Exercise 3 solution

    def find_all(data, pattern):
        offsets = []
        start = 0
        while True:
            i = data.find(pattern, start)
            if i == -1:
                return offsets
            offsets.append(i)
            start = i + 1

    print(find_all(b"abcabcab", b"ab"))   # [0, 3, 6]
    print(find_all(b"aaaa", b"aa"))       # [0, 1, 2]
    print(find_all(b"xyz", b"q"))         # []

(Run as written; the output matched the comments.)

Explanation: bytes.find(pattern, start) returns the offset of the next match
at or after start, or -1 when there is none. Restarting the search at i + 1
(rather than i + len(pattern)) also finds overlapping matches, which is why
b"aaaa" contains "aa" at three offsets. Use i + len(pattern) instead if you
want non-overlapping matches only.