structure members of packed structures

[Richard Henderson]

> On Tuesday, July 15, 2003, at 03:59  PM, Richard Henderson wrote:
> >Personally, I have no problem with the address being taken and giving
> >SIGBUS when used (or, gasp, fixed up by UAC fault software).

Just so no one gets the wrong idea, IMO the absolutely best 
solution is to make the unalignedness part of the type, and
to propagate that as necessary.

Failing that I think there's no good solution, only choices
ranging from bad to worse.


r~