Subject: Re: [boost] Any interest in bitstream class?
From: Paul Long (plong_at_[hidden])
Date: 2013-06-28 18:50:07
On 6/28/2013 3:44 PM, Klaim - Joël Lamotte wrote:
> I meant platform A uses one endianness, platform B uses another, but if i
> put data from either into the bitstream, the bitstream will use one
> specific endianness whatever the platform.
> Which is basically what you describe (the network packet case indeed) if I
> understood correctly.
Correct, but it is important to note that integrals are always encoded
big endian in the bitstream. I suppose it's possible that someone would
want some other endian, but that's got to be an obscure use case.
Actually, I can think of an example. Years ago, I worked on a WinTel,
little-endian product, and a colleague didn't use any serialization for
passing objects between two network components. The thinking was, pshaw,
it'll always be a Windows product, so why bother with serialization.
Then, of course, Apple paid us to port it to their 68000, big-endian
platform while maintaining interoperability. In a use case like this, it
might be useful to support other-than-big endianness. What do you think?
Should the bitstream library support multiple endian schemes in the bit
stream or is big endian enough?
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk