The query below selects a sequence consisting of three
Unicode code points:

U+0055 LATIN CAPITAL LETTER U
U+031B COMBINING HORN
U+0323 COMBINING DOT BELOW

They make all together a single letter equivalent to
"U+1EF0 LATIN CAPITAL LETTER U WITH HORN AND DOT BELOW"
and occupy a single cell in the terminal screen.
The combining marks are printed on top and below the letter U.

The code in "mysql" client erroneously considers combining marks
as normal characters, each occupying its own cell in the screen.
As a result, any sequence with combining marks prints a broken
table with the rightmost border put on a wrong place:

mysql> select _ucs2 0x0055031B0323;

+----------------------+

| _ucs2 0x0055031B0323 |

+----------------------+

| Ự                  |

+----------------------+

1 row in set (0.00 sec)

Note, the border is correct in case of the equivalent pre-composed character:

mysql> select _ucs2 0x1EF0;

+--------------+

| _ucs2 0x1EF0 |

+--------------+

| Ự            |

+--------------+

1 row in set (0.00 sec)