Problem Description

Little M is happily typing code when he suddenly finds that the judging machine for code submissions is not responding.

You checked the backend data. Please use the existing data to determine which test points he did not pass.

It is known that the judging result and whether a test point is passed have the following relationship:

Input Format

The first line contains a positive integer $n$, indicating the number of judging records.

Lines $2$ to $n+1$ each contain a string, representing the judging result of test point $1 \sim n$ in order. The data guarantees that each string is one of $\{\texttt{AC},\texttt{WA},\texttt{TLE},\texttt{MLE}\}$.

Output Format

Output one line containing $m$ positive integers. The $i$-th number $a_i$ indicates that test point $a_i$ was not passed.

In particular, you need to ensure that these $m$ numbers are in increasing order.

5
AC
WA
AC
TLE
MLE

2 4 5

5
AC
AC
AC
MLE
TLE

4 5

Hint

Constraints

For all testdata, $1 \le n \le 10^5$, and the input strings are guaranteed to be in $\{\texttt{AC},\texttt{WA},\texttt{TLE},\texttt{MLE}\}$.

This problem uses bundled tests: all test points with the same constraints are bundled into one $\text{Subtask}$.

The additional limits for each test point are shown in the table below.

Translated by ChatGPT 5