Cheat Engine

[bknight2602]

ms = createMemScan()
--Need to search for Array Of Bytes
--Barret and Tifa exists from beginning of game
--AOB for Barrett is 22 41 52 52 45 54 FF FF FF before character named, after naming 22 41 52 52 45 54 FF FF
--AOB for Tifa is 34 49 46 41 FF FF FF FF FF before character named, after naming 34 49 46 41 FF FF
memscan_firstScan(ms, soExactValue, vtByteArray, rtRounded, "22 41 52 52 45 54 FF FF", "", 0x01000000, 0x7fffffff, "*X*C*W", fsmNotAligned, "", true, false, false, false);
memscan_waitTillDone(ms);
fl = createFoundList(ms);
foundlist_initialize(fl);
foundlist_getCount(fl);
print("num of search hits", foundlist_getCount(fl));--Output 2, correct
for x = 0, foundlist_getCount(fl)-1 do--Items are in the hex format (0125D60F) without 0x
memrec1 = foundlist_getAddress(fl, x);
--print("memrec1 from foundlist_getAddress(fl, x)", memrec1);
memrec2 = "0x" .. memrec1;
print("memrec1 concatenated with '0x'", memrec2);-- Output 0x1DAC7EC, 0x01EF72C4, correct
tifaaddress = memrec2 + 0x84;--Items are in the number format of the hex summation
--print("Search record number", x, " test Tifa address", tifaaddress);
tifaaddress = string.format('%x', tifaaddress);--Items in number format converted to hex format without 0x
tifaaddress = "0x" .. tifaaddress;
print("Tifa test address as hex format", tifaaddress);--);-- Output 0x1dac870, 0x01ef72c4, correct
print(readBytes(tifaaddress, 6));--ReadBytes(address, bytecount) returns an array of integers--Output 52 73 70 65 255 255, correct for both
--bytestring = readBytes(tifaaddress, 6);--Attempted these two lines output 52, when true used after 6 resulted in blank output
--print("bytestring", bytestring);
if readBytes(tifaaddress, 6) == "52, 73, 70, 65, 255, 255" then--"52 73 70 65 255 255" is the integer representation of "34 49 46 41 FF FF"
byteoffset = tonumber(memrec2);--Original format n = tonumber("0x" .. s);
print("Correct '22 41 52 52 45 54 FF FF' to recalculate table", byteoffset);--Output blank, so IF statement is never true
end;
end;
1. During the process I found two different ways to use readBytes one without a t/f after the bytecount and one with a t/f. What is the difference and uses?
2. Correct syntax for searching for AOB?

[Dark Byte]

The one without a true returns the bytes as different results

e.g:

[bknight2602]

Ah that's why it only printed one result. The 0x01000000 just eliminates informational addresses used by the game (Final Fantasy VII).
I guess I wasn't clear in the second question, once I have the two addresses what is the correct syntax to search another location for a set of bytes?
if readBytes(tifaaddress, 6) == "52, 73, 70, 65, 255, 255" then
since the print(readBytes(tifaaddress, 6)) resulted in 52 73 70 65 255 255
I did try to compare "52 73 70 65 255 255", and that didn't work either.

Would it be better to have two memscans and then compare the resulting foundlist?

Edit: I believe that the game vaiables start at 0x01000000

[mgr.inz.Player]

[bknight2602]

ms = createMemScan()
--Need to search for Array Of Bytes
--Barret and Tifa exists from beginning of game
--AOB for Barrett is 22 41 52 52 45 54 FF FF FF before character named, after naming 22 41 52 52 45 54 FF FF
--AOB for Tifa is 34 49 46 41 FF FF FF FF FF before character named, after naming 34 49 46 41 FF FF
memscan_firstScan(ms, soExactValue, vtByteArray, rtRounded, "22 41 52 52 45 54 FF FF", "", 0x01000000, 0x7fffffff, "*X*C*W", fsmNotAligned, "", true, false, false, false);
memscan_waitTillDone(ms);
fl = createFoundList(ms);
foundlist_initialize(fl);
foundlist_getCount(fl);--Found 9 occurreneces
for x = 0, foundlist_getCount(fl)-1 do--Items are in the hex format (0125D60F) without 0x
memrec1 = foundlist_getAddress(fl, x);
memrec2 = "0x" .. memrec1;
print("Barret found concatenated with '0x'", memrec2);--Found 9 addresses identical to a normal search in CE
tifaaddress = memrec2 + 0x84;--Items are in the number format of the hex summation
tifaaddress = string.format('%x', tifaaddress);--Items in number format converted to hex format without 0x
tifaaddress = "0x" .. tifaaddress;);--Found 9 addresses identical to a normal search in CE
print(table.concat(readBytes(tifaaddress,6,true),', '));--Output 73, 70, 65, 255, 255 for all 9 occurrences so IF statement never true
if table.concat(readBytes(tifaaddress,6,true),', ') == "52, 73, 70, 65, 255, 255" then
byteoffset = tonumber(memrec2);--Original format n = tonumber("0x" .. s);
print("memrec2 from table", byteoffset);
--print("Correct '22 41 52 52 45 54 FF FF' to recalculate table", byteoffset);
end;
end;
--The remainder of code will be tried after part 1 is correct
--addresslist = getAddressList();
--memrec3 = addresslist_getMemoryRecordByDescription(addresslist, "Orrin");
--orrinaddress = memoryrecord_getAddress(memrec3);--Item in num format
--for x = 0, addresslist_getCount(addresslist)-1 do
--memrec4 = addresslist_getMemoryRecordByID(addresslist, x);
--if byteoffset > orrinaddress then
--memoryrecord_setAddress(memrec4, memoryrecord_getAddress(memrec4) + byteoffset - orrinaddress);
--elseif byteoffset < orrinaddress then
--memoryrecord_setAddress(memrec4, memoryrecord_getAddress(memrec4) - byteoffset + orrinaddress);
--else
--end;
--end;

