> The paper on braided CRCs is in the zlib distribution: doc/crc-doc.1.0.pdf   Thanks for a tip, I read that and now it's bit clearer to me.    Maybe it is slightly faster, but I wouldn't bet dollars against nuts it's exact in my case as 7zip-crc also uses slice-by-8 algorithm and is actually faster.  Despite that the question is how do I use zlib crc32()? It doesn't give me correct result.   My 'rig' is this:   ~~~  crc = -1  while (buffer, length = read_data()):      {      crc = crcfunc(crc, buffer, length)      }  crc = ~crc  ~~~   This doesn't work with `crc32_z();'