Inconsistent behavior on db:prepare task

[morgoth]

The db:prepare task works different for PG adapter than SQLite3 one.

db:prepare task introduced in #35768 in case of missing database is supposed to create it, load schema and seeds.

This is working fine on PostgreSQL database, however when using SQLite3 database, seeds are not loaded.
This is happening, because db:migrate does not raise ActiveRecord::NoDatabaseError which db:prepare relies on

rails/activerecord/lib/active_record/railties/databases.rake

Line 272 in eece0bf

rescue ActiveRecord::NoDatabaseError

Not sure what the proper fix would be. Raising ActiveRecord::NoDatabaseError on SQLite3 would be backward incompatible change, so maybe we should add some internal method for existence check of db, that could be used instead of relying on raising error?
/cc @robertomiranda

[robertomiranda]

@morgoth thanks for raising up this issue, I've been thinking about this and probably a db_exists? method on the different adapters will work better.

[itsWill]

The source of this issue is the same as the one for this one: #32914

The fact that Sqlite will silently create a new database instead of throwing an error upon connecting to a non existing database.

For what its worth, on my approach to solve the above issue I used the db_exists? method @robertomiranda mentioned above: #36025 and I think it worked quite well.

Note: my PR doesn't solve this issue, but having the db_exists? method across adapters is a good step to solving both.

[rails-bot]

This issue has been automatically marked as stale because it has not been commented on for at least three months.
The resources of the Rails team are limited, and so we are asking for your help.
If you can still reproduce this error on the 6-0-stable branch or on master, please reply with all of the information you have about it in order to keep the issue open.
Thank you for all your contributions.

[morgoth]

it's still an issue being worked on #36620 but got a little stale there ;)

[rails-bot]

This issue has been automatically marked as stale because it has not been commented on for at least three months.
The resources of the Rails team are limited, and so we are asking for your help.
If you can still reproduce this error on the 6-0-stable branch or on master, please reply with all of the information you have about it in order to keep the issue open.
Thank you for all your contributions.