Working with bytes in Python 2.7

Question

I'm trying to assemble a byte array in Python for a signature that resembles the following:

4 bytes that represent the length, in bytes, of String A.
String A
4 bytes that represent the length, in bytes, of String B.
String B
4 bytes that represent the length, in bytes, of the Long A value.
Long A

String A + B are utf-8 which I converted to utf-8 using unicode(string, 'utf-8')

I've tried converting each item to a byte array and joining them using the plus symbol e.g.

bytearray(len(a)) + bytearray(a, "utf-8")...

I've also stried using struct.pack e.g.

struct.pack("i", len(a)) + bytearray(access_token, "utf-8")...

But nothing seems to generate a valid signature. Is this the right way to make the above byte array in Python?

FWIW, `unicode(some_string, 'utf-8')` converts the byte string `some_string` _to_ a Unicode string. The `'utf-8'` arg says that the source string is encoded in utf-8. — PM 2Ring, Sep 23 '15 at 13:10
You should also mention whether those length values need to be formatted as big-endian or little-endian. I assume this is for some Net thing, so you probably want big-endian, but it's nice to be explicit. :) — PM 2Ring, Sep 23 '15 at 13:15
possible duplicate of [How to convert integer value to array of four bytes in python](http://stackoverflow.com/questions/6187699/how-to-convert-integer-value-to-array-of-four-bytes-in-python) — Peter Wood, Sep 23 '15 at 13:15
I think you need [`'!I'`](https://docs.python.org/2/library/struct.html#format-strings) instead of `'i'` in your packing format. — Peter Wood, Sep 23 '15 at 13:17

score 2 · Answer 1 · edited Sep 23 '15 at 13:36

The last question is about endianness of the 4 byte lengths, but you can easily control it with the struct module.

I would use

def dopack(A, B, LongA):
    fmt='!'  # for network order, use < for little endian, > for big endian, = for native
    fmt += 'i'

    buf = pack(fmt, len(A))
    buf += A
    buf += pack(fmt, len(B))
    buf += B
    b = bytes(LongA)
    buf += pack(fmt, len(b))
    buf += B

In this way, the LongA value is coded in ASCII, but it is easier, and you can just do int(b) do convert it back to a long.

Working with bytes in Python 2.7

1 Answers1