Re: [Uncc-dev] Re: playing with uncc

[Stefan Weyergraf]

Hi,
I'm one of the authors of the belowmentioned program HTE. I'm writing here 
because you cc'ed me.

> > Another good idea is integrate uncc concept in HTE editor () because, in
> > my mind it is not really useful to uncc a full software
>
>  indeed
>  in the uncc we have in mind, a user will be able to decompile a
> single function.
Is it uncc's exact goal to decompile single functions ?

> >but it can be
> > really useful to work on specific functions found/extracted with
> >HTE.
>
>  atm all that i can say, uncc is gpl, if you find a way to include it
> in hte i am pleased. but for a while me and megabug will code it as a
> standalone software
We are currently rewriting (!) the whole program (ie. HT Editor) and next is 
the machine-analyser engine. The released/old versions of the program include 
support for finding branches and subroutines, data accesses, entrypoints 
mainly.
The rewritten program shall also include argument recognition and type 
propagation for subroutine arguments using some database for common system 
libraries (like IDA does).

I see the following problems integration uncc with HT:
1) your code is C, ours is C++
2) your code is for 386 targets only, ours is (more or less) generic
3) we want the user to have *complete* control of the target program 
(low-level). decompilation (when replacing disassembly) takes this away.

It probably doesn't make sense to combine the two programs at a source code 
level. The rewritten HT however will feature virtual terminals (VT100) within 
it's own windowing system as well as spawning of subshells.
Someone (eg. me) could write code to invoke uncc in a subshell and display 
it's output in a window.
This would be a very useful application.

Cheers,
Stefan Weyergraf

PS: How much decompilation do you intend ? Have you had a look at 
boomerang.sf.net ?
Or at Cifuentes' 
http://www.itee.uq.edu.au/~cristina/dcc/decompilation_thesis.ps.gz ?