activerecord 5.2.7 → 6.0.6.1

data/CHANGELOG.md

@@ -1,1046 +1,1449 @@
-## Rails 5.2.7 (March 10, 2022) ##
+## Rails 6.0.6.1 (January 17, 2023) ##
 
-*   No changes.
+*   Make `sanitize_as_sql_comment` more strict
 
+    Though this method was likely never meant to take user input, it was
+    attempting sanitization. That sanitization could be bypassed with
+    carefully crafted input.
 
-## Rails 5.2.6.3 (March 08, 2022) ##
+    This commit makes the sanitization more robust by replacing any
+    occurrances of "/*" or "*/" with "/ *" or "* /". It also performs a
+    first pass to remove one surrounding comment to avoid compatibility
+    issues for users relying on the existing removal.
 
-*   No changes.
+    This also clarifies in the documentation of annotate that it should not
+    be provided user input.
 
+    [CVE-2023-22794]
 
-## Rails 5.2.6.2 (February 11, 2022) ##
 
-*   No changes.
+## Rails 6.0.6 (September 09, 2022) ##
 
+*   Symbol is allowed by default for YAML columns
 
-## Rails 5.2.6.1 (February 11, 2022) ##
+    *Étienne Barrié*
 
-*   No changes.
 
+## Rails 6.0.5.1 (July 12, 2022) ##
 
-## Rails 5.2.6 (May 05, 2021) ##
+*   Change ActiveRecord::Coders::YAMLColumn default to safe_load
 
-*   No changes.
+    This adds two new configuration options The configuration options are as
+    follows:
+    
+    * `config.active_storage.use_yaml_unsafe_load`
+    
+    When set to true, this configuration option tells Rails to use the old
+    "unsafe" YAML loading strategy, maintaining the existing behavior but leaving
+    the possible escalation vulnerability in place.  Setting this option to true
+    is *not* recommended, but can aid in upgrading.
+    
+    * `config.active_record.yaml_column_permitted_classes`
+    
+    The "safe YAML" loading method does not allow all classes to be deserialized
+    by default.  This option allows you to specify classes deemed "safe" in your
+    application.  For example, if your application uses Symbol and Time in
+    serialized data, you can add Symbol and Time to the allowed list as follows:
+    
+    ```
+    config.active_record.yaml_column_permitted_classes = [Symbol, Date, Time]
+    ```
+
+    [CVE-2022-32224]
 
 
-## Rails 5.2.5 (March 26, 2021) ##
+## Rails 6.0.5 (May 09, 2022) ##
 
 *   No changes.
 
 
-## Rails 5.2.4.6 (May 05, 2021) ##
+## Rails 6.0.4.8 (April 26, 2022) ##
 
 *   No changes.
 
 
-## Rails 5.2.4.5 (February 10, 2021) ##
+## Rails 6.0.4.7 (March 08, 2022) ##
 
-*   Fix possible DoS vector in PostgreSQL money type
+*   No changes.
 
-    Carefully crafted input can cause a DoS via the regular expressions used
-    for validating the money format in the PostgreSQL adapter.  This patch
-    fixes the regexp.
 
-    Thanks to @dee-see from Hackerone for this patch!
+## Rails 6.0.4.6 (February 11, 2022) ##
 
-    [CVE-2021-22880]
+*   No changes.
 
-    *Aaron Patterson*
+
+## Rails 6.0.4.5 (February 11, 2022) ##
+
+*   No changes.
 
 
-## Rails 5.2.4.4 (September 09, 2020) ##
+## Rails 6.0.4.4 (December 15, 2021) ##
 
 *   No changes.
 
 
-## Rails 5.2.4.3 (May 18, 2020) ##
+## Rails 6.0.4.3 (December 14, 2021) ##
 
 *   No changes.
 
-## Rails 5.2.4.2 (March 19, 2020) ##
+
+## Rails 6.0.4.2 (December 14, 2021) ##
 
 *   No changes.
 
 
-## Rails 5.2.4.1 (December 18, 2019) ##
+## Rails 6.0.4.1 (August 19, 2021) ##
 
 *   No changes.
 
 
-## Rails 5.2.4 (November 27, 2019) ##
+## Rails 6.0.4 (June 15, 2021) ##
 
-*   Fix circular `autosave: true` causes invalid records to be saved.
+*   Only warn about negative enums if a positive form that would cause conflicts exists.
 
-    Prior to the fix, when there was a circular series of `autosave: true`
-    associations, the callback for a `has_many` association was run while
-    another instance of the same callback on the same association hadn't
-    finished running. When control returned to the first instance of the
-    callback, the instance variable had changed, and subsequent associated
-    records weren't saved correctly. Specifically, the ID field for the
-    `belongs_to` corresponding to the `has_many` was `nil`.
+    Fixes #39065.
 
-    Fixes #28080.
+    *Alex Ghiculescu*
 
-    *Larry Reid*
+*   Allow the inverse of a `has_one` association that was previously autosaved to be loaded.
 
-*   PostgreSQL: Fix GROUP BY with ORDER BY virtual count attribute.
+    Fixes #34255.
 
-    Fixes #36022.
+    *Steven Weber*
+
+*   Reset statement cache for association if `table_name` is changed.
+
+    Fixes #36453.
 
     *Ryuta Kamizono*
 
-*   Fix sqlite3 collation parsing when using decimal columns.
+*   Type cast extra select for eager loading.
 
-    *Martin R. Schuster*
+    *Ryuta Kamizono*
 
-*   Make ActiveRecord `ConnectionPool.connections` method thread-safe.
+*   Prevent collection associations from being autosaved multiple times.
 
-    Fixes #36465.
+    Fixes #39173.
 
-    *Jeff Doering*
+    *Eugene Kenny*
 
-*   Assign all attributes before calling `build` to ensure the child record is visible in
-    `before_add` and `after_add` callbacks for `has_many :through` associations.
+*   Resolve issue with insert_all unique_by option when used with expression index.
 
-    Fixes #33249.
+    When the `:unique_by` option of `ActiveRecord::Persistence.insert_all` and
+    `ActiveRecord::Persistence.upsert_all` was used with the name of an expression index, an error
+    was raised. Adding a guard around the formatting behavior for the `:unique_by` corrects this.
 
-    *Ryan H. Kerr*
+    Usage:
+
+    ```ruby
+    create_table :books, id: :integer, force: true do |t|
+      t.column :name, :string
+      t.index "lower(name)", unique: true
+    end
 
+    Book.insert_all [{ name: "MyTest" }], unique_by: :index_books_on_lower_name
+    ```
 
-## Rails 5.2.3 (March 27, 2019) ##
+    Fixes #39516.
 
-*   Fix different `count` calculation when using `size` with manual `select` with DISTINCT.
+    *Austen Madden*
 
-    Fixes #35214.
+*   Fix preloading for polymorphic association with custom scope.
 
-    *Juani Villarejo*
+    *Ryuta Kamizono*
 
-*   Fix prepared statements caching to be enabled even when query caching is enabled.
+*   Allow relations with different SQL comments in the `or` method.
+
+    *Takumi Shotoku*
+
+*   Resolve conflict between counter cache and optimistic locking.
+
+    Bump an Active Record instance's lock version after updating its counter
+    cache. This avoids raising an unnecessary `ActiveRecord::StaleObjectError`
+    upon subsequent transactions by maintaining parity with the corresponding
+    database record's `lock_version` column.
+
+    Fixes #16449.
+
+    *Aaron Lipman*
+
+*   Fix through association with source/through scope which has joins.
 
     *Ryuta Kamizono*
 
-*   Don't allow `where` with invalid value matches to nil values.
+*   Fix through association to respect source scope for includes/preload.
 
-    Fixes #33624.
+    *Ryuta Kamizono*
+
+*   Fix eager load with Arel joins to maintain the original joins order.
 
     *Ryuta Kamizono*
 
-*   Restore an ability that class level `update` without giving ids.
+*   Fix group by count with eager loading + order + limit/offset.
+
+    *Ryuta Kamizono*
 
-    Fixes #34743.
+*   Fix left joins order when merging multiple left joins from different associations.
 
     *Ryuta Kamizono*
 
-*   Fix join table column quoting with SQLite.
+*   Fix index creation to preserve index comment in bulk change table on MySQL.
 
-    *Gannon McGibbon*
+    *Ryuta Kamizono*
 
-*   Ensure that `delete_all` on collection proxy returns affected count.
+*   Change `remove_foreign_key` to not check `:validate` option if database
+    doesn't support the feature.
 
     *Ryuta Kamizono*
 
-*   Reset scope after delete on collection association to clear stale offsets of removed records.
+*   Fix the result of aggregations to maintain duplicated "group by" fields.
 
-    *Gannon McGibbon*
+    *Ryuta Kamizono*
+
+*   Do not return duplicated records when using preload.
+
+    *Bogdan Gusiev*
 
 
-## Rails 5.2.2.1 (March 11, 2019) ##
+## Rails 6.0.3.7 (May 05, 2021) ##
 
 *   No changes.
 
 
-## Rails 5.2.2 (December 04, 2018) ##
+## Rails 6.0.3.6 (March 26, 2021) ##
 
-*   Do not ignore the scoping with query methods in the scope block.
+*   No changes.
 
