lots of features & bugfixes both

1 files changed, 6 insertions, 6 deletions

diff --git a/utf8.h b/utf8.h
index 01c3336..164279b 100644
--- a/utf8.h
+++ b/utf8.h
@@ -6,6 +6,9 @@
 #include "wrmr.h"
 
 #define UTF8_INVALID 0xFFFD /* replacement character */
+#define UTF8_CP_LEN_BITS   ((uint64_t)0xFFEAA550000)
+#define UTF8_CP_SHIFT(cp)  ((32 - stdc_leading_zeros((uint32_t)(cp))) << 1)
+#define UTF8_CP_LEN(cp)    (1 + ((UTF8_CP_LEN_BITS >> UTF8_CP_SHIFT(cp)) & 3))
 
 u32 utf8_decode_len(const char *src, u32 ch_count);
 u32 utf8_encode_len(const u32 *src, u32 cp_count);
@@ -18,9 +21,6 @@ u32 utf8_decode_at(const char *s, u32 i, u32 n);
 #include <stdbit.h>
 
 /* packed array of 2-bit lengths for codepoints 0..10FFFF */
-#define UTF8_CP_LEN_BITS   ((uint64_t)0xFFEAA550000)
-#define UTF8_CP_SHIFT(cp)  ((32 - stdc_leading_zeros((uint32_t)(cp))) << 1)
-#define UTF8_CP_LEN(cp)    (1 + ((UTF8_CP_LEN_BITS >> UTF8_CP_SHIFT(cp)) & 3))
 
 u32 utf8_encode_len(const u32 *src, u32 cp_count) {
 	u32 len = 0;
@@ -72,14 +72,14 @@ void utf8_decode(u32 *dst, const char *src, u32 cp_count) {
 
 u32 utf8_decode_at(const char *s, u32 i, u32 n) {
 	if (i >= n) return 0;
-	u32 cp = (u8)s[i++];
+	u32 cp = (u8)s[i];
 	u32 b = stdc_leading_ones((u8)cp);
 	if (!b) return cp;
 	u32 end = i + b - 1;
 	if (end >= n) return 0;
 	cp &= 0xff >> b;
-	while (i < end) {
-		u8 c = s[i++];
+	while (++i <= end) {
+		u8 c = s[i];
 		cp = (cp << 6) | (c & 0x3f);
 	}
 	return cp;