barrust / hashmap Goto Github PK

Add a full suite of unittests using [minunit]https://github.com/barrust/c-utils/blob/master/src/minunit.h) from the c-utils library.

Bonus: add code coverage and continuous integration

The following results in a segmentation fault:

HashMap h;
hashmap_init(&h);
hashmap_stats(&h);

The same is true after clearing the hashmap. Likely caused by malloc'ing memory based on the number of used nodes which is 0.

Look into replacing strncpy with memcpy since all lengths are predefined and are known to not overlap.

This should have an easier test suite to check that everything is working correctly

Due to this being used with other, similar, libraries the HashFunction type should be made unique.

Make it usable with C++

The repository's organization should be upated and releases should be provided.

I am running an OpenMP program with 2 threads. One thread sets some values and the other tries to read it but I don't get any set values back in the other thread

// thread 1 sets
int i = 0;
                for (i = 0; i < 10; i++)
                {
                    omp_set_lock(&writelock);
                    char key = i + '0';
                    printf("++setting key %d to val %d \n", key, i);
                    //#pragma omp critcal
                    hashmap_set_int(&h, &key, i);
                    omp_unset_lock(&writelock);
                }

               // thread 2 reading code
                sleep(5);
                int i = 0;
                for (i = 0; i < 10; i++)
                {
                    omp_set_lock(&writelock);
                    char key = i + '0';
                    int *hVal;
                    //#pragma omp critical
                    hVal = hashmap_get(&h, &key);
                    omp_unset_lock(&writelock);
                    if (hVal)
                    {
                        printf("hval is %d\n", *hVal);
                    }
                    else
                    {
                        printf("not found\n");
                    }
                }

HashMap h; is done at the top outside of main and hashmap_init(&h); is done in main

am I doing something wrong?

If the unified header build is needed, then it should be more intelligent as to how it collects includes and the likes. Otherwise, it is probably not needed.