-    *Ryuta Kamizono*
 
-*   Allow aliased attributes to be used in `#update_columns` and `#update`.
+## Rails 6.0.3.5 (February 10, 2021) ##
 
-    *Gannon McGibbon*
+*   Fix possible DoS vector in PostgreSQL money type
 
-*   Allow spaces in postgres table names.
+    Carefully crafted input can cause a DoS via the regular expressions used
+    for validating the money format in the PostgreSQL adapter.  This patch
+    fixes the regexp.
 
-    Fixes issue where "user post" is misinterpreted as "\"user\".\"post\"" when quoting table names with the postgres
-    adapter.
+    Thanks to @dee-see from Hackerone for this patch!
 
-    *Gannon McGibbon*
+    [CVE-2021-22880]
 
-*   Cached columns_hash fields should be excluded from ResultSet#column_types
+    *Aaron Patterson*
 
-    PR #34528 addresses the inconsistent behaviour when attribute is defined for an ignored column. The following test
-    was passing for SQLite and MySQL, but failed for PostgreSQL:
 
-    ```ruby
-    class DeveloperName < ActiveRecord::Type::String
-      def deserialize(value)
-        "Developer: #{value}"
-      end
-    end
+## Rails 6.0.3.4 (October 07, 2020) ##
 
-    class AttributedDeveloper < ActiveRecord::Base
-      self.table_name = "developers"
+*   No changes.
 
-      attribute :name, DeveloperName.new
 
-      self.ignored_columns += ["name"]
-    end
+## Rails 6.0.3.3 (September 09, 2020) ##
 
-    developer = AttributedDeveloper.create
-    developer.update_column :name, "name"
+*   No changes.
 
-    loaded_developer = AttributedDeveloper.where(id: developer.id).select("*").first
-    puts loaded_developer.name # should be "Developer: name" but it's just "name"
-    ```
 
-    *Dmitry Tsepelev*
+## Rails 6.0.3.2 (June 17, 2020) ##
 
-*   Values of enum are frozen, raising an error when attempting to modify them.
+*   No changes.
 
-    *Emmanuel Byrd*
 
-*   `update_columns` now correctly raises `ActiveModel::MissingAttributeError`
-    if the attribute does not exist.
+## Rails 6.0.3.1 (May 18, 2020) ##
 
-    *Sean Griffin*
+*   No changes.
 
-*   Do not use prepared statement in queries that have a large number of binds.
 
-    *Ryuta Kamizono*
+## Rails 6.0.3 (May 06, 2020) ##
 
-*   Fix query cache to load before first request.
+*   Recommend applications don't use the `database` kwarg in `connected_to`
+
+    The database kwarg in `connected_to` was meant to be used for one-off scripts but is often used in requests. This is really dangerous because it re-establishes a connection every time. It's deprecated in 6.1 and will be removed in 6.2 without replacement. This change soft deprecates it in 6.0 by removing documentation.
 
     *Eileen M. Uchitelle*
 
-*   Fix collection cache key with limit and custom select to avoid ambiguous timestamp column error.
+*   Fix support for PostgreSQL 11+ partitioned indexes.
 
-    Fixes #33056.
+    *Sebastián Palma*
 
-    *Federico Martinez*
+*   Add support for beginless ranges, introduced in Ruby 2.7.
 
-*   Fix duplicated record creation when using nested attributes with `create_with`.
+    *Josh Goodall*
 
-    *Darwin Wu*
+*   Fix insert_all with enum values
 
-*   Fix regression setting children record in parent `before_save` callback.
+    Fixes #38716.
 
-    *Guo Xiang Tan*
+    *Joel Blum*
 
-*   Prevent leaking of user's DB credentials on `rails db:create` failure.
+*   Regexp-escape table name for MS SQL
 
-    *bogdanvlviv*
+    Add `Regexp.escape` to one method in ActiveRecord, so that table names with regular expression characters in them work as expected. Since MS SQL Server uses "[" and "]" to quote table and column names, and those characters are regular expression characters, methods like `pluck` and `select` fail in certain cases when used with the MS SQL Server adapter.
 
-*   Clear mutation tracker before continuing the around callbacks.
+    *Larry Reid*
 
-    *Yuya Tanaka*
+*   Store advisory locks on their own named connection.
 
-*   Prevent deadlocks when waiting for connection from pool.
+    Previously advisory locks were taken out against a connection when a migration started. This works fine in single database applications but doesn't work well when migrations need to open new connections which results in the lock getting dropped.
 
-    *Brent Wheeldon*
+    In order to fix this we are storing the advisory lock on a new connection with the connection specification name `AdisoryLockBase`. The caveat is that we need to maintain at least 2 connections to a database while migrations are running in order to do this.
 
-*   Avoid extra scoping when using `Relation#update` that was causing this method to change the current scope.
+    *Eileen M. Uchitelle*, *John Crepezzi*
 
-    *Ryuta Kamizono*
+*   Ensure `:reading` connections always raise if a write is attempted.
 
-*   Fix numericality validator not to be affected by custom getter.
+    Now Rails will raise an `ActiveRecord::ReadOnlyError` if any connection on the reading handler attempts to make a write. If your reading role needs to write you should name the role something other than `:reading`.
 
-    *Ryuta Kamizono*
+    *Eileen M. Uchitelle*
 
-*   Fix bulk change table ignores comment option on PostgreSQL.
+*   Enforce fresh ETag header after a collection's contents change by adding
+    ActiveRecord::Relation#cache_key_with_version. This method will be used by
+    ActionController::ConditionalGet to ensure that when collection cache versioning
+    is enabled, requests using ConditionalGet don't return the same ETag header
+    after a collection is modified. Fixes #38078.
 
-    *Yoshiyuki Kinjo*
+    *Aaron Lipman*
 
+*   A database URL can now contain a querystring value that contains an equal sign. This is needed to support passing PostgresSQL `options`.
 
-## Rails 5.2.1.1 (November 27, 2018) ##
+     *Joshua Flanagan*
 
-*   No changes.
+*   Retain explicit selections on the base model after applying `includes` and `joins`.
 
+    Resolves #34889.
 
-## Rails 5.2.1 (August 07, 2018) ##
+    *Patrick Rebsch*
 
-*   PostgreSQL: Support new relkind for partitioned tables.
 
-    Fixes #33008.
+## Rails 6.0.2.2 (March 19, 2020) ##
 
-    *Yannick Schutz*
+*   No changes.
 
-*   Rollback parent transaction when children fails to update.
 
-    *Guillaume Malette*
+## Rails 6.0.2.1 (December 18, 2019) ##
 
-*   Fix default value for MySQL time types with specified precision.
+*   No changes.
 
-    *Nikolay Kondratyev*
 
-*   Fix `touch` option to behave consistently with `Persistence#touch` method.
+## Rails 6.0.2 (December 13, 2019) ##
 
-    *Ryuta Kamizono*
+*   Share the same connection pool for primary and replica databases in the
+    transactional tests for the same database.
 
-*   Fix `save` in `after_create_commit` won't invoke extra `after_create_commit`.
+    *Edouard Chin*
 
-    Fixes #32831.
+*   Fix the preloader when one record is fetched using `after_initialize`
+    but not the entire collection.
 
-    *Ryuta Kamizono*
+    *Bradley Price*
 
-*   Fix logic on disabling commit callbacks so they are not called unexpectedly when errors occur.
+*   Fix collection callbacks not terminating when `:abort` is thrown.
 
-    *Brian Durand*
+    *Edouard Chin*, *Ryuta Kamizono*
 
-*   Fix parent record should not get saved with duplicate children records.
+*   Correctly deprecate `where.not` working as NOR for relations.
 
-    Fixes #32940.
+    12a9664 deprecated where.not working as NOR, however
+    doing a relation query like `where.not(relation: { ... })`
+    wouldn't be properly deprecated and `where.not` would work as
+    NAND instead.
 
-    *Santosh Wadghule*
+    *Edouard Chin*
+
+*   Fix `db:migrate` task with multiple databases to restore the connection
+    to the previous database.
+
+    The migrate task iterates and establish a connection over each db
+    resulting in the last one to be used by subsequent rake tasks.
+    We should reestablish a connection to the connection that was
+    established before the migrate tasks was run
 
-*   Fix that association's after_touch is not called with counter cache.
+    *Edouard Chin*
 
-    Fixes #31559.
+*   Fix multi-threaded issue for `AcceptanceValidator`.
 
     *Ryuta Kamizono*
 
-*   `becomes` should clear the mutation tracker which is created in `after_initialize`.
 
-    Fixes #32867.
+## Rails 6.0.1 (November 5, 2019) ##
 
-    *Ryuta Kamizono*
+*    Common Table Expressions are allowed on read-only connections.
 
-*   Allow a belonging to parent object to be created from a new record.
+     *Chris Morris*
 
-    *Jolyon Pawlyn*
+*    New record instantiation respects `unscope`.
 
-*   Fix that building record with assigning multiple has_one associations
-    wrongly persists through record.
+     *Ryuta Kamizono*
 
-    Fixes #32511.
+*    Fixed a case where `find_in_batches` could halt too early.
 
-    *Sam DeCesare*
+     *Takayuki Nakata*
 
-*   Fix relation merging when one of the relations is going to skip the
-    query cache.
+*    Autosaved associations always perform validations when a custom validation
+     context is used.
 
-    *James Williams*
+     *Tekin Suleyman*
 
+*    `sql.active_record` notifications now include the `:connection` in
+     their payloads.
 
