Decoding an old mail protocol called B2F

I am struggling to decode a weird email protocol called B2F: https://www.winlink.org/B2F

What I have to do is this:

Load the entire file, which includes compressed (binary) attachments.

I then have to start reading lines, until I reach an empty line. The is the demarcation between the headers and the body of the message.

Then I have to read X characters (not lines) (where X is a value mentioned in the headers) of the body + an additional LF.

After that last LF, I have to read the binary file into a stream, so I can then decompress it.

Loading into a TStringList does not allow me to locate the position of the binaries. Loading into a TMemoryStream, I am not sure how I can read strings from this, and afterwards know the exact position from where I can copy the binaries into another stream for decompresion.

I also looked at TStringStream but I don't quite understand how that works.

NOTE: All data is in ASCII, not Unicode.

Maybe I am making it too complicated for myself?

A TFileStream or TMemoryStream will suffice. Break up each step into separate operations. Don't try to do everything all at once.

First, write a function (or use a 3rd party function ¹) that reads 1 byte at a time from a TStream until it reaches a [CR]LF terminator, and then returns the read bytes as a String, omitting the terminator.

^{1: such as Indy's ReadLnFromStream() function.}

Then, load the file into the TStream.

Then, call that reading function in a loop, storing each string in a TStringList or other suitable container ², until a blank string is returned. That's the end of the headers.

^{2: such as Indy's TIdHeaderList class.}

Now, find the Body entry in the header list, parse out its count and read the specified number of bytes from the TStream, plus an extra [CR]LF terminator.

Now, iterate the header list, and for each File entry, parse out its count and read the specified number of bytes from the TStream, plus an extra [CR]LF terminator. Decompress the bytes as needed.