Add uprop_get_dm()

1 files changed, 115 insertions, 0 deletions

diff --git a/gen/prop/dm b/gen/prop/dm
new file mode 100755
index 0000000..5d2d98e
--- /dev/null
+++ b/gen/prop/dm
@@ -0,0 +1,115 @@
+#!/usr/bin/python3
+
+import math
+
+from lib import *
+
+
+longest = 0
+
+def parse(file: str) -> list[str]:
+	global longest
+
+	xs = ['{}'] * 0x110000
+	with open(file, 'r') as f:
+		for line in f.readlines():
+			if len(line.strip()) == 0 or line[0] == '#':
+				continue
+
+			parts = line.split(';')
+			if parts[5] == '':
+				continue
+			n = int(parts[0], 16)
+			xs[n] = (
+				'_('
+				+ ', '.join(f'0x{x}' for x in parts[5].split() if '<' not in x)
+				+ ')'
+			)
+			longest = max(longest, len(xs[n]))
+	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/dm; DO NOT EDIT. */
+
+#include <stdint.h>
+
+#include "macros.h"
+#include "unicode/prop.h"
+
+#define _(...) \\
+	{(const rune []){__VA_ARGS__}, lengthof(((const rune []){__VA_ARGS__}))}
+''')
+
+	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, longest + 1)
+	print(f'static const struct rview 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(' ' * (longest + 1 - len(c[i*ppc + j])), end='')
+			if i < blksize // ppc - 1:
+				print()
+		print('},')
+	print('};')
+
+	print()
+
+	print(f'''\
+struct rview
+uprop_get_dm(rune ch)
+{{
+	static thread_local rune hack;
+	struct rview rv =  stage2[stage1[ch / {blksize}]][ch % {blksize}];
+	if (rv.p != nullptr)
+		return rv;
+	hack = ch;
+	return (struct rview){{&hack, 1}};
+}}''')
+
+def main() -> None:
+	cwd_init()
+	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 * 16  # (rune *) + size_t
+		sz = sz_s1 + sz_s2
+
+		if sz < smallest:
+			smallest = sz
+			blksize = bs
+
+	Cs = [tuple(x) for x in chunks(xs, blksize)]
+	with open('lib/unicode/prop/uprop_get_dm.c', 'w') as f:
+		sys.stdout = f
+		genfile(Cs, blksize)
+
+	report_size(len(xs), smallest)
+
+if __name__ == '__main__':
+	main()