Database collations

[RichardAnderson]

Pulls the collation and charset information from the database on install.
Syncs the charset information to existing DBs
Provides the ability to specify the charset.
Partially implements synching of DBs.

• Test with MySQL
• Test with MariaDB
• Test with PostgreSQL

[saeedvaziry]

We need a rollback mechanism and let the user to retry if this fails

[saeedvaziry]

Please don't 😂

[saeedvaziry]

@RichardAnderson I added some unit tests. they might fail now but it is expected to pass

[slaFFik]

utf8mb3 is being phased out in MySQL, so it's better to switch to utf8mb4 for future compatibility.

Also, iirc emojis can't be stored correctly with utf8mb3, plus there are issues with some Chinese/Japanese/Korean characters.

[saeedvaziry]

@slaFFik this PR uses utf8mb3 as default choice already.

[RichardAnderson]

utf8mb3 is being phased out in MySQL, so it's better to switch to utf8mb4 for future compatibility.

Also, iirc emojis can't be stored correctly with utf8mb3, plus there are issues with some Chinese/Japanese/Korean characters.

This is true, however, we are attempting to support 3 versions of MySQL, so uft8mb4 is not supported on all of them. I guess for later version we should change the default collation, but at the moment, nothing in the service can identify the version, it's treated generically.