C / C++ / MFC

18-Jan-07 18:29

I saw so many articles in codeproject all are working under _MBCS , but wHEN define UNICODE,_UNICODE the encrypt and decypt is not working.
plz Help me out

Rajesh R Subramanian18-Jan-07 19:12

18-Jan-07 19:12

kiranin wrote:
the encrypt and decypt

What is encrypt and decrypt that you are talking about? Are you taking care of string handling accordingly when in UNICODE build?

Nobody can give you wiser advice than yourself. - Cicero
ப்ரம்மா

kiranin18-Jan-07 19:17

18-Jan-07 19:17

am storing usernames and passwords in a file. I want encrypt password and store it in file. When after retrive the password i want to decrypt and compare. This is not working when UNICODE defined application.

Rajesh R Subramanian18-Jan-07 19:19

18-Jan-07 19:19

Can you show a relevant code snippet?

Nobody can give you wiser advice than yourself. - Cicero
ப்ரம்மா

kiranin18-Jan-07 19:23

18-Jan-07 19:23

BOOL EncryptString(LPTSTR szPassword,LPTSTR szEncryptPwd,LPTSTR szKey)
{	
	BOOL bResult = TRUE;	
	HKEY hRegKey = NULL;	
	HCRYPTPROV hProv = NULL;	
	HCRYPTKEY hKey = NULL;
	HCRYPTKEY hXchgKey = NULL;	
	HCRYPTHASH hHash = NULL;	
	DWORD dwLength;
	// Get handle to user default provider.
	if (CryptAcquireContext(&hProv, NULL, NULL, PROV_RSA_FULL, 0))	
	{
		// Create hash object.		
		if (CryptCreateHash(hProv, CALG_MD5, 0, 0, &hHash))		
		{
			// Hash password string.			
			dwLength = sizeof(TCHAR)*_tcslen(szKey);
			if (CryptHashData(hHash, (BYTE *)szKey, dwLength, 0))			
			{
				// Create block cipher session key based on hash of the password.
				if (CryptDeriveKey(hProv, MY_ENCRYPT, hHash, CRYPT_EXPORTABLE, &hKey))				
				{
					// Determine number of bytes to encrypt at a time.
					dwLength = sizeof(TCHAR)*_tcslen(szPassword);					
					// Allocate memory.
					BYTE *pbBuffer = (BYTE *)malloc(dwLength);					
					if (pbBuffer != NULL)					
					{
						memcpy(pbBuffer, szPassword, dwLength);						
						// Encrypt data
						if (CryptEncrypt(hKey, 0, TRUE, 0, pbBuffer, &dwLength, dwLength)) 						
						{
							// return encrypted string
							memcpy(szEncryptPwd, pbBuffer, dwLength);

						}	
						else						
						{							
							bResult = FALSE;						
						}						
						// Free memory
						free(pbBuffer);					
					}
					else					
					{						
						bResult = FALSE;					
					}
					CryptDestroyKey(hKey);  // Release provider handle.				
				}				
				else				
				{
					// Error during CryptDeriveKey!					
					bResult = FALSE;				
				}			
			}			
			else			
			{
				// Error during CryptHashData!				
				bResult = FALSE;			
			}
			CryptDestroyHash(hHash); 
			// Destroy session key.		
		}		
		else		
		{
			// Error during CryptCreateHash!			
			bResult = FALSE;		
		}
		CryptReleaseContext(hProv, 0);	
	}	
	return bResult;
}


