Use IndexSet instead of Vec for codepoints

This ensures that codepoints are unique, but also respects the original
order in which they appear in the file.

Author: cmyr

Cargo.toml

@@ -27,6 +27,7 @@ quick-xml = { version = "0.26.0", features = ["serialize"] }
 rayon = { version = "1.3.0", optional = true }
 kurbo = { version = "0.9.0", optional = true }
 thiserror = "1.0"
+indexmap = {version = "1.9.2", features = ["serde"] }
 
 [dependencies.druid]
 default-features = false

src/glyph/mod.rs

@@ -13,6 +13,7 @@ use crate::error::ConvertContourError;
 
 #[cfg(feature = "druid")]
 use druid::{Data, Lens};
+use indexmap::IndexSet;
 
 use crate::error::{ErrorKind, GlifLoadError, GlifWriteError, StoreError};
 use crate::name::Name;
@@ -38,7 +39,7 @@ pub struct Glyph {
     /// A collection of glyph Unicode code points.
     ///
     /// The first entry defines the primary Unicode value for this glyph.
-    pub codepoints: Vec<char>,
+    pub codepoints: IndexSet<char>,
     /// Arbitrary glyph note.
     pub note: Option<String>,
     /// A collection of glyph guidelines.
@@ -115,7 +116,7 @@ impl Glyph {
             name,
             height: 0.0,
             width: 0.0,
-            codepoints: Vec::new(),
+            codepoints: Default::default(),
             note: None,
             guidelines: Vec::new(),
             anchors: Vec::new(),

src/glyph/parse.rs

@@ -409,7 +409,7 @@ impl<'names> GlifParser<'names> {
                         .map_err(|_| value.to_string())
                         .and_then(|n| char::try_from(n).map_err(|_| value.to_string()))
                         .map_err(|_| ErrorKind::BadHexValue)?;
-                    self.glyph.codepoints.push(chr);
+                    self.glyph.codepoints.insert(chr);
                 }
                 _other => return Err(ErrorKind::UnexpectedAttribute.into()),
             }

src/glyph/tests.rs

@@ -1,3 +1,5 @@
+use indexmap::indexset;
+
 use super::parse::parse_glyph;
 use super::*;
 use crate::write::QuoteChar;
@@ -868,3 +870,30 @@ fn has_component_with_base() {
     assert!(glyph.has_component_with_base("dieresis"));
     assert!(!glyph.has_component_with_base("Z"));
 }
+
+#[test]
+fn deduplicate_unicodes2() {
+    let data = r#"
+<?xml version="1.0" encoding="UTF-8"?>
+<glyph name="period" format="2">
+  <unicode hex="0065"/>
+  <unicode hex="0066"/>
+  <unicode hex="0065"/>
+  <unicode hex="0067"/>
+</glyph>
+"#;
+    let mut glyph = parse_glyph(data.as_bytes()).unwrap();
+    assert_eq!(glyph.codepoints, indexset!['e', 'f', 'g'].into());
+
+    glyph.codepoints = indexset!['e', 'f', 'e', 'g'].into();
+    let data2 = glyph.encode_xml().unwrap();
+    let data2 = std::str::from_utf8(&data2).unwrap();
+    let data2_expected = r#"<?xml version="1.0" encoding="UTF-8"?>
+<glyph name="period" format="2">
+	<unicode hex="0065"/>
+	<unicode hex="0066"/>
+	<unicode hex="0067"/>
+</glyph>
+"#;
+    assert_eq!(data2, data2_expected);
+}

src/layer.rs

@@ -594,6 +594,8 @@ impl Default for Layer {
 
 #[cfg(test)]
 mod tests {
+    use indexmap::indexset;
+
     use crate::DataRequest;
 
     use super::*;
@@ -613,7 +615,7 @@ mod tests {
         let glyph = layer.get_glyph("A").expect("failed to load glyph 'A'");
         assert_eq!(glyph.height, 0.);
         assert_eq!(glyph.width, 1190.);
-        assert_eq!(glyph.codepoints, vec!['A']);
+        assert_eq!(glyph.codepoints, indexset!['A']);
     }
 
     #[test]

tests/save.rs

@@ -1,5 +1,6 @@
 //! Testing saving files.
 
+use indexmap::indexset;
 use norad::{Font, FormatVersion, Glyph, Identifier, Plist};
 use plist::Value;
 
@@ -24,7 +25,7 @@ fn save_default() {
 fn save_new_file() {
     let mut my_ufo = Font::new();
     let mut my_glyph = Glyph::new("A");
-    my_glyph.codepoints = vec!['A'];
+    my_glyph.codepoints = indexset!['A'];
     my_glyph.note = Some("I did a glyph!".into());
     let mut plist = Plist::new();
     plist.insert("my-cool-key".into(), plist::Value::Integer(420_u32.into()));
@@ -42,7 +43,7 @@ fn save_new_file() {
     let loaded = Font::load(dir).unwrap();
     assert!(loaded.default_layer().get_glyph("A").is_some());
     let glyph = loaded.default_layer().get_glyph("A").unwrap();
-    assert_eq!(glyph.codepoints, vec!['A']);
+    assert_eq!(glyph.codepoints, indexset!['A']);
     let lib_val = glyph.lib.get("my-cool-key").and_then(|val| val.as_unsigned_integer());
     assert_eq!(lib_val, Some(420));
 }