Re: Make Instruction Emulation Code More Tolerant of Bad Addresses

[Alexandre Julliard]

Robert Shearman  writes:
> Therefore this patch aids in debugging. It isn't perfect because of
> x86's variable length instructions, but it is fairly unlikely that an
> instruction will straddle a page boundary and the first page be
> readable and the second not.

Putting an exception handler around the whole function would probably
be a better choice, this way it would catch data accesses too.