On Thu, 4 May 2000, Richard Earnshaw wrote:

> Well the biggest culprit used to be the ethernet header code (a struct 
> containing 6 octets that effectively assumed that sizeof would return 6).  
> But that was fixed in February last year by the looks of things (by adding 
> the "packed" attribute).  So maybe the hard work has already been done for 
> the common cases: the only way to find out is to try it :-)

OK.  I'll do a test build without the structure padding hack and see how
it turns out.  It may take till the weekend for me to get around to it,
though.

-- 
Ben Harris