|
Members in Shoutbox
None.
Shoutbox Search
Shoutbox Commands
/w [name] > Whisper
/r > Reply to last whisper /me > Marks as action Shoutbox Information
Moderators may delete any and all shouts at will.
|
Global Shoutbox
Please log in to shout.
[2016-4-16. : 8:43 am] O)FaRTy1billion[MM] -- Clokr_Clokr_ shouted: FaRTy1billion I was expecting this kind of problems to appear. Even if you had no space, you can in some situations remove previous instructions, replace them with some call and place them in a new function. It's only a problem because apparently 256 flingies isn't good enough. ![]() [2016-4-16. : 8:43 am] O)FaRTy1billion[MM] -- I'm replacing them with like mov eax,ecx / call func_AA, which contains only movzx eax,word ptr ds:[eax*2 + 6644F8] / retn[2016-4-16. : 8:42 am] Clokr_ -- FaRTy1billionFaRTy1billion shouted: I can do it in 3 really tiny functions. xD One that movs to eax, one that movs to ecx, and one that movs to edx. Then the 7 bytes can be replaced with the appropriate mov e_x,e_x and the 5 byte call to the function I was expecting this kind of problems to appear. Even if you had no space, you can in some situations remove previous instructions, replace them with some call and place them in a new function.[2016-4-16. : 8:42 am] O)FaRTy1billion[MM] -- It's fine, I just found some memory I can dump my 9-byte, 2-opcode functions in ![]() [2016-4-16. : 8:31 am] Neiv -- FaRTy1billionFaRTy1billion shouted: ya, movzx thing,word ptr [thing*2 + stuff] is a byte bigger than movzx thing, byte ptr [thing + stuff] Ah, I didn't realize that it had to be multiplied by 2 D:[2016-4-16. : 8:31 am] Clokr_ -- FaRTy1billionFaRTy1billion shouted: Clokr_ Images.dat has some other stuff to do but shouldn't be too hard ... the one that will be hard is units.dat I was talking about convincing you ![]() [2016-4-16. : 8:18 am] O)FaRTy1billion[MM] -- I can do it in 3 really tiny functions. xD One that movs to eax, one that movs to ecx, and one that movs to edx. Then the 7 bytes can be replaced with the appropriate mov e_x,e_x and the 5 byte call to the function[2016-4-16. : 7:50 am] O)FaRTy1billion[MM] -- that movzx is 7 bytes, a call is 5, and a mov thing,thing appears to be 2, so that seems the best solution[2016-4-16. : 7:48 am] O)FaRTy1billion[MM] -- lol I could replace them to a call to a function that just does MOVZX EDX,BYTE PTR DS:[EAX+6644F9] and returns[2016-4-16. : 7:45 am] O)FaRTy1billion[MM] -- and a couple of these are in a place where it might be tricky to just push the rest of the function down 1 byte[2016-4-16. : 7:42 am] O)FaRTy1billion[MM] -- ya, movzx thing,word ptr [thing*2 + stuff] is a byte bigger than movzx thing, byte ptr [thing + stuff][2016-4-16. : 7:40 am] O)FaRTy1billion[MM] -- NeivNeiv shouted: And changing byte to word with movzx seems to be the same size, though a regular mov with word requires an additional instruction byte D: I'm getting it a byte bigger[2016-4-16. : 7:33 am] O)FaRTy1billion[MM] -- Clokr_Clokr_ shouted: FaRTy1billion I knew it would work Images.dat has some other stuff to do but shouldn't be too hard ... the one that will be hard is units.dat![]() [2016-4-16. : 7:32 am] O)FaRTy1billion[MM] -- lil-Infernolil-Inferno shouted: REAL NIGGA HOURS WHO DA FUCC UP? I' mabout to go to bed ol[2016-4-16. : 7:11 am] Clokr_ -- FaRTy1billionFaRTy1billion shouted: Well thanks clokr_ for interesting me in this. I spent a whole day on it I knew it would work ![]() ![]() [2016-4-16. : 5:45 am] O)FaRTy1billion[MM] -- oh, GetCurrentProcess has all access, so that should work[2016-4-16. : 5:40 am] O)FaRTy1billion[MM] -- What would really be easy, though, is to reduce the number of entries. ![]() [2016-4-16. : 5:37 am] O)FaRTy1billion[MM] -- I