Retry with an Attempt Limit

Simulate a callable whose ordered outcomes are provided in outcomes. Each entry is either FAILURE or SUCCESS:value.
Invoke outcomes from left to right, using at most maxAttempts total attempts. Stop at the first success and return SUCCESS:value|attempts. If every permitted attempt fails, return FAILED|attempts.

Function
retry(outcomes: String[], maxAttempts: int) → String

Examples
Example 1
outcomes = ["FAILURE","FAILURE","SUCCESS:ready"]
maxAttempts = 3
return = "SUCCESS:ready|3"
The third permitted attempt succeeds, so no later outcome is inspected.
Example 2
outcomes = ["FAILURE","SUCCESS:late"]
maxAttempts = 1
return = "FAILED|1"
The attempt budget is exhausted before the success.

Constraints
1 <= outcomes.length <= 100000.
1 <= maxAttempts <= outcomes.length.
Every entry is exactly FAILURE or starts with SUCCESS:.
A success value contains printable ASCII characters and may be empty.

What it really reduces to: a single loop from index 0 to maxAttempts-1. At each step, check if the entry starts with SUCCESS:. If it does, return SUCCESS:value|attempts, where attempts is i+1 and the value is everything after the first 7 characters. If the loop ends with no success, return FAILED|maxAttempts. That's O(maxAttempts) time and O(1) extra space. The pitfalls are small but real. The success value can be empty, so SUCCESS: must return SUCCESS:|k. The value may contain colons or pipes, so slice by prefix length instead of splitting on a delimiter. Don't look past maxAttempts even if a success sits right after it, as Example 2 shows. The constraint maxAttempts <= outcomes.length means you never run off the end. If you freeze on the formatting during the live OA, StealthCoder can hand you the clean loop.

Drill it cold or hedge it with StealthCoder. Either way, don't walk into the OA hoping you remember the trick.