My assumption is the original author was writing the timestamp with the wrong endianness in the first place, which is easily done when not using bitstreams (see http://www.jenkinssoftware.com/raknet/manual/creatingpackets.html). The original patch causes the endianness to be inverted twice when using bitstreams, breaking reception of these timestamps.
Reverting https://github.com/larku/RakNet/pull/15
My assumption is the original author was writing the timestamp with the wrong endianness in the first place, which is easily done when not using bitstreams (see http://www.jenkinssoftware.com/raknet/manual/creatingpackets.html). The original patch causes the endianness to be inverted twice when using bitstreams, breaking reception of these timestamps.