Break text runs by unicode script

components/gfx/Cargo.toml

@@ -23,6 +23,7 @@ serde_macros = "0.5"
 smallvec = "0.1"
 string_cache = "0.1"
 time = "0.1.12"
+unicode-script = { version = "0.1", features = ["harfbuzz"] }
 
 [dependencies.plugins]
 path = "../plugins"

components/gfx/font.rs

@@ -18,6 +18,7 @@ use style::properties::style_structs::Font as FontStyle;
 use text::Shaper;
 use text::glyph::{GlyphId, GlyphStore};
 use text::shaping::ShaperMethods;
+use unicode_script::Script;
 use util::cache::HashCache;
 use util::geometry::Au;
 
@@ -117,6 +118,8 @@ pub struct ShapingOptions {
     pub letter_spacing: Option<Au>,
     /// Spacing to add between each word. Corresponds to the CSS 2.1 `word-spacing` property.
     pub word_spacing: Au,
+    /// The Unicode script property of the characters in this run.
+    pub script: Script,
     /// Various flags.
     pub flags: ShapingFlags,
 }

components/gfx/lib.rs

@@ -72,6 +72,7 @@ extern crate smallvec;
 extern crate string_cache;
 extern crate style;
 extern crate time;
+extern crate unicode_script;
 extern crate url;
 
 

components/gfx/text/shaping/harfbuzz.rs

@@ -215,6 +215,8 @@ impl ShaperMethods for Shaper {
                 HB_DIRECTION_LTR
             });
 
+            hb_buffer_set_script(hb_buffer, options.script.to_hb_script());
+
             hb_buffer_add_utf8(hb_buffer,
                                text.as_ptr() as *const c_char,
                                text.len() as c_int,

components/layout/Cargo.toml

@@ -81,3 +81,4 @@ serde = "0.6"
 serde_macros = "0.5"
 serde_json = "0.5"
 unicode-bidi = "0.2"
+unicode-script = { version = "0.1", features = ["harfbuzz"] }

components/layout/lib.rs

@@ -53,6 +53,7 @@ extern crate smallvec;
 extern crate string_cache;
 extern crate style;
 extern crate unicode_bidi;
+extern crate unicode_script;
 extern crate url;
 
 #[macro_use]

components/layout/text.rs

@@ -23,6 +23,7 @@ use style::computed_values::{white_space};
 use style::properties::ComputedValues;
 use style::properties::style_structs::Font as FontStyle;
 use unicode_bidi::{is_rtl, process_text};
+use unicode_script::{get_script, Script};
 use util::geometry::Au;
 use util::linked_list::split_off_head;
 use util::logical_geometry::{LogicalSize, WritingMode};
@@ -204,8 +205,22 @@ impl TextRunScanner {
                         None => 0
                     };
 
+                    // Break the run if the new character has a different explicit script than the
+                    // previous characters.
+                    //
+                    // TODO: Special handling of paired punctuation characters.
+                    // http://www.unicode.org/reports/tr24/#Common
+                    let script = get_script(character);
+                    let compatible_script = is_compatible(script, run_info.script);
+                    if compatible_script && !is_specific(run_info.script) && is_specific(script) {
+                        run_info.script = script;
+                    }
+
                     // Now, if necessary, flush the mapping we were building up.
-                    if run_info.font_index != font_index || run_info.bidi_level != bidi_level {
+                    if run_info.font_index != font_index ||
+                       run_info.bidi_level != bidi_level ||
+                       !compatible_script
+                    {
                         if end_position > start_position {
                             mapping.flush(&mut mappings,
                                           &mut run_info,
@@ -226,6 +241,7 @@ impl TextRunScanner {
                         }
                         run_info.font_index = font_index;
                         run_info.bidi_level = bidi_level;
+                        run_info.script = script;
                     }
 
                     // Consume this character.
@@ -269,12 +285,14 @@ impl TextRunScanner {
             let options = ShapingOptions {
                 letter_spacing: letter_spacing,
                 word_spacing: word_spacing,
+                script: Script::Common,
                 flags: flags,
             };
 
             // FIXME(https://github.com/rust-lang/rust/issues/23338)
             run_info_list.into_iter().map(|run_info| {
                 let mut options = options;
+                options.script = run_info.script;
                 if is_rtl(run_info.bidi_level) {
                     options.flags.insert(RTL_FLAG);
                 }
@@ -440,6 +458,8 @@ struct RunInfo {
     character_length: usize,
     /// The bidirection embedding level of this text run.
     bidi_level: u8,
+    /// The Unicode script property of this text run.
+    script: Script,
 }
 
 impl RunInfo {
@@ -450,6 +470,7 @@ impl RunInfo {
             font_index: 0,
             character_length: 0,
             bidi_level: 0,
+            script: Script::Common,
         }
     }
 }
@@ -599,3 +620,13 @@ struct ScannedTextRun {
     run: Arc<TextRun>,
     insertion_point: Option<CharIndex>,
 }
+
+/// Can a character with script `b` continue a text run with script `a`?
+fn is_compatible(a: Script, b: Script) -> bool {
+    a == b || !is_specific(a) || !is_specific(b)
+}
+
+/// Returns true if the script is not invalid or inherited.
+fn is_specific(script: Script) -> bool {
+    script != Script::Common && script != Script::Inherited
+}

components/servo/Cargo.lock

@@ -612,6 +612,7 @@ dependencies = [
  "string_cache 0.1.13 (registry+https://github.com/rust-lang/crates.io-index)",
  "style 0.0.1",
  "time 0.1.32 (registry+https://github.com/rust-lang/crates.io-index)",
+ "unicode-script 0.1.1 (registry+https://github.com/rust-lang/crates.io-index)",
  "url 0.2.37 (registry+https://github.com/rust-lang/crates.io-index)",
  "util 0.0.1",
 ]
@@ -960,6 +961,7 @@ dependencies = [
  "string_cache_plugin 0.1.7 (registry+https://github.com/rust-lang/crates.io-index)",
  "style 0.0.1",
  "unicode-bidi 0.2.2 (registry+https://github.com/rust-lang/crates.io-index)",
+ "unicode-script 0.1.1 (registry+https://github.com/rust-lang/crates.io-index)",
  "url 0.2.37 (registry+https://github.com/rust-lang/crates.io-index)",
  "util 0.0.1",
 ]
@@ -1778,6 +1780,14 @@ dependencies = [
  "matches 0.1.2 (registry+https://github.com/rust-lang/crates.io-index)",
 ]
 
+[[package]]
+name = "unicode-script"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+dependencies = [
+ "harfbuzz-sys 0.1.0 (registry+https://github.com/rust-lang/crates.io-index)",
+]
+
 [[package]]
 name = "unreachable"
 version = "0.0.2"