Charset/collation on event stream tables

[darrylhein]

Is there a reason the event stream tables are utf8 instead utf8mb4?

And related, do they need to be case sensitive? ie, utf8_bin vs utf8mb4_unicode_ci (just curious)

[prolic]

utf8mb4 is a good idea for the stream-tables itself, but who uses emojies and stuff like this on the stream name? anyway, we can do this I think.

The utf8_bin collation compares strings based purely on their Unicode code point values, so it is the fastest because no Unicode normalization is applied to the strings. And who needs case insensitivity for stream names anyway?

[prolic]

@darrylhein can you check mariadb as well?