Consider using utf8mb4 (utf8mb4_unicode_ci) in MySQL/MariaDB

[ghost]

For performance reasons (or maybe just by mistake) someone in MySQL decided to break UTF-8 and not support the whole 4-bytes, instead they used only 3-bytes and thus MySQL was notorious for not supporting many UTF-8 characters.

Time has passed since then and both MySQL and MariaDB support the full UTF-8 set. But because the charset name "utf8" implied the broken set, the correct set is instead named as "utf8mb4", so that now matches what the rest of the world knows as UTF-8.

Please consider converting your tables to utf8mb4 (collate utf8mb4_unicode_ci), so that the full UTF-8 set is supported.

I believe there was a similar discussion with issue #6535

Thank you.

[alecpl]

Yes, this is planned for 1.5.

[alecpl]

Done.

[ghost]

Thank you for the prompt response and fix.

I'd like add an important note for posterity reasons. In older MySQL/MariaDB 5.x versions, there is an issue with VARCHAR(255). A column defined as such, will break the conversion because of a limit in the index.

There are several solutions:

• upgrade to a newer MySQL/MariaDB version,
• increase the index limit
• alter the table to VARCHAR(190), if I remember correctly.