Decoding base64 file contents between PHP and C#

Question

I need to serve an AES encrypted, base64 encoded file from PHP to a C# client (Mono, on various platforms). I've successfully got the AES encryption/decryption working fine but as soon as I attempt the base64 encoding/decoding I run into trouble. Both the examples below have the AES disabled, so that shouldn't be a factor.

My simplest test case, a Hello World string, works fine:

PHP serving output-

// Save encoded data to file
$data = base64_encode("Hello encryption world!!");
$file = fopen($targetPath, 'w');
fwrite($file, $data);
fclose($file);

// Later on, serve the file
header("Pragma: public");
header("Expires: 0");
header("Cache-Control: must-revalidate, post-check=0, pre-check=0");
header("Cache-Control: private",false);
header("Content-Type: application/octet-stream");
header("Content-Disposition: attachment; filename=".basename($product->PackageFilename($packageId)));
header("Content-Transfer-Encoding: binary");
header("Content-Length: ".filesize($targetPath));
ob_clean();
flush();
$handle = fopen($targetPath, "r");
fpassthru($handle);
fclose($handle);

C# decoding and using-

StreamReader reader = new StreamReader(stream);
char[] buffer = DecodeBuffer;
string decoded = "";
int read = 0;
while (0 < (read = reader.Read(buffer, 0, DecodeBufferSize)))
{
    byte[] decodedBytes = Convert.FromBase64CharArray(buffer, 0, read);
    decoded += System.Text.Encoding.UTF8.GetString(decodedBytes);
}
Log(decoded); // Correctly logs "Hello encryption world!!"

However once I start trying to do the same thing with the contents of a file, a FormatException: Invalid character found is thrown by Convert.FromBase64CharArray:

PHP serving output-

// Save encoded data to file
$data = base64_encode(file_get_contents($targetPath));
$file = fopen($targetPath, 'w');
fwrite($file, $data);
fclose($file);

// Later on, serve the file
// Same as above

C# decoding and using-

using (Stream file = File.Open(zipPath, FileMode.Create))
{
using (StreamReader reader = new StreamReader(stream))
{
    char[] buffer = DecodeBuffer;
    byte[] decodedBytes;
    int read = 0;
    while (0 < (read = reader.Read(buffer, 0, DecodeBufferSize)))
    {
        // Throws FormatException: Invalid character found
                    decodedBytes = Convert.FromBase64CharArray(buffer, 0, read);
        file.Write(decodedBytes, 0, decodedBytes.Length);
    }
}
}

Is there some kind of additional processing that should be done on larger data for base64 to be valid? Is it perhaps just not appropriate to be doing this with large binary data-and if so how else would you prevent potential problems with characters unsafe for transmission?

Answer 1

Your reading Base64 text code is not correct.

• Base64 is text, so consider using text reader instead
• Base64 may contain new lines/white spaces. Ie it is custom to have split whole Base64 encoded value into 70-80 character long strings.

To verify if data in the file is correct read whole file as string ( StreamReader.ReadToEnd ) and convert to byte array ( Convert.FromBase64String ).

If file contains valid Base64 data and you can't read it as single string you should implement your own Base64 decoding or manually read correct number of non white space characters (multiple of 4) and decode such chunks.

Answer 2

Base64 encoding converts 3 octets into 4 encoded characters. Thus, the length of the data you provide for decoding needs to be a multiple of 4.

First, ensure that DecodeBufferSize is such a multiple of 4. Next, since StreamReader.Read does not guarantee that all the requested bytes will be read, you should continue reading into buffer until either it has been filled, or the end of the stream been reached.