Use more optimal perfect hash table

Author: nielsdos

ext/mbstring/libmbfl/mbfl/mbfl_encoding.c

@@ -146,179 +146,168 @@ static const mbfl_encoding *mbfl_encoding_ptr_list[] = {
 
 /* The following perfect hashing table was amended from gperf, and hashing code was generated using gperf.
  * The table was amended to refer to the table above such that it is lighter for the data cache.
- * Command used: gperf encodings.txt --readonly-tables --null-strings --ignore-case
- * The encodings.txt contains all the contents of the name fields of the mbfl_encoding_ptr_list table. */
+ * You can use the generate_name_perfect_hash_table.php script to help generate the necessary lookup tables. */
 
-static const int8_t mbfl_encoding_ptr_list_after_hashing[231] = {
-	-1, -1,
-	61,
+static const int8_t mbfl_encoding_ptr_list_after_hashing[] = {
+	-1, -1, -1,
 	66,
+	-1, -1, -1, -1,
+	-1, -1, -1, -1,
+	9,
+	-1, -1, -1, -1,
+	-1,
 	23,
-	73,
+	18,
 	59,
-	-1,
-	1,
+	78,
 	-1, -1, -1,
-	11,
+	39,
+	73,
+	71,
+	40,
 	-1,
-	5,
-	9,
+	75,
 	-1,
-	10,
-	38,
+	67,
+	35,
 	-1,
+	76,
+	1,
+	61,
+	46,
+	-1,
+	21,
+	-1,
+	11,
+	-1, -1,
 	52,
 	54,
+	64,
+	12,
+	10,
 	-1,
-	2,
-	40,
-	46,
-	27,
-	76,
-	26,
-	-1,
+	3,
 	49,
 	57,
+	69,
+	70,
+	-1, -1,
+	74,
+	50,
+	58,
+	63,
+	65,
 	-1,
-	75,
+	14,
+	30,
+	53,
+	5,
+	-1, -1, -1,
+	13,
 	-1,
+	48,
+	56,
+	60,
+	-1,
+	8,
+	36,
+	20,
 	47,
 	55,
-	78,
-	36,
+	6,
 	-1,
-	50,
-	58,
-	8,
-	-1, -1,
-	69,
-	39,
 	7,
-	-1, -1,
-	64,
-	67,
-	-1, -1,
-	30,
-	48,
-	56,
-	-1, -1, -1,
-	35,
-	74,
-	-1, -1,
-	24,
-	53,
-	62,
-	43,
-	-1, -1,
+	-1,
+	27,
 	45,
-	22,
-	-1, -1, -1,
-	6,
-	3,
-	-1, -1, -1,
-	18,
-	71,
-	-1, -1, -1,
-	21,
+	26,
+	31,
 	-1,
 	37,
-	-1,
-	4,
-	60,
-	25,
-	-1, -1,
-	72,
+	17,
+	24,
 	51,
 	-1,
-	44,
-	29,
+	33,
+	15,
 	-1,
+	16,
+	-1,
+	38,
+	-1, -1,
 	28,
-	0,
+	34,
 	-1,
-	14,
-	31,
-	63,
-	12,
+	72,
+	-1, -1, -1,
+	0,
 	-1,
-	13,
-	33,
-	-1, -1,
-	68,
-	-1, -1, -1, -1,
+	29,
 	-1, -1,
-	20,
-	-1, -1, -1, -1,
-	-1, -1, -1,
+	4,
+	19,
+	22,
 	77,
 	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	65,
-	-1, -1, -1, -1,
-	70,
+	32,
+	2,
 	-1, -1, -1, -1,
 	-1,
-	41,
+	43,
 	-1, -1, -1, -1,
-	-1,
-	17,
-	-1, -1, -1,
+	-1, -1,
 	42,
-	16,
+	-1,
+	41,
 	-1, -1, -1, -1,
 	-1, -1, -1, -1,
 	-1, -1, -1, -1,
-	15,
 	-1, -1, -1, -1,
-	34,
+	-1,
+	25,
 	-1, -1, -1, -1,
 	-1, -1, -1,
-	32,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
-	-1, -1, -1, -1,
+	44,
 	-1, -1, -1, -1,
 	-1, -1, -1, -1,
 	-1, -1, -1, -1,
 	-1,
-	19
+	68,
+	-1, -1, -1, -1,
+	-1, -1, -1,
+	62
 };
 
 static unsigned int mbfl_name2encoding_perfect_hash(const char *str, size_t len)
 {
 	static const unsigned char asso_values[] =
 	{
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231,   5, 231, 231,   0,  50,
-			5,  15,  35,  10,  20,  75,   0,  45, 231, 231,
-		231, 231, 231, 231, 231,  80,   5,   0,   0,   0,
-		 75,  75,   0,   0,  15,  70,   0,   5,   0,   0,
-		 25,  55,  30,   0,  10,   0, 231,  25, 231, 231,
-			0, 231, 231, 231, 231, 231, 231,  80,   5,   0,
-			0,   0,  75,  75,   0,   0,  15,  70,   0,   5,
-			0,   0,  25,  55,  30,   0,  10,   0, 231,  25,
-		231, 231,   0, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231, 231, 231, 231, 231,
-		231, 231, 231, 231, 231, 231
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189,   7, 189, 189,   7,  49,
+		0,  42,  35,  14,  21,  56,   7,  28, 189, 189,
+		189, 189, 189, 189, 189,  35,   7,   0,   0,  14,
+		0,  21,  28,   0,  84,  14,   0,  56,   0,   0,
+		7,   0,  14,   7,  56,   0, 189,  56, 189, 189,
+		7, 189, 189, 189, 189, 189, 189,  35,   7,   0,
+		0,  14,   0,  21,  28,   0,  84,  14,   0,  56,
+		0,   0,   7,   0,  14,   7,  56,   0, 189,  56,
+		189, 189,   7, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189, 189, 189, 189, 189,
+		189, 189, 189, 189, 189, 189
 	};
 	unsigned int hval = len;
 
@@ -377,7 +366,7 @@ const mbfl_encoding *mbfl_name2encoding_ex(const char *name, size_t name_len)
 	/* Use perfect hash lookup for name */
 	if (name_len <= NAME_HASH_MAX_NAME_LENGTH && name_len >= NAME_HASH_MIN_NAME_LENGTH) {
 		unsigned int key = mbfl_name2encoding_perfect_hash(name, name_len);
-		if (key <= sizeof(mbfl_encoding_ptr_list_after_hashing)) {
+		if (key < sizeof(mbfl_encoding_ptr_list_after_hashing) / sizeof(mbfl_encoding_ptr_list_after_hashing[0])) {
 			int8_t offset = mbfl_encoding_ptr_list_after_hashing[key];
 			if (offset >= 0) {
 				encoding = mbfl_encoding_ptr_list + offset;