[MDEV-26113] Replace MariaDB Connector/C 3.1 by MariaDB Connector/C 3.2 for MariaDB Server 10.2+ Created: 2021-07-08  Updated: 2021-07-27  Resolved: 2021-07-27

Status: Closed
Project: MariaDB Server
Component/s: libmariadb
Fix Version/s: N/A

Type: Task Priority: Blocker
Reporter: Ralf Gebhardt Assignee: Georg Richter
Resolution: Won't Fix Votes: 0
Labels: None

Attachments: File MDEV-26113.patch    

 Description   

To be able to EOL MariaDB Connector/C 3.1 we need to verify if it can be replaced by 3.2 for MariaDB 10.2, 10.3, 10.4 and 10.5.
C/C 3.2 is already used by 10.6

 Comments   
Comment by Georg Richter [ 2021-07-12 ]

Patch attached

Comment by Sergei Golubchik [ 2021-07-12 ]

two questions:

 static void test_bug11766854()
 
 {
 
-  struct st_mysql_client_plugin *plugin;
 
+  struct st_mysql_client_plugin *plugin= NULL;

and in two places

-  DIE_UNLESS(strcmp(character_set_name_4, "utf8") == 0);
 
+  DIE_UNLESS(strcmp(character_set_name_4, "utf8mb3") == 0);

why is that?

Comment by Georg Richter [ 2021-07-13 ]

1) Right, not needed
2) Since C/C 3.2 mysql_character_set_name returns utf8mb3 instead of utf8 (MDEV-8334)

Comment by Sergei Golubchik [ 2021-07-19 ]
  1. There are two nearly identical functions, mysql_get_charset_by_nr and mysql_find_charset_nr. It'd be good to remove one implementation and let one of those functions call the other
  2. same for mysql_get_charset_by_name and mysql_find_charset_name
  3. doing utf8mb3 for all servers below 10.6 doesn't look like a good idea. What about doing this conditionally, depending on MARIADB_VERSION_ID ?
Comment by Sergei Golubchik [ 2021-07-27 ]

At the end the utf8 changes seem to be too big for such a change. We'll keep C/C 3.1 in old server releases.

Generated at Thu Feb 08 09:42:50 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.