slim parley_data

Author: robertbastian

parley/src/analysis/cluster.rs

@@ -3,8 +3,8 @@
 
 use alloc::vec::Vec;
 
-use crate::analysis::CompositeProps;
 use icu_normalizer::properties::{CanonicalComposition, CanonicalDecomposition, Decomposed};
+use parley_data::Properties;
 
 /// The maximum number of characters in a single cluster.
 const MAX_CLUSTER_SIZE: usize = 32;
@@ -117,7 +117,7 @@ impl CharCluster {
 
     #[inline(always)]
     fn contributes_to_shaping(ch: char) -> bool {
-        let props = CompositeProps.properties(ch as u32);
+        let props = Properties::get(ch);
         crate::analysis::contributes_to_shaping(props.general_category(), props.script())
     }
 

parley/src/analysis/mod.rs

@@ -5,6 +5,7 @@ pub(crate) mod cluster;
 
 use alloc::vec::Vec;
 use core::marker::PhantomData;
+use parley_data::Properties;
 
 use crate::resolve::RangedStyle;
 use crate::{Brush, LayoutContext, WordBreak};
@@ -13,7 +14,6 @@ use icu_properties::CodePointMapData;
 use icu_properties::props::{BidiMirroringGlyph, GeneralCategory, GraphemeClusterBreak, Script};
 use icu_segmenter::options::{LineBreakOptions, LineBreakWordOption, WordBreakInvariantOptions};
 use icu_segmenter::{LineSegmenter, WordSegmenter};
-use parley_data::CompositeProps;
 
 #[derive(Copy, Clone, Debug, PartialEq, Eq)]
 pub(crate) struct CharInfo {
@@ -361,7 +361,7 @@ pub(crate) fn analyze_text<B: Brush>(lcx: &mut LayoutContext<B>, text: &str) {
         // character's index, but we need our iterators to align, and the rest are simply
         // character-indexed.
         .fold(false, |is_mandatory_linebreak, (boundary, ch)| {
-            let properties = CompositeProps.properties(ch as u32);
+            let properties = Properties::get(ch);
             let script = properties.script();
             let grapheme_cluster_break = properties.grapheme_cluster_break();
             let bidi_class = properties.bidi_class();

parley_data/Cargo.toml

@@ -10,12 +10,11 @@ publish = false
 
 [features]
 default = ["baked"]
-baked = ["dep:zerovec"]
-datagen = []
+baked = ["dep:icu_collections", "dep:zerovec"]
 
 [dependencies]
 icu_properties = { workspace = true }
-icu_collections = { workspace = true }
+icu_collections = { workspace = true, optional = true }
 zerovec = { workspace = true, optional = true }
 
 [lints]

parley_data/README.md

@@ -9,7 +9,6 @@
 ## Cargo features
 
 - `baked` *(default)* embeds pre-generated ICU4X and composite data from `src/generated`, enabling use in `no_std` targets without a filesystem.
-- `datagen` enables code so the crate can participate in regeneration workflows. This feature is intended for developer use when refreshing the baked data and principally used by `../parley_data_gen`.
 
 ## Regenerating the baked data
 

parley_data/src/generated/composite/mod.rs

@@ -12,19 +12,6 @@ use icu_properties::props::{BidiClass, GeneralCategory, GraphemeClusterBreak, Sc
 #[cfg(feature = "baked")]
 pub mod generated;
 
-/// Lookup for [`Properties`]
-#[derive(Clone, Debug, Copy)]
-pub struct CompositeProps;
-
-#[cfg(feature = "baked")]
-impl CompositeProps {
-    /// Returns the properties for a given character.
-    #[inline(always)]
-    pub fn properties(&self, ch: u32) -> Properties {
-        Properties(generated::COMPOSITE.get32(ch))
-    }
-}
-
 /// Unicode character properties relevant for text analysis.
 #[derive(Copy, Clone, Debug)]
 pub struct Properties(u32);
@@ -51,8 +38,13 @@ impl Properties {
     const IS_MANDATORY_LINE_BREAK_SHIFT: u32 =
         Self::IS_REGION_INDICATOR_SHIFT + Self::IS_REGION_INDICATOR_BITS;
 
-    /// Packs the given arguments into a single u32.
-    #[cfg(feature = "datagen")]
+    #[cfg(feature = "baked")]
+    /// Returns the properties for a given character.
+    pub fn get(ch: char) -> Self {
+        Self(generated::COMPOSITE.get(ch))
+    }
+
+    /// Creates a new [`Properties`] from the given properties
     pub fn new(
         script: Script,
         gc: GeneralCategory,
@@ -81,7 +73,7 @@ impl Properties {
     }
 
     #[inline(always)]
-    fn get(&self, shift: u32, bits: u32) -> u32 {
+    fn bits(&self, shift: u32, bits: u32) -> u32 {
         (self.0 >> shift) & ((1 << bits) - 1)
     }
 
