Switch to two-face for extra syntect syntaxes (#2937)

Cargo.lock

@@ -2645,6 +2645,17 @@ version = "0.19.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "49d64318d8311fc2668e48b63969f4343e0a85c4a109aa8460d6672e364b8bd1"
 
+[[package]]
+name = "two-face"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "37bed2135b2459c7eefba72c906d374697eb15949c205f2f124e3636a46b5eeb"
+dependencies = [
+ "once_cell",
+ "serde",
+ "syntect",
+]
+
 [[package]]
 name = "typed-arena"
 version = "2.0.2"
@@ -2692,6 +2703,7 @@ dependencies = [
  "toml",
  "tracing",
  "ttf-parser",
+ "two-face",
  "typed-arena",
  "typst-macros",
  "typst-syntax",

Cargo.toml

@@ -104,6 +104,7 @@ tracing-error = "0.2"
 tracing-flame = "0.2.0"
 tracing-subscriber = "0.3.17"
 ttf-parser = "0.19.2"
+two-face = { version = "0.3.0", default-features = false, features = ["syntect-fancy"] }
 typed-arena = "2"
 unicode_names2 = "1.2"
 unicode-bidi = "0.3.13"

NOTICE

@@ -215,7 +215,7 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 ================================================================================
 
 ================================================================================
-Syntaxes in `crates/typst-library/assets/syntect.bin` are generated from syntax
+Syntaxes embedded within the `two-face` crate are generated from syntax
 definitions which can be found here:
 https://github.com/sharkdp/bat/tree/master/assets/syntaxes
 

crates/typst/Cargo.toml

@@ -54,6 +54,7 @@ time = { workspace = true }
 toml = { workspace = true }
 tracing = { workspace = true }
 ttf-parser = { workspace = true }
+two-face = { workspace = true }
 typed-arena = { workspace = true }
 unicode-bidi = { workspace = true }
 unicode-math-class = { workspace = true }

crates/typst/src/text/raw.rs

@@ -734,31 +734,10 @@ fn parse_theme(
 
 /// The syntect syntax definitions.
 ///
-/// Code for syntax set generation is below. The `syntaxes` directory is from
+/// Syntax set is generated from the syntaxes from the `bat` project
 /// <https://github.com/sharkdp/bat/tree/master/assets/syntaxes>
-///
-/// ```ignore
-/// fn main() {
-///     let mut builder = syntect::parsing::SyntaxSet::load_defaults_nonewlines().into_builder();
-///     builder.add_from_folder("syntaxes/02_Extra", false).unwrap();
-///     syntect::dumps::dump_to_file(&builder.build(), "syntect.bin").unwrap();
-/// }
-/// ```
-///
-/// The following syntaxes are disabled due to compatibility issues:
-/// ```text
-/// syntaxes/02_Extra/Assembly (ARM).sublime-syntax
-/// syntaxes/02_Extra/Elixir/Regular Expressions (Elixir).sublime-syntax
-/// syntaxes/02_Extra/JavaScript (Babel).sublime-syntax
-/// syntaxes/02_Extra/LiveScript.sublime-syntax
-/// syntaxes/02_Extra/PowerShell.sublime-syntax
-/// syntaxes/02_Extra/SCSS_Sass/Syntaxes/Sass.sublime-syntax
-/// syntaxes/02_Extra/SLS/SLS.sublime-syntax
-/// syntaxes/02_Extra/VimHelp.sublime-syntax
-/// syntaxes/02_Extra/cmd-help/syntaxes/cmd-help.sublime-syntax
-/// ```
 pub static RAW_SYNTAXES: Lazy<syntect::parsing::SyntaxSet> =
-    Lazy::new(|| syntect::dumps::from_binary(include_bytes!("../../assets/syntect.bin")));
+    Lazy::new(two_face::syntax::extra_no_newlines);
 
 /// The default theme used for syntax highlighting.
 pub static RAW_THEME: Lazy<synt::Theme> = Lazy::new(|| synt::Theme {