-## Rails 5.2.0 (April 09, 2018) ##
+     *Eugene Kenny*
 
-*   MySQL: Support mysql2 0.5.x.
+*    A rollback encountered in an `after_commit` callback does not reset
+     previously-committed record state.
 
-    *Aaron Stone*
+     *Ryuta Kamizono*
 
-*   Apply time column precision on assignment.
+*    Fixed that join order was lost when eager-loading.
 
-    PR #20317 changed the behavior of datetime columns so that when they
-    have a specified precision then on assignment the value is rounded to
-    that precision. This behavior is now applied to time columns as well.
+     *Ryuta Kamizono*
 
-    Fixes #30301.
+*   `DESCRIBE` queries are allowed on read-only connections.
 
-    *Andrew White*
+    *Dylan Thacker-Smith*
 
-*   Normalize time column values for SQLite database.
+*   Fixed that records that had been `inspect`ed could not be marshaled.
 
-    For legacy reasons, time columns in SQLite are stored as full datetimes
-    because until #24542 the quoting for time columns didn't remove the date
-    component. To ensure that values are consistent we now normalize the
-    date component to 2001-01-01 on reading and writing.
+    *Eugene Kenny*
 
-    *Andrew White*
+*   The connection pool reaper thread is respawned in forked processes. This
+    fixes that idle connections in forked processes wouldn't be reaped.
 
-*   Ensure that the date component is removed when quoting times.
+    *John Hawthorn*
 
-    PR #24542 altered the quoting for time columns so that the date component
-    was removed however it only removed it when it was 2001-01-01. Now the
-    date component is removed irrespective of what the date is.
+*   The memoized result of `ActiveRecord::Relation#take` is properly cleared
+    when `ActiveRecord::Relation#reset` or `ActiveRecord::Relation#reload`
+    is called.
 
-    *Andrew White*
+    *Anmol Arora*
 
-*   Fix `dependent: :destroy` issue for has_one/belongs_to relationship where
-    the parent class was getting deleted when the child was not.
+*   Fixed the performance regression for `primary_keys` introduced MySQL 8.0.
 
-    Fixes #32022.
+    *Hiroyuki Ishii*
 
-    *Fernando Gorodscy*
+*   `insert`, `insert_all`, `upsert`, and `upsert_all` now clear the query cache.
 
-*   Whitelist `NULLS FIRST` and `NULLS LAST` in order clauses too.
+    *Eugene Kenny*
 
-    *Xavier Noria*
+*   Call `while_preventing_writes` directly from `connected_to`.
 
-*   Fix that after commit callbacks on update does not triggered when optimistic locking is enabled.
+    In some cases application authors want to use the database switching middleware and make explicit calls with `connected_to`. It's possible for an app to turn off writes and not turn them back on by the time we call `connected_to(role: :writing)`.
 
-    *Ryuta Kamizono*
+    This change allows apps to fix this by assuming if a role is writing we want to allow writes, except in the case it's explicitly turned off.
 
-*   Fix `#columns_for_distinct` of MySQL and PostgreSQL to make
-    `ActiveRecord::FinderMethods#limited_ids_for` use correct primary key values
-    even if `ORDER BY` columns include other table's primary key.
+    *Eileen M. Uchitelle*
+
+*   Improve detection of ActiveRecord::StatementTimeout with mysql2 adapter in the edge case when the query is terminated during filesort.
+
+    *Kir Shatrov*
 
-    Fixes #28364.
 
-    *Takumi Kagiyama*
+## Rails 6.0.0 (August 16, 2019) ##
 
-*   Make `reflection.klass` raise if `polymorphic?` not to be misused.
+*   Preserve user supplied joins order as much as possible.
 
-    Fixes #31876.
+    Fixes #36761, #34328, #24281, #12953.
 
     *Ryuta Kamizono*
 
-*   PostgreSQL: Allow pg-1.0 gem to be used with Active Record.
+*   Make the DATABASE_URL env variable only affect the primary connection. Add new env variables for multiple databases.
 
-    *Lars Kanis*
+    *John Crepezzi*, *Eileen Uchitelle*
 
-*   Deprecate `expand_hash_conditions_for_aggregates` without replacement.
-    Using a `Relation` for performing queries is the prefered API.
+*   Add a warning for enum elements with 'not_' prefix.
 
-    *Ryuta Kamizono*
+        class Foo
+          enum status: [:sent, :not_sent]
+        end
 
-*   Fix not expanded problem when passing an Array object as argument to the where method using `composed_of` column.
+    *Edu Depetris*
 
-    ```
-    david_balance = customers(:david).balance
-    Customer.where(balance: [david_balance]).to_sql
+*   Make currency symbols optional for money column type in PostgreSQL
 
-    # Before: WHERE `customers`.`balance` = NULL
-    # After : WHERE `customers`.`balance` = 50
-    ```
+    *Joel Schneider*
 
-    Fixes #31723.
 
-    *Yutaro Kanagawa*
+## Rails 6.0.0.rc2 (July 22, 2019) ##
 
-*   Fix `count(:all)` with eager loading and having an order other than the driving table.
+*   Add database_exists? method to connection adapters to check if a database exists.
 
-    Fixes #31783.
+    *Guilherme Mansur*
+
+*   PostgreSQL: Fix GROUP BY with ORDER BY virtual count attribute.
+
+    Fixes #36022.
 
     *Ryuta Kamizono*
 
-*   Clear the transaction state when an Active Record object is duped.
+*   Make ActiveRecord `ConnectionPool.connections` method thread-safe.
 
-    Fixes #31670.
+    Fixes #36465.
 
-    *Yuriy Ustushenko*
+    *Jeff Doering*
 
-*   Support for PostgreSQL foreign tables.
+*   Fix sqlite3 collation parsing when using decimal columns.
 
-    *fatkodima*
+    *Martin R. Schuster*
 
-*   Fix relation merger issue with `left_outer_joins`.
+*   Fix invalid schema when primary key column has a comment.
 
-    *Mehmet Emin İNAÇ*
+    Fixes #29966.
 
-*   Don't allow destroyed object mutation after `save` or `save!` is called.
+    *Guilherme Goettems Schneider*
 
-    *Ryuta Kamizono*
+*   Fix table comment also being applied to the primary key column.
 
-*   Take into account association conditions when deleting through records.
+    *Guilherme Goettems Schneider*
 
-    Fixes #18424.
+*   Fix merging left_joins to maintain its own `join_type` context.
 
-    *Piotr Jakubowski*
+    Fixes #36103.
 
-*   Fix nested `has_many :through` associations on unpersisted parent instances.
+    *Ryuta Kamizono*
 
-    For example, if you have
 
-        class Post < ActiveRecord::Base
-          belongs_to :author
-          has_many :books, through: :author
-          has_many :subscriptions, through: :books
-        end
+## Rails 6.0.0.rc1 (April 24, 2019) ##
 
-        class Author < ActiveRecord::Base
-          has_one :post
-          has_many :books
-          has_many :subscriptions, through: :books
-        end
+*   Add `touch` option to `has_one` association.
 
-        class Book < ActiveRecord::Base
-          belongs_to :author
-          has_many :subscriptions
-        end
+    *Abhay Nikam*
 
-        class Subscription < ActiveRecord::Base
-          belongs_to :book
-        end
+*   Deprecate `where.not` working as NOR and will be changed to NAND in Rails 6.1.
 
-    Before:
+    ```ruby
+    all = [treasures(:diamond), treasures(:sapphire), cars(:honda), treasures(:sapphire)]
+    assert_equal all, PriceEstimate.all.map(&:estimate_of)
+    ```
 
-    If `post` is not persisted, then `post.subscriptions` will be empty.
+    In Rails 6.0:
 
-    After:
+    ```ruby
+    sapphire = treasures(:sapphire)
+
+    nor = all.reject { |e|
+      e.estimate_of_type == sapphire.class.polymorphic_name
+    }.reject { |e|
+      e.estimate_of_id == sapphire.id
+    }
+    assert_equal [cars(:honda)], nor
+
+    without_sapphire = PriceEstimate.where.not(
+      estimate_of_type: sapphire.class.polymorphic_name, estimate_of_id: sapphire.id
+    )
+    assert_equal nor, without_sapphire.map(&:estimate_of)
+    ```
 
-    If `post` is not persisted, then `post.subscriptions` can be set and used
-    just like it would if `post` were persisted.
+    In Rails 6.1:
 
-    Fixes #16313.
+    ```ruby
+    sapphire = treasures(:sapphire)
 
-    *Zoltan Kiss*
+    nand = all - [sapphire]
+    assert_equal [treasures(:diamond), cars(:honda)], nand
 
-*   Fixed inconsistency with `first(n)` when used with `limit()`.
-    The `first(n)` finder now respects the `limit()`, making it consistent
-    with `relation.to_a.first(n)`, and also with the behavior of `last(n)`.
+    without_sapphire = PriceEstimate.where.not(
+      estimate_of_type: sapphire.class.polymorphic_name, estimate_of_id: sapphire.id
+    )
+    assert_equal nand, without_sapphire.map(&:estimate_of)
+    ```
 
-    Fixes #23979.
+    *Ryuta Kamizono*
 
-    *Brian Christian*
+*   Fix dirty tracking after rollback.
 
-*   Use `count(:all)` in `HasManyAssociation#count_records` to prevent invalid
-    SQL queries for association counting.
+    Fixes #15018, #30167, #33868.
 
