Small changes (valgrind)

README.md

@@ -142,3 +142,6 @@ Please, if you use this tool in an academic setting cite the following paper:
 
 \[4\] Nong, G., Practical linear-time O(1)-workspace suffix sorting for constant alphabets, ACM Trans. Inform. Syst., vol. 31, no. 3, pp. 1–15, 2013
 
+## Thanks:
+
+Thanks to Giovanni Manzini by helpful suggestions and debugging.

experiments/Makefile

@@ -1,5 +1,5 @@
 CC = gcc
-CFLAGS += -Wall 
+CFLAGS += -Wall
 #CFLAGS += -g -O0
 CFLAGS += -D_FILE_OFFSET_BITS=64 -m64 -O3 -fomit-frame-pointer -Wno-char-subscripts 
 
@@ -59,4 +59,4 @@ run:
 	./gsais $(DIR) $(INPUT) $(K) $(MODE) $(LCP_COMPUTE) $(DA_COMPUTE) $(BWT_COMPUTE) $(CHECK) $(OUTPUT)
 
 valgrind:
-	valgrind --tool=memcheck --leak-check=full --track-origins=yes ./gsais $(DIR) $(INPUT) $(K) $(MODE) $(LCP_COMPUTE) $(DA_COMPUTE) $(BWT_COMPUTE) $(CHECK) $(OUTPUT)
+	valgrind -q --tool=memcheck --leak-check=full --track-origins=yes --show-leak-kinds=all ./gsais $(DIR) $(INPUT) $(K) $(MODE) $(LCP_COMPUTE) $(DA_COMPUTE) $(BWT_COMPUTE) $(CHECK) $(OUTPUT)

experiments/lib/utils.c

@@ -105,7 +105,7 @@ int_text* cat_int(unsigned char** R, int k, int_t *n){
 	}
 
 	str_int[l++]=0;
-        if(*n>l){
+  if(*n>l){
 		str_int = (int_text*) realloc(str_int, (l)*sizeof(int_text));
 		printf("N = %" PRIdN "\n", l);
 	}
@@ -152,8 +152,8 @@ unsigned char* cat_char(unsigned char** R, int k, int_t *n){
 	}
 
 	str[l++]=0;
-        if(*n>l){
-		str = (unsigned char*) realloc(str, (l)*sizeof(unsigned char));
+  if(*n>l){
+		str = (unsigned char*) realloc(str, (l+1)*sizeof(unsigned char));
 		printf("N = %" PRIdN "\n", l);
 	}
 	*n = l;

gsacak.c

@@ -5,10 +5,7 @@
 const uint_t EMPTY_k=((uint_t)1)<<(sizeof(uint_t)*8-1); 
 
 // get s[i] at a certain level
-
-//#if M64 && LARGE_ALPHABET==0
 #define chr(i) (cs==sizeof(int_t)?((int_t*)s)[i]:(cs==sizeof(int_text)?((int_text*)s)[i]:((unsigned char *)s)[i]))
-//define chr(i) (cs==sizeof(int_t)?((int_t*)s)[i]:((unsigned char *)s)[i])
 
 #define true 1
 #define false 0
@@ -638,7 +635,7 @@ void induceSAl0_generalized_LCP_DA(uint_t *SA, int_t *LCP, int_t *DA,
 
       if(LCP[i]==I_MIN){ //is a L/S-seam position
   	  int_t l=0;
-	  if(SA[bkt[chr(SA[i])]-1]<n-1)	
+		  if(SA[bkt[chr(SA[i])]-1]<n-1)	
    	    while(chr(SA[i]+l)==chr(SA[bkt[chr(SA[i])]-1]+l))++l;
   	  LCP[i]=l;
       }
@@ -755,44 +752,44 @@ void induceSAs0_generalized_LCP_DA(uint_t *SA, int_t* LCP, int_t* DA,
     if(SA[i]>0) {
       j=SA[i]-1;
       if(chr(j)<=chr(j+1) && bkt[chr(j)]<i)// induce S-type
-        if(chr(j)!=separator){
-	  SA[bkt[chr(j)]]=j;
-      DA[bkt[chr(j)]]=DA[i];
+      if(chr(j)!=separator){
+	    SA[bkt[chr(j)]]=j;
+        DA[bkt[chr(j)]]=DA[i];
 
-          #if RMQ_S == 1
+        #if RMQ_S == 1
   	    if(LCP[bkt[chr(j)]+1]>=0) 
   	      LCP[bkt[chr(j)]+1]=M[chr(j)]+1;
-  	  
+
 	    if(LCP[bkt[chr(j)]]>0) 
   	      LCP[bkt[chr(j)]]=I_MAX;
 
