c48}', NSM), ('\u{c4a}', '\u{c4d}', NSM), ('\u{c55}', '\u{c56}', NSM), + ('\u{c58}', '\u{c5a}', L), ('\u{c5d}', '\u{c5d}', L), ('\u{c60}', '\u{c61}', L), ('\u{c62}', + '\u{c63}', NSM), ('\u{c66}', '\u{c6f}', L), ('\u{c77}', '\u{c77}', L), ('\u{c78}', '\u{c7e}', + ON), ('\u{c7f}', '\u{c80}', L), ('\u{c81}', '\u{c81}', NSM), ('\u{c82}', '\u{c8c}', L), + ('\u{c8e}', '\u{c90}', L), ('\u{c92}', '\u{ca8}', L), ('\u{caa}', '\u{cb3}', L), ('\u{cb5}', + '\u{cb9}', L), ('\u{cbc}', '\u{cbc}', NSM), ('\u{cbd}', '\u{cc4}', L), ('\u{cc6}', '\u{cc8}', + L), ('\u{cca}', '\u{ccb}', L), ('\u{ccc}', '\u{ccd}', NSM), ('\u{cd5}', '\u{cd6}', L), + ('\u{cdd}', '\u{cde}', L), ('\u{ce0}', '\u{ce1}', L), ('\u{ce2}', '\u{ce3}', NSM), ('\u{ce6}', + '\u{cef}', L), ('\u{cf1}', '\u{cf2}', L), ('\u{d00}', '\u{d01}', NSM), ('\u{d02}', '\u{d0c}', + L), ('\u{d0e}', '\u{d10}', L), ('\u{d12}', '\u{d3a}', L), ('\u{d3b}', '\u{d3c}', NSM), + ('\u{d3d}', '\u{d40}', L), ('\u{d41}', '\u{d44}', NSM), ('\u{d46}', '\u{d48}', L), ('\u{d4a}', + '\u{d4c}', L), ('\u{d4d}', '\u{d4d}', NSM), ('\u{d4e}', '\u{d4f}', L), ('\u{d54}', '\u{d61}', + L), ('\u{d62}', '\u{d63}', NSM), ('\u{d66}', '\u{d7f}', L), ('\u{d81}', '\u{d81}', NSM), + ('\u{d82}', '\u{d83}', L), ('\u{d85}', '\u{d96}', L), ('\u{d9a}', '\u{db1}', L), ('\u{db3}', + '\u{dbb}', L), ('\u{dbd}', '\u{dbd}', L), ('\u{dc0}', '\u{dc6}', L), ('\u{dca}', '\u{dca}', + NSM), ('\u{dcf}', '\u{dd1}', L), ('\u{dd2}', '\u{dd4}', NSM), ('\u{dd6}', '\u{dd6}', NSM), + ('\u{dd8}', '\u{ddf}', L), ('\u{de6}', '\u{def}', L), ('\u{df2}', '\u{df4}', L), ('\u{e01}', + '\u{e30}', L), ('\u{e31}', '\u{e31}', NSM), ('\u{e32}', '\u{e33}', L), ('\u{e34}', '\u{e3a}', + NSM), ('\u{e3f}', '\u{e3f}', ET), ('\u{e40}', '\u{e46}', L), ('\u{e47}', '\u{e4e}', NSM), + ('\u{e4f}', '\u{e5b}', L), ('\u{e81}', '\u{e82}', L), ('\u{e84}', '\u{e84}', L), ('\u{e86}', + '\u{e8a}', L), ('\u{e8c}', '\u{ea3}', L), ('\u{ea5}', '\u{ea5}', L), ('\u{ea7}', '\u{eb0}', L), + ('\u{eb1}', '\u{eb1}', NSM), ('\u{eb2}', '\u{eb3}', L), ('\u{eb4}', '\u{ebc}', NSM), ('\u{ebd}', + '\u{ebd}', L), ('\u{ec0}', '\u{ec4}', L), ('\u{ec6}', '\u{ec6}', L), ('\u{ec8}', '\u{ecd}', + NSM), ('\u{ed0}', '\u{ed9}', L), ('\u{edc}', '\u{edf}', L), ('\u{f00}', '\u{f17}', L), + ('\u{f18}', '\u{f19}', NSM), ('\u{f1a}', '\u{f34}', L), ('\u{f35}', '\u{f35}', NSM), ('\u{f36}', + '\u{f36}', L), ('\u{f37}', '\u{f37}', NSM), ('\u{f38}', '\u{f38}', L), ('\u{f39}', '\u{f39}', + NSM), ('\u{f3a}', '\u{f3d}', ON), ('\u{f3e}', '\u{f47}', L), ('\u{f49}', '\u{f6c}', L), + ('\u{f71}', '\u{f7e}', NSM), ('\u{f7f}', '\u{f7f}', L), ('\u{f80}', '\u{f84}', NSM), ('\u{f85}', + '\u{f85}', L), ('\u{f86}', '\u{f87}', NSM), ('\u{f88}', '\u{f8c}', L), ('\u{f8d}', '\u{f97}', + NSM), ('\u{f99}', '\u{fbc}', NSM), ('\u{fbe}', '\u{fc5}', L), ('\u{fc6}', '\u{fc6}', NSM), + ('\u{fc7}', '\u{fcc}', L), ('\u{fce}', '\u{fda}', L), ('\u{1000}', '\u{102c}', L), ('\u{102d}', + '\u{1030}', NSM), ('\u{1031}', '\u{1031}', L), ('\u{1032}', '\u{1037}', NSM), ('\u{1038}', + '\u{1038}', L), ('\u{1039}', '\u{103a}', NSM), ('\u{103b}', '\u{103c}', L), ('\u{103d}', + '\u{103e}', NSM), ('\u{103f}', '\u{1057}', L), ('\u{1058}', '\u{1059}', NSM), ('\u{105a}', + '\u{105d}', L), ('\u{105e}', '\u{1060}', NSM), ('\u{1061}', '\u{1070}', L), ('\u{1071}', + '\u{1074}', NSM), ('\u{1075}', '\u{1081}', L), ('\u{1082}', '\u{1082}', NSM), ('\u{1083}', + '\u{1084}', L), ('\u{1085}', '\u{1086}', NSM), ('\u{1087}', '\u{108c}', L), ('\u{108d}', + '\u{108d}', NSM), ('\u{108e}', '\u{109c}', L), ('\u{109d}', '\u{109d}', NSM), ('\u{109e}', + '\u{10c5}', L), ('\u{10c7}', '\u{10c7}', L), ('\u{10cd}', '\u{10cd}', L), ('\u{10d0}', + '\u{1248}', L), ('\u{124a}', '\u{124d}', L), ('\u{1250}', '\u{1256}', L), ('\u{1258}', + '\u{1258}', L), ('\u{125a}', '\u{125d}', L), ('\u{1260}', '\u{1288}', L), ('\u{128a}', + '\u{128d}', L), ('\u{1290}', '\u{12b0}', L), ('\u{12b2}', '\u{12b5}', L), ('\u{12b8}', + '\u{12be}', L), ('\u{12c0}', '\u{12c0}', L), ('\u{12c2}', '\u{12c5}', L), ('\u{12c8}', + '\u{12d6}', L), ('\u{12d8}', '\u{1310}', L), ('\u{1312}', '\u{1315}', L), ('\u{1318}', + '\u{135a}', L), ('\u{135d}', '\u{135f}', NSM), ('\u{1360}', '\u{137c}', L), ('\u{1380}', + '\u{138f}', L), ('\u{1390}', '\u{1399}', ON), ('\u{13a0}', '\u{13f5}', L), ('\u{13f8}', + '\u{13fd}', L), ('\u{1400}', '\u{1400}', ON), ('\u{1401}', '\u{167f}', L), ('\u{1680}', + '\u{1680}', WS), ('\u{1681}', '\u{169a}', L), ('\u{169b}', '\u{169c}', ON), ('\u{16a0}', + '\u{16f8}', L), ('\u{1700}', '\u{1711}', L), ('\u{1712}', '\u{1714}', NSM), ('\u{1715}', + '\u{1715}', L), ('\u{171f}', '\u{1731}', L), ('\u{1732}', '\u{1733}', NSM), ('\u{1734}', + '\u{1736}', L), ('\u{1740}', '\u{1751}', L), ('\u{1752}', '\u{1753}', NSM), ('\u{1760}', + '\u{176c}', L), ('\u{176e}', '\u{1770}', L), ('\u{1772}', '\u{1773}', NSM), ('\u{1780}', + '\u{17b3}', L), ('\u{17b4}', '\u{17b5}', NSM), ('\u{17b6}', '\u{17b6}', L), ('\u{17b7}', + '\u{17bd}', NSM), ('\u{17be}', '\u{17c5}', L), ('\u{17c6}', '\u{17c6}', NSM), ('\u{17c7}', + '\u{17c8}', L), ('\u{17c9}', '\u{17d3}', NSM), ('\u{17d4}', '\u{17da}', L), ('\u{17db}', + '\u{17db}', ET), ('\u{17dc}', '\u{17dc}', L), ('\u{17dd}', '\u{17dd}', NSM), ('\u{17e0}', + '\u{17e9}', L), ('\u{17f0}', '\u{17f9}', ON), ('\u{1800}', '\u{180a}', ON), ('\u{180b}', + '\u{180d}', NSM), ('\u{180e}', '\u{180e}', BN), ('\u{180f}', '\u{180f}', NSM), ('\u{1810}', + '\u{1819}', L), ('\u{1820}', '\u{1878}', L), ('\u{1880}', '\u{1884}', L), ('\u{1885}', + '\u{1886}', NSM), ('\u{1887}', '\u{18a8}', L), ('\u{18a9}', '\u{18a9}', NSM), ('\u{18aa}', + '\u{18aa}', L), ('\u{18b0}', '\u{18f5}', L), ('\u{1900}', '\u{191e}', L), ('\u{1920}', + '\u{1922}', NSM), ('\u{1923}', '\u{1926}', L), ('\u{1927}', '\u{1928}', NSM), ('\u{1929}', + '\u{192b}', L), ('\u{1930}', '\u{1931}', L), ('\u{1932}', '\u{1932}', NSM), ('\u{1933}', + '\u{1938}', L), ('\u{1939}', '\u{193b}', NSM), ('\u{1940}', '\u{1940}', ON), ('\u{1944}', + '\u{1945}', ON), ('\u{1946}', '\u{196d}', L), ('\u{1970}', '\u{1974}', L), ('\u{1980}', + '\u{19ab}', L), ('\u{19b0}', '\u{19c9}', L), ('\u{19d0}', '\u{19da}', L), ('\u{19de}', + '\u{19ff}', ON), ('\u{1a00}', '\u{1a16}', L), ('\u{1a17}', '\u{1a18}', NSM), ('\u{1a19}', + '\u{1a1a}', L), ('\u{1a1b}', '\u{1a1b}', NSM), ('\u{1a1e}', '\u{1a55}', L), ('\u{1a56}', + '\u{1a56}', NSM), ('\u{1a57}', '\u{1a57}', L), ('\u{1a58}', '\u{1a5e}', NSM), ('\u{1a60}', + '\u{1a60}', NSM), ('\u{1a61}', '\u{1a61}', L), ('\u{1a62}', '\u{1a62}', NSM), ('\u{1a63}', + '\u{1a64}', L), ('\u{1a65}', '\u{1a6c}', NSM), ('\u{1a6d}', '\u{1a72}', L), ('\u{1a73}', + '\u{1a7c}', NSM), ('\u{1a7f}', '\u{1a7f}', NSM), ('\u{1a80}', '\u{1a89}', L), ('\u{1a90}', + '\u{1a99}', L), ('\u{1aa0}', '\u{1aad}', L), ('\u{1ab0}', '\u{1ace}', NSM), ('\u{1b00}', + '\u{1b03}', NSM), ('\u{1b04}', '\u{1b33}', L), ('\u{1b34}', '\u{1b34}', NSM), ('\u{1b35}', + '\u{1b35}', L), ('\u{1b36}', '\u{1b3a}', NSM), ('\u{1b3b}', '\u{1b3b}', L), ('\u{1b3c}', + '\u{1b3c}', NSM), ('\u{1b3d}', '\u{1b41}', L), ('\u{1b42}', '\u{1b42}', NSM), ('\u{1b43}', + '\u{1b4c}', L), ('\u{1b50}', '\u{1b6a}', L), ('\u{1b6b}', '\u{1b73}', NSM), ('\u{1b74}', + '\u{1b7e}', L), ('\u{1b80}', '\u{1b81}', NSM), ('\u{1b82}', '\u{1ba1}', L), ('\u{1ba2}', + '\u{1ba5}', NSM), ('\u{1ba6}', '\u{1ba7}', L), ('\u{1ba8}', '\u{1ba9}', NSM), ('\u{1baa}', + '\u{1baa}', L), ('\u{1bab}', '\u{1bad}', NSM), ('\u{1bae}', '\u{1be5}', L), ('\u{1be6}', + '\u{1be6}', NSM), ('\u{1be7}', '\u{1be7}', L), ('\u{1be8}', '\u{1be9}', NSM), ('\u{1bea}', + '\u{1bec}', L), ('\u{1bed}', '\u{1bed}', NSM), ('\u{1bee}', '\u{1bee}', L), ('\u{1bef}', + '\u{1bf1}', NSM), ('\u{1bf2}', '\u{1bf3}', L), ('\u{1bfc}', '\u{1c2b}', L), ('\u{1c2c}', + '\u{1c33}', NSM), ('\u{1c34}', '\u{1c35}', L), ('\u{1c36}', '\u{1c37}', NSM), ('\u{1c3b}', + '\u{1c49}', L), ('\u{1c4d}', '\u{1c88}', L), ('\u{1c90}', '\u{1cba}', L), ('\u{1cbd}', + '\u{1cc7}', L), ('\u{1cd0}', '\u{1cd2}', NSM), ('\u{1cd3}', '\u{1cd3}', L), ('\u{1cd4}', + '\u{1ce0}', NSM), ('\u{1ce1}', '\u{1ce1}', L), ('\u{1ce2}', '\u{1ce8}', NSM), ('\u{1ce9}', + '\u{1cec}', L), ('\u{1ced}', '\u{1ced}', NSM), ('\u{1cee}', '\u{1cf3}', L), ('\u{1cf4}', + '\u{1cf4}', NSM), ('\u{1cf5}', '\u{1cf7}', L), ('\u{1cf8}', '\u{1cf9}', NSM), ('\u{1cfa}', + '\u{1cfa}', L), ('\u{1d00}', '\u{1dbf}', L), ('\u{1dc0}', '\u{1dff}', NSM), ('\u{1e00}', + '\u{1f15}', L), ('\u{1f18}', '\u{1f1d}', L), ('\u{1f20}', '\u{1f45}', L), ('\u{1f48}', + '\u{1f4d}', L), ('\u{1f50}', '\u{1f57}', L), ('\u{1f59}', '\u{1f59}', L), ('\u{1f5b}', + '\u{1f5b}', L), ('\u{1f5d}', '\u{1f5d}', L), ('\u{1f5f}', '\u{1f7d}', L), ('\u{1f80}', + '\u{1fb4}', L), ('\u{1fb6}', '\u{1fbc}', L), ('\u{1fbd}', '\u{1fbd}', ON), ('\u{1fbe}', + '\u{1fbe}', L), ('\u{1fbf}', '\u{1fc1}', ON), ('\u{1fc2}', '\u{1fc4}', L), ('\u{1fc6}', + '\u{1fcc}', L), ('\u{1fcd}', '\u{1fcf}', ON), ('\u{1fd0}', '\u{1fd3}', L), ('\u{1fd6}', + '\u{1fdb}', L), ('\u{1fdd}', '\u{1fdf}', ON), ('\u{1fe0}', '\u{1fec}', L), ('\u{1fed}', + '\u{1fef}', ON), ('\u{1ff2}', '\u{1ff4}', L), ('\u{1ff6}', '\u{1ffc}', L), ('\u{1ffd}', + '\u{1ffe}', ON), ('\u{2000}', '\u{200a}', WS), ('\u{200b}', '\u{200d}', BN), ('\u{200e}', + '\u{200e}', L), ('\u{200f}', '\u{200f}', R), ('\u{2010}', '\u{2027}', ON), ('\u{2028}', + '\u{2028}', WS), ('\u{2029}', '\u{2029}', B), ('\u{202a}', '\u{202a}', LRE), ('\u{202b}', + '\u{202b}', RLE), ('\u{202c}', '\u{202c}', PDF), ('\u{202d}', '\u{202d}', LRO), ('\u{202e}', + '\u{202e}', RLO), ('\u{202f}', '\u{202f}', CS), ('\u{2030}', '\u{2034}', ET), ('\u{2035}', + '\u{2043}', ON), ('\u{2044}', '\u{2044}', CS), ('\u{2045}', '\u{205e}', ON), ('\u{205f}', + '\u{205f}', WS), ('\u{2060}', '\u{2064}', BN), ('\u{2066}', '\u{2066}', LRI), ('\u{2067}', + '\u{2067}', RLI), ('\u{2068}', '\u{2068}', FSI), ('\u{2069}', '\u{2069}', PDI), ('\u{206a}', + '\u{206f}', BN), ('\u{2070}', '\u{2070}', EN), ('\u{2071}', '\u{2071}', L), ('\u{2074}', + '\u{2079}', EN), ('\u{207a}', '\u{207b}', ES), ('\u{207c}', '\u{207e}', ON), ('\u{207f}', + '\u{207f}', L), ('\u{2080}', '\u{2089}', EN), ('\u{208a}', '\u{208b}', ES), ('\u{208c}', + '\u{208e}', ON), ('\u{2090}', '\u{209c}', L), ('\u{20a0}', '\u{20cf}', ET), ('\u{20d0}', + '\u{20f0}', NSM), ('\u{2100}', '\u{2101}', ON), ('\u{2102}', '\u{2102}', L), ('\u{2103}', + '\u{2106}', ON), ('\u{2107}', '\u{2107}', L), ('\u{2108}', '\u{2109}', ON), ('\u{210a}', + '\u{2113}', L), ('\u{2114}', '\u{2114}', ON), ('\u{2115}', '\u{2115}', L), ('\u{2116}', + '\u{2118}', ON), ('\u{2119}', '\u{211d}', L), ('\u{211e}', '\u{2123}', ON), ('\u{2124}', + '\u{2124}', L), ('\u{2125}', '\u{2125}', ON), ('\u{2126}', '\u{2126}', L), ('\u{2127}', + '\u{2127}', ON), ('\u{2128}', '\u{2128}', L), ('\u{2129}', '\u{2129}', ON), ('\u{212a}', + '\u{212d}', L), ('\u{212e}', '\u{212e}', ET), ('\u{212f}', '\u{2139}', L), ('\u{213a}', + '\u{213b}', ON), ('\u{213c}', '\u{213f}', L), ('\u{2140}', '\u{2144}', ON), ('\u{2145}', + '\u{2149}', L), ('\u{214a}', '\u{214d}', ON), ('\u{214e}', '\u{214f}', L), ('\u{2150}', + '\u{215f}', ON), ('\u{2160}', '\u{2188}', L), ('\u{2189}', '\u{218b}', ON), ('\u{2190}', + '\u{2211}', ON), ('\u{2212}', '\u{2212}', ES), ('\u{2213}', '\u{2213}', ET), ('\u{2214}', + '\u{2335}', ON), ('\u{2336}', '\u{237a}', L), ('\u{237b}', '\u{2394}', ON), ('\u{2395}', + '\u{2395}', L), ('\u{2396}', '\u{2426}', ON), ('\u{2440}', '\u{244a}', ON), ('\u{2460}', + '\u{2487}', ON), ('\u{2488}', '\u{249b}', EN), ('\u{249c}', '\u{24e9}', L), ('\u{24ea}', + '\u{26ab}', ON), ('\u{26ac}', '\u{26ac}', L), ('\u{26ad}', '\u{27ff}', ON), ('\u{2800}', + '\u{28ff}', L), ('\u{2900}', '\u{2b73}', ON), ('\u{2b76}', '\u{2b95}', ON), ('\u{2b97}', + '\u{2bff}', ON), ('\u{2c00}', '\u{2ce4}', L), ('\u{2ce5}', '\u{2cea}', ON), ('\u{2ceb}', + '\u{2cee}', L), ('\u{2cef}', '\u{2cf1}', NSM), ('\u{2cf2}', '\u{2cf3}', L), ('\u{2cf9}', + '\u{2cff}', ON), ('\u{2d00}', '\u{2d25}', L), ('\u{2d27}', '\u{2d27}', L), ('\u{2d2d}', + '\u{2d2d}', L), ('\u{2d30}', '\u{2d67}', L), ('\u{2d6f}', '\u{2d70}', L), ('\u{2d7f}', + '\u{2d7f}', NSM), ('\u{2d80}', '\u{2d96}', L), ('\u{2da0}', '\u{2da6}', L), ('\u{2da8}', + '\u{2dae}', L), ('\u{2db0}', '\u{2db6}', L), ('\u{2db8}', '\u{2dbe}', L), ('\u{2dc0}', + '\u{2dc6}', L), ('\u{2dc8}', '\u{2dce}', L), ('\u{2dd0}', '\u{2dd6}', L), ('\u{2dd8}', + '\u{2dde}', L), ('\u{2de0}', '\u{2dff}', NSM), ('\u{2e00}', '\u{2e5d}', ON), ('\u{2e80}', + '\u{2e99}', ON), ('\u{2e9b}', '\u{2ef3}', ON), ('\u{2f00}', '\u{2fd5}', ON), ('\u{2ff0}', + '\u{2ffb}', ON), ('\u{3000}', '\u{3000}', WS), ('\u{3001}', '\u{3004}', ON), ('\u{3005}', + '\u{3007}', L), ('\u{3008}', '\u{3020}', ON), ('\u{3021}', '\u{3029}', L), ('\u{302a}', + '\u{302d}', NSM), ('\u{302e}', '\u{302f