BOOL DecryptString(LPTSTR szEncryptPwd,LPTSTR szPassword,LPTSTR szKey) 
{	
	BOOL bResult = TRUE;	
	HCRYPTPROV hProv = NULL;		
	HCRYPTKEY hKey = NULL;		
	HCRYPTKEY hXchgKey = NULL;
	HCRYPTHASH hHash = NULL;
	TCHAR szPasswordTemp[32] = _T("");
	DWORD dwLength;
	// Get handle to user default provider.
	if (CryptAcquireContext(&hProv, NULL, NULL, PROV_RSA_FULL, 0))		
	{
		// Create hash object.			
		if (CryptCreateHash(hProv, CALG_MD5, 0, 0, &hHash))
		{				
			// Hash password string.
			dwLength = sizeof(TCHAR)*_tcslen(szKey);
			if (CryptHashData(hHash, (BYTE *)szKey, dwLength, 0))				
			{
				// Create block cipher session key based on hash of the password.
				if (CryptDeriveKey(
					hProv, MY_ENCRYPT, hHash, CRYPT_EXPORTABLE, &hKey))					
				{
					// we know the encrypted password and the length
					dwLength = sizeof(TCHAR)*_tcslen(szEncryptPwd);						
					// copy encrypted password to temporary TCHAR
					_tcscpy(szPasswordTemp,szEncryptPwd);
					if (!CryptDecrypt(
							hKey, 0, TRUE, 0, (BYTE *)szPasswordTemp, &dwLength))
						bResult = FALSE;						
					CryptDestroyKey(hKey);  // Release provider handle.					
					// copy decrypted password to outparameter
					_tcscpy(szPassword,szPasswordTemp);
				}					
				else					
				{
					// Error during CryptDeriveKey!						
					bResult = FALSE;					
				}				
			}				
			else
			{					
				// Error during CryptHashData!					
				bResult = FALSE;				
			}
			CryptDestroyHash(hHash); // Destroy session key.			
		}			
		else			
		{
			// Error during CryptCreateHash!				
			bResult = FALSE;			
		}
		CryptReleaseContext(hProv, 0);		
	}		
	return bResult;
}

These are the functions am using to encrypt and decrypt. These are working fine when i define _MBCS, but if UNICODE deifned the functions are not working

Rajesh R Subramanian18-Jan-07 19:30

18-Jan-07 19:30

I asked you for relevant[^] code!

And from what I could make from this code, it should work with unicode since TCHAR, _tcslen, etc are being used. Are you just adding _UNICODE or are you replacing _MBCS with _UNICODE?

Nobody can give you wiser advice than yourself. - Cicero
ப்ரம்மா

kiranin18-Jan-07 19:35

18-Jan-07 19:35

i replaced _MBCS with _UNICODE ,UNICODE.
These two functions are failing to encrypt and decrypt even though THCAR ,_tcslen , etc used, thats why i posted a question here...

Rajesh R Subramanian18-Jan-07 19:57

18-Jan-07 19:57

No no, I didn't mean that. There are plenty of external dependencies. For instance, the following: CryptAcquireContext, CryptCreateHash, CryptHashData, etc.

So I do not have a clue of what is going on. Did you try debugging to analyse what is going on in the program?

PS: Replacing _MBCS with _UNICODE is sufficient. _UNICODE implicitly defines UNICODE and so you do not need to define UNICODE separately.

Nobody can give you wiser advice than yourself. - Cicero
ப்ரம்மா

kiranin18-Jan-07 20:04

18-Jan-07 20:04

yeah i debugged

memcpy(pbBuffer, szPassword, dwLength);						
						// Encrypt data
		if (CryptEncrypt(hKey, 0, TRUE, 0, pbBuffer, &dwLength, dwLength)) 			//problem may here i think			
		{
		// return encrypted string
		memcpy(szEncryptPwd, pbBuffer, dwLength);

may be CryptEncrypt function or CryptDeriveKey these two functions
and/or memcpy() function and length these are the main concerns

Rajesh R Subramanian18-Jan-07 20:07

18-Jan-07 20:07

Use wmemcpy while working with Unicode strings. memcpy is for working with non-unicode strings. Sorry that I missed it earlier.

Nobody can give you wiser advice than yourself. - Cicero
ப்ரம்மா

Mark Salsbery19-Jan-07 5:59

Mark Salsbery

19-Jan-07 5:59

You'll need to go through the code and make SURE that all your "dwLength" (and similar) variables
are both initialized and used properly.

The Cryptxxx() APIs only deal in bytes - Unicode doesn't change that. Make sure you are using
counts calculated in TCHARs or in bytes where applicable.

wmemcpy doesn't apply here either Smile | :)

Mark

Mark Salsbery19-Jan-07 6:06

Mark Salsbery

19-Jan-07 6:06

How is this going to work:

dwLength = sizeof(TCHAR)*_tcslen(szEncryptPwd);

If szEncryptPwd is encrypted there's not necessarily an ASCIIZ at the end so _tcslen()
won't work. You need a way to store the encrypted length somewhere or use a pre-defined length.

*EDIT* The encrypt side looks ok but the decrypt isn't going to work as coded.
Also, not all encryption algorithms produce the same number of encrypted bytes as unencrypted
source bytes, but you've assumed that in your encryption function. Be careful Smile | :)

You can check the size required with an extra call to CryptEncrypt() with the pbData param set to
NULL.

Mark