Gitlab::CurrentSettings will create ApplicationSetting.current if not
present which means we don't have to use `&.`. We can also more easily
use stub_application_setting in specs
This replaces the use of fake_application_settings with
`::ApplicationSetting.build`_from_defaults. The reason is that
`fake_application_settings` doesn't have the custom accessors that
`ApplicationSetting` has, e.g. `#commit_email_hostname`, thus this
can lead to unexpected `nil` values which comes from the database
column instead of `.default_commit_email_hostname` returned by
`ApplicationSetting#commit_email_hostname`.
Using `::ApplicationSetting.build_from_defaults` should be safe as it
doesn't try to `INSERT` a DB record, in contrary to
`::ApplicationSetting.create_from_defaults` which we used to use, and
which created issues that the introduction of
`fake_application_settings` tried to resolve (575dced5).
Signed-off-by: Rémy Coutable <remy@rymai.me>
master was failing because `ApplicationSetting.create_from_defaults` attempted
to write to a column that did not exist in the database. This occurred in a
`rake db:migrate` task, which was unable to perform the migration that would
have added the missing column in the first place.
In 9.3 RC2, we also had a bug where password sign-ins were disabled because
there were many pending migrations. The problem occurred because
`fake_application_settings` was being returned with an OpenStruct that did not
include the predicate method `signup_enabled?`. As a result, the value would
erroneously return `nil` instead of `true`. This commit uses the values of the
defaults to mimic this behavior.
This commit also refactors some of the logic to be clearer.
If Redis were not running or USE_DB were set to false, the
application settings retrieval would fail completely. This
change only attempts to use the cache if the system actually
wants to connect to the DB and rescues any failures in talking to
Redis.
Closes#17557