-    *Klas Eskilson*
+    *Ryuta Kamizono*
 
-*   Fix to invoke callbacks when using `update_attribute`.
+*   Add `ActiveRecord::Relation#cache_version` to support recyclable cache keys via
+    the versioned entries in `ActiveSupport::Cache`. This also means that
+    `ActiveRecord::Relation#cache_key` will now return a stable key that does not
+    include the max timestamp or count any more.
 
-    *Mike Busch*
+    NOTE: This feature is turned off by default, and `cache_key` will still return
+    cache keys with timestamps until you set `ActiveRecord::Base.collection_cache_versioning = true`.
+    That's the setting for all new apps on Rails 6.0+
 
-*   Fix `count(:all)` to correctly work `distinct` with custom SELECT list.
+    *Lachlan Sylvester*
 
-    *Ryuta Kamizono*
+*   Fix dirty tracking for `touch` to track saved changes.
 
-*   Using subselect for `delete_all` with `limit` or `offset`.
+    Fixes #33429.
 
-    *Ryuta Kamizono*
+    *Ryuta Kamzono*
+
+*   `change_column_comment` and `change_table_comment` are invertible only if
+    `to` and `from` options are specified.
 
-*   Undefine attribute methods on descendants when resetting column
-    information.
+    *Yoshiyuki Kinjo*
 
-    *Chris Salzberg*
+*   Don't call commit/rollback callbacks when a record isn't saved.
 
-*   Log database query callers.
+    Fixes #29747.
 
-    Add `verbose_query_logs` configuration option to display the caller
-    of database queries in the log to facilitate N+1 query resolution
-    and other debugging.
+    *Ryuta Kamizono*
 
-    Enabled in development only for new and upgraded applications. Not
-    recommended for use in the production environment since it relies
-    on Ruby's `Kernel#caller_locations` which is fairly slow.
+*   Fix circular `autosave: true` causes invalid records to be saved.
 
-    *Olivier Lacan*
+    Prior to the fix, when there was a circular series of `autosave: true`
+    associations, the callback for a `has_many` association was run while
+    another instance of the same callback on the same association hadn't
+    finished running. When control returned to the first instance of the
+    callback, the instance variable had changed, and subsequent associated
+    records weren't saved correctly. Specifically, the ID field for the
+    `belongs_to` corresponding to the `has_many` was `nil`.
 
-*   Fix conflicts `counter_cache` with `touch: true` by optimistic locking.
+    Fixes #28080.
 
-    ```
-    # create_table :posts do |t|
-    #   t.integer :comments_count, default: 0
-    #   t.integer :lock_version
-    #   t.timestamps
-    # end
-    class Post < ApplicationRecord
-    end
+    *Larry Reid*
 
-    # create_table :comments do |t|
-    #   t.belongs_to :post
-    # end
-    class Comment < ApplicationRecord
-      belongs_to :post, touch: true, counter_cache: true
-    end
-    ```
+*   Raise `ArgumentError` for invalid `:limit` and `:precision` like as other options.
 
     Before:
-    ```
-    post = Post.create!
-    # => begin transaction
-         INSERT INTO "posts" ("created_at", "updated_at", "lock_version")
-         VALUES ("2017-12-11 21:27:11.387397", "2017-12-11 21:27:11.387397", 0)
-         commit transaction
-
-    comment = Comment.create!(post: post)
-    # => begin transaction
-         INSERT INTO "comments" ("post_id") VALUES (1)
-
-         UPDATE "posts" SET "comments_count" = COALESCE("comments_count", 0) + 1,
-         "lock_version" = COALESCE("lock_version", 0) + 1 WHERE "posts"."id" = 1
-
-         UPDATE "posts" SET "updated_at" = '2017-12-11 21:27:11.398330',
-         "lock_version" = 1 WHERE "posts"."id" = 1 AND "posts"."lock_version" = 0
-         rollback transaction
-    # => ActiveRecord::StaleObjectError: Attempted to touch a stale object: Post.
-
-    Comment.take.destroy!
-    # => begin transaction
-         DELETE FROM "comments" WHERE "comments"."id" = 1
-
-         UPDATE "posts" SET "comments_count" = COALESCE("comments_count", 0) - 1,
-         "lock_version" = COALESCE("lock_version", 0) + 1 WHERE "posts"."id" = 1
-
-         UPDATE "posts" SET "updated_at" = '2017-12-11 21:42:47.785901',
-         "lock_version" = 1 WHERE "posts"."id" = 1 AND "posts"."lock_version" = 0
-         rollback transaction
-    # => ActiveRecord::StaleObjectError: Attempted to touch a stale object: Post.
+
+    ```ruby
+    add_column :items, :attr1, :binary,   size: 10      # => ArgumentError
+    add_column :items, :attr2, :decimal,  scale: 10     # => ArgumentError
+    add_column :items, :attr3, :integer,  limit: 10     # => ActiveRecordError
+    add_column :items, :attr4, :datetime, precision: 10 # => ActiveRecordError
     ```
 
     After:
-    ```
-    post = Post.create!
-    # => begin transaction
-         INSERT INTO "posts" ("created_at", "updated_at", "lock_version")
-         VALUES ("2017-12-11 21:27:11.387397", "2017-12-11 21:27:11.387397", 0)
-         commit transaction
-
-    comment = Comment.create!(post: post)
-    # => begin transaction
-         INSERT INTO "comments" ("post_id") VALUES (1)
-
-         UPDATE "posts" SET "comments_count" = COALESCE("comments_count", 0) + 1,
-         "lock_version" = COALESCE("lock_version", 0) + 1,
-         "updated_at" = '2017-12-11 21:37:09.802642' WHERE "posts"."id" = 1
-         commit transaction
-
-    comment.destroy!
-    # => begin transaction
-         DELETE FROM "comments" WHERE "comments"."id" = 1
-
-         UPDATE "posts" SET "comments_count" = COALESCE("comments_count", 0) - 1,
-         "lock_version" = COALESCE("lock_version", 0) + 1,
-         "updated_at" = '2017-12-11 21:39:02.685520' WHERE "posts"."id" = 1
-         commit transaction
+
+    ```ruby
+    add_column :items, :attr1, :binary,   size: 10      # => ArgumentError
+    add_column :items, :attr2, :decimal,  scale: 10     # => ArgumentError
+    add_column :items, :attr3, :integer,  limit: 10     # => ArgumentError
+    add_column :items, :attr4, :datetime, precision: 10 # => ArgumentError
     ```
 
-    Fixes #31199.
+    *Ryuta Kamizono*
 
-    *bogdanvlviv*
+*   Association loading isn't to be affected by scoping consistently
+    whether preloaded / eager loaded or not, with the exception of `unscoped`.
 
-*   Add support for PostgreSQL operator classes to `add_index`.
+    Before:
 
-    Example:
+    ```ruby
+    Post.where("1=0").scoping do
+      Comment.find(1).post                   # => nil
+      Comment.preload(:post).find(1).post    # => #<Post id: 1, ...>
+      Comment.eager_load(:post).find(1).post # => #<Post id: 1, ...>
+    end
+    ```
+
+    After:
+
+    ```ruby
+    Post.where("1=0").scoping do
+      Comment.find(1).post                   # => #<Post id: 1, ...>
+      Comment.preload(:post).find(1).post    # => #<Post id: 1, ...>
+      Comment.eager_load(:post).find(1).post # => #<Post id: 1, ...>
+    end
+    ```
 
-        add_index :users, :name, using: :gist, opclass: { name: :gist_trgm_ops }
+    Fixes #34638, #35398.
 
-    *Greg Navis*
+    *Ryuta Kamizono*
 
-*   Don't allow scopes to be defined which conflict with instance methods on `Relation`.
+*   Add `rails db:prepare` to migrate or setup a database.
 
-    Fixes #31120.
+    Runs `db:migrate` if the database exists or `db:setup` if it doesn't.
 
-    *kinnrot*
+    *Roberto Miranda*
 
-*   Add new error class `QueryCanceled` which will be raised
-    when canceling statement due to user request.
+*   Add `after_save_commit` callback as shortcut for `after_commit :hook, on: [ :create, :update ]`.
 
-    *Ryuta Kamizono*
+    *DHH*
 
-*   Add `#up_only` to database migrations for code that is only relevant when
-    migrating up, e.g. populating a new column.
+*   Assign all attributes before calling `build` to ensure the child record is visible in
+    `before_add` and `after_add` callbacks for `has_many :through` associations.
 
-    *Rich Daley*
+    Fixes #33249.
 
-*   Require raw SQL fragments to be explicitly marked when used in
-    relation query methods.
+    *Ryan H. Kerr*
 
-    Before:
-    ```
-    Article.order("LENGTH(title)")
-    ```
+*   Add `ActiveRecord::Relation#extract_associated` for extracting associated records from a relation.
 
-    After:
     ```
-    Article.order(Arel.sql("LENGTH(title)"))
+    account.memberships.extract_associated(:user)
+    # => Returns collection of User records
     ```
 
-    This prevents SQL injection if applications use the [strongly
-    discouraged] form `Article.order(params[:my_order])`, under the
-    mistaken belief that only column names will be accepted.
+    *DHH*
+
+*   Add `ActiveRecord::Relation#annotate` for adding SQL comments to its queries.
 
