在C中访问结构成员时出现段错误

Question

I'm trying to implement a trie for storing words in C, but I'm getting a segmentation fault when trying to acess a struct member.

The code is below:

#include <stdbool.h>
#include <ctype.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>

#define ALPHABET_SIZE 27
#define SIZE 45

//Trie data structure declaration
typedef struct _dictionary {
    bool is_word;
    char letter;
    struct _dictionary *children[ALPHABET_SIZE];
} dicto;

dicto *DICT;

//Function prototypes
void insert(char *string);
int toIndex(char s);

int main() {
    FILE *fp = fopen("small", "r");
    if (fp == NULL) {
        printf("Could not open file\n");
        return 1;
    }

    char word[46];

    while (fgets(word, sizeof(word), fp)) {
        insert(word);
        if (feof(fp)) {
            return 0;
        }
    }
    return 2;
}

//Inserts word into trie
void insert(char *string) {
    dicto *trav; //Pointer to walk through the trie
    trav = DICT;
    for (int n = 0; n = strlen(string); n++) {
        if (trav->children[toIndex(string[n])] == NULL) {
            trav->children[toIndex(string[n])] = malloc(sizeof(DICT));
            trav->letter = string[n];
            trav = trav->children[toIndex(string[n])];
        } else {
            trav->letter = string[n];
            trav = trav->children[toIndex(string[n])];
        }

        if (trav->letter == '\0') {
            trav->is_word = true;
        }
    }
    return;
}

/**
 * Output alphabetic index from given input
*/
int toIndex(char s) {
    s = toupper(s);
    int index = s - 65;
    return index;
}

I've tried debugging it with Valgrind and GDB . The output from Valgrind is:

==1979== Memcheck, a memory error detector
==1979== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
==1979== Using Valgrind-3.11.0 and LibVEX; rerun with -h for copyright info
==1979== Command: ./test_function1
==1979== 
==1979== Invalid read of size 4
==1979==    at 0x8048684: insert (in /home/test_function1)
==1979==    by 0x80485F7: main (in /home/test_function1)
==1979==  Address 0xffffff00 is not stack'd, malloc'd or (recently) free'd
==1979== 
==1979== 
==1979== Process terminating with default action of signal 11 (SIGSEGV)
==1979==  Access not within mapped region at address 0xFFFFFF00
==1979==    at 0x8048684: insert (in /home/test_function1)
==1979==    by 0x80485F7: main (in /home/test_function1)
==1979==  If you believe this happened as a result of a stack
==1979==  overflow in your program's main thread (unlikely but
==1979==  possible), you can try to increase the size of the
==1979==  main thread stack using the --main-stacksize= flag.
==1979==  The main thread stack size used in this run was 8388608.
==1979== 
==1979== HEAP SUMMARY:
==1979==     in use at exit: 344 bytes in 1 blocks
==1979==   total heap usage: 2 allocs, 1 frees, 4,440 bytes allocated
==1979== 
==1979== LEAK SUMMARY:
==1979==    definitely lost: 0 bytes in 0 blocks
==1979==    indirectly lost: 0 bytes in 0 blocks
==1979==      possibly lost: 0 bytes in 0 blocks
==1979==    still reachable: 344 bytes in 1 blocks
==1979==         suppressed: 0 bytes in 0 blocks
==1979== Reachable blocks (those to which a pointer was found) are not shown.
==1979== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==1979== 
==1979== For counts of detected and suppressed errors, rerun with: -v
==1979== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
Segmentation fault (core dumped)

And by running GDB, looks like the error comes from line 54:

if (trav->children[toIndex(string[n])] == NULL)

No idea on what might be happening.

Answer 1

This is just a quick answer regarding one of the possible issues with the code in the question. I didn't read through the whole thing.

After the following allocation, the memory is full of junk data:

trav->children[toIndex(string[n])] = malloc(sizeof(dicto));

You would be better off either using calloc (which guarantees the memory to be zeroed out):

trav->children[toIndex(string[n])] = calloc(sizeof(dicto), 1);

Or zero out the data yourself:

trav->children[toIndex(string[n])] = malloc(sizeof(dicto));
 memset(trav->children[toIndex(string[n])], 0, sizeof(dicto));

If you keep the junk data in the memory, than the following condition might be false even when it should be true:

if(trav->children[toIndex(string[n])] == NULL)

PS

Also, sizeof(DICT) is the size of the pointer , NOT the structure . You might consider sizeof(*DICT) or sizeof(dicto) .

Answer 2

There are multiple problems in your code:

• testing feof(fp) does not do what you think, it is actually unnecessary as fgets() will return NULL at end of file.

• the loop for (int n = 0; n = strlen(string); n++) never ends as n is recomputed as the length of the string at each iteration, Use this instead:

   for (int n = 0, len = strlen(string); n < len; n++) { 

• when you allocate a new node, you must initialize the structure, otherwise you may have undefined behavior as the memory block returned by malloc() is uninitialized. Use calloc() instead.

• the toIndex() function does not necessarily return a value in the range 0 to 26 . You should not hard-code the value of 'A' and you should test is the character is indeed a letter.

Here is a modified version:

#include <stdbool.h>
#include <ctype.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>

#define ALPHABET_SIZE 27
#define SIZE 45

//Trie data structure declaration
typedef struct _dictionary {
    bool is_word;
    char letter;
    struct _dictionary *children[ALPHABET_SIZE];
} dicto;

dicto *DICT;

//Function prototypes
void insert(char *string);
int toIndex(char s);

int main(void) {
    char word[SIZE + 1];
    FILE *fp = fopen("small", "r");
    if (fp == NULL) {
        printf("Could not open file\n");
        return 1;
    }

    while (fgets(word, sizeof(word), fp)) {
        insert(word);
    }
    return 0;
}

//Inserts word into trie
void insert(char *string) {
    dicto *trav = DICT; //Pointer to walk through the trie

    for (int n = 0, len = strlen(string); n < len; n++) {
        int index = toIndex(string[n]);
        if (trav->children[index] == NULL) {
            trav->children[index] = malloc(sizeof(DICT));
        }
        trav->letter = string[n];
        trav = trav->children[index];
    }
    trav->is_word = true;
}

/**
 * Output alphabetic index from given input (assuming ASCII)
 */
int toIndex(char c) {
    if (c >= 'a' && c <= 'z')
        return c - 'a';
    if (c >= 'A' && c <= 'Z')
        return c - 'A';
    return 26;  /* not a letter */
}