Cheat Engine

Birdi · Posted: Fri May 28, 2021 5:09 am Post subject: Subtracting Floats

I'm completely lost, but this is a bit more complicated than I thought it would need to be.. unless I'm just being an idiot but I've been trying to do this for hours and I'm at my limit.

Just trying to store the difference in hits between the previous and current HP bars of a boss shield, as "shield_hit"

There's plenty of comments in the attached image so it should be explanatory enough, but "shield_hit" always contains 0? The actual shield value decreases so I'm not sure what the issue is.. I've tried so many different approaches to this and I just can't get it. My best guess is Line 86's pop EDX interrupting the cmp? Not too familiar with the conditions but I think I recall seeing it was okay to do this way.. otherwise EDX would need to be popped under code2 which won't always have it pushed to the stack.. maybe a better way I can do that section.
Help super appreciated.

ParkourPenguin · Posted: Fri May 28, 2021 1:55 pm Post subject:

You don't reset the stats if the enemy changes.

You should be pushing and popping the entire register- not just the lower 32 bits. Most modifications to 32-bit registers zero extend to the upper 32 bits. If the game is using the upper 32 bits for something, e.g. a pointer, then the game would crash since you didn't restore it properly.

You seem to have no idea what's a pointer and what's a value. "boss_shield" is a symbol that I presume represents an address in memory allocated by CE (either by an alloc or by a label defined under an alloc). "[boss_shield]" is a pointer you store with "mov [boss_shield],rdi" - it is not a float value as you seem to think it is with the instructions "mov eax,[boss_shield]" and "cmp [boss_shield],edx".

You shouldn't be trying to compare floating point values with integer cmp instructions. Coincidentally, it usually works out anyway.

I can't think of any reason anyone should ever push the stack pointer (rsp) onto the stack.

Birdi · Posted: Fri May 28, 2021 2:13 pm Post subject:

Thanks for the feedback.. just trying to work with what I know and clearly not well enough, haha.

I originally attempted to just take rsp+50's value as the damage, but for whatever reason in doing so it's the previous shield hp, before damage (I should just check what's in xmm0 I guess).. so I tried this stupid workaround to just math it out.

Unfortunately a lot of my learning has been through secondhand examples, brief documentation and a lot of trial and error, so I appreciate you telling me what's wrong with all this. I'll have another proper go at this later on, thanks again.

ParkourPenguin · Posted: Fri May 28, 2021 2:43 pm Post subject:

I'd put a breakpoint at the instruction "subss xmm0,[rsp+50]", look at what xmm0 is, step over the instruction, then look at how it changed.

Breakpoints are really nice for learning what code does. Learning how to use them should be one of the first things you do. (unlike assembly, basic operations with breakpoints can be learned through trial and error easily)