*   Drop MySQL 5.5 support.

    MySQL 5.5 is the only version that does not support datetime with precision,
    which we have supported in the core. Now we support MySQL 5.6.4 or later, which
    is the first version to support datetime with precision.

    *Ryuta Kamizono*

*   Make Active Record asynchronous queries compatible with transactional fixtures.

    Previously transactional fixtures would disable asynchronous queries, because transactional
    fixtures impose all queries use the same connection.

    Now asynchronous queries will use the connection pinned by transactional fixtures, and behave
    much closer to production.

    *Jean Boussier*

*   Deserialize binary data before decrypting

    This ensures that we call `PG::Connection.unescape_bytea` on PostgreSQL before decryption.

    *Donal McBreen*

*   Ensure `ActiveRecord::Encryption.config` is always ready before access.

    Previously, `ActiveRecord::Encryption` configuration was deferred until `ActiveRecord::Base`
    was loaded. Therefore, accessing `ActiveRecord::Encryption.config` properties before
    `ActiveRecord::Base` was loaded would give incorrect results.

    `ActiveRecord::Encryption` now has its own loading hook so that its configuration is set as
    soon as needed.

    When `ActiveRecord::Base` is loaded, even lazily, it in turn triggers the loading of
    `ActiveRecord::Encryption`, thus preserving the original behavior of having its config ready
    before any use of `ActiveRecord::Base`.

    *Maxime Réty*

*   Add `TimeZoneConverter#==` method, so objects will be properly compared by
    their type, scale, limit & precision.

    Address #52699.

    *Ruy Rocha*

*   Add support for SQLite3 full-text-search and other virtual tables.

    Previously, adding sqlite3 virtual tables messed up `schema.rb`.

    Now, virtual tables can safely be added using `create_virtual_table`.

    *Zacharias Knudsen*

*   Support use of alternative database interfaces via the `database_cli` ActiveRecord configuration option.

    ```ruby
    Rails.application.configure do
      config.active_record.database_cli = { postgresql: "pgcli" }
    end
    ```

    *T S Vallender*

*   Add support for dumping table inheritance and native partitioning table definitions for PostgeSQL adapter

    *Justin Talbott*

*   Infer default `:inverse_of` option for `delegated_type` definitions.

    ```ruby
    class Entry < ApplicationRecord
      delegated_type :entryable, types: %w[ Message ]
      # => defaults to inverse_of: :entry
    end
    ```

    *Sean Doyle*

*   Add support for `ActiveRecord::Point` type casts using `Hash` values

    This allows `ActiveRecord::Point` to be cast or serialized from a hash
    with `:x` and `:y` keys of numeric values, mirroring the functionality of
    existing casts for string and array values. Both string and symbol keys are
    supported.

    ```ruby
    class PostgresqlPoint < ActiveRecord::Base
      attribute :x, :point
      attribute :y, :point
      attribute :z, :point
    end

    val = PostgresqlPoint.new({
      x: '(12.34, -43.21)',
      y: [12.34, '-43.21'],
      z: {x: '12.34', y: -43.21}
    })
    ActiveRecord::Point.new(12.32, -43.21) == val.x == val.y == val.z
    ```

    *Stephen Drew*

*   Replace `SQLite3::Database#busy_timeout` with `#busy_handler_timeout=`.

    Provides a non-GVL-blocking, fair retry interval busy handler implementation.

    *Stephen Margheim*

*   SQLite3Adapter: Translate `SQLite3::BusyException` into `ActiveRecord::StatementTimeout`.

    *Matthew Nguyen*

*   Include schema name in `enable_extension` statements in `db/schema.rb`.

    The schema dumper will now include the schema name in generated
    `enable_extension` statements if they differ from the current schema.

    For example, if you have a migration:

    ```ruby
    enable_extension "heroku_ext.pgcrypto"
    enable_extension "pg_stat_statements"
    ```

    then the generated schema dump will also contain:

    ```ruby
    enable_extension "heroku_ext.pgcrypto"
    enable_extension "pg_stat_statements"
    ```

    *Tony Novak*

*   Fix `ActiveRecord::Encryption::EncryptedAttributeType#type` to return
    actual cast type.

    *Vasiliy Ermolovich*

*   SQLite3Adapter: Bulk insert fixtures.

    Previously one insert command was executed for each fixture, now they are
    aggregated in a single bulk insert command.

    *Lázaro Nixon*

*   PostgreSQLAdapter: Allow `disable_extension` to be called with schema-qualified name.

    For parity with `enable_extension`, the `disable_extension` method can be called with a schema-qualified
    name (e.g. `disable_extension "myschema.pgcrypto"`). Note that PostgreSQL's `DROP EXTENSION` does not
    actually take a schema name (unlike `CREATE EXTENSION`), so the resulting SQL statement will only name
    the extension, e.g. `DROP EXTENSION IF EXISTS "pgcrypto"`.

    *Tony Novak*

*   Make `create_schema` / `drop_schema` reversible in migrations.

    Previously, `create_schema` and `drop_schema` were irreversible migration operations.

    *Tony Novak*

*   Support batching using custom columns.

    ```ruby
    Product.in_batches(cursor: [:shop_id, :id]) do |relation|
      # do something with relation
    end
    ```

    *fatkodima*

*   Use SQLite `IMMEDIATE` transactions when possible.

    Transactions run against the SQLite3 adapter default to IMMEDIATE mode to improve concurrency support and avoid busy exceptions.

    *Stephen Margheim*

*   Raise specific exception when a connection is not defined.

    The new `ConnectionNotDefined` exception provides connection name, shard and role accessors indicating the details of the connection that was requested.

    *Hana Harencarova*, *Matthew Draper*

*   Delete the deprecated constant `ActiveRecord::ImmutableRelation`.

    *Xavier Noria*

*   Fix duplicate callback execution when child autosaves parent with `has_one` and `belongs_to`.

    Before, persisting a new child record with a new associated parent record would run `before_validation`,
    `after_validation`, `before_save` and `after_save` callbacks twice.

    Now, these callbacks are only executed once as expected.

    *Joshua Young*

*   `ActiveRecord::Encryption::Encryptor` now supports a `:compressor` option to customize the compression algorithm used.

    ```ruby
    module ZstdCompressor
      def self.deflate(data)
        Zstd.compress(data)
      end

      def self.inflate(data)
        Zstd.decompress(data)
      end
    end

    class User
      encrypts :name, compressor: ZstdCompressor
    end
    ```

    You disable compression by passing `compress: false`.

    ```ruby
    class User
      encrypts :name, compress: false
    end
    ```

    *heka1024*

*   Add condensed `#inspect` for `ConnectionPool`, `AbstractAdapter`, and
    `DatabaseConfig`.

    *Hartley McGuire*

*   Add `.shard_keys`, `.sharded?`, & `.connected_to_all_shards` methods.

    ```ruby
    class ShardedBase < ActiveRecord::Base
        self.abstract_class = true

        connects_to shards: {
          shard_one: { writing: :shard_one },
          shard_two: { writing: :shard_two }
        }
    end

    class ShardedModel < ShardedBase
    end

    ShardedModel.shard_keys => [:shard_one, :shard_two]
    ShardedModel.sharded? => true
    ShardedBase.connected_to_all_shards { ShardedModel.current_shard } => [:shard_one, :shard_two]
    ```

    *Nony Dutton*

*   Add a `filter` option to `in_order_of` to prioritize certain values in the sorting without filtering the results
    by these values.

    *Igor Depolli*

*   Fix an issue where the IDs reader method did not return expected results
    for preloaded associations in models using composite primary keys.

    *Jay Ang*

*   Allow to configure `strict_loading_mode` globally or within a model.

    Defaults to `:all`, can be changed to `:n_plus_one_only`.

    *Garen Torikian*

*   Add `ActiveRecord::Relation#readonly?`.

    Reflects if the relation has been marked as readonly.

    *Theodor Tonum*

*   Improve `ActiveRecord::Store` to raise a descriptive exception if the column is not either
    structured (e.g., PostgreSQL +hstore+/+json+, or MySQL +json+) or declared serializable via
    `ActiveRecord.store`.

    Previously, a `NoMethodError` would be raised when the accessor was read or written:

        NoMethodError: undefined method `accessor' for an instance of ActiveRecord::Type::Text

    Now, a descriptive `ConfigurationError` is raised:

        ActiveRecord::ConfigurationError: the column 'metadata' has not been configured as a store.
          Please make sure the column is declared serializable via 'ActiveRecord.store' or, if your
          database supports it, use a structured column type like hstore or json.

    *Mike Dalessio*

*   Fix inference of association model on nested models with the same demodularized name.

    E.g. with the following setup:

    ```ruby
    class Nested::Post < ApplicationRecord
      has_one :post, through: :other
    end
    ```

    Before, `#post` would infer the model as `Nested::Post`, but now it correctly infers `Post`.

    *Joshua Young*

*   Add public method for checking if a table is ignored by the schema cache.

    Previously, an application would need to reimplement `ignored_table?` from the schema cache class to check if a table was set to be ignored. This adds a public method to support this and updates the schema cache to use that directly.

    ```ruby
    ActiveRecord.schema_cache_ignored_tables = ["developers"]
    ActiveRecord.schema_cache_ignored_table?("developers")
    => true
    ```

    *Eileen M. Uchitelle*

Please check [7-2-stable](https://github.com/rails/rails/blob/7-2-stable/activerecord/CHANGELOG.md) for previous changes.