-    Raw SQL strings will now cause a deprecation warning, which will
-    become an UnknownAttributeReference error in Rails 6.0. Applications
-    can opt in to the future behavior by setting `allow_unsafe_raw_sql`
-    to `:disabled`.
+    For example:
 
-    Common and judged-safe string values (such as simple column
-    references) are unaffected:
     ```
-    Article.order("title DESC")
+    Post.where(id: 123).annotate("this is a comment").to_sql
+    # SELECT "posts".* FROM "posts" WHERE "posts"."id" = 123 /* this is a comment */
     ```
 
-    *Ben Toews*
+    This can be useful in instrumentation or other analysis of issued queries.
 
-*   `update_all` will now pass its values to `Type#cast` before passing them to
-    `Type#serialize`. This means that `update_all(foo: 'true')` will properly
-    persist a boolean.
+    *Matt Yoho*
 
-    *Sean Griffin*
+*   Support Optimizer Hints.
 
-*   Add new error class `StatementTimeout` which will be raised
-    when statement timeout exceeded.
+    In most databases, a way to control the optimizer is by using optimizer hints,
+    which can be specified within individual statements.
 
-    *Ryuta Kamizono*
+    Example (for MySQL):
 
-*   Fix `bin/rails db:migrate` with specified `VERSION`.
-    `bin/rails db:migrate` with empty VERSION behaves as without `VERSION`.
-    Check a format of `VERSION`: Allow a migration version number
-    or name of a migration file. Raise error if format of `VERSION` is invalid.
-    Raise error if target migration doesn't exist.
+        Topic.optimizer_hints("MAX_EXECUTION_TIME(50000)", "NO_INDEX_MERGE(topics)")
+        # SELECT /*+ MAX_EXECUTION_TIME(50000) NO_INDEX_MERGE(topics) */ `topics`.* FROM `topics`
 
-    *bogdanvlviv*
+    Example (for PostgreSQL with pg_hint_plan):
 
-*   Fixed a bug where column orders for an index weren't written to
-    `db/schema.rb` when using the sqlite adapter.
+        Topic.optimizer_hints("SeqScan(topics)", "Parallel(topics 8)")
+        # SELECT /*+ SeqScan(topics) Parallel(topics 8) */ "topics".* FROM "topics"
 
-    Fixes #30902.
+    See also:
 
-    *Paul Kuruvilla*
+    * https://dev.mysql.com/doc/refman/8.0/en/optimizer-hints.html
+    * https://pghintplan.osdn.jp/pg_hint_plan.html
+    * https://docs.oracle.com/en/database/oracle/oracle-database/12.2/tgsql/influencing-the-optimizer.html
+    * https://docs.microsoft.com/en-us/sql/t-sql/queries/hints-transact-sql-query?view=sql-server-2017
+    * https://www.ibm.com/support/knowledgecenter/en/SSEPGG_11.1.0/com.ibm.db2.luw.admin.perf.doc/doc/c0070117.html
 
-*   Remove deprecated method `#sanitize_conditions`.
+    *Ryuta Kamizono*
 
-    *Rafael Mendonça França*
+*   Fix query attribute method on user-defined attribute to be aware of typecasted value.
 
-*   Remove deprecated method `#scope_chain`.
+    For example, the following code no longer return false as casted non-empty string:
 
-    *Rafael Mendonça França*
+    ```
+    class Post < ActiveRecord::Base
+      attribute :user_defined_text, :text
+    end
 
-*   Remove deprecated configuration `.error_on_ignored_order_or_limit`.
+    Post.new(user_defined_text: "false").user_defined_text? # => true
+    ```
 
-    *Rafael Mendonça França*
+    *Yuji Kamijima*
 
-*   Remove deprecated arguments from `#verify!`.
+*   Quote empty ranges like other empty enumerables.
 
-    *Rafael Mendonça França*
+    *Patrick Rebsch*
 
-*   Remove deprecated argument `name` from `#indexes`.
+*   Add `insert_all`/`insert_all!`/`upsert_all` methods to `ActiveRecord::Persistence`,
+    allowing bulk inserts akin to the bulk updates provided by `update_all` and
+    bulk deletes by `delete_all`.
 
-    *Rafael Mendonça França*
+    Supports skipping or upserting duplicates through the `ON CONFLICT` syntax
+    for PostgreSQL (9.5+) and SQLite (3.24+) and `ON DUPLICATE KEY UPDATE` syntax
+    for MySQL.
 
-*   Remove deprecated method `ActiveRecord::Migrator.schema_migrations_table_name`.
+    *Bob Lail*
 
-    *Rafael Mendonça França*
+*   Add `rails db:seed:replant` that truncates tables of each database
+    for current environment and loads the seeds.
 
-*   Remove deprecated method `supports_primary_key?`.
+    *bogdanvlviv*, *DHH*
 
-    *Rafael Mendonça França*
+*   Add `ActiveRecord::Base.connection.truncate` for SQLite3 adapter.
 
-*   Remove deprecated method `supports_migrations?`.
+    *bogdanvlviv*
 
-    *Rafael Mendonça França*
+*   Deprecate mismatched collation comparison for uniqueness validator.
 
-*   Remove deprecated methods `initialize_schema_migrations_table` and `initialize_internal_metadata_table`.
+    Uniqueness validator will no longer enforce case sensitive comparison in Rails 6.1.
+    To continue case sensitive comparison on the case insensitive column,
+    pass `case_sensitive: true` option explicitly to the uniqueness validator.
 
-    *Rafael Mendonça França*
+    *Ryuta Kamizono*
 
-*   Raises when calling `lock!` in a dirty record.
+*   Add `reselect` method. This is a short-hand for `unscope(:select).select(fields)`.
 
-    *Rafael Mendonça França*
+    Fixes #27340.
 
-*   Remove deprecated support to passing a class to `:class_name` on associations.
+    *Willian Gustavo Veiga*
 
-    *Rafael Mendonça França*
+*   Add negative scopes for all enum values.
 
-*   Remove deprecated argument `default` from `index_name_exists?`.
+    Example:
 
-    *Rafael Mendonça França*
+        class Post < ActiveRecord::Base
+          enum status: %i[ drafted active trashed ]
+        end
 
-*   Remove deprecated support to `quoted_id` when typecasting an Active Record object.
+        Post.not_drafted # => where.not(status: :drafted)
+        Post.not_active  # => where.not(status: :active)
+        Post.not_trashed # => where.not(status: :trashed)
 
-    *Rafael Mendonça França*
+    *DHH*
 
-*   Fix `bin/rails db:setup` and `bin/rails db:test:prepare` create wrong
-    ar_internal_metadata's data for a test database.
+*   Fix different `count` calculation when using `size` with manual `select` with DISTINCT.
+
+    Fixes #35214.
+
+    *Juani Villarejo*
+
+
+## Rails 6.0.0.beta3 (March 11, 2019) ##
+
+*   No changes.
+
+
+## Rails 6.0.0.beta2 (February 25, 2019) ##
+
+*   Fix prepared statements caching to be enabled even when query caching is enabled.
+
+    *Ryuta Kamizono*
+
+*   Ensure `update_all` series cares about optimistic locking.
+
+    *Ryuta Kamizono*
+
+*   Don't allow `where` with non numeric string matches to 0 values.
+
+    *Ryuta Kamizono*
+
+*   Introduce `ActiveRecord::Relation#destroy_by` and `ActiveRecord::Relation#delete_by`.
+
+    `destroy_by` allows relation to find all the records matching the condition and perform
+    `destroy_all` on the matched records.
+
+    Example:
+
+        Person.destroy_by(name: 'David')
+        Person.destroy_by(name: 'David', rating: 4)
+
+        david = Person.find_by(name: 'David')
+        david.posts.destroy_by(id: [1, 2, 3])
+
+    `delete_by` allows relation to find all the records matching the condition and perform
+    `delete_all` on the matched records.
+
+    Example:
+
+        Person.delete_by(name: 'David')
+        Person.delete_by(name: 'David', rating: 4)
+
+        david = Person.find_by(name: 'David')
+        david.posts.delete_by(id: [1, 2, 3])
+
+    *Abhay Nikam*
+
+*   Don't allow `where` with invalid value matches to nil values.
+
+    Fixes #33624.
+
+    *Ryuta Kamizono*
+
+*   SQLite3: Implement `add_foreign_key` and `remove_foreign_key`.
+
+    *Ryuta Kamizono*
+
+*   Deprecate using class level querying methods if the receiver scope
+    regarded as leaked. Use `klass.unscoped` to avoid the leaking scope.
+
+    *Ryuta Kamizono*
+
+*   Allow applications to automatically switch connections.
+
+    Adds a middleware and configuration options that can be used in your
+    application to automatically switch between the writing and reading
+    database connections.
+
+    `GET` and `HEAD` requests will read from the replica unless there was
+    a write in the last 2 seconds, otherwise they will read from the primary.
+    Non-get requests will always write to the primary. The middleware accepts
+    an argument for a Resolver class and an Operations class where you are able
+    to change how the auto-switcher works to be most beneficial for your
+    application.
+
+    To use the middleware in your application you can use the following
+    configuration options:
 
-    Before:
     ```
-    $ RAILS_ENV=test rails dbconsole
-    > SELECT * FROM ar_internal_metadata;
-    key|value|created_at|updated_at
-    environment|development|2017-09-11 23:14:10.815679|2017-09-11 23:14:10.815679
+    config.active_record.database_selector = { delay: 2.seconds }
+    config.active_record.database_resolver = ActiveRecord::Middleware::DatabaseSelector::Resolver
+    config.active_record.database_resolver_context = ActiveRecord::Middleware::DatabaseSelector::Resolver::Session
     ```
 
