• d-bugmail puremagic.com May 20 2008
• "Janice Caron" <caron800 googlemail.com> May 20 2008
• Ingo Oeser <ioe-news rameria.de> May 24 2008

d-bugmail puremagic.com writes:

http://d.puremagic.com/issues/show_bug.cgi?id=2117

           Summary: Please add more byteswapping support
           Product: D
           Version: 2.014
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Phobos
        AssignedTo: bugzilla digitalmars.com
        ReportedBy: jlquinn optonline.net


Right now there is limited support for byteswapping.  std.intrinsic includes
bswap(uint), and std.system includes a predefined constant for system
endianness.

It would be really helpful to have bswap equivalents for all basic datatypes,
as well as routines to swap between native endianness and explicit big or
little endian, ala ntohl, ntohs, htonl, htons

A related thing that would come in handy is byteswapping mmap support.


-- 

"Janice Caron" <caron800 googlemail.com> writes:

On 20/05/2008, d-bugmail puremagic.com <d-bugmail puremagic.com> wrote:
  It would be really helpful to have bswap equivalents for all basic datatypes,


Essentially, bswap(T)(T x)


  as well as routines to swap between native endianness and explicit big or
  little endian, ala ntohl, ntohs, htonl, htons


I never did understand the difference between ntoh* and hton*. :-)

Ingo Oeser <ioe-news rameria.de> writes:

Janice Caron wrote:
 I never did understand the difference between ntoh* and hton*. :-)


It is a semantic difference only, not a behavioral one.
You can deduce the expected and actual endianess 
when using the semantically correct transformation. 

This helps (static) checkers like e.g. sparse on Linux.


Best Regards

Ingo Oeser