diff options
| author | 2026-02-05 07:30:22 -0500 | |
|---|---|---|
| committer | 2026-02-05 07:30:22 -0500 | |
| commit | 80d679d84701a63c0fd3eccd7d3a71a32fab8943 (patch) | |
| tree | 0f50ddce3770bd4d95dd4b045d023f84c4cd5fa8 | |
| parent | Delete upper and lower.zig (diff) | |
| parent | Merge branch 'master' into emoji (diff) | |
| download | zg-80d679d84701a63c0fd3eccd7d3a71a32fab8943.tar.gz zg-80d679d84701a63c0fd3eccd7d3a71a32fab8943.tar.xz zg-80d679d84701a63c0fd3eccd7d3a71a32fab8943.zip | |
Merge remote-tracking branch 'jacob/emoji' into no-allocation
Thanks again, Jacob!
| -rw-r--r-- | README.md | 28 | ||||
| -rw-r--r-- | build.zig | 27 | ||||
| -rw-r--r-- | codegen/emoji.zig | 146 | ||||
| -rw-r--r-- | src/Emoji.zig | 132 |
4 files changed, 333 insertions, 0 deletions
| @@ -610,6 +610,34 @@ test "Scripts" { | |||
| 610 | } | 610 | } |
| 611 | ``` | 611 | ``` |
| 612 | 612 | ||
| 613 | ## Emoji | ||
| 614 | |||
| 615 | To get information about emoji and emoji-like characters, use the `Emoji` module. | ||
| 616 | |||
| 617 | In your `build.zig`: | ||
| 618 | |||
| 619 | ```zig | ||
| 620 | exe.root_module.addImport("Emoji", zg.module("Emoji")); | ||
| 621 | ``` | ||
| 622 | |||
| 623 | In your code: | ||
| 624 | |||
| 625 | ```zig | ||
| 626 | const Emoji = @import("Emoji"); | ||
| 627 | |||
| 628 | test "Emoji" { | ||
| 629 | const emoji = try Emoji.init(allocator); | ||
| 630 | defer emoji.deinit(allocator); | ||
| 631 | |||
| 632 | try expect(emoji.isEmoji(0x1F415)); // 🐕 | ||
| 633 | try expect(emoji.isEmojiPresentation(0x1F408)); // 🐈 | ||
| 634 | try expect(emoji.isEmojiModifier(0x1F3FF)); // 🏿 | ||
| 635 | try expect(emoji.isEmojiModifierBase(0x1F977)); // 🥷 | ||
| 636 | try expect(emoji.isEmojiComponent(0x1F9B0)); // 🦰 | ||
| 637 | try expect(emoji.isExtendedPictographic(0x1F005)); // 🀅 | ||
| 638 | } | ||
| 639 | ``` | ||
| 640 | |||
| 613 | ## Limits | 641 | ## Limits |
| 614 | 642 | ||
| 615 | Iterators, and fragment types such as `CodePoint`, `Grapheme` and | 643 | Iterators, and fragment types such as `CodePoint`, `Grapheme` and |
| @@ -54,6 +54,16 @@ pub fn build(b: *std.Build) void { | |||
| 54 | const run_gbp_gen_exe = b.addRunArtifact(gbp_gen_exe); | 54 | const run_gbp_gen_exe = b.addRunArtifact(gbp_gen_exe); |
| 55 | const gbp_gen_out = run_gbp_gen_exe.addOutputFileArg("gbp.zig"); | 55 | const gbp_gen_out = run_gbp_gen_exe.addOutputFileArg("gbp.zig"); |
| 56 | 56 | ||
| 57 | // Emoji | ||
| 58 | const emoji_gen_exe = b.addExecutable(.{ | ||
| 59 | .name = "emoji", | ||
| 60 | .root_source_file = b.path("codegen/emoji.zig"), | ||
| 61 | .target = b.graph.host, | ||
| 62 | .optimize = .Debug, | ||
| 63 | }); | ||
| 64 | const run_emoji_gen_exe = b.addRunArtifact(emoji_gen_exe); | ||
| 65 | const emoji_gen_out = run_emoji_gen_exe.addOutputFileArg("emoji.bin.z"); | ||
| 66 | |||
| 57 | const wbp_gen_exe = b.addExecutable(.{ | 67 | const wbp_gen_exe = b.addExecutable(.{ |
| 58 | .name = "wbp", | 68 | .name = "wbp", |
| 59 | .root_module = b.createModule(.{ | 69 | .root_module = b.createModule(.{ |
| @@ -262,6 +272,22 @@ pub fn build(b: *std.Build) void { | |||
| 262 | }); | 272 | }); |
| 263 | const grapheme_tr = b.addRunArtifact(grapheme_t); | 273 | const grapheme_tr = b.addRunArtifact(grapheme_t); |