-    After:
+    To change the database selection strategy, pass a custom class to the
+    configuration options:
+
     ```
-    $ RAILS_ENV=test rails dbconsole
-    > SELECT * FROM ar_internal_metadata;
-    key|value|created_at|updated_at
-    environment|test|2017-09-11 23:14:10.815679|2017-09-11 23:14:10.815679
+    config.active_record.database_selector = { delay: 10.seconds }
+    config.active_record.database_resolver = MyResolver
+    config.active_record.database_resolver_context = MyResolver::MyCookies
     ```
 
-    Fixes #26731.
+    *Eileen M. Uchitelle*
 
-    *bogdanvlviv*
+*   MySQL: Support `:size` option to change text and blob size.
 
-*   Fix longer sequence name detection for serial columns.
+    *Ryuta Kamizono*
 
-    Fixes #28332.
+*   Make `t.timestamps` with precision by default.
 
     *Ryuta Kamizono*
 
-*   MySQL: Don't lose `auto_increment: true` in the `db/schema.rb`.
 
-    Fixes #30894.
+## Rails 6.0.0.beta1 (January 18, 2019) ##
+
+*   Remove deprecated `#set_state` from the transaction object.
+
+    *Rafael Mendonça França*
+
+*   Remove deprecated `#supports_statement_cache?` from the database adapters.
+
+    *Rafael Mendonça França*
+
+*   Remove deprecated `#insert_fixtures` from the database adapters.
+
+    *Rafael Mendonça França*
+
+*   Remove deprecated `ActiveRecord::ConnectionAdapters::SQLite3Adapter#valid_alter_table_type?`.
+
+    *Rafael Mendonça França*
+
+*   Do not allow passing the column name to `sum` when a block is passed.
+
+    *Rafael Mendonça França*
+
+*   Do not allow passing the column name to `count` when a block is passed.
+
+    *Rafael Mendonça França*
+
+*   Remove delegation of missing methods in a relation to arel.
+
+    *Rafael Mendonça França*
+
+*   Remove delegation of missing methods in a relation to private methods of the class.
+
+    *Rafael Mendonça França*
+
+*   Deprecate `config.active_record.sqlite3.represent_boolean_as_integer`.
+
+    *Rafael Mendonça França*
+
+*   Change `SQLite3Adapter` to always represent boolean values as integers.
+
+    *Rafael Mendonça França*
+
+*   Remove ability to specify a timestamp name for `#cache_key`.
+
+    *Rafael Mendonça França*
+
+*   Remove deprecated `ActiveRecord::Migrator.migrations_path=`.
+
+    *Rafael Mendonça França*
+
+*   Remove deprecated `expand_hash_conditions_for_aggregates`.
+
+    *Rafael Mendonça França*
+
+*   Set polymorphic type column to NULL on `dependent: :nullify` strategy.
+
+    On polymorphic associations both the foreign key and the foreign type columns will be set to NULL.
+
+    *Laerti Papa*
+
+*   Allow permitted instance of `ActionController::Parameters` as argument of `ActiveRecord::Relation#exists?`.
+
+    *Gannon McGibbon*
+
+*   Add support for endless ranges introduces in Ruby 2.6.
+
+    *Greg Navis*
+
+*   Deprecate passing `migrations_paths` to `connection.assume_migrated_upto_version`.
 
     *Ryuta Kamizono*
 
-*   Fix `COUNT(DISTINCT ...)` for `GROUP BY` with `ORDER BY` and `LIMIT`.
+*   MySQL: `ROW_FORMAT=DYNAMIC` create table option by default.
 
-    Fixes #30886.
+    Since MySQL 5.7.9, the `innodb_default_row_format` option defines the default row
+    format for InnoDB tables. The default setting is `DYNAMIC`.
+    The row format is required for indexing on `varchar(255)` with `utf8mb4` columns.
 
     *Ryuta Kamizono*
 
-*   PostgreSQL `tsrange` now preserves subsecond precision.
+*   Fix join table column quoting with SQLite.
 
-    PostgreSQL 9.1+ introduced range types, and Rails added support for using
-    this datatype in Active Record. However, the serialization of
-    `PostgreSQL::OID::Range` was incomplete, because it did not properly
-    cast the bounds that make up the range. This led to subseconds being
-    dropped in SQL commands:
+    *Gannon McGibbon*
 
-    Before:
+*   Allow disabling scopes generated by `ActiveRecord.enum`.
 
-        connection.type_cast(tsrange.serialize(range_value))
-        # => "[2010-01-01 13:30:00 UTC,2011-02-02 19:30:00 UTC)"
+    *Alfred Dominic*
 
-    Now:
+*   Ensure that `delete_all` on collection proxy returns affected count.
 
-        connection.type_cast(tsrange.serialize(range_value))
-        # => "[2010-01-01 13:30:00.670277,2011-02-02 19:30:00.745125)"
+    *Ryuta Kamizono*
 
-    *Thomas Cannon*
+*   Reset scope after delete on collection association to clear stale offsets of removed records.
 
-*   Passing a `Set` to `Relation#where` now behaves the same as passing an
-    array.
+    *Gannon McGibbon*
 
-    *Sean Griffin*
+*   Add the ability to prevent writes to a database for the duration of a block.
 
-*   Use given algorithm while removing index from database.
+    Allows the application to prevent writes to a database. This can be useful when
+    you're building out multiple databases and want to make sure you're not sending
+    writes when you want a read.
 
-    Fixes #24190.
+    If `while_preventing_writes` is called and the query is considered a write
+    query the database will raise an exception regardless of whether the database
+    user is able to write.
 
-    *Mehmet Emin İNAÇ*
+    This is not meant to be a catch-all for write queries but rather a way to enforce
+    read-only queries without opening a second connection. One purpose of this is to
+    catch accidental writes, not all writes.
 
-*   Update payload names for `sql.active_record` instrumentation to be
-    more descriptive.
+    *Eileen M. Uchitelle*
 
-    Fixes #30586.
+*   Allow aliased attributes to be used in `#update_columns` and `#update`.
 
-    *Jeremy Green*
+    *Gannon McGibbon*
 
-*   Add new error class `LockWaitTimeout` which will be raised
-    when lock wait timeout exceeded.
+*   Allow spaces in postgres table names.
 
-    *Gabriel Courtemanche*
+    Fixes issue where "user post" is misinterpreted as "\"user\".\"post\"" when quoting table names with the postgres adapter.
 
-*   Remove deprecated `#migration_keys`.
+    *Gannon McGibbon*
 
-    *Ryuta Kamizono*
+*   Cached `columns_hash` fields should be excluded from `ResultSet#column_types`.
 
-*   Automatically guess the inverse associations for STI.
+    PR #34528 addresses the inconsistent behaviour when attribute is defined for an ignored column. The following test
+    was passing for SQLite and MySQL, but failed for PostgreSQL:
 
-    *Yuichiro Kaneko*
+    ```ruby
+    class DeveloperName < ActiveRecord::Type::String
+      def deserialize(value)
+        "Developer: #{value}"
+      end
+    end
 
-*   Ensure `sum` honors `distinct` on `has_many :through` associations.
+    class AttributedDeveloper < ActiveRecord::Base
+      self.table_name = "developers"
 
-    Fixes #16791.
+      attribute :name, DeveloperName.new
 
-    *Aaron Wortham*
+      self.ignored_columns += ["name"]
+    end
 
-*   Add `binary` fixture helper method.
+    developer = AttributedDeveloper.create
+    developer.update_column :name, "name"
 
