#!/usr/bin/python3 import math from lib import * def parse(file: str) -> list[bool]: xs = ['GC_CN'] * 0x110000 with open(file, 'r') as f: for line in f.readlines(): parts = line.split(';') parts[0] = int(parts[0], 16) if 'First' in parts[1]: lo = parts[0] elif 'Last' in parts[1]: hi = parts[0] for i in range(lo, hi + 1): xs[i] = f'GC_{parts[2].upper()}' else: xs[parts[0]] = f'GC_{parts[2].upper()}' return xs def genfile(cs: list[tuple[bool, ...]], blksize: int) -> None: Cs = cs cs = list(dict.fromkeys(Cs)) print('''\ /* This file is autogenerated by gen/prop/gc; DO NOT EDIT. */ #include "unicode/prop.h" ''') print(f'static constexpr {typename(len(cs) - 1)} stage1[] = {{') for i, c in enumerate(Cs): print(f'%c%{len(str(len(cs) - 1))}d,' % ('\t' if i % 16 == 0 else ' ', cs.index(c)), end='') if i % 16 == 15: print() print('};') print() ppc = columns(blksize, 6) print(f'static constexpr enum uprop_gc stage2[][{blksize}] = {{') for c in cs: for i in range(blksize // ppc): print('\t{' if i == 0 else '\t ', end='') for j in range(ppc): print(c[i*ppc + j], end='') if i < blksize // ppc - 1 or j < ppc - 1: print(',', end='') if j < ppc - 1: print(' ', end='') if i < blksize // ppc - 1: print() print('},') print('};') print() print(f'''\ enum uprop_gc uprop_get_gc(rune ch) {{ return stage2[stage1[ch / {blksize}]][ch % {blksize}]; }}''') def main() -> None: cwd_init() sys.stdout = open('lib/unicode/prop/uprop_get_gc.c', 'w') xs = parse('data/UnicodeData') blksize = -1 smallest = math.inf for bs in powers_of_2(): if bs > len(xs): break Cs = [tuple(x) for x in chunks(xs, bs)] cs = set(Cs) sz_s1 = len(Cs) * isize(len(cs) - 1) sz_s2 = len(cs) * bs * 4 sz = sz_s1 + sz_s2 if sz < smallest: smallest = sz blksize = bs Cs = [tuple(x) for x in chunks(xs, blksize)] genfile(Cs, blksize) report_size(len(xs), smallest) if __name__ == '__main__': main()