Kudos to @zzak for taking over SDoc and make it work with RDoc 4.

Currently Active Record can be configured via the environment variable `DATABASE_URL` or by manually injecting a hash of values which is what Rails does, reading in `database.yml` and setting Active Record appropriately. Active Record expects to be able to use `DATABASE_URL` without the use of Rails, and we cannot rip out this functionality without deprecating. This presents a problem though when both config is set, and a `DATABASE_URL` is present. Currently the `DATABASE_URL` should "win" and none of the values in `database.yml` are used. This is somewhat unexpected to me if I were to set values such as `pool` in the `production:` group of `database.yml` they are ignored.

There are many ways that active record initiates a connection today:

- Stand Alone (without rails)
  - `rake db:<tasks>`
  - ActiveRecord.establish_connection
 
- With Rails
  - `rake db:<tasks>`
  - `rails <server> | <console>`
  - `rails dbconsole`


We should make all of these behave exactly the same way. The best way to do this is to put all of this logic in one place so it is guaranteed to be used.

Here is my prosed matrix of how this behavior should work:

```
No database.yml
No DATABASE_URL
=> Error
```

```
database.yml present
No DATABASE_URL
=> Use database.yml configuration
```

```
No database.yml
DATABASE_URL present
=> use DATABASE_URL configuration
```

```
database.yml present
DATABASE_URL present
=> Merged into `url` sub key. If both specify `url` sub key, the `database.yml` `url`
   sub key "wins". If other paramaters `adapter` or `database` are specified in YAML,
   they are discarded as the `url` sub key "wins".
```

### Implementation

Current implementation uses `ActiveRecord::Base.configurations` to resolve and merge all connection information before returning. This is achieved through a utility class: `ActiveRecord::ConnectionHandling::MergeAndResolveDefaultUrlConfig`.

To understand the exact behavior of this class, it is best to review the behavior in activerecord/test/cases/connection_adapters/connection_handler_test.rb though it should match the above proposal.

* Test for: Invalid plugin name, because of reserved rails word.
* Test for: Invalid plugin name because of constant name that is already in use.

Only config.autoload_paths is frozen, so add the preview_path
to ActiveSupport::Dependencies.autoload_paths directly in an
after_initialize block. Also protect against a blank preview_path
being added to autoload_paths which can cause a serious slowdown
as Dir[] tries to load all *_preview.rb files under /

Fixes #13372

See https://github.com/rails/rails/pull/13463#issuecomment-31480799 for full conversation.

* Move check from generated helper to test_help.rb, so that all
  applications can benefit
* Rather than just raising when the test schema has pending migrations,
  try to load in the schema and only raise if there are pending
  migrations afterwards
* Opt out of the check by setting
  config.active_record.maintain_test_schema = false
* Deprecate db:test:* tasks. The test helper is now fully responsible
  for maintaining the test schema, so we don't need rake tasks for this.
  This is also a speed improvement since we're no longer reloading the
  test database on every call to "rake test".

By using the URL sub key in the `database.yml` by default we are exposing the ability to set other attributes such as `pool` or `reap_frequency` without need of modifying the URL to contain non-connection specific information.

Right now `db:drop` depends on `load_config` since so when `db:drop` gets executed `load_config` gets run. `db:structure:load` depends on `[:environment, :load_config]`. So before it runs, it executes `environment` but because `load_config` has already executed it is skipped. Note `db:load_config` is "invoke"-d twice, but only "execute"-d once:

```
** Invoke db:drop (first_time)
** Invoke db:load_config (first_time)
** Execute db:load_config
** Execute db:drop
** Invoke db:structure:load (first_time)
** Invoke environment (first_time)
** Execute environment
** Invoke db:load_config
** Execute db:structure:load
```

The fix for this is making sure that the environment is run before any `load_config`:

```
** Invoke environment (first_time)
** Execute environment
** Invoke db:drop (first_time)
** Invoke db:load_config (first_time)
** Execute db:load_config
** Execute db:drop
** Invoke db:structure:load (first_time)
** Invoke environment
** Invoke db:load_config
** Execute db:structure:load
```

Prior to #13463 when `DATABASE_URL` was set, Rails automagically used that value instead of the database.yml. There are tests in dbs_test that expect this to still be true. After that PR, `RAILS_DATABASE_URL` is expected to be read into the YAML file via ERB, this PR fixes that behavior.

Note: this does not entirely fix the tests. It seems that `ActiveRecord::Tasks::DatabaseTasks.current_config` does not process the url string correctly (convert it into a hash), and ` ActiveRecord::Tasks::DatabaseTasks.structure_load(current_config, filename)` as well as other methods in `DatabaseTasks` expect a hash.

It seems like we should involve the resolver somewhere in this process to correctly convert the database url, I do not know the best place for that /cc @josevalim

A better solution has been pushed to master.

This reverts commit 959cfcef.

Tests are failing due to missing env var on master https://travis-ci.org/rails/rails/jobs/15930622#L641

This adds an environment variable `ENV['RAILS_SECRET_KEY_BASE']` so these tests will pass.

The build is broken: https://travis-ci.org/rails/rails/builds/15824530

This commit fixes it.

The problem: Sqlite expects the `database` part to be an absolute path. That prompted this change to be committed to master: https://github.com/rails/rails/commit/fbb79b517f3127ba620fedd01849f9628b78d6ce

This change provides correct behavior. Unfortunately tests were introduced in https://github.com/rails/rails/commit/971d5107cd4cd08c22a85d34546f4ba03ed5c925 that were relying on the incorrect behavior. We can avoid the fix by changing to another database url such as `mysql` or `postgresql`

In addition to fixing the failure, the assertions are changed so that the "expected" value comes before "actual" value.

Ref: https://github.com/seattlerb/minitest/blob/master/History.txt

We will default this option to true from now on to ensure users properly
handle their list of available locales whenever necessary. This option
was added as a security measure and thus Rails will follow it defaulting
to secure option.

Also improve the handling of I18n config options in its railtie, taking
the new enforce_available_locales option into account, by setting it as
the last one in the process. This ensures no other configuration will
trigger a deprecation warning due to that setting.

This will avoid false positives where caching? is nil and should be
false

before the application

This test is a regression test to
b068e20b that changed the load order of
Action View initializers.