Optimize out bounds check in UHC decoder

This gives a 25% speed boost for conversion operations on long strings
(~10,000 codepoints). For shorter strings, the speed boost is less
(as the input gets smaller, it is progressively swamped more and more
by the overhead of entering and exiting the conversion function).

When benchmarking string conversion speed, we are measuring not only
the speed of the decoder, but also the time which it takes to re-encode
the string in another encoding like UTF-8 or UTF-16. So the performance
increase for functions which only need to decode but not re-encode the
input string will be much more than 25%.

Author: alexdowad

ext/mbstring/libmbfl/filters/mbfilter_uhc.c

@@ -196,12 +196,16 @@ static size_t mb_uhc_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf,
 	unsigned char *p = *in, *e = p + *in_len;
 	uint32_t *out = buf, *limit = buf + bufsize;
 
+	e--; /* Stop the main loop 1 byte short of the end of the input */
+
 	while (p < e && out < limit) {
 		unsigned char c = *p++;
 
 		if (c < 0x80) {
 			*out++ = c;
-		} else if (c > 0x80 && c < 0xFE && c != 0xC9 && p < e) {
+		} else if (c > 0x80 && c < 0xFE && c != 0xC9) {
+			/* We don't need to check p < e here; it's not possible that this pointer dereference
+			 * will be outside the input string, because of e-- above */
 			unsigned char c2 = *p++;
 			if (c2 < 0x41 || c2 == 0xFF) {
 				*out++ = MBFL_BAD_INPUT;
@@ -227,7 +231,13 @@ static size_t mb_uhc_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf,
 		}
 	}
 
-	*in_len = e - p;
+	/* Finish up last byte of input string if there is one */
+	if (p == e && out < limit) {
+		unsigned char c = *p++;
+		*out++ = (c < 0x80) ? c : MBFL_BAD_INPUT;
+	}
+
+	*in_len = e - p + 1;
 	*in = p;
 	return out - buf;
 }