It doesn't you do it in the Encrypt function you presented.

The encrypt subroutine I pasted above was from from the BlowfishSimple class. It comes as part of the BlowfishNET download. I was comparing the libraries to see how the two encrypt and decrypt and I believe they pad differently. That is the reason I pasted the subroutine here. I thought I had mentioned it.

As far using the BlowfishSimple library/class, it pretty much boils down to making the following calls:

bfs = new BlowfishSimple(key);
enc = bfs.Encrypt(plaintext);
[download]

So, as you can see that is why I keep querying about how the padding is done in BlowfishSimple, because that is what I thought is different, but then, I am not sure.

BlowfishSimple:
$fill = $unpadded_len % $block_size;

Crypt::CBC (by default):
$fill = $pad_size;

		size of padding
		0	1	2	3	4	5	6	7
padded with	BlowfishSimple	7	6	5	4	3	2	1	0	(For cyphers with a blocksize of 8)
	Crypt::CBC	0	1	2	3	4	5	6	7

Update: I accidentally had the note about blocksize dependence attached to Crypt::CBC. Fixed.

One problem that you may have is that the NET function ConvertFromBase64 does not give the same result as perl MIME::Base64. The results look very similar but doing a hex dump of the perl and the NET versions shows that they are different. My ramblings on this subject are below. Don't treat it as character data ??? A true Base64 encoding function would be: Create a byte array which is of an appropriate size, then populate it by taking each pair of characters in turn and using Byte.Parse with the number style of HexNumber. http://bytes.com/topic/c-sharp/answers/234105-base64-encode-characters-127-a Base64 encoding turns 3 bytes into 4? http://bytes.com/topic/c-sharp/answers/224907-base64-string-byte-array-whats-length-ratio