-          #elif RMQ_S == 2
-            int_t min = I_MAX, end = top-1; 
+        #elif RMQ_S == 2
+          int_t min = I_MAX, end = top-1; 
 
-  	    int_t last=last_occ[chr(j)];
-            while(STACK[end].idx<=last) end--;
+          int_t last=last_occ[chr(j)];
+          while(STACK[end].idx<=last) end--;
 
-            min=STACK[(end+1)].lcp;
-            last_occ[chr(j)] = i;
+          min=STACK[(end+1)].lcp;
+          last_occ[chr(j)] = i;
 
-  	    if(LCP[bkt[chr(j)]+1]>=0) 
-              LCP[bkt[chr(j)]+1]=min+1;
-          #endif
+  	      if(LCP[bkt[chr(j)]+1]>=0) 
+            LCP[bkt[chr(j)]+1]=min+1;
+        #endif
 
 
-          #if RMQ_S == 1
-  	  M[chr(j)] = I_MAX;
-          #endif
+        #if RMQ_S == 1
+          M[chr(j)] = I_MAX;
+        #endif
 
-          bkt[chr(j)]--;
+        bkt[chr(j)]--;
 
-  	  if(SA[bkt[chr(j)]]!=U_MAX) {//L/S-seam
-            int_t l=0;	
-            while(chr(SA[bkt[chr(j)]+1]+l)==chr(SA[bkt[chr(j)]]+l))++l;
-            LCP[bkt[chr(j)]+1]=l;
-  	  }
-  	}
-     }
+        if(SA[bkt[chr(j)]]!=U_MAX) {//L/S-seam
+              int_t l=0;	
+              while(chr(SA[bkt[chr(j)]+1]+l)==chr(SA[bkt[chr(j)]]+l))++l;
+              LCP[bkt[chr(j)]+1]=l;
+        }
+      }
+    }
 
     if(LCP[i]<0) LCP[i]=0;
     #if RMQ_S == 1
@@ -2505,8 +2502,10 @@ int sacak_int(int_text *s, uint_t *SA, uint_t n, uint_t k){
 int gsacak(unsigned char *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n){
   if((s == NULL) || (SA == NULL) || (n < 0)) return -1;
 
+  int i;
+  for(i=0; i<n; i++) SA[i]=LCP[i]=DA[i]=0;
+
   #if EMPTY_STRING
-    int i;
     for(i=0; i<n-1; i++) if(s[i]==1 && s[i+1]==1) return -2; 
   #endif  
 
@@ -2523,6 +2522,9 @@ int gsacak(unsigned char *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n){
 int gsacak_int(int_text *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n, uint_t k){
   if((s == NULL) || (SA == NULL) || (n < 0)) return -1;
 
+  int i;
+  for(i=0; i<n; i++) SA[i]=LCP[i]=DA[i]=0;
+
   if((LCP == NULL) && (DA == NULL))
 	return gSACA_K((uint_t*)s, SA, n, k, sizeof(int_text), 1, 0);
   else if (DA == NULL)

gsacak.h

@@ -51,60 +51,56 @@
 	#define I_MAX	INT64_MAX
 	#define I_MIN	INT64_MIN
 #else
-	typedef int32_t	int_t;
-	typedef uint32_t	uint_t;
+	typedef int32_t int_t;
+	typedef uint32_t uint_t;
 	#define PRIdN	PRId32
 	#define U_MAX	UINT32_MAX
 	#define I_MAX	INT32_MAX
 	#define I_MIN	INT32_MIN
 #endif
 
 /*! @option type of s[0,n-1] for integer alphabets 
- *	@constraint sizeof(int_t) >= sizeof(int_text) 
+ *
+ *  @constraint sizeof(int_t) >= sizeof(int_text) 
  */
-typedef uint32_t	int_text;	//4N bytes for s[0..n-1]
+typedef uint32_t int_text;	//4N bytes for s[0..n-1]
 #define PRIdT	PRIu32
 
 /******************************************************************************/
 
 /** @brief computes the suffix array of string s[0..n-1] 
  *
- *	@param s	input string with s[n-1]=0
- *	@param SA		suffix array 
- *	@param n	string length
- *	@return -1 if an error occured, otherwise the depth of the recursive calls.
+ *  @param s	input string with s[n-1]=0
+ *  @param SA		suffix array 
+ *  @param n	string length
+ *  @return -1 if an error occured, otherwise the depth of the recursive calls.
  */
 int sacak(unsigned char *s, uint_t *SA, uint_t n);
 
 /** @brief computes the suffix array of string s[0..n-1]
- *	@param k	alphabet size
+ *
+ *  @param k	alphabet size
  */
 int sacak_int(int_text *s, uint_t *SA, uint_t n, uint_t k);
 