| 264 | 274 | ||
| 275 | // Emoji module | ||
| 276 | const emoji = b.addModule("Emoji", .{ | ||
| 277 | .root_source_file = b.path("src/Emoji.zig"), | ||
| 278 | .target = target, | ||
| 279 | .optimize = optimize, | ||
| 280 | }); | ||
| 281 | emoji.addAnonymousImport("emoji", .{ .root_source_file = emoji_gen_out }); | ||
| 282 | |||
| 283 | const emoji_t = b.addTest(.{ | ||
| 284 | .name = "Emoji", | ||
| 285 | .root_module = emoji, | ||
| 286 | .target = target, | ||
| 287 | .optimize = optimize, | ||
| 288 | }); | ||
| 289 | const emoji_tr = b.addRunArtifact(emoji_t); | ||
| 290 | |||
| 265 | // Word Breaking | 291 | // Word Breaking |
| 266 | const words = b.addModule("Words", .{ | 292 | const words = b.addModule("Words", .{ |
| 267 | .root_source_file = b.path("src/Words.zig"), | 293 | .root_source_file = b.path("src/Words.zig"), |
| @@ -499,6 +525,7 @@ pub fn build(b: *std.Build) void { | |||
| 499 | test_step.dependOn(&code_point_tr.step); | 525 | test_step.dependOn(&code_point_tr.step); |
| 500 | test_step.dependOn(&display_width_tr.step); | 526 | test_step.dependOn(&display_width_tr.step); |
| 501 | test_step.dependOn(&grapheme_tr.step); | 527 | test_step.dependOn(&grapheme_tr.step); |
| 528 | test_step.dependOn(&emoji_tr.step); | ||
| 502 | test_step.dependOn(&words_tr.step); | 529 | test_step.dependOn(&words_tr.step); |
| 503 | test_step.dependOn(&ascii_tr.step); | 530 | test_step.dependOn(&ascii_tr.step); |
| 504 | test_step.dependOn(&ccc_data_tr.step); | 531 | test_step.dependOn(&ccc_data_tr.step); |
diff --git a/codegen/emoji.zig b/codegen/emoji.zig new file mode 100644 index 0000000..0a4dbe6 --- /dev/null +++ b/codegen/emoji.zig | |||
| @@ -0,0 +1,146 @@ | |||
| 1 | const std = @import("std"); | ||
| 2 | const builtin = @import("builtin"); | ||
| 3 | |||
| 4 | pub const Emoji = packed struct { | ||
| 5 | Emoji: bool = false, | ||
| 6 | Emoji_Presentation: bool = false, | ||
| 7 | Emoji_Modifier: bool = false, | ||
| 8 | Emoji_Modifier_Base: bool = false, | ||
| 9 | Emoji_Component: bool = false, | ||
| 10 | Extended_Pictographic: bool = false, | ||
| 11 | }; | ||
| 12 | |||
| 13 | const block_size = 256; | ||
| 14 | const Block = [block_size]u6; | ||
| 15 | |||
| 16 | comptime { | ||
| 17 | if (@bitSizeOf(u6) != @bitSizeOf(Emoji)) { | ||
| 18 | @compileError("Emoji doesn't have expected bit size."); | ||
| 19 | } | ||
| 20 | } | ||
| 21 | |||
| 22 | const BlockMap = std.HashMap( | ||
| 23 | Block, | ||
| 24 | u16, | ||
| 25 | struct { | ||
| 26 | pub fn hash(_: @This(), k: Block) u64 { | ||
| 27 | var hasher = std.hash.Wyhash.init(0); | ||
| 28 | std.hash.autoHashStrat(&hasher, k, .DeepRecursive); | ||
| 29 | return hasher.final(); | ||
| 30 | } | ||
| 31 | |||
| 32 | pub fn eql(_: @This(), a: Block, b: Block) bool { | ||
| 33 | return std.mem.eql(u6, &a, &b); | ||
| 34 | } | ||
| 35 | }, | ||
| 36 | std.hash_map.default_max_load_percentage, | ||
| 37 | ); | ||
| 38 | |||
| 39 | pub fn main() !void { | ||
| 40 | var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator); | ||
| 41 | defer arena.deinit(); | ||
| 42 | const allocator = arena.allocator(); | ||
| 43 | |||
| 44 | var emoji_map = std.AutoHashMap(u21, Emoji).init(allocator); | ||
