pset5 (check) - can't pass word to hash function - error with var type

Question

I'm really struggling with this one! I've just about got through to the last section I have to complete and I have two issues that I know of (I can't get 'check' to work so I don't know if 'load' is working properly really - even though 'size' is working fine so I'm hoping!) The first problem is that I couldn't pass the char 'word' provided by speller into my hash function because it is a const char * and my function needs a char * (my error said "passing const char * to parameter of type 'char *' discards qualifiers")?

So, I just duplicated the hash function (djb) & changed the value to a const char * thinking that would work - now I'm getting a segfault... So how do I pass word into my hash function to get the array location? The walkthrough didn't mention this issue!!

I'm also getting valgrind errors saying I have an uninitialised value at line 174...I've tried to make sure 'value' is initialised even though I don't know if that's what was needed but it still says the same - I can't see what else I can do with 'hashtable' so I'm stuck!!!

I would really appreciate any pointers....(LOL!) thank you!! :-)

The code is here: http://pastebin.com/Rm5XTAvL

Answer 1

The compiler wanting const char* might indicate you tried to test the hash function using a string constant.

Not sure what your test for spaces is meant to do, there should not be spaces in a word. Maybe strlen is a way to go?

A hash table does not have a head. The head of the linked list is already stored in hashtable[loc % len] (the %len is for mapping the integer to range 0..len-1, you used it in load, but not in check).

load should return false on failing to load the dict (e.g. file could not be opened), and true when being successful. You return 2 at some place.

You might not want to pass LENGTH to the hash function. Maybe use strlen(word) instead.

Inside the load loop, you create two nodes (two nodes for one word?). You then either put both as a new linked list in the bin (again, two nodes for one word?), or create a linked list the other way around and forget about both nodes (that never happens, as the condition is always false), or forget both nodes in memory leak space without doing anything.

There is no use for a head node, and the whole thing

//check if the index given points to an empty 'bucket' and if so set a head pointer at hashtable[value] pointing to the first node containing the word.
if (hashtable[value] == NULL)
    {
    hashtable[value] = head;
    head->next = new_node;
    }
//or add the node containing the word as the first word after the existing head node
else if  
    (hashtable[value] == head)    
        {
        new_node->next = head;
        head = new_node;
        }

could be just

    new_node->next = hashtable[value];
    hashtable[value] = new_node;

Also, I don't get your final if (new_node == NULL).

In unload, you should use len, not sizeof(hashtable), the latter is the former, but times a pointer's length in bytes. Also, there is no last node in the list to be freed. Remember, you loop until cursor == NULL, so you then call free(NULL);.