From 961d2ec2f9075f30e9264ed2db6c394cfa5967f9 Mon Sep 17 00:00:00 2001
From: Jose Colon Rodriguez
Date: Mon, 25 Mar 2024 20:05:36 -0400
Subject: CaseData

---
 codegen/upper.zig | 58 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 58 insertions(+)
 create mode 100644 codegen/upper.zig

(limited to 'codegen/upper.zig')

diff --git a/codegen/upper.zig b/codegen/upper.zig
new file mode 100644
index 0000000..6fea608
--- /dev/null
+++ b/codegen/upper.zig
@@ -0,0 +1,58 @@
+const std = @import("std");
+const builtin = @import("builtin");
+
+pub fn main() !void {
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+    defer arena.deinit();
+    const allocator = arena.allocator();
+
+    // Process UnicodeData.txt
+    var in_file = try std.fs.cwd().openFile("data/unicode/UnicodeData.txt", .{});
+    defer in_file.close();
+    var in_buf = std.io.bufferedReader(in_file.reader());
+    const in_reader = in_buf.reader();
+
+    var args_iter = try std.process.argsWithAllocator(allocator);
+    defer args_iter.deinit();
+    _ = args_iter.skip();
+    const output_path = args_iter.next() orelse @panic("No output file arg!");
+
+    const compressor = std.compress.deflate.compressor;
+    var out_file = try std.fs.cwd().createFile(output_path, .{});
+    defer out_file.close();
+    var out_comp = try compressor(allocator, out_file.writer(), .{ .level = .best_compression });
+    defer out_comp.deinit();
+    const writer = out_comp.writer();
+
+    const endian = builtin.cpu.arch.endian();
+    var line_buf: [4096]u8 = undefined;
+
+    lines: while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| {
+        if (line.len == 0) continue;
+
+        var field_iter = std.mem.splitScalar(u8, line, ';');
+        var cps: [2]u24 = undefined;
+
+        var i: usize = 0;
+        while (field_iter.next()) |field| : (i += 1) {
+            switch (i) {
+                0 => cps[0] = try std.fmt.parseInt(u24, field, 16),
+
+                12 => {
+                    // Simple uppercase mapping
+                    if (field.len == 0) continue :lines;
+                    cps[1] = try std.fmt.parseInt(u24, field, 16);
+                },
+
+                2 => if (line[0] == '<') continue :lines,
+
+                else => {},
+            }
+        }
+
+        for (&cps) |cp| try writer.writeInt(u24, cp, endian);
+    }
+
+    try writer.writeInt(u24, 0, endian);
+    try out_comp.flush();
+}
-- 
cgit v1.2.3