It's not so difficult to produce two md5-collided exe files

[Git]

B&W - I don't know Wangs stuff yet, but I did some collaborative work on this a few years ago, and it most definitely was not easy to match the CRC's of two exes, where one is an arbitrarily hacked version of the other. It is possible under some fairly simple circumstances but far from easy. We did produce a tool, but it's use was so limited it was not worth doing anything with it. My dim memory reminds me it involved adding a resource to the hacked exe and then altering specific bytes in that resource to match the good exes CRC. I'll see if I still have any source or notes.

Git

[BlackWhite]

Originally Posted by Git

B&W - I don't know Wangs stuff yet, but I did some collaborative work on this a few years ago, and it most definitely was not easy to match the CRC's of two exes, where one is an arbitrarily hacked version of the other. It is possible under some fairly simple circumstances but far from easy. We did produce a tool, but it's use was so limited it was not worth doing anything with it. My dim memory reminds me it involved adding a resource to the hacked exe and then altering specific bytes in that resource to match the good exes CRC. I'll see if I still have any source or notes.

Git

Yes, to build two different EXE files with the same CRC or the same MD5
is nearly impossible.
My method is actually the combination of two exe files, the combined
exe choose to run the internal exe1 or exe2 according to the bit difference
between md5_1 & md5_2. So it's much easier.