Search code examples
c++archivewinrarself-extracting

How can I identify a WinRAR SFX using a binary signature of some sort?


I'm looking for an answer similar to the one in this question, but for WinRAR instead of 7zip. Essentially I want an authoritative signature that I can say at least "All WinRAR generated self-extracting executables created with the Default SFX file in version 420 will have bytes 0x15, 0xa1, 0x45, 0xcc, 0x21, 0x98 at location 0x00027400, and other NON SFX files are unlikely to have this signature". Even better, if the same signature can be found in all versions of the WinRAR SFX files, even if they are in different locations.

By nature of how SFX's work, they will be very similar, except for the archive portion of the file, but some strings make poor identifiers (e.x., "This program cannot be run in DOS mode" is in every SFX, but it happens to also be in most other Windows executables)

Currently my method of determining a signature is to look at various versions of the WinRAR SFX, and finding sequences of 4 or 6 bytes that all files have in common. Unfortunately, there are a lot of these, making it difficult to pick one.


Solution

  • From TechNote.txt in the WinRAR installation folder:

    1. To process an SFX archive you need to skip the SFX module searching for the marker block in the archive. There is no marker block sequence (0x52 0x61 0x72 0x21 0x1a 0x07 0x00) in the SFX module itself.

    This you already indicated yourself: the unrar source code has in archive.cpp the code how it reads in a file. The maximum size of the SFX part will never be larger than the number defined in rardefs.hpp:

    #define  MAXSFXSIZE        0x100000