path: root/mysql-test/main/ctype_latin1.test

#
# Tests with the latin1 character set
#
--disable_warnings
drop table if exists t1;
--enable_warnings

#
# WL 1494: Treat latin1 as cp1252 for unicode conversion
#

SET NAMES latin1;
CREATE TABLE t1 (a char(1) character set latin1);
INSERT INTO t1 VALUES (0x00),(0x01),(0x02),(0x03),(0x04),(0x05),(0x06),(0x07);
INSERT INTO t1 VALUES (0x08),(0x09),(0x0A),(0x0B),(0x0C),(0x0D),(0x0E),(0x0F);
INSERT INTO t1 VALUES (0x10),(0x11),(0x12),(0x13),(0x14),(0x15),(0x16),(0x17);
INSERT INTO t1 VALUES (0x18),(0x19),(0x1A),(0x1B),(0x1C),(0x1D),(0x1E),(0x1F);
INSERT INTO t1 VALUES (0x20),(0x21),(0x22),(0x23),(0x24),(0x25),(0x26),(0x27);
INSERT INTO t1 VALUES (0x28),(0x29),(0x2A),(0x2B),(0x2C),(0x2D),(0x2E),(0x2F);
INSERT INTO t1 VALUES (0x30),(0x31),(0x32),(0x33),(0x34),(0x35),(0x36),(0x37);
INSERT INTO t1 VALUES (0x38),(0x39),(0x3A),(0x3B),(0x3C),(0x3D),(0x3E),(0x3F);
INSERT INTO t1 VALUES (0x40),(0x41),(0x42),(0x43),(0x44),(0x45),(0x46),(0x47);
INSERT INTO t1 VALUES (0x48),(0x49),(0x4A),(0x4B),(0x4C),(0x4D),(0x4E),(0x4F);
INSERT INTO t1 VALUES (0x50),(0x51),(0x52),(0x53),(0x54),(0x55),(0x56),(0x57);
INSERT INTO t1 VALUES (0x58),(0x59),(0x5A),(0x5B),(0x5C),(0x5D),(0x5E),(0x5F);
INSERT INTO t1 VALUES (0x60),(0x61),(0x62),(0x63),(0x64),(0x65),(0x66),(0x67);
INSERT INTO t1 VALUES (0x68),(0x69),(0x6A),(0x6B),(0x6C),(0x6D),(0x6E),(0x6F);
INSERT INTO t1 VALUES (0x70),(0x71),(0x72),(0x73),(0x74),(0x75),(0x76),(0x77);
INSERT INTO t1 VALUES (0x78),(0x79),(0x7A),(0x7B),(0x7C),(0x7D),(0x7E),(0x7F);
INSERT INTO t1 VALUES (0x80),(0x81),(0x82),(0x83),(0x84),(0x85),(0x86),(0x87);
INSERT INTO t1 VALUES (0x88),(0x89),(0x8A),(0x8B),(0x8C),(0x8D),(0x8E),(0x8F);
INSERT INTO t1 VALUES (0x90),(0x91),(0x92),(0x93),(0x94),(0x95),(0x96),(0x97);
INSERT INTO t1 VALUES (0x98),(0x99),(0x9A),(0x9B),(0x9C),(0x9D),(0x9E),(0x9F);
INSERT INTO t1 VALUES (0xA0),(0xA1),(0xA2),(0xA3),(0xA4),(0xA5),(0xA6),(0xA7);
INSERT INTO t1 VALUES (0xA8),(0xA9),(0xAA),(0xAB),(0xAC),(0xAD),(0xAE),(0xAF);
INSERT INTO t1 VALUES (0xB0),(0xB1),(0xB2),(0xB3),(0xB4),(0xB5),(0xB6),(0xB7);
INSERT INTO t1 VALUES (0xB8),(0xB9),(0xBA),(0xBB),(0xBC),(0xBD),(0xBE),(0xBF);
INSERT INTO t1 VALUES (0xC0),(0xC1),(0xC2),(0xC3),(0xC4),(0xC5),(0xC6),(0xC7);
INSERT INTO t1 VALUES (0xC8),(0xC9),(0xCA),(0xCB),(0xCC),(0xCD),(0xCE),(0xCF);
INSERT INTO t1 VALUES (0xD0),(0xD1),(0xD2),(0xD3),(0xD4),(0xD5),(0xD6),(0xD7);
INSERT INTO t1 VALUES (0xD8),(0xD9),(0xDA),(0xDB),(0xDC),(0xDD),(0xDE),(0xDF);
INSERT INTO t1 VALUES (0xE0),(0xE1),(0xE2),(0xE3),(0xE4),(0xE5),(0xE6),(0xE7);
INSERT INTO t1 VALUES (0xE8),(0xE9),(0xEA),(0xEB),(0xEC),(0xED),(0xEE),(0xEF);
INSERT INTO t1 VALUES (0xF0),(0xF1),(0xF2),(0xF3),(0xF4),(0xF5),(0xF6),(0xF7);
INSERT INTO t1 VALUES (0xF8),(0xF9),(0xFA),(0xFB),(0xFC),(0xFD),(0xFE),(0xFF);

