Thread: Unique Files

Is it possible to identify if files are the same if they have the same md5 code? Erm, just came to mind, how do you calculate the md5 of a file?

Not sure, but you might look at this and this.

ok, I took a look at
http://userpages.umbc.edu/~mabzug1/cs/md5/md5.html

And I'm trying to get the Cpp code to work, I get tons of errors and not sure why. All I do is include the md5.cc and md5.hh in my main.cpp and compile and I get a shyteload of errors.

Visual C++ .Net 2k3
Windows XP

Ok so I found in MSDN this:
http://msdn.microsoft.com/library/de...ClassTopic.asp
(imagine life without copy & paste)

I'm a little confused though; I've never used namespaces before. How do I include the System.Security.Cryptography namespace?
namespace System.Security.Cryptography
doesn't work.

Also, will I need to add the mscorlib.dll file in my project somehow?

> Is it possible to identify if files are the same if they have the same md5 code?
No, that just tells you that the files are probably the same.

The probablility is however pretty high. md5 is a 128 bit hash, and the probability that two different files have the same md5 hash is 1 in pow(2,128)

> and I get a shyteload of errors.
Well posting a few of them may be helpful to the rest of us.

Originally posted by Salem
>> and I get a shyteload of errors.
Well posting a few of them may be helpful to the rest of us.

Well, I kinda dumped that code and found out that there are functions already in Security namespace, but I still get an error with this code:

Code:

#define WIN32_LEAN_AND_MEAN
#include <windows.h>

#using <mscorlib.dll>
using namespace System;
using namespace System::IO;
using namespace System::Security;
using namespace System::Security::Cryptography;

INT WINAPI WinMain(HINSTANCE hInst, HINSTANCE hPInst, LPSTR lpszCmdLine, INT iCmdShow)
{
	MSG msgMain;

	while(GetMessage(&msgMain, 0, 0, 0))
	{
		TranslateMessage(&msgMain);
		DispatchMessage(&msgMain);
	}

	MD5 *str = new MD5CryptoServiceProvider();

	return (int)msgMain.wParam;
}

Code:

Error:
main.cpp(4): fatal error C1190: managed targeted code
requires '#using <mscorlib.dll>' and '/clr' option

Ok so I changed the project option to Using Managed Extentions or something like that. No errors now.

damnit, I can't get that to work, never used namespaces before, too much new stuff, back to using the example c++ code given.

Ok, back to using the example code, all I did was include md5.h and md5.cpp into my project.

Code:

[i]Errors:[/]
md5.h(52): error C2061: syntax error : identifier 'istream'

md5.h(54): error C2061: syntax error : identifier 'ifstream'

md5.h(54): error C2535: 'void MD5::update(void)' : member
function already defined or declared

md5.h(60): error C2143: syntax error : missing ')' before '&'

That's only a few of the errors I get. From a glance, all errors are in md5.h

How about sticking to regular C++ before trying to getting yourself horribly tangled in all that vs.net "managed" stuff?


> md5.h(52): error C2061: syntax error : identifier 'istream'
Perhaps a simple C++
using namespace std;
placed at the start of md5.h

Yea, that's why I stopped using the .net shyte. I tried adding the namespace, and it didn't change anything. Could you try the same thing I did? Download the file, extract md5.hh and md5.cc and include them in your project. Then for main.cpp all I did was have a WinMain() function with nothing in it (except the message handling).

Originally posted by neandrake
Is it possible to identify if files are the same if they have the same md5 code? Erm, just came to mind, how do you calculate the md5 of a file?

Also, you may want to move to SHA-1 or RIPEMD-160. MD5 has been demonstrated to be weaker against collisions in some sense, which is why most cryptographic systems are supporting MD5 only for backwards compatibility.

In other words, SHA-1 and RIPEMD-160 are better in that collisions (two different messages producing the same digest) seems to be less. At least that is the current thought.

You might try this library:
http://www.eskimo.com/~weidai/cryptlib.html

of course, it may be overkill.

Very overkill

There aren't any sites that show how to implement these hashes in c++? Maybe in another language that can be easily converted to c++

http://www.codeproject.com/cpp/csha1.asp?print=true