|
UCA1400 collations (added by MDEV-27009) currently use Unicode-5.2.0 case folding tables.
They should use Unicode-14.0.0 tables instead.
The difference (see attached diff-520-1400.diff) between these two files:
shows that a few hundred new case folding mapping pairs where added in these letter scripts:
Cyrillic, Gergian, Cherokee, Glagolitic, Coptic, Latin, Osage, Vithkuqi, Old Hungarian, Warang Citi, Medefaidrin, Adlam.
This SQL script demonstrates the out-dated case folding:
CREATE OR REPLACE TABLE t1 (a VARCHAR(10) CHARACTER SET utf8 COLLATE uca1400_ai_ci);
|
# Insert letters appeared in Unicode-6.1 (released in January 2012)
|
INSERT INTO t1 VALUES (_ucs2 0xA792) /* U+A792 LATIN CAPITAL LETTER C WITH BAR */;
|
INSERT INTO t1 VALUES (_ucs2 0xA793) /* U+A793 LATIN SMALL LETTER C WITH BAR */;
|
SELECT HEX(a), HEX(LOWER(a)), HEX(UPPER(a)), a, LOWER(a), UPPER(a) FROM t1;
|
+--------+---------------+---------------+------+----------+----------+
|
| HEX(a) | HEX(LOWER(a)) | HEX(UPPER(a)) | a | LOWER(a) | UPPER(a) |
|
+--------+---------------+---------------+------+----------+----------+
|
| EA9E92 | EA9E92 | EA9E92 | Ꞓ | Ꞓ | Ꞓ |
|
| EA9E93 | EA9E93 | EA9E93 | ꞓ | ꞓ | ꞓ |
|
+--------+---------------+---------------+------+----------+----------+
|
The above two characters (first appeared in Unicode-6.1) are expected to map to each other by functions UPPER and LOWER.
|
diff-500-1400.diff