From 1e4ad99311611a7fbb5cdfe4b1b79202381dc717 Mon Sep 17 00:00:00 2001
From: "Luke T. Shumaker" <lukeshu@lukeshu.com>
Date: Sun, 13 Apr 2025 22:13:43 -0600
Subject: Split lib9p into lib9p_core and lib9p_srv

---
 lib9p/core_utf8.h | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100644 lib9p/core_utf8.h

(limited to 'lib9p/core_utf8.h')

diff --git a/lib9p/core_utf8.h b/lib9p/core_utf8.h
new file mode 100644
index 0000000..636d4eb
--- /dev/null
+++ b/lib9p/core_utf8.h
@@ -0,0 +1,34 @@
+/* lib9p/core_utf8.h - Internal UTF-8 validation
+ *
+ * Copyright (C) 2024-2025  Luke T. Shumaker <lukeshu@lukeshu.com>
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+#ifndef _LIB9P_CORE_UTF8_H_
+#define _LIB9P_CORE_UTF8_H_
+
+static inline bool _is_valid_utf8(uint8_t *str, size_t len, bool forbid_nul) {
+	uint32_t ch;
+	uint8_t chlen;
+	assert(str);
+	for (size_t pos = 0; pos < len;) {
+		if      ((str[pos] & 0b10000000) == 0b00000000) { ch = str[pos] & 0b01111111; chlen = 1; }
+		else if ((str[pos] & 0b11100000) == 0b11000000) { ch = str[pos] & 0b00011111; chlen = 2; }
+		else if ((str[pos] & 0b11110000) == 0b11100000) { ch = str[pos] & 0b00001111; chlen = 3; }
+		else if ((str[pos] & 0b11111000) == 0b11110000) { ch = str[pos] & 0b00000111; chlen = 4; }
+		else return false;
+		if ((ch == 0 && (chlen != 1 || forbid_nul)) || pos + chlen > len) return false;
+		for (uint8_t i = 1; i < chlen; i++) {
+			if ((str[pos+i] & 0b11000000) != 0b10000000) return false;
+			ch = (ch << 6) | (str[pos+i] & 0b00111111);
+		}
+		if (ch > 0x10FFFF) return false;
+		pos += chlen;
+	}
+	return true;
+}
+
+#define is_valid_utf8(str, len)                 _is_valid_utf8(str, len, false)
+#define is_valid_utf8_without_nul(str, len)     _is_valid_utf8(str, len, true)
+
+#endif /* _LIB9P_CORE_UTF8_H_ */
-- 
cgit v1.2.3-2-g168b