utils: Add utf8_check function

Add 'utf8_check' function to check if a given string is utf8 encoded. Signed-off-by: James Roy <rruuaanng@outlook.com>
zephyrproject-rtos · Nov 2, 2024 · ed28e46 · ed28e46
1 parent 2f23313
commit ed28e46
Show file tree

Hide file tree

Showing 3 changed files with 53 additions and 0 deletions.
diff --git a/include/zephyr/sys/util.h b/include/zephyr/sys/util.h
@@ -685,6 +685,16 @@ char *utf8_trunc(char *utf8_str);
  */
 char *utf8_lcpy(char *dst, const char *src, size_t n);
 
+/**
+ * @brief Checks if the given string @p str is UTF-8 encode.
+ *
+ * @param str Target string
+ * @param len The length of string @p str
+ *
+ * @return true if @p str is UTF-8 encode, and false otherwise.
+ */
+bool utf8_check(const uint8_t *str, int len);
+
 #define __z_log2d(x) (32 - __builtin_clz(x) - 1)
 #define __z_log2q(x) (64 - __builtin_clzll(x) - 1)
 #define __z_log2(x) (sizeof(__typeof__(x)) > 4 ? __z_log2q(x) : __z_log2d(x))

diff --git a/lib/utils/utf8.c b/lib/utils/utf8.c
@@ -7,6 +7,7 @@
 #include <stdint.h>
 #include <string.h>
 #include <zephyr/sys/__assert.h>
+#include <zephyr/sys/util.h>
 
 #define ASCII_CHAR 0x7F
 #define SEQUENCE_FIRST_MASK 0xC0
@@ -79,3 +80,36 @@ char *utf8_lcpy(char *dst, const char *src, size_t n)
 
 	return dst;
 }
+
+bool utf8_check(uint8_t *str, int len)
+{
+	int i = 0;
+	int nbyte = 0;
+
+	if (len < 0) {
+		errno = EINVAL;
+		return false;
+	}
+
+	while (i < len) {
+		if (str[i] <= 0x7F && str[i] >= 0x00) {
+			i++;
+			continue;
+		} else {
+			if (str[i] <= 0xDF && str[i] >= 0xC2) {
+				nbyte = 2;
+			} else if (str[i] <= 0xEF && str[i] >= 0xE0) {
+				nbyte = 3;
+			} else if (str[i] <= 0xF4 && str[i] >= 0xF0) {
+				nbyte = 4;
+			} else {
+				return false;
+			}
+		}
+		if (i + nbyte > len) {
+			return false;
+		}
+		i += nbyte;
+	}
+	return true;
+}
diff --git a/tests/unit/util/main.c b/tests/unit/util/main.c
@@ -867,6 +867,15 @@ ZTEST(util, test_utf8_lcpy_truncated)
 	zassert_str_equal(dest_str, expected_result, "Failed to copy");
 }
 
+ZTEST(util, test_utf8_check)
+{
+	uint8_t s1[] = "€€€";
+	uint8_t s2[] = "\xe2";
+
+	zassert_true(utf8_check(s1, strlen(s1)), "Failed to check");
+	zassert_false(utf8_check(s2, strlen(s2)), "Failed to check");
+}
+
 ZTEST(util, test_utf8_lcpy_not_truncated)
 {
 	/* dest_str size is based on storing 3 * € plus the null terminator  */