}', L), ('\u{3030}', '\u{3030}', ON), ('\u{3031}', + '\u{3035}', L), ('\u{3036}', '\u{3037}', ON), ('\u{3038}', '\u{303c}', L), ('\u{303d}', + '\u{303f}', ON), ('\u{3041}', '\u{3096}', L), ('\u{3099}', '\u{309a}', NSM), ('\u{309b}', + '\u{309c}', ON), ('\u{309d}', '\u{309f}', L), ('\u{30a0}', '\u{30a0}', ON), ('\u{30a1}', + '\u{30fa}', L), ('\u{30fb}', '\u{30fb}', ON), ('\u{30fc}', '\u{30ff}', L), ('\u{3105}', + '\u{312f}', L), ('\u{3131}', '\u{318e}', L), ('\u{3190}', '\u{31bf}', L), ('\u{31c0}', + '\u{31e3}', ON), ('\u{31f0}', '\u{321c}', L), ('\u{321d}', '\u{321e}', ON), ('\u{3220}', + '\u{324f}', L), ('\u{3250}', '\u{325f}', ON), ('\u{3260}', '\u{327b}', L), ('\u{327c}', + '\u{327e}', ON), ('\u{327f}', '\u{32b0}', L), ('\u{32b1}', '\u{32bf}', ON), ('\u{32c0}', + '\u{32cb}', L), ('\u{32cc}', '\u{32cf}', ON), ('\u{32d0}', '\u{3376}', L), ('\u{3377}', + '\u{337a}', ON), ('\u{337b}', '\u{33dd}', L), ('\u{33de}', '\u{33df}', ON), ('\u{33e0}', + '\u{33fe}', L), ('\u{33ff}', '\u{33ff}', ON), ('\u{3400}', '\u{4dbf}', L), ('\u{4dc0}', + '\u{4dff}', ON), ('\u{4e00}', '\u{a48c}', L), ('\u{a490}', '\u{a4c6}', ON), ('\u{a4d0}', + '\u{a60c}', L), ('\u{a60d}', '\u{a60f}', ON), ('\u{a610}', '\u{a62b}', L), ('\u{a640}', + '\u{a66e}', L), ('\u{a66f}', '\u{a672}', NSM), ('\u{a673}', '\u{a673}', ON), ('\u{a674}', + '\u{a67d}', NSM), ('\u{a67e}', '\u{a67f}', ON), ('\u{a680}', '\u{a69d}', L), ('\u{a69e}', + '\u{a69f}', NSM), ('\u{a6a0}', '\u{a6ef}', L), ('\u{a6f0}', '\u{a6f1}', NSM), ('\u{a6f2}', + '\u{a6f7}', L), ('\u{a700}', '\u{a721}', ON), ('\u{a722}', '\u{a787}', L), ('\u{a788}', + '\u{a788}', ON), ('\u{a789}', '\u{a7ca}', L), ('\u{a7d0}', '\u{a7d1}', L), ('\u{a7d3}', + '\u{a7d3}', L), ('\u{a7d5}', '\u{a7d9}', L), ('\u{a7f2}', '\u{a801}', L), ('\u{a802}', + '\u{a802}', NSM), ('\u{a803}', '\u{a805}', L), ('\u{a806}', '\u{a806}', NSM), ('\u{a807}', + '\u{a80a}', L), ('\u{a80b}', '\u{a80b}', NSM), ('\u{a80c}', '\u{a824}', L), ('\u{a825}', + '\u{a826}', NSM), ('\u{a827}', '\u{a827}', L), ('\u{a828}', '\u{a82b}', ON), ('\u{a82c}', + '\u{a82c}', NSM), ('\u{a830}', '\u{a837}', L), ('\u{a838}', '\u{a839}', ET), ('\u{a840}', + '\u{a873}', L), ('\u{a874}', '\u{a877}', ON), ('\u{a880}', '\u{a8c3}', L), ('\u{a8c4}', + '\u{a8c5}', NSM), ('\u{a8ce}', '\u{a8d9}', L), ('\u{a8e0}', '\u{a8f1}', NSM), ('\u{a8f2}', + '\u{a8fe}', L), ('\u{a8ff}', '\u{a8ff}', NSM), ('\u{a900}', '\u{a925}', L), ('\u{a926}', + '\u{a92d}', NSM), ('\u{a92e}', '\u{a946}', L), ('\u{a947}', '\u{a951}', NSM), ('\u{a952}', + '\u{a953}', L), ('\u{a95f}', '\u{a97c}', L), ('\u{a980}', '\u{a982}', NSM), ('\u{a983}', + '\u{a9b2}', L), ('\u{a9b3}', '\u{a9b3}', NSM), ('\u{a9b4}', '\u{a9b5}', L), ('\u{a9b6}', + '\u{a9b9}', NSM), ('\u{a9ba}', '\u{a9bb}', L), ('\u{a9bc}', '\u{a9bd}', NSM), ('\u{a9be}', + '\u{a9cd}', L), ('\u{a9cf}', '\u{a9d9}', L), ('\u{a9de}', '\u{a9e4}', L), ('\u{a9e5}', + '\u{a9e5}', NSM), ('\u{a9e6}', '\u{a9fe}', L), ('\u{aa00}', '\u{aa28}', L), ('\u{aa29}', + '\u{aa2e}', NSM), ('\u{aa2f}', '\u{aa30}', L), ('\u{aa31}', '\u{aa32}', NSM), ('\u{aa33}', + '\u{aa34}', L), ('\u{aa35}', '\u{aa36}', NSM), ('\u{aa40}', '\u{aa42}', L), ('\u{aa43}', + '\u{aa43}', NSM), ('\u{aa44}', '\u{aa4b}', L), ('\u{aa4c}', '\u{aa4c}', NSM), ('\u{aa4d}', + '\u{aa4d}', L), ('\u{aa50}', '\u{aa59}', L), ('\u{aa5c}', '\u{aa7b}', L), ('\u{aa7c}', + '\u{aa7c}', NSM), ('\u{aa7d}', '\u{aaaf}', L), ('\u{aab0}', '\u{aab0}', NSM), ('\u{aab1}', + '\u{aab1}', L), ('\u{aab2}', '\u{aab4}', NSM), ('\u{aab5}', '\u{aab6}', L), ('\u{aab7}', + '\u{aab8}', NSM), ('\u{aab9}', '\u{aabd}', L), ('\u{aabe}', '\u{aabf}', NSM), ('\u{aac0}', + '\u{aac0}', L), ('\u{aac1}', '\u{aac1}', NSM), ('\u{aac2}', '\u{aac2}', L), ('\u{aadb}', + '\u{aaeb}', L), ('\u{aaec}', '\u{aaed}', NSM), ('\u{aaee}', '\u{aaf5}', L), ('\u{aaf6}', + '\u{aaf6}', NSM), ('\u{ab01}', '\u{ab06}', L), ('\u{ab09}', '\u{ab0e}', L), ('\u{ab11}', + '\u{ab16}', L), ('\u{ab20}', '\u{ab26}', L), ('\u{ab28}', '\u{ab2e}', L), ('\u{ab30}', + '\u{ab69}', L), ('\u{ab6a}', '\u{ab6b}', ON), ('\u{ab70}', '\u{abe4}', L), ('\u{abe5}', + '\u{abe5}', NSM), ('\u{abe6}', '\u{abe7}', L), ('\u{abe8}', '\u{abe8}', NSM), ('\u{abe9}', + '\u{abec}', L), ('\u{abed}', '\u{abed}', NSM), ('\u{abf0}', '\u{abf9}', L), ('\u{ac00}', + '\u{d7a3}', L), ('\u{d7b0}', '\u{d7c6}', L), ('\u{d7cb}', '\u{d7fb}', L), ('\u{e000}', + '\u{fa6d}', L), ('\u{fa70}', '\u{fad9}', L), ('\u{fb00}', '\u{fb06}', L), ('\u{fb13}', + '\u{fb17}', L), ('\u{fb1d}', '\u{fb1d}', R), ('\u{fb1e}', '\u{fb1e}', NSM), ('\u{fb1f}', + '\u{fb28}', R), ('\u{fb29}', '\u{fb29}', ES), ('\u{fb2a}', '\u{fb4f}', R), ('\u{fb50}', + '\u{fd3d}', AL), ('\u{fd3e}', '\u{fd4f}', ON), ('\u{fd50}', '\u{fdce}', AL), ('\u{fdcf}', + '\u{fdcf}', ON), ('\u{fdf0}', '\u{fdfc}', AL), ('\u{fdfd}', '\u{fdff}', ON), ('\u{fe00}', + '\u{fe0f}', NSM), ('\u{fe10}', '\u{fe19}', ON), ('\u{fe20}', '\u{fe2f}', NSM), ('\u{fe30}', + '\u{fe4f}', ON), ('\u{fe50}', '\u{fe50}', CS), ('\u{fe51}', '\u{fe51}', ON), ('\u{fe52}', + '\u{fe52}', CS), ('\u{fe54}', '\u{fe54}', ON), ('\u{fe55}', '\u{fe55}', CS), ('\u{fe56}', + '\u{fe5e}', ON), ('\u{fe5f}', '\u{fe5f}', ET), ('\u{fe60}', '\u{fe61}', ON), ('\u{fe62}', + '\u{fe63}', ES), ('\u{fe64}', '\u{fe66}', ON), ('\u{fe68}', '\u{fe68}', ON), ('\u{fe69}', + '\u{fe6a}', ET), ('\u{fe6b}', '\u{fe6b}', ON), ('\u{fe70}', '\u{fefe}', AL), ('\u{feff}', + '\u{feff}', BN), ('\u{ff01}', '\u{ff02}', ON), ('\u{ff03}', '\u{ff05}', ET), ('\u{ff06}', + '\u{ff0a}', ON), ('\u{ff0b}', '\u{ff0b}', ES), ('\u{ff0c}', '\u{ff0c}', CS), ('\u{ff0d}', + '\u{ff0d}', ES), ('\u{ff0e}', '\u{ff0f}', CS), ('\u{ff10}', '\u{ff19}', EN), ('\u{ff1a}', + '\u{ff1a}', CS), ('\u{ff1b}', '\u{ff20}', ON), ('\u{ff21}', '\u{ff3a}', L), ('\u{ff3b}', + '\u{ff40}', ON), ('\u{ff41}', '\u{ff5a}', L), ('\u{ff5b}', '\u{ff65}', ON), ('\u{ff66}', + '\u{ffbe}', L), ('\u{ffc2}', '\u{ffc7}', L), ('\u{ffca}', '\u{ffcf}', L), ('\u{ffd2}', + '\u{ffd7}', L), ('\u{ffda}', '\u{ffdc}', L), ('\u{ffe0}', '\u{ffe1}', ET), ('\u{ffe2}', + '\u{ffe4}', ON), ('\u{ffe5}', '\u{ffe6}', ET), ('\u{ffe8}', '\u{ffee}', ON), ('\u{fff9}', + '\u{fffd}', ON), ('\u{10000}', '\u{1000b}', L), ('\u{1000d}', '\u{10026}', L), ('\u{10028}', + '\u{1003a}', L), ('\u{1003c}', '\u{1003d}', L), ('\u{1003f}', '\u{1004d}', L), ('\u{10050}', + '\u{1005d}', L), ('\u{10080}', '\u{100fa}', L), ('\u{10100}', '\u{10100}', L), ('\u{10101}', + '\u{10101}', ON), ('\u{10102}', '\u{10102}', L), ('\u{10107}', '\u{10133}', L), ('\u{10137}', + '\u{1013f}', L), ('\u{10140}', '\u{1018c}', ON), ('\u{1018d}', '\u{1018e}', L), ('\u{10190}', + '\u{1019c}', ON), ('\u{101a0}', '\u{101a0}', ON), ('\u{101d0}', '\u{101fc}', L), ('\u{101fd}', + '\u{101fd}', NSM), ('\u{10280}', '\u{1029c}', L), ('\u{102a0}', '\u{102d0}', L), ('\u{102e0}', + '\u{102e0}', NSM), ('\u{102e1}', '\u{102fb}', EN), ('\u{10300}', '\u{10323}', L), ('\u{1032d}', + '\u{1034a}', L), ('\u{10350}', '\u{10375}', L), ('\u{10376}', '\u{1037a}', NSM), ('\u{10380}', + '\u{1039d}', L), ('\u{1039f}', '\u{103c3}', L), ('\u{103c8}', '\u{103d5}', L), ('\u{10400}', + '\u{1049d}', L), ('\u{104a0}', '\u{104a9}', L), ('\u{104b0}', '\u{104d3}', L), ('\u{104d8}', + '\u{104fb}', L), ('\u{10500}', '\u{10527}', L), ('\u{10530}', '\u{10563}', L), ('\u{1056f}', + '\u{1057a}', L), ('\u{1057c}', '\u{1058a}', L), ('\u{1058c}', '\u{10592}', L), ('\u{10594}', + '\u{10595}', L), ('\u{10597}', '\u{105a1}', L), ('\u{105a3}', '\u{105b1}', L), ('\u{105b3}', + '\u{105b9}', L), ('\u{105bb}', '\u{105bc}', L), ('\u{10600}', '\u{10736}', L), ('\u{10740}', + '\u{10755}', L), ('\u{10760}', '\u{10767}', L), ('\u{10780}', '\u{10785}', L), ('\u{10787}', + '\u{107b0}', L), ('\u{107b2}', '\u{107ba}', L), ('\u{10800}', '\u{1091e}', R), ('\u{1091f}', + '\u{1091f}', ON), ('\u{10920}', '\u{10a00}', R), ('\u{10a01}', '\u{10a03}', NSM), ('\u{10a04}', + '\u{10a04}', R), ('\u{10a05}', '\u{10a06}', NSM), ('\u{10a07}', '\u{10a0b}', R), ('\u{10a0c}', + '\u{10a0f}', NSM), ('\u{10a10}', '\u{10a37}', R), ('\u{10a38}', '\u{10a3a}', NSM), ('\u{10a3b}', + '\u{10a3e}', R), ('\u{10a3f}', '\u{10a3f}', NSM), ('\u{10a40}', '\u{10ae4}', R), ('\u{10ae5}', + '\u{10ae6}', NSM), ('\u{10ae7}', '\u{10b38}', R), ('\u{10b39}', '\u{10b3f}', ON), ('\u{10b40}', + '\u{10cff}', R), ('\u{10d00}', '\u{10d23}', AL), ('\u{10d24}', '\u{10d27}', NSM), ('\u{10d28}', + '\u{10d2f}', R), ('\u{10d30}', '\u{10d39}', AN), ('\u{10d3a}', '\u{10e5f}', R), ('\u{10e60}', + '\u{10e7e}', AN), ('\u{10e7f}', '\u{10eaa}', R), ('\u{10eab}', '\u{10eac}', NSM), ('\u{10ead}', + '\u{10f2f}', R), ('\u{10f30}', '\u{10f45}', AL), ('\u{10f46}', '\u{10f50}', NSM), ('\u{10f51}', + '\u{10f59}', AL), ('\u{10f5a}', '\u{10f81}', R), ('\u{10f82}', '\u{10f85}', NSM), ('\u{10f86}', + '\u{10fff}', R), ('\u{11000}', '\u{11000}', L), ('\u{11001}', '\u{11001}', NSM), ('\u{11002}', + '\u{11037}', L), ('\u{11038}', '\u{11046}', NSM), ('\u{11047}', '\u{1104d}', L), ('\u{11052}', + '\u{11065}', ON), ('\u{11066}', '\u{1106f}', L), ('\u{11070}', '\u{11070}', NSM), ('\u{11071}', + '\u{11072}', L), ('\u{11073}', '\u{11074}', NSM), ('\u{11075}', '\u{11075}', L), ('\u{1107f}', + '\u{11081}', NSM), ('\u{11082}', '\u{110b2}', L), ('\u{110b3}', '\u{110b6}', NSM), ('\u{110b7}', + '\u{110b8}', L), ('\u{110b9}', '\u{110ba}', NSM), ('\u{110bb}', '\u{110c1}', L), ('\u{110c2}', + '\u{110c2}', NSM), ('\u{110cd}', '\u{110cd}', L), ('\u{110d0}', '\u{110e8}', L), ('\u{110f0}', + '\u{110f9}', L), ('\u{11100}', '\u{11102}', NSM), ('\u{11103}', '\u{11126}', L), ('\u{11127}', + '\u{1112b}', NSM), ('\u{1112c}', '\u{1112c}', L), ('\u{1112d}', '\u{11134}', NSM), ('\u{11136}', + '\u{11147}', L), ('\u{11150}', '\u{11172}', L), ('\u{11173}', '\u{11173}', NSM), ('\u{11174}', + '\u{11176}', L), ('\u{11180}', '\u{11181}', NSM), ('\u{11182}', '\u{111b5}', L), ('\u{111b6}', + '\u{111be}', NSM), ('\u{111bf}', '\u{111c8}', L), ('\u{111c9}', '\u{111cc}', NSM), ('\u{111cd}', + '\u{111ce}', L), ('\u{111cf}', '\u{111cf}', NSM), ('\u{111d0}', '\u{111df}', L), ('\u{111e1}', + '\u{111f4}', L), ('\u{11200}', '\u{11211}', L), ('\u{11213}', '\u{1122e}', L), ('\u{1122f}', + '\u{11231}', NSM), ('\u{11232}', '\u{11233}', L), ('\u{11234}', '\u{11234}', NSM), ('\u{11235}', + '\u{11235}', L), ('\u{11236}', '\u{11237}', NSM), ('\u{11238}', '\u{1123d}', L), ('\u{1123e}', + '\u{1123e}', NSM), ('\u{11280}', '\u{11286}', L), ('\u{11288}', '\u{11288}', L), ('\u{1128a}', + '\u{1128d}', L), ('\u{1128f}', '\u{1129d}', L), ('\u{1129f}', '\u{112a9}', L), ('\u{112b0}', + '\u{112de}', L), ('\u{112df}', '\u{112df}', NSM), ('\u{112e0}', '\u{112e2}', L), ('\u{112e3}', + '\u{112ea}', NSM), ('\u{112f0}', '\u{112f9}', L), ('\u{11300}', '\u{11301}', NSM), ('\u{11302}', + '\u{11303}', L), ('\u{11305}', '\u{1130c}', L), ('\u{1130f}', '\u{11310}', L), ('\u{11313}', + '\u{11328}', L), ('\u{1132a}', '\u{11330}', L), ('\u{11332}', '\u{11333}', L), ('\u{11335}', + '\u{11339}', L), ('\u{1133b}', '\u{1133c}', NSM), ('\u{1133d}', '\u{1133f}', L), ('\u{11340}', + '\u{11340}', NSM), ('\u{11341}', '\u{11344}', L), ('\u{11347}', '\u{11348}', L), ('\u{1134b}', + '\u{1134d}', L), ('\u{11350}', '\u{11350}', L), ('\u{11357}', '\u{11357}', L), ('\u{1135d}', + '\u{11363}', L), ('\u{11366}', '\u{1136c}', NSM), ('\u{11370}', '\u{11374}', NSM), ('\u{11400}', + '\u{11437}', L), ('\u{11438}', '\u{1143f}', NSM), ('\u{11440}', '\u{11441}', L), ('\u{11442}', + '\u{11444}', NSM), ('\u{11445}', '\u{11445}', L), ('\u{11446}', '\u{11446}', NSM), ('\u{11447}', + '\u{1145b}', L), ('\u{1145d}', '\u{1145d}', L), ('\u{1145e}', '\u{1145e}', NSM), ('\u{1145f}', + '\u{11461}', L), ('\u{11480}', '\u{114b2}', L), ('\u{114b3}', '\u{114b8}', NSM), ('\u{114b9}', + '\u{114b9}', L), ('\u{114ba}', '\u{114ba}', NSM), ('\u{114bb}', '\u{114be}', L), ('\u{114bf}', + '\u{114c0}', NSM), ('\u{114c1}', '\u{114c1}', L), ('\u{114c2}', '\u{114c3}', NSM), ('\u{114c4}', + '\u{114c7}', L), ('\u{114d0}', '\u{114d9}', L), ('\u{11580}', '\u{115b1}', L), ('\u{115b2}', + '\u{115b5}', NSM), ('\u{115b8}', '\u{115bb}', L), ('\u{115bc}', '\u{115bd}', NSM), ('\u{115be}', + '\u{115be}', L), ('\u{115bf}', '\u{115c0}', NSM), ('\u{115c1}', '\u{115db}', L), ('\u{115dc}', + '\u{115dd}', NSM), ('\u{11600}', '\u{11632}', L), ('\u{11633}', '\u{1163a}', NSM), ('\u{1163b}', + '\u{1163c}', L), ('\u{1163d}', '\u{1163d}', NSM), ('\u{1163e}', '\u{1163e}', L), ('\u{1163f}', + '\u{11640}', NSM), ('\u{11641}', '\u{11644}', L), ('\u{11650}', '\u{11659}', L), ('\u{11660}', + '\u{1166c}', ON), ('\u{11680}', '\u{116aa}', L), ('\u{116ab}', '\u{116ab}', NSM), ('\u{116ac}', + '\u{116ac}', L), ('\u{116ad}', '\u{116ad}', NSM), ('\u{116ae}', '\u{116af}', L), ('\u{116b0}', + '\u{116b5}', NSM), ('\u{116b6}', '\u{116b6}', L), ('\u{116b7}', '\u{116b7}', NSM), ('\u{116b8}', + '\u{116b9}', L), ('\u{116c0}', '\u{116c9}', L), ('\u{11700}', '\u{1171a}', L), ('\u{1171d}', + '\u{1171f}', NSM), ('\u{11720}', '\u{11721}', L), ('\u{11722}', '\u{11725}', NSM), ('\u{11726}', + '\u{11726}', L), ('\u{11727}', '\u{1172b}', NSM), ('\u{11730}', '\u{11746}', L), ('\u{11800}', + '\u{1182e}', L), ('\u{1182f}', '\u{11837}', NSM), ('\u{11838}', '\u{11838}', L), ('\u{11839}', + '\u{1183a}', NSM), ('\u{1183b}', '\u{1183b}', L), ('\u{118a0}', '\u{118f2}', L), ('\u{118ff}', + '\u{11906}', L), ('\u{11909}', '\u{11909}', L), ('\u{1190c}', '\u{11913}', L), ('\u{11915}', + '\u{11916}', L), ('\u{11918}', '\u{11935}', L), ('\u{11937}', '\u{11938}', L), ('\u{1193b}', + '\u{1193c}', NSM), ('\u{1193d}', '\u{1193d}', L), ('\u{1193e}', '\u{1193e}', NSM), ('\u{1193f}', + '\u{11942}', L), ('\u{11943}', '\u{11943}', NSM), ('\u{11944}', '\u{11946}', L), ('\u{11950}', + '\u{11959}', L), ('\u{119a0}', '\u{119a7}', L), ('\u{119aa}', '\u{119d3}', L), ('\u{119d4}', + '\u{119d7}', NSM), ('\u{119da}', '\u{119db}', NSM), ('\u{119dc}', '\u{119df}', L), ('\u{119e0}', + '\u{119e0}', NSM), ('\u{119e1}', '\u{119e4}', L), ('\u{11a00}', '\u{11a00}', L), ('\u{11a01}', + '\u{11a06}', NSM), ('\u{11a07}', '\u{11a08}', L), ('\u{11a09}', '\u{11a0a}', NSM), ('\u{11a0b}', + '\u{11a32}', L), ('\u{11a33}', '\u{11a38}', NSM), ('\u{11a39}', '\u{11a3a}', L), ('\u{11a3b}', + '\u{11a3e}', NSM), ('\u{11a3f}', '\u{11a46}', L), ('\u{11a47}', '\u{11a47}', NSM), ('\u{11a50}', + '\u{11a50}', L), ('\u{11a51}', '\u{11a56}', NSM), ('\u{11a57}', '\u{11a58}', L), ('\u{11a59}', + '\u{11a5b}', NSM), ('\u{11a5c}', '\u{11a89}', L), ('\u{11a8a}', '\u{11a96}', NSM), ('\u{11a97}', + '\u{11a97}', L), ('\u{11a98}', '\u{11a99}', NSM), ('\u{11a9a}', '\u{11aa2}', L), ('\u{11ab0}', + '\u{11af8}', L), ('\u{11c00}', '\u{11c08}', L), ('\u{11c0a}', '\u{11c2f}', L), ('\u{11c30}', + '\u{11c36}', NSM), ('\u{11c38}', '\u{11c3d}', NSM), ('\u{11c3e}', '\u{11c45}', L), ('\u{11c50}', + '\u{11c6c}', L), ('\u{11c70}', '\u{11c8f}', L), ('\u{11c92}', '\u{11ca7}', NSM), ('\u{11ca9}', + '\u{11ca9}', L), ('\u{11caa}', '\u{11cb0}', NSM), ('\u{11cb1}', '\u{11cb1}', L), ('\u{11cb2}', + '\u{11cb3}', NSM), ('\u{11cb4}', '\u{11cb4}', L), ('\u{11cb5}', '\u{11cb6}', NSM), ('\u{11d00}', + '\u{11d06}', L), ('\u{11d08}', '\u{11d09}', L), ('\u{11d0b}', '\u{11d30}', L), ('\u{11d31}', + '\u{11d36}', NSM), ('\u{11d3a}', '\u{11d3a}', NSM), ('\u{11d3c}', '\u{11d3d}', NSM), + ('\u{11d3f}', '\u{11d45}', NSM), ('\u{11d46}', '\u{11d46}', L), ('\u{11d47}', '\u{11d47}', NSM), + ('\u{11d50}', '\u{11d59}', L), ('\u{11d60}', '\u{11d65}', L), ('\u{11d67}', '\u{11d68}', L), + ('\u{11d6a}', '\u{11d8e}', L), ('\u{11d90}', '\u{11d91}', NSM), ('\u{11d93}', '\u{11d94}', L), + ('\u{11d95}', '\u{11d95}', NSM), ('\u{11d96}', '\u{11d96}', L), ('\u{11d97}', '\u{11d97}', NSM), + ('\u{11d98}', '\u{11d98}', L), ('\u{11da0}', '\u{11da9}', L), ('\u{11ee0}', '\u{11ef2}', L), + ('\u{11ef3}', '\u{11ef4}', NSM), ('\u{11ef5}', '\u{11ef8}', L), ('\u{11fb0}', '\u{11fb0}', L), + ('\u{11fc0}', '\u{11fd4}', L), ('\u{11fd5}', '\u{11fdc}', ON), ('\u{11fdd}', '\u{11fe0}', ET), + ('\u{11fe1}', '\u{11ff1}', ON), ('\u{11fff}', '\u{12399}', L), ('\u{12400}', '\u{1246e}', L), + ('\u{12470}', '\u{12474}', L), ('\u{12480}', '\u{12543}', L), ('\u{12f90}', '\u{12ff2}', L), + ('\u{13000}', '\u{1342e}', L), ('\u{13430}', '\u{13438}', L), ('\u{14400}', '\u{14646}', L), + ('\u{16800}', '\u{16a38}', L), ('\u{16a40}', '\u{16a5e}', L), ('\u{16a60}', '\u{16a69}', L), + ('\u{16a6e}', '\u{16abe}', L), ('\u{16ac0}', '\u{16ac9}', L), ('\u{16ad0}', '\u{16aed}', L), + ('\u{16af0}', '\u{16af4}', NSM), ('\u{16af5}', '\u{16af5}', L), ('\u{16b00}', '\u{16b2f}', L), + ('\u{16b30}', '\u{16b36}', NSM), ('\u{16b37}', '\u{16b45}', L), ('\u{16b50}', '\u{16b59}', L), + ('\u{16b5b}', '\u{16b61}', L), ('\u{16b63}', '\u{16b77}', L), ('\u{16b7d}', '\u{16b8f}', L), + ('\u{16e40}', '\u{16e9a}', L), ('\u{16f00}', '\u{16f4a}', L), ('\u{16f4f}', '\u{16f4f}', NSM), + ('\u{16f50}', '\u{16f87}', L), ('\u{16f8f}', '\u{16f92}', NSM), ('\u{16f93}', '\u{16f9f}', L), + ('\u{16fe0}', '\u{16fe1}', L), ('\u{16fe2}', '\u{16fe2}', ON), ('\u{16fe3}', '\u{16fe3}', L), + ('\u{16fe4}', '\u{16fe4}', NSM), ('\u{16ff0}', '\u{16ff1}', L), ('\u{17000}', '\u{187f7}', L), + ('\u{18800}', '\u{18cd5}', L), ('\u{18d00}', '\u{18d08}', L), ('\u{1aff0}', '\u{1aff3}', L), + ('\u{1aff5}', '\u{1affb}', L), ('\u{1affd}', '\u{1affe}', L), ('\u{1b000}', '\u{1b122}', L), + ('\u{1b150}', '\u{1b152}', L), ('\u{1b164}', '\u{1b167}', L), ('\u{1b170}', '\u{1b2fb}', L), + ('\u{1bc00}', '\u{1bc6a}', L), ('\u{1bc70}', '\u{1bc7c}', L), ('\u{1bc80}', '\u{1bc88}', L), + ('\u{1bc90}', '\u{1bc99}', L), ('\u{1bc9c}', '\u{1bc9c}', L), ('\u{1bc9d}', '\u{1bc9e}', NSM), + ('\u{1bc9f}', '\u{1bc9f}', L), ('\u{1bca0}', '\u{1bca3}', BN), ('\u{1cf00}', '\u{1cf2d}', NSM), + ('\u{1cf30}', '\u{1cf46}', NSM), ('\u{1cf50}', '\u{1cfc3}', L), ('\u{1d000}', '\u{1d0f5}', L), + ('\u{1d100}', '\u{1d126}', L), ('\u{1d129}', '\u{1d166}', L), ('\u{1d167}', '\u{1d169}', NSM), + ('\u{1d16a}', '\u{1d172}', L), ('\u{1d173}', '\u{1d17a}', BN), ('\u{1d17b}', '\u{1d182}', NSM), + ('\u{1d183}', '\u{1d184}', L), ('\u{1d185}', '\u{1d18b}', NSM), ('\u{1d18c}', '\u{1d1a9}', L), + ('\u{1d1aa}', '\u{1d1ad}', NSM), ('\u{1d1ae}', '\u{1d1e8}', L), ('\u{1d1e9}', '\u{1d1ea}', ON), + ('\u{1d200}', '\u{1d241}', ON), ('\u{1d242}', '\u{1d244}', NSM), ('\u{1d245}', '\u{1d245}', ON), + ('\u{1d2e0}', '\u{1d2f3}', L), ('\u{1d300}', '\u{1d356}', ON), ('\u{1d360}', '\u{1d378}', L), + ('\u{1d400}', '\u{1d454}', L), ('\u{1d456}', '\u{1d49c}', L), ('\u{1d49e}', '\u{1d49f}', L), + ('\u{1d4a2}', '\u{1d4a2}', L), ('\u{1d4a5}', '\u{1d4a6}', L), ('\u{1d4a9}', '\u{1d4ac}', L), + ('\u{1d4ae}', '\u{1d4b9}', L), ('\u{1d4bb}', '\u{1d4bb}', L), ('\u{1d4bd}', '\u{1d4c3}', L), + ('\u{1d4c5}', '\u{1d505}', L), ('\u{1d507}', '\u{1d50a}', L), ('\u{1d50d}', '\u{1d514}', L), + ('\u{1d516}', '\u{1d51c}', L), ('\u{1d51e}', '\u{1d539}', L), ('\u{1d53b}', '\u{1d53e}', L), + ('\u{1d540}', '\u{1d544}', L), ('\u{1d546}', '\u{1d546}', L), ('\u{1d54a}', '\u{1d550}', L), + ('\u{1d552}', '\u{1d6a5}', L), ('\u{1d6a8}', '\u{1d6da}', L), ('\u{1d6db}', '\u{1d6db}', ON), + ('\u{1d6dc}', '\u{1d714}', L), ('\u{1d715}', '\u{1d715}', ON), ('\u{1d716}', '\u{1d74e}', L), + ('\u{1d74f}', '\u{1d74f}', ON), ('\u{1d750}', '\u{1d788}', L), ('\u{1d789}', '\u{1d789}', ON), + ('\u{1d78a}', '\u{1d7c2}', L), ('\u{1d7c3}', '\u{1d7c3}', ON), ('\u{1d7c4}', '\u{1d7cb}', L), + ('\u{1d7ce}', '\u{1d7ff}', EN), ('\u{1d800}', '\u{1d9ff}', L), ('\u{1da00}', '\u{1da36}', NSM), + ('\u{1da37}', '\u{1da3a}', L), ('\u{1da3b}', '\u{1da6c}', NSM), ('\u{1da6d}', '\u{1da74}', L), + ('\u{1da75}', '\u{1da75}', NSM), ('\u{1da76}', '\u{1da83}', L), ('\u{1da84}', '\u{1da84}', NSM), + ('\u{1da85}', '\u{1da8b}', L), ('\u{1da9b}', '\u{1da9f}', NSM), ('\u{1daa1}', '\u{1daaf}', NSM), + ('\u{1df00}', '\u{1df1e}', L), ('\u{1e000}', '\u{1e006}', NSM), ('\u{1e008}', '\u{1e018}', NSM), + ('\u{1e01b}', '\u{1e021}', NSM), ('\u{1e023}', '\u{1e024}', NSM), ('\u{1e026}', '\u{1e02a}', + NSM), ('\u{1e100}', '\u{1e12c}', L), ('\u{1e130}', '\u{1e136}', NSM), ('\u{1e137}', '\u{1e13d}', + L), ('\u{1e140}', '\u{1e149}', L), ('\u{1e14e}', '\u{1e14f}', L), ('\u{1e290}', '\u{1e2ad}', L), + ('\u{1e2ae}', '\u{1e2ae}', NSM), ('\u{1e2c0}', '\u{1e2eb}', L), ('\u{1e2ec}', '\u{1e2ef}', NSM), + ('\u{1e2f0}', '\u{1e2f9}', L), ('\u{1e2ff}', '\u{1e2ff}', ET), ('\u{1e7e0}', '\u{1e7e6}', L), + ('\u{1e7e8}', '\u{1e7eb}', L), ('\u{1e7ed}', '\u{1e7ee}', L), ('\u{1e7f0}', '\u{1e7fe}', L), + ('\u{1e800}', '\u{1e8cf}', R), ('\u{1e8d0}', '\u{1e8d6}', NSM), ('\u{1e8d7}', '\u{1e943}', R), + ('\u{1e944}', '\u{1e94a}', NSM), ('\u{1e94b}', '\u{1ec70}', R), ('\u{1ec71}', '\u{1ecb4}', AL), + ('\u{1ecb5}', '\u{1ed00}', R), ('\u{1ed01}', '\u{1ed3d}', AL), ('\u{1ed3e}', '\u{1edff}', R), + ('\u{1ee00}', '\u{1eeef}', AL), ('\u{1eef0}', '\u{1eef1}', ON), ('\u{1eef2}', '\u{1eeff}', AL), + ('\u{1ef00}', '\u{1efff}', R), ('\u{1f000}', '\u{1f02b}', ON), ('\u{1f030}', '\u{1f093}', ON), + ('\u{1f0a0}', '\u{1f0ae}', ON), ('\u{1f0b1}', '\u{1f0bf}', ON), ('\u{1f0c1}', '\u{1f0cf}', ON), + ('\u{1f0d1}', '\u{1f0f5}', ON), ('\u{1f100}', '\u{1f10a}', EN), ('\u{1f10b}', '\u{1f10f}', ON), + ('\u{1f110}', '\u{1f12e}', L), ('\u{1f12f}', '\u{1f12f}', ON), ('\u{1f130}', '\u{1f169}', L), + ('\u{1f16a}', '\u{1f16f}', ON), ('\u{1f170}', '\u{1f1ac}', L), ('\u{1f1ad}', '\u{1f1ad}', ON), + ('\u{1f1e6}', '\u{1f202}', L), ('\u{1f210}', '\u{1f23b}', L), ('\u{1f240}', '\u{1f248}', L), + ('\u{1f250}', '\u{1f251}', L), ('\u{1f260}', '\u{1f265}', ON), ('\u{1f300}', '\u{1f6d7}', ON), + ('\u{1f6dd}', '\u{1f6ec}', ON), ('\u{1f6f0}', '\u{1f6fc}', ON), ('\u{1f700}', '\u{1f773}', ON), + ('\u{1f780}', '\u{1f7d8}', ON), ('\u{1f7e0}', '\u{1f7eb}', ON), ('\u{1f7f0}', '\u{1f7f0}', ON), + ('\u{1f800}', '\u{1f80b}', ON), ('\u{1f810}', '\u{1f847}', ON), ('\u{1f850}', '\u{1f859}', ON), + ('\u{1f860}', '\u{1f887}', ON), ('\u{1f890}', '\u{1f8ad}', ON), ('\u{1f8b0}', '\u{1f8b1}', ON), + ('\u{1f900}', '\u{1fa53}', ON), ('\u{1fa60}', '\u{1fa6d}', ON), ('\u{1fa70}', '\u{1fa74}', ON), + ('\u{1fa78}', '\u{1fa7c}', ON), ('\u{1fa80}', '\u{1fa86}', ON), ('\u{1fa90}', '\u{1faac}', ON), + ('\u{1fab0}', '\u{1faba}', ON), ('\u{1fac0}', '\u{1fac5}', ON), ('\u{1fad0}', '\u{1fad9}', ON), + ('\u{1fae0}', '\u{1fae7}', ON), ('\u{1faf0}', '\u{1faf6}', ON), ('\u{1fb00}', '\u{1fb92}', ON), + ('\u{1fb94}', '\u{1fbca}', ON), ('\u{1fbf0}', '\u{1fbf9}', EN), ('\u{20000}', '\u{2a6df}', L), + ('\u{2a700}', '\u{2b738}', L), ('\u{2b740}', '\u{2b81d}', L), ('\u{2b820}', '\u{2cea1}', L), + ('\u{2ceb0}', '\u{2ebe0}', L), ('\u{2f800}', '\u{2fa1d}', L), ('\u{30000}', '\u{3134a}', L), + ('\u{e0001}', '\u{e0001}', BN), ('\u{e0020}', '\u{e007f}', BN), ('\u{e0100}', '\u{e01ef}', NSM), + ('\u{f0000}', '\u{ffffd}', L), ('\u{100000}', '\u{10fffd}', L) +]; + diff --git a/vendor/unicode-bidi/src/data_source.rs b/vendor/unicode-bidi/src/data_source.rs new file mode 100644 index 000000000..3958a255e --- /dev/null +++ b/vendor/unicode-bidi/src/data_source.rs @@ -0,0 +1,16 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +use crate::BidiClass; + +/// This trait abstracts over a data source that is able to produce the Unicode Bidi class for a given +/// character +pub trait BidiDataSource { + fn bidi_class(&self, c: char) -> BidiClass; +} diff --git a/vendor/unicode-bidi/src/deprecated.rs b/vendor/unicode-bidi/src/deprecated.rs new file mode 100644 index 000000000..ec3b84f67 --- /dev/null +++ b/vendor/unicode-bidi/src/deprecated.rs @@ -0,0 +1,92 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! This module holds deprecated assets only. + +use alloc::vec::Vec; + +use super::*; + +/// Find the level runs within a line and return them in visual order. +/// +/// NOTE: This implementation is incomplete. The algorithm needs information about the text, +/// including original `BidiClass` property of each character, to be able to perform correctly. +/// Please see [`BidiInfo::visual_runs()`](../struct.BidiInfo.html#method.visual_runs) for the +/// improved implementation. +/// +/// `line` is a range of bytes indices within `levels`. +/// +/// <http://www.unicode.org/reports/tr9/#Reordering_Resolved_Levels> +#[deprecated( + since = "0.3.0", + note = "please use `BidiInfo::visual_runs()` instead." +)] +pub fn visual_runs(line: Range<usize>, levels: &[Level]) -> Vec<LevelRun> { + assert!(line.start <= levels.len()); + assert!(line.end <= levels.len()); + + let mut runs = Vec::new(); + + // Find consecutive level runs. + let mut start = line.start; + let mut run_level = levels[start]; + let mut min_level = run_level; + let mut max_level = run_level; + + for (i, &new_level) in levels.iter().enumerate().take(line.end).skip(start + 1) { + if new_level != run_level { + // End of the previous run, start of a new one. + runs.push(start..i); + start = i; + run_level = new_level; + + min_level = min(run_level, min_level); + max_level = max(run_level, max_level); + } + } + runs.push(start..line.end); + + let run_count = runs.len(); + + // Re-order the odd runs. + // <http://www.unicode.org/reports/tr9/#L2> + + // Stop at the lowest *odd* level. + min_level = min_level.new_lowest_ge_rtl().expect("Level error"); + + while max_level >= min_level { + // Look for the start of a sequence of consecutive runs of max_level or higher. + let mut seq_start = 0; + while seq_start < run_count { + if levels[runs[seq_start].start] < max_level { + seq_start += 1; + continue; + } + + // Found the start of a sequence. Now find the end. + let mut seq_end = seq_start + 1; + while seq_end < run_count { + if levels[runs[seq_end].start] < max_level { + break; + } + seq_end += 1; + } + + // Reverse the runs within this sequence. + runs[seq_start..seq_end].reverse(); + + seq_start = seq_end; + } + max_level + .lower(1) + .expect("Lowering embedding level below zero"); + } + + runs +} diff --git a/vendor/unicode-bidi/src/explicit.rs b/vendor/unicode-bidi/src/explicit.rs new file mode 100644 index 000000000..e9d579fd6 --- /dev/null +++ b/vendor/unicode-bidi/src/explicit.rs @@ -0,0 +1,195 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! 