BYTE, char data corruption

-- If you don't know what a byte is, don't bother reading. --

My program reads/writes the bytes of a program. I specify the bytes to be written [e.g. 0x01 0x02 0x03 0x04] and it writes them to the address.

The problem I'm having is: They must be formatted as char when calling my write function, they can't be BYTEs. My bytes get corrupted, even if I initially declare them char.

I'll have a byte, 0x8E. The char will make it 0xFFFFFFFFFFFFFF8E.

I've tried to Variable = Variable%256; it, but it doesn't change the outcome. 0xFFFFFFFFFFFFFF8E is -24 as an integer, which isn't a valid char. This makes my function fail.

I need a way to either preserve the byte, or rid the char of the additional Fs.

0xFFFFFFFFFFFFFF8E is an integer, not a char or a BYTE (unsigned char). you need to post some code. How are you writing them to the file?

Quote:

---

-- If you don't know what a byte is, don't bother reading. --

---

Nice way to get people to help you.

The problem is really simple.

Code:

---

int i = 0xFFFFFFFFFFFFFF8E;
char *byte = (char*)&i;
for(int j=0; j < 4; j++)
  cout<<byte[j];

---

>-- If you don't know what a byte is, don't bother reading. --
I'm willing to bet that we know more about what a byte is than you do. ;)

I prefer nibbles myself :)

nybbles, I think. :)

I've seen it spelt both ways so /shrug

Quote:

---

Originally Posted by Thantos

Code:

---

int i = 0xFFFFFFFFFFFFFF8E;
char *byte = (char*)&i;
for(int j=0; j < 4; j++)
  cout<<byte[j];

---

---

Isn't 0xFFFFFFFFFFFFFF8E a 64-bit value?

>> >-- If you don't know what a byte is, don't bother reading. --
>> I'm willing to bet that we know more about what a byte is than you do.

Not everyone here on this board knows what a byte is.

0xFFFFFFFFFFFFFF8E is indeed a 64 bit value.

Quote:

---

Originally Posted by JaWiB

Isn't 0xFFFFFFFFFFFFFF8E a 64-bit value?

---

*butt cover* yeah but I was supposing a system with a 16bit BYTE....

and besides its easily fixed:

Code:

---

for(int j=0; j < sizeof i; j++)

---

>Not everyone here on this board knows what a byte is.
That much is obvious, judging from the frequency of questions about it and incorrect responses. ;)

I added that comment because last time I spoke of a byte here, the only responses I got were "What's a byte"?

That's because a BYTE is not a native type. And it does not necessarily mean an octet. So without adding context, it's marginally as informative as FOO.

Anyway, I could bypass this problem if I could get my struct to accept a byte array instead of a char array.

Code:

---

struct input
{
        UINT_PTR processid;
        void *address;
        unsigned short int bytestowrite;
        BYTE bytes[300];
};

---

Code:

---

input Line = { GetCurrentProcessId(), Address, Length, _Buffer };

---

Buffer is a BYTE[300].

error C2440: 'initializing' : cannot convert from 'unsigned char [300]' to 'unsigned char'

Changing both types to char[300] is accepted, but "corrupts" it.

This gets sent through DeviceIoControl, so I can't declare them pointers, I need the actual value.

Quote:

---

Originally Posted by Denethor2000

error C2440: 'initializing' : cannot convert from 'unsigned char [300]' to 'unsigned char'

---

What's _Buffer? According to the error, it's an 'unsigned char'.

> input Line = { GetCurrentProcessId(), Address, Length, _Buffer };
You can't assign arrays - use memcpy()

Having BYTE *bytes; in your struct would work, but that wouldn't be making a copy of the buffer.

Quote:

---

Originally Posted by dwks

What's _Buffer? According to the error, it's an 'unsigned char'.

---

Yeah, I edited that in a minute later, sorry about that.

It's a BYTE array.

