Thanks @b_b. I went ahead and tested it.
SPIP 4.4.25 is working successfully with MySQL 8.4.11 in my test environment.
My test configuration:
- SPIP 4.4.25
- PHP 8.5.11
- MySQL 8.4.11
- mysqli/mysqlnd
- MySQL authentication: caching_sha2_password
- Apache 2.4.68
Instead of testing with a new empty installation, I used a local copy of a real production SPIP site with its production-derived database containing 84 tables, as well as its actual templates, plugins, IMG directory and content.
I successfully tested:
- homepage, articles, rubriques, authors and comments
- SPIP pretty URLs
- admin authentication and the private area
- the private interface in both English and Persian, including switching between the two
- editing an existing article
- creating and publishing new articles
- assigning authors and keywords
- uploading article images/logos
- GD/WebP image processing
- multilingual content, including creating an English test article and associating it as the translation of a Persian test article
So I have tested both database reads and writes successfully against MySQL 8.4.11.
During preparation of the local environment, I encountered some issues related to the test environment itself, including cache directories, the PHP memory limit, Apache URL rewriting and GD WebP support. Once those were configured correctly, the problems disappeared. So far, I haven’t encountered a problem that appears to be caused by MySQL 8.4.
For transparency, I actively used AI assistance to help set up, troubleshoot and run this test environment and to structure the compatibility checks. The results reported here, however, are based on the actual behavior I observed from SPIP 4.4.25 running against MySQL 8.4.11, including successful reads, writes, publishing, image processing and multilingual article operations.
Therefore, my result so far is:
SPIP 4.4.25 + PHP 8.5.11 + MySQL 8.4.11: working successfully in this production-derived test environment.
I am nevertheless going to be conservative before upgrading the production MySQL server. I want first to investigate why the local SPIP copy is noticeably slower than the production site and test the other applications that use the same MySQL server. I also have WordPress sites and a Django application on that server, so I want to perform representative tests of those applications as well.
This is especially important in my case because I do not want to treat the production upgrade from MySQL 8.0 to 8.4 as something that I can simply undo if an unexpected application problem appears afterward.
I also noticed a few aspects of the SPIP private-interface design, particularly around some multilingual controls, that I think could be clearer or more intuitive. That is unrelated to MySQL 8.4 compatibility, however, so I think it is better discussed separately another time.
Thanks for encouraging me to test this. I hope these results are useful when considering whether the documented MySQL compatibility range can be updated.