 /******************************************************************************/
 
 /** @brief Computes the suffix array SA (LCP, DA) of T^cat in s[0..n-1]
  *
- *	@param s		input concatenated string, using separators s[i]=1 and with s[n-1]=0
- *	@param SA		suffix array 
- *	@param LCP	LCP array 
- *	@param DA		Document array
- *	@param n	string length
- * 
- *	@return depth of the recursive calls.
+ *  @param s		input concatenated string, using separators s[i]=1 and with s[n-1]=0
+ *  @param SA		suffix array 
+ *  @param LCP	LCP array 
+ *  @param DA		Document array
+ *  @param n	string length
+ *  
+ *  @return depth of the recursive calls.
  */
 int gsacak(unsigned char *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n);
 
 /** @brief Computes the suffix array SA (LCP, DA) of T^cat in s[0..n-1]
  *
- *	@param s		input concatenated string, using separators s[i]=1 and with s[n-1]=0
- *	@param SA		suffix array 
- *	@param LCP	LCP array 
- *	@param DA		Document array
- *	@param n	string length
- *	@param K	alphabet size
- * 
- *	@return depth of the recursive calls.
+ *  @param s    input concatenated string, with s[n-1]=0
+ *  @param K	alphabet size
  */
 int gsacak_int(int_text *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n, uint_t k);
 

gsais.c

@@ -1,8 +1,7 @@
 #include "gsais.h"
 
-//#if M64 && LARGE_ALPHABET==0
+// get s[i] at a certain level
 #define chr(i) (cs==sizeof(int_t)?((int_t*)s)[i]:(cs==sizeof(int_text)?((int_text*)s)[i]:((unsigned char *)s)[i]))
-//#define chr(i) (cs==sizeof(int_t)?((int_t*)s)[i]:((unsigned char *)s)[i])
 
 #define tget(i) ( (t[(i)/8]&mask[(i)%8]) ? 1 : 0 )
 #define tset(i, b) t[(i)/8]=(b) ? (mask[(i)%8]|t[(i)/8]) : ((~mask[(i)%8])&t[(i)/8])

gsais.h

@@ -44,26 +44,26 @@
 #endif
 
 #if M64
-	typedef int64_t	int_t;
-	typedef uint64_t	uint_t;
+	typedef int64_t int_t;
+	typedef uint64_t uint_t;
 	#define PRIdN	PRId64
 	#define U_MAX	UINT64_MAX
 	#define I_MAX	INT64_MAX
 	#define I_MIN	INT64_MIN
 #else
-	typedef int32_t	int_t;
-	typedef uint32_t	uint_t;
+	typedef int32_t int_t;
+	typedef uint32_t uint_t;
 	#define PRIdN	PRId32
 	#define U_MAX UINT32_MAX
 	#define I_MAX INT32_MAX
 	#define I_MIN INT32_MIN
 #endif
 
-
 /*! @option type of s[0,n-1] for integer alphabets 
- *	@constraint sizeof(int_t) >= sizeof(int_text) 
+ *
+ *  @constraint sizeof(int_t) >= sizeof(int_text) 
  */
-typedef uint32_t	int_text;	//4N bytes for s[0..n-1]
+typedef uint32_t int_text;	//4N bytes for s[0..n-1]
 #define PRIdT	PRIu32
 
 /******************************************************************************/
@@ -78,6 +78,7 @@ typedef uint32_t	int_text;	//4N bytes for s[0..n-1]
 int sais(unsigned char *s, uint_t *SA, uint_t n);
 
 /** @brief computes the suffix array of string s[0..n-1]
+ *
  *  @param k	alphabet size
  */
 int sais_int(int_text *s, uint_t *SA, uint_t n, uint_t k);
@@ -91,21 +92,15 @@ int sais_int(int_text *s, uint_t *SA, uint_t n, uint_t k);
  *  @param LCP 	LCP array 
  *  @param DA 	Document array
  *  @param n	string length
- * 
+ *  
  *  @return depth of the recursive calls.
  */
 int gsais(unsigned char *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n);
 
 /** @brief Computes the suffix array SA (LCP, DA) of T^cat in s[0..n-1]
  *
- *  @param s    input concatenated string, using separators s[i]=1 and with s[n-1]=0
- *  @param SA 	suffix array 
- *  @param LCP 	LCP array 
- *  @param DA 	Document array
- *  @param n	string length
+ *  @param s    input concatenated string, with s[n-1]=0
  *  @param K	alphabet size
- * 
- *  @return depth of the recursive calls.
  */
 int gsais_int(int_text *s, uint_t *SA, int_t *LCP, int_t *DA, uint_t n, uint_t k);