C Segmentation Fault (core dump) with linked lists

Question

I am using a linked list which i named word_entry. The struct for the word_entry is the following:

struct word_entry

{
        char *unique_word ;
        int word_count ;
        struct word_entry *next;
        struct word_entry *prev;
} ;

And this is how I am using the code:

struct word_entry *head = NULL;
struct word_entry *curr = NULL;
int read_file( char *file_name )
{
        int initialized = 0;//was head initialized?
        char ch;
        FILE *file;
        struct word_entry *we = malloc(sizeof(struct word_entry));
        we->unique_word = malloc(sizeof(char));
        we->prev = NULL;
        we->next = NULL;
        char *s1 = malloc(sizeof(char));
        char *s2 = s1;
        file = fopen(file_name, "r");//opens the file

        if(!file){
                return 0;//file not opened
        }else{

                while((ch = fgetc(file))!= EOF){
                        if(ch >= 'A' && ch <= 'z'){
                                *s2++ = ch;
                                continue;
                        }
                        if(strlen(s1)>0){
                                if(!initialized){
                                        head = malloc(sizeof(struct word_entry));
                                        curr = head;
                                        initialized = 1;
                                }
                                *s2 = '\0';
                                strcpy(we->unique_word,s1);
                                we->word_count = 1;
                                curr = we;
                                printf("%s\n", head->unique_word);
                                s1 = malloc(sizeof(char));
                                s2 = s1;
                        }
                }
                return 1;
        }
        return 0; //file not opened
}

I am not sure why curr = head does not modify the head when I modify curr. I would like help on figuring out to how implement my code so that whenever i modify curr it also modifies the information in head.

Answer 1

You allocate memory to these char *'s-

       we->unique_word = malloc(sizeof(char));        // sizeof(char) is 1 
       ....
       char *s1 = malloc(sizeof(char));               

Even you read one character, there is no space for '\0' and when you pass them to strlen, strcpy or print them with %s causes undefined behaviour.

You need to allocate more memory.

Note - You allocate memory in every iteration but never free any of these. This causes memory leak .

I am not sure why curr = head does not modify the head when I modify curr

That is because curr points to head . If you do curr=we , curr stops pointing to head and points to we. This won't change head.

Answer 2

In your example, you use malloc with a sizeof(char) to allocate a char pointer (I assume?).

This is unneccessary, you don't need to allocate a char *, you need to initalize a block of memory this char pointer will point too.

What you're doing here is trying to write a block of memory into a single byte

strcpy(we->unique_word,s1);

unique_word is a pointer to a block of memory allocated of 1 byte size - you're writing out of the bounds you've allocated.

a more correct way to do it would be:

 we->unique_word = (char *)malloc(strlen(s1) + 1); // For the null terminator

 if (!we->unique_word) { /* Handle error where memory was not allocated */ }

 strcpy(we->unique_word,s1);