MDEV-32336 deb default config - use collation-server = utf8mb4_uca1400_ai_ci

[grooverdan]

• The Jira issue number for this PR is: MDEV-32336

Description

utf8mb4_general_ci has been outdated for a while. Lets use our modern standard.

How can this PR be tested?

my_print_defaults --mysqld on install in debian

Basing the PR against the correct MariaDB version

• This is a new feature and the PR is based against the latest MariaDB development branch.
• [] This is a bug fix and the PR is based against the earliest maintained branch in which the bug can be reproduced.

PR quality check

• I checked the CODING_STANDARDS.md file and my PR conforms to this where appropriate.
• For any trivial modifications to the PR, I am ok with the reviewer making the changes themselves.

[ottok]

The migration to utf-8 was done in Debian by having Debian defaults that override the upstream defaults (latin1) as they seem outdated and unsuitable for most users. According to https://mariadb.com/kb/en/server-system-variables/#character_set_server and https://mariadb.com/kb/en/server-system-variables/#collation_server this still seems to be the case.

Would it not be the best solution to simply remove these customizations in debian/ and default to upstream server defaults, and ensure upstream has modern and sensible values?

If changing upstream default is not an option, the change in debian/ should be properly documented. The Jira https://jira.mariadb.org/browse/MDEV-32336 does not explain why the new default should be specifically utf8mb4_uca1400_ai_ci nor does the commit message explain why this change should be done to this value:

MDEV-32336 deb default config - use collation-server =
 utf8mb4_uca1400_ai_ci

utf8mb4_general_ci has been outdated for a while. Lets use our modern
standard.

See tips 2 & 3 in about good commit message. The commit message should say something about why utf8mb4_uca1400_ai_ci is the best value and why this change is done now in 11.3 and not say in 11.4 as https://jira.mariadb.org/browse/MDEV-25829 says.

[grooverdan]

Good point on commit message. Later and actually a standard is the general criteria however I'll find some better words.

Looking back on why utf8mb4_general_ci, was copied from utf8 comment from 2012 in 438ed04 in and in 7c2079f was made non-commented, but updated to utf8mb4.

MDEV-25829, I'm not sure is a genuine target version of 11.4, but I hadn't looked for it either.

I'd certainly welcome better upstream defaults too.

11.3 was chosen as only non-packaged releases have been done on this so I was assuming its still compatible and not causing packaging regressions.

[illuusio]

What is the real difference between utf8mb4_uca1400_ai_ci and utf8mb4_general_ci. Is it just newer standard or all the pages are supported?

[grooverdan]

Mainly from https://stackoverflow.com/questions/766809/whats-the-difference-between-utf8-general-ci-and-utf8-unicode-ci - utf8mb4_general_ci isn't a standard, it was just a (slightly) quicker (and dirty) implementation.

[ottok]

You probably want to have the git commit message updated with the text you posted in PR comments here, and use as title something like "MDEV-32336: Use utf8mb4_uca1400_ai_ci as default collation in Debian"

[grooverdan]

Acceptable?

[illuusio]

Is that correct new notation that I don't know utf8mb4=uca1400_ai_ci. I suppose so. I have to admit look very weird but if it work then I'm cool with it.