[#MDEV-8922] Bug#20238729 MySQL sometimes produces no warning when it's unable to interpret a character in a given character set

Comment by Alexander Barkov [ 2017-10-13 ]

This patch adds warnings in some cases where they are questionable (but probably tolerable):

Inserting a string literal with bad multi-byte sequences into a BLOB:

set names utf8; -- notice, not utf8mb4

drop table if exists t1;

create table t1 (a blob);

insert into t1 values ('a?b'); -- a string with a non-BMP character in the middle (utf8mb4)

show warnings;

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

| Level   | Code | Message                                 |

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

| Warning | 1300 | Invalid utf8 character string: 'F09F98' |

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

The string can be a result of mysql_escape_string(), and it goes to a BLOB. It's not used as a text string.

In the above example, a non-BMP character was replaced to the question mark, because JIRA does not support non-BMP characters.

Inserting a 8-bit string literal with unassigned characters into a BLOB:

set names cp1251;

drop table if exists t1;

create table t1 (a blob);

set @stmt=CONCAT('INSERT INTO t1 VALUES ("', 0x98,'")'); -- 0x98 is not assigned in cp1251

prepare stmt from @stmt;

show warnings;

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

| Level   | Code | Message                               |

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

| Warning | 1300 | Invalid cp1251 character string: '98' |

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

As in the example above, the string can be result of mysql_escap_string(), and it goes to a BLOB. It's not used as a text string.

Using a 8-bit HEX character string literal with unassigned characters

set names cp1251;

set @stmt=CONCAT('SELECT HEX(_cp1251"', 0x98,'") AS x');

prepare stmt from @stmt;

show warnings;

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

| Level   | Code | Message                               |

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

| Warning | 1300 | Invalid cp1251 character string: '98' |

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

In this example, the literal IS used as a text (notice the character set introducer). But it should be OK to use unassigned characters in a single character set environment (when the client, the server, and all tables and columns use the same character set).

[MDEV-8922] Bug#20238729 MySQL sometimes produces no warning when it's unable to interpret a character in a given character set Created: 2015-10-08 Updated: 2022-09-08
Status:	Open
Project:	MariaDB Server
Component/s:	Character Sets
Affects Version/s:	5.5, 10.0, 10.1
Fix Version/s:	5.5, 10.1