-    *Atsushi Yoshida*
+    loaded_developer = AttributedDeveloper.where(id: developer.id).select("*").first
+    puts loaded_developer.name # should be "Developer: name" but it's just "name"
+    ```
 
-*   When using `Relation#or`, extract the common conditions and put them before the OR condition.
+    *Dmitry Tsepelev*
 
-    *Maxime Handfield Lapointe*
+*   Make the implicit order column configurable.
 
-*   `Relation#or` now accepts two relations who have different values for
-    `references` only, as `references` can be implicitly called by `where`.
+    When calling ordered finder methods such as `first` or `last` without an
+    explicit order clause, ActiveRecord sorts records by primary key. This can
+    result in unpredictable and surprising behaviour when the primary key is
+    not an auto-incrementing integer, for example when it's a UUID. This change
+    makes it possible to override the column used for implicit ordering such
+    that `first` and `last` will return more predictable results.
 
-    Fixes #29411.
+    Example:
 
-    *Sean Griffin*
+        class Project < ActiveRecord::Base
+          self.implicit_order_column = "created_at"
+        end
 
-*   `ApplicationRecord` is no longer generated when generating models. If you
-    need to generate it, it can be created with `rails g application_record`.
+    *Tekin Suleyman*
 
-    *Lisa Ugray*
+*   Bump minimum PostgreSQL version to 9.3.
 
-*   Fix `COUNT(DISTINCT ...)` with `ORDER BY` and `LIMIT` to keep the existing select list.
+    *Yasuo Honda*
 
-    *Ryuta Kamizono*
+*   Values of enum are frozen, raising an error when attempting to modify them.
+
+    *Emmanuel Byrd*
 
-*   When a `has_one` association is destroyed by `dependent: destroy`,
-    `destroyed_by_association` will now be set to the reflection, matching the
-    behaviour of `has_many` associations.
+*   Move `ActiveRecord::StatementInvalid` SQL to error property and include binds as separate error property.
 
-    *Lisa Ugray*
+    `ActiveRecord::ConnectionAdapters::AbstractAdapter#translate_exception_class` now requires `binds` to be passed as the last argument.
 
-*   Fix `unscoped(where: [columns])` removing the wrong bind values.
+    `ActiveRecord::ConnectionAdapters::AbstractAdapter#translate_exception` now requires `message`, `sql`, and `binds` to be passed as keyword arguments.
 
-    When the `where` is called on a relation after a `or`, unscoping the column of that later `where` removed
-    bind values used by the `or` instead. (possibly other cases too)
+    Subclasses of `ActiveRecord::StatementInvalid` must now provide `sql:` and `binds:` arguments to `super`.
+
+    Example:
 
     ```
-    Post.where(id: 1).or(Post.where(id: 2)).where(foo: 3).unscope(where: :foo).to_sql
-    # Currently:
-    #     SELECT "posts".* FROM "posts" WHERE ("posts"."id" = 2 OR "posts"."id" = 3)
-    # With fix:
-    #     SELECT "posts".* FROM "posts" WHERE ("posts"."id" = 1 OR "posts"."id" = 2)
+    class MySubclassedError < ActiveRecord::StatementInvalid
+      def initialize(message, sql:, binds:)
+        super(message, sql: sql, binds: binds)
+      end
+    end
     ```
 
-    *Maxime Handfield Lapointe*
+    *Gannon McGibbon*
 
-*   Values constructed using multi-parameter assignment will now use the
-    post-type-cast value for rendering in single-field form inputs.
+*   Add an `:if_not_exists` option to `create_table`.
 
-    *Sean Griffin*
+    Example:
 
-*   `Relation#joins` is no longer affected by the target model's
-    `current_scope`, with the exception of `unscoped`.
+        create_table :posts, if_not_exists: true do |t|
+          t.string :title
+        end
+
+    That would execute:
+
+        CREATE TABLE IF NOT EXISTS posts (
+          ...
+        )
+
+    If the table already exists, `if_not_exists: false` (the default) raises an
+    exception whereas `if_not_exists: true` does nothing.
+
+    *fatkodima*, *Stefan Kanev*
 
-    Fixes #29338.
+*   Defining an Enum as a Hash with blank key, or as an Array with a blank value, now raises an `ArgumentError`.
+
+    *Christophe Maximin*
+
+*   Adds support for multiple databases to `rails db:schema:cache:dump` and `rails db:schema:cache:clear`.
+
+    *Gannon McGibbon*
+
+*   `update_columns` now correctly raises `ActiveModel::MissingAttributeError`
+    if the attribute does not exist.
 
     *Sean Griffin*
 
-*   Change sqlite3 boolean serialization to use 1 and 0.
+*   Add support for hash and URL configs in database hash of `ActiveRecord::Base.connected_to`.
 
-    SQLite natively recognizes 1 and 0 as true and false, but does not natively
-    recognize 't' and 'f' as was previously serialized.
+    ````
+    User.connected_to(database: { writing: "postgres://foo" }) do
+      User.create!(name: "Gannon")
+    end
 
-    This change in serialization requires a migration of stored boolean data
-    for SQLite databases, so it's implemented behind a configuration flag
-    whose default false value is deprecated.
+    config = { "adapter" => "sqlite3", "database" => "db/readonly.sqlite3" }
+    User.connected_to(database: { reading: config }) do
+      User.count
+    end
+    ````
 
-    *Lisa Ugray*
+    *Gannon McGibbon*
 
-*   Skip query caching when working with batches of records (`find_each`, `find_in_batches`,
-    `in_batches`).
+*   Support default expression for MySQL.
 
-    Previously, records would be fetched in batches, but all records would be retained in memory
-    until the end of the request or job.
+    MySQL 8.0.13 and higher supports default value to be a function or expression.
 
-    *Eugene Kenny*
+    https://dev.mysql.com/doc/refman/8.0/en/create-table.html
 
-*   Prevent errors raised by `sql.active_record` notification subscribers from being converted into
-    `ActiveRecord::StatementInvalid` exceptions.
+    *Ryuta Kamizono*
 
-    *Dennis Taylor*
+*   Support expression indexes for MySQL.
 
-*   Fix eager loading/preloading association with scope including joins.
+    MySQL 8.0.13 and higher supports functional key parts that index
+    expression values rather than column or column prefix values.
 
-    Fixes #28324.
+    https://dev.mysql.com/doc/refman/8.0/en/create-index.html
 
     *Ryuta Kamizono*
 
-*   Fix transactions to apply state to child transactions.
+*   Fix collection cache key with limit and custom select to avoid ambiguous timestamp column error.
 
-    Previously, if you had a nested transaction and the outer transaction was rolledback, the record from the
-    inner transaction would still be marked as persisted.
+    Fixes #33056.
 
-    This change fixes that by applying the state of the parent transaction to the child transaction when the
-    parent transaction is rolledback. This will correctly mark records from the inner transaction as not persisted.
+    *Federico Martinez*
 
-    *Eileen M. Uchitelle*, *Aaron Patterson*
+*   Add basic API for connection switching to support multiple databases.
 
-*   Deprecate `set_state` method in `TransactionState`.
+    1) Adds a `connects_to` method for models to connect to multiple databases. Example:
 
-    Deprecated the `set_state` method in favor of setting the state via specific methods. If you need to mark the
-    state of the transaction you can now use `rollback!`, `commit!` or `nullify!` instead of
-    `set_state(:rolledback)`, `set_state(:committed)`, or `set_state(nil)`.
+    ```
+    class AnimalsModel < ApplicationRecord
+      self.abstract_class = true
 
-    *Eileen M. Uchitelle*, *Aaron Patterson*
+      connects_to database: { writing: :animals_primary, reading: :animals_replica }
+    end
 