3.3.2 Explicit Levels and Directions +//! +//! <http://www.unicode.org/reports/tr9/#Explicit_Levels_and_Directions> + +use alloc::vec::Vec; + +use super::char_data::{ + is_rtl, + BidiClass::{self, *}, +}; +use super::level::Level; + +/// Compute explicit embedding levels for one paragraph of text (X1-X8). +/// +/// `processing_classes[i]` must contain the `BidiClass` of the char at byte index `i`, +/// for each char in `text`. +#[cfg_attr(feature = "flame_it", flamer::flame)] +pub fn compute( + text: &str, + para_level: Level, + original_classes: &[BidiClass], + levels: &mut [Level], + processing_classes: &mut [BidiClass], +) { + assert_eq!(text.len(), original_classes.len()); + + // <http://www.unicode.org/reports/tr9/#X1> + let mut stack = DirectionalStatusStack::new(); + stack.push(para_level, OverrideStatus::Neutral); + + let mut overflow_isolate_count = 0u32; + let mut overflow_embedding_count = 0u32; + let mut valid_isolate_count = 0u32; + + for (i, c) in text.char_indices() { + match original_classes[i] { + // Rules X2-X5c + RLE | LRE | RLO | LRO | RLI | LRI | FSI => { + let last_level = stack.last().level; + + // X5a-X5c: Isolate initiators get the level of the last entry on the stack. + let is_isolate = match original_classes[i] { + RLI | LRI | FSI => true, + _ => false, + }; + if is_isolate { + levels[i] = last_level; + match stack.last().status { + OverrideStatus::RTL => processing_classes[i] = R, + OverrideStatus::LTR => processing_classes[i] = L, + _ => {} + } + } + + let new_level = if is_rtl(original_classes[i]) { + last_level.new_explicit_next_rtl() + } else { + last_level.new_explicit_next_ltr() + }; + if new_level.is_ok() && overflow_isolate_count == 0 && overflow_embedding_count == 0 + { + let new_level = new_level.unwrap(); + stack.push( + new_level, + match original_classes[i] { + RLO => OverrideStatus::RTL, + LRO => OverrideStatus::LTR, + RLI | LRI | FSI => OverrideStatus::Isolate, + _ => OverrideStatus::Neutral, + }, + ); + if is_isolate { + valid_isolate_count += 1; + } else { + // The spec doesn't explicitly mention this step, but it is necessary. + // See the reference implementations for comparison. + levels[i] = new_level; + } + } else if is_isolate { + overflow_isolate_count += 1; + } else if overflow_isolate_count == 0 { + overflow_embedding_count += 1; + } + } + + // <http://www.unicode.org/reports/tr9/#X6a> + PDI => { + if overflow_isolate_count > 0 { + overflow_isolate_count -= 1; + } else if valid_isolate_count > 0 { + overflow_embedding_count = 0; + loop { + // Pop everything up to and including the last Isolate status. + match stack.vec.pop() { + None + | Some(Status { + status: OverrideStatus::Isolate, + .. + }) => break, + _ => continue, + } + } + valid_isolate_count -= 1; + } + let last = stack.last(); + levels[i] = last.level; + match last.status { + OverrideStatus::RTL => processing_classes[i] = R, + OverrideStatus::LTR => processing_classes[i] = L, + _ => {} + } + } + + // <http://www.unicode.org/reports/tr9/#X7> + PDF => { + if overflow_isolate_count > 0 { + continue; + } + if overflow_embedding_count > 0 { + overflow_embedding_count -= 1; + continue; + } + if stack.last().status != OverrideStatus::Isolate && stack.vec.len() >= 2 { + stack.vec.pop(); + } + // The spec doesn't explicitly mention this step, but it is necessary. + // See the reference implementations for comparison. + levels[i] = stack.last().level; + } + + // Nothing + B | BN => {} + + // <http://www.unicode.org/reports/tr9/#X6> + _ => { + let last = stack.last(); + levels[i] = last.level; + match last.status { + OverrideStatus::RTL => processing_classes[i] = R, + OverrideStatus::LTR => processing_classes[i] = L, + _ => {} + } + } + } + + // Handle multi-byte characters. + for j in 1..c.len_utf8() { + levels[i + j] = levels[i]; + processing_classes[i + j] = processing_classes[i]; + } + } +} + +/// Entries in the directional status stack: +struct Status { + level: Level, + status: OverrideStatus, +} + +#[derive(PartialEq)] +enum OverrideStatus { + Neutral, + RTL, + LTR, + Isolate, +} + +struct DirectionalStatusStack { + vec: Vec<Status>, +} + +impl DirectionalStatusStack { + fn new() -> Self { + DirectionalStatusStack { + vec: Vec::with_capacity(Level::max_explicit_depth() as usize + 2), + } + } + + fn push(&mut self, level: Level, status: OverrideStatus) { + self.vec.push(Status { level, status }); + } + + fn last(&self) -> &Status { + self.vec.last().unwrap() + } +} diff --git a/vendor/unicode-bidi/src/format_chars.rs b/vendor/unicode-bidi/src/format_chars.rs new file mode 100644 index 000000000..5b19b5e76 --- /dev/null +++ b/vendor/unicode-bidi/src/format_chars.rs @@ -0,0 +1,42 @@ +// Copyright 2017 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! Directional Formatting Characters +//! +//! <http://www.unicode.org/reports/tr9/#Directional_Formatting_Characters> + +// == Implicit == +/// ARABIC LETTER MARK +pub const ALM: char = '\u{061C}'; +/// LEFT-TO-RIGHT MARK +pub const LRM: char = '\u{200E}'; +/// RIGHT-TO-LEFT MARK +pub const RLM: char = '\u{200F}'; + +// == Explicit Isolates == +/// LEFT‑TO‑RIGHT ISOLATE +pub const LRI: char = '\u{2066}'; +/// RIGHT‑TO‑LEFT ISOLATE +pub const RLI: char = '\u{2067}'; +/// FIRST STRONG ISOLATE +pub const FSI: char = '\u{2068}'; +/// POP DIRECTIONAL ISOLATE +pub const PDI: char = '\u{2069}'; + +// == Explicit Embeddings and Overrides == +/// LEFT-TO-RIGHT EMBEDDING +pub const LRE: char = '\u{202A}'; +/// RIGHT-TO-LEFT EMBEDDING +pub const RLE: char = '\u{202B}'; +/// POP DIRECTIONAL FORMATTING +pub const PDF: char = '\u{202C}'; +/// LEFT-TO-RIGHT OVERRIDE +pub const LRO: char = '\u{202D}'; +/// RIGHT-TO-LEFT OVERRIDE +pub const RLO: char = '\u{202E}'; diff --git a/vendor/unicode-bidi/src/implicit.rs b/vendor/unicode-bidi/src/implicit.rs new file mode 100644 index 000000000..bf37f17e6 --- /dev/null +++ b/vendor/unicode-bidi/src/implicit.rs @@ -0,0 +1,239 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! 3.3.4 - 3.3.6. Resolve implicit levels and types. + +use alloc::vec::Vec; +use core::cmp::max; + +use super::char_data::BidiClass::{self, *}; +use super::level::Level; +use super::prepare::{not_removed_by_x9, removed_by_x9, IsolatingRunSequence, LevelRun}; + +/// 3.3.4 Resolving Weak Types +/// +/// <http://www.unicode.org/reports/tr9/#Resolving_Weak_Types> +#[cfg_attr(feature = "flame_it", flamer::flame)] +pub fn resolve_weak(sequence: &IsolatingRunSequence, processing_classes: &mut [BidiClass]) { + // FIXME (#8): This function applies steps W1-W6 in a single pass. This can produce + // incorrect results in cases where a "later" rule changes the value of `prev_class` seen + // by an "earlier" rule. We should either split this into separate passes, or preserve + // extra state so each rule can see the correct previous class. + + // FIXME: Also, this could be the cause of increased failure for using longer-UTF-8 chars in + // conformance tests, like BidiTest:69635 (AL ET EN) + + let mut prev_class = sequence.sos; + let mut last_strong_is_al = false; + let mut et_run_indices = Vec::new(); // for W5 + + // Like sequence.runs.iter().flat_map(Clone::clone), but make indices itself clonable. + fn id(x: LevelRun) -> LevelRun { + x + } + let mut indices = sequence + .runs + .iter() + .cloned() + .flat_map(id as fn(LevelRun) -> LevelRun); + + while let Some(i) = indices.next() { + match processing_classes[i] { + // <http://www.unicode.org/reports/tr9/#W1> + NSM => { + processing_classes[i] = match prev_class { + RLI | LRI | FSI | PDI => ON, + _ => prev_class, + }; + } + EN => { + if last_strong_is_al { + // W2. If previous strong char was AL, change EN to AN. + processing_classes[i] = AN; + } else { + // W5. If a run of ETs is adjacent to an EN, change the ETs to EN. + for j in &et_run_indices { + processing_classes[*j] = EN; + } + et_run_indices.clear(); + } + } + // <http://www.unicode.org/reports/tr9/#W3> + AL => processing_classes[i] = R, + + // <http://www.unicode.org/reports/tr9/#W4> + ES | CS => { + let next_class = indices + .clone() + .map(|j| processing_classes[j]) + .find(not_removed_by_x9) + .unwrap_or(sequence.eos); + processing_classes[i] = match (prev_class, processing_classes[i], next_class) { + (EN, ES, EN) | (EN, CS, EN) => EN, + (AN, CS, AN) => AN, + (_, _, _) => ON, + } + } + // <http://www.unicode.org/reports/tr9/#W5> + ET => { + match prev_class { + EN => processing_classes[i] = EN, + _ => et_run_indices.push(i), // In case this is followed by an EN. + } + } + class => { + if removed_by_x9(class) { + continue; + } + } + } + + prev_class = processing_classes[i]; + match prev_class { + L | R => { + last_strong_is_al = false; + } + AL => { + last_strong_is_al = true; + } + _ => {} + } + if prev_class != ET { + // W6. If we didn't find an adjacent EN, turn any ETs into ON instead. + for j in &et_run_indices { + processing_classes[*j] = ON; + } + et_run_indices.clear(); + } + } + + // W7. If the previous strong char was L, change EN to L. + let mut last_strong_is_l = sequence.sos == L; + for run in &sequence.runs { + for i in run.clone() { + match processing_classes[i] { + EN if last_strong_is_l => { + processing_classes[i] = L; + } + L => { + last_strong_is_l = true; + } + R | AL => { + last_strong_is_l = false; + } + _ => {} + } + } + } +} + +/// 3.3.5 Resolving Neutral Types +/// +/// <http://www.unicode.org/reports/tr9/#Resolving_Neutral_Types> +#[cfg_attr(feature = "flame_it", flamer::flame)] +pub fn resolve_neutral( + sequence: &IsolatingRunSequence, + levels: &[Level], + processing_classes: &mut [BidiClass], +) { + let e: BidiClass = levels[sequence.runs[0].start].bidi_class(); + let mut indices = sequence.runs.iter().flat_map(Clone::clone); + let mut prev_class = sequence.sos; + + while let Some(mut i) = indices.next() { + // N0. Process bracket pairs. + // TODO + + // Process sequences of NI characters. + let mut ni_run = Vec::new(); + if is_NI(processing_classes[i]) { + // Consume a run of consecutive NI characters. + ni_run.push(i); + let mut next_class; + loop { + match indices.next() { + Some(j) => { + i = j; + if removed_by_x9(processing_classes[i]) { + continue; + } + next_class = processing_classes[j]; + if is_NI(next_class) { + ni_run.push(i); + } else { + break; + } + } + None => { + next_class = sequence.eos; + break; + } + }; + } + + // N1-N2. + // + // <http://www.unicode.org/reports/tr9/#N1> + // <http://www.unicode.org/reports/tr9/#N2> + let new_class = match (prev_class, next_class) { + (L, L) => L, + (R, R) + | (R, AN) + | (R, EN) + | (AN, R) + | (AN, AN) + | (AN, EN) + | (EN, R) + | (EN, AN) + | (EN, EN) => R, + (_, _) => e, + }; + for j in &ni_run { + processing_classes[*j] = new_class; + } + ni_run.clear(); + } + prev_class = processing_classes[i]; + } +} + +/// 3.3.6 Resolving Implicit Levels +/// +/// Returns the maximum embedding level in the paragraph. +/// +/// <http://www.unicode.org/reports/tr9/#Resolving_Implicit_Levels> +#[cfg_attr(feature = "flame_it", flamer::flame)] +pub fn resolve_levels(original_classes: &[BidiClass], levels: &mut [Level]) -> Level { + let mut max_level = Level::ltr(); + + assert_eq!