On Sun, Oct 30, 2005 at 02:30:14PM -0800, Chuck Silvers wrote:
> oh... looking further I see that someone added an optional siginfo_t to the
> KTR_PSIG data, which is of course completely hardware-specific.  sigh.
> that does make this a bit more complicated.  ideally, these SIG records
> would contain both something to indicate the format of the siginfo_t
> (ie. some unique ID for the MACHINE_ARCH) and a length so that it can
> be skipped over even if the ID isn't know.

er, nevermind.  siginfo_t is constant-size, I was thinking of ucontext_t,
which came in around the same time and I mentally associate with the
siginfo stuff for some reason.

-Chuck