got all the addresses for all the dat files, so it should be simple to apply this to any of them (units.dat has a ton of other stuff thatd need to be changed though)[2016-4-16. : 5:36 am] O)FaRTy1billion[MM] -- Well thanks clokr_ for interesting me in this. I spent a whole day on it ![]() [2016-4-16. : 5:35 am] O)FaRTy1billion[MM] -- units.dat will be way harder, but it at least proves that there isn't any stupid checks or limitations[2016-4-16. : 5:32 am] O)FaRTy1billion[MM] -- I really need to make it just an mpqdraft plugin. Having to run the mpq in mpqdraft and then patching it with a separate program is dumb[2016-4-16. : 4:27 am] jjf28 -- free hookers: https://github.com/jjf28/ScQuickLaunch/blob/master/ScQuickLaunch/ProcessModder.h https://github.com/jjf28/ScQuickLaunch/blob/master/ScQuickLaunch/ProcessModder.cpp[2016-4-16. : 4:19 am] O)FaRTy1billion[MM] -- it's because I was doing it from a separate process. I don't know how to modify the current process to be able to write to itself ![]() [2016-4-16. : 4:15 am] O)FaRTy1billion[MM] -- I don't remember, I just have code that lets me write to SC. ![]() [2016-4-16. : 4:15 am] Neiv -- Thought I think you have to use VirtualProtectEx even if you write from another process?[2016-4-16. : 4:14 am] Neiv -- FaRTy1billionFaRTy1billion shouted: Neiv I dunno how to do that VirtualProtect https://msdn.microsoft.com/en-us/library/windows/desktop/aa366898%28v=vs.85%29.aspx[2016-4-16. : 4:14 am] O)FaRTy1billion[MM] -- NeivNeiv shouted: obviously after getting write access to the code I dunno how to do that[2016-4-16. : 4:13 am] Neiv -- I've fixed a bunch of buggy diablo 2 plugins which assumed that dlls are always at same address <.<[2016-4-16. : 4:13 am] Neiv -- And patching bw is easy, the executable doesn't get relocated unlike dlls do ![]() [2016-4-16. : 4:12 am] Neiv -- WriteProcessMemory and such should work even with using the current process, but you can just cast a integer to a pointer and write to that[2016-4-16. : 4:12 am] O)FaRTy1billion[MM] -- my EUDEnable source folder is a mess. There's like 4 different programs in there xD[2016-4-16. : 4:10 am] O)FaRTy1billion[MM] -- I know how to overwrite code when working from an external process ![]() [2016-4-16. : 4:10 am] Neiv -- Yea, they are being loaded in sc's process when the Initialize function is called[2016-4-16. : 4:09 am] O)FaRTy1billion[MM] -- Also will I have to do stuff to overwrite the code? lol[2016-4-16. : 4:08 am] O)FaRTy1billion[MM] -- Do MPQDraft plugins share memory with SC? I don't remember[2016-4-16. : 4:02 am] O)FaRTy1billion[MM] -- because if this doesn't work / ends up being really complicated somehow then it won't be worth doing it now[2016-4-16. : 4:00 am] Neiv -- And changing byte to word with movzx seems to be the same size, though a regular mov with word requires an additional instruction byte[2016-4-16. : 3:59 am] Neiv -- 256 flingies aren't enough If that place is only one with 8-bit ids then it's not too bad[2016-4-16. : 3:58 am] O)FaRTy1billion[MM] -- Changing units.dat flingy value to 2 bytes is trivial, and then just change a handful of MOVZX E_X,BYTE PTR DS:[E_X+6644F8] to WORD .... I don't know the instructions well enough to know if that will change the size of the opcodes ![]() [2016-4-16. : 3:57 am] O)FaRTy1billion[MM] -- Although with flingys in particular it would be really easy to change that to be 16-bit[2016-4-16. : 3:56 am] O)FaRTy1billion[MM] -- So there still will be limits with just replacing pointers, but you could still add 46 or whatever flingies[2016-4-16. : 3:55 am] O)FaRTy1billion[MM] -- LoveLessLoveLess shouted: I just noticed, Leeroy has been absent from SEN for half a year. What hell is this. yes |