For mdx format, Adler-32 is computed twice for zlib decompression

[shenlebantongying]

Code here can be optimized:

goldendict-ng/src/dict/mdictparser.cc

Lines 250 to 257 in 2b98b50

	// zlib compression
	decompressedBlock = zlibDecompress( buf, size );
	if ( !checkAdler32( decompressedBlock.constData(), decompressedBlock.size(), checksum ) ) {
	gdWarning( "MDict: parseCompressedBlock: zlib: checksum does not match" );
	return false;
	}
	break;

The z_stream in zlibDecompress has a field adler

goldendict-ng/src/decompress.cc

Line 10 in 2b98b50

z_stream zs;

See zlib's doc's inflate

otherwise it sets strm->adler to the Adler-32 checksum of all output produced so far (that is, total_out bytes)

In other words, zlib's inflate will be rolling a alder for final checksum and there is no need to go over the decompressed data to compute it again.