Bug 1883464 - Update unicode-bidi to latest upstream revision, to get…

… changes from servo/unicode-bidi#131. r=platform-i18n-reviewers,supply-chain-reviewers,gregtatum

No change in behavior, just internal performance optimizations.

Differential Revision: https://phabricator.services.mozilla.com/D203729

.cargo/config.in

@@ -105,9 +105,9 @@ git = "https://github.com/servo/rust-cssparser"
 rev = "aaa966d9d6ae70c4b8a62bb5e3a14c068bb7dff0"
 replace-with = "vendored-sources"
 
-[source."git+https://github.com/servo/unicode-bidi?rev=d43c644c388659d093cc5d1648a970daeef692d5"]
+[source."git+https://github.com/servo/unicode-bidi?rev=ca612daf1c08c53abe07327cb3e6ef6e0a760f0c"]
 git = "https://github.com/servo/unicode-bidi"
-rev = "d43c644c388659d093cc5d1648a970daeef692d5"
+rev = "ca612daf1c08c53abe07327cb3e6ef6e0a760f0c"
 replace-with = "vendored-sources"
 
 

Cargo.toml

@@ -188,7 +188,7 @@ rure = { path = "third_party/rust/rure" }
 # To-be-published changes.
 cssparser = { git = "https://github.com/servo/rust-cssparser", rev = "aaa966d9d6ae70c4b8a62bb5e3a14c068bb7dff0" }
 cssparser-macros = { git = "https://github.com/servo/rust-cssparser", rev = "aaa966d9d6ae70c4b8a62bb5e3a14c068bb7dff0" }
-unicode-bidi = { git = "https://github.com/servo/unicode-bidi", rev = "d43c644c388659d093cc5d1648a970daeef692d5" }
+unicode-bidi = { git = "https://github.com/servo/unicode-bidi", rev = "ca612daf1c08c53abe07327cb3e6ef6e0a760f0c" }
 
 # Other overrides
 chardetng = { git = "https://github.com/hsivonen/chardetng", rev = "3484d3e3ebdc8931493aa5df4d7ee9360a90e76b" }

supply-chain/audits.toml

@@ -4024,7 +4024,7 @@ delta = "0.3.14 -> 0.3.15"
 [[audits.unicode-bidi]]
 who = "Jonathan Kew <[email protected]>"
 criteria = "safe-to-deploy"
-delta = "0.3.15 -> 0.3.15@git:d43c644c388659d093cc5d1648a970daeef692d5"
+delta = "0.3.15 -> 0.3.15@git:ca612daf1c08c53abe07327cb3e6ef6e0a760f0c"
 importable = false
 
 [[audits.unicode-ident]]

