no need to realloc buffer, since we've already created a MAX_SNAPLEN sized
buffer and copied the packet over. fixes #134
Note: still not really sure why this dies, I think in some situations,
realloc() is free()'ing the original ptr and creating a new one, but that
isn't being tracked (not updating *packet in tcpedit_packet()) and so I try to
realloc() that original ptr again and boom. Seems to only happen on some
versions of linux (like RH/CentOS) and not OS X, Unbuntu, etc- I guess
something to do with how glibc manages memory blocks.