Unsigned and signed integers are represented as they are represented in the core X protocol. Single and double precision floating point numbers are sent and received in IEEE floating point format. The X byte stream and network specifications make it impossible for the client to assure that double precision floating point numbers will be naturally aligned within the transport buffers of the server. For those architectures that require it, the server or client must copy those floating point numbers to a properly aligned buffer before using them.
Byte swapping on the encapsulated OpenGL byte stream is performed by the server using the same rule as the core X protocol. Single precision floating point values are swapped in the same way that 32-bit integers are swapped. Double precision floating point values are potentially swapped across all 8 bytes.