Piped text to binary

[rudyloo]

Hi Guys

Quick question.

What do you think the best way is to break up pipped data coming into a C++ programming from command line into x number of bytes (eg. 5 bytes at a time) for manipulation?

I have the following at the moment, but not sure if this is correct and how to handle the tail of the message where i may have less than 5 bytes / characters in the data block.

        while (!cin.eof())
        {
            int size = 5;
            
            char *buffer = new char[1];
            buffer[0] = '\0';
            cout << buffer << "\n";
            cin.read(buffer, size);

eg. 1234567

Would show me
12345
67
45

The 45 should not be there!

Thanks

[Philip Nicoletti]

1) You are only allocating 1 space in buffer, instead of 5

2) In while loops, checking eof() almost always leads to
outputting extra values, since eof() does not become true
until an actual attempt to read past eof is made.

3) A safer way:

while (cin.read(buffer, size))
{
   // do something with buffer.
   // you can use the gcount() member function
   // to see how many characters were read
}

[rudyloo]

Hi,

With the following method, for "1234567"

I would only see 12345

67 isn't shown.

Thanks

 char *buffer = new char[5];
        
 while (cin.read(buffer, 5))
 {
     cout << buffer << "\n";
 }

[2kaud]

Try this

const int NOBYTES = 5;
char buffer[NOBYTES + 1] = {0};
        
	while (cin.get(buffer, NOBYTES + 1))
		cout << buffer << endl;

[rudyloo]

Hi

The last example works great - but only for 1 line.

I'm sorry if my example data was misleading.

I could have something like:

1234567
Hello This Is
A Test

I would need to ready 5 characters aka bytes at a time.

Thanks

[2kaud]

How do you want to deal with <LF>? Do you want the above to show as

12345
67Hel
lo Th
Is A 
Test

or

12345
67
Hello
This 
Is
A Tes
t

[rudyloo]

Bottom version.

[2kaud]

OK. Try this

const int NOBYTES = 5;
string cline;
        
	while (getline(cin, cline))
		for (size_t pos = 0; pos < cline.length(); pos += NOBYTES)
			cout << cline.substr(pos, NOBYTES) << endl;

[rudyloo]

This works.

I need to do manipulation with the chucks of strings. Isnt it best to use chars to do this? I need to mangle the block using bit operation.

[2kaud]

This works.

I need to do manipulation with the chucks of strings. Isnt it best to use chars to do this? I need to mangle the block using bit operation.

You can acess the chars in the string using iterators or [] or if you really must, use the class function .c_str() to return a pointer of type const char*. See http://www.cplusplus.com/reference/string/string/

What exactly are you wanting to do with the blocks of text?

[rudyloo]

I want to read 5 characters in at a time, and then XOR it with a password of 5 characters long. And then output the mangled block of 5 characters to file or screen.

So, I would want to retain all the special characters as well. eg. \n \t spaces etc..

Eg. XOR each character of "ABCDE" with "MYWORD". A AND M, and then B and Y, and then C and W etc..

[2kaud]

MYWORD is 6 chars!

You want to leave 'whitespace' chars un-mangled and xor the others. What about other special chars such as punctuation etc or do you only want to mangle letters - what about numbers?

PS This isn't an assignment is it?

[rudyloo]

From what I understand ASCII is a subset to Unicode. I would ned to handle anything in the ASCII table. So I assume this means spaces, and numbers.

[2kaud]

This will mangle all text entered. The string mangle must be at least NOBYTES in length.

const string mangle = "MYXOR";
const int NOBYTES = 5;
string cline;
        
	while (getline(cin, cline))
		for (size_t pos = 0; pos < cline.length(); pos += NOBYTES) {
			string mang = cline.substr(pos, NOBYTES);
			for (size_t m = 0; m < mang.length(); ++m)
				mang[m] ^= mangle[m];

			cout << mang << endl;
		}

[rudyloo]

Hi, thanks for all your help.

I have it reading line by line and then XORing char by char given the mangle string.

I can output the encrypted file fine. However re-encrypting (decrypting) the encrypted file doenst give me the original piped in data. (Which it should).

I am thinking when I XOR, I probably get non standard characters (letters, numbers, basic symbols), and probably have \n \0 in the encrypted data. Therefore when I try to reprocess the encrypted data, it processes the data line by line - which doesn't match up.

So, perhaps reading 1 line at a time isn't the way to go.

Thoughts?