| 45 | defer emoji_map.deinit(); | ||
| 46 | |||
| 47 | var line_buf: [4096]u8 = undefined; | ||
| 48 | |||
| 49 | // Process Emoji | ||
| 50 | var in_file = try std.fs.cwd().openFile("data/unicode/emoji/emoji-data.txt", .{}); | ||
| 51 | defer in_file.close(); | ||
| 52 | var in_buf = std.io.bufferedReader(in_file.reader()); | ||
| 53 | const in_reader = in_buf.reader(); | ||
| 54 | |||
| 55 | while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| { | ||
| 56 | if (line.len == 0 or line[0] == '#') continue; | ||
| 57 | const no_comment = if (std.mem.indexOfScalar(u8, line, '#')) |octo| line[0..octo] else line; | ||
| 58 | |||
| 59 | var field_iter = std.mem.tokenizeAny(u8, no_comment, "; "); | ||
| 60 | var current_code: [2]u21 = undefined; | ||
| 61 | |||
| 62 | var i: usize = 0; | ||
| 63 | while (field_iter.next()) |field| : (i += 1) { | ||
| 64 | switch (i) { | ||
| 65 | 0 => { | ||
| 66 | // Code point(s) | ||
| 67 | if (std.mem.indexOf(u8, field, "..")) |dots| { | ||
| 68 | current_code = .{ | ||
| 69 | try std.fmt.parseInt(u21, field[0..dots], 16), | ||
| 70 | try std.fmt.parseInt(u21, field[dots + 2 ..], 16), | ||
| 71 | }; | ||
| 72 | } else { | ||
| 73 | const code = try std.fmt.parseInt(u21, field, 16); | ||
| 74 | current_code = .{ code, code }; | ||
| 75 | } | ||
| 76 | }, | ||
| 77 | 1 => { | ||
| 78 | // Prop | ||
| 79 | const prop = std.meta.stringToEnum(std.meta.FieldEnum(Emoji), field) orelse return error.InvalidProp; | ||
| 80 | for (current_code[0]..current_code[1] + 1) |code| { | ||
| 81 | const cp: u21 = @intCast(code); | ||
| 82 | const gop = try emoji_map.getOrPut(cp); | ||
| 83 | if (!gop.found_existing) gop.value_ptr.* = .{}; | ||
| 84 | switch (prop) { | ||
| 85 | inline else => |tag| { | ||
| 86 | @field(gop.value_ptr.*, @tagName(tag)) = true; | ||
| 87 | }, | ||
| 88 | } | ||
| 89 | } | ||
| 90 | }, | ||
| 91 | else => {}, | ||
| 92 | } | ||
| 93 | } | ||
| 94 | } | ||
| 95 | |||
| 96 | var blocks_map = BlockMap.init(allocator); | ||
| 97 | defer blocks_map.deinit(); | ||
| 98 | |||
| 99 | var stage1 = std.ArrayList(u16).init(allocator); | ||
| 100 | defer stage1.deinit(); | ||
| 101 | |||
| 102 | var stage2 = std.ArrayList(u6).init(allocator); | ||
| 103 | defer stage2.deinit(); | ||
| 104 | |||
| 105 | var block: Block = [_]u6{0} ** block_size; | ||
| 106 | var block_len: u16 = 0; | ||
| 107 | |||
| 108 | for (0..0x110000) |i| { | ||
| 109 | const cp: u21 = @intCast(i); | ||
| 110 | const emoji = emoji_map.get(cp) orelse Emoji{}; | ||
| 111 | |||
| 112 | block[block_len] = @bitCast(emoji); | ||
| 113 | block_len += 1; | ||
| 114 | |||
| 115 | if (block_len < block_size and cp != 0x10ffff) continue; | ||
| 116 | |||
| 117 | const gop = try blocks_map.getOrPut(block); | ||
| 118 | if (!gop.found_existing) { | ||
| 119 | gop.value_ptr.* = @intCast(stage2.items.len); | ||
| 120 | try stage2.appendSlice(&block); | ||
| 121 | } | ||
| 122 | |||
| 123 | try stage1.append(gop.value_ptr.*); | ||
| 124 | block_len = 0; | ||
| 125 | } | ||
| 126 | |||
| 127 | var args_iter = try std.process.argsWithAllocator(allocator); | ||
| 128 | defer args_iter.deinit(); | ||
| 129 | _ = args_iter.skip(); | ||
| 130 | const output_path = args_iter.next() orelse @panic("No output file arg!"); | ||