(original_classes.len(), levels.len()); + for i in 0..levels.len() { + match (levels[i].is_rtl(), original_classes[i]) { + (false, AN) | (false, EN) => levels[i].raise(2).expect("Level number error"), + (false, R) | (true, L) | (true, EN) | (true, AN) => { + levels[i].raise(1).expect("Level number error") + } + (_, _) => {} + } + max_level = max(max_level, levels[i]); + } + + max_level +} + +/// Neutral or Isolate formatting character (B, S, WS, ON, FSI, LRI, RLI, PDI) +/// +/// <http://www.unicode.org/reports/tr9/#NI> +#[allow(non_snake_case)] +fn is_NI(class: BidiClass) -> bool { + match class { + B | S | WS | ON | FSI | LRI | RLI | PDI => true, + _ => false, + } +} diff --git a/vendor/unicode-bidi/src/level.rs b/vendor/unicode-bidi/src/level.rs new file mode 100644 index 000000000..f2e0d99ab --- /dev/null +++ b/vendor/unicode-bidi/src/level.rs @@ -0,0 +1,385 @@ +// Copyright 2017 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! Bidi Embedding Level +//! +//! See [`Level`](struct.Level.html) for more details. +//! +//! <http://www.unicode.org/reports/tr9/#BD2> + +use alloc::string::{String, ToString}; +use alloc::vec::Vec; +use core::convert::{From, Into}; + +use super::char_data::BidiClass; + +/// Embedding Level +/// +/// Embedding Levels are numbers between 0 and 126 (inclusive), where even values denote a +/// left-to-right (LTR) direction and odd values a right-to-left (RTL) direction. +/// +/// This struct maintains a *valid* status for level numbers, meaning that creating a new level, or +/// mutating an existing level, with the value smaller than `0` (before conversion to `u8`) or +/// larger than 125 results in an `Error`. +/// +/// <http://www.unicode.org/reports/tr9/#BD2> +#[derive(Copy, Clone, Debug, Eq, Ord, PartialEq, PartialOrd)] +#[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))] +pub struct Level(u8); + +pub const LTR_LEVEL: Level = Level(0); +pub const RTL_LEVEL: Level = Level(1); + +const MAX_DEPTH: u8 = 125; +/// During explicit level resolution, embedding level can go as high as `max_depth`. +pub const MAX_EXPLICIT_DEPTH: u8 = MAX_DEPTH; +/// During implicit level resolution, embedding level can go as high as `max_depth + 1`. +pub const MAX_IMPLICIT_DEPTH: u8 = MAX_DEPTH + 1; + +/// Errors that can occur on Level creation or mutation +#[derive(Debug, PartialEq)] +pub enum Error { + /// Out-of-range (invalid) embedding level number. + OutOfRangeNumber, +} + +impl Level { + /// New LTR level with smallest number value (0). + #[inline] + pub fn ltr() -> Level { + LTR_LEVEL + } + + /// New RTL level with smallest number value (1). + #[inline] + pub fn rtl() -> Level { + RTL_LEVEL + } + + /// Maximum depth of the directional status stack during implicit resolutions. + pub fn max_implicit_depth() -> u8 { + MAX_IMPLICIT_DEPTH + } + + /// Maximum depth of the directional status stack during explicit resolutions. + pub fn max_explicit_depth() -> u8 { + MAX_EXPLICIT_DEPTH + } + + // == Inquiries == + + /// Create new level, fail if number is larger than `max_depth + 1`. + #[inline] + pub fn new(number: u8) -> Result<Level, Error> { + if number <= MAX_IMPLICIT_DEPTH { + Ok(Level(number)) + } else { + Err(Error::OutOfRangeNumber) + } + } + + /// Create new level, fail if number is larger than `max_depth`. + #[inline] + pub fn new_explicit(number: u8) -> Result<Level, Error> { + if number <= MAX_EXPLICIT_DEPTH { + Ok(Level(number)) + } else { + Err(Error::OutOfRangeNumber) + } + } + + // == Inquiries == + + /// The level number. + #[inline] + pub fn number(&self) -> u8 { + self.0 + } + + /// If this level is left-to-right. + #[inline] + pub fn is_ltr(&self) -> bool { + self.0 % 2 == 0 + } + + /// If this level is right-to-left. + #[inline] + pub fn is_rtl(&self) -> bool { + self.0 % 2 == 1 + } + + // == Mutators == + + /// Raise level by `amount`, fail if number is larger than `max_depth + 1`. + #[inline] + pub fn raise(&mut self, amount: u8) -> Result<(), Error> { + match self.0.checked_add(amount) { + Some(number) => { + if number <= MAX_IMPLICIT_DEPTH { + self.0 = number; + Ok(()) + } else { + Err(Error::OutOfRangeNumber) + } + } + None => Err(Error::OutOfRangeNumber), + } + } + + /// Raise level by `amount`, fail if number is larger than `max_depth`. + #[inline] + pub fn raise_explicit(&mut self, amount: u8) -> Result<(), Error> { + match self.0.checked_add(amount) { + Some(number) => { + if number <= MAX_EXPLICIT_DEPTH { + self.0 = number; + Ok(()) + } else { + Err(Error::OutOfRangeNumber) + } + } + None => Err(Error::OutOfRangeNumber), + } + } + + /// Lower level by `amount`, fail if number goes below zero. + #[inline] + pub fn lower(&mut self, amount: u8) -> Result<(), Error> { + match self.0.checked_sub(amount) { + Some(number) => { + self.0 = number; + Ok(()) + } + None => Err(Error::OutOfRangeNumber), + } + } + + // == Helpers == + + /// The next LTR (even) level greater than this, or fail if number is larger than `max_depth`. + #[inline] + pub fn new_explicit_next_ltr(&self) -> Result<Level, Error> { + Level::new_explicit((self.0 + 2) & !1) + } + + /// The next RTL (odd) level greater than this, or fail if number is larger than `max_depth`. + #[inline] + pub fn new_explicit_next_rtl(&self) -> Result<Level, Error> { + Level::new_explicit((self.0 + 1) | 1) + } + + /// The lowest RTL (odd) level greater than or equal to this, or fail if number is larger than + /// `max_depth + 1`. + #[inline] + pub fn new_lowest_ge_rtl(&self) -> Result<Level, Error> { + Level::new(self.0 | 1) + } + + /// Generate a character type based on a level (as specified in steps X10 and N2). + #[inline] + pub fn bidi_class(&self) -> BidiClass { + if self.is_rtl() { + BidiClass::R + } else { + BidiClass::L + } + } + + pub fn vec(v: &[u8]) -> Vec<Level> { + v.iter().map(|&x| x.into()).collect() + } +} + +/// If levels has any RTL (odd) level +/// +/// This information is usually used to skip re-ordering of text when no RTL level is present +#[inline] +pub fn has_rtl(levels: &[Level]) -> bool { + levels.iter().any(|&lvl| lvl.is_rtl()) +} + +impl Into<u8> for Level { + /// Convert to the level number + #[inline] + fn into(self) -> u8 { + self.number() + } +} + +impl From<u8> for Level { + /// Create level by number + #[inline] + fn from(number: u8) -> Level { + Level::new(number).expect("Level number error") + } +} + +/// Used for matching levels in conformance tests +impl<'a> PartialEq<&'a str> for Level { + #[inline] + fn eq(&self, s: &&'a str) -> bool { + *s == "x" || *s == self.0.to_string() + } +} + +/// Used for matching levels in conformance tests +impl<'a> PartialEq<String> for Level { + #[inline] + fn eq(&self, s: &String) -> bool { + self == &s.as_str() + } +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn test_new() { + assert_eq!(Level::new(0), Ok(Level(0))); + assert_eq!(Level::new(1), Ok(Level(1))); + assert_eq!(Level::new(10), Ok(Level(10))); + assert_eq!(Level::new(125), Ok(Level(125))); + assert_eq!(Level::new(126), Ok(Level(126))); + assert_eq!(Level::new(127), Err(Error::OutOfRangeNumber)); + assert_eq!(Level::new(255), Err(Error::OutOfRangeNumber)); + } + + #[test] + fn test_new_explicit() { + assert_eq!(Level::new_explicit(0), Ok(Level(0))); + assert_eq!(Level::new_explicit(1), Ok(Level(1))); + assert_eq!(Level::new_explicit(10), Ok(Level(10))); + assert_eq!(Level::new_explicit(125), Ok(Level(125))); + assert_eq!(Level::new_explicit(126), Err(Error::OutOfRangeNumber)); + assert_eq!(Level::new_explicit(255), Err(Error::OutOfRangeNumber)); + } + + #[test] + fn test_is_ltr() { + assert_eq!(Level(0).is_ltr(), true); + assert_eq!(Level(1).is_ltr(), false); + assert_eq!(Level(10).is_ltr(), true); + assert_eq!(Level(11).is_ltr(), false); + assert_eq!(Level(124).is_ltr(), true); + assert_eq!(Level(125).is_ltr(), false); + } + + #[test] + fn test_is_rtl() { + assert_eq!(Level(0).is_rtl(), false); + assert_eq!(Level(1).is_rtl(), true); + assert_eq!(Level(10).is_rtl(), false); + assert_eq!(Level(11).is_rtl(), true); + assert_eq!(Level(124).is_rtl(), false); + assert_eq!(Level(125).is_rtl(), true); + } + + #[test] + fn test_raise() { + let mut level = Level::ltr(); + assert_eq!(level.number(), 0); + assert!(level.raise(100).is_ok()); + assert_eq!(level.number(), 100); + assert!(level.raise(26).is_ok()); + assert_eq!(level.number(), 126); + assert!(level.raise(1).is_err()); // invalid! + assert!(level.raise(250).is_err()); // overflow! + assert_eq!(level.number(), 126); + } + + #[test] + fn test_raise_explicit() { + let mut level = Level::ltr(); + assert_eq!(level.number(), 0); + assert!(level.raise_explicit(100).is_ok()); + assert_eq!(level.number(), 100); + assert!(level.raise_explicit(25).is_ok()); + assert_eq!(level.number(), 125); + assert!(level.raise_explicit(1).is_err()); // invalid! + assert!(level.raise_explicit(250).is_err()); // overflow! + assert_eq!(level.number(), 125); + } + + #[test] + fn test_lower() { + let mut level = Level::rtl(); + assert_eq!(level.number(), 1); + assert!(level.lower(1).is_ok()); + assert_eq!(level.number(), 0); + assert!(level.lower(1).is_err()); // underflow! + assert!(level.lower(250).is_err()); // underflow! + assert_eq!(level.number(), 0); + } + + #[test] + fn test_has_rtl() { + assert_eq!(has_rtl(&Level::vec(&[0, 0, 0])), false); + assert_eq!(has_rtl(&Level::vec(&[0, 1, 0])), true); + assert_eq!(has_rtl(&Level::vec(&[0, 2, 0])), false); + assert_eq!(has_rtl(&Level::vec(&[0, 125, 0])), true); + assert_eq!(has_rtl(&Level::vec(&[0, 126, 0])), false); + } + + #[test] + fn test_into() { + let level = Level::rtl(); + let number: u8 = level.into(); + assert_eq!(1u8, number); + } + + #[test] + fn test_vec() { + assert_eq!( + Level::vec(&[0, 1, 125]), + vec![Level(0), Level(1), Level(125)] + ); + } + + #[test] + fn test_str_eq() { + assert_eq!(Level::vec(&[0, 1, 4, 125]), vec!["0", "1", "x", "125"]); + assert_ne!(Level::vec(&[0, 1, 4, 125]), vec!["0", "1", "5", "125"]); + } + + #[test] + fn test_string_eq() { + assert_eq!( + Level::vec(&[0, 1, 4, 125]), + vec!["0".to_string(), "1".to_string(), "x".to_string(), "125".to_string()] + ); + } +} + +#[cfg(all(feature = "serde", test))] +mod serde_tests { + use super::*; + use serde_test::{assert_tokens, Token}; + + #[test] + fn test_statics() { + assert_tokens( + &Level::ltr(), + &[Token::NewtypeStruct { name: "Level" }, Token::U8(0)], + ); + assert_tokens( + &Level::rtl(), + &[Token::NewtypeStruct { name: "Level" }, Token::U8(1)], + ); + } + + #[test] + fn test_new() { + let level = Level::new(42).unwrap(); + assert_tokens( + &level, + &[Token::NewtypeStruct { name: "Level" }, Token::U8(42)], + ); + } +} diff --git a/vendor/unicode-bidi/src/lib.rs b/vendor/unicode-bidi/src/lib.rs new file mode 100644 index 000000000..bda9dd8ec --- /dev/null +++ b/vendor/unicode-bidi/src/lib.rs @@ -0,0 +1,1066 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! This crate implements the [Unicode Bidirectional Algorithm][tr9] for display of mixed +//! right-to-left and left-to-right text. It is written in safe Rust, compatible with the +//! current stable release. +//! +//! ## Example +//! +//! ```rust +//! # #[cfg(feature = "hardcoded-data")] { +//! use unicode_bidi::BidiInfo; +//! +//! // This example text is defined using `concat!` because some browsers +//! // and text editors have trouble displaying bidi strings. +//! let text = concat![ +//! "א", +//! "ב", +//! "ג", +//! "a", +//! "b", +//! "c", +//! ]; +//! +//! // Resolve embedding levels within the text. Pass `None` to detect the +//! // paragraph level automatically. +//! let bidi_info = BidiInfo::new(&text, None); +//! +//! // This paragraph has embedding level 1 because its first strong character is RTL. +//! assert_eq!(bidi_info.paragraphs.len(), 1); +//! let para = &bidi_info.paragraphs[0]; +//! assert_eq!(para.level.number(), 1); +//! assert_eq!(para.level.is_rtl(), true); +//! +//! // Re-ordering is done after wrapping each paragraph into a sequence of +//! // lines. For this example, I'll just use a single line that spans the +//! // entire paragraph. +//! let line = para.range.clone(); +//! +//! let display = bidi_info.reorder_line(para, line); +//! assert_eq!