Thanks for the tip, but only outputs the last 5 of the 6 bytes, not sure why

EDIT: Changed
print(table.concat(readBytes(tifaaddress,6,true),', '0,5));--Output 52, 73, 70, 65, 255, 255 for all 9 occurrences
if table.concat(readBytes(tifaaddress,6,true),', '0,5) == "52, 73, 70, 65, 255, 255" then
byteoffset = tonumber(memrec2 + 0x84);--Original format n = tonumber("0x" .. s);--Change this line as it was incorrect by 0x84
print("memrec2 from table", byteoffset);Output 9 correct Tifas addresses

On two second half of code.

[bknight2602]

byteoffset from previous section of code = 116290112

[bknight2602]

Well I did finally get the code to run, but MANY questions arise.

[Dark Byte]

1:
Lua tables start their index at 1
C/Pascal arrays start their index at 0, and the element indexer is a C/Pascal object.
Mostly ce lua uses pascal/c arrays, but in a few rare cases like the memoryrecord_getAddress it returns a lua table

2:
You use addresslist_getMemoryRecordByID() instead of addresslist_getMemoryRecord()
Each cheat entry has an unique ID that never changes no matter where you drag it to or what addresses you put in front
If you use addresslist_getMemoryRecord() you get the index in the list (counted from 0)

3: Depends on the game and emulator. Perhaps it's used as an default initializer, or one is used by the emulator internally and the other is used by the game where when the game wishes to read the address it gives the other one instead (speed improvement)

[bknight2602]

1. Ok
2. Ok, I did not understand the implication of each instruction.
3. Ok
Final code for Final Fantasy VII using the emulator psxfin.exe, as it executes properly at least twice, works after Barret is "named" during the initial sequence of the game not before AND Barret must not be re-named.
Onto next game, thanks for the help.

[bknight2602]

The above code worked using one lap, but didn't work using another.
The error was pointing to a memory block that obviously doesn't exist in the second lap. So I need an error trapping routine to continue the program if it finds an error.

This is the line that failed:
if table.concat(readBytes(secondbarretaddress,8,true),', ',0) == "34, 65, 82, 82, 69, 84, 255, 255" then
In this case the loop fails on the third of 7 finds. The error reports bad argument #1 to 'concat' (table expected, got nil)

I looked in the code list and found errorOnLookupFailure(state):If set to true (default) address lokups in string form will raise an error if it can not be looked up...(Useful for pointers that don't work 100% of the time).
Since that command only exists in 6.2, I installed it, pulled up a previous code that worked with 6.1 and got an errror with the line listed above. Changed it to read from 1 instead of 0 and the code performed, BUT the addresses were changed incorrectly.
This instance the name correct location is 0x67BC7EC which is
108775404, the table address was 0x64BC7EC which is 105629676.
The address correction should be current address +108775404(correct location) - 105629676(current table address) that should inrease all the addresses by 3145728 (0x300000) which is what the desired results should be, and was with ver 6.1. With 6.2 the addresses looked like this:
108614448, 108614452 etc. Very confusing until I applied a calculator and the numbers become the equivalent to the associated hex addresses.

It appears to me that the procedure, memoryrecord_setAddress(memrec4, memoryrecord_getAddress(memrec4) + byteoffset - barretaddress) now treats these as hex addresses, whereas 6.1 treated them as number addresses. IF this is correct then all my coding built with 6.1 is incorrect with 6.2. This does seem to be inconsistent.

[bknight2602]

Attempting to fix the code but getting incorrect address recalculations

[Dark Byte]

It's a bug in 6.2 where it detects that the input can be parsed as a string, and thus handles it as a string input (problem with lua_isstring() is that it returns true for integers)

you need to convert the new address to a cheat engine interpretable address

So, instead of

[bknight2602]

Great that worked almost 100%, as first two addresses are incorrect, but that may have been my tinkering during the debug.

1. What does errorOnLookupFailure(false) do exactly? Just continue with next statement?
2. Did you see the question of adding the function statement? As this will lead to a form used to perform some other tasks including the recalculation of addresses.

[Dark Byte]

1: Instead of giving an error and stopping the script it makes it return nil (check for that manually)

2: Did the execute button have the functionname in it's "onclick" event tab ?

[bknight2602]

No the form hasn't been tried yet with 6.2. The function was put into the Lua script: Cheat Table.
I was tinkering with one form on 6.1, but now that has been negated. Two projects for the form would be:
1. One button to recalculate the addresses that will use the function RecalculateAddresses()

2. One button to change the value of a block of addresses, but I could not get the code right as I couldn't get writeInterger(address, value), writeIntergerLocal(address, value) or doValueChange(addresslist).

The task intended is:

Table Item 1
Table Item 2
...
Table Item n--Change these values to ff. Don't know how to find n address for a for loop as m is a known value
Table Items n+m
...
Table Item last

Edited: What I meant to say, I can find that n record by the table description as done previously, but I don't know how to access the nth table record and then write vaues to the mth record.