RE: Byte-swapping at load time

["Henderson, Michael D" <michael.d.henderson@...>]

Adam D. Moss wrote:
>Luiz Henrique de Figueiredo wrote:
>> We have moved towards the simplest dump/undump pair, with the 
>> intention of providing an external tool that could rewrite
>> precompiled files, which would include byte swapping and more.
>
> That is, an application on platform B simply wants to load scripts
> compiled on platform A.  In this case there's little benefit (in
> fact it's pretty inconvenient) to have a tool that converts from
> one format to another - the app really just wants to load them,
> ideally (in the case of big data-scripts) in one pass.

I'm sorry if this takes the conversation off topic, but I'm creating a
server application that farms out processing requests to agents running
on multiple machines by sending compiled chunks. Currently it's only
sending data between Linux x86 boxes, but if I want to use agents
running on a Sparc or somesuch, my server application has to know the
byte ordering of the machine that the agent is running on so as to pass
the chunk in the correct format? Would there be difference between 32
bit and 64 bit architectures, too?

Thanks,
Mike