@@ -92,7 +84,7 @@ impl Properties {
             clippy::cast_possible_truncation,
             reason = "script data only occupies SCRIPT_BITS bits; we cast to `u16` to fulfil the `from_icu4c_value` contract."
         )]
-        Script::from_icu4c_value(self.get(Self::SCRIPT_SHIFT, Self::SCRIPT_BITS) as u16)
+        Script::from_icu4c_value(self.bits(Self::SCRIPT_SHIFT, Self::SCRIPT_BITS) as u16)
     }
 
     /// Returns the general category for the character.
@@ -102,7 +94,7 @@ impl Properties {
             clippy::cast_possible_truncation,
             reason = "general category data only occupies GC_BITS bits."
         )]
-        GeneralCategory::try_from(self.get(Self::GC_SHIFT, Self::GC_BITS) as u8).unwrap()
+        GeneralCategory::try_from(self.bits(Self::GC_SHIFT, Self::GC_BITS) as u8).unwrap()
     }
 
     /// Returns the grapheme cluster break for the character.
@@ -112,7 +104,7 @@ impl Properties {
             clippy::cast_possible_truncation,
             reason = "cluster break data only occupies GCB_BITS bits"
         )]
-        GraphemeClusterBreak::from_icu4c_value(self.get(Self::GCB_SHIFT, Self::GCB_BITS) as u8)
+        GraphemeClusterBreak::from_icu4c_value(self.bits(Self::GCB_SHIFT, Self::GCB_BITS) as u8)
     }
 
     /// Returns the bidirectional class for the character.
@@ -122,13 +114,13 @@ impl Properties {
             clippy::cast_possible_truncation,
             reason = "bidi class data only occupies BIDI_BITS bits"
         )]
-        BidiClass::from_icu4c_value(self.get(Self::BIDI_SHIFT, Self::BIDI_BITS) as u8)
+        BidiClass::from_icu4c_value(self.bits(Self::BIDI_SHIFT, Self::BIDI_BITS) as u8)
     }
 
     /// Returns whether the character is an emoji or pictograph.
     #[inline(always)]
     pub fn is_emoji_or_pictograph(&self) -> bool {
-        self.get(
+        self.bits(
             Self::IS_EMOJI_OR_PICTOGRAPH_SHIFT,
             Self::IS_EMOJI_OR_PICTOGRAPH_BITS,
         ) != 0
@@ -137,7 +129,7 @@ impl Properties {
     /// Returns whether the character is a variation selector.
     #[inline(always)]
     pub fn is_variation_selector(&self) -> bool {
-        self.get(
+        self.bits(
             Self::IS_VARIATION_SELECTOR_SHIFT,
             Self::IS_VARIATION_SELECTOR_BITS,
         ) != 0
@@ -146,7 +138,7 @@ impl Properties {
     /// Returns whether the character is a region indicator.
     #[inline(always)]
     pub fn is_region_indicator(&self) -> bool {
-        self.get(
+        self.bits(
             Self::IS_REGION_INDICATOR_SHIFT,
             Self::IS_REGION_INDICATOR_BITS,
         ) != 0
@@ -155,7 +147,7 @@ impl Properties {
     /// Returns whether the character is a mandatory linebreak.
     #[inline(always)]
     pub fn is_mandatory_linebreak(&self) -> bool {
-        self.get(
+        self.bits(
             Self::IS_MANDATORY_LINE_BREAK_SHIFT,
             Self::IS_MANDATORY_LINE_BREAK_BITS,
         ) != 0

parley_data/src/generated/mod.rs

@@ -9,11 +9,11 @@ repository.workspace = true
 publish = false
 
 [dependencies]
-databake = { workspace = true, features = ["derive"] }
+databake = { workspace = true }
 icu_codepointtrie_builder = { workspace = true, features = ["wasm"] }
 icu_collections = { workspace = true, features = ["databake"] }
 icu_properties = { workspace = true, features = ["compiled_data"] }
-parley_data = { workspace = true, default-features = false, features = ["datagen"] }
+parley_data = { workspace = true }
 
 [lints]
 workspace = true

parley_data/src/lib.rs

@@ -54,11 +54,7 @@ pub fn generate(out: std::path::PathBuf) {
         }
         .build();
 
-        let composite_dir = out.join("composite");
-        if !composite_dir.exists() {
-            std::fs::create_dir(&composite_dir).unwrap();
-        }
-        let mut file = BufWriter::new(std::fs::File::create(composite_dir.join("mod.rs")).unwrap());
+        let mut file = BufWriter::new(std::fs::File::create(out.join("mod.rs")).unwrap());
 
         writeln!(&mut file, "{COPYRIGHT_HEADER}").unwrap();
         writeln!(&mut file, "/// Backing data for the `CompositeProps`").unwrap();
@@ -77,10 +73,4 @@ pub fn generate(out: std::path::PathBuf) {
         )
         .unwrap();
     }
-
-    let mut file = BufWriter::new(std::fs::File::create(out.join("mod.rs")).unwrap());
-
-    writeln!(&mut file, "{COPYRIGHT_HEADER}").unwrap();
-    writeln!(&mut file, "mod composite;").unwrap();
-    writeln!(&mut file, "pub use composite::*;").unwrap();
 }