Better performance for unique slug checking

[calvinl]

As of now, generating a unique slug means checking the slug against the database each time and then incrementing the count, checking again, and so on, until an unused slug is found. For slugs with a high count, this can mean checking against the database many times.

One method is to parse the suffix's count and increment once from there. However, this would not work with a custom generator (if #3 is implemented). Another way is to suffix with a UUID. A third way would be to allow both as default strategies for unique slugging, but this could have repercussions if the strategy is changed sometime during the lifetime of the application.

[calvinl]

Addressed in v3.0.0