| 131 | |||
| 132 | const compressor = std.compress.flate.deflate.compressor; | ||
| 133 | var out_file = try std.fs.cwd().createFile(output_path, .{}); | ||
| 134 | defer out_file.close(); | ||
| 135 | var out_comp = try compressor(.raw, out_file.writer(), .{ .level = .best }); | ||
| 136 | const writer = out_comp.writer(); | ||
| 137 | |||
| 138 | const endian = builtin.cpu.arch.endian(); | ||
| 139 | try writer.writeInt(u16, @intCast(stage1.items.len), endian); | ||
| 140 | for (stage1.items) |i| try writer.writeInt(u16, i, endian); | ||
| 141 | |||
| 142 | try writer.writeInt(u16, @intCast(stage2.items.len), endian); | ||
| 143 | for (stage2.items) |i| try writer.writeInt(u8, i, endian); | ||
| 144 | |||
| 145 | try out_comp.flush(); | ||
| 146 | } | ||
diff --git a/src/Emoji.zig b/src/Emoji.zig new file mode 100644 index 0000000..75b44c2 --- /dev/null +++ b/src/Emoji.zig | |||
| @@ -0,0 +1,132 @@ | |||
| 1 | const std = @import("std"); | ||
| 2 | const builtin = @import("builtin"); | ||
| 3 | const mem = std.mem; | ||
| 4 | const Allocator = mem.Allocator; | ||
| 5 | const compress = std.compress; | ||
| 6 | const unicode = std.unicode; | ||
| 7 | |||
| 8 | const CodePoint = @import("code_point").CodePoint; | ||
| 9 | const CodePointIterator = @import("code_point").Iterator; | ||
| 10 | |||
| 11 | s1: []u16 = undefined, | ||
| 12 | s2: []u6 = undefined, | ||
| 13 | |||
| 14 | const Emoji = @This(); | ||
| 15 | |||
| 16 | // This must be an exact match of `Emoji` from `codegen/emoji.zig`. | ||
| 17 | pub const Properties = packed struct { | ||
| 18 | Emoji: bool = false, | ||
| 19 | Emoji_Presentation: bool = false, | ||
| 20 | Emoji_Modifier: bool = false, | ||
| 21 | Emoji_Modifier_Base: bool = false, | ||
| 22 | Emoji_Component: bool = false, | ||
| 23 | Extended_Pictographic: bool = false, | ||
| 24 | }; | ||
| 25 | |||
| 26 | pub fn init(allocator: Allocator) Allocator.Error!Emoji { | ||
| 27 | var emoji = Emoji{}; | ||
| 28 | try emoji.setup(allocator); | ||
| 29 | return emoji; | ||
| 30 | } | ||
| 31 | |||
| 32 | pub fn setup(emoji: *Emoji, allocator: Allocator) Allocator.Error!void { | ||
| 33 | const decompressor = compress.flate.inflate.decompressor; | ||
| 34 | const in_bytes = @embedFile("emoji"); | ||
| 35 | var in_fbs = std.io.fixedBufferStream(in_bytes); | ||
| 36 | var in_decomp = decompressor(.raw, in_fbs.reader()); | ||
| 37 | var reader = in_decomp.reader(); | ||
| 38 | |||
| 39 | const endian = builtin.cpu.arch.endian(); | ||
| 40 | |||
| 41 | const s1_len: u16 = reader.readInt(u16, endian) catch unreachable; | ||
| 42 | emoji.s1 = try allocator.alloc(u16, s1_len); | ||
| 43 | errdefer allocator.free(emoji.s1); | ||
| 44 | for (0..s1_len) |i| emoji.s1[i] = reader.readInt(u16, endian) catch unreachable; | ||
| 45 | |||
| 46 | const s2_len: u16 = reader.readInt(u16, endian) catch unreachable; | ||
| 47 | emoji.s2 = try allocator.alloc(u6, s2_len); | ||
| 48 | errdefer allocator.free(emoji.s2); | ||
| 49 | for (0..s2_len) |i| emoji.s2[i] = @intCast(reader.readInt(u8, endian) catch unreachable); | ||
| 50 | } | ||
| 51 | |||
| 52 | pub fn deinit(emoji: *const Emoji, allocator: Allocator) void { | ||
| 53 | allocator.free(emoji.s1); | ||
