Cheat Engine

Fresco · Posted: Mon Feb 20, 2012 12:39 pm Post subject:

i have finally managed to write my own exception handler code and fully understand it ... i think

however now the code works like before for the levels that have "the feature on" ... but for those level who don't have it, there's still a cheat/code bug.

now the game doesn't crash at all (because of the exception handler) and therefore doesn't close at the very beginning of level load 1% ...
now the game just stop's loading at 90% and in task manager i see "the process is not responding"
ce is not attached to the process

Dark Byte · Posted: Mon Feb 20, 2012 2:17 pm Post subject:

Looks like it is in an infinite loop

If you also implemented an counter to see how many exceptions have happened you could check that to see if it's looping on the exception handler or something else

Make sure the "onException" part that the exception handler makes the code jump to does not raise an exception again.

and most importantly, do not make the exception handler cause an exception itself.

If you used any push or esp modifying code before the first mov eax,[esp+4] you need to take that into account

edit:
Make sure you ONLY handle exceptions caused by your code. Do not handle those generated by the game.

Fresco · Posted: Tue Feb 21, 2012 7:51 am Post subject:

doing some break tracing to original code ...
at 90% of no feature level load the code continuously writes to multiple values ... however ...my pointer point's nowhere.

here's what i did in my code:

Dark Byte · Posted: Tue Feb 21, 2012 8:06 am Post subject:

Fresco · Posted: Tue Feb 21, 2012 8:53 am Post subject:

i really cant get it work ...
that's my code
i only hide the processname and the offsets.

Dark Byte · Posted: Tue Feb 21, 2012 9:49 am Post subject:

hmm, really no idea.

Change one of the "mov eax,[eax]" lines to "mov eax,[00000000]"

And see if it also crashes in levels where that feature is present ?
If it does there is a problem with the exception handler

And you really don't have another script running that might cause the crash instead ?

Fresco · Posted: Tue Feb 21, 2012 9:56 am Post subject:

but then the pointer would not point anymore to my featureaddress in a level where there is the feature

Dark Byte · Posted: Tue Feb 21, 2012 9:58 am Post subject:

True, but then you can see if the exception is handled properly or not (Sometimes you have to cause an expected bug to debug stuff)

Also, is there any other place where you write to the memory ? (e.g another script or a cheat table entry ? )

Dark Byte · Posted: Tue Feb 21, 2012 10:00 am Post subject:

Nevermind. I just saw the mistake you made

Fresco · Posted: Tue Feb 21, 2012 10:15 am Post subject:

no when i debug the code .. only one code is enabled...
but ... there are two codes and both does the same thing ... so i enable two auto assemble entries; one for code1 and one for code2; ... again .. code1 and code2 does the same thing writes on the same addresses, but nopping one and leaving the other just would not accomplish my goal ... (maybe if the developers put two code that does the same thing ... they had their reasons ... so that's why i just don't want to nop the other)
also because if i just nop the other ... my value would just freeze to: not write 1 ... therefore 1 can still be written in a very short amount of time ...
(i could have just freeze the pointer for that without any script)

... also ... i think that in a very 1 millisecond ... the pointer point somewhere for a very little time .. then returns back to nowhere; when it comes to a level that don't have the feature. how can i be sure of that ? ....
i mean is there any method to check or log ... during playtime ... on what addresses the pointer pointed ?

PS: forgot to say ... that if i nop code2 and leave code 1 the level will load but player will not be placed at level-start but somewhere else in the level and will be surrounded by enemy. i guess that's because the pointer at the very first of level load .. for a 10 milliseconds points to an address that should be 1, and it's not because my code tells not to write 1

[edit]
it worked !! Smile

it was because of push and pop ... the stak

Fresco · Posted: Wed Feb 22, 2012 5:00 am Post subject:

now this is very very weird.

for the level that have the feature ... works.
for the level who don't have it, also works.
but now there's a level where if the feature gets turned on the mission fails.
now my code, as you see above, checks the address on which the code is writing then checks if the code writes 1 ... if yes ... my code say to write 0.
... the pointer points correctly to the address of the feature, and in the level's who have the feature, if it gets turned on ... the feature address is never set to 1 ... in the levels who don't have it ... runs original code ...
but in this new level where, if the feature gets turned on i fail the mission, the feature can till go to 1 for a millisecond ...
now i tried to search for what writes on the address and also what accesses it ... i checked the registers and NO instruction was writing 1 in it...
recapping ...
level 1 = my code NEVER lets feature set to 1.
level 2 = my code runs original code because level 2 don't have feature.
level 3 = pointer points correctly, no code writes 1, but feature address can still go to 1 for a very short amount of time (the needed to fail the mission).

strangely if i nop the code, works, ... but then other features will be disabled as well and i certainly don't want this to happen.

Dark Byte · Posted: Wed Feb 22, 2012 5:13 am Post subject:

This is quite game specific so you'll have to figure out yourself what is happening here.
Perhaps you could find out the address where the current level is stored and apply that info to your code?