third_party/rust/unicode-bidi/.cargo-checksum.json

@@ -1 +1 @@
-{"files":{".appveyor.yml":"15bdeea0e836ac2ccbb259cde1509a0673a73300e90e970f3e533b189234b6fd",".github/workflows/main.yml":"e0bee93284a8b39c9d419038bfa72a6389ebdae39ce55c40624e764ac1c98a9e",".rustfmt.toml":"168c973274f3f5946e90cac6ae0f017d0832a5c830872d9d3b9b387ad6c1a81e","AUTHORS":"1ff3a7c8519b29544bb28ba9b1e7502df0cb764051fb9a1172e60006aa2b8dcc","COPYRIGHT":"edb20b474f6cbd4f4db066b54a9e0f687d0009d309412a63431189b59b8e2a07","Cargo.lock":"8842f03d0fcea88aa1546244d0455834732603175b293218f8e9a9f44c297b7c","Cargo.toml":"099454ebee9b081080e1521eccbe447db30b17ac36e9e655ed1d0d1e20e657fb","LICENSE-APACHE":"a60eea817514531668d7e00765731449fe14d059d3249e0bc93b36de45f759f2","LICENSE-MIT":"7b63ecd5f1902af1b63729947373683c32745c16a10e8e6292e2e2dcd7e90ae0","README.md":"81d3dc6d894a68894d982760b0a907f9dcbb0da179a8063ed9de1d2257518957","src/char_data/mod.rs":"8cbdcaacddb3dd9b70d615693fa73d0e7dca6332102a95f0d3ce447df7645284","src/char_data/tables.rs":"8adf126131f573a3b6d2c35849c1cc13c831c9b55c4d3fcb5a3961d8ed7a0d44","src/data_source.rs":"36fa0785e51c549c1f72f09040cfe515b848d1b23fb30d469770a6b4b17b49df","src/deprecated.rs":"370189b26e6ad604f01f4310805919ad6c4197af96301de4eea39f397cb02604","src/explicit.rs":"a6f87a31cbbb2810dba4c60866b9815b99d28acb1949aec91d7f7bf19ee80d59","src/format_chars.rs":"678399fec3f4bfaf4093f38cfdb8956288313386dc3511dab9fb58164e8dc01b","src/implicit.rs":"eedc82a7724bd4e4e2759ea6f480c33fa236253ad9059e884d19dcb2271d4fcb","src/level.rs":"367c56346f0a60e0b040b083c00b718d197f4f91f2d98460e0d10af49d3ecdc8","src/lib.rs":"686173269694f42ac21ef76f924a63cacb2aec2e50be345df12e23577a95468c","src/prepare.rs":"02a86c82a3823c917b47f8605d54f7266ca95c1c9f7c8e3c2efcbaf47abd8647","src/utf16.rs":"eaa01b3c93fae0d0399755e4ef328cb83f6df49c9f9021f37ed41fddf021a5e7"},"package":null}
+{"files":{".appveyor.yml":"15bdeea0e836ac2ccbb259cde1509a0673a73300e90e970f3e533b189234b6fd",".github/workflows/main.yml":"e0bee93284a8b39c9d419038bfa72a6389ebdae39ce55c40624e764ac1c98a9e",".rustfmt.toml":"168c973274f3f5946e90cac6ae0f017d0832a5c830872d9d3b9b387ad6c1a81e","AUTHORS":"1ff3a7c8519b29544bb28ba9b1e7502df0cb764051fb9a1172e60006aa2b8dcc","COPYRIGHT":"edb20b474f6cbd4f4db066b54a9e0f687d0009d309412a63431189b59b8e2a07","Cargo.lock":"8842f03d0fcea88aa1546244d0455834732603175b293218f8e9a9f44c297b7c","Cargo.toml":"099454ebee9b081080e1521eccbe447db30b17ac36e9e655ed1d0d1e20e657fb","LICENSE-APACHE":"a60eea817514531668d7e00765731449fe14d059d3249e0bc93b36de45f759f2","LICENSE-MIT":"7b63ecd5f1902af1b63729947373683c32745c16a10e8e6292e2e2dcd7e90ae0","README.md":"81d3dc6d894a68894d982760b0a907f9dcbb0da179a8063ed9de1d2257518957","src/char_data/mod.rs":"8cbdcaacddb3dd9b70d615693fa73d0e7dca6332102a95f0d3ce447df7645284","src/char_data/tables.rs":"8adf126131f573a3b6d2c35849c1cc13c831c9b55c4d3fcb5a3961d8ed7a0d44","src/data_source.rs":"36fa0785e51c549c1f72f09040cfe515b848d1b23fb30d469770a6b4b17b49df","src/deprecated.rs":"f94c0e75dec7e70cb9802e26b7f82fe618dcdd50e9973927bacd4eccc6899c62","src/explicit.rs":"86c3c55bf2cc90aab1411aac6cf05de505ca74e44a76fe829572dd7dc4dd2aa3","src/format_chars.rs":"678399fec3f4bfaf4093f38cfdb8956288313386dc3511dab9fb58164e8dc01b","src/implicit.rs":"8d5b003464aee3f333785c6170a884945251f39601e4ea658e669a2ad575d588","src/level.rs":"ce1eaa9940f1b90bc59aba296488b8cd128aefeb4b6b2e3ecc34da26c569150b","src/lib.rs":"9dff9c105f481a03823de6ad9a0a11733af019649ae211644061d5a525670244","src/prepare.rs":"aeb8b88cfb2d2e6b74473f5903205dd3683d57abcc8801de7b9fdea6a432a0fe","src/utf16.rs":"12ee177127a0b5b0350a1fcc1edf7387c26b51ec5654f724629aab723881c313"},"package":null}

third_party/rust/unicode-bidi/src/deprecated.rs

@@ -9,8 +9,6 @@
 
 //! 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.

third_party/rust/unicode-bidi/src/explicit.rs

@@ -11,28 +11,43 @@
 //!
 //! <http://www.unicode.org/reports/tr9/#Explicit_Levels_and_Directions>
 
+#[cfg(feature = "smallvec")]
+use smallvec::{smallvec, SmallVec};
+
 use super::char_data::{
     is_rtl,
     BidiClass::{self, *},
 };
 use super::level::Level;
+use super::prepare::removed_by_x9;
+use super::LevelRunVec;
 use super::TextSource;
 
-/// Compute explicit embedding levels for one paragraph of text (X1-X8).
+/// Compute explicit embedding levels for one paragraph of text (X1-X8), and identify
+/// level runs (BD7) for use when determining Isolating Run Sequences (X10).
 ///
 /// `processing_classes[i]` must contain the `BidiClass` of the char at byte index `i`,
 /// for each char in `text`.
+///
+/// `runs` returns the list of level runs (BD7) of the text.
 #[cfg_attr(feature = "flame_it", flamer::flame)]
 pub fn compute<'a, T: TextSource<'a> + ?Sized>(
     text: &'a T,
     para_level: Level,
     original_classes: &[BidiClass],
     levels: &mut [Level],
     processing_classes: &mut [BidiClass],
+    runs: &mut LevelRunVec,
 ) {
     assert_eq!(text.len(), original_classes.len());
 
     // <http://www.unicode.org/reports/tr9/#X1>
+    #[cfg(feature = "smallvec")]
+    let mut stack: SmallVec<[Status; 8]> = smallvec![Status {
+        level: para_level,
+        status: OverrideStatus::Neutral,
+    }];
+    #[cfg(not(feature = "smallvec"))]
     let mut stack = vec![Status {
         level: para_level,
         status: OverrideStatus::Neutral,
@@ -42,6 +57,9 @@ pub fn compute<'a, T: TextSource<'a> + ?Sized>(
     let mut overflow_embedding_count = 0u32;
     let mut valid_isolate_count = 0u32;
 