#
# 0x81 0x8D 0x8F 0x90 0x9D are undefined in cp1252
#
SELECT 
  hex(a), 
  hex(@u:=convert(a using utf8)),
  hex(@l:=convert(@u using latin1)),
  a=@l FROM t1;
DROP TABLE t1;

#
# Bug#13145: A table named "C-cedilla" can't be dropped.
# Accept extended cp1252 letters as valid identifiers.
# This test partially checks that "ctype" array is correct
# for cp1252 extended characters 0x80-0x9F.
#
# 0x83 0x0192  #LATIN SMALL LETTER F WITH HOOK
# 0x8A 0x0160  #LATIN CAPITAL LETTER S WITH CARON
# 0x8C 0x0152  #LATIN CAPITAL LIGATURE OE
# 0x8E 0x017D  #LATIN CAPITAL LETTER Z WITH CARON
# 0x9A 0x0161  #LATIN SMALL LETTER S WITH CARON
# 0x9C 0x0153  #LATIN SMALL LIGATURE OE
# 0x9E 0x017E  #LATIN SMALL LETTER Z WITH CARON
# 0x9F 0x0178  #LATIN CAPITAL LETTER Y WITH DIAERESIS
#
SELECT 1 as ƒ, 2 as Š, 3 as Œ, 4 as Ž, 5 as š, 6 as œ, 7 as ž, 8 as Ÿ;

#
# Bug #6737: REGEXP gives wrong result with case sensitive collation
#
select 'a' regexp 'A' collate latin1_general_ci;
select 'a' regexp 'A' collate latin1_general_cs;
select 'a' regexp 'A' collate latin1_bin;


SET @test_character_set= 'latin1';
SET @test_collation= 'latin1_swedish_ci';
-- source include/ctype_common.inc

SET collation_connection='latin1_swedish_ci';
-- source include/ctype_filesort.inc
-- source include/ctype_like_escape.inc
-- source include/ctype_str_to_date.inc
-- source include/ctype_pad_space.inc
SET collation_connection='latin1_bin';
-- source include/ctype_filesort.inc
-- source include/ctype_like_escape.inc
-- source include/ctype_pad_space.inc

#
# Bug#8041
# An unknown character (e.g. 0x84) should result in ERROR,
# It was treated like a space character earlier.
# Howerver, it should still work fine as a string part.
--error 1064
CREATE TABLE „a (a int);
SELECT '„a' as str;


#
# Bug#18321: Can't store EuroSign with latin1_german1_ci and latin1_general_ci
# The problem was in latin1->utf8->latin1 round trip.
#
set @str= _latin1 'ABC €°§ß²³µ~ äöüÄÖÜ áéíóú ÀÈÌÒÙ @ abc';
SELECT convert(@str collate latin1_bin using utf8);
SELECT convert(@str collate latin1_general_ci using utf8);
SELECT convert(@str collate latin1_german1_ci using utf8);
SELECT convert(@str collate latin1_danish_ci using utf8);
SELECT convert(@str collate latin1_spanish_ci using utf8);
SELECT convert(@str collate latin1_german2_ci using utf8);
SELECT convert(@str collate latin1_swedish_ci using utf8);

# End of 4.1 tests

SET NAMES latin1;
--disable_warnings
DROP TABLE IF EXISTS `abcÿdef`;
--enable_warnings
CREATE TABLE `abcÿdef` (i int);
INSERT INTO `abcÿdef` VALUES (1);
INSERT INTO abcÿdef VALUES (2);
SELECT * FROM `abcÿdef`;
SELECT * FROM abcÿdef;
DROP TABLE `abcÿdef`;
#
# Bug#29499 Converting 'del' from ascii to Unicode results in 'question mark'
#
select hex(cast(_ascii 0x7f as char(1) character set latin1));

