How to Deal with Special Characters in Communication Frames?

Question

I am implementing a very simple communication system based in serial transmission and message frames, small, roughly and loosely adapted from Modbus, with 0x3A as starting byte, 0x0D 0x0A sequence as ending bytes.

Most of the action happens when the serial bus detects the 0x3A byte, and stops until the 0x0D 0x0A sequence is found.

I would wish a little advice about:

• How should escape "special ending symbols", such as 0x0A and 0x0D inside a frame? for example, if one of the data in a frame happens to be equal to 0x0A (an address?).
• How should I deal with frames with random raw data, i.e. data included in a frame "as is", without any additional conversion, when the special ending byte happens as data?

I noted ASCII represents digits symbols with the 0x30 offset, should I do something like that in order to represent raw data, at the obvious expense of losing bytes? Do this have a possible real advantage in here?

Summarizing: Should I stop interpreting the ending bytes as ending bytes and treat all frames as given (predefined or header defined) length?

Answer 1

As an 'escaping scheme' a special escape character (ESC) can be inserted before each control character (start-byte, end-byte) that is part of the data. When a reciever receives an ESC the next byte is not an control character but part of the data instead. When ESC itself is part of the data ESC is also inserted before it.

For example with ESC 0x10 and data 0x00 0x10 0x3A 0x00 0x0D the transmitted message should become: 0x3A 0x00 0x10 0x10 0x10 0x3A 0x00 0x10 0x0D 0x0D 0x0A.