+    let mut current_run_level = Level::ltr();
+    let mut current_run_start = 0;
+
     for (i, len) in text.indices_lengths() {
         let last = stack.last().unwrap();
 
@@ -173,6 +191,26 @@ pub fn compute<'a, T: TextSource<'a> + ?Sized>(
             levels[i + j] = levels[i];
             processing_classes[i + j] = processing_classes[i];
         }
+
+        // Identify level runs to be passed to prepare::isolating_run_sequences().
+        if i == 0 {
+            // Initialize for the first (or only) run.
+            current_run_level = levels[i];
+        } else {
+            // Check if we need to start a new level run.
+            // <https://www.unicode.org/reports/tr9/#BD7>
+            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;
+            }
+        }
+    }
+
+    // Append the trailing level run, if non-empty.
+    if levels.len() > current_run_start {
+        runs.push(current_run_start..levels.len());
     }
 }
 

third_party/rust/unicode-bidi/src/implicit.rs

@@ -9,6 +9,7 @@
 
 //! 3.3.4 - 3.3.6. Resolve implicit levels and types.
 
+#[cfg(not(feature = "smallvec"))]
 use alloc::vec::Vec;
 use core::cmp::max;
 #[cfg(feature = "smallvec")]
@@ -250,6 +251,11 @@ pub fn resolve_weak<'a, T: TextSource<'a> + ?Sized>(
     }
 }
 
+#[cfg(feature = "smallvec")]
+type BracketPairVec = SmallVec<[BracketPair; 8]>;
+#[cfg(not(feature = "smallvec"))]
+type BracketPairVec = Vec<BracketPair>;
+
 /// 3.3.5 Resolving Neutral Types
 ///
 /// <http://www.unicode.org/reports/tr9/#Resolving_Neutral_Types>
@@ -273,7 +279,14 @@ pub fn resolve_neutral<'a, D: BidiDataSource, T: TextSource<'a> + ?Sized>(
 
     // > Identify the bracket pairs in the current isolating run sequence according to BD16.
     // We use processing_classes, not original_classes, due to BD14/BD15
-    let bracket_pairs = identify_bracket_pairs(text, data_source, sequence, processing_classes);
+    let mut bracket_pairs = BracketPairVec::new();
+    identify_bracket_pairs(
+        text,
+        data_source,
+        sequence,
+        processing_classes,
+        &mut bracket_pairs,
+    );
 
     // > For each bracket-pair element in the list of pairs of text positions
     //
@@ -493,8 +506,8 @@ fn identify_bracket_pairs<'a, T: TextSource<'a> + ?Sized, D: BidiDataSource>(
     data_source: &D,
     run_sequence: &IsolatingRunSequence,
     original_classes: &[BidiClass],
-) -> Vec<BracketPair> {
-    let mut ret = vec![];
+    bracket_pairs: &mut BracketPairVec,
+) {
     #[cfg(feature = "smallvec")]
     let mut stack = SmallVec::<[(char, usize, usize); 8]>::new();
     #[cfg(not(feature = "smallvec"))]
@@ -544,7 +557,7 @@ fn identify_bracket_pairs<'a, T: TextSource<'a> + ?Sized, D: BidiDataSource>(
                                 start_run: element.2,
                                 end_run: run_index,
                             };
-                            ret.push(pair);
+                            bracket_pairs.push(pair);
 
                             // > Pop the stack through the current stack element inclusively.
                             stack.truncate(stack_index);
@@ -557,8 +570,7 @@ fn identify_bracket_pairs<'a, T: TextSource<'a> + ?Sized, D: BidiDataSource>(
     }
     // > Sort the list of pairs of text positions in ascending order based on
     // > the text position of the opening paired bracket.
-    ret.sort_by_key(|r| r.start);
-    ret
+    bracket_pairs.sort_by_key(|r| r.start);
 }
 
 /// 3.3.6 Resolving Implicit Levels
@@ -567,11 +579,11 @@ fn identify_bracket_pairs<'a, T: TextSource<'a> + ?Sized, D: BidiDataSource>(
 ///
 /// <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 {
+pub fn resolve_levels(processing_classes: &[BidiClass], levels: &mut [Level]) -> Level {
     let mut max_level = Level::ltr();
-    assert_eq!(original_classes.len(), levels.len());
+    assert_eq!(processing_classes.len(), levels.len());
     for i in 0..levels.len() {
-        match (levels[i].is_rtl(), original_classes[i]) {
+        match (levels[i].is_rtl(), processing_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")

third_party/rust/unicode-bidi/src/level.rs

@@ -17,10 +17,7 @@ use alloc::{
     string::{String, ToString},
     vec::Vec,
 };
-use core::{
-    convert::{From, Into},
-    slice,
-};
+use core::slice;
 
 use super::char_data::BidiClass;