[#MDEV-27042] UCA: Resetting contractions to ignorable does not work well

I patch the character set configuration file Index.xml as follows:

diff --git a/Index.xml.orig b/Index.xml

index cec3bfc..c69047e 100644

--- a/Index.xml.orig

+++ b/Index.xml

@@ -540,6 +540,19 @@ To make maintaining easier please:

     <flag>binary</flag>

     <flag>compiled</flag>

   </collation>

+  <collation name="utf8mb3_phone_ci" id="352">

+    <rules>

+      <reset>\u0000</reset>

+        <i>\u0020</i> <!-- space -->

+        <i>\u0028</i> <!-- left parenthesis -->

+        <i>\u0029</i> <!-- right parenthesis -->

+        <i>\u002B</i> <!-- plus -->

+        <i>\u002D</i> <!-- hyphen -->

+        <i>tel.</i>

+    </rules>

+  </collation>

 </charset>

 <charset name="ucs2">

I.e. I want to make ignorable:

Now I run this script:

CREATE OR REPLACE TABLE t1

  phone VARCHAR(64) CHARACTER SET utf8mb3 COLLATE utf8mb3_phone_ci

);

INSERT INTO t1 VALUES ('123'),('tel.123');

SELECT * FROM t1 WHERE phone='123';

+-------+

| phone |

+-------+

| 123   |

+-------+

Looks wrong. It should return both lines.

Now I run:

SELECT phone, HEX(WEIGHT_STRING(phone)) FROM t1;

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

| phone   | HEX(WEIGHT_STRING(phone)) |

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

| 123     | 0E2A0E2B0E2C              |

| tel.123 |                           |

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

It also looks wrong: the weight string in the second line should be equal to the weight string in the first line.

[MDEV-27042] UCA: Resetting contractions to ignorable does not work well Created: 2021-11-14 Updated: 2021-11-29 Resolved: 2021-11-24
Status:	Closed
Project:	MariaDB Server
Component/s:	Character Sets
Affects Version/s:	10.2, 10.3, 10.4, 10.5, 10.6, 10.7, 10.8
Fix Version/s:	10.8.0