Interesting case of ghost address (CE with Yuzu)

makotech222 · Post by **makotech222** » Thu Jun 04, 2020 4:48 pm

Reposting from cheatengine forum. Didn't get an answer there.

Context: Using Cheat engine 7.1 (and 6.7) on latest YUZU release on D(iablo) 3

I'm using CE to find and freeze character's mana pool. I'm encountering a very weird issue i've never seen before. It seems that the CE scan sometimes skips over the address I'm looking for.

When I run the scan and get lucky, Ill find the Mana address (float) at value 200. I can freeze it and everything works fine. However, if I scan for the value again, CE will not show it in the search list! It seems to be skipping it over for some reason. I can fiddle with the Start/Stop addresses and sometimes get it to work again, but, usually the address is missing. I've tried fiddling with multiple settings like Fast Scan, but nothing seems to be working.

Most of the time, CE can't find the address, which is a problem when I boot up the game for the first time. I'm not sure whats triggering or causing this, so I'm posting here to get some hints of whats going on and why CE would miss the address.

TimFun13 · Post by **TimFun13** » Thu Jun 04, 2020 9:49 pm

It might just be that the address isn't allocated tell you do something to manipulate the mana value. You should just look in to code injection so you don't have to find the value every time you restart.

makotech222 · Post by **makotech222** » Thu Jun 04, 2020 10:51 pm

Well, if i'm lucky, the first search shows the value I need. But most of the time, the search skips over it, even AFTER i've found it already and saved it.

I'm trying to work up to code injection, but having difficulties with AOB and finding the value makes it even more difficult.

Post by **Dread_Pony_Roberts** » Sat Jun 13, 2020 11:56 pm

You're not alone, in Halo MCC I've been dealing with the same issue of being unable to scan for a value even though my script has already found it. I plan to do much more thorough testing to confirm it and submit it in a bug report, but mainly wanted to let you know that you are not the only one to come across this.

Edit. I've got it repeatable and can confirm it, I'll even make a video on it.

Post by **SunBeam** » Thu Jun 18, 2020 2:59 pm

Please confirm you have [B]MEM_MAPPED[/B] enabled under Scan Settings. If not, tick it and try scanning again. Worth mentioning, I guess :)

makotech222 · Post by **makotech222** » Thu Jun 18, 2020 3:14 pm

Yeah i have MEM_MAPPED checked. Also tried with all 3 checked.

Post by **Dread_Pony_Roberts** » Thu Jun 18, 2020 3:39 pm

If the error is related to the one I reported, it has been addressed and you can download the patch now from the CE patreon page (you have to be a patron of course) or wait for the next official update.

makotech222 · Post by **makotech222** » Thu Jun 18, 2020 7:37 pm

Thanks for the notification. Its possibly it may be relevant to my case, looks like it might be. I wonder what the bug was exactly.

Post by **Dread_Pony_Roberts** » Thu Jun 18, 2020 9:04 pm

There is a bug in CE's scan routine that causes it to fail on 2GB+ memoryregions when doing an unknown initial value scan. (Signed size variable)

Normally games don't allocate chunks of memory in that size, but some emulators, and ported games (that likely have an emulator built-in) do, in which case when you do an unknown initial value scan, it won't find the value

makotech222 · Post by **makotech222** » Fri Jun 19, 2020 12:30 am

Hmm, might not be my case, then. I'm not doing unknown initial value scan :P

Karlos_007 · Post by **Karlos_007** » Thu Oct 01, 2020 6:22 am

makotech222 wrote: ↑
Fri Jun 19, 2020 12:30 am
Hmm, might not be my case, then. I'm not doing unknown initial value scan

Also happens with aob scans on the gui and inside scripts.
Specially on Yuzu emulator since it uses big chunks of memory.
That's why some of my tables don't work for other people.
Source: [Link]

FearLess Cheat Engine