X-Git-Url: http://mj.ucw.cz/gitweb/?a=blobdiff_plain;f=lib%2Fhashfunc.c;h=8fca5f4139254cd211bd71eb9a4edb04fe4d90aa;hb=029a7f6703b0a0a102f7dff6c25ff1d419d75610;hp=980b4077f673b339386ed7a30e0b4e696d818531;hpb=b0cbde12a98264c2629b52f618678e7f1eacf097;p=libucw.git

diff --git a/lib/hashfunc.c b/lib/hashfunc.c
index 980b4077..8fca5f41 100644
--- a/lib/hashfunc.c
+++ b/lib/hashfunc.c
@@ -1,11 +1,14 @@
 /*
- *	Hyper-super-meta-alt-control-shift extra fast str_len() and hash_*()
- *	routines
+ *	UCW Library -- Hyper-super-meta-alt-control-shift extra fast
+ *	str_len() and hash_*() routines
  *
  *	It is always at least as fast as the classical strlen() routine and for
  *	strings longer than 100 characters, it is substantially faster.
  *
  *	(c) 2002, Robert Spalek <robert@ucw.cz>
+ *
+ *	This software may be freely distributed and used according to the terms
+ *	of the GNU Lesser General Public License.
  */
 
 #include "lib/lib.h"
@@ -14,7 +17,7 @@
 
 /* The number of bits the hash in the function hash_*() is rotated by after
  * every pass.  It should be prime with the word size.  */
-#define	SHIFT_BITS	5
+#define	SHIFT_BITS	7
 
 /* A bit-mask which clears higher bytes than a given threshold.  */
 static uns mask_higher_bits[sizeof(uns)];
@@ -37,19 +40,21 @@ hashfunc_init(void)
 static inline uns CONST
 str_len_uns(uns x)
 {
-	const uns sub = ((uns) -1) / 0xff;
+	const uns sub = ~0U / 0xff;
 	const uns and = sub * 0x80;
 	uns a, i;
 	byte *bytes;
-	a = (x ^ (x - sub)) & and;
-	/* 
+	a = ~x & (x - sub) & and;
+	/*
 	 * x_2 = x - 0x01010101;
-	 * x_3 = x ^ x_2;
+	 * x_3 = ~x & x_2;
 	 * a = x_3 & 0x80808080;
 	 *
-	 * If no byte of x is in {0, 0x80}, then the highest bit of each byte
-	 * of x_2 is the same as of x.  Hence x_3 has all these highest bits
-	 * cleared.  If a == 0, then we are sure there is no zero byte in x.
+	 * If all bytes of x are nonzero, then the highest bit of each byte of
+	 * x_2 is lower or equal to the corresponding bit of x.  Hence x_3 has
+	 * all these highest bits cleared (the target bit is set iff the source
+	 * bit has changed from 0 to 1).  If a == 0, then we are sure there is
+	 * no zero byte in x.
 	 */
 	if (!a)
 		return sizeof(uns);
@@ -59,7 +64,7 @@ str_len_uns(uns x)
 }
 
 inline uns
-str_len_aligned(const byte *str)
+str_len_aligned(const char *str)
 {
 	const uns *u = (const uns *) str;
 	uns len = 0;
@@ -73,7 +78,7 @@ str_len_aligned(const byte *str)
 }
 
 inline uns
-hash_string_aligned(const byte *str)
+hash_string_aligned(const char *str)
 {
 	const uns *u = (const uns *) str;
 	uns hash = 0;
@@ -107,7 +112,7 @@ hash_block_aligned(const byte *str, uns len)
 
 #ifndef	CPU_ALLOW_UNALIGNED
 uns
-str_len(const byte *str)
+str_len(const char *str)
 {
 	uns shift = UNALIGNED_PART(str, uns);
 	if (!shift)
@@ -124,11 +129,12 @@ str_len(const byte *str)
 }
 
 uns
-hash_string(const byte *str)
+hash_string(const char *str)
 {
-	uns shift = UNALIGNED_PART(str, uns);
+	const byte *s = str;
+	uns shift = UNALIGNED_PART(s, uns);
 	if (!shift)
-		return hash_string_aligned(str);
+		return hash_string_aligned(s);
 	else
 	{
 		uns hash = 0;
@@ -144,9 +150,9 @@ hash_string(const byte *str)
 #endif
 			if (!modulo)
 				hash = ROL(hash, SHIFT_BITS);
-			if (!str[i])
+			if (!s[i])
 				break;
-			hash ^= str[i] << (shift * 8);
+			hash ^= s[i] << (shift * 8);
 		}
 		return hash;
 	}
@@ -183,8 +189,9 @@ hash_block(const byte *str, uns len)
 #endif
 
 uns
-hash_string_nocase(const byte *str)
+hash_string_nocase(const char *str)
 {
+	const byte *s = str;
 	uns hash = 0;
 	uns i;
 	for (i=0; ; i++)
@@ -198,9 +205,9 @@ hash_string_nocase(const byte *str)
 #endif
 		if (!modulo)
 			hash = ROL(hash, SHIFT_BITS);
-		if (!str[i])
+		if (!s[i])
 			break;
-		hash ^= Cupcase(str[i]) << (shift * 8);
+		hash ^= Cupcase(s[i]) << (shift * 8);
 	}
 	return hash;
 }