Re: In search of the perfect Disassembler

I did some research on the web about this Betov and now I think I understand what we are facing here...


For the longest time, Rene was claiming that 100% automatic (and
perfect) disassembly was possible and RosAsm was going to do that.
This, despite the fact that several people around here pointed out that
disassembly reduces to the halting problem (which is unsolvable).


Well, maybe he has solved the halting problem!
Didn't you consider this possibility?
Betov kicks Gödel's ass badly!


And even if that's all they *did* care about, the novelty of "two
clicks disassembly/assembly" wears off real quick when you realize that
it isn't practical at all. There will *always* be mistakes in the
produced code if you're disassembling a non-trivial application. If
automatic disassembly *were* possible, programs like IDAPro (which have
been under serious development for many, many years) would have cracked
that nut by now. But as has been pointed out many times around here, it
is *not* possible to do an automatic disassembly.  An automatic
disassembly makes a good "first pass" for the code, to give you
something to work with in interactive mode (and IDAPro does a decent
job of this), but it will *never* be sufficient by itself except for
simple programs (like the ones Rene uses to demo how great his
disassembler is).


Yes I know about this but I often regretted than the wonderful IDA wasn't at least even trying to dump a MASM-assemblable with binary sections inside, because it would have worked for me in some occurences (of course a bit of TXL or a plugin can do the trick)


Also note: when Rene announces that his disassembler is complete, I'll
be able to break it in about two minutes. It's not hard to confuse an
automatic disassembler.


It's not hard to confuse a programmer either :)


Actually, a discussion of PS2 VU instruction scheduling and MIPS 5900
cache prefetching is a totally appropriate subject to discuss around
here (and many people would find it interesting). Far more appropriate
than most of Rene's (Betov's) points.


Yuck, are you an insane pervert, Randy? I only do this stuff to be paid!
(and I'm not doing it anymore :) )

Cheers,
Randy Hyde


Cheers,

Bertrand
.

Relevant Pages

  • Re: Interesting Correspondence with the Manager (Pierre Vandevenne) of www.datarescue.com
    ... > You might want to take a look at RosAsm Disassembler- ... > with what commercial Apps can do, but the less i can say, ... it is also evident that an Automatic DisAssembler ...
    (alt.lang.asm)
  • Re: RosAsm disassembler is still not working
    ... Given the RosAsm development team's track record, ... the RosAsm disassembler in order to handle those cases, ... I've not had much of a problem recompiling IDAPro output files. ... assemblers who have a fair library of routines will stick with those ...
    (alt.lang.asm)
  • Re: Need reviews of HLA Adventure
    ... So you really don't know what IDAPro does, ... how much better your disassembler is than IDAPro. ... > 2) I base my words on the ones the main Author of IDA ... > better than RosAsm Disassembler, ...
    (alt.lang.asm)
  • Renes Revised History of Assembly Language
    ... remember back when you were promising an automatic disassembler in just ... >> automatic disassembler isn't going to cut it and you're going to have ... RosAsm does not even try to disassemble Z80 code. ... You don't really need the ability to reassemble the source ...
    (alt.lang.asm)
  • Re: Disassembler
    ... > Is there an equally capable disassembler for NASM under Linux? ... There is IDAPro, which is actually a bit better than the "under development" ... RosAsm disassembler at this point. ...
    (alt.lang.asm)