[Pdx-pm] PowerPC and little-endian

[Austin Schutz]

On Tue, Feb 28, 2006 at 11:11:40PM -0800, jerry gay wrote:
> On 2/28/06, Eric Wilhelm <scratchcomputing at gmail.com> wrote:
> > Hi all,
> >
> > Since I'm in the minority of non-mac owners around here, maybe you all
> > can fill me in on this whole dual-mode software-switched rumor about
> > the powerpc being able to run in either big or little endian mode.  I
> > need to unpack some floats and doubles in a cross-platform way and yeah
> > I could do that in pure perl with some of the portable pack/unpack
> > commands and some byte twiddling, but that's going to be slow and so
> > I'll then end up maintaining two pieces of code, so I was thinking I
> > could compile perl to run in little-endian mode.
> >
> from `perldoc -f pack`:
> 
>     pack TEMPLATE,LIST
>     [snip]
>                 n   An unsigned short in "network" (big-endian) order.
>                 N   An unsigned long in "network" (big-endian) order.
>                 v   An unsigned short in "VAX" (little-endian) order.
>                 V   An unsigned long in "VAX" (little-endian) order.
>                       (These 'shorts' and 'longs' are _exactly_ 16 bits and
>                        _exactly_ 32 bits, respectively.)
> 
> maybe that will suffice? it will pack/unpack in the respective byte
> order on any platform, no byte twiddling needed.

	There's a big note about how they decided not to tackle floats
and doubles:

        *              Real numbers (floats and doubles) are in the native
                       machine format only; due to the multiplicity of float‐
                       ing formats around, and the lack of a standard "net‐
                       work" representation, no facility for interchange has
                       been made.  This means that packed floating point data
                       written on one machine may not be readable on another −
                       even if both use IEEE floating point arithmetic (as the
                       endian‐ness of the memory representation is not part of
                       the IEEE spec).  See also perlport.


	Personally I think that's a cop out. The java people figured it
out (java.lang.float.floatToIntBits()).
	On the other hand, what is it I hear people say when someone
complains about a lacking feature? Oh yeah.. "send code". Well, maybe I
will.

http://babbage.cs.qc.edu/courses/cs341/IEEE-754references.html

	Austin