1 files changed, 32 insertions, 42 deletions

diff --git a/src/system/utf8.c b/src/system/utf8.c
index cb14691..35b0f2f 100644
--- a/src/system/utf8.c
+++ b/src/system/utf8.c
@@ -1,54 +1,42 @@
 #include "../../include/mgl/system/utf8.h"
 
+static inline bool utf8_get_codepoint_length(unsigned char b, size_t *codepoint_length) {
+    if((b & 0x80) == 0) {
+        *codepoint_length = 1;
+        return true;
+    } else if((b & 0xE0) == 0xC0) {
+        *codepoint_length = 2;
+        return true;
+    } else if((b & 0xF0) == 0xE0) {
+        *codepoint_length = 3;
+        return true;
+    } else if((b & 0xF8) == 0xF0) {
+        *codepoint_length = 4;
+        return true;
+    } else {
+        return false;
+    }
+}
+
 /* TODO: Optimize (remove branching, etc) */
-bool mgl_utf8_is_valid(const unsigned char *str, size_t size) {
-    size_t i = 0;
-    while(i < size) {
-        size_t codepoint_length = 0;
-        const unsigned char b = str[i];
+bool mgl_utf8_decode(const unsigned char *str, size_t size, uint32_t *decoded_codepoint, size_t *codepoint_length) {
+    if(size == 0)
+        return false;
 
-        if((b & 0x80) == 0)
-            codepoint_length = 1;
-        else if((b & 0xE0) == 0xC0)
-            codepoint_length = 2;
-        else if((b & 0xF0) == 0xE0)
-            codepoint_length = 3;
-        else if((b & 0xF8) == 0xF0)
-            codepoint_length = 4;
-        else
-            return false;
+    size_t clen;
+    if(!utf8_get_codepoint_length(str[0], &clen))
+        return false;
 
-        const size_t next = i + codepoint_length;
-        if(next > size)
-            return false;
+    if(size < clen)
+        return false;
 
-        /* TODO: Remove this overflow check? */
-        /* Check overflow */
-        if(next <= i)
+    for(size_t i = 1; i < clen; ++i) {
+        if((str[i] & 0xC0) != 0x80)
             return false;
-
-        ++i;
-        for(; i < next; ++i) {
-            if((str[i] & 0xC0) != 0x80)
-                return false;
-        }
     }
-    return true;
-}
-
-static inline size_t utf8_get_codepoint_length(unsigned char b) {
-    const unsigned int length1 = b >> 7;
-    const unsigned int length2 = length1 & ((b & 0x40) >> 6);
-    const unsigned int length3 = length2 & ((b & 0x20) >> 5);
-    const unsigned int length4 = length3 & ((b & 0x10) >> 4);
-    return (length1 ^ 0x01) + length1 + length2 + length3 + length4;
-}
 
-/* TODO: Optimize (remove branching, etc) */
-size_t mgl_utf8_decode(const unsigned char *str, uint32_t *decoded_codepoint) {
-    const size_t length = utf8_get_codepoint_length(str[0]);
     uint32_t codepoint;
-    switch(length) {
+    switch(clen) {
         case 1:
             codepoint =  (uint32_t)(str[0] & 0x7F);
             break;
@@ -68,6 +56,8 @@ size_t mgl_utf8_decode(const unsigned char *str, uint32_t *decoded_codepoint) {
             codepoint |=  (uint32_t)(str[3] & 0x3F);
             break;
     }
+
+    *codepoint_length = clen;
     *decoded_codepoint = codepoint;
-    return length;
+    return true;
 }