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/numeric.zig | 45 ++++++++++++++++++++------------------------- 1 file changed, 20 insertions(+), 25 deletions(-) (limited to 'codegen/numeric.zig') diff --git a/codegen/numeric.zig b/codegen/numeric.zig index 80fc7c2..ad8490c 100644 --- a/codegen/numeric.zig +++ b/codegen/numeric.zig @@ -1,5 +1,6 @@ const std = @import("std"); const builtin = @import("builtin"); +const mem = std.mem; const block_size = 256; const Block = [block_size]u8; @@ -15,7 +16,7 @@ const BlockMap = std.HashMap( } pub fn eql(_: @This(), a: Block, b: Block) bool { - return std.mem.eql(u8, &a, &b); + return mem.eql(u8, &a, &b); } }, std.hash_map.default_max_load_percentage, @@ -31,17 +32,17 @@ pub fn main() !void { var line_buf: [4096]u8 = undefined; - // Process DerivedCombiningClass.txt - var num_file = try std.fs.cwd().openFile("data/unicode/extracted/DerivedNumericType.txt", .{}); - defer num_file.close(); - var num_buf = std.io.bufferedReader(num_file.reader()); - const num_reader = num_buf.reader(); + // Process DerivedNumericType.txt + var in_file = try std.fs.cwd().openFile("data/unicode/extracted/DerivedNumericType.txt", .{}); + defer in_file.close(); + var in_buf = std.io.bufferedReader(in_file.reader()); + const in_reader = in_buf.reader(); - while (try num_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| { + while (try in_reader.readUntilDelimiterOrEof(&line_buf, '\n')) |line| { if (line.len == 0 or line[0] == '#') continue; - const no_comment = if (std.mem.indexOfScalar(u8, line, '#')) |octo| line[0..octo] else line; + const no_comment = if (mem.indexOfScalar(u8, line, '#')) |octo| line[0..octo] else line; - var field_iter = std.mem.tokenizeAny(u8, no_comment, "; "); + var field_iter = mem.tokenizeAny(u8, no_comment, "; "); var current_code: [2]u21 = undefined; var i: usize = 0; @@ -49,7 +50,7 @@ pub fn main() !void { switch (i) { 0 => { // Code point(s) - if (std.mem.indexOf(u8, field, "..")) |dots| { + if (mem.indexOf(u8, field, "..")) |dots| { current_code = .{ try std.fmt.parseInt(u21, field[0..dots], 16), try std.fmt.parseInt(u21, field[dots + 2 ..], 16), @@ -61,25 +62,19 @@ pub fn main() !void { }, 1 => { // Numeric type - if (std.mem.eql(u8, field, "Numeric")) { - for (current_code[0]..current_code[1] + 1) |cp| { - const gop = try flat_map.getOrPut(@intCast(cp)); - if (!gop.found_existing) gop.value_ptr.* = 0; - gop.value_ptr.* |= 1; - } - } else if (std.mem.eql(u8, field, "Digit")) { - for (current_code[0]..current_code[1] + 1) |cp| { - const gop = try flat_map.getOrPut(@intCast(cp)); - if (!gop.found_existing) gop.value_ptr.* = 0; - gop.value_ptr.* |= 2; - } - } else if (std.mem.eql(u8, field, "Decimal")) { + var bit: u8 = 0; + + if (mem.eql(u8, field, "Numeric")) bit = 1; + if (mem.eql(u8, field, "Digit")) bit = 2; + if (mem.eql(u8, field, "Decimal")) bit = 4; + + if (bit != 0) { for (current_code[0]..current_code[1] + 1) |cp| { const gop = try flat_map.getOrPut(@intCast(cp)); if (!gop.found_existing) gop.value_ptr.* = 0; - gop.value_ptr.* |= 4; + gop.value_ptr.* |= bit; } - } else continue; + } }, else => {}, } -- cgit v1.2.3