Code:

---

BYTE _Buffer[300];

---

Code:

---

input Line = { GetCurrentProcessId(), Address, Length, _Buffer };

---

What the hell is the above? what are those curly-braces doing there? First calls GetCurrentProcessID() and assigns the return value to Line (which is wrong because Input is not a DWORD). but what's the purpose of Address, Length, and _Buffer in that line?

Quote:

---

Originally Posted by Ancient Dragon

Code:

---

input Line = { GetCurrentProcessId(), Address, Length, _Buffer };

---

What the hell is the above? what are those curly-braces doing there? First calls GetCurrentProcessID() and assigns the return value to Line (which is probably wrong if Input is not a DWORD). but what's the purpose of Address, Length, and _Buffer in that line?

---

It's a struct. Which I included in the same post.

It gets passed to a driver with DeviceIoControl. Which has nothing to do with my problem.

Probably to initialize the remaining members of the structure.

Code:

---

struct s {
    int x;
    float f;
};

s one;
s two = {1};
s three = {1, 2.0};

---

[edit]You beat me to it.[/edit]

Quote:

---

Which has nothing to do with my problem.

---

It might, depending on what _Buffer is. It's [300], right?

Quote:

---

Originally Posted by Denethor2000

It's a struct. Which I included in the same post..

---

Yes, I reread your previous post and found that out. You can't assign members of a struct like that -- each member requires its own assign statement.

Code:

---

Line.processid = GetProcessID();
Line.address = Address;
Line.bytestowrite = Length;
// The next line may not work if sizeof(Line.bytes) > siezof(_Buffer)
memcpy(Line.bytes,_Buffer,sizeof(Line.bytes));

---

You can too assign a struct like that. How about you re-read my post:

Quote:

---

Originally Posted by dwks

Code:

---

struct s {
    int x;
    float f;
};

s one;
s two = {1};
s three = {1, 2.0};

---

---

And what's siezof? :rolleyes:

Quote:

---

Originally Posted by dwks

You can too assign a struct like that. How about you re-read my post:

---

That obscure trick won't work with his struct because it contains character arrays which cannot be copied like that.

Quote:

---

That obscure trick

---

I use it all the time.

Quote:

---

won't work with his struct because it contains character arrays

---

Yes, you're right. (It does work for char*, not for char[].) But

Quote:

---

but what's the purpose of Address, Length, and _Buffer in that line?

---

Address and Length are valid.

Quote:

---

Originally Posted by dwks

I use it all the time.

---

You are one of a very select few who do. I've never see that before in over 20 years.

Quote:

---

Originally Posted by dwks

Yes, you're right. (It does work for char*, not for char[].)

---

well, it really does not work for char* either -- unless all you want copied is the address and not the object to which it points.

technically you aren't doing assignment. You are doing initalization.

Denethor2000 instead of rationing out your limitations to us how about you give us all the info in the beginning?

A BYTE is nothing more then a fancy unsigned char.

Quote:

---

Changing both types to char[300] is accepted, but "corrupts" it.

---

how does it corrupt it?

Quote:

---

You are one of a very select few who do. I've never see that before in over 20 years.

---

What? Really? That's amazing.

Quote:

---

well, it really does work for char* either -- unless all you want copied is the address and not the object to which it points.

---

It does work? :) Oh, you mean doesn't. It does work, because all I want copied is the address.

Quote:

---

Originally Posted by dwks

It does work? :) Oh, you mean doesn't. It does work, because all I want copied is the address.

---

That was a typo that I later corrected. :rolleyes: If all you want is the address, then great, it will work for you.

Quote:

---

You are one of a very select few who do. I've never see that before in over 20 years.

---

It's the only way to initialize (as Thantos says) a struct. It's the same as with arrays. Do you never do that either?

This code sets all elements to zero:

Code:

---

int x[10] = {0};

---

Quote:

---

If all you want is the address, then great, it will work for you.