(display, concat![ +//! "a", +//! "b", +//! "c", +//! "ג", +//! "ב", +//! "א", +//! ]); +//! # } // feature = "hardcoded-data" +//! ``` +//! +//! # Features +//! +//! - `std`: Enabled by default, but can be disabled to make `unicode_bidi` +//! `#![no_std]` + `alloc` compatible. +//! - `hardcoded-data`: Enabled by default. Includes hardcoded Unicode bidi data and more convenient APIs. +//! - `serde`: Adds [`serde::Serialize`] and [`serde::Deserialize`] +//! implementations to relevant types. +//! +//! [tr9]: <http://www.unicode.org/reports/tr9/> + +#![forbid(unsafe_code)] +#![no_std] +// We need to link to std to make doc tests work on older Rust versions +#[cfg(feature = "std")] +extern crate std; +#[macro_use] +extern crate alloc; + +pub mod data_source; +pub mod deprecated; +pub mod format_chars; +pub mod level; + +mod char_data; +mod explicit; +mod implicit; +mod prepare; + +pub use crate::char_data::{BidiClass, UNICODE_VERSION}; +pub use crate::data_source::BidiDataSource; +pub use crate::level::{Level, LTR_LEVEL, RTL_LEVEL}; +pub use crate::prepare::LevelRun; + +#[cfg(feature = "hardcoded-data")] +pub use crate::char_data::{bidi_class, HardcodedBidiData}; + +use alloc::borrow::Cow; +use alloc::string::String; +use alloc::vec::Vec; +use core::cmp::{max, min}; +use core::iter::repeat; +use core::ops::Range; + +use crate::format_chars as chars; +use crate::BidiClass::*; + +#[derive(PartialEq, Debug)] +pub enum Direction { + Ltr, + Rtl, + Mixed, +} + +/// Bidi information about a single paragraph +#[derive(Debug, PartialEq)] +pub struct ParagraphInfo { + /// The paragraphs boundaries within the text, as byte indices. + /// + /// TODO: Shrink this to only include the starting index? + pub range: Range<usize>, + + /// The paragraph embedding level. + /// + /// <http://www.unicode.org/reports/tr9/#BD4> + pub level: Level, +} + +impl ParagraphInfo { + /// Gets the length of the paragraph in the source text. + pub fn len(&self) -> usize { + self.range.end - self.range.start + } +} + +/// Initial bidi information of the text. +/// +/// Contains the text paragraphs and `BidiClass` of its characters. +#[derive(PartialEq, Debug)] +pub struct InitialInfo<'text> { + /// The text + pub text: &'text str, + + /// The BidiClass of the character at each byte in the text. + /// If a character is multiple bytes, its class will appear multiple times in the vector. + pub original_classes: Vec<BidiClass>, + + /// The boundaries and level of each paragraph within the text. + pub paragraphs: Vec<ParagraphInfo>, +} + +impl<'text> InitialInfo<'text> { + /// Find the paragraphs and BidiClasses in a string of text. + /// + /// <http://www.unicode.org/reports/tr9/#The_Paragraph_Level> + /// + /// Also sets the class for each First Strong Isolate initiator (FSI) to LRI or RLI if a strong + /// character is found before the matching PDI. If no strong character is found, the class will + /// remain FSI, and it's up to later stages to treat these as LRI when needed. + /// + /// The `hardcoded-data` Cargo feature (enabled by default) must be enabled to use this. + #[cfg_attr(feature = "flame_it", flamer::flame)] + #[cfg(feature = "hardcoded-data")] + pub fn new(text: &str, default_para_level: Option<Level>) -> InitialInfo<'_> { + Self::new_with_data_source(&HardcodedBidiData, text, default_para_level) + } + + /// Find the paragraphs and BidiClasses in a string of text, with a custom [`BidiDataSource`] + /// for Bidi data. If you just wish to use the hardcoded Bidi data, please use [`InitialInfo::new()`] + /// instead (enabled with tbe default `hardcoded-data` Cargo feature) + /// + /// <http://www.unicode.org/reports/tr9/#The_Paragraph_Level> + /// + /// Also sets the class for each First Strong Isolate initiator (FSI) to LRI or RLI if a strong + /// character is found before the matching PDI. If no strong character is found, the class will + /// remain FSI, and it's up to later stages to treat these as LRI when needed. + #[cfg_attr(feature = "flame_it", flamer::flame)] + pub fn new_with_data_source<'a, D: BidiDataSource>( + data_source: &D, + text: &'a str, + default_para_level: Option<Level>, + ) -> InitialInfo<'a> { + let mut original_classes = Vec::with_capacity(text.len()); + + // The stack contains the starting byte index for each nested isolate we're inside. + let mut isolate_stack = Vec::new(); + let mut paragraphs = Vec::new(); + + let mut para_start = 0; + let mut para_level = default_para_level; + + #[cfg(feature = "flame_it")] + flame::start("InitialInfo::new(): iter text.char_indices()"); + + for (i, c) in text.char_indices() { + let class = data_source.bidi_class(c); + + #[cfg(feature = "flame_it")] + flame::start("original_classes.extend()"); + + original_classes.extend(repeat(class).take(c.len_utf8())); + + #[cfg(feature = "flame_it")] + flame::end("original_classes.extend()"); + + match class { + B => { + // P1. Split the text into separate paragraphs. The paragraph separator is kept + // with the previous paragraph. + let para_end = i + c.len_utf8(); + paragraphs.push(ParagraphInfo { + range: para_start..para_end, + // P3. If no character is found in p2, set the paragraph level to zero. + level: para_level.unwrap_or(LTR_LEVEL), + }); + // Reset state for the start of the next paragraph. + para_start = para_end; + // TODO: Support defaulting to direction of previous paragraph + // + // <http://www.unicode.org/reports/tr9/#HL1> + para_level = default_para_level; + isolate_stack.clear(); + } + + L | R | AL => { + match isolate_stack.last() { + Some(&start) => { + if original_classes[start] == FSI { + // X5c. If the first strong character between FSI and its matching + // PDI is R or AL, treat it as RLI. Otherwise, treat it as LRI. + for j in 0..chars::FSI.len_utf8() { + original_classes[start + j] = + if class == L { LRI } else { RLI }; + } + } + } + + None => { + if para_level.is_none() { + // P2. Find the first character of type L, AL, or R, while skipping + // any characters between an isolate initiator and its matching + // PDI. + para_level = Some(if class != L { RTL_LEVEL } else { LTR_LEVEL }); + } + } + } + } + + RLI | LRI | FSI => { + isolate_stack.push(i); + } + + PDI => { + isolate_stack.pop(); + } + + _ => {} + } + } + if para_start < text.len() { + paragraphs.push(ParagraphInfo { + range: para_start..text.len(), + level: para_level.unwrap_or(LTR_LEVEL), + }); + } + assert_eq!(original_classes.len(), text.len()); + + #[cfg(feature = "flame_it")] + flame::end("InitialInfo::new(): iter text.char_indices()"); + + InitialInfo { + text, + original_classes, + paragraphs, + } + } +} + +/// Bidi information of the text. +/// +/// The `original_classes` and `levels` vectors are indexed by byte offsets into the text. If a +/// character is multiple bytes wide, then its class and level will appear multiple times in these +/// vectors. +// TODO: Impl `struct StringProperty<T> { values: Vec<T> }` and use instead of Vec<T> +#[derive(Debug, PartialEq)] +pub struct BidiInfo<'text> { + /// The text + pub text: &'text str, + + /// The BidiClass of the character at each byte in the text. + pub original_classes: Vec<BidiClass>, + + /// The directional embedding level of each byte in the text. + pub levels: Vec<Level>, + + /// The boundaries and paragraph embedding level of each paragraph within the text. + /// + /// TODO: Use SmallVec or similar to avoid overhead when there are only one or two paragraphs? + /// Or just don't include the first paragraph, which always starts at 0? + pub paragraphs: Vec<ParagraphInfo>, +} + +impl<'text> BidiInfo<'text> { + /// Split the text into paragraphs and determine the bidi embedding levels for each paragraph. + /// + /// + /// The `hardcoded-data` Cargo feature (enabled by default) must be enabled to use this. + /// + /// TODO: In early steps, check for special cases that allow later steps to be skipped. like + /// text that is entirely LTR. See the `nsBidi` class from Gecko for comparison. + /// + /// TODO: Support auto-RTL base direction + #[cfg_attr(feature = "flame_it", flamer::flame)] + #[cfg(feature = "hardcoded-data")] + pub fn new(text: &str, default_para_level: Option<Level>) -> BidiInfo<'_> { + Self::new_with_data_source(&HardcodedBidiData, text, default_para_level) + } + + /// Split the text into paragraphs and determine the bidi embedding levels for each paragraph, with a custom [`BidiDataSource`] + /// for Bidi data. If you just wish to use the hardcoded Bidi data, please use [`BidiInfo::new()`] + /// instead (enabled with tbe default `hardcoded-data` Cargo feature). + /// + /// TODO: In early steps, check for special cases that allow later steps to be skipped. like + /// text that is entirely LTR. See the `nsBidi` class from Gecko for comparison. + /// + /// TODO: Support auto-RTL base direction + #[cfg_attr(feature = "flame_it", flamer::flame)] + pub fn new_with_data_source<'a, D: BidiDataSource>( + data_source: &D, + text: &'a str, + default_para_level: Option<Level>, + ) -> BidiInfo<'a> { + let InitialInfo { + original_classes, + paragraphs, + .. + } = InitialInfo::new_with_data_source(data_source, text, default_para_level); + + let mut levels = Vec::<Level>::with_capacity(text.len()); + let mut processing_classes = original_classes.clone(); + + for para in ¶graphs { + let text = &text[para.range.clone()]; + let original_classes = &original_classes[para.range.clone()]; + let processing_classes = &mut processing_classes[para.range.clone()]; + + let new_len = levels.len() + para.range.len(); + levels.resize(new_len, para.level); + let levels = &mut levels[para.range.clone()]; + + explicit::compute( + text, + para.level, + original_classes, + levels, + processing_classes, + ); + + let sequences = prepare::isolating_run_sequences(para.level, original_classes, levels); + for sequence in &sequences { + implicit::resolve_weak(sequence, processing_classes); + implicit::resolve_neutral(sequence, levels, processing_classes); + } + implicit::resolve_levels(processing_classes, levels); + + assign_levels_to_removed_chars(para.level, original_classes, levels); + } + + BidiInfo { + text, + original_classes, + paragraphs, + levels, + } + } + + /// Re-order a line based on resolved levels and return only the embedding levels, one `Level` + /// per *byte*. + #[cfg_attr(feature = "flame_it", flamer::flame)] + pub fn reordered_levels(&self, para: &ParagraphInfo, line: Range<usize>) -> Vec<Level> { + let (levels, _) = self.visual_runs(para, line); + levels + } + + /// Re-order a line based on resolved levels and return only the embedding levels, one `Level` + /// per *character*. + #[cfg_attr(feature = "flame_it", flamer::flame)] + pub fn reordered_levels_per_char( + &self, + para: &ParagraphInfo, + line: Range<usize>, + ) -> Vec<Level> { + let levels = self.reordered_levels(para, line); + self.text.char_indices().map(|(i, _)| levels[i]).collect() + } + + /// Re-order a line based on resolved levels and return the line in display order. + #[cfg_attr(feature = "flame_it", flamer::flame)] + pub fn reorder_line(&self, para: &ParagraphInfo, line: Range<usize>) -> Cow<'text, str> { + let (levels, runs) = self.visual_runs(para, line.clone()); + + // If all isolating run sequences are LTR, no reordering is needed + if runs.iter().all(|run| levels[run.start].is_ltr()) { + return self.text[line].into(); + } + + let mut result = String::with_capacity(line.len()); + for run in runs { + if levels[run.start].is_rtl() { + result.extend(self.text[run].chars().rev()); + } else { + result.push_str(&self.text[run]); + } + } + result.into() + } + + /// Find the level runs within a line and return them in visual order. + /// + /// `line` is a range of bytes indices within `levels`. + /// + /// <http://www.unicode.org/reports/tr9/#Reordering_Resolved_Levels> + #[cfg_attr(feature = "flame_it", flamer::flame)] + pub fn visual_runs( + &self, + para: &ParagraphInfo, + line: Range<usize>, + ) -> (Vec<Level>, Vec<LevelRun>) { + assert!(line.start <= self.levels.len()); + assert!(line.end <= self.levels.len()); + + let mut levels = self.levels.clone(); + let line_classes = &self.original_classes[line.clone()]; + let line_levels = &mut levels[line.clone()]; + + // Reset some whitespace chars to paragraph level. + // <http://www.unicode.org/reports/tr9/#L1> + let line_str: &str = &self.text[line.clone()]; + let mut reset_from: Option<usize> = Some(0); + let mut reset_to: Option<usize> = None; + for (i, c) in line_str.char_indices() { + match line_classes[i] { + // Ignored by X9 + RLE | LRE | RLO | LRO | PDF | BN => {} + // Segment separator, Paragraph separator + B | S => { + assert_eq!