Cheat Engine

[oddgamer]

So many, many moons ago, I asked the wonderful people here to help me make a custom type for searching for "ticks", to make finding time-related values easier in Rimworld (yes, there are other, easier ways to cheat this game, yes I could turn on the dev mode or edit the save file or any of a billion other things, thank you, I'm aware, I'm having fun doing it this way, moving on).

Time in RimWorld is measured in ticks. 60 ticks to a real-world second (which is shown sometimes), 2500 to an in-game hour, 60,000 to an in-game day (24 in-game hours), and 3,600,000 to an in-game year (60 in-game days), all stored as a 4-byte signed integer. When I want to, say, find the age of someone that is 42 years old, I need to bust out the calculator and figure out what 41 x 3,600,000 and 43 x that are in order to get the possible range of values. Which is frustrating. The idea here would be that the number displayed would be a float or double (probably float) that represents the integer number divided by those various values (so different types for second, hour, day, year).

This one, for finding seconds of time left and from which I adapted the others, does not work. I don't think it ever worked. I didn't write this, I don't know who did (can't find it on here). I gave it a value of 50, which is nearly 60 and so the number reported should be almost 1 (specifically it should be 0.833333). Instead what I get is 1.401298464E-45. Not even close. Can anyone help fix this?

[ParkourPenguin]

[oddgamer]

[Csimbi]

[ParkourPenguin]

No. The ConvertRoutine function takes an int (really 4-bytes of arbitrary data), the ConvertBackRoutine function is passed an int, and the native data stored in memory is a 4-byte integer. Everything is 4 bytes.

Setting ByteSize to 8 is weird if the native type is a 4-byte integer (it should be 4), but that won't cause any bad things to happen. CE just gives the functions more data than they need in the input / output parameters. In theory, if the 4-byte value is the last 4 bytes in a scannable region of memory, then it wouldn't be found; however, that's very unlikely to happen.

[Csimbi]

So, multipling two ints can never be longer than an int? That's news to me.

[ParkourPenguin]

That custom type isn't multiplying two ints. It's multiplying two floats. `mulss` means "multiply scalar single", with "single" referring to floats.

There is a problem with `cvttss2si` (convert / truncate scalar single to signed integer) if the truncated float is outside the valid range of the signed integer output type. This isn't likely to be relevant here: the float we're converting to an integer was created from an integer already in the game. If there was a chance of the float-to-int conversion overflowing, then what happens in the vanilla game when the integer wraps around is far more concerning.

Of course, users could always enter ridiculous values, but if they do that, it's their fault.

[Csimbi]