home *** CD-ROM | disk | FTP | other *** search
- From: markh@csd4.csd.uwm.edu (Mark William Hopkins)
- Newsgroups: alt.sources
- Subject: Non-trivial self-reproducing programs
- Message-ID: <10458@uwm.edu>
- Date: 23 Mar 91 06:39:58 GMT
-
- In article <1991Mar19.165839.15308@isc.rit.edu> rxg3321@isc.rit.edu (R.X. Getter ) writes:
- >hi,
- > I am looking for a c program which outputs itself. Has anyone
- >seen one or heard of one? (No, it cannot read its own source code.)
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
- But can it decompile its own object code?
-
- Generally the challenge is to take any non-trivial C program generator (meaning
- a decompiler) and make it self-reproducing.
-
- Of course, since this (otherwise) general-purpose program will be formatting
- its generated source code in a fixed way and naming its variables in a fixed
- way, it should be written conforming to those restrictions if it is going to
- ever be able to reproduce a copy of itself from its own object code.
-
- Equally obvious is that you can write it any way you want (and in any
- language), compile it into object code and then run the object code on itself.
- The generated source code should then be syntatically self-reproducing
- (assuming the original was semantically self-reproducing) when re-compiled and
- run on its own object code. If not, you can repeat the process until it
- converges to a fixed point...
-
- And most obvious of all is that you can start with ANY decompiler and do this
- provided you can compile the decompiler, so it can decompile its object code...
-