[rescue] Aviator

[Mouse]

> I ftp'd a tar file of this to one of our colleagues but haven't heard
> what his experience was.B  I think he was trying to gin up a system
> old enough to run it.

That was me.  My apologies for not reporting my results, as meagre as
they are.

I tried to run it under NetBSD's SunOS compatability mode.  It started
and promptly crashed with a memory error; the machine with the data on
it is turned off right now and I'm across the city from it (at work),
so I can't give exact details, but from memory gdb told me the
instruction stream looked something like

	ld	[%l1], %o0
	ld	[%l1+1], %o1

which cannot possibly ever work unless kernel alignment fixups are on.
So either (a) it's expecting kernel alignment fixups but not requesting
them in any way NetBSD recognizes, (b) the code is somehow corrupted,
(c) this is code that shouldn't be executed at all but somehow is, (d)
I misread or am misremembering, or ... well, I haven't come up with any
other possibilities yet.

I'm fairly sure I have a SunOS 4.x distribution CD somewhere, but after
having moved repeatedly over the last few years, I no longer know where
it is to any useful degree of precision.  I've been considering
possibilities like arranging for kernel alignment fixups, or running it
under emulation and getting an instruction-level trace which I can
examine to see if it makes sense (of course, that may well produce too
voluminous a trace to be useful - it crashed fast on human timescales,
but that could well be many millions of instructions).

Between work and family issues, I haven't had the leisure yet to dig
into this as far as I'd like.  It's well up there on my list, though,
and I would love to hear from anyone else who does anything with this
stuff.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse at rodents-montreal.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B