(reset_to, None); + reset_to = Some(i + c.len_utf8()); + if reset_from == None { + reset_from = Some(i); + } + } + // Whitespace, isolate formatting + WS | FSI | LRI | RLI | PDI => { + if reset_from == None { + reset_from = Some(i); + } + } + _ => { + reset_from = None; + } + } + if let (Some(from), Some(to)) = (reset_from, reset_to) { + for level in &mut line_levels[from..to] { + *level = para.level; + } + reset_from = None; + reset_to = None; + } + } + if let Some(from) = reset_from { + for level in &mut line_levels[from..] { + *level = para.level; + } + } + + // Find consecutive level runs. + let mut runs = Vec::new(); + let mut start = line.start; + let mut run_level = levels[start]; + let mut min_level = run_level; + let mut max_level = run_level; + + for (i, &new_level) in levels.iter().enumerate().take(line.end).skip(start + 1) { + if new_level != run_level { + // End of the previous run, start of a new one. + runs.push(start..i); + start = i; + run_level = new_level; + min_level = min(run_level, min_level); + max_level = max(run_level, max_level); + } + } + runs.push(start..line.end); + + let run_count = runs.len(); + + // Re-order the odd runs. + // <http://www.unicode.org/reports/tr9/#L2> + + // Stop at the lowest *odd* level. + min_level = min_level.new_lowest_ge_rtl().expect("Level error"); + + while max_level >= min_level { + // Look for the start of a sequence of consecutive runs of max_level or higher. + let mut seq_start = 0; + while seq_start < run_count { + if self.levels[runs[seq_start].start] < max_level { + seq_start += 1; + continue; + } + + // Found the start of a sequence. Now find the end. + let mut seq_end = seq_start + 1; + while seq_end < run_count { + if self.levels[runs[seq_end].start] < max_level { + break; + } + seq_end += 1; + } + + // Reverse the runs within this sequence. + runs[seq_start..seq_end].reverse(); + + seq_start = seq_end; + } + max_level + .lower(1) + .expect("Lowering embedding level below zero"); + } + + (levels, runs) + } + + /// If processed text has any computed RTL levels + /// + /// This information is usually used to skip re-ordering of text when no RTL level is present + #[inline] + pub fn has_rtl(&self) -> bool { + level::has_rtl(&self.levels) + } +} + +/// Contains a reference of `BidiInfo` and one of its `paragraphs`. +/// And it supports all operation in the `Paragraph` that needs also its +/// `BidiInfo` such as `direction`. +#[derive(Debug)] +pub struct Paragraph<'a, 'text> { + pub info: &'a BidiInfo<'text>, + pub para: &'a ParagraphInfo, +} + +impl<'a, 'text> Paragraph<'a, 'text> { + pub fn new(info: &'a BidiInfo<'text>, para: &'a ParagraphInfo) -> Paragraph<'a, 'text> { + Paragraph { info, para } + } + + /// Returns if the paragraph is Left direction, right direction or mixed. + pub fn direction(&self) -> Direction { + let mut ltr = false; + let mut rtl = false; + for i in self.para.range.clone() { + if self.info.levels[i].is_ltr() { + ltr = true; + } + + if self.info.levels[i].is_rtl() { + rtl = true; + } + } + + if ltr && rtl { + return Direction::Mixed; + } + + if ltr { + return Direction::Ltr; + } + + Direction::Rtl + } + + /// Returns the `Level` of a certain character in the paragraph. + pub fn level_at(&self, pos: usize) -> Level { + let actual_position = self.para.range.start + pos; + self.info.levels[actual_position] + } +} + +/// Assign levels to characters removed by rule X9. +/// +/// The levels assigned to these characters are not specified by the algorithm. This function +/// assigns each one the level of the previous character, to avoid breaking level runs. +#[cfg_attr(feature = "flame_it", flamer::flame)] +fn assign_levels_to_removed_chars(para_level: Level, classes: &[BidiClass], levels: &mut [Level]) { + for i in 0..levels.len() { + if prepare::removed_by_x9(classes[i]) { + levels[i] = if i > 0 { levels[i - 1] } else { para_level }; + } + } +} + +#[cfg(test)] +#[cfg(feature = "hardcoded-data")] +mod tests { + use super::*; + + #[test] + fn test_initial_text_info() { + let text = "a1"; + assert_eq!( + InitialInfo::new(text, None), + InitialInfo { + text, + original_classes: vec![L, EN], + paragraphs: vec![ParagraphInfo { + range: 0..2, + level: LTR_LEVEL, + },], + } + ); + + let text = "غ א"; + assert_eq!( + InitialInfo::new(text, None), + InitialInfo { + text, + original_classes: vec![AL, AL, WS, R, R], + paragraphs: vec![ParagraphInfo { + range: 0..5, + level: RTL_LEVEL, + },], + } + ); + + let text = "a\u{2029}b"; + assert_eq!( + InitialInfo::new(text, None), + InitialInfo { + text, + original_classes: vec![L, B, B, B, L], + paragraphs: vec![ + ParagraphInfo { + range: 0..4, + level: LTR_LEVEL, + }, + ParagraphInfo { + range: 4..5, + level: LTR_LEVEL, + }, + ], + } + ); + + let text = format!("{}א{}a", chars::FSI, chars::PDI); + assert_eq!( + InitialInfo::new(&text, None), + InitialInfo { + text: &text, + original_classes: vec![RLI, RLI, RLI, R, R, PDI, PDI, PDI, L], + paragraphs: vec![ParagraphInfo { + range: 0..9, + level: LTR_LEVEL, + },], + } + ); + } + + #[test] + #[cfg(feature = "hardcoded-data")] + fn test_process_text() { + let text = "abc123"; + assert_eq!( + BidiInfo::new(text, Some(LTR_LEVEL)), + BidiInfo { + text, + levels: Level::vec(&[0, 0, 0, 0, 0, 0]), + original_classes: vec![L, L, L, EN, EN, EN], + paragraphs: vec![ParagraphInfo { + range: 0..6, + level: LTR_LEVEL, + },], + } + ); + + let text = "abc אבג"; + assert_eq!( + BidiInfo::new(text, Some(LTR_LEVEL)), + BidiInfo { + text, + levels: Level::vec(&[0, 0, 0, 0, 1, 1, 1, 1, 1, 1]), + original_classes: vec![L, L, L, WS, R, R, R, R, R, R], + paragraphs: vec![ParagraphInfo { + range: 0..10, + level: LTR_LEVEL, + },], + } + ); + assert_eq!( + BidiInfo::new(text, Some(RTL_LEVEL)), + BidiInfo { + text, + levels: Level::vec(&[2, 2, 2, 1, 1, 1, 1, 1, 1, 1]), + original_classes: vec![L, L, L, WS, R, R, R, R, R, R], + paragraphs: vec![ParagraphInfo { + range: 0..10, + level: RTL_LEVEL, + },], + } + ); + + let text = "אבג abc"; + assert_eq!( + BidiInfo::new(text, Some(LTR_LEVEL)), + BidiInfo { + text, + levels: Level::vec(&[1, 1, 1, 1, 1, 1, 0, 0, 0, 0]), + original_classes: vec![R, R, R, R, R, R, WS, L, L, L], + paragraphs: vec![ParagraphInfo { + range: 0..10, + level: LTR_LEVEL, + },], + } + ); + assert_eq!( + BidiInfo::new(text, None), + BidiInfo { + text, + levels: Level::vec(&[1, 1, 1, 1, 1, 1, 1, 2, 2, 2]), + original_classes: vec![R, R, R, R, R, R, WS, L, L, L], + paragraphs: vec![ParagraphInfo { + range: 0..10, + level: RTL_LEVEL, + },], + } + ); + + let text = "غ2ظ א2ג"; + assert_eq!( + BidiInfo::new(text, Some(LTR_LEVEL)), + BidiInfo { + text, + levels: Level::vec(&[1, 1, 2, 1, 1, 1, 1, 1, 2, 1, 1]), + original_classes: vec![AL, AL, EN, AL, AL, WS, R, R, EN, R, R], + paragraphs: vec![ParagraphInfo { + range: 0..11, + level: LTR_LEVEL, + },], + } + ); + + let text = "a א.\nג"; + assert_eq!( + BidiInfo::new(text, None), + BidiInfo { + text, + original_classes: vec![L, WS, R, R, CS, B, R, R], + levels: Level::vec(&[0, 0, 1, 1, 0, 0, 1, 1]), + paragraphs: vec![ + ParagraphInfo { + range: 0..6, + level: LTR_LEVEL, + }, + ParagraphInfo { + range: 6..8, + level: RTL_LEVEL, + }, + ], + } + ); + + // BidiTest:69635 (AL ET EN) + let bidi_info = BidiInfo::new("\u{060B}\u{20CF}\u{06F9}", None); + assert_eq!(bidi_info.original_classes, vec![AL, AL, ET, ET, ET, EN, EN]); + } + + #[test] + #[cfg(feature = "hardcoded-data")] + fn test_bidi_info_has_rtl() { + // ASCII only + assert_eq!(BidiInfo::new("123", None).has_rtl(), false); + assert_eq!(BidiInfo::new("123", Some(LTR_LEVEL)).has_rtl(), false); + assert_eq!(BidiInfo::new("123", Some(RTL_LEVEL)).has_rtl(), false); + assert_eq!(BidiInfo::new("abc", None).has_rtl(), false); + assert_eq!(BidiInfo::new("abc", Some(LTR_LEVEL)).has_rtl(), false); + assert_eq!(BidiInfo::new("abc", Some(RTL_LEVEL)).has_rtl(), false); + assert_eq!(BidiInfo::new("abc 123", None).has_rtl(), false); + assert_eq!(BidiInfo::new("abc\n123", None).has_rtl(), false); + + // With Hebrew + assert_eq!(BidiInfo::new("אבּג", None).has_rtl(), true); + assert_eq!(BidiInfo::new("אבּג", Some(LTR_LEVEL)).has_rtl(), true); + assert_eq!(BidiInfo::new("אבּג", Some(RTL_LEVEL)).has_rtl(), true); + assert_eq!(BidiInfo::new("abc אבּג", None).has_rtl(), true); + assert_eq!(BidiInfo::new("abc\nאבּג", None).has_rtl(), true); + assert_eq!(BidiInfo::new("אבּג abc", None).has_rtl(), true); + assert_eq!(BidiInfo::new("אבּג\nabc", None).has_rtl(), true); + assert_eq!(BidiInfo::new("אבּג 123", None).has_rtl(), true); + assert_eq!(BidiInfo::new("אבּג\n123", None).has_rtl(), true); + } + + #[cfg(feature = "hardcoded-data")] + fn reorder_paras(text: &str) -> Vec<Cow<'_, str>> { + let bidi_info = BidiInfo::new(text, None); + bidi_info + .paragraphs + .iter() + .map(|para| bidi_info.reorder_line(para, para.range.clone())) + .collect() + } + + #[test] + #[cfg(feature = "hardcoded-data")] + fn test_reorder_line() { + // Bidi_Class: L L L B L L L B L L L + assert_eq!( + reorder_paras("abc\ndef\nghi"), + vec!["abc\n", "def\n", "ghi"] + ); + + // Bidi_Class: L L EN B L L EN B L L EN + assert_eq!( + reorder_paras("ab1\nde2\ngh3"), + vec!["ab1\n", "de2\n", "gh3"] + ); + + // Bidi_Class: L L L B AL AL AL + assert_eq!(reorder_paras("abc\nابج"), vec!["abc\n", "جبا"]); + + // Bidi_Class: AL AL AL B L L L + assert_eq!(reorder_paras("ابج\nabc"), vec!["\nجبا", "abc"]); + + assert_eq!(reorder_paras("1.-2"), vec!["1.-2"]); + assert_eq!(reorder_paras("1-.2"), vec!["1-.2"]); + assert_eq!(reorder_paras("abc אבג"), vec!["abc גבא"]); + + // Numbers being weak LTR characters, cannot reorder strong RTL + assert_eq!(reorder_paras("123 אבג"), vec!["גבא 123"]); + + assert_eq!(reorder_paras("abc\u{202A}def"), vec!["abc\u{202A}def"]); + + assert_eq!( + reorder_paras("abc\u{202A}def\u{202C}ghi"), + vec!["abc\u{202A}def\u{202C}ghi"] + ); + + assert_eq!( + reorder_paras("abc\u{2066}def\u{2069}ghi"), + vec!["abc\u{2066}def\u{2069}ghi"] + ); + + // Testing for RLE Character + assert_eq!( + reorder_paras("\u{202B}abc אבג\u{202C}"), + vec!["\u{202B}\u{202C}גבא abc"] + ); + + // Testing neutral characters + assert_eq!(reorder_paras("אבג? אבג"), vec!["גבא ?גבא"]); + + // Testing neutral characters with special case + assert_eq!(reorder_paras("A אבג?"), vec!["A גבא?"]); + + // Testing neutral characters with Implicit RTL Marker + assert_eq!(reorder_paras("A אבג?\u{200F}"), vec!["A \u{200F}?גבא"]); + assert_eq!(reorder_paras("אבג abc"), vec!["abc גבא"]); + assert_eq!( + reorder_paras("abc\u{2067}.-\u{2069}ghi"), + vec!["abc\u{2067}-.\u{2069}ghi"] + ); + + assert_eq!( + reorder_paras("Hello, \u{2068}\u{202E}world\u{202C}\u{2069}!"), + vec!["Hello, \u{2068}\u{202E}\u{202C}dlrow\u{2069}!"] + ); + + // With mirrorable characters in RTL run + assert_eq!(reorder_paras("א(ב)ג."), vec![".ג)ב(א"]); + + // With mirrorable characters on level boundry + assert_eq!(reorder_paras("אב(גד[&ef].)gh"), vec!["ef].)gh&[דג(בא"]); + } + + fn reordered_levels_for_paras(text: &str) -> Vec<Vec<Level>> { + let bidi_info = BidiInfo::new(text, None); + bidi_info + .paragraphs + .iter() + .map(|para| bidi_info.reordered_levels(para, para.range.clone())) + .collect() + } + + fn reordered_levels_per_char_for_paras(text: &str) -> Vec<Vec<Level>> { + let bidi_info = BidiInfo::new(text, None); + bidi_info + .paragraphs + .iter() + .map(|para| bidi_info.reordered_levels_per_char(para, para.range.clone())) + .collect() + } + + #[test] + #[cfg(feature = "hardcoded-data")] + fn test_reordered_levels() { + // BidiTest:946 (LRI PDI) + let text = "\u{2067}\u{2069}"; + assert_eq!( + reordered_levels_for_paras(text), + vec![Level::vec(&[0, 0, 0, 0, 0, 0])] + ); + assert_eq!( + reordered_levels_per_char_for_paras(text), + vec![Level::vec(&[0, 0])] + ); + + let text = "aa טֶ"; + let bidi_info = BidiInfo::new(text, None); + assert_eq!( + bidi_info.reordered_levels(&bidi_info.paragraphs[0], 3..7), + Level::vec(&[0, 0, 0, 1, 1, 1, 1]), + ) + + /* TODO + /// BidiTest:69635 (AL ET EN) + let text = "\u{060B}\u{20CF}\u{06F9}"; + assert_eq!( + reordered_levels_for_paras(text), + vec![Level::vec(&[1, 1, 1, 1, 1, 2, 2])] + ); + assert_eq!( + reordered_levels_per_char_for_paras(text), + vec![Level::vec(&[1, 1, 2])] + ); + */ + + /* TODO + // BidiTest:291284 (AN RLI PDF R) + assert_eq!( + reordered_levels_per_char_for_paras("\u{0605}\u{2067}\u{202C}\u{0590}"), + vec![&["2", "0", "x", "1"]] + ); + */ + } + + #[test] + fn test_paragraph_info_len() { + let text = "hello world"; + let bidi_info = BidiInfo::new(text, None); + assert_eq!(bidi_info.paragraphs.len(), 1); + assert_eq!(bidi_info.paragraphs[0].len(), text.len()); + + let text2 = "How are you"; + let whole_text = format!("{}\n{}", text, text2); + let bidi_info = BidiInfo::new(&whole_text, None); + assert_eq!(bidi_info.paragraphs.len(), 2); + + // The first paragraph include the paragraph separator. + // TODO: investigate if the paragraph separator character + // should not be part of any paragraph. + assert_eq!(bidi_info.paragraphs[0].len(), text.len() + 1); + assert_eq!(bidi_info.paragraphs[1].len(), text2.len()); + } + + #[test] + fn test_direction() { + let ltr_text = "hello world"; + let rtl_text = "أهلا بكم"; + let all_paragraphs = format!("{}\n{}\n{}{}", ltr_text, rtl_text, ltr_text, rtl_text); + let bidi_info = BidiInfo::new(&all_paragraphs, None); + assert_eq!(bidi_info.paragraphs.len(), 3); + let p_ltr = Paragraph::new(&bidi_info, &bidi_info.paragraphs[0]); + let p_rtl = Paragraph::new(&bidi_info, &bidi_info.paragraphs[1]); + let p_mixed = Paragraph::new(&bidi_info, &bidi_info.paragraphs[2]); + assert_eq!(p_ltr.direction(), Direction::Ltr); + assert_eq!(p_rtl.direction(), Direction::Rtl); + assert_eq!(p_mixed.direction(), Direction::Mixed); + } + + #[test] + fn test_edge_cases_direction() { + // No paragraphs for empty text. + let empty = ""; + let bidi_info = BidiInfo::new(empty, Option::from(RTL_LEVEL)); + assert_eq!