Re: reverse engineer c or java

[crispin () WIREX COM (Crispin Cowan)]

Jeff Bachtel wrote:

> The problem is that with optimizing compilers, a given output of the
> compiler has an infinite (or thereabouts ;) number of possible source
> programs.

It is truly infinite.  The number of source programs is isomorphic to
the
number of rational numbers (insert Math undergrad degree here :-)

> The fact is, that a decompiler can produce perfectly valid C code, that
> makes no sense to normal humans.

Agreed.

> This does not take into account self-modifying code, which would
> require a decompiler coupled with a simulation engine, and logic to
> detect flow between possible states of the program, and the
> algorythmns used.

Which is the basis for Turing's Halting problem (insert grad course in
computability theory here :-)

The decompiler would not notice; it would produce a C program that
computes
pointers that point into the data space, writes to them, and then
assigns
(*foo)() = (void *) and jumps to foo().  You're right; it won't make
sense
to humans, unless the humans can understand the byte sequence written to
the pinter.

Crispin