-*   Deprecate delegating to `arel` in `Relation`.
+    class Dog < AnimalsModel
+      # connected to both the animals_primary db for writing and the animals_replica for reading
+    end
+    ```
 
-    *Ryuta Kamizono*
+    2) Adds a `connected_to` block method for switching connection roles or connecting to
+    a database that the model didn't connect to. Connecting to the database in this block is
+    useful when you have another defined connection, for example `slow_replica` that you don't
+    want to connect to by default but need in the console, or a specific code block.
 
-*   Query cache was unavailable when entering the `ActiveRecord::Base.cache` block
-    without being connected.
+    ```
+    ActiveRecord::Base.connected_to(role: :reading) do
+      Dog.first # finds dog from replica connected to AnimalsBase
+      Book.first # doesn't have a reading connection, will raise an error
+    end
+    ```
+
+    ```
+    ActiveRecord::Base.connected_to(database: :slow_replica) do
+      SlowReplicaModel.first # if the db config has a slow_replica configuration this will be used to do the lookup, otherwise this will throw an exception
+    end
+    ```
 
-    *Tsukasa Oishi*
+    *Eileen M. Uchitelle*
 
-*   Previously, when building records using a `has_many :through` association,
-    if the child records were deleted before the parent was saved, they would
-    still be persisted. Now, if child records are deleted before the parent is saved
-    on a `has_many :through` association, the child records will not be persisted.
+*   Enum raises on invalid definition values
 
-    *Tobias Kraze*
+    When defining a Hash enum it can be easy to use `[]` instead of `{}`. This
+    commit checks that only valid definition values are provided, those can
+    be a Hash, an array of Symbols or an array of Strings. Otherwise it
+    raises an `ArgumentError`.
 
-*   Merging two relations representing nested joins no longer transforms the joins of
-    the merged relation into LEFT OUTER JOIN.
+    Fixes #33961
 
-    Example:
+    *Alberto Almagro*
+
+*   Reloading associations now clears the Query Cache like `Persistence#reload` does.
 
     ```
-    Author.joins(:posts).merge(Post.joins(:comments))
-    # Before the change:
-    #=> SELECT ... FROM authors INNER JOIN posts ON ... LEFT OUTER JOIN comments ON...
+    class Post < ActiveRecord::Base
+      has_one :category
+      belongs_to :author
+      has_many :comments
+    end
 
-    # After the change:
-    #=> SELECT ... FROM authors INNER JOIN posts ON ... INNER JOIN comments ON...
+    # Each of the following will now clear the query cache.
+    post.reload_category
+    post.reload_author
+    post.comments.reload
     ```
 
-    *Maxime Handfield Lapointe*
+    *Christophe Maximin*
 
-*   `ActiveRecord::Persistence#touch` does not work well when optimistic locking enabled and
-    `locking_column`, without default value, is null in the database.
+*   Added `index` option for `change_table` migration helpers.
+    With this change you can create indexes while adding new
+    columns into the existing tables.
 
-    *bogdanvlviv*
+    Example:
 
-*   Fix destroying existing object does not work well when optimistic locking enabled and
-    `locking_column` is null in the database.
+        change_table(:languages) do |t|
+          t.string :country_code, index: true
+        end
 
-    *bogdanvlviv*
+    *Mehmet Emin İNAÇ*
 
-*   Use bulk INSERT to insert fixtures for better performance.
+*   Fix `transaction` reverting for migrations.
 
-    *Kir Shatrov*
+    Before: Commands inside a `transaction` in a reverted migration ran uninverted.
+    Now: This change fixes that by reverting commands inside `transaction` block.
 
-*   Prevent creation of bind param if casted value is nil.
+    *fatkodima*, *David Verhasselt*
 
-    *Ryuta Kamizono*
+*   Raise an error instead of scanning the filesystem root when `fixture_path` is blank.
+
+    *Gannon McGibbon*, *Max Albrecht*
 
-*   Deprecate passing arguments and block at the same time to `count` and `sum` in `ActiveRecord::Calculations`.
+*   Allow `ActiveRecord::Base.configurations=` to be set with a symbolized hash.
+
+    *Gannon McGibbon*
+
+*   Don't update counter cache unless the record is actually saved.
+
+    Fixes #31493, #33113, #33117.
 
     *Ryuta Kamizono*
 
-*   Loading model schema from database is now thread-safe.
+*   Deprecate `ActiveRecord::Result#to_hash` in favor of `ActiveRecord::Result#to_a`.
 
-    Fixes #28589.
+    *Gannon McGibbon*, *Kevin Cheng*
 
-    *Vikrant Chaudhary*, *David Abdemoulaie*
+*   SQLite3 adapter supports expression indexes.
 
-*   Add `ActiveRecord::Base#cache_version` to support recyclable cache keys via the new versioned entries
-    in `ActiveSupport::Cache`. This also means that `ActiveRecord::Base#cache_key` will now return a stable key
-    that does not include a timestamp any more.
+    ```
+    create_table :users do |t|
+      t.string :email
+    end
+
+    add_index :users, 'lower(email)', name: 'index_users_on_email', unique: true
+    ```
 
-    NOTE: This feature is turned off by default, and `#cache_key` will still return cache keys with timestamps
-    until you set `ActiveRecord::Base.cache_versioning = true`. That's the setting for all new apps on Rails 5.2+
+    *Gray Kemmey*
 
-    *DHH*
+*   Allow subclasses to redefine autosave callbacks for associated records.
 
-*   Respect `SchemaDumper.ignore_tables` in rake tasks for databases structure dump.
+    Fixes #33305.
 
-    *Rusty Geldmacher*, *Guillermo Iguaran*
+    *Andrey Subbota*
 
-*   Add type caster to `RuntimeReflection#alias_name`.
+*   Bump minimum MySQL version to 5.5.8.
 
-    Fixes #28959.
+    *Yasuo Honda*
 
-    *Jon Moss*
+*   Use MySQL utf8mb4 character set by default.
 
-*   Deprecate `supports_statement_cache?`.
+    `utf8mb4` character set with 4-Byte encoding supports supplementary characters including emoji.
+    The previous default 3-Byte encoding character set `utf8` is not enough to support them.
+
+    *Yasuo Honda*
+
+*   Fix duplicated record creation when using nested attributes with `create_with`.
+
+    *Darwin Wu*
+
+*   Configuration item `config.filter_parameters` could also filter out
+    sensitive values of database columns when calling `#inspect`.
+    We also added `ActiveRecord::Base::filter_attributes`/`=` in order to
+    specify sensitive attributes to specific model.
+
+    ```
+    Rails.application.config.filter_parameters += [:credit_card_number, /phone/]
+    Account.last.inspect # => #<Account id: 123, name: "DHH", credit_card_number: [FILTERED], telephone_number: [FILTERED] ...>
+    SecureAccount.filter_attributes += [:name]
+    SecureAccount.last.inspect # => #<SecureAccount id: 42, name: [FILTERED], credit_card_number: [FILTERED] ...>
+    ```
+
+    *Zhang Kang*, *Yoshiyuki Kinjo*
+
+*   Deprecate `column_name_length`, `table_name_length`, `columns_per_table`,
+    `indexes_per_table`, `columns_per_multicolumn_index`, `sql_query_length`,
+    and `joins_per_query` methods in `DatabaseLimits`.
 
     *Ryuta Kamizono*
 
-*   Raise error `UnknownMigrationVersionError` on the movement of migrations
-    when the current migration does not exist.
+*   `ActiveRecord::Base.configurations` now returns an object.
 
-    *bogdanvlviv*
+    `ActiveRecord::Base.configurations` used to return a hash, but this
+    is an inflexible data model. In order to improve multiple-database
+    handling in Rails, we've changed this to return an object. Some methods
+    are provided to make the object behave hash-like in order to ease the
+    transition process. Since most applications don't manipulate the hash
+    we've decided to add backwards-compatible functionality that will throw
+    a deprecation warning if used, however calling `ActiveRecord::Base.configurations`
+    will use the new version internally and externally.
 
-*   Fix `bin/rails db:forward` first migration.
+    For example, the following `database.yml`:
 
-    *bogdanvlviv*
+    ```
+    development:
+      adapter: sqlite3
+      database: db/development.sqlite3
+    ```
+
+    Used to become a hash:
+
+    ```
+    { "development" => { "adapter" => "sqlite3", "database" => "db/development.sqlite3" } }
+    ```
 
-*   Support Descending Indexes for MySQL.
+    Is now converted into the following object:
+
+    ```
+    #<ActiveRecord::DatabaseConfigurations:0x00007fd1acbdf800 @configurations=[
+      #<ActiveRecord::DatabaseConfigurations::HashConfig:0x00007fd1acbded10 @env_name="development",
+        @spec_name="primary", @config={"adapter"=>"sqlite3", "database"=>"db/development.sqlite3"}>
+      ]
+    ```
+
+    Iterating over the database configurations has also changed. Instead of
+    calling hash methods on the `configurations` hash directly, a new method `configs_for` has
+    been provided that allows you to select the correct configuration. `env_name` and
+    `spec_name` arguments are optional. For example, these return an array of
+    database config objects for the requested environment and a single database config object
+    will be returned for the requested environment and specification name respectively.
+
+    ```
+    ActiveRecord::Base.configurations.configs_for(env_name: "development")
+    ActiveRecord::Base.configurations.configs_for(env_name: "development", spec_name: "primary")
+    ```
+
+    *Eileen M. Uchitelle*, *Aaron Patterson*
 
-    MySQL 8.0.1 and higher supports descending indexes: `DESC` in an index definition is no longer ignored.
-    See https://dev.mysql.com/doc/refman/8.0/en/descending-indexes.html.
+*   Add database configuration to disable advisory locks.
+
+    ```
+    production:
+      adapter: postgresql
+      advisory_locks: false
+    ```
+
+    *Guo Xiang*
+
+*   SQLite3 adapter `alter_table` method restores foreign keys.
+
+    *Yasuo Honda*
+
+*   Allow `:to_table` option to `invert_remove_foreign_key`.
+
+    Example:
+
+       remove_foreign_key :accounts, to_table: :owners
+
+    *Nikolay Epifanov*, *Rich Chen*
+
+*   Add environment & load_config dependency to `bin/rake db:seed` to enable
+    seed load in environments without Rails and custom DB configuration
+
+    *Tobias Bielohlawek*
+
+*   Fix default value for mysql time types with specified precision.
+
+    *Nikolay Kondratyev*
+
+*   Fix `touch` option to behave consistently with `Persistence#touch` method.
 
     *Ryuta Kamizono*
 
-*   Fix inconsistency with changed attributes when overriding Active Record attribute reader.
+*   Migrations raise when duplicate column definition.
 
-    *bogdanvlviv*
+    Fixes #33024.
+
+    *Federico Martinez*
+
+*   Bump minimum SQLite version to 3.8
+
+    *Yasuo Honda*
+
+*   Fix parent record should not get saved with duplicate children records.
+
+    Fixes #32940.
+
+    *Santosh Wadghule*
+
+*   Fix logic on disabling commit callbacks so they are not called unexpectedly when errors occur.
+
+    *Brian Durand*
+
+*   Ensure `Associations::CollectionAssociation#size` and `Associations::CollectionAssociation#empty?`
+    use loaded association ids if present.
+
+    *Graham Turner*
+
+*   Add support to preload associations of polymorphic associations when not all the records have the requested associations.
 
-*   When calling the dynamic fixture accessor method with no arguments, it now returns all fixtures of this type.
-    Previously this method always returned an empty array.
+    *Dana Sherson*
 
-    *Kevin McPhillips*
+*   Add `touch_all` method to `ActiveRecord::Relation`.
+
+    Example:
+
+        Person.where(name: "David").touch_all(time: Time.new(2020, 5, 16, 0, 0, 0))
+
+    *fatkodima*, *duggiefresh*
+
+*   Add `ActiveRecord::Base.base_class?` predicate.
+
+    *Bogdan Gusiev*
+
+*   Add custom prefix/suffix options to `ActiveRecord::Store.store_accessor`.
+
+    *Tan Huynh*, *Yukio Mizuta*
+
+*   Rails 6 requires Ruby 2.5.0 or newer.
+
+    *Jeremy Daer*, *Kasper Timm Hansen*
+
+*   Deprecate `update_attributes`/`!` in favor of `update`/`!`.
+
+    *Eddie Lebow*
+
+*   Add `ActiveRecord::Base.create_or_find_by`/`!` to deal with the SELECT/INSERT race condition in
+    `ActiveRecord::Base.find_or_create_by`/`!` by leaning on unique constraints in the database.
+
+    *DHH*
+
+*   Add `Relation#pick` as short-hand for single-value plucks.
+
+    *DHH*
 
 
-Please check [5-1-stable](https://github.com/rails/rails/blob/5-1-stable/activerecord/CHANGELOG.md) for previous changes.
+Please check [5-2-stable](https://github.com/rails/rails/blob/5-2-stable/activerecord/CHANGELOG.md) for previous changes.