(bidi_info.paragraphs.len(), 0); + // The paragraph separator will take the value of the default direction + // which is left to right. + let empty = "\n"; + let bidi_info = BidiInfo::new(empty, None); + assert_eq!(bidi_info.paragraphs.len(), 1); + let p = Paragraph::new(&bidi_info, &bidi_info.paragraphs[0]); + assert_eq!(p.direction(), Direction::Ltr); + // The paragraph separator will take the value of the given initial direction + // which is left to right. + let empty = "\n"; + let bidi_info = BidiInfo::new(empty, Option::from(LTR_LEVEL)); + assert_eq!(bidi_info.paragraphs.len(), 1); + let p = Paragraph::new(&bidi_info, &bidi_info.paragraphs[0]); + assert_eq!(p.direction(), Direction::Ltr); + + // The paragraph separator will take the value of the given initial direction + // which is right to left. + let empty = "\n"; + let bidi_info = BidiInfo::new(empty, Option::from(RTL_LEVEL)); + assert_eq!(bidi_info.paragraphs.len(), 1); + let p = Paragraph::new(&bidi_info, &bidi_info.paragraphs[0]); + assert_eq!(p.direction(), Direction::Rtl); + } + + #[test] + fn test_level_at() { + let ltr_text = "hello world"; + let rtl_text = "أهلا بكم"; + let all_paragraphs = format!("{}\n{}\n{}{}", ltr_text, rtl_text, ltr_text, rtl_text); + let bidi_info = BidiInfo::new(&all_paragraphs, None); + assert_eq!(bidi_info.paragraphs.len(), 3); + + let p_ltr = Paragraph::new(&bidi_info, &bidi_info.paragraphs[0]); + let p_rtl = Paragraph::new(&bidi_info, &bidi_info.paragraphs[1]); + let p_mixed = Paragraph::new(&bidi_info, &bidi_info.paragraphs[2]); + + assert_eq!(p_ltr.level_at(0), LTR_LEVEL); + assert_eq!(p_rtl.level_at(0), RTL_LEVEL); + assert_eq!(p_mixed.level_at(0), LTR_LEVEL); + assert_eq!(p_mixed.info.levels.len(), 54); + assert_eq!(p_mixed.para.range.start, 28); + assert_eq!(p_mixed.level_at(ltr_text.len()), RTL_LEVEL); + } +} + +#[cfg(all(feature = "serde", test))] +mod serde_tests { + use super::*; + use serde_test::{assert_tokens, Token}; + + #[test] + fn test_levels() { + let text = "abc אבג"; + let bidi_info = BidiInfo::new(text, None); + let levels = bidi_info.levels; + assert_eq!(text.as_bytes().len(), 10); + assert_eq!(levels.len(), 10); + assert_tokens( + &levels, + &[ + Token::Seq { len: Some(10) }, + Token::NewtypeStruct { name: "Level" }, + Token::U8(0), + Token::NewtypeStruct { name: "Level" }, + Token::U8(0), + Token::NewtypeStruct { name: "Level" }, + Token::U8(0), + Token::NewtypeStruct { name: "Level" }, + Token::U8(0), + Token::NewtypeStruct { name: "Level" }, + Token::U8(1), + Token::NewtypeStruct { name: "Level" }, + Token::U8(1), + Token::NewtypeStruct { name: "Level" }, + Token::U8(1), + Token::NewtypeStruct { name: "Level" }, + Token::U8(1), + Token::NewtypeStruct { name: "Level" }, + Token::U8(1), + Token::NewtypeStruct { name: "Level" }, + Token::U8(1), + Token::SeqEnd, + ], + ); + } +} diff --git a/vendor/unicode-bidi/src/prepare.rs b/vendor/unicode-bidi/src/prepare.rs new file mode 100644 index 000000000..7b952361a --- /dev/null +++ b/vendor/unicode-bidi/src/prepare.rs @@ -0,0 +1,368 @@ +// Copyright 2015 The Servo Project Developers. See the +// COPYRIGHT file at the top-level directory of this distribution. +// +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your +// option. This file may not be copied, modified, or distributed +// except according to those terms. + +//! 3.3.3 Preparations for Implicit Processing +//! +//! <http://www.unicode.org/reports/tr9/#Preparations_for_Implicit_Processing> + +use alloc::vec::Vec; +use core::cmp::max; +use core::ops::Range; + +use super::level::Level; +use super::BidiClass::{self, *}; + +/// A maximal substring of characters with the same embedding level. +/// +/// Represented as a range of byte indices. +pub type LevelRun = Range<usize>; + +/// Output of `isolating_run_sequences` (steps X9-X10) +#[derive(Debug, PartialEq)] +pub struct IsolatingRunSequence { + pub runs: Vec<LevelRun>, + pub sos: BidiClass, // Start-of-sequence type. + pub eos: BidiClass, // End-of-sequence type. +} + +/// Compute the set of isolating run sequences. +/// +/// An isolating run sequence is a maximal sequence of level runs such that for all level runs +/// except the last one in the sequence, the last character of the run is an isolate initiator +/// whose matching PDI is the first character of the next level run in the sequence. +/// +/// Note: This function does *not* return the sequences in order by their first characters. +#[cfg_attr(feature = "flame_it", flamer::flame)] +pub fn isolating_run_sequences( + para_level: Level, + original_classes: &[BidiClass], + levels: &[Level], +) -> Vec<IsolatingRunSequence> { + let runs = level_runs(levels, original_classes); + + // Compute the set of isolating run sequences. + // <http://www.unicode.org/reports/tr9/#BD13> + let mut sequences = Vec::with_capacity(runs.len()); + + // When we encounter an isolate initiator, we push the current sequence onto the + // stack so we can resume it after the matching PDI. + let mut stack = vec![Vec::new()]; + + for run in runs { + assert!(run.len() > 0); + assert!(!stack.is_empty()); + + let start_class = original_classes[run.start]; + let end_class = original_classes[run.end - 1]; + + let mut sequence = if start_class == PDI && stack.len() > 1 { + // Continue a previous sequence interrupted by an isolate. + stack.pop().unwrap() + } else { + // Start a new sequence. + Vec::new() + }; + + sequence.push(run); + + if let RLI | LRI | FSI = end_class { + // Resume this sequence after the isolate. + stack.push(sequence); + } else { + // This sequence is finished. + sequences.push(sequence); + } + } + // Pop any remaning sequences off the stack. + sequences.extend(stack.into_iter().rev().filter(|seq| !seq.is_empty())); + + // Determine the `sos` and `eos` class for each sequence. + // <http://www.unicode.org/reports/tr9/#X10> + sequences + .into_iter() + .map(|sequence: Vec<LevelRun>| { + assert!(!sequence.is_empty()); + + let start_of_seq = sequence[0].start; + let end_of_seq = sequence[sequence.len() - 1].end; + let seq_level = levels[start_of_seq]; + + #[cfg(test)] + for run in sequence.clone() { + for idx in run { + if not_removed_by_x9(&original_classes[idx]) { + assert_eq!(seq_level, levels[idx]); + } + } + } + + // Get the level of the last non-removed char before the runs. + let pred_level = match original_classes[..start_of_seq] + .iter() + .rposition(not_removed_by_x9) + { + Some(idx) => levels[idx], + None => para_level, + }; + + // Get the level of the next non-removed char after the runs. + let succ_level = if let RLI | LRI | FSI = original_classes[end_of_seq - 1] { + para_level + } else { + match original_classes[end_of_seq..] + .iter() + .position(not_removed_by_x9) + { + Some(idx) => levels[end_of_seq + idx], + None => para_level, + } + }; + + IsolatingRunSequence { + runs: sequence, + sos: max(seq_level, pred_level).bidi_class(), + eos: max(seq_level, succ_level).bidi_class(), + } + }) + .collect() +} + +/// Finds the level runs in a paragraph. +/// +/// <http://www.unicode.org/reports/tr9/#BD7> +fn level_runs(levels: &[Level], original_classes: &[BidiClass]) -> Vec<LevelRun> { + assert_eq!(levels.len(), original_classes.len()); + + let mut runs = Vec::new(); + if levels.is_empty() { + return runs; + } + + let mut current_run_level = levels[0]; + let mut current_run_start = 0; + for i in 1..levels.len() { + if !removed_by_x9(original_classes[i]) && levels[i] != current_run_level { + // End the last run and start a new one. + runs.push(current_run_start..i); + current_run_level = levels[i]; + current_run_start = i; + } + } + runs.push(current_run_start..levels.len()); + + runs +} + +/// Should this character be ignored in steps after X9? +/// +/// <http://www.unicode.org/reports/tr9/#X9> +pub fn removed_by_x9(class: BidiClass) -> bool { + match class { + RLE | LRE | RLO | LRO | PDF | BN => true, + _ => false, + } +} + +// For use as a predicate for `position` / `rposition` +pub fn not_removed_by_x9(class: &BidiClass) -> bool { + !removed_by_x9(*class) +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn test_level_runs() { + assert_eq!(level_runs(&Level::vec(&[]), &[]), &[]); + assert_eq!( + level_runs(&Level::vec(&[0, 0, 0, 1, 1, 2, 0, 0]), &[L; 8]), + &[0..3, 3..5, 5..6, 6..8] + ); + } + + // From <http://www.unicode.org/reports/tr9/#BD13> + #[rustfmt::skip] + #[test] + fn test_isolating_run_sequences() { + + // == Example 1 == + // text1·RLE·text2·PDF·RLE·text3·PDF·text4 + // index 0 1 2 3 4 5 6 7 + let classes = &[L, RLE, L, PDF, RLE, L, PDF, L]; + let levels = &[0, 1, 1, 1, 1, 1, 1, 0]; + let para_level = Level::ltr(); + let mut sequences = isolating_run_sequences(para_level, classes, &Level::vec(levels)); + sequences.sort_by(|a, b| a.runs[0].clone().cmp(b.runs[0].clone())); + assert_eq!( + sequences.iter().map(|s| s.runs.clone()).collect::<Vec<_>>(), + vec![vec![0..2], vec![2..7], vec![7..8]] + ); + + // == Example 2 == + // text1·RLI·text2·PDI·RLI·text3·PDI·text4 + // index 0 1 2 3 4 5 6 7 + let classes = &[L, RLI, L, PDI, RLI, L, PDI, L]; + let levels = &[0, 0, 1, 0, 0, 1, 0, 0]; + let para_level = Level::ltr(); + let mut sequences = isolating_run_sequences(para_level, classes, &Level::vec(levels)); + sequences.sort_by(|a, b| a.runs[0].clone().cmp(b.runs[0].clone())); + assert_eq!( + sequences.iter().map(|s| s.runs.clone()).collect::<Vec<_>>(), + vec![vec![0..2, 3..5, 6..8], vec![2..3], vec![5..6]] + ); + + // == Example 3 == + // text1·RLI·text2·LRI·text3·RLE·text4·PDF·text5·PDI·text6·PDI·text7 + // index 0 1 2 3 4 5 6 7 8 9 10 11 12 + let classes = &[L, RLI, L, LRI, L, RLE, L, PDF, L, PDI, L, PDI, L]; + let levels = &[0, 0, 1, 1, 2, 3, 3, 3, 2, 1, 1, 0, 0]; + let para_level = Level::ltr(); + let mut sequences = isolating_run_sequences(para_level, classes, &Level::vec(levels)); + sequences.sort_by(|a, b| a.runs[0].clone().cmp(b.runs[0].clone())); + assert_eq!( + sequences.iter().map(|s| s.runs.clone()).collect::<Vec<_>>(), + vec![vec![0..2, 11..13], vec![2..4, 9..11], vec![4..6], vec![6..8], vec![8..9]] + ); + } + + // From <http://www.unicode.org/reports/tr9/#X10> + #[rustfmt::skip] + #[test] + fn test_isolating_run_sequences_sos_and_eos() { + + // == Example 1 == + // text1·RLE·text2·LRE·text3·PDF·text4·PDF·RLE·text5·PDF·text6 + // index 0 1 2 3 4 5 6 7 8 9 10 11 + let classes = &[L, RLE, L, LRE, L, PDF, L, PDF, RLE, L, PDF, L]; + let levels = &[0, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 0]; + let para_level = Level::ltr(); + let mut sequences = isolating_run_sequences(para_level, classes, &Level::vec(levels)); + sequences.sort_by(|a, b| a.runs[0].clone().cmp(b.runs[0].clone())); + + // text1 + assert_eq!( + &sequences[0], + &IsolatingRunSequence { + runs: vec![0..2], + sos: L, + eos: R, + } + ); + + // text2 + assert_eq!( + &sequences[1], + &IsolatingRunSequence { + runs: vec![2..4], + sos: R, + eos: L, + } + ); + + // text3 + assert_eq!( + &sequences[2], + &IsolatingRunSequence { + runs: vec![4..6], + sos: L, + eos: L, + } + ); + + // text4 text5 + assert_eq!( + &sequences[3], + &IsolatingRunSequence { + runs: vec![6..11], + sos: L, + eos: R, + } + ); + + // text6 + assert_eq!( + &sequences[4], + &IsolatingRunSequence { + runs: vec![11..12], + sos: R, + eos: L, + } + ); + + // == Example 2 == + // text1·RLI·text2·LRI·text3·PDI·text4·PDI·RLI·text5·PDI·text6 + // index 0 1 2 3 4 5 6 7 8 9 10 11 + let classes = &[L, RLI, L, LRI, L, PDI, L, PDI, RLI, L, PDI, L]; + let levels = &[0, 0, 1, 1, 2, 1, 1, 0, 0, 1, 0, 0]; + let para_level = Level::ltr(); + let mut sequences = isolating_run_sequences(para_level, classes, &Level::vec(levels)); + sequences.sort_by(|a, b| a.runs[0].clone().cmp(b.runs[0].clone())); + + // text1·RLI·PDI·RLI·PDI·text6 + assert_eq!( + &sequences[0], + &IsolatingRunSequence { + runs: vec![0..2, 7..9, 10..12], + sos: L, + eos: L, + } + ); + + // text2·LRI·PDI·text4 + assert_eq!( + &sequences[1], + &IsolatingRunSequence { + runs: vec![2..4, 5..7], + sos: R, + eos: R, + } + ); + + // text3 + assert_eq!( + &sequences[2], + &IsolatingRunSequence { + runs: vec![4..5], + sos: L, + eos: L, + } + ); + + // text5 + assert_eq!( + &sequences[3], + &IsolatingRunSequence { + runs: vec![9..10], + sos: R, + eos: R, + } + ); + } + + #[test] + fn test_removed_by_x9() { + let rem_classes = &[RLE, LRE, RLO, LRO, PDF, BN]; + let not_classes = &[L, RLI, AL, LRI, PDI]; + for x in rem_classes { + assert_eq!(removed_by_x9(*x), true); + } + for x in not_classes { + assert_eq!(removed_by_x9(*x), false); + } + } + + #[test] + fn test_not_removed_by_x9() { + let non_x9_classes = &[L, R, AL, EN, ES, ET, AN, CS, NSM, B, S, WS, ON, LRI, RLI, FSI, PDI]; + for x in non_x9_classes { + assert_eq!(not_removed_by_x9(&x), true); + } + } +} |