/* It reads just enough bytes from the input stream that are needed to decode the next Huffman code (and absolutely no more). It works by trying to fully decode a */
/* Huffman code by using whatever bits are currently present in the bit buffer. If this fails, it reads another byte, and tries again until it succeeds or until the */
/* The slow path is only executed at the very end of the input buffer. */
/* v1.16: The original macro handled the case at the very end of the passed-in input buffer, but we also need to handle the case where the user passes in 1+zillion bytes */
/* following the deflate data and our non-conservative read-ahead path won't kick in here on this code. This is much trickier. */
/* Ensure the output buffer's size is a power of 2, unless the output buffer is large enough to hold the entire output file (in which case it doesn't matter). */