Add --combined option

src/app.rs

@@ -199,6 +199,12 @@ pub fn app() -> App<'static, 'static> {
     let ucd_dir = Arg::with_name("ucd-dir")
         .required(true)
         .help("Directory containing the Unicode character database files.");
+    let flag_combined = Arg::with_name("combined").long("combined").help(
+        "Emit a single table with all included codepoint ranges. You might \
+        want to use this option when checking if characters belong to a \
+        subset of categories, since only one table will need to be checked. \
+        Searching the combined table should be simpler and more efficient.",
+    );
 
     // Subcommands.
     let cmd_bidi_class = SubCommand::with_name("bidi-class")
@@ -213,6 +219,7 @@ pub fn app() -> App<'static, 'static> {
         .arg(flag_chars.clone())
         .arg(flag_trie_set.clone())
         .arg(flag_short_names.clone())
+        .arg(flag_combined.clone())
         .arg(
             Arg::with_name("enum").long("enum").help(
                 "Emit a single table that maps codepoints to bidi class.",
@@ -252,6 +259,7 @@ pub fn app() -> App<'static, 'static> {
         .arg(flag_name("GENERAL_CATEGORY"))
         .arg(flag_chars.clone())
         .arg(flag_trie_set.clone())
+        .arg(flag_combined.clone())
         .arg(
             Arg::with_name("enum").long("enum").help(
                 "Emit a single table that maps codepoints to categories.",
@@ -285,6 +293,7 @@ pub fn app() -> App<'static, 'static> {
         .arg(flag_name("SCRIPT"))
         .arg(flag_chars.clone())
         .arg(flag_trie_set.clone())
+        .arg(flag_combined.clone())
         .arg(
             Arg::with_name("enum")
                 .long("enum")
@@ -362,6 +371,7 @@ pub fn app() -> App<'static, 'static> {
             .arg(flag_name("JOINING_TYPE"))
             .arg(flag_chars.clone())
             .arg(flag_trie_set.clone())
+            .arg(flag_combined.clone())
             .arg(Arg::with_name("enum").long("enum").help(
                 "Emit a single table that maps codepoints to joining type.",
             ))

src/bidi_class.rs

@@ -140,6 +140,8 @@ pub fn command(args: ArgMatches<'_>) -> Result<()> {
     } else if args.is_present("rust-enum") {
         let variants = by_type.keys().map(String::as_str).collect::<Vec<_>>();
         wtr.ranges_to_rust_enum(args.name(), &variants, &by_type)?;
+    } else if args.is_present("combined") {
+        wtr.ranges_to_combined(args.name(), &by_type)?;
     } else {
         wtr.names(by_type.keys())?;
         for (name, set) in by_type {

src/general_category.rs

@@ -43,6 +43,8 @@ pub fn command(args: ArgMatches<'_>) -> Result<()> {
     } else if args.is_present("rust-enum") {
         let variants = bycat.keys().map(String::as_str).collect::<Vec<_>>();
         wtr.ranges_to_rust_enum(args.name(), &variants, &bycat)?;
+    } else if args.is_present("combined") {
+        wtr.ranges_to_combined(args.name(), &bycat)?;
     } else {
         wtr.names(bycat.keys().filter(|n| filter.contains(n)))?;
         for (name, set) in bycat {

src/joining_type.rs

@@ -62,6 +62,8 @@ pub fn command(args: ArgMatches<'_>) -> Result<()> {
     } else if args.is_present("rust-enum") {
         let variants = by_type.keys().map(String::as_str).collect::<Vec<_>>();
         wtr.ranges_to_rust_enum(args.name(), &variants, &by_type)?;
+    } else if args.is_present("combined") {
+        wtr.ranges_to_combined(args.name(), &by_type)?;
     } else {
         wtr.names(by_type.keys())?;
         for (name, set) in by_type {

src/script.rs

@@ -31,6 +31,8 @@ pub fn command_script(args: ArgMatches<'_>) -> Result<()> {
         let mut variants = vec!["Unknown"];
         variants.extend(by_name.keys().map(String::as_str));
         wtr.ranges_to_rust_enum(args.name(), &variants, &by_name)?;
+    } else if args.is_present("combined") {
+        wtr.ranges_to_combined(args.name(), &by_name)?;
     } else {
         wtr.names(by_name.keys().filter(|n| filter.contains(n)))?;
         for (name, set) in by_name {

src/writer.rs

@@ -330,6 +330,24 @@ impl Writer {
         Ok(())
     }
 
+    /// Write a map that combines codepoint ranges into a single table.
+    ///
+    /// The given map should be a map from the variant value to the set of
+    /// codepoints that have that value.
+    pub fn ranges_to_combined(
+        &mut self,
+        name: &str,
+        enum_map: &BTreeMap<String, BTreeSet<u32>>,
+    ) -> Result<()> {
+        let mut set = BTreeSet::new();
+        for other_set in enum_map.values() {
+            set.extend(other_set.iter().cloned());
+        }
+        self.ranges(name, &set)?;
+        self.wtr.flush()?;
+        Ok(())
+    }
+
     fn ranges_to_enum_slice<S>(
         &mut self,
         name: &str,