Normalization and case folding

Both of which deserve some further attention.

1 files changed, 13 insertions, 29 deletions

diff --git a/src/HangulData.zig b/src/HangulData.zig
index cae8b97..9c17421 100644
--- a/src/HangulData.zig
+++ b/src/HangulData.zig
@@ -9,39 +9,23 @@ pub const Syllable = enum {
     T,
 };
 
-s1: []u16 = undefined,
-s2: []u3 = undefined,
-
-const Hangul = @This();
-
-pub fn init(allocator: mem.Allocator) !Hangul {
-    const in_bytes = @embedFile("hangul");
-    var in_fbs = std.io.fixedBufferStream(in_bytes);
-    var reader = in_fbs.reader();
-
-    const endian = builtin.cpu.arch.endian();
-    var hangul = Hangul{};
-
-    const stage_1_len: u16 = try reader.readInt(u16, endian);
-    hangul.s1 = try allocator.alloc(u16, stage_1_len);
-    errdefer allocator.free(hangul.s1);
-    for (0..stage_1_len) |i| hangul.s1[i] = try reader.readInt(u16, endian);
-
-    const stage_2_len: u16 = try reader.readInt(u16, endian);
-    hangul.s2 = try allocator.alloc(u3, stage_2_len);
-    errdefer allocator.free(hangul.s2);
-    for (0..stage_2_len) |i| hangul.s2[i] = @intCast(try reader.readInt(u8, endian));
+const Data = struct {
+    s1: []const u16 = undefined,
+    s2: []const u3 = undefined,
+};
 
-    return hangul;
-}
+const hangul = hangul_data: {
+    const data = @import("hangul");
+    break :hangul_data Data{
+        .s1 = &data.s1,
+        .s2 = &data.s2,
+    };
+};
 
-pub fn deinit(hangul: *const Hangul, allocator: mem.Allocator) void {
-    allocator.free(hangul.s1);
-    allocator.free(hangul.s2);
-}
+const Hangul = @This();
 
 /// Returns the Hangul syllable type for `cp`.
-pub fn syllable(hangul: *const Hangul, cp: u21) Syllable {
+pub fn syllable(cp: u21) Syllable {
     return @enumFromInt(hangul.s2[hangul.s1[cp >> 8] + (cp & 0xff)]);
 }