You've already forked linux-packaging-mono
Rewrite with hard-coded offsets into the PE file format to discern if a binary is PE32 or PE32+, and then to determine if it contains a "CLR Data Directory" entry that looks valid. Tested with PE32 and PE32+ compiled Mono binaries, PE32 and PE32+ native binaries, and a random assortment of garbage files. Former-commit-id: 9e7ac86ec84f653a2f79b87183efd5b0ebda001b
OID information is generated via a series of perl scripts. In order, the full list of commands to run are: perl objects.pl objects.txt obj_mac.num ../../include/openssl/nid.h perl obj_dat.pl ../../include/openssl/nid.h obj_dat.h perl obj_xref.pl obj_mac.num obj_xref.txt > obj_xref.h objects.txt contains the list of all built-in OIDs. It is processed by objects.pl to output obj_mac.num and nid.h. obj_mac.num is the list of NID values for each OID. This is an input/output parameter so NID values are stable across regenerations. nid.h is the header which defines macros for all the built-in OIDs in C. nid.h is read by obj_dat.pl to generate obj_dat.h. obj_dat.h contains the ASN1_OBJECTs corresponding to built-in OIDs themselves along with lookup tables for search by short name, OID, etc. obj_mac.num and obj_xref.txt are read by obj_xref.pl to generate obj_xref.h. obj_xref.txt links signature OIDs to corresponding public key algorithms and digests. obj_xref.h contains lookup tables for querying this information in both directions. Dependency graph: objects.txt | V [objects.pl] <--+ / \ | V V | nid.h obj_mac.num obj_xref.txt | \ / V V V [obj_dat.pl] [obj_xref.pl] | | V V obj_dat.h obj_xref.h