Eric Blake wrote:

> Jim Meyering <jim <at> meyering.net> writes:
>
>> >    if (new_table == NULL)
>> >      return false;
>> > +  if (new_table->n_buckets == table->n_buckets)
>> > +    return true;
>
> Aargh.  Ten minutes after I push, I finally see my memory leak.
>
> Obviously, new_table needs to be freed before returning true.

Hah!  I should have looked at more than the patch.

Is that code path already exercised by test-hash.c?
If so, I would have seen it eventually, via valgrind.


Reply via email to