--echo End of 5.0 tests


--echo #
--echo # Bug#58022 ... like ... escape export_set ( ... ) crashes when export_set returns warnings
--echo #
SELECT '' LIKE '' ESCAPE EXPORT_SET(1, 1, 1, 1, '');

--echo End of 5.1 tests

                                                                               
--echo #
--echo # Start of 5.5 tests
--echo #

--source include/ctype_numconv.inc

--echo #
--echo # Bug#11764503 (Bug#57341) Query in EXPLAIN EXTENDED shows wrong characters
--echo #
# Test latin1 client erroneously started with --default-character-set=utf8
# EXPLAIN output should still be pretty readable.
# We're using 'ó' (\xC3\xB3) as a magic sequence:
# - it's "LATIN CAPITAL LETTER A WITH TILDE ABOVE + SUPERSCRIPT 3" in latin1
# - it's "LATIN SMALL LETTER O WITH ACUTE ABOVE" in utf8.
SET NAMES utf8;
EXPLAIN EXTENDED SELECT 'abcdó', _latin1'abcdó', _utf8'abcdó';
# Test normal latin1
SET NAMES latin1;
EXPLAIN EXTENDED SELECT 'abcdó', _latin1'abcdó', _utf8'abcdó';


--echo #
--echo # End of 5.5 tests
--echo #



--echo #
--echo # Start of 5.6 tests
--echo #

--echo #
--echo # WL#3664 WEIGHT_STRING
--echo #

set @@collation_connection=latin1_swedish_ci;
--source include/weight_string.inc
--source include/weight_string_euro.inc
--source include/weight_string_l1.inc

set @@collation_connection=latin1_bin;
--source include/weight_string.inc
--source include/weight_string_euro.inc
--source include/weight_string_l1.inc

set @@collation_connection=latin1_general_cs;
--source include/weight_string.inc
--source include/weight_string_euro.inc
--source include/weight_string_l1.inc

set @@collation_connection=binary;
--source include/weight_string.inc
--source include/weight_string_euro.inc
--source include/weight_string_l1.inc

set names latin1;

--echo #
--echo # End of 5.6 tests
--echo #

--echo #
--echo # Start of 10.0 tests
--echo #

let $ctype_unescape_combinations=selected;
--source include/ctype_unescape.inc

SET NAMES latin1;
--source include/ctype_like_cond_propagation.inc

SET NAMES latin1 COLLATE latin1_bin;
--source include/ctype_like_cond_propagation.inc


--echo #
--echo # MDEV-6752 Trailing incomplete characters are not replaced to question marks on conversion
--echo #
SET NAMES utf8, character_set_connection=latin1;
SELECT 'Â';
SELECT HEX('Â');
SELECT HEX(CAST('Â' AS CHAR CHARACTER SET utf8));
SELECT HEX(CAST('Â' AS CHAR CHARACTER SET latin1));
SELECT HEX(CONVERT('Â' USING utf8));
SELECT HEX(CONVERT('Â' USING latin1));
SELECT 'Âx';
SELECT HEX('Âx');
SELECT HEX(CAST('Âx' AS CHAR CHARACTER SET utf8));
SELECT HEX(CAST('Âx' AS CHAR CHARACTER SET latin1));
SELECT HEX(CONVERT('Âx' USING utf8));
SELECT HEX(CONVERT('Âx' USING latin1));
SET NAMES utf8;
CREATE TABLE t1 (a VARCHAR(10) CHARACTER SET latin1);
INSERT IGNORE INTO t1 VALUES ('Â'),('Â#');
SHOW WARNINGS;
SELECT HEX(a),a FROM t1;
DROP TABLE t1;

--echo #
--echo # MDEV-7629 Regression: Bit and hex string literals changed column names in 10.0.14
--echo #
SELECT _latin1 0x7E, _latin1 X'7E', _latin1 B'01111110';


--echo #
--echo # MDEV-9886 Illegal mix of collations with a view comparing a field to a binary constant
--echo #

SET NAMES latin1;
CREATE TABLE t1 (a TEXT CHARACTER SET latin1);
INSERT INTO t1 VALUES (0xC0);
SELECT a<>0xEE5D AS a FROM t1;
CREATE VIEW v1 AS SELECT a<>0xC0 AS a FROM t1;
SHOW CREATE VIEW v1;
SELECT * FROM v1;
DROP VIEW v1;
DROP TABLE t1;


