B. bHTML Tables Analisys
time limit per test
2 seconds
memory limit per test
256 megabytes
input
stdin
output
stdout

In this problem is used an extremely simplified version of HTML table markup. Please use the statement as a formal document and read it carefully.

A string is a bHTML table, if it satisfies the grammar: 


TABLE ::= <table>ROWS</table>
ROWS ::= ROW | ROW ROWS
ROW ::= <tr>CELLS</tr>
CELLS ::= CELL | CELL CELLS
CELL ::= <td></td> | <td>TABLE</td>

Blanks in the grammar are only for purposes of illustration, in the given data there will be no spaces. The bHTML table is very similar to a simple regular HTML table in which meet only the following tags : "table", "tr", "td", all the tags are paired and the table contains at least one row and at least one cell in each row. Have a look at the sample tests as examples of tables.

As can be seen, the tables may be nested. You are given a table (which may contain other(s)). You need to write a program that analyzes all the tables and finds the number of cells in each of them. The tables are not required to be rectangular.

Input

For convenience, input data can be separated into non-empty lines in an arbitrary manner. The input data consist of no more than 10 lines. Combine (concatenate) all the input lines into one, to get a text representation s of the specified table. String s corresponds to the given grammar (the root element of grammar is TABLE), its length does not exceed 5000. Only lower case letters are used to write tags. There are no spaces in the given string s.

Output

Print the sizes of all the tables in the non-decreasing order.

Examples
Input
<table><tr><td></td></tr></table>
Output
1


Input
<table>
<tr>
<td>
<table><tr><td></td></tr><tr><td></
td
></tr><tr
><td></td></tr><tr><td></td></tr></table>
</td>
</tr>
</table>
Output
1 4


Input
<table><tr><td>
<table><tr><td>
<table><tr><td>
<table><tr><td></td><td></td>
</tr><tr><td></td></tr></table>
</td></tr></table>
</td></tr></table>
</td></tr></table>
Output
1 1 1 3