7 Feb
1999
7 Feb
'99
11:44 p.m.
On 8 Feb 99 at 0:36, Ethan Dicks wrote:
Ahh, but neither could compare to the obtruseness and expressiveness of
the ST bomb images, up to 13 of them, each number of them having multiple fault
reasons. Or the fault #s you would have to translate to another # to look it
up in the fault chart.
ciao
lwalker(a)interlog.com
> Surly the AmigaDOS Guru meditation was both
earlier and more obscure?
Earlier? Probably. The Guru Meditation number dated from 1985 (at least
for those of us outside of the original program development). Obscure?
I think not. You got two numbers: a 32-bit error code, most of which were
expressions of MC68000 exception numbers (like odd address trap or bus
error) and a 32-bit process address of the offending process. If you cared
to, you could enter a ROM-based debugger (via the serial port at 9600 baud)
and walk through memory to see why things went astray. Most people just swore
and clicked the left button.
Later, there were some great exception handlers that would translate the
32-bit error code into something meaningful inside the flashing red box.
At first, we just looked up the number in the Rom Kernel Manuals (RKMs).
Anyone who programmed an Amiga in the first five years got used to the
most common 80% of the Guru numbers anyway. Usually it was something
stupid like bolixing up an address register or accessing non-existent
memory (except AmigaDOS 1.0 and 1.1 when a malloc failure Guru'ed the machine).
-ethan
(Amiga owner since 1986)