XF 2.0 Would changing collation from utf8mb4_general_ci to a language specific one make problems?

[sbj]

I want to know if we can change the collation of our Xenforo database to utf8mb4_german_ci or utf8mb4_turkish_ci without having negative side effects?

This would be done mainly to make sure having an accent-sensitive database.

My knowledge is very limited in this kind of thing, so I am not sure if a conversion could be done and if so if it would solve the problem? And can we go back whenever we want to general_ci again?

 

[Xon]

This may affect username matching when doing user mentions, and you can probably get yourself in trouble when modifying template names. Otherwise it should be safe, but I'ld wait for an official XF answer

 

[sbj]

This may affect username matching when doing user mentions,

Can you give an example?

and you can probably get yourself in trouble when modifying template names.

Example?

Otherwise it should be safe, but I'ld wait for an official XF answer

Specially, what I want to know is if we can go back to general_ci again, if problems occur.
@Chris D @Mike

 

[Chris D]

We chose general_ci specifically because it solves a bunch of edge cases we came across when initially considering unicode_ci. Going back would be possible, but it may depend on whether you end up with data created that the other collation would consider duplicate.

If you make any changes, it would generally be against our advice and not something we would be able to provide support with. If you still want to consider it, I'd recommend an extended period of testing using the new collation on a test database.