Problem

During debugging in One Drop services, a useful first step is finding where an observed execution trace stops matching the expected control flow. Given two execution logs as arrays of strings, return the earliest event index where they diverge after normalizing nested function scopes.

Each log entry is one of:

• "CALL:name" — enter a function scope
• "RETURN:name" — exit a function scope
• "SET:key=value" — record a state update in the current scope

Two logs are considered equivalent up to an index if, after processing entries in order, they produce the same stack of active function calls and the same latest values for keys visible in the current scope chain. Return the smallest index i where the states differ immediately after processing expected[i] and observed[i]. If the logs have different lengths but match through the shorter one, return the shorter length. If they never diverge, return -1.

Formal Specification

Implement find_first_divergence(expected, observed) where:

• expected: list of strings
• observed: list of strings
• Returns: integer divergence index

Examples

Example 1

• Input: expected = ["CALL:A", "SET:x=1", "RETURN:A"], observed = ["CALL:A", "SET:x=2", "RETURN:A"]
• Output: 1
• Explanation: After index 1, visible state differs because x has different values.

Example 2

• Input: expected = ["CALL:A", "SET:x=1"], observed = ["CALL:A", "SET:x=1", "RETURN:A"]
• Output: 2
• Explanation: The shorter log is fully matched, but lengths differ.

Constraints

• 0 <= len(expected), len(observed) <= 10^4
• Each entry length is at most 100
• Function names and keys are non-empty alphanumeric strings
• Input logs are individually well-formed