XPost: comp.theory, sci.logic, sci.math
From: 643-408-1753@kylheku.com
On 2025-11-27, Richard Damon wrote:
> On 11/27/25 10:21 AM, olcott wrote:
>> The DD that halts is not the same DD that is
>> an input to HHH(DD). The input to HHH(DD)
>> specifies non-halting behavior that HHH
>> recognizes and terminates.
>
> Then somebody is lying.
Yes; Olcott lying to himself through code.
HHH makes this function call:
u32 Root = Init_Halts_HH(&Aborted, &execution_trace, &decoded, &code_end,
(u32)P,
&master_state, &slave_state, &slave_stack);
execution_trace is pointlessly doubly indirected; it holds a pointer
aimed at a static area right inside the coe of HHH. The above function
deos this:
if (**execution_trace == 0x90909090)
{
**Aborted = 0;
**execution_trace = (u32)Allocate(sizeof(Decoded_Line_Of_Code) * 10000);
Output((char*)"\nBegin Local Halt Decider Simulation "
"Execution Trace Stored at:", **execution_trace);
return 1;
}
return 0;
(You can see here that the argument should just be execution_trace,
and the allocation *execution_trace = ...; Ths execution_trace points
into the code of HHH.)
So the first call to HHH allocates the execution trace and changes
the static value in the code space from 0x90909090 to a different value.
In ths case 1 is returned. Otherwise nothing is allocated and 0 is
returned.
This return value is captured as Root and changes the behavior of the
first HHH instance compared to the others.
The first HHH instance performs the abort check. The other instances
are free running simulation steppers.
This changes DD from non-terminating to terminating. When HHH changes
its behavior form aborting and returning 0, to indefinitely simulating,
then DD changes its behavior from terminating to nonterminating.
> As DD is supposed to call HHH with a representation of itself.
Unfortunately, DD's representation is tied up that of HHH,
and HHH is self-modifying code.
Needless to say, you cannot do that; you can't use self-modifying code
to change a terminating test case to non-terminating and then claim that
0 is the correct return value.
--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
Mastodon: @Kazinator@mstdn.ca
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
