Indexer stores data as big endian

[AHSauge]

The indexer code currently uses QDataStream to read and write data. This defaults to big endian byte order. While this is great for compatibility and data exchange, it also adds overhead for any little endian machine (x86, ARM etc.). That overhead seems pointless for files that are exclusively local. I therefore suggest to use native byte order and speed up the indexer about 5% as a result.