| 54 | allocator.free(emoji.s2); | ||
| 55 | } | ||
| 56 | |||
| 57 | /// Lookup the emoji properties for a code point. | ||
| 58 | fn properties(emoji: Emoji, cp: u21) Properties { | ||
| 59 | return @bitCast(emoji.s2[emoji.s1[cp >> 8] + (cp & 0xff)]); | ||
| 60 | } | ||
| 61 | |||
| 62 | pub fn isEmoji(emoji: Emoji, cp: u21) bool { | ||
| 63 | return properties(emoji, cp).Emoji; | ||
| 64 | } | ||
| 65 | |||
| 66 | pub fn isEmojiPresentation(emoji: Emoji, cp: u21) bool { | ||
| 67 | return properties(emoji, cp).Emoji_Presentation; | ||
| 68 | } | ||
| 69 | |||
| 70 | pub fn isEmojiModifier(emoji: Emoji, cp: u21) bool { | ||
| 71 | return properties(emoji, cp).Emoji_Modifier; | ||
| 72 | } | ||
| 73 | |||
| 74 | pub fn isEmojiModifierBase(emoji: Emoji, cp: u21) bool { | ||
| 75 | return properties(emoji, cp).Emoji_Modifier_Base; | ||
| 76 | } | ||
| 77 | |||
| 78 | pub fn isEmojiComponent(emoji: Emoji, cp: u21) bool { | ||
| 79 | return properties(emoji, cp).Emoji_Component; | ||
| 80 | } | ||
| 81 | |||
| 82 | pub fn isExtendedPictographic(emoji: Emoji, cp: u21) bool { | ||
| 83 | return properties(emoji, cp).Extended_Pictographic; | ||
| 84 | } | ||
| 85 | |||
| 86 | test "isEmoji" { | ||
| 87 | const emoji = try Emoji.init(std.testing.allocator); | ||
| 88 | defer emoji.deinit(std.testing.allocator); | ||
| 89 | |||
| 90 | try std.testing.expect(emoji.isEmoji(0x1F415)); // 🐕 | ||
| 91 | try std.testing.expect(!emoji.isEmoji(0x3042)); // あ | ||
| 92 | } | ||
| 93 | |||
| 94 | test "isEmojiPresentation" { | ||
| 95 | const emoji = try Emoji.init(std.testing.allocator); | ||
| 96 | defer emoji.deinit(std.testing.allocator); | ||
| 97 | |||
| 98 | try std.testing.expect(emoji.isEmojiPresentation(0x1F408)); // 🐈 | ||
| 99 | try std.testing.expect(!emoji.isEmojiPresentation(0x267E)); // ♾ | ||
| 100 | } | ||
| 101 | |||
| 102 | test "isEmojiModifier" { | ||
| 103 | const emoji = try Emoji.init(std.testing.allocator); | ||
| 104 | defer emoji.deinit(std.testing.allocator); | ||
| 105 | |||
| 106 | try std.testing.expect(emoji.isEmojiModifier(0x1F3FF)); // 🏿 | ||
| 107 | try std.testing.expect(!emoji.isEmojiModifier(0x1F385)); // 🎅 | ||
| 108 | } | ||
| 109 | |||
| 110 | test "isEmojiModifierBase" { | ||
| 111 | const emoji = try Emoji.init(std.testing.allocator); | ||
| 112 | defer emoji.deinit(std.testing.allocator); | ||
| 113 | |||
| 114 | try std.testing.expect(emoji.isEmojiModifierBase(0x1F977)); // 🥷 | ||
| 115 | try std.testing.expect(!emoji.isEmojiModifierBase(0x1F4F8)); // 📸 | ||
| 116 | } | ||
| 117 | |||
| 118 | test "isEmojiComponent" { | ||
| 119 | const emoji = try Emoji.init(std.testing.allocator); | ||
| 120 | defer emoji.deinit(std.testing.allocator); | ||
| 121 | |||
| 122 | try std.testing.expect(emoji.isEmojiComponent(0x1F9B0)); // 🦰 | ||
| 123 | try std.testing.expect(!emoji.isEmojiComponent(0x1F9B5)); // 🦵 | ||
| 124 | } | ||
| 125 | |||
| 126 | test "isExtendedPictographic" { | ||
| 127 | const emoji = try Emoji.init(std.testing.allocator); | ||
| 128 | defer emoji.deinit(std.testing.allocator); | ||
| 129 | |||
| 130 | try std.testing.expect(emoji.isExtendedPictographic(0x1F005)); // 🀅 | ||
| 131 | try std.testing.expect(!emoji.isExtendedPictographic(0x2A)); // * | ||
| 132 | } | ||