Find out the real file type

Question

I am working on an ASP web page that handles file uploads. Only certain types of files are allowed to be uploaded, like .XLS, .XML, .CSV, .TXT, .PDF, .PPT, etc.

I

Answer 1

The following C++ code could help you:

//-1 : File Does not Exist or no access
//0 : not an office document
//1 : (General) MS office 2007
//2 : (General) MS office older than 2007
//3 : MS office 2003 PowerPoint presentation
//4 : MS office 2003 Excel spreadsheet
//5 : MS office applications or others 
int IsOffice2007OrOlder(wchar_t * fileName)
{
    int iRet = 0;
    byte msgFormatChk2007[8]    = {0x50, 0x4B, 0x03, 0x04, 0x14, 0x00, 0x06, 0x00};     //offset 0 for office 2007 documents
    byte possibleMSOldOffice[8] = {0xD0, 0xCF, 0x11, 0xE0, 0xA1, 0xB1, 0x1A, 0xE1};     //offset 0 for possible office 2003 documents

    byte msgFormatChkXLSPPT[4]  = {0xFD, 0xFF, 0xFF, 0xFF};     // offset 512: xls, ppt: FD FF FF FF 
    byte msgFormatChkOnlyPPT[4] = {0x00, 0x6E, 0x1E, 0xF0};     // offset 512: another ppt offset PPT   
    byte msgFormatChkOnlyDOC[4] = {0xEC, 0xA5, 0xC1, 0x00};     //offset 512: EC A5 C1 00 
    byte msgFormatChkOnlyXLS[8] = {0x09, 0x08, 0x10, 0x00, 0x00, 0x06, 0x05, 0x00};     //offset 512: XLS

    int iMsgChk = 0;
    HANDLE fileHandle = CreateFile(fileName, GENERIC_READ,
        FILE_SHARE_READ, NULL, OPEN_EXISTING, FILE_ATTRIBUTE_READONLY, NULL  );
    if(INVALID_HANDLE_VALUE == fileHandle) 
    { 
        return -1; 
    }

    byte buff[20];
    DWORD bytesRead;
    iMsgChk = 1;
    if(0 == ReadFile(fileHandle, buff, 8, &bytesRead, NULL )) 
    { 
        return -1; 
    }

    if(buff[0] == msgFormatChk2007[0]) 
    {
        while(buff[iMsgChk] == msgFormatChk2007[iMsgChk] && iMsgChk < 9)
            iMsgChk++;

        if(iMsgChk >= 8) {  
            iRet = 1; //office 2007 file format
        }
    } 
    else if(buff[0] == possibleMSOldOffice[0])
    {
        while(buff[iMsgChk] == possibleMSOldOffice[iMsgChk] && iMsgChk < 9)
            iMsgChk++;

        if(iMsgChk >= 8)
        {   
            //old office file format, check 512 offset further in order to filter out real office format
            iMsgChk = 1;
            SetFilePointer(fileHandle, 512, NULL, FILE_BEGIN);
            if(ReadFile(fileHandle, buff, 8, &bytesRead, NULL ) == 0) { return 0; }

            if(buff[0] == msgFormatChkXLSPPT[0])
            {
                while(buff[iMsgChk] == msgFormatChkXLSPPT[iMsgChk] && iMsgChk < 5)
                    iMsgChk++;

                if(iMsgChk == 4)
                    iRet = 2;
            }
            else if(buff[iMsgChk] == msgFormatChkOnlyDOC[iMsgChk])
            {
                while(buff[iMsgChk] == msgFormatChkOnlyDOC[iMsgChk] && iMsgChk < 5)
                    iMsgChk++;
                if(iMsgChk == 4)
                    iRet = 2;

            }
            else if(buff[0] == msgFormatChkOnlyPPT[0])
            {
                while(buff[iMsgChk] == msgFormatChkOnlyPPT[iMsgChk] && iMsgChk < 5)
                    iMsgChk++;

                if(iMsgChk == 4)
                    iRet = 3;
            }
            else if(buff[0] == msgFormatChkOnlyXLS[0])
            {

                while(buff[iMsgChk] == msgFormatChkOnlyXLS[iMsgChk] && iMsgChk < 9)
                    iMsgChk++;

                if(iMsgChk == 9)
                    iRet = 4;
            } 

            if(0 == iRet){
                iRet = 5;
            }
        }
    }


    CloseHandle(fileHandle);

    return iRet;
}