Quote from: Daniel3D on August 09, 2021, 08:47:42 AM
Well, even with comparing the two versions (was has de bit flipt and the other not and knowing the offset. Two hours of looking at it in various ways in different programs (hex and binary editor's) I still don't have a clue... Dyslexia really gets in my way here...
first:
my mcga.exe is minor/micro different to clvn's version - i think he used a different EXEPACK unpacker - the file are nearly the same - except a block of ~100 Null-Bytes at the exe header-end (is padding, not functional relevant) - will recheck that after my vacation - which starts in 4h
UPDATE: UNP.EXE gives the exact same result as clvn - i was using a different exepack-unpacker (also filed a bug report)
second:
i know the code position where the check is for the mcga.exe (~some bytes early as clvn's said - due to my different sized exe-header)
by in need to analyze the other gfx versions - there is not a easy searchable hex pattern due to relocation (the normal dos exe loading process involves patching of segment adresses in code based on the exe-header-relocation table - which is complete different for the exes) - its the same code but not equaly placed in the executables