--echo #
--echo # End of 10.0 tests
--echo #

--echo #
--echo # Start of 10.1 tests
--echo #

--echo #
--echo # MDEV-8661 Wrong result for SELECT..WHERE a='a' AND a='a' COLLATE latin1_bin
--echo #
SET NAMES latin1;
CREATE TABLE t1 (a CHAR(10));
INSERT INTO t1 VALUES ('a'),('A');
SELECT * FROM t1 WHERE a='a' AND a='a' COLLATE latin1_bin;
DROP TABLE t1;

CREATE TABLE t1 (a VARCHAR(10) CHARACTER SET latin1 COLLATE latin1_swedish_ci);
INSERT INTO t1 VALUES ('a'),('A');
SELECT * FROM t1 WHERE a='a' COLLATE latin1_bin AND a='A' COLLATE latin1_swedish_ci;
DROP TABLE t1;

--echo #
--echo # MDEV-8679 Equal field propagation is not used for VARCHAR when it safely could
--echo #
CREATE TABLE t1 (a VARCHAR(10));
INSERT INTO t1 VALUES ('10'),('11'),('12');
EXPLAIN EXTENDED SELECT * FROM t1 WHERE a='10' AND IF(a='10',1,0)=1;
EXPLAIN EXTENDED SELECT * FROM t1 WHERE a='10' AND CASE WHEN a='10' THEN 1 ELSE 0 END;
DROP TABLE t1;

--echo #
--echo # MDEV-8680 Wrong result for SELECT..WHERE a IN ('a' COLLATE latin1_bin,'b') AND a='a'
--echo #
SET NAMES latin1;
CREATE TABLE t1 (a VARCHAR(10));
INSERT INTO t1 VALUES ('a'),('A');
SELECT * FROM t1 WHERE a IN ('a' COLLATE latin1_bin,'b');
SELECT * FROM t1 WHERE a IN ('a' COLLATE latin1_bin,'b') AND a='a';
EXPLAIN EXTENDED SELECT * FROM t1 WHERE a IN ('a' COLLATE latin1_bin,'b') AND a='a';
DROP TABLE t1;

--echo #
--echo # MDEV-8698 Wrong result for SELECT..WHERE a BETWEEN 'a' AND 'c' COLLATE latin1_bin;
--echo #
SET NAMES latin1;
CREATE TABLE t1 (a VARCHAR(10));
INSERT INTO t1 VALUES ('a'),('A'),('b'),('B'),('c'),('C');
SELECT * FROM t1 WHERE a BETWEEN 'a' AND 'c' COLLATE latin1_bin;
SELECT * FROM t1 WHERE a BETWEEN 'a' AND 'c' COLLATE latin1_bin AND a='a';
EXPLAIN EXTENDED SELECT * FROM t1 WHERE a BETWEEN 'a' AND 'c' COLLATE latin1_bin AND a='a';
DROP TABLE t1;

--echo #
--echo # MDEV-8707 Wrong result for SELECT..WHERE varchar_column=DATE'2001-01-01' AND varchar_column='2001-01-01'
--echo #
SET NAMES latin1;
CREATE TABLE t1 (a VARCHAR(40));
INSERT INTO t1 VALUES ('2001-01-01'),('2001-01-01x');
SELECT * FROM t1 WHERE a=DATE'2001-01-01' AND a='2001-01-01';
SELECT * FROM t1 WHERE a='2001-01-01' AND a=DATE'2001-01-01';
SELECT * FROM t1 WHERE (a,a)=('2001-01-01x',DATE'2001-01-01');
SELECT * FROM t1 WHERE (a,a)=(DATE'2001-01-01','2001-01-01x');
SELECT * FROM t1 WHERE (a,a)=('2001-01-01',DATE'2001-01-01');
SELECT * FROM t1 WHERE (a,a)=(DATE'2001-01-01','2001-01-01');
DROP TABLE t1;
CREATE TABLE t1 (a ENUM('2001-01-01','2001-01-01x'));
INSERT INTO t1 VALUES ('2001-01-01'),('2001-01-01x');
SELECT * FROM t1 WHERE a=DATE'2001-01-01' AND a='2001-01-01';
SELECT * FROM t1 WHERE a='2001-01-01' AND a=DATE'2001-01-01';
SELECT * FROM t1 WHERE (a,a)=('2001-01-01x',DATE'2001-01-01');
SELECT * FROM t1 WHERE (a,a)=(DATE'2001-01-01','2001-01-01x');
SELECT * FROM t1 WHERE (a,a)=('2001-01-01',DATE'2001-01-01');
SELECT * FROM t1 WHERE (a,a)=(DATE'2001-01-01','2001-01-01');
DROP TABLE t1;
CREATE TABLE t1 (a VARCHAR(40),b VARCHAR(40));
INSERT INTO t1 VALUES ('2001-01-01','2001-01-01x');
SELECT * FROM t1 WHERE a=b AND a=DATE'2001-01-01';
DROP TABLE t1;
CREATE TABLE t1 (a ENUM('2001-01-01','2001-01-01x'),b ENUM('2001-01-01','2001-01-01x'));
INSERT INTO t1 VALUES ('2001-01-01','2001-01-01x');
SELECT * FROM t1 WHERE a=b AND a=DATE'2001-01-01';
DROP TABLE t1;