---

The address is what I would expect, since it's a pointer. :)

Quote:

---

Originally Posted by dwks

It's the only way to initialize (as Thantos says) a struct. It's the same as with arrays. Do you never do that either?

---

No. Never like that, except maybe in an array of structures when initializers contain no function calls.

Quote:

---

Originally Posted by dwks

This code sets all elements to zero:

Code:

---

int x[10] = {0};

---

---

Do that all the time, but that's not the same as initializing a structure.

Quote:

---

but that's not the same as initializing a structure.

---

It's pretty much the same. They both initialize things. The only difference is that with structures, the type you're initializing can change.

Quote:

---

Originally Posted by Thantos

A BYTE is nothing more then a fancy unsigned char.

---

In C++ language, yes. I use BYTE because it's what I'm actually dealing with: bytes. Less confusing to me, and whoever I share my code with.

Quote:

---

how does it corrupt it?

---

Changing it to a char makes 50% of my bytes turn from 0x0000008E to 0xFFFFFF8E.

Use & then to elininate the extra Fs.

[edit]

Quote:

---

Changing it to a char

---

How? With a cast?[/edit]

Quote:

---

Originally Posted by Denethor2000

Changing it to a char makes 50% of my bytes turn from 0x0000008E to 0xFFFFFF8E.

---

Impossible -- one byte cannot contain that big a number. The max falue of a byte is 0xff. you are probably seeing the output of a debugger, which is not necessarily the same as what is actually stored in the memory location.

Quote:

---

In C++ language, yes. I use BYTE because it's what I'm actually dealing with: bytes. Less confusing to me, and whoever I share my code with

---

Obviously its causing more problems then its helping.

However you still haven't given up all the information. Or hell even a full code example. if all you need to do is write it out to a file use write() on a binary file

Perhaps he should be "changing" it to an unsigned char.

Quote:

---

Impossible -- one byte cannot contain that big a number. The max falue of a byte is 0xff.

---

What if you were dealing with 16 bit bytes? 32 bit bytes? They'd have a high value then that.

BTW I'm guessing we all are assuming you are using the BYTE specified in the windows header. Can someone tell me if its a #define or typedef?

Quote:

---

Originally Posted by Thantos

What if you were dealing with 16 bit bytes? 32 bit bytes? They'd have a high value then that.

BTW I'm guessing we all are assuming you are using the BYTE specified in the windows header. Can someone tell me if its a #define or typedef?

---

You have a point -- but have you ever heard of such an animal? Not to say there isn't one, but I doubt the OP is using it.

Quote:

---

Originally Posted by windef.h

typedef unsigned char BYTE;

---

Bytes are 2 bits. Words are 4 bits. Dwords are 8 bits. Qwords are 16 bits.

All bytes are 2bit, 0xFF. Stop asking if my bytes are 32bit or 16bit... They're neither, obviously.

Anyway, using memcpy lets me copy my array into the struct, and it works correctly. Thanks for the help. <3

Quote:

---

Originally Posted by Denethor2000

Bytes are 2 bits. Words are 4 bits. Dwords are 8 bits. Qwords are 16 bits.

All bytes are 2bit, 0xFF. Stop asking if my bytes are 32bit or 16bit... They're neither, obviously.

Anyway, using memcpy lets me copy my array into the struct, and it works correctly. Thanks for the help. <3

---

No. on MS-Windows and *nix a byte is 8 bits. What you probably mean is a nibble, which is half the size of a byte.

-- If you don't know what a byte is, don't bother reading. --

I guess you should never have started this thread, since you don't know what a byte is either :D

Ok you are seriously in need of some education then. Bytes are the lowest addressible memory element. By standard they are required to contain at least 8 bits. So whereever you got your size definations, burn it.

Doh, you're right. 1 bit for every digit in the binary. I was thinking 1 bit for each digit in the hex, I don't know why.

Anyway, I'm talking about standard 8bit bytes.

Sorry about that.