From: Pavel Charvat <pavel.charvat@netcentrum.cz>
Date: Wed, 5 Apr 2006 08:49:33 +0000 (+0200)
Subject: New version of SubStrings analyser...
X-Git-Tag: holmes-import~650^2~14
X-Git-Url: http://mj.ucw.cz/gitweb/?a=commitdiff_plain;h=8106838654534920f5e919604a437a990038c860;p=libucw.git

New version of SubStrings analyser...
- KMP uses continuous read of fastbuf streams
- the tag is OR combination of groups masks
---

diff --git a/lib/kmp.h b/lib/kmp.h
index 9577e8d3..fd37a6e3 100644
--- a/lib/kmp.h
+++ b/lib/kmp.h
@@ -161,7 +161,8 @@ kmp_search_internal(struct kmp *kmp, byte *str, uns len, struct list *nonzeroes,
   /* For every found string with id ID, it increments freq[ID].
    * Also, it finds the longest among the leftmost matches.  */
 {
-	byte *str_end = str + len;
+  	if (!len)
+	  return NULL;
 	kmp_state_t s = 0;
 	kmp_char_t c = KMP_CONTROL_CHAR;
 	struct kmp_transition tr, **prev;
@@ -195,7 +196,7 @@ kmp_search_internal(struct kmp *kmp, byte *str, uns len, struct list *nonzeroes,
 		}
 		if (eof)
 			break;
-		if (str >= str_end)
+		if (!--len)
 			c = 0;
 		else
 			kmp_get_char((const byte **)&str, &c, kmp->modify_flags);
@@ -210,9 +211,9 @@ kmp_search_internal(struct kmp *kmp, byte *str, uns len, struct list *nonzeroes,
 }
 
 static inline void
-kmp_search(struct kmp *kmp, const byte *str, struct list *nonzeroes, struct kmp_result *freq)
+kmp_search(struct kmp *kmp, const byte *str, uns len, struct list *nonzeroes, struct kmp_result *freq)
 {
-	kmp_search_internal(kmp, (byte*) str, strlen(str), nonzeroes, freq, NULL);
+	kmp_search_internal(kmp, (byte*) str, len, nonzeroes, freq, NULL);
 }
 
 static inline byte *