--echo #
--echo # MDEV-8695 Wrong result for SELECT..WHERE varchar_column='a' AND CRC32(varchar_column)=3904355907
--echo #
CREATE TABLE t1 (a VARCHAR(10) COLLATE latin1_bin);
INSERT INTO t1 VALUES ('a'),('a ');
SELECT a, LENGTH(a), CRC32(a) FROM t1 WHERE CRC32(a)=3904355907;
SELECT a, LENGTH(a), CRC32(a) FROM t1 WHERE a='a' AND CRC32(a)=3904355907;
--echo # 'a' should not be propagated into CRC32(a)
EXPLAIN EXTENDED SELECT a, LENGTH(a), CRC32(a) FROM t1 WHERE a='a' AND CRC32(a)=3904355907;
SELECT a, HEX(a) FROM t1 WHERE HEX(a)='61';
SELECT *,HEX(a) FROM t1 WHERE a='a' AND HEX(a)='61';
--echo # 'a' should not be propagated into HEX(a)
EXPLAIN EXTENDED SELECT *,HEX(a) FROM t1 WHERE a='a' AND HEX(a)='61';
SELECT * FROM t1 WHERE a='a';
SELECT * FROM t1 WHERE LENGTH(a)=2;
SELECT * FROM t1 WHERE a='a' AND LENGTH(a)=2;
--echo # 'a' should not be propagated into LENGTH(a)
EXPLAIN EXTENDED SELECT * FROM t1 WHERE a='a' AND LENGTH(a)=2;
DROP TABLE t1;

--echo #
--echo # MDEV-8712 Wrong result for SELECT..WHERE latin1_bin_column=_latin1'a' AND latin1_bin_column='A'
--echo #
SET NAMES latin1;
CREATE TABLE t1 (a VARCHAR(20) COLLATE latin1_bin);
INSERT INTO t1 VALUES ('a'),('b');
SELECT * FROM t1 WHERE a='A';
SELECT * FROM t1 WHERE a='A' AND a=_latin1'a';
SELECT * FROM t1 WHERE a=_latin1'a' AND a='A';
SELECT * FROM t1 WHERE a=_latin1'A';
SELECT * FROM t1 WHERE a=_latin1'A' AND a=_latin1'a';
SELECT * FROM t1 WHERE a=_latin1'a' AND a=_latin1'A';
DROP TABLE t1;

--echo #
--echo # MDEV-8253 EXPLAIN SELECT prints unexpected characters
--echo #
SET NAMES latin1;
CREATE TABLE t1 (a DECIMAL(10,1),b DECIMAL(10,1),c VARCHAR(10),d VARCHAR(10));
INSERT INTO t1 VALUES (1.5,1.5,'1','1'),(3.5,3.5,'3','3');
EXPLAIN EXTENDED
SELECT * FROM t1 WHERE COALESCE(c,0)='3 ' AND COALESCE(d,0)=COALESCE(c,0);
DROP TABLE t1;

--echo #
--echo # MDEV-15005 ASAN: stack-buffer-overflow in my_strnncollsp_simple
--echo #

SET NAMES latin1;
SELECT CONVERT(1, CHAR) IN ('100', 10, '101');
SELECT CONVERT(1, CHAR) IN ('100', 10, '1');
SELECT CONVERT(1, CHAR) IN ('100', '10', '1');

--echo #
--echo # End of 10.1 tests
--echo #

--echo #
--echo # Start of 10.2 tests
--echo #

--echo #
--echo # MDEV-9711 NO PAD Collatons
--echo #
SET character_set_connection=latin1;
let $coll='latin1_swedish_nopad_ci';
let $coll_pad='latin1_swedish_ci';
--source include/ctype_pad_all_engines.inc

let $coll='latin1_nopad_bin';
let $coll_pad='latin1_bin';
--source include/ctype_pad_all_engines.inc

SET NAMES latin1;
--source include/ctype_like_range_mdev14350.inc

--echo #
--echo # End of 10.2 tests
--echo #