stopping point for tonight

unicode/utf8.c

@@ -12,7 +12,7 @@ int main(int argc, char *argv[]){
 	int i;
 	int l; /* line counter */
 	char *n;
-	long int utf32_lit;
+	long unsigned int utf32_lit;
 	unsigned char utf8_bytes[8];
 
 	c = '\0';
@@ -50,20 +50,21 @@ int main(int argc, char *argv[]){
 				return 0;
 			continue;
 		}
-		switch(i =	(utf32_lit < 65536)
-			+ (utf32_lit < 2048)
-			+ (utf32_lit < 128)){
+		/* something with the bit math is broken */
+		switch(i =	(utf32_lit < 0x10000)
+			+ (utf32_lit < 0x0800)
+			+ (utf32_lit < 0x0080)){
 		case 0: utf8_bytes[0] =
-			((utf32_lit >> 18) & 7) + 240; /* 11110xxx */
+			0xF0 + ((utf32_lit >> 18) & 0x07); /* 11110xxx */
 		case 1:	utf8_bytes[1] = i == 1
-			? ((utf32_lit >> 12) & 15) + 224 /* 1110xxxx */
-			: ((utf32_lit >> 12) & 63) + 80; /* 10xxxxxx */
+			? 0xE0 + ((utf32_lit >> 12) & 0x0F) /* 1110xxxx */
+			: 0x50 + ((utf32_lit >> 12) & 0x3F); /* 10xxxxxx */
 		case 2:	utf8_bytes[2] = i == 2
-			? ((utf32_lit >> 6) & 31) + 192 /* 110xxxxx */
-			: ((utf32_lit >> 6) & 63) + 80; /* 10xxxxxx */
+			? 0xC0 + ((utf32_lit >> 6) & 0x1F) /* 110xxxxx */
+			: 0x50 + ((utf32_lit >> 6) & 0x3F); /* 10xxxxxx */
 		case 3:	utf8_bytes[3] = i == 3
-			? utf8_bytes[3] = utf32_lit & 127 /* 0xxxxxxx */
-			: (utf32_lit & 63) + 80; /* 10xxxxxx */
+			? utf8_bytes[3] = utf32_lit & 0x7F /* 0xxxxxxx */
+			: 0x50 + (utf32_lit & 0x3F); /* 10xxxxxx */
 		}
 		for( ; i < 4; ++i)
 			putc(utf8_bytes[i], stdout);