#!/usr/bin/python3 import math import sys from lib import * def parse(file: str) -> list[bool]: xs = [False] * 0x110000 sys.argv[1] = ({ 'Composition_Exclusion': '', 'Indic_Conjunct_Break': 'InCB;', }).get(sys.argv[1], sys.argv[1]) with open(file, 'r') as f: for line in f.readlines(): if ( len(line.strip()) == 0 or line[0] == '#' or sys.argv[1] not in line ): continue parts = [ int(x, 16) for x in ( line .split('#')[0] .split(';')[0] .strip() .split('..') ) ] for i in range(parts[0], parts[len(parts) - 1] + 1): xs[i] = True 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/bool-props; DO NOT EDIT. */ #include "bitset.h" #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() bcnt = blksize // 8 bpc = columns(bcnt, 5) print(f'static constexpr unsigned char stage2[][{bcnt}] = {{') for c in cs: x = sum(map(lambda x: x[1] << x[0], enumerate(c))) for i in range(bcnt // bpc): print('\t{' if i == 0 else '\t ', end='') for j in range(bpc): print('0x%02X' % (x & 0xFF), end='') x >>= 8 if i < bcnt // bpc - 1 or j < bpc - 1: print(',', end='') if j < bpc - 1: print(' ', end='') if i < bcnt // bpc - 1: print() print('},') print('};') print() print(f'''\ bool uprop_is_{sys.argv[2]}(rune ch) {{ return TESTBIT(stage2[stage1[ch / {blksize}]], ch % {blksize}); }}''') def main() -> None: if len(sys.argv) != 4: print('Usage: bool-props.py name shortname file', file=sys.stderr) exit(1) xs = parse(sys.argv[3]) 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 sz = sz_s1 + sz_s2 if sz < smallest: smallest = sz blksize = bs Cs = [tuple(x) for x in chunks(xs, blksize)] genfile(Cs, blksize) sys.argv[0] = sys.argv[2] report_size